From “squatters” to citizens? Slum dwellers, developers, land sharing and power in Phnom Penh, Cambodia

PDF

Download

Open document

Transcript (if available)

Content

FROM “SQUATTERS” TO CITIZENS?
SLUM DWELLERS, DEVELOPERS, LAND SHARING AND POWER
IN PHNOM PENH, CAMBODIA

by
Paul Ewoud Rabé

A Dissertation Presented to the
FACULTY OF THE USC SCHOOL OF POLICY, PLANNING,
AND DEVELOPMENT
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF POLICY, PLANNING AND DEVELOPMENT

December 2009

Copyright 2009 Paul Ewoud Rabé

ii

EPIGRAPH
“They’re knocking our life and times away!” said an elderly Mancunian. We stood
together gazing over a wilderness on which still another vast slum had been razed, and he spoke
in grief. A kind of culture unlikely to rise again had gone in the rubble, and he knew it. But most
of the young who left such places had no such regrets: the old ways their fathers had accepted had
long grown insupportable; better by far the ‘cliff’ dwellings of modern Manchester.

From: Robert Roberts,
The Classic Slum: Salford Life in the First Quarter of the Century

iii

DEDICATION
I dedicate this study to the current and former inhabitants of Borei Keila, Dey Krahom,
Railway A and Railway B—the four case studies of this dissertation—whose settlements went
through profound changes during the time period of this study (2003 to 2009), and whose lives
were uprooted in the name of “development”.

iv

ACKNOWLEDGMENTS
My gratitude is due, first and foremost, to Professor Eric J. Heikkila, the Chair of my
Doctoral Committee, for his wise counsel, always salient technical advice, enduring patience and
occasional prodding—all of which were instrumental in helping me to complete this dissertation.
I also want to express special thanks to my two research assistants and former colleagues,
Chap Samoeun and Mok Chanly, for their tireless translation work during the structured interview
process as well as their professional research assistance and good company during the many
hours and days we spent together in Borei Keila, Dey Krahom, the Railway settlements, and
many relocation sites outside Phnom Penh.
Last but certainly not least, my heartfelt thanks are due to my family and many good
friends—across several continents—who gave me the critical moral, material and editorial
support without which I would not have been able to get through the many years required to
undertake this study.

v

TABLE OF CONTENTS
EPIGRAPH ...................................................................................................................................... ii
DEDICATION ................................................................................................................................ iii
ACKNOWLEDGMENTS .............................................................................................................. iv
LIST OF TABLES ......................................................................................................................... vii
LIST OF FIGURES ........................................................................................................................ ix
LIST OF TEXT BOXES ................................................................................................................. xi
ABBREVIATIONS ....................................................................................................................... xii
ABSTRACT ................................................................................................................................... xv
NOTE ON CURRENCIES .......................................................................................................... xvii
CHAPTER 1: INTRODUCTION .................................................................................................... 1
1.1 Announcing A Slum Upgrading Policy ................................................................. 1
1.2 Physical Upgrading As A Metaphor For Empowerment ....................................... 4
1.3 Assessing Power And “Citizenship” ...................................................................... 9
1.4 Structure Of The Dissertation .............................................................................. 14

CHAPTER 2: HISTORICAL CONTEXT ..................................................................................... 17
2.1 Overview of Chapter 2 ......................................................................................... 17
2.2 Historical Images of the Khmers ......................................................................... 18
2.3 Principal Themes in the Literature ....................................................................... 22
2.4 Pre-Angkorian and Angkorian Cambodia ........................................................... 27
2.5 A New Khmer Capital at Phnom Penh ................................................................ 40
2.6 Phnom Penh under the French Protectorate ......................................................... 43
2.7 Post-Independence and the “Sangkum” Period .................................................... 53
2.8 The Democratic Kampuchea Regime .................................................................. 66
2.9 The People’s Republic of Kampuchea Period ..................................................... 75
2.10 The Paris Peace Treaty and the Transition from War to Peace ........................... 84
2.11 A Sharply Contested Urban Space ....................................................................... 97

vi

CHAPTER 3: THE CASE STUDY SETTLEMENTS ................................................................ 111
3.1 Overview of Chapter 3 ....................................................................................... 111
3.2 A “Housing Breakthrough” for Phnom Penh’s Poor ......................................... 112
3.3 Announcing Four Land Sharing Projects ........................................................... 114
3.4 The Political Backdrop to Land Sharing ............................................................ 116
3.5 The Economic Backdrop to Land Sharing ......................................................... 129
3.6 The Case Study Settlement of Borei Keila ........................................................ 149
3.7 The Case Study Settlement of Dey Krahom ...................................................... 167
3.8 The Case Study Settlements of Railway A and Railway B ............................... 182

CHAPTER 4: THEORETICAL FRAMEWORK ........................................................................ 196
4.1 Overview of Chapter 4 ....................................................................................... 196
4.2 Slums of Despair ................................................................................................ 198
4.3 Slums of Hope ................................................................................................... 217
4.4 A Synthesis View: Foucault’s Concept of Capillary Power .............................. 238
4.5 The Slum as Sphere of Force Relations ............................................................. 254

CHAPTER 5: RESEARCH DESIGN AND METHODOLOGY ................................................ 260
5.1 Overview of Chapter 5 ....................................................................................... 260
5.2 The EVLN Model in Theoretical Perspective ................................................... 261
5.3 Presenting an Adapted EVLN Model ................................................................ 270
5.4 Theoretical Propositions .................................................................................... 285
5.5 Data Collection .................................................................................................. 292

CHAPTER 6: DATA ANALYSIS .............................................................................................. 317
6.1 Overview of Chapter 6 ....................................................................................... 317
6.2 General Profile of Respondents ......................................................................... 318
6.3 A Common Measure of Relative Dissatisfaction ............................................... 333
6.4 EVLN Measures in the ‘Independent Assessment’ ........................................... 336
6.5 EVLN Measures in the Decision-Tree Model ................................................... 361
6.6 Are Former Residents “Tiebout Exiters”? ......................................................... 371

CHAPTER 7: CONCLUSION .................................................................................................... 391
7.1 Overview of Chapter 7 ....................................................................................... 391
7.2 The Abandonment of the Slum Upgrading Campaign....................................... 393
7.3 “Citizenship as Status”, Only for a Select Few .................................................. 397
7.4 “Citizenship as Practice” Unfulfilled ................................................................. 405
7.5 Power From Above, Power From Below ........................................................... 414
7.6 Assessing Six Years of “Land Sharing” in Phnom Penh ................................... 425

GLOSSARY ................................................................................................................................ 433
BIBLIOGRAPHY ........................................................................................................................ 435
APPENDIX A: AERIAL PHOTOGRAPHS OF THE CASE STUDY SETTLEMENTS .......... 452
APPENDIX B: QUESTIONNAIRE FOR CURRENT RESIDENTS ......................................... 456
APPENDIX C: QUESTIONNAIRE FOR FORMER RESIDENTS ........................................... 471

vii

LIST OF TABLES
Table 1: Housing rights groups’ assumptions about anticipated benefits for slum
dwellers as a result of the May 2003 slum upgrading announcement ............................... 6
Table 2: Official population and area statistics of the four land sharing
settlements ..................................................................................................................... 115
Table 3: Overview of the Borei Keila land sharing agreement .................................................... 155
Table 4: A rudimentary classification of Foucauldian forms of structure and
agency power ................................................................................................................. 245
Table 5: Hypothesized relationships between determinants responses to
dissatisfaction ................................................................................................................ 291
Table 6: Number of former and current residents interviewed by settlement .............................. 295
Table 7: Sub-groups of respondents and their assumed EVLN responses ................................... 303
Table 8: Total number of persons interviewed by sub-group, category and
settlement ....................................................................................................................... 311
Table 9: Rationale for Respondents’ Perceptions of Tenure Security or Insecurity .................... 324
Table 10: Rationale for respondents’ perceptions of tenure security or insecurity ...................... 326
Table 11: Component items of relative dissatisfaction and results for each item ........................ 336
Table 12: Frequencies of dominant EVLN strategies for dissatisfied respondents ..................... 348
Table 13: Testing EVLN hypotheses by population sub-group ................................................... 349
Table 14: Frequencies of dominant EVLN strategies as measured by three
indicators of the independent assessment of EVLN ................................................... 351
Table 15: Frequency results for component items of Indicator 1 ................................................ 352
Table 16: Frequency results for component items of Indicator 2 ................................................ 354

viii

Table 17: Respondents’ feelings about the redevelopment plans in their
settlements................................................................................................................... 355
Table 18: Frequency results for component items of Indicator 3 ................................................ 359
Table 19: Test of the EVLN Model on Dissatisfied Current Respondents .................................. 370
Table 20: Test of the EVLN model on former respondents ......................................................... 382
Table 21: Selected domination and covert resistance tactics in the case study
settlements................................................................................................................... 420

ix

LIST OF FIGURES
Figure 1: Railway B settlement shows off its social land concession award ................................... 9
Figure 2: Illustration of the central research question .................................................................... 14
Figure 3: Illustration of dissertation topics by chapter .................................................................. 14
Figure 4: The Fifth Anniversary Celebration of UPDF ............................................................... 113
Figure 5: Artist’s impression of the Borei Keila community buildings ....................................... 122
Figure 6: Borei Keila, from athletes’ village to crowded slum .................................................... 151
Figure 7: Slum conditions in Borei Keila .................................................................................... 153
Figure 8: Map of Borei Keila, showing the location of new community buildings ..................... 157
Figure 9: The new community housing in Borei Keila ................................................................ 165
Figure 10: The settlement of Dey Krahom .................................................................................. 169
Figure 11: The two faces of Dey Krahom.................................................................................... 176
Figure 12: The final clearance of the Dey Krahom site ............................................................... 179
Figure 13: The Dey Krahom relocation site................................................................................. 180
Figure 14: The settlement of Railway B ...................................................................................... 183
Figure 15: The settlement of Railway A ...................................................................................... 184
Figure 16: The Tang Krosang relocation site for residents of the Railway settlements .............. 189
Figure 17: The basis for agency: “there is no power without resistance” .................................... 251
Figure 18: Imagining the Foucauldian sphere of force relations ................................................. 253
Figure 19: Prototypical response types in the original EVLN model (Lyons and
Lowery 1986 and 1989) ............................................................................................. 271

x

Figure 20: EVLN responses in my adapted Lyons and Lowery model ....................................... 273
Figure 21: Selected anticipated EVLN responses by slum dwellers in Phnom Penh .................. 285
Figure 22: Decision tree model of responses to dissatisfaction among slum dwellers ................ 289
Figure 23: Key indicators in the independent approach to assessing EVLN ............................... 338
Figure 24: Completed decision-tree for current residents ............................................................ 367
Figure 25: Overview of reasons for leaving by former residents ................................................. 374
Figure 26: Completed decision-tree for former residents ............................................................ 380
Figure 27: Overview of Tiebout exit by measure and by former settlement ............................... 386
Figure 28: Aerial photo of Phnom Penh, showing the location of the four case
study settlements ........................................................................................................ 452
Figure 29: Orthophoto of Borei Keila settlement ........................................................................ 453
Figure 30: Orthophoto of Dey Krahom settlement ...................................................................... 454
Figure 31: Orthophoto of Railway A and B settlements .............................................................. 455

xi

LIST OF TEXT BOXES
Box 1: Public Edict No. 1, Phnom Penh Cleanliness and Beautification Committee, 1958 .......... 62
Box 2: Public Edict No. 2, Phnom Penh Cleanliness and Beautification Committee, 1958 .......... 62
Box 3: Public Edict No. 3, Phnom Penh Cleanliness and Beautification Committee, 1958 .......... 62
Box 4: Is participation in community savings schemes an expression of loyalty or voice? ........ 341
Box 5: The thin line between voluntary and involuntary exit ...................................................... 384

xii

ABBREVIATIONS
1

ACHR Asian Coalition for Housing Rights (Thailand)
ASEAN Association of Southeast Asian Nations
BABSEA Bridges Across Borders South-East Asia
COHRE Center on Housing Rights and Evictions
CGDK Coalition Government of Democratic Kampuchea (political alliance against
the People’s Republic of Kampuchea regime)
CPK Communist Party of Kampuchea
CPP Cambodian People’s Party
CR Current residents
DK Democratic Kampuchea (official name of Cambodia under the Khmers
Rouges regime, 1975-1979)
ESCR United Nations Committee on Economic, Social and Cultural Rights
EWMI East-West Management Institute, Inc.
FR Former residents
FUNCINPEC Front Uni National pour un Cambodge Indépendant, Neutre, Pacifique et
Coopératif (United National Front for an Independent, Neutral, Peaceful and
Cooperative Cambodia)
FUNSK Front Uni National pour le Salut (United Front for National Salvation)

1
Abbreviations of public and private institutions and political organizations refer to Cambodian terms,
unless stated otherwise.

xiii

GANEFO Games of the New Emerging Forces (non-aligned nations’ sporting event,
held in Phnom Penh in 1966)
KPNLF Khmer People’s National Liberation Front
LICADHO The Cambodian League for the Promotion and Defense of Human Rights
LMAP Land Management and Administration Project
MLMUPC Ministry of Land Management, Urban Planning and Construction
MoEYS Ministry of Education, Youth and Sports
MPP Municipality of Phnom Penh
NGO Non-Governmental Organization
NIS National Institute of Statistics, Ministry of Planning
NPRS National Poverty Reduction Strategy
NSDF National Slum Dwellers Federation (India)
OCIC Overseas Cambodian Investment Corporation
OHCHR Office of the High Commissioner for Human Rights (United Nations)
OPL Overall Poverty Line
PILAP Public Interest Legal Advocacy Project
PPUPRP Phnom Penh Urban Poverty Reduction Project of UN-Habitat and UNDP
(1996–2004)
PRK People’s Republic of Kampuchea (Vietnamese occupation period: 1979-
1989)
RGC Royal Government of Cambodia
RUFA Royal University of Fine Arts, Phnom Penh
RUPP Royal University of Phnom Penh

xiv

SDI Slum/Shack Dwellers International (international network of housing rights
organizations)
SLC Social land concession
SNC Supreme National Council of Cambodia (governing body during political
transition in the early 1990s)
SPARC Society for the Promotion of Area Resource Centers (Indian NGO)
SRSG Special Representative of the (U.N.) Secretary General for Human Rights in
Cambodia
SUPF Solidarity for the Urban Poor Federation (Cambodian people’s organization)
UNDP United Nations Development Program
UN-Habitat United Nations Center for Human Settlements (formerly UNCHS)
UNHCR United Nations High Commission for Refugees
UNTAC United Nations Transitional Authority for Cambodia (UN caretaker
authority, 1991-3)
UPDF Urban Poor Development Fund
URC Urban Resource Center
USAID United States Agency for International Development
USG Urban Sector Group

xv

ABSTRACT
This study examines the struggle of residents of four slums in the heart of the Cambodian
capital, Phnom Penh, to remain in the city in the face of commercial redevelopment pressures
during the first decade of the 21
st
century.
In 2003, following a decade of forced evictions of the poor from the city center, the
Royal Government of Cambodia broke with the past and authorized the redevelopment of the four
city center slums as pilots for a major slum upgrading campaign, through the technique of land
sharing.
The study assesses the extent to which the redevelopment projects succeeded in providing
slum dwellers with the basic legal rights promised to them as part of upgrading, which can be
regarded as residents’ core citizenship contract with the state—a “citizenship of status”. Second,
the study examines the extent to which slum residents were involved in shaping the
redevelopment plans, which can be regarded as a reflection of their “power” and agency—a
“citizenship of practice”. The assessment of “citizenship of practice” is based on two theoretical
approaches. The first is guided by the propositions of capillary power by Michel Foucault (1978
et al.), for whom power is diffuse and omnipresent, with all actors in the role of subjects as well
as agents. The second expands on the typology of options of exit, voice and loyalty developed by
the economist Albert Hirschman (1970), and later refined by the political scientists William
Lyons and David Lowery (1986 and 1989).
The study concludes that the slum dwellers’ “battle for living space” took the form of an
elaborate charade. Commercial developers, with the backing of the Municipality and local

xvi

authorities, dressed up their slum redevelopment projects with an appearance of inevitability to
convince residents that they had the legal right and technical ability to develop the settlements as
they saw fit. In response, residents presented an appearance of compliance, even as—below the
surface—they were engaged in manifold resistance tactics, not necessarily to oppose the projects,
but at least to take from them what little they could for themselves.

xvii

NOTE ON CURRENCIES
Currencies cited in this dissertation, unless specified otherwise, are Cambodian riel (riel)
and U.S. dollars ($). The U.S. dollar is the de facto principal currency used in Cambodia,
particularly in Phnom Penh, and particularly for larger transactions.
The exchange rate used in this dissertation is: $1.00 = 4,000 riel. This represents an
average exchange rate during the focal period of this study (2003 to 2009).

1

CHAPTER 1: INTRODUCTION
1.1 Announcing a Slum Upgrading Policy
In May 2003 an apparent breakthrough occurred for slum dwellers in Cambodia’s capital
city, Phnom Penh. After more than a decade of forced evictions and poorly planned relocations
of urban poor residents to barren fields outside the city, Prime Minister Hun Sen publicly
announced a campaign to “upgrade 100 poor settlements in one year”. Henceforth, slum
residents would no longer be forcibly displaced from their inner city settlements—where most
had been living for decades—but they would be regarded instead as legitimate land occupants.
Over the course of the next five years, almost all of the city’s more than 500 poor settlements
would ostensibly be improved on site and their residents would receive land titles.
The Council of Ministers of Cambodia announced that there would be four pilot projects
to implement what it called the new “social land concession” policy in Phnom Penh. The four
pilots were to be “land sharing” projects, in the city center settlements of Borei Keila, Dey
Krahom, Railway A and Railway B. As a form of redevelopment, land sharing has attracted
interest among housing professionals for its ability to accommodate development without
displacing existing residents, mainly by agreeing to divide (“share”) a plot of disputed land, so
that a developer is given the right to build on one portion of the site and land occupants are re-
housed on another portion, with a promise of secure tenure on their new plots. The land sharing
technique has been carried out relatively successfully—albeit on a small scale, and on a relatively
ad hoc basis—in several urban settlements in Bangkok and several other Asian cities in the
1980s. In recent years, land sharing has reappeared in a more planned fashion in several

2

government housing programs in Southeast Asia as a way to help low-income communities stay
in city centers
2
.
1.1.1 A “Battle for Living Space”
The official “amnesty” for urban slum dwellers in Phnom Penh was remarkable for the
fact that it occurred in the midst of a strong recovery in the city. After decades of civil war and
political strife in Cambodia, Phnom Penh went through a period of reconstruction between 1993
and 2000, focused on the rebuilding of infrastructure and basic services. This was followed by a
“burst of major urban development projects”, rapid growth and urbanization (World Bank 2009,
104-5). The city center became the focus of large-scale land development and real estate
activity.
Urban expansion in the capital had unleashed a “battle for living space”
3
. The increased
land development activities directly and indirectly threatened the homes and livelihoods of more
than a quarter of the city’s estimated population of 1.2 million who were living in 569 recorded
“urban poor settlements”
4
, i.e. those areas characterized by precarious conditions. Most of these
residents enjoyed possession rights according to the Cambodian Land Law of 2001. In spite of
their legal rights, however, between 1990 and 2003 an estimated 61,500 people were forcibly

2
One of the largest government schemes to include the land sharing technique is the Baan Mankong
program in Thailand, which is helping to upgrade thousands of informal settlements across that country.
3
Asian Coalition for Housing Rights, “Cause for Hope in the Battle for Living Space in Asia’s Cities”,
Housing by People in Asia (Bangkok), Newsletter No. 15 (October 2003): 1.
4
The figure is based on a May 2003 survey conducted by the Solidarity for the Urban Poor Federation
(SUPF), as quoted in UPDF, “2003 Survey of Poor Settlements”, UPDF Newsletter (Phnom Penh, May
2003): 12.

3

removed from city slums in order to accommodate the urban expansion and new infill
development
5
.
Evicted residents included those who had moved to Phnom Penh in various waves of
migration following the end of the Democratic Kampuchea (Khmers Rouges) regime in 1979.
After the Paris Peace Accords of 1991 and national elections in 1993, which established the
foundations for a political settlement and the present constitutional monarchy, an additional
stream of rural migrants had been attracted to the city as a result of a combination of “push” and
“pull” factors: employment prospects in the rural economy of Cambodia were declining, while
the economy of the capital was expanding
6
.
The series of evictions in the 1990s and early 2000s sowed the seeds of future
confrontations. Most redevelopment projects in Phnom Penh took place without the provision for
decent housing alternatives for low-income people affected, thus setting the stage for a future
cycle of land occupation and further evictions. Past experience with relocations in Phnom Penh
(and further afield) was instructive: it demonstrated that evictions from city centers, where the
urban poor had livelihood opportunities, to marginal land on the urban periphery, where they
have few such opportunities, were bound to increase poverty. Moreover, experience showed that
a large proportion of evicted families would attempt to return to the city center in search of a
place to live and make ends meet. By promising to upgrade and improve existing urban poor
settlements, which represented an important reservoir of low-income housing in the city center,
the slum upgrading campaign therefore raised the prospect of breaking this vicious cycle.

5
COHRE, BABSEA and Licadho, Preparatory Document for the Universal Periodic Review Assessment
for the Office of the High Commissioner for Human Rights (Phnom Penh: COHRE, BABSEA and Licadho,
2009), 5.
6
Department of Sociology, Royal University of Phnom Penh. Livelihoods of the Urban Poor in Phnom
Penh, Working Paper No. 3 (Phnom Penh: RUPP, 2005).

4

1.2 Physical Upgrading as a Metaphor for Empowerment
In development policy and practice, the goals of secure tenure and slum improvement
have traditionally been considered narrowly as “housing problems”, requiring technical solutions.
In this dissertation, however, I consider the underlying problem of slums
7
to be political. The
battle for living space must be understood as a “struggle for enfranchisement”, for access to
housing, to secure tenure, to livelihoods, to public services, and ultimately, to public recognition.
Arjun Appadurai (2004) has characterized slums as the “most public drama of
disenfranchisement”, with the politics of housing as “the single-most critical site of a politics of
citizenship” (Appadurai, 72). The battle for living space is not only a struggle for a physical
space but also a political space—a place in the city, as equal “citizens”. What is ultimately at
stake in the competition for land thrown up by urban redevelopment is recognition and
empowerment, with the achievement of secure tenure and public services marking the residents’
evolution from mere “squatters” to fully-fledged citizens.
In Phnom Penh the political significance of the new slum upgrading campaign was
indicated by the words of one of the city’s Vice-Governors at the time:

7
A note on terminology: In this dissertation I use the relatively neutral, technical terms “slum” and “urban
poor settlement” to refer to the residential quarters of the poor in Phnom Penh. Use of the term “slum” is
consistent with the technical definition of the United Nations in its Millennium Development Goals
(MDGs) for urban poor areas. According to this definition, “slums” are areas with a high incidence of poor
sanitation; unsafe water supply; poor structural quality of housing; insecure tenure; and “insufficient living
area”. In the Khmer language, “slums” are known principally as “poor communities” (sahakum neak krai
kroh). I have attempted to avoid usage of two traditional descriptions of slums in the literature, as
“informal” or “illegal” settlements, as these terms do not accurately describe the legal or factual situation of
most of the residents of these areas in the Phnom Penh context.

5

“The city does not only belong to the rich. It belongs to all of us, so we should all
be involved in improving it. Now we have a lot of work to do. We have to sit
down and set concrete plans for this upgrading program together”
8
.
1.2.1 A Breakthrough for Phnom Penh’s Poor?
The pilot land sharing projects in Phnom Penh appeared to signal two progressive new
developments in the government’s approach to the (urban) poor. First, they seemed to signal that
the government finally acknowledged the land rights of many of the residents of urban slums. In
a report to the United Nations Committee on Economic, Social and Cultural Rights (ESCR), the
government indicated that the land sharing projects would help the “communities to build houses
on their legally owned land”
9
. The second hopeful development was that the government
appeared interested to pilot a new mechanism to deliver social housing, through the private
sector. In the same report to the ESCR, the government indicated, with reference to the land
sharing projects, that “the Royal Government of Cambodia is not able to build houses for its
people. But it allows private companies and the people to build houses by respecting the
construction law”
10
.
Housing rights groups and international donors interpreted the new slum upgrading
campaign, including the land sharing pilots, as representing a “breakthrough for the urban
poor”
11
. They assumed that through community-led upgrading efforts, supported by a coalition of
local and international community-based organizations, non-governmental organizations and
donors, slum dwellers would achieve better living conditions, a new “citizenship” contract in

8
Vice-Governor of Phnom Penh, Chev Kim Heng, quoted in UPDF and SUPF, Community News (Phnom
Penh), Issue No. 2 (June 2003): 1.
9
Royal Government of Cambodia, Cambodia State Party Report to the United Nations Committee on
Economic, Social and Cultural Rights (Phnom Penh, 10 November 2008): para. 537.
10
Ibid., para. 538.
11
UPDF and SUPF, “Housing Breakthrough for Phnom Penh’s Poor”, Community News (Phnom Penh),
Issue No. 2 (June 2003): 1.

6

their relations with the state, and greater agency through collective power. The presumed benefits
(both technical and political) of the new policy for the urban poor are summarized in Table 1.

Situation of slum dwellers,
before May 2003
Assumed situation of slum dwellers,
after May 2003
Official discourse
toward slum
dwellers
“Squatters” and “anarchists” “Citizens”
Assumptions about
slum dwellers’
power
Individual “powerlessness” Individual and collective empowerment
Official policy
toward slums
Forced eviction and relocation On-site upgrading with full basic services,
or relocation based on fair and just
compensation
Legal status of
residents
(“Citizenship as
Status”)
Insecure tenure and absent public
services
Regularized legal tenure and access to
public services (water and electricity)
Residents’ housing
strategies
(“Citizenship as
Practice”, through
voice)
Traditional individual coping
strategies, including: patronage;
withdrawal; flight; speculation
Individual and collective action, including:
community organizing; community
development activities, including saving
and surveying; building links with local
and international organizations
Table 1: Housing rights groups’ assumptions about anticipated benefits for slum dwellers as a result
of the May 2003 slum upgrading announcement

First, the state’s policy towards existing informal settlements would no longer revolve
around forced evictions and relocations, but towards upgrading on site, including the provision of
formal basic services. Relocations of slum dwellers would only take place in cases where
informal settlements stood in the way of planned civic projects, such as roads, parks and drainage
lines. In those cases, the Prime Minister pledged that the government would help identify
suitable relocation sites in areas nearby job opportunities.
Second, at the level of official discourse, slum dwellers would no longer be labeled as
“squatters” or “anarchists”, as they were known in the past, but as “citizens” who had a proper

7

place in the city. The change in official discourse would have an immediate effect in improving
the self-confidence and political power of tens of thousands of urban poor households.
Third, in terms of legal status, the new policy indicated that slum dwellers would no
longer be considered as illegal land occupants, but as legal possessors of land they had occupied,
in many cases since the end of the Democratic Kampuchea (Khmers Rouges) period in 1979.
The upgrading schemes could perhaps even be construed as an implicit recognition of the
possession right of informal land occupants as enshrined in the Land Law 2001—although this
was never made explicit. The Prime Minister’s pledge to pursue relocation only in case of
conflicts with planned civic projects appeared to respect Article 44 of the Cambodian
Constitution and Article 5 of the Land Law 2001, which require fair and just compensation in
advance in case of expropriation for projects in the public interest
12
.
Fourth, the upgrading policy would mark a radical shift in slum dwellers’ housing
strategies. New forms of individual and collective action would replace the uncoordinated
individual strategies of the past, which include such traditional passive resistance strategies as
flight, withdrawal, and involvement in patronage structures, or such active strategies as
speculation. In the new upgrading context, local community based organizations such as
Solidarity for the Urban Poor Federation (SUPF) and the Urban Poor Development Fund (UPDF)
would help slum residents to organize and carry out a wide range of development activities as
collectives, including saving, surveying, and putting in basic infrastructure. They would help
slum communities to “scale up” settlement improvement so that it becomes a city-wide process.
Moreover, UPDF would help slum dwellers to build linkages with local and international

12
Article 44 of the Constitution holds that “The right to confiscate possessions from any person shall be
exercised only in the public interest as provided for under law and shall required fair and just compensation
in advance”.

8

organizations for “joint projects and mutual learning”
13
, through the international network of
housing rights organizations, Slum/Shack Dwellers International (SDI).
Finally, perhaps the most important assumption of housing rights organizations and
donors was that slum dwellers in Phnom Penh would achieve new power to tackle pressing
problems in the slums, including the poor slum environmental conditions and services. This
power would be expressed through both individual and collective structures rather than
uncoordinated individual initiatives, and would be based on a new kind of activism which, to cite
Arjun Appadurai, enables poor communities “to engage in partnerships with more powerful
agencies—urban, regional, national and multilateral—that purport to be concerned with poverty
and with citizenship” (Appadurai 2001, 42).
1.2.2 Or a Continuation of Past Practices?
Amid all the optimism surrounding the central government’s support for the new slum
upgrading campaign, there were plenty of reasons to doubt that the campaign would mark a new
chapter in the empowerment of Phnom Penh’s slum dwellers. First, many of the same structural
forces that led to exclusion of the urban poor and the eviction of slum dwellers during the 1990s
remained in place after May 2003. The underlying economic rationale for evictions of the poor
persisted: as a result of the continued economic boom in Phnom Penh, the pressure on lands
occupied by the urban poor remained strong, and even seemed to increase due to growing foreign
investment and the arrival of foreign developers in the capital.

13
UPDF and SUPF, “Housing Breakthrough for Phnom Penh’s Poor”: 1.

9

Second, despite the increase in support
from community organizations and international
non-governmental organizations during the
1990s, and despite the growth of savings
initiatives and other community activities, most
slum populations appeared too fragmented to
mount a successful collective effort to resist the
eviction forces of private developers and the
authorities.
Third, many commentators questioned
whether Prime Minister Hun Sen’s slum
upgrading announcement marked a true policy
shift or was merely a strategic political
calculation ahead of the July 2003 national elections—an effort by the ruling Cambodian
People’s Party (CPP) to win the political support of voters in the hundreds of slum settlements in
Phnom Penh.
But in the heady months following the Prime Minister’s slum upgrading announcement,
most slum residents in Phnom Penh chose to overlook these doubts. In the four land sharing pilot
areas, photographs of the Prime Minister inspecting their communities’ land sharing plans
appeared, as a new mascot, all over the settlements and in many a home (see Figure 1).
1.3 Assessing Power and “Citizenship”
This study examines the degree to which the May 2003 slum upgrading policy can indeed
be considered a true “breakthrough” for Phnom Penh’s slum dwellers. It does so by examining

A framed photograph of the Prime Minister
inspecting the community land sharing plan in
2003 prominently adorned the entrance to
Railway B settlement (Author’s photo, April 2006)
Figure 1: Railway B settlement shows off its
social land concession award

10

different forms of power and citizenship. First, I assess the extent to which the new policy
delivered to slum dwellers the promised basic legal rights as part of slum upgrading and land
sharing, which I argue can be regarded as part of the citizenship contract with the state—a
citizenship of status. In concrete terms, I suggest that the most important legal rights for slum
dwellers include access to public services and secure tenure in the form of a legal title, or (in the
case of voluntary relocation) fair and just compensation.
Second, I assess the degree of agency and level of empowerment of slum dwellers, which
I suggest can be regarded as a reflection of slum dwellers’ own political action—a citizenship of
practice. I evaluate the implications of the new policy for slum dwellers by examining as case
studies the four land sharing pilot projects (in Borei Keila, Dey Krahom, Railway A and Railway
B) that were selected as pilots of the slum upgrading campaign.
My central research question is formulated as follows:
Did the four land sharing projects announced as pilots of the Government’s slum
upgrading program in Phnom Penh constitute the “breakthrough” for the poor
that housing rights agencies anticipated in May 2003? Specifically, did the land
sharing projects mark a real shift for their residents from “squatters” (connoting
illegal and disempowered status) to “citizens” (connoting legal, empowered and
equal status)?
My power framework is based on a power relations approach, inspired by the ideas on
“capillary power” of the French philosopher Michel Foucault, which hold that power flows in
many different directions—not merely from top to bottom. A Foucauldian power relations
approach provides me with my main hypothesis about the nature of power in urban poor
settlements during the redevelopment battle, which is that residents of informal settlements are
not only passive recipients of actions and policies of other stakeholders (particularly, public
authorities and private developers), but that through their various forms of resistance, they can be
agents as well as subjects of power.

11

I argue that a power relations approach demonstrates that the residents of the four pilot
settlements already possessed a diverse set of instruments of power. These are expressed in the
various tactics deployed by residents in their interactions with public authorities and private
developers: some tactics empower residents, while others ultimately act to disempower them.
The slum must therefore be understood as a “sphere of force relations”, where power is exercised
from innumerable points, “diffuse and capillary, omnipresent, and both productive and
repressive” (Foucault 1978, 94-7). The outcome of the redevelopment struggle (the gain or lose
question) ultimately depends on the effectiveness of each party’s strategies at any one given point
in time—and may therefore be in constant flux.
1.3.1 Assessing Citizenship of Status
I assess the effectiveness of the new slum upgrading policy against two forms of
“citizenship”, corresponding with two classic traditions of citizenship in the literature. The first is
the social rights tradition, which views citizenship as status. This is a rights-based perspective on
citizenship, based on T. H. Marshall’s (1950) classic formulation of citizenship as full
membership of a community, with all the rights that membership entails, including civil, political
and social rights. The second tradition is a more active interpretation of citizenship that
emphasizes human agency and political participation (Lister, 23, 36-7). This tradition
emphasizes citizenship as practice, involving individuals organizing with others, as political
actors, in pursuit of common goals (Lister, 6, 24-5).
The approach combining both types of citizenship, which acknowledges the interplay of
structural forces (the role of the state in bestowing rights and entitlements) and agency (an active
citizenship that stresses individual responsibilities and collective action) is borrowed from Ruth
Lister (2003). This approach contains many of the basic elements of “capillary power”
propagated by Foucault. While Lister holds up citizenship as practice as a higher form of

12

citizenship than citizenship as status because it is active rather than passive, she clarifies that the
two forms of citizenship are not diametrically opposed, but represent opposite ends of a spectrum:
To be a citizen, in the legal and sociological sense, means to enjoy the rights of
citizenship necessary for agency and social and political participation. To act as
a citizen involves fulfilling the full potential of the status. Those who do not
fulfill that potential do not cease to be citizens; moreover, in practice
participation tends to be more of a continuum than an all or nothing affair and
people might participate more or less at different points in the life-course (Lister,
42) (italics added).
Both measures of citizenship currently elude residents of the four case study slums; I
argue that the land sharing projects present a major test case and “battleground” for the
achievement of slum dwellers’ citizenship rights as well as practices. The Prime Minister’s slum
upgrading policy promised residents of the four land sharing settlements a new citizenship of
status: in theory it gave slum dwellers access to basic rights, such as legal status and (implicitly)
an acknowledgment of their possession right to the land they were occupying in central Phnom
Penh, through the promise of free replacement housing as part of a land sharing agreement. I will
make an assessment of slum dwellers’ citizenship of status by asking a relatively straightforward
question:
At the end of the slum redevelopment process, did residents of the four land
sharing settlements obtain the basic legal rights promised to them by the Prime
Minister in May 2003, as measured by the achievement of secure tenure and/or
fair and just compensation, and adequate public services?
1.3.2 Assessing Citizenship of Practice through the EVLN Model
In order to assess slum dwellers’ attainment of the second, more ambitious form of
citizenship—citizenship as practice—I will make use of the classic typology of “exit, voice and
loyalty”, originally conceived by the economist Albert O. Hirschman (1970) as individual
responses to decline in firms, organizations and states. The typology was later extended by the
political scientists William Lyons and David Lowery (1986 and 1989) to examine the responses
to dissatisfaction of ordinary residents with local government services in the United States, based

13

on earlier work by social psychologists. Lyons and Lowery added a fourth response to
Hirschman’s original typology—the response of “neglect”. The result is a combined exit, voice,
loyalty and neglect (EVLN) model, describing four types of consumer behavior (in the case of
Hirschman) or political participation (in the case of Lyons and Lowery) that may arise in
response to a situation of dissatisfaction.
The Hirschman and Lyons and Lowery EVLN behaviors provide a useful set of proxy
indicators with which to operationalize notions of empowerment and citizenship in the
Cambodian context. I present an adapted EVLN model according to which two responses—voice
and loyalty—represent constructive responses to dissatisfaction and citizenship. Voice is an
indicator of “active citizenship” and collective power, given that it represents the behavior of
politically engaged citizens. The response of loyalty offers a proxy for passive but constructive
behavior of slum dwellers in response to dissatisfaction, or in my framework, “passive
citizenship”. Exit and neglect, the two destructive behaviors in the EVLN model, resemble a
form of “anti-citizenship”, as they are not oriented towards cooperation with the local government
or towards repairing the cause of dissatisfaction but are aimed instead at flight or withdrawal
from the local jurisdiction.
My central research question (illustrated in Figure 2) can now be translated in terms of
the EVLN model, and reframed as an inquiry into the degree of “active citizenship” (citizenship
as practice) among slum dwellers in Phnom Penh:
Did the four land sharing projects signal a milestone for slum residents, in the
sense that they paved the way for ‘voice’ (signifying individual and collective
power and active citizenship) at the expense of traditional coping strategies such
as ‘exit’ (signifying individual power and a threat to citizenship), ‘loyalty’
(signifying collective disempowerment and passive citizenship) and ‘neglect’
(signifying individual disempowerment and non-citizenship)?

14

Two forms of
“citizenship”
As a result of the land
sharing process, did
residents of the case
study settlements
achieve basic legal
rights?
In residents’ EVLN
responses, did ‘voice’
(active citizenship)
triumph over ‘loyalty’
(passive citizenship)
and ‘exit’ and ‘neglect’
(anti-citizenship)?
Citizenship
as status
Citizenship
as practice

Figure 2: Illustration of the central research question

1.4 Structure of the Dissertation
The analysis is organized into the following sections, as illustrated in Figure 3.
Historical trends affecting the urban poor in
Cambodia and Phnom Penh;
Current slum redevelopment forces
Resistance tactics of
slum dwellers
Exit Voice
Loyalty Neglect
External Forces
Internal Forces
Theoretical Framework
The Slum as “Battlefield”
Chapter 4
Chapters 2 & 3
Chapters 5 & 6

Figure 3: Illustration of dissertation topics by chapter

15

Chapter 2 traces the struggle by the urban poor in Cambodia for access to land, housing
and “citizenship” in historical perspective. I examine how ordinary Cambodians have been
recorded in the literature and cultivated in myths by historians, colonial administrators and public
officials, and I attempt to uncover the coping strategies and active resistances that poor
Cambodians have deployed throughout history to deal with, and overcome, pressure from
dominant classes.
Chapter 3 examines the political and economic backdrop to Prime Minister Hun Sen’s
“breakthrough” slum upgrading announcement in May 2003. I provide a description of the four
pilot land sharing settlements that are the case studies of this dissertation, and I describe how
political and economic realities since 2003 have shaped the development of the land sharing plans
over time.
Chapter 4 outlines theoretical approaches to power in the context of slum dwellers, in
both the literature and in development practice. Specifically, I am interested in theoretical
perspectives on the importance of structure and agency in shaping the nature of slums and the
survival strategies of slum dwellers. In doing so, I investigate the slum as a social construct,
shaped by competing discourses. An analysis of these discourses has direct relevance to
development practice, as these discourses contain within them their own sets of assumptions
about the slum and its residents, which in turn determine the kinds of policies and programs that
are directed at slums.
Chapter 5 details the research design and methodology that I use to analyze the shape and
outcome of the power struggle of slum dwellers in the four pilot settlements. I operationalize the
concepts of citizenship and empowerment using an adapted version of the EVLN model of
William Lyons and David Lowery, and I present the main theoretical propositions flowing from
this adapted model. In the second part of this chapter I describe my approach to data collection

16

and my questionnaire design, and I highlight the possible bias inherent in the research
methodology.
Chapter 6 presents an analysis of data from structured interviews conducted with a
representative sample of 152 “current residents” and 31 “former residents” of the four land
sharing settlements. The objective of the interviews was to examine whether residents deployed
exit, voice, loyalty or neglect behaviors—or a combination of those—to address their
dissatisfaction with redevelopment in their settlements. First, I provide a profile of residents’
socio-economic status and perceptions of tenure status and satisfaction levels. Next, I identify the
extent of relative dissatisfaction among current residents, as a basis for the implementation of the
EVLN model. Finally, in order to assess the predictive capacity of the Lyons and Lowery EVLN
model, I attempt to identify the EVLN strategies of current residents through two different
methods— through the Lyons and Lowery decision-tree model (1989) as well as through an
“independent assessment” based on three objective indicators of respondents’ behavior.
Chapter 7 returns to answer the central research questions of the study: did the four land
sharing pilot projects constitute the “breakthrough” for the urban poor that housing rights
agencies anticipated in May 2003? Specifically, did the slum dwellers attain greater “citizenship
as status” as a result of the social land concession announcement, and did they attain greater
“citizenship as practice” through the slum redevelopment process? I examine two additional
questions, which aim to link the findings from the Phnom Penh case studies to development
theory and practice. First, does the application of the EVLN theoretical model add to an
understanding of the forces surrounding slum redevelopment in the Phnom Penh context, and
beyond? And second, what are the lessons of this failure for slum upgrading policy?

17

CHAPTER 2: HISTORICAL CONTEXT
2.1 Overview of Chapter 2
The struggle by the urban poor for access to land, housing and citizenship in the slums of
modern-day Phnom Penh is only the latest incarnation of a long “struggle for access” by ordinary
Cambodians, including the poor and other historically subordinate classes, throughout the
country’s tumultuous history. In order to place the current struggle by the urban poor in a wider
historical context, Chapter 2 examines the position of ordinary Cambodians, including their
housing conditions, land tenure status and resistance mechanisms, from the pre-Angkorian period
to the early 2000s, as they have been recorded in the literature and cultivated in myths by
historians, colonial administrators and Cambodian public officials.
This Chapter is divided into two parts. The first part (covering sections 2.2 and 2.3)
traces common perceptions and “stereotypes” of the Khmers that appear in myths and historical
themes. The second part (covering sections 2.4 to 2.11) describes the position of ordinary
Cambodians during key periods in Khmer history. The historical literature on Cambodia has
tended to depict ordinary Cambodians as passive and downtrodden as a result of centuries of
adversity, and stuck in rigid social hierarchies that have limited their scope for action. This
Chapter argues that this narrative ignores the many forms of coping strategies and active
resistances that ordinary Cambodians have deployed throughout history to deal with, and
overcome, pressure from dominant classes.

18

2.2 Historical Images of the Khmers
Much of what is known about Cambodian history comes from the writings of foreigners
(Mabbett and Chandler, 13), particularly Westerners. Historical impressions of Cambodia have
long been influenced by three popular stereotypes or myths about Khmer cultural characteristics.
This first is the myth of the Cambodian as peasant. Closely linked to this is a second myth, which
holds that cities and city life are foreign to Cambodians. The third is the myth of the passive and
changeless mindset of the Khmers. Over time these myths assumed the power of “discourses”, in
the sense implied by Foucault: through a process of “production of knowledge” (Revel, 39) they
began to inform policy towards the subordinated classes, most notably during the French colonial
period and after independence, during the 1960s and during the Democratic Kampuchea period.
2.2.1 The Myth of the Cambodian as Peasant
Because a large majority of Cambodians lived in rural areas and engaged in subsistence
agriculture, the Khmers were considered to be, at heart, a village folk. Jean Delvert’s classic
study of the Cambodians (1961) was significantly entitled Le Paysan Cambodgien (The
Cambodian Peasant), as Delvert concluded that “the Cambodian is a peasant” and that “to study
the Cambodians is therefore to study the Cambodian peasant” (Delvert, 32). Delvert based his
observation on the fact that, in the late 1950s, 92 percent of the total population lived in rural
areas and that, perhaps even more important, agriculture was the “only truly national activity”—
for while “foreigners” (principally the Chinese and Vietnamese minorities) dominated most
professions, it was exclusively Cambodians who tilled the soil (Delvert, 31). In a similar vein,
Charles Meyer (1971), in his history of post-independence Cambodia, Derrière le Sourire Khmer,
suggested that the “classic Khmers”, the descendents of warriors and the founders of the
Angkorian empire, were now reduced simply to an “army of rice farmers” (neak sray), burdened

19

by the interferences of a new administrative and political class composed principally of the
descendants of Chinese immigrants (Meyer, 22).
Even when ordinary Cambodians did migrate to the cities, Delvert suggested that their
urban existence was only temporary, and that even the physical traces of their urban life were
impermanent. Delvert described human habitat in Cambodian villages as “unstable”: the houses
are made from impermanent organic materials and are easy to dismantle. The Cambodian
character itself he described as “somewhat nomadic”, with peasants used to moving around for
economic, family and religious reasons (Delvert, 198-9). The nomadic character of the villager in
turn shapes urban life: Delvert described many of the Cambodian inhabitants of Phnom Penh as
peasants who migrate temporarily to the city for unskilled, manual work, principally as “coolies”,
cyclo drivers and dockworkers. For Delvert, this was reflected in the village-like character of
many settlements in Phnom Penh and the provincial cities. Cambodian quarters of these cities had
a “rural character”, with families keeping domestic animals, and with houses on stilts surrounded
by fruit trees and land for cultivating crops. Delvert concluded that “even in the cities, the
Cambodian is a rural person” (Delvert, 31-2).
2.2.2 The Myth of the “Un-Khmer” City
From the myth of the Cambodian as peasant there followed another myth: Cambodians
were considered ill-adapted to life in cities, and as a consequence, for a long time cities were
considered “un-Khmer”. This stereotype ignored the fact that, at its height in the 10
th
and 11
th

centuries, the capital of the ancient Khmer empire at Angkor had been the largest low-density
urban conglomeration in the world (Higham, 161).
At the end of the 19
th
century, when the French arrived in Cambodia, they found
Cambodians working on the land and foreigners trading in Phnom Penh (Igout, 6). The foreign
population outnumbered the Khmer population in Phnom Penh: the Chinese population alone at

20

that time was larger than the ethnic Khmer population (at 22,000 versus 16,000) (Igout, 9). The
myth of the rural Khmer and the “foreign” city was born. The dominance of foreigners continued
until well into the 20
th
century. Delvert claims that in 1950 Cambodian nationals accounted for
only 150,000 out of 363,000 inhabitants of Phnom Penh (Delvert, 31). It was not until the years
after independence in 1953 that Phnom Penh started for the first time to become a “truly
Cambodian city”, at least in terms of demographics. The growth in the Khmer population came
about partly because of national development, with the development of infrastructure and
education nationwide, and the growth in government jobs, which were reserved for Cambodians.
But it was also due to growing insecurity in the countryside, which caused refugees from the
countryside and rural migrants to inflate the overall population in Phnom Penh (Delvert, 32).
The myth of the “un-Khmer” city—juxtaposed against an idealized existence in rural
villages—was also nurtured by several post-independence intellectuals and governments in
Cambodia in their nation-building schemes. This was particularly the case for the political Left,
for whom the inhabitants of urban areas represented a parasitic capitalist class (often of Chinese
origin) who were undermining the productive labors of a peasantry and lower class that
represented simpler Khmer values such as “truth and faithfulness” (Hou 1982, 135-9). For
Norodom Sihanouk, the “true people” of Cambodia were to be found in rural society—“genuine,
hard-working, loyal and faithful” peasant men and women whom he contrasted with the
scheming, “corrupt” and “powerful” political class to be found in the cities (Jeldres 2005, 39-43).
2.2.3 The Myth of Changelessness
Yet another myth followed from the peasant stereotype. In Western writings
Cambodians were typecast as “sedentary” and “static” (Edwards 2006, 421), as well as
conservative, prone to inertia, and docile. All of these traits appeared to be characteristics of a
deeply rural society that was unchanging (Chandler 2003, 2). Some observers offered other

21

variations on the same theme: in their account of social organization and power structures in rural
Cambodia, Jan Ovesen, Ing-Britt Trankell and Joakim Öjendal (1996) suggested that what
appeared to Westerners as “conservatism” and an unwillingness to accommodate change was in
fact a function of the Khmer quest for order, “for restoring and/or upholding the ideal social and
cosmological order which is a prominent feature of Khmer culture and world view” (Ovesen,
Trankell and Öjendal, 36). And David Chandler (1982) argued that Cambodians were a
“backward-looking people”, in the sense that
their social conducts were based on ideas, techniques and phrases which were
passed along through time and space like heirlooms, with the result that people
were continually reliving, repeating or restoring what was past—in ceremonial
terms, in adages, and in the agricultural cycle (Chandler 1982, 54).
Westerners who came with a mission to bring development to Cambodia contrasted the
attachment to structure inherent in the “conservatism of Khmer culture” with the forces of agency
inherent in development efforts. Thus, Ovesen, Trankell and Öjendal questioned whether the
Khmer world view and the development world view were even compatible (Ovesen, Trankell and
Öjendal, 83). Many concluded that they were not: Jean-Claude Pomonti and Serge Thion (1971)
argued that the “economic apathy” of the peasantry, “who live more or less in a situation of
autarky”, was one of the main causes of the under-development of Cambodia (Pomoni and Thion,
23).
A Cambodian perspective on this common stereotype of the Khmer was offered by the
future revolutionary leader of the Democratic Kampuchea regime, Khieu Samphan (1959), in his
doctoral thesis on the nature of Cambodia’s economic and industrial development. Khieu
suggests that an explanation for the so-called “apathy” of the Khmer race lies in the economic and
social structures of Khmer society, and particularly, in the “’tyrannical’ influence of mandarins
and feudal lords which diminished people to nothing more than ‘abstract beings’”, and which
“forbade Khmer peasants and artisans from developing their full potential”. Under these feudal

22

conditions, Khieu argues, the Khmer peasant is just as apathetic as any French serf living in the
middle ages was at the time (Khieu, 33).
2.3 Principal Themes in the Literature
Historians have identified numerous themes appearing as threads throughout Cambodian
history. Two themes in particular are frequently depicted as “burdens of the past”
14
that mark the
lives of ordinary Cambodians to this day: the legacy of warfare and upheaval, and the strict
hierarchy and patron-client structures that pervade Khmer society. Both themes are often
presented as (political, social and economic) structures that envelop ordinary Cambodians and
from which they find it hard to escape. However, there is a third theme that offers the historically
subordinated class a new environment with more scope for individual initiative and “agency”: the
rapid economic growth of Phnom Penh and the re-urbanization of the capital city since the late
1980s.
2.3.1 A Legacy of Warfare and Upheaval
Ian Mabbett and David Chandler (1995) claims that war has been a “normal state of
society” and a factor that conditioned society throughout Cambodia’s history (Mabbett and
Chandler, 156). Elizabeth Becker (1998) writes that Cambodia is a country with a “tradition of
violence”, long accustomed to “quarrelsome, despotic rulers” (Becker, xv). War and civil
disturbance are relatively fresh in the minds of most Cambodians: Khmers Rouges forces were
defeated in remote areas of Cambodia only in the late-1990s, and as recently as the late-1990s,
residents of Phnom Penh lived through armed conflict between political parties. For Jan Ovesen,

14
The term is borrowed from Charles Meyer (1971), who referred to “le poids du passé” to describe certain
elements of Cambodia’s past that continue to influence its current institutions and the “spirit of its people”
(Meyer, 44-5).

23

Ing-Britt Trankell and Joakim Öjendal (1996) the legacy of warfare for rural Cambodians is both
tangible (resulting in the large-scale migration to the cities as well as reduced economic
opportunities, due to land mines in the countryside) as well as intangible (including the
psychological scars of lingering fear and insecurity, and domestic violence) (Ovesen, Trankell
and Öjendal, 29-31). For the anthropologist Fabienne Luco (2002) Cambodia’s “long succession
of battles, internal conflicts, rebellions, insurrections, territorial partitions and supervision by
foreign countries” partly explains a “prevailing feeling of insecurity”. It also shapes attitudes to
those in positions of authority: “in a shattered society, people long for personal safety and greatly
fear disturbing the established order”. In addition, ordinary Cambodians prefer to rely solely on
themselves and keep contacts with the authorities to a strict minimum (Luco, 14).
2.3.2 Hierarchy, Patron-Client Relations and Paternalism
Almost all Western historians of Cambodia have commented on the highly stratified
nature of Khmer society, and on the importance of patron-client relations in underpinning socio-
economic relations. David Chandler (2003) refers to the “pervasiveness of patronage and
hierarchical terminology in Cambodian thinking, politics, and social relations”, whereby, “for
most of Cambodian history…people in power were thought (by themselves and nearly everyone
else) to be more meritorious than other people” (Chandler 2003, 2).
Historians of ancient Cambodia commonly ascribe the stratification of Khmer society to
the “authoritarian character” of ancient Khmer kings who, in the Hindu tradition, ruled as Gods
on earth (devarajas). Khmer kings appeared not only as fathers over their subjects, but also as
masters (gurus) over their servants (Sahai, 33-4). Fabienne Luco, quoting Stanley Tambiah
(1976), likens Khmer society to a “galaxy society”, with Khmer kings having long ruled at the
center of a “constellation of mandarins and princely estates, all linked by a contract based on
family relations: father, children and grand-children”. The contract implied royal protection (to

24

guarantee prosperity, administer justice and fight wars) in return for popular allegiance (in the
form of offerings, tributes, duties and the supply of manpower in the event of war). Luco draws
parallels between this ancient political arrangement and the manner in which the present “King
father” (Samdech Euv, or “Highness father”) Norodom Sihanouk still addresses his people as
“grandchildren” (kon chaw) (Luco, 15-6). The “grandchildren”, meanwhile, have long seemed
willing to perpetuate this arrangement: May Ebihara (1968), in her ground-breaking
anthropological study of village life in Cambodia, remarked that villagers adopted public
statements of then-Prince Norodom Sihanouk unquestioningly because “they believe that there is
no wisdom greater than that of the prince” (Ebihara, 548-9 quoting Smith, 643).
Anthropologists and political scientists have long emphasized the importance of patron-
client ties in societies across Asia, and Cambodia is no exception (Hughes 2003, 18-9). James C.
Scott (1972) documented extensively the nature of patron-client ties in the Southeast Asian
context, which he suggested were an important structural principle of Southeast Asian politics.
Scott defines these vertical linkages, at their most basic level, as
an informal cluster consisting of a power figure who is in a position to give
security, inducements, or both, and his personal followers who, in return for such
benefits, contribute their loyalty and personal assistance to the patron’s designs
(Scott 1972, 92).
In the modern economic and political context of Southeast Asia, defined by the
availability of new economic resources and electoral democracy, the nature of patron-client
structures has changed. Patron-client relations are now more “closely linked to the national level,
with jobs, cash and favors flowing down the network, and votes or support flowing upward”
(Scott 1972, 105-6). In modern Cambodia, a new power elite has replaced kings as chief
patrons. According to Caroline Hughes (2003) the economic liberalization at the end of the
1980s represented a turning point for patron-client relations: power-holders considered the new
economic opportunities as offering a “matrix of resources that could shore up exclusionary

25

loyalties within the weak state apparatus, and reducing the field of action for resistance in rural
villages, as a means to strengthen the state militarily and politically” (Hughes 2003, 19).
Most historians seem to accept that the traditional patronage structures of Khmer society
rendered ordinary Cambodians passive and elevated loyalty to a critical survival strategy—at
least in rural areas. Paul Collard (1925), resident mayor of Phnom Penh during the early years of
the French protectorate, wrote that the Khmer system of patron-client relations “undermined any
instinct [the common man may have] for private property and the cultivation of crops”. The
“oppression of the mandarins” was responsible for “impairing in the [Khmer] race any innate
penchant for resistance and initiative” (Collard, 78). More than forty years later, Ebihara (1968)
remarked that the peasants in her village of Svay in Kandal province remained passive recipients
of, rather than active agitators for, any improvements in their lives. Villagers and local officials
treated district-level officials, acting as local mediators of central government, as “big people”
(neak thom) of clearly superior rank, demanding of respect and distance, rather than as their
representatives (Ebihara, 512-3, 541). Chandler suggests that, as a result of their widespread
acceptance of the status quo, Cambodians have gone through, in Marxist terms, “centuries of
mystification” (Chandler 2003, 2)—or, in James C. Scott’s terms, a form of passivity that
assumes at least a fatalistic acceptance of [the] social order and perhaps even an active complicity
(Scott 1985, 39).
2.3.3 Capital City versus Countryside
More recently, a third theme has begun to attract the attention of historians—the rapid
development of Phnom Penh and the growing disparity between the capital city and the
countryside (srok). David Chandler argues that the contrast between Phnom Penh and its
hinterland first became significant in the late 1890s, when the French started to develop the city
as a capital of their protectorate. Whereas in the rural areas French officials found “old habits of

26

patronage, dependence, violence, fatalism and corruption largely unchanged from year to year”,
the capital city started to enjoy the first of many real estate booms (Chandler 2003, 146-8).
At key points throughout the 20
th
century Phnom Penh had occupied a role as magnet and
at times refuge for poor migrants from the provinces. This migration was often seasonal or
situational, in response to push factors such as hardship or armed conflict in the countryside. The
political economist Caroline Hughes (2003) argues that the opening up of the Cambodian
economy in 1989 ushered in a period of rapid economic development in Phnom Penh that created
new opportunities for the poor, both economically and politically. Urban economic growth
resulted in new employment opportunities, particularly in private garment factories, construction
and the sex industry. These opportunities complemented the familiar seasonal employment forms
of the past, such as cyclo driving and motorcycle taxi (motodop) driving (Hughes 2003, 183). At
the same time, new channels were established for “substantive political participation by the poor”
as well as “new social forces flexing political muscles for the first time”. The net effect, after a
decade of economic expansion in Phnom Penh, was “an urban economy which differs radically
from the more easily monopolized arena of the rural village” (Hughes 2003, 173).
The age-old patron-client ties that had characterized village life in Cambodia for
centuries were taking on new forms in the Phnom Penh of the 1990s and beyond. In contrast to
the village, the city offered clients “multiple sources of patronage, as well as places to hide when
patrons were crossed or abandoned” (Hughes 2003, 176). The result, for Hughes, is the
“depersonalization” and “commoditization” of patronage links in Phnom Penh, as these are now
bought and sold on the open market to the highest bidder, rather than—as in the past—awarded to
political loyalists. Hughes concludes that this development has resulted in a loss of political
control from the center and—in the newly opened up political space—the urban population has

27

“proved to be significantly more unruly than the rural population in its voting patterns and in its
predilection for mobilization against government policies” (Hughes 2003, 176).
2.4 Pre-Angkorian and Angkorian Cambodia
During the first eight centuries AD Cambodia, like most other areas of early Southeast
Asia, consisted of a multitude of small principalities that competed for hegemony. These small
states were “fiefs run by families of warrior barons” (Mabbett and Chandler, 87) which
considered themselves minor rajas or kings, each of which had its own court and elite and
“entourages” of people who grew food and could be called upon to go to war for these states
(Chandler 2003, 17). In this pre-Angkorian period political unity was provided by these minor
kings, and central authority was weak (Sahai, 140). The principalities shared a “despotic
language of politics and control”, as the minor kings saw themselves as absolute rulers and as
rivals, independent from each other (Chandler 2003, 26).
A significant transition occurred in the 9
th
and 10
th
centuries, when a foundation for
durable political unity in Cambodia started to be established during the reigns of King
Jayavarman II and his son Jayavarman III. This period marked the beginning of the end of a long
era of “jostling principalities” (Mabbett and Chandler, 85, 88) and the transition to “some sort of
self-aware community” (Chandler 2003, 36), and eventually to pan-Khmer regime and empire
based at Angkor. Unification was an “irregular and fitful process”, however, and not a “once-for-
all achievement.” There would be many reverses and serious challenges to unity throughout the
Angkorian period (Mabbett and Chandler, 85, 88, 92, 99). Integration between the rival
principalities was a gradual process that is attributed to a combination of factors, including
population growth, victories in war or protracted periods of peace, and increased wet-rice

28

technology and the mobilization of manpower instead of a dependence (solely) on subsistence
agriculture and coastal trade, as in the pre-Angkorian period (Chandler 2003, 26-27).
2.4.1 The Foundation of an Empire at Angkor
The founding of the capital of the new unified kingdom at Angkor is commonly placed in
the year 802 AD, when a consecration ceremony was held for King Jayavarman II on the plateau
of Phnom Kulen (the “mountain of litchis”), at which the King became the state’s supreme ruler
and universal monarch (Chandler 2003, 34). The Angkorian complex of monuments and temples
became the religious and ritual center of the kingdom, and it radiated spiritual power (Higham,
151). Its location between the northern shore of the great lake Tonle Sap and the quarries of the
plateau of Kulen enabled a limitless supply of fish and building materials, while regular flooding
encouraged the cultivation of rice. Agricultural surpluses sustained the kingdom (Higham, 53).
Historians agree that the cultivation of rice was the “solid foundation” of the civilization at
Angkor, but there are opposing theories about how the rice was produced and what the
significance of this was for the kingdom (Higham, 160). According to the French scholars
Bernard-Philippe Groslier (1979) and Jacques Dumarçay (1998), Angkor was a “hydraulic city”
par excellence, with a sophisticated irrigation system centered on the large reservoirs (barays)
that are found adjacent to the historical temple complex. In this view, the rise of Angkor was
predicated on the success of its irrigation system, and its fall was determined by the siltation of
this system, and by difficulties with its maintenance (Higham, 156, 161).
The opposing view, held by Philip Stott (1992), W.J. van Liere (1982), and Robert Acker
(1997), rules out the central importance of irrigation for Angkor, and concludes that other factors
were more important contributors to the rise and eventual decline of Angkorian civilization.
Based on an examination of the various dams on the Siem Reap river, van Liere claims that their
total capacity could never have been used to irrigate more than four thousand hectares (van Liere,

29

46)
15
. Given that few channels for distribution of water to surrounding fields have been
discovered in Angkor, Ian Harris suggests a more symbolic function for the barays at Angkor,
namely, that these were not simply (or even mainly) “reservoirs”, but part of a system evoking the
oceans surrounding Mount Meru in Hindu and Buddhist cosmology (Harris 2006, 11).
2.4.2 Ancient Loyalty and Patronage Systems
In terms of its organization, Angkorian society appears to have been very stratified. At
the top was the king—a supreme and absolute ruler with divine characteristics—and below the
king society was divided into free people and servants or “slaves”. In practice, however, David
Chandler (2003) points out that the situation was more complex, and that ancient inscriptions
reveal “webs of relationships, responsibilities, and expectations, within which everyone appears
to have been entangled” (Chandler 2003, 47).
Kings were praised as embodiments of virtue, occupying the very top of society because
of their merit and power. But more than that, kings were considered superhuman and even
divine. In terms of its political institutions and administration, Sachchidanand Sahai (1970)
argues that a distinguishing feature of the Angkorian period was the institution of the devaraja (or
god-king) cult
16
(Sahai, 149-150). The origin and significance of the devaraja is not entirely
clear to historians: Ian Harris (2006) claims that while earlier scholars (such as Georges Coedès
and Pierre Dupont) understood the devaraja to be either the deified king himself, or an image of
Shiva standing in the king’s stead, modern scholars believe the devaraja to be either a special

15
W. J. van Liere, “Was Angkor a Hydraulic City?” Ruam Botkwam Prawatisat 4 (1982): 36-48, quoted in
Ian Harris, Cambodian Buddhism: History and Practice (Chiang Mai, Thailand: Silkworm Books, 2006),
243
16
The cult is known as devaraja in Sanskrit, and kamraten jagat ta raja in ancient Khmer (Mabbett and
Chandler, 89).

30

mobile image (Hermann Kulke, 1993) or some sort of sacred fire (Hiram Woodward, 2001)
(Harris 2006, 11-12).
For ordinary Cambodians who had experienced the transformation of their political
loyalties from the earlier monarchies to the ideal of political unity, the devaraja cult redefined
these loyalties in ritual terms (Mabbett and Chandler, 88). The early Angkorian kings, starting
with Jayavarman II, were considered devarajas or semi-divine god-kings with absolute secular
powers and the benediction of the gods (Mabbett and Chandler, 89). King Jayavarman II set a
precedent and projected his court as the center of the kingdom, and a representation of heaven
(Higham, 53). The devaraja cult linked the Khmer monarchs to the Hindu god-king Shiva, and
thus sacralized the Angkorian kings, though historians believe that later in the Angkorian era, the
devaraja cult was no longer the definitive royal cult and did not have a unique role in making the
sacred link to kingship (Chandler 2003, 36, citing Hermann Kulke 1978).
The French historian Georges Coedès (1947) points out that the divinity cult existed not
only at the level of kings, but that a “cult of big men” flourished in Angkor between the 9
th
and
12
th
centuries. This cult entailed the erection of statues where the image of the person depicted is
represented by divine traits. These deified statues were not only of kings but also of “persons of
distinction”, both men and women (Coedès 1947, 47-8). The gods depicted were both Hindu and
Buddhist, demonstrating a remarkable syncretism, particularly later in the Angkorian period,
when Buddhism took hold of the kingdom. The statues were considered a “living tomb” and
were meant to ensure the cult of these persons’ image, and the image of their parents, in durable
materials (Coedès 1947, 54, 60). Both the devaraja cult and the “cult of big men” were
“aristocratic cults” that united the people in adoration of their kings and chiefs (Coedès 1947, 66).
The loyalty cults marked the physical landscape of Angkor. According to Mabbett and
Chandler, a common physical expression of the “quest for legitimacy” was a group of monuments

31

consisting of shrines as a tribute to previous rulers and their wives, and in the middle, a pyramid
representing the mountain of the gods (the mythical Mount Meru, which in Hindu cosmology lies
at the center of the universe), which would house the king’s own relics after his death (Mabbett
and Chandler, 96). Inscriptions show that during the first two centuries of Angkor kings and
notables would earn merit by endowing religious foundations (Mabbett and Chandler, 97).
The devaraja cult lost most of its power with the advent of Buddhism in the later years of
the Angkorian period. But several scholars and historians argue that the devaraja cult had such
deep roots in the court that it still has a hidden influence and lasting significance in modern
Cambodian history and politics. One example of this hidden influence in modern Cambodia,
cited by Coedès, is a Buddha standing at the foot of the altar of the Silver Pagoda in the Royal
Palace in Phnom Penh. The Buddha is bedecked in the crown jewels of King Norodom, who
reigned in the late 19
th
century. The statue has exactly the former king’s body size—in other
words, it is not just of the Buddha but it represents an image of the devaraja, representing the
“lord of the universe as royalty” (Coedès, 66-7). Elizabeth Becker (1998) draws a more modern
parallel with the Angkorian-era cult, claiming that Norodom Sihanouk portrayed himself as an
embodiment of a devaraja and treated Cambodia as a his own paradise, taking it upon himself to
“protect” his country and keep out “unwanted foreign or modern influences that might disrupt the
largely rural, Buddhist life in his kingdom” (Becker, 5).
In spite of their semi-divine status, Angkorian kings also had to govern their realm. The
king was the chief magistrate of his kingdom (Sahai, 151): the king remained the court of last
appeal in Cambodia up until the 20
th
century (Chandler 2003, 47). To manage all these
responsibilities Angkorian kings had to be political operators. They ruled by extending networks
of patronage and mutual obligations outward from the palace, at first through close associates and

32

family members but becoming diffuse—and more dependent on local power holders—at the
edges of the kingdom (Chandler 2003, 48).
In the economic realm, kings put in place large development schemes, such as the
construction and maintenance of irrigation works; managed the religious foundations by granting
them land and slaves; bestowed titles on high officials; and conducted foreign relations (Chandler
2003, 47). Of all the economic schemes, the hydraulic schemes of Angkor were perhaps the
most famous, as they were linked to the political power structure of the kings. In this sense some
scholars argue that the Angkorian state was a classic example of the Marxist concept of an
“Asiatic mode of production” (Chandler 2003, 43, 54; Vickery 1999, 286), according to which
“oriental” states organized mass labor for great public works schemes, especially water works.
Politically and economically, then, the water works enabled the state to dominate the mass of
intermediate producers and peasants (Wittfogel, 360-2). In the Angkorian case, the “Asiatic
mode of production” was sustained at its peak in the 11
th
century under King Suryavarman I,
whose large-scale hydraulic schemes were linked to increased bureaucratic centralization
(Chandler 2003, 43).
2.4.3 The Fate of Ordinary People in Ancient Cambodia
Relatively little is known about the lives of ordinary people in Angkorian Cambodia.
Most of the textual evidence of the period comes from about 1,200 stone inscriptions in Sanskrit
and Old Khmer, which were found in sanctuaries and relate principally to the support and
administration of religious foundations. The inscriptions give little insight into the lives of
ordinary people during the Angkorian period; they do show, however, that slaves were listed as
commodities (Chandler 2003, 48). The inscriptions were written largely by and for the elite, as
their purpose was to directly or indirectly reinforce status, power and ownership as they affected,
and were affected by, royalty, the elite, the religious establishment, the bureaucracy, and the

33

judiciary (Lustig, Evans, and Richards, 1-6). The inscriptions must be interpreted with care, as
much of what was written about the deeds and qualities of the rulers and founders of the religious
foundations may be “somewhat fanciful” (Lustig, Evans, and Richards, 7).
The only still surviving historical account of daily life in ancient Cambodia was by a
Chinese emissary, Zhou Daguan (or Chou Ta-Kuan), who spent a year in the (by then already
declining) Khmer empire in the late-13
th
century, as part of a diplomatic mission sent by Timur
Khan to the throne of King Indravarman III. Zhou observed that the housing conditions in the
empire reflected the highly stratified nature of Angkorian society. Ordinary people lived in
thatch houses, in simple conditions, with “no tables, seats, basins or buckets”, while “noble or
rich households use silver or sometimes even gold for their receptacles” (Zhou, 81). People of
princely rank lived in bigger houses, on account of their higher merit:
The dwellings of the princes and principal officials have a completely different
layout and dimension from those of the people. All the outlying buildings are
covered with thatch; only the family temple and the principal apartments can be
covered in tiles. The official rank of each person determines the size of the
houses. Ordinary people only have thatch roofs, and would not dare to place the
least tile there. The size of their houses depends on individual wealth, but the
people would never dare imitate the arrangement of the dwellings of the nobility
(Zhou, 23-4).
Despite the stratification in Angkorian society and the inequalities inherent in the
“Asiatic mode of production”, Michael Vickery (1999) argues that the “Asiatic mode of
production” characteristic of the Angkorian state was “at the time a progressive social formation
which in addition to luxuries for the rulers no doubt provided the base population with all that
they imagined they needed and certainly with adequate food” (Vickery 1999, 288).
David Chandler suggests that most people were “subjects” rather than free people, as they
were at the disposition of patrons, who had the right to sell them (Chandler 2003, 47-8). Many of
the subjects can be considered as servants or “slaves”, who in ancient inscriptions were listed as
commodities. In practice, though, the distinction between slaves and free people was not always

34

so clear: there were many categories of slaves, including slaves who owned slaves, and slaves
who married into the royal family. On the other hand, there were also free people who were
disposed of by others, just as slaves (Chandler 2003, 47, citing Mabbett 1983
17
). Zhou Daguan
confirmed the presence of “slaves” in Angkorian Cambodia, whom he referred to as “savages”
taken from the “mountain vastnesses”, who dared not enter their masters’ homes: “Savages are
purchased to serve as slaves. People who have many have more than a hundred; those with few
between ten and twenty; only the very poor have none at all” (Zhou, 39).
2.4.4 Land Acquisition by the Plough
The king was the largest property owner in Angkor-era Cambodia, with powers akin to
the powers of eminent domain of a feudal lord in medieval Europe (Sahai, 148). According to
ancient Khmer law, as recorded and translated in the late 19
th
century by Adhémare Leclère and
compiled in the “Cambodian Codes”
18
, all land in the Kingdom of Cambodia belonged to the
king, who was the only legal “property owner” of all land in Cambodia (Delvert, 488). Despite
his overall powers, however, the king could not be considered as “absolute owner” of all
properties in his kingdom; private (use) rights were respected (Sahai, 146-7).
There was no system of de jure individual property rights. Instead, Cambodian custom
recognized the principle of “land acquisition by the plough”: a use right to land was established
for peasants if land was exploited for agricultural use. Possession that was “continuous, public
and in good faith” amounted to a de facto property right for the individual (Cambodian Codes,
volume 617, cited by Delvert, 489). The use right was recognized when peasants informed the
authorities of it. In turn, the authorities had a right to levy taxes on the produce of the land

17
Ian W. Mabbitt, “Some Remarks on the Present State of Knowledge about Slavery at Angkor”, quoted in
Anthony Reid (ed.), Slavery, Bondage and Dependency in Southeast Asia (St. Lucia, Australia, 1983).
18
Adhémare Leclère, Les Codes Cambodgiens (Paris: Ernest Leroux, 1898).

35

(Pomonti and Thion, 28) although the Codes specified that the taxation should be minimal. In
inhabited areas, such as in a village, the occupation of vacant land, either for residential purposes
or for cultivation, was tolerated only with a prior declaration from the authorities, whose duty it
was to measure the plot of land
19
.
The de facto acquisition and possession right of land “by the plough” was the
consequence of Buddhist law imposed on the throne, which stipulated that land is “God’s land,
entrusted to the king, and released by the king to all men who have a use of it”
20
. The individual
possessor had the possibility to pass on this land through inheritance. He could also rent out his
field to others, but if the land was left fallow for three consecutive years, then the land returned to
the king (Cambodian Codes, volumes 374, 407 and 617, cited by Delvert, 489).
In rural areas, other recorded ways to obtain a “right” to a piece of land, besides
customary possession through use, were through royal concessions, donations or other forms of
privileges (Goulin, 98). In theory, the king had a right to recover the land when he needed it, but
in practice, this right was rarely exercised (Pomonti and Thion, 28). Land in itself was of little
value; ancient Khmer law and custom dictated that it was human labor that added value to the
land, which reflects the fact that throughout most of Cambodian history there was always more
land than people to cultivate the land. Landlessness among the peasantry was in fact not
recognized to be a serious problem by historians, including Delvert, who remarked in 1961 that
even “in the most densely populated regions, 90 to 100 percent of peasants are property owners”
(Delvert, 501). Jean-Claude Pomonti and Serge Thion remarked as recently as 1971 that overall

19
A. Boudillon, Le Régime de la Propriété Foncière en Indochine, Rapport au Ministre des Colonies
(Paris, 1915), 109.
20
Note by Prince Yukanthor to the President of the French Council, 1899, as quoted in Delvert, 489.

36

population density was not high, and that there were still substantial areas of the country that
were ‘empty’” (Pomonti and Thion, 24).
2.4.5 An Early System of Land Concessions
In the absence of a monetary system, there was an extensive system of land concessions
(also referred to as land grants) in Angkor-era Cambodia. Land concessions were an essential
form of currency as well as a way to reward loyalty. Kings rewarded civil servants and political
allies with land concessions as well as court titles and duties (Higham, 151). Land concessions
represented “tenure salary” or fiefs for civil servants, in exchange for which they professed their
loyalty to the king and vowed to serve him. In addition, the king gave land concessions and
sometimes even villages and entire towns as rewards to his spouses, to other members of the
royal family, to military officers, and to other allies (Sahai, 143-5). The legal conditions of land
concessions are not entirely clear, but it appears that property concessions were given out for free.
These properties could be passed on through inheritance, and properties and villages could be
sold, although the king reserved the right in principle to revoke the concessions (Sahai, 144).
Government officers were expressly mandated to ensure that only vacant lands were included in
land grants (Sahai, 146-7).
Beneficiaries of a land concession invested capital in the form of buffaloes and cattle,
seeds and ploughs, thereby bringing new land into production. A proportion of this new land was
usually donated to a state temple. In addition, beneficiaries usually founded a family or lineage
temple (Higham, 152). Large numbers of priests and workers would be attached to these temples.
Taxes on properties were one of the state’s principal revenue sources. In a society
without a monetary system, taxes were paid in kind. A cadastre recorded the dimensions of each
parcel and its revenue. In principle, each property was taxable, although in many cases collective
taxes, for instance on an entire village, were imposed instead. Certain institutions, such as

37

religious foundations, were granted immunity from taxes. Apparently, charges were heavy, and if
contributors could not pay they were subject to property confiscations and heavy penalties (Sahai,
114-5, 119, 147, 151).
2.4.6 The Rise and Fall of Angkor
Starting in the 11
th
century, Angkor became a major empire, as its population expanded
and its economy grew more complex. During the reign of King Suryavarman (1113 to 1145-
1150), Angkor was at the peak of its glory and it experienced its greatest physical expansion. This
was a period of large construction projects, including hydraulic systems and ornate temples,
which were built to show the “wealth and piety of benefactors past and present” (Mabbett and
Chandler, 106). The most prominent temple built during this time was the national shrine of
Angkor Wat (“city monastery”)—a grand expression of royal power meant to command
admiration and awe among Khmers and attract divine favors for the kingdom as a result of the
ruler’s piety (Mabbett and Chandler, 103-5).
Two measures were instituted to attempt to manage the expansion of Angkor. First, the
state brought the religious foundations under a state-supervised system dominated by a group of
state temples endowed by the king. These foundations had become wealthy and influential, as
monks and priests became advisors to the court and as their temples housed many people and
were possibly chief providers of education (Mabbett and Chandler, 102). Second, to cope with
population growth the state sponsored what Mabbett and Chandler call a system of “ecclesiastical
colonization”, triggered by a king’s order to grant a new district. The endowment of temples in
the new district became the precursor for the founding of new settlements, as each temple led to
the establishment of a “community of incumbents, artificers, servants and bonded labor granted
by benefactors to supply its needs” (Mabbett and Chandler, 102). Inscriptions from the first two

38

centuries of Angkor show that new villages were established by the deliberate transplanting of
groups of people (Mabbett and Chandler, 97).
Despite the appearance of stronger central authority at Angkor, the Khmer empire in fact
suffered “endemic instability”, both at the top and at the bottom. Unclear rules of succession and
factionalism at the center created conflict within the ruling dynasty. At the local level,
meanwhile, there was a “constant problem” of maintaining control over the provinces and
managing the kingdom’s “centrifugal tendencies” (Higham, 54, 154). The political unity of the
empire remained fragile, as its authority was continually challenged by the power of small
principalities and local potentates. Village elders and notables retained a strong influence over
affairs in their area. Sahai suggests that the “excessive use of land concessions” was possibly a
key factor that contributed to the eventual collapse of the Angkorian kingdom. Many of the
beneficiaries of land concessions included regional military chiefs, and their landholdings gave
them a power base that allowed them to manipulate local political instabilities to their advantage,
thus posing a direct threat to central authority. There is evidence (in the form of epigraphs) of
numerous local revolts fomented by these regional army chiefs (Sahai, 150-1). According to
Michael Vickery (1977) the growing independence of some of the outlying regions was
facilitated by the neglect of the highway system that connected Angkor with the outer provinces
(Vickery 1977, 515).
Another development during this period that contributed to growing local power was the
increased importance of land ownership, which was a direct function of greater competition for
land due to population pressure. Inscriptions from the period show numerous cases of contested
land ownership (Mabbett and Chandler, 102). The widespread phenomenon of de facto property
“owners” enabled a degree of independence from the central government.

39

In addition to the internal weaknesses of Angkor, starting in the 13
th
century broader
economic, social and political changes were underway in the entire Southeast Asian region that
would eventually contribute to a shift of power away from Angkor. These changes included,
first, the growth of Theravada Buddhism at the expense of “state-sponsored and cast enhancing
Hindu cults”. A second change was the rise of Siamese power to the west, and the corresponding
loss of Khmer political control in areas to the northwest of Angkor. Another significant
development was the rapid expansion of Chinese maritime trade with Southeast Asia (Chandler
2003, 70-8). Of all these factors the third external factor might have been the most important
21
.
For Michael Vickery the shift away from Angkor represented a shift of emphasis from “inland
agrarian activities” to Cambodia’s “integration in the China-Southeast Asia trading network”.
This meant a gradual shift of power from landlocked Angkor to two new centers, Ayuthayya in
Siam and Phnom Penh-Longvek
22
in southeastern Cambodia, both of which were located on
lower courses of major rivers with good access to the sea (Vickery 1977, 511, 522).
The final trigger for the abandonment of Angkor was probably the lengthy siege and
subsequent sacking of the city in 1431 by the “vigorous” Siamese kingdoms of Sukhothai and
Ayutthaya (Higham, 161). By that time, however, Cambodia had already entered its “middle
period” (Chandler 2003, 79), which was characterized by a shift of the capital from Angkor to the
southeast of the country, where new river capitals were established, first at Longvek, then at
Oudong, and finally at Phnom Penh, towards the beginning of the 15
th
century.

21
The reasons for the decline of Angkor are still being debated today. Other theories for the decline that are
advanced by historians include the collapse of the irrigation system due to siltation and maintenance
problems, and the fact that the introduction of Theravada Buddhism—in place of the traditional worship of
Shiva linked with ancestors—removed the need for state temples at Angkor (Higham, 161).
22
Longvek was one of the first capitals in southeastern Cambodia, before Phnom Penh.

40

2.5 A New Khmer Capital at Phnom Penh
The strategic location of Phnom Penh, at the confluence of three rivers (the Mekong, the
Tonle Sap and the Bassac—a site which Cambodians call chaktomuk, or four faces, and which the
French would later call the quatre bras), was the subject of an ancient legend—a site where a
wooden Buddha with four faces was apparently found floating in the water in a tree trunk and
refused to be taken along by the currents beyond that point.
While Phnom Penh’s unique physical location on a crossroads of major rivers was
considered auspicious and good for maritime trade, it also presented challenges for development.
The site is at the center of an enormous watershed created by the three rivers to the east, and a
complex of natural floodplains, marshes, seasonal ponds (boeungs) and tributaries (preks) to the
north, south and west of the city. The land was subject to flooding, erosion and change from the
fast-flowing current of the Mekong river, the unpredictable water levels from one year to the
other, and the deposits of large quantities of sediment. Any human settlement in this site,
therefore, was vulnerable to constant environmental risk (Goulin, 54-6). As a result of these
natural hazards early settlement in Phnom Penh was confined, in linear fashion, to the banks of
the Bassac and Tonle Sap rivers, and to a small village grouped around the higher-lying phnom
(Grant Ross 2005, 5-7)
23
. The French urban geographer Christian Goulin claims that, in the pre-
colonial period, there was no urban planning in Phnom Penh (Goulin, 101), although there is
evidence that King Ponhea Yat, the first king to relocate the capital to Phnom Penh in the 15
th

century, embarked on some public works. The king apparently put in place a drainage system
with protective ditches, filled in zones vulnerable to flooding, and had fresh water brought in

23
The Phnom is the small mount after which the city takes its name.

41

through a canal. In this sense, according to Michel Igout, he appeared to be a “worthy heir” to the
hydraulic engineers of Angkor (Igout, 2).
Early European travelers to Phnom Penh described the poor condition of the wooden and
bamboo houses (Tully, 32). Richer families had houses constructed mainly of bamboo and
hardwoods, while houses of poorer families were made mostly of straw and leaves. In later years
kings permitted Chinese traders to use bricks to construct their shop houses (Goulin, 98). Public
hygiene was considered poor, with “sordid streets” and garbage accumulating under the houses.
Animals were crowded in among the human population and in people’s homes at night to prevent
theft (Igout, 5-6).
By the mid-19
th
century the French naturalist Henri Mouhot estimated the city’s
population at 10,000 permanent residents (who lived on land) and an additional “transient
population” of double that size who lived on boats in the Tonle Sap river (Tully, 32 quoting
Mouhot 1864)
24
. By this time Phnom Penh was a flourishing trading center, and among its
population were many foreigners, particularly Chinese and Vietnamese, but also groups of
Malays, Siamese, Indians and Laotians (Igout, 5). Nevertheless, despite its cosmopolitan
population Phnom Penh was described as no more than a “big village” (Goulin, 2), both as a
result of its physical aspect (the majority of its dwellings being straw huts) as well as its small
population.
For a majority of the population in early Phnom Penh land tenure appears to have been
quite insecure. Urban land distribution was highly regulated, and kings took an active role in
distributing plots of land. All land occupants were tenants of the crown: possession was subject
to the approval of the king, and “subject to local customs” (Goulin, 99). Around the royal palace

24
Henri Mouhot, Travels in the Central Parts of Indo-China (Siam), Cambodia and Laos during the Years
1858, 1859 and 1860, vol. 1 (London: John Murray, 1864), 225.

42

and along the riverbanks, land occupants required royal authorization to build, and if
authorization was given, it was usually subject to the condition that the occupants vacate the land
after a period of ten years and return the property to the crown. The occupants of these plots of
land paid land rent or fees, but were nevertheless always vulnerable to the threat of eviction. The
consequence of this insecure tenure was that land occupants constructed with the minimum
amount of cost, and the city stagnated (Goulin, 99). Blancot adds that as land filling increased, an
“ancient law” of eviction (déguerpissement) became applicable, which held that land occupants
were obliged to move off the land (Blancot 1992, 9).
In the 18
th
and 19
th
centuries Cambodia experienced a period of great turbulence, as the
country found itself “between the tiger and the crocodile”—between two (by now) more powerful
neighbors, the Siamese and Vietnamese, who were threatening the country from both sides. For
the Khmers the rivalry between the two bigger powers manifested itself as a “whipsaw” between
Siamese and Vietnamese influence, and as a rivalry between pro-Siamese and pro-Vietnamese
factions at the Khmer court (Chandler 2003, 82). Throughout the 18
th
century Cambodia was
little more than a vassal state of Siam. By the early 19
th
century Vietnamese dominance had
increased, and Cambodia went through a period where it paid dual vassalage to both Siam and
Vietnam (Osborne 1997, 9-10). Cambodia had become weak, and it had lost territory. In the 18
th

century the Vietnamese had annexed parts of the lower Mekong delta (so-called “lower
Cambodia” or Kampuchea Krom), and in the 1790s the Siamese had seized Angkor and the
province of Battambang in the west (Tully, 11). By the mid-1800s the very survival of Cambodia
was in doubt. In desperation, King Ang Duong, who was crowned in 1847, sought to establish an
“ill-defined relationship” with France (Osborne 1997, 12, 30), which had just established a
foothold in neighboring Cochinchina, the southern provinces of Vietnam. Cambodia’s overtures
to France would eventually lead to the imposition of a French protectorate.

43

2.6 Phnom Penh under the French Protectorate
On 11 August 1863 the successor of King Ang Duong, King Norodom, signed a treaty
with the French that marked the beginning of a 90-year French protectorate over Cambodia. The
treaty was a “desperate attempt” by the king to “staunch the hemorrhage of sovereignty” to the
Siamese and Vietnamese that had begun a century earlier (Tully, 11). Under the treaty Napoleon
III extended his protection to the Cambodian king, in return for wide-ranging commercial,
territorial and religious rights. The protectorate was to be overseen by a French Résident
Supérieur, who received the rank of a grand mandarin of the Cambodian court. Despite the
seemingly innocuous nature of the treaty for Cambodia, the Kingdom had become a de facto
colony of France (Tully, 17-8).
Most historians have described the encounter between the Khmers and their new
colonists as a clash between a worldview steeped in centuries of tradition and a rational,
development oriented world view. John Tully, for example, writes that “Cambodians…saw life
differently than did post-enlightenment Europeans.” The French deemed it “self-evident that
their reforms would be very popular with those they considered would benefit from them”. Yet
the Khmers saw their own traditional customs and laws as legitimated by many hundreds, perhaps
thousands, of years of practice. The French approached Cambodia with “orientalist notions”:
their mission was “essentially grounded in the unexamined assumption of European superiority
based on legal-rational authority”. But for the Khmers, however, “Buddhism, with its emphasis
on merit…taught them resignation and held out the hope for a better lot in a later life” (Tully, 69).
The French considered the Cambodian political system to be despotic: the powers of the
King and the nobility over the common people were absolute and unlimited. Corrupt mandarins
were considered to be milking the peasants in order to live in wasteful opulence. Furthermore, the
French considered the Cambodian administrative system to be ramshackle, unjust, wasteful and

44

labyrinthine (Tully, 34, 67-8). These observations, among other factors, led the French to adopt a
“great, sustaining conviction that their presence was both right and necessary” and to be driven to
promise a “better, more civilized life for the ‘native people’” (Osborne 1997, 33). Milton
Osborne claims that this belief in a mission civilisatrice was a dominant theme of French colonial
theory in both Cochinchina and Cambodia—a theme that was upheld by colonial administrators
as well as by historians and missionaries of the time (Osborne 1997, 33, 36).
2.6.1 The First Decades of French Rule (1863 to 1890)
The Cambodian protectorate started as an after-thought for the French: their main priority
lay in neighboring Cochinchina, which offered them far larger rewards. Phnom Penh became the
capital of the new protectorate, but during the French colonial period it would always be eclipsed
by Saigon, which was the capital of the French colonial construct of “Indochina” and a colonial
showcase compared to the relative backwater of Phnom Penh.
The French would not begin to attempt real change in their new protectorate until the
1880s (Tully, 29), mainly because during the first two decades the protectorate lacked financial
powers and civil servants. The first priority of the colonial administrators as well as King
Norodom was to embark on major public works in Phnom Penh, in order to expand the city and
help rid it of its “sordid image” and give it the “appearance of a capital city” (Igout, 7). The
decision to expand Phnom Penh required battling the natural elements, and thus represented a
major break with the Khmer tradition of “composing with water” by which human construction
was essentially adapted to the flow and variations of the water (Grant Ross 2005, 7). To enable
development to take place, the site at Phnom Penh required systematic filling in of land and
construction of dikes, and constant stabilization of river banks and flood control. Moreover, the
rivers had to be regularly dredged to make navigation possible in the dry season (Goulin, 3).

45

There were, however, two major obstacles to these plans during the first two decades of
the protectorate. The first was that the French were dependent on land concessions given to them
by King Norodom, as a result of the ancient Khmer property system whereby the kings were the
sole property owners. The second obstacle was that royal finances at the time were not sufficient
to fill in the land behind the narrow strip of housing along the Tonle Sap required to expand the
city (Igout, 7).
In 1884 the French established the Municipality of Phnom Penh, and with this, the city
acquired its first administrative, financing and planning powers. The same year, the French
abolished the royal monopoly on property ownership in the Kingdom and introduced inalienable
private property rights. For the first time in Cambodia, properties would be registered as
individual parcels and recorded in a cadastre
25
. The Protectorate would henceforth control all
land transfers in the urban area (Igout, 8). Although in principle this enabled the French to buy
and speculate with land, in practice the system of private property rights was only introduced
gradually, so that it was not until the 1920s that significant purchases of land by French
companies and private citizens started taking place, following the adoption of the French Civil
Code in 1920 (Pomonti and Thion, 28-9).
In the city center King Norodom ordered the large-scale construction of shop houses
(compartiments) made of permanent materials, including brick walls and tiled roofs, in order to
modernize his capital. The shop houses combined residential and commercial functions, and were
built particularly for the Chinese market (Igout, 7). Housing conditions in the poorer “native”
quarters, meanwhile, were still abysmal, characterized by a “massing of huts and cabins around

25
Christiane Blancot and Aline Hetreau-Pottier, “1863-1953: Une ville nouvelle dans un site d’occupation
ancienne”, quoted in Atelier parisien d’urbanisme, Phnom Penh: développement urbain et patrimoine, ed.
N. Starkman (Paris: Ministry of Culture, 1997), 32.

46

fetid, stagnant ponds” and piles of uncollected garbage. These conditions were made worse by
demographic pressures, as thousands of new migrants had been attracted to the city as a result of
an increase in economic activity. Numerous contagious diseases flourished in the poor
environmental conditions, and the straw huts of the lower income population were constantly
threatened by fires (Igout, 8-9).
2.6.2 The “Modernization” of the City (1890 to 1906)
The French soon introduced another measure that broke with Cambodian customary law:
the concept of a distinction between the public and private domain. As a result, the separation
between public and private space became the rule. This imposed physical limits whereas before
land in Cambodia had always been considered an open and public space
26
.
In 1890 the French Résident Huyn de Vernéville embarked on an ambitious program to
“modernize” Phnom Penh through large-scale public works and urban planning schemes. Much
effort was invested in the drainage of many of the boeungs behind the riverfront to accommodate
future growth, by digging a semi-circular canal linked to the Tonle Sap river. Another project
widened the main road (the Grand’ Rue) along the Tonle Sap river, so that the city acquired a true
riverfront, with quays along the river. A stretch of land along the banks of the river was declared
public domain and off-limits for construction
27
. This was followed by the first “city
beautification” projects and the construction of public buildings to accommodate the
administration of the protectorate.
By the early 1900s the erstwhile “sordid village” had become a more attractive
administrative and political center, equipped with modern amenities (Igout, 9). A drinking water

26
Ibid.
27
Ibid., 31.

47

network was put in place in 1895, which brought in filtered water from the Mekong. By 1901,
Phnom Penh had an electricity network and public lighting. A rudimentary sewage network was
put in place for parts of the city—although not for the “native” quarters, so that a large segment
of the population continued to resort to traditional forms of sanitation, with the Tonle Sap river
and the boeungs serving as repositories for human waste. During de Vernéville’s tenure Phnom
Penh’s population increased from between 25,000-30,000 to 50,000. Almost half of the city’s
residents were Chinese (Igout, 9).
Urban plans divided Phnom Penh into three separate city districts, by ethnic group. The
northernmost quarter, located around the phnom, housed the French; the central quarter, around
the central market, housed the Chinese; and the southernmost quarter, around the royal palace,
was reserved for the “native” population. Two other districts would be added later: an
“Annamite” quarter to house the Vietnamese, to the west of the Cambodian quarter; and a district
housing Vietnamese catholics, to the north of the European quarter. The ethnic zoning plan
apparently reflected a desire by the king for each ethnic group to manage their own affairs
28
.
2.6.3 Expansion of the City (1906 to 1953)
During the first decade of the 20
th
century, under the reign of Kings Sisowath and
Monivong, Phnom Penh continued to expand westward and southward as more and more boeungs
were filled in, thus creating more land for urban expansion (Igout, 11). Large public works
projects continued. The Mekong and Tonle Sap rivers were made navigable for maritime traffic
up to Phnom Penh. The system of bridges and canals in Phnom Penh, which characterized the
early part of the protectorate, made way for a network of boulevards. The train station was built
and a railway line to Battambang inaugurated in 1932.

28
Ibid., 33-4.

48

The city’s population also continued to expand, reaching 108,000 in 1939. A massive
increase of rural migrants accounted for a large part of the city’s population growth during this
period. A sizeable portion of the migrants were a “floating population” of temporary migrants: in
1921 approximately one-sixth of the population of Phnom Penh did not have a stable home
(Goulin, 18). A major challenge became integrating all these migrants and their rural ways of life
into the fabric of the city. New “villages” of straw huts (paillottes) sprung up all around the
periphery of the city, and along the banks of the Tonle Sap and Bassac rivers. These areas were
classified in urban plans of the authorities as “underprivileged” and “temporary districts”, as they
had no access to the road network or to urban services, such as clean water and electricity
(Goulin, 107)—and presumably they never would have in the future.
Around the turn of the century urban planning developments in France influenced
practices in the French colonies. Public spaces and streets in Phnom Penh were planned or
redesigned following uniform standards applicable at the time in France. Thus, new streets were
planned according to the “Haussmann model” in Paris, i.e., as straight boulevards with a standard
width of 20 meters, flanked by pavements and trees. Public buildings copied architectural models
in vogue in France
29
. Under the 1919 Cornudet Law in France, French cities had to put in place
urban expansion and city beautification plans. This law also applied to the French colonies in
Indochina. The first plans under this law in Phnom Penh were developed in 1924, and legal
frameworks and city planning regulations for their implementation were adopted in 1928, in line
with French legislation
30
.
In 1950, just before independence, the first master plan (“controlling plan”) was
prepared for Phnom Penh by the Municipality’s Urban Technical Service. The plan was used as

29
Ibid., 32-3.
30
Ibid., 35.

49

a guide for urban regulations adopted after independence in 1956, and it became the main
strategic instrument guiding urban development—and large urban projects—from the time of
independence through the 1970s (Igout, 15).
As Phnom Penh’s economy expanded, and as land in the city center grew more scarce
and expensive, two developments took place that would be echoed at the end of the 20
th
century:
these were large-scale evictions of the urban poor and the beginning of active land speculation.
In order to implement their urban plans and carry out their public works projects, the
French authorities carried out a “systematic policy of evictions” (déguerpissements) (Goulin,
109). Most of the evictees were urban poor residents living in the “underprivileged” and
“temporary districts”, principally along the banks of the Tonle Sap and Bassac rivers, where the
straw huts of the poor had to make way for the construction of a deep water port, jetties, an
industrial zone, and a river port (Goulin, 109-10). Starting in 1915 the Municipality of Phnom
Penh planned several resettlement sites (referred to as “reserves”) outside the city center, where
people displaced by evictions “necessitated by the application of urban plans” would receive plots
of land where they could build their own houses, following regulations laid down by the
Municipality (Goulin, 107). In addition to the original evictees, the planned reserves later also
accommodated “spontaneous” arrivals—rural migrants to the city who housed themselves
directly in the reserves instead of first squatting in the city center (Goulin, 109). The planned
“reserves” were quite possibly the first examples of resettlement sites for the urban poor in
Phnom Penh
31
.
On 31 October 1924 the colonial authorities introduced an order (arrêté) that prohibited
foreigners (particularly the Chinese) from owning private property. This kicked off the first

31
It is not clear from historical records how many of these planned resettlement villages were ever in use,
or how many residents actually moved there.

50

large-scale land speculation activities in Phnom Penh. Christian Goulin (1966) argues that the
arrêté created an instant market for Cambodians to purchase plots of land at low prices and sell
them at higher prices to foreigners. The local speculators subsequently settled in the urban
periphery, where land was still cheap (Goulin, 113).
2.6.4 The Legacy of the “Mission Civilisatrice” in Phnom Penh
During the 90 year interval of the French protectorate the Cambodian capital had evolved
from what early travelers described as a mere string of huts, surrounded by water, to a “modern
city in the making” (Igout, 14), built on top of boeungs that French engineers had progressively
filled in over the years. In his assessment of the French legacy in Cambodia, Norodom Sihanouk
argued that the French had put an end to feudalism and, for the first time in Khmer history,
rendered all Cambodians equal before the law. This included peasants: as a result of the private
property regime introduced in 1884, peasants were given deeds of ownership for the land they
farmed (Jeldres, 40). John Tully suggests that the protectorate delivered several other positive
achievements for Cambodia; at least three of these directly benefited the capital city
32
. First, the
French ensured a period of protracted peace and stability, which provided the foundation for the
capital’s growth. Second, colonial authorities planted the seeds of a modern city, through
investment in physical improvements, including public works projects, roads and
communications. And third, through its mass vaccination program the protectorate helped to
prevent the spread of endemic diseases, such as smallpox, cholera, dysentery and the plague
(Tully, 487-9).

32
In addition to the three achievements mentioned, two others included the preservation of Cambodian
sovereignty and political unity, and the restoration of certain aspects of Khmer culture, most notably the
restoration of Angkor’s glory (Tully, 487).

51

More broadly, however, several Western historians argue that the mission civilisatrice
failed in one of its key objectives: to improve the plight of the poor by changing the nature of
power relations in Khmer society. Even if the French abolished serfdom by law, in practice
however, David Ayres (2003) argues that the French authorities “reinforced the country’s
hierarchical social order and its associated concepts of power, authority, and patronage” by
choosing to align themselves with the monarchy (Ayres, 187). John Tully goes even further,
arguing that the protectorate “probably strengthened the existing tendencies in traditional
Cambodian society towards autocracy and the exclusion of the general population from decision-
making” (Tully, 488). They did so, among others, by not allowing elections and by “never
regarding themselves as accountable to the Khmer people” (Tully, 490-1).
David Ayres argues that the fruits of the protectorate’s mission civilisatrice were limited
to Phnom Penh: while the infrastructure investments, building programs and public works
schemes gave Phnom Penh a taste of the future, “this path to the future hardly extended beyond
Cambodia’s capital”. The countryside, where the majority of the population lived, was largely
ignored and left unchanged (Ayres, 185-6). The result was the start of a growing divide between
Phnom Penh and the rest of the country.
John Tully suggests that, even in Phnom Penh, the fruits of development were not evenly
spread and that, in certain sectors, the “indifference and parsimony” of the colonial authorities
made a mockery of the pretensions of the mission civilisatrice (Tully, 228). With the exception
of the vaccination program, one of these neglected sectors was public health. In Phnom Penh,
environmental conditions in the housing areas of the “indigenous population” remained abysmal.
An annual report from the Phnom Penh municipal hygiene service in 1923 reported that the city’s
rapid growth made it impossible for the authorities to provide sufficient housing for the
population. The result was severe overcrowding in the “indigenous quarters”. Landlords

52

typically partitioned houses and shop house floors into several rooms that lacked light and air, in
contravention of the protectorate’s housing regulations
33
. A report by a colonial inspector on
sanitary conditions in 1930 mentions that while the straw and bamboo houses of the poorer
classes were in good condition when they were new, they rapidly deteriorated with age, and took
on a “lamentable aspect”. Hygienic conditions within the homes were described as shocking, and
conducive to diseases such as dysentery and cholera
34
. The colonial authorities found it difficult
to tackle the housing problems of the poor. The municipal hygiene service cited internal failures:
the agency lacked sufficient staff to undertake regular controls; the sanctions which it meted out
to landlords were “derisory”; and agency staff were often themselves lax in enforcing the rules.
Moreover, the hygiene service indicated that landlords showed “passive resistance” by simply
flouting housing regulations
35
. An annual report from the following year noted that the hygiene
service faced enormous resistance from the “forces of inertia” among the local population
36
.
Another legacy of the French protectorate in Phnom Penh was a city increasingly divided
along ethnic and socio-economic lines. As a result of several decades of planning policy and
widespread evictions, as well as more “spontaneous” developments such as ongoing rural
migration and land speculation, the central areas of Phnom Penh were inhabited principally by
Chinese (as well as some Vietnamese and Europeans), while the outlying urban and peri-urban
areas were inhabited by Khmers. Most, though not all, Khmers living in the urban periphery

33
Ville de Phnom Penh, Rapport annuel du Service d’Hygiène et de prophylaxie générale au Cambodge, in
National Archives of Cambodia (Phnom Penh: Protectorat du Cambodge, 1923), 2-3.
34
Directeur Local de la Santé, Rapport au Monsieur l’Inspecteur des Services Sanitaires et Médicaux de
l’Indochine, in National Archives of Cambodia (Phnom Penh: Direction de la Santé, 1930), 14-5.
35
Ville de Phnom Penh, Rapport annuel du Service d’Hygiène et de prophylaxie générale au Cambodge, in
National Archives of Cambodia (Phnom Penh: Protectorat du Cambodge, 1923), 4-5.
36
Ville de Phnom Penh, Rapport annuel du Service d’Hygiène et de prophylaxie générale au Cambodge, in
National Archives of Cambodia (Phnom Penh: Protectorat du Cambodge, 1924), 11.

53

were poor and still engaged in agricultural activities (Goulin, 114-5). At the end of the 19
th

century Prince Norodom Yukanthor blamed the introduction by the French of private property
rights in 1884 for the increasing socio-economic divisions in Phnom Penh society:
The people suffer…from the transfer of power to the Résident Supérieur. Before,
all Cambodian lands belonged by law to the King. De facto, land belonged to he
who occupied it…this was a consequence of obligations imposed on the throne
by Buddhist law: the land of God, entrusted to the King, put by him at the
disposition of all men who need it. And this without restrictions of any kind.
You have established property…you have created the poor
37
.
On 13 March 1945, as Japanese forces occupied Phnom Penh, King Sihanouk declared
independence. With France under pressure in the Second World War, French rule in Indochina
had effectively come to an end (Chandler 2003, 170, 173), even though the French would duly
return after the war. It would take eight more years before Cambodia could celebrate true
independence from France. Phnom Penh’s most glorious days lay ahead, as the capital city of a
newly independent nation. But the specter of war loomed on the horizon, and this was to have
big consequences for the capital.
2.7 Post-Independence and the “Sangkum” Period
On 9 November 1953 Phnom Penh officially became the capital of an independent state.
Under the stewardship of its new Head of State, King Norodom Sihanouk
38
, Cambodia entered a
seventeen-year period often described as a “golden age” for the country: a time of relative

37
Prince Norodom Yukanthor, extracts from personal diary quoted in Ministère de l’Information du
Gouvernement Royal du Cambodge, “Cambodge” (Hong Kong: n.p., 1962), 49, quoted in Christian
Goulin, Phnom Penh: Etude de géographie urbaine (Paris: Centre de Documentation et de Recherches sur
l'Asie du Sud-Est et le Monde Indonésien, 1966), 115.
38
A note on terminology: At different periods in history Norodom Sihanouk has carried the title of King
(1953-1955; 1991-2005), Prince (1955-1975) and (his present-day honorific, after the accession to the
throne of his son, King Norodom Sihamoni) “King-Father” (2005 to the present).

54

stability and prosperity before the outbreak of another long period of warfare and political
upheaval.
In March 1955 Norodom Sihanouk abdicated the throne in favor of his father, Prince
Norodom Suramarit, and established a new political platform, the Sangkum Reastr Niyum (widely
referred to simply as the Sangkum movement), defined by Sihanouk in his own words as a
“people’s community”
39
that would be his “chief instrument in forging the national unity which
[he] had set as [his] goal”. Sihanouk declared that the Sangkum movement would transcend all
political parties and ideologies and “give birth to a truly democratic, equalitarian [sic], and
socialist Cambodia, to restore the past greatness of [the] motherland” (Jeldres, 55).
André Tong (1972) suggests that Sihanouk’s reasons for abdicating were less noble than
he made appear, and that his aim in forming the Sangkum was in fact to dominate political life in
Cambodia (Tong, 74-6). Within months several political parties had dissolved their organizations
and advised their members to join the Sangkum (Jeldres, 55). At the September 1955 legislative
elections the Sangkum movement won all seats in the National Assembly. In December 1957 a
constitutional amendment proclaimed that the National Assembly should henceforth address only
those matters that were raised by six-monthly national congresses, which were held under the
auspices of the Sangkum (Osborne 1994, 98). Behind a façade of democracy, political life in the
post-independence period thus came to be dominated by one man (Tong, 83).
2.7.1 An Age of Benevolent Paternalism
For ordinary Cambodians, the Sangkum regime promised a radical break with the past.
After centuries of “remote power indifferent to the needs of the people”, from the reigns of the

39
The Sangkum Reastr Niyum is commonly translated as “People’s Socialist Community” (Serge Thion
and Jean-Claude Pomonti, 1971; André Tong, 1972; Bernard Hamel, 1993; Milton Osborne, 1994; David
Chandler, 2003; David Ayres, 2003; Ian Harris, 2005), and sometimes also described as “Buddhist
socialism” (Charles Meyer, 1971; David Chandler 2003).

55

semi-divine Angkorian kings to the technocratic French protectorate, Sihanouk, as President of
the Sangkum, offered a new discourse of “benevolent paternalistic authority”
40
. Sihanouk
emphasized his identification with Cambodia’s poor, whom he called the “little people” (reas
touch tach)—the “loyal Khmer population” for whom he declared to feel a “paternal affection”
(Jeldres, 39-40). In return the “little people” (particularly older people) venerated Sihanouk as
someone with “supernatural powers” and “agreed to play supporting roles” (Chandler 2003, 200).
Sihanouk claimed that he abdicated the throne in order to abandon its “pomp and pageantry” and
instead, to “devote [his] whole time and energies to serve the people and their well-being”
(Jeldres, 54).
The Sangkum instituted several unprecedented pro-poor measures. Perhaps the most
remarkable was the system of semi-annual national congresses, which Sihanouk established to
give the people a direct role in the decision-making process (Jeldres, 58). The congresses, held in
Phnom Penh, were open to all members of the public—both men and women. Many attendees
were bused in from the provinces (Osborne 1994, 98). Their format was always the same:
following a presentation on the issues for discussion by Sihanouk, as President of the Sangkum,
people were invited to ask questions, and offer their comments, suggestions or criticisms, and
afterwards, to vote on the issues presented through a show of hands. Over the years the national
congresses introduced some historic measures with direct relevance to the poor, such as
institutionalization of the vote for women; the use of Khmer (not French) as the language of
public institutions; and the possibility to remove legislators from office (Meyer, 138). The
national congresses remained a fixture of Cambodian political life until the end of 1969, although

40
Smith, Roger, “Cambodia”, in Governments and Politics of Southeast Asia, ed. G. Kahin (Ithaca, NY:
Cornell University Press, 1964), quoted in May Ebihara, Svay: a Khmer Village in Cambodia (Ph.D. diss.,
Columbia University, 1968), 545.

56

Charles Meyer argues that after 1966 they had lost their “democratic spirit” and had become
approval mechanisms for the increasingly autocratic decisions of Sihanouk (Meyer, 138-9).
Another democratic innovation of the Sangkum period was a system of provincial
assemblies, which Sihanouk had established (through a constitutional amendment in 1957) as
consultative bodies to link the rural population and central government, and to encourage
peasants to participate in public affairs. After only two years, however, the provincial assemblies
ceased to be operational. Charles Meyer argues that this was because of the lack of political
maturity of the masses and the opposition of provincial cadres and “mandarins” who refused to
let themselves be controlled by the new assemblies (Meyer, 139-40).
Other pro-poor features of the Sangkum movement were Sihanouk’s large budgetary
allocations to education (amounting to over 20 percent of the national budget in some years), and
the frequent tours Sihanouk made to the countryside, which put him in touch with hundreds of
thousands of ordinary people (Chandler 2003, 199).
But while the Sangkum movement gave ordinary Cambodians unprecedented
opportunities to participate in political life, and arguably offered the poor more “access” to power
than at any other prior period in Cambodian history, Sihanouk’s benevolent paternalism at the
same time borrowed heavily from themes in ancient Khmer history. These references to the past
represented a return to the status quo, and undermined any budding sense of citizenship poor
Cambodians might have felt. For Bernard Hamel (1993) Sihanouk remained above all a God
king or devaraja in the purest Angkorian tradition, despite his abdication from the throne. As a
result of his growing “demagogy, absolutism and repression” Sihanouk kept the Cambodian
people at a stage of evolution that, politically at least, resembled the 12
th
century (Hamel, 29-30).
Other historians have remarked that, for all his attempts at democratic reform, Sihanouk
remained a Khmer monarch in the ancient tradition. Jean-Claude Pomonti and Serge Thion noted

57

that Sihanouk surrounded himself with the kind of veneration expected by ancient Khmer kings.
People were afraid to touch him, and could only approach him on their knees (Pomonti and
Thion, 35). According to Charles Meyer, Sihanouk wanted to serve the people and be very close
to them, but at the same time he did not want [others] to forget that he was a prince and former
king (Meyer, 143). For André Tong, the Sangkum regime took to new heights the ancient theme
of paternalism in Cambodian history. Sihanouk gave himself the titles of “Lord father” (Samdech
Euv) and, following his abdication, “Lord the ancient king” (Samdech Preah Upayuvareach). As
the titles indicate, Sihanouk considered himself “father of the nation”; the Cambodian people
were his subjects and “children”. Tong suggests that Sihanouk’s paternalism was motivated not
so much by affection for his subjects, as Sihanouk considered Cambodians to be “lesser beings”,
but by the need to impose his own personal political agenda (Tong, 76).
For Charles Meyer, the Sangkum movement was ultimately reduced to “Sihanoukism”.
What started off as a platform to institutionalize an “egalitarian and socialist democracy” in
Cambodia ended up “limiting all perspectives for the development of political thought” as
members were expected to give their unconditional adhesion to Sihanouk. Meyer concludes that
the Sangkum’s “Buddhist socialism” ultimately boiled down to no more than “a few traditional
charitable practices” (Meyer, 150-1). Moreover, Meyer highlights the broader clash between
Sihanouk’s ambitious social program and the legacy of ancient Khmer institutions: up until 1960,
Sihanouk sincerely wanted to give power to the people but the new, audacious forms of
grassroots democracy he attempted to institutionalize were ultimately incompatible with the
“medieval, immovable monarchy” that he represented (Meyer, 142). Public perceptions and
expectations were perhaps also resistant to change: ordinary Khmers never forgot that Sihanouk
was a descendant and heir of the kings of Angkor, and therefore took after the gods (Meyer, 148).

58

2.7.2 An Urban Renaissance in Phnom Penh
During the Sangkum period Phnom Penh transcended once and for all its status as
colonial city under the shadow of Saigon and became a city in its own right—the capital city of a
proud new sovereign nation, an “administrative and symbolic hub of the kingdom…industrious
and commercially vital, projecting an image of a country that was undergoing rapid expansion”
41
.
Norodom Sihanouk had ambitious plans to turn Phnom Penh into a capital city “worthy
of Angkor Thom”
42
, the walled Angkorian city constructed during the reign of King Jayavarman
VII starting in the latter part of the 12
th
century. With this aim Sihanouk embarked on a grand
program—dubbed by André Tong as a “politics of prestige” (Tong, 85)—to plan and implement
public works schemes and urban projects for Phnom Penh in the areas of infrastructure, culture,
city beautification, sports, and housing. The Sangkum period inspired its own architecture (a
“new Khmer architecture”), which was a modernist style suited to the post-colonial nation. A
new national identity had to be developed, and in the capital and a few other cities in Cambodia
this was achieved through the construction of emblematic modern buildings, adapted to the
Khmer context, including national monuments, universities, ministries and a national sports
complex (Vann Molyvann 2003, 157).
Among the largest urban schemes during the Sangkum period were infrastructure
projects. Phnom Penh became the center of the road network in Cambodia, and a new rail line
connected Phnom Penh with the deep-sea port at Sihanoukville. In Phnom Penh the first bridge
across the Tonle Sap river was constructed, as was a new bridge to replace the old Monivong
bridge across the Bassac river. The largest single urban project in Phnom Penh at this time was

41
H.M. Norodom Sihamoni, King of Cambodia, quoted in preface of Helen Grant Ross and Darryl Collins,
Building Cambodia: New Khmer Architecture 1953-1970 (Bangkok: The Key Publisher, 2006), xvii.
42
H.M. Norodom Sihanouk, Urbanisme et Tourisme, Album No. 6 of a series on the achievements of the
Sangkum, Reastr Niyum (Phnom Penh: Rama Printing International, 1994), 1, 5.

59

the reclamation and development of the Bassac riverfront, a large area at the confluence of the
Tonle Sap and Bassac rivers, starting in 1964. The riverfront was planned as a residential and
cultural zone of Phnom Penh, and was to accommodate apartments, an exhibition hall, a museum,
a school of art, music and classical dance, a municipal theater, a tourism authority, an art gallery,
a large international hotel, a nautical club and a marina
43
.
After independence the physical expansions of Phnom Penh through the filling in of
boeungs had largely ceased. In the 1950s and 1960s the city became denser and there was an
increased construction of mid-level buildings. Modern shop houses of three or four floors were
built across the city. There was an expansion of public facilities, including new hospitals,
schools, and universities (Igout, 16). New cultural landmarks appeared, such as the Preah
Suramarit Theater (also known as the Tonle Bassac Theater) on the new Bassac riverfront,
designed by the new Khmer architect (and minister in the Sangkum regime at the time) Vann
Molyvann and constructed in 1966, with a seating capacity of 1,200.
In 1963 Sihanouk started a campaign to develop sports in the Kingdom. This entailed the
construction of forty provincial stadiums and hundreds of athletic fields (Meyer, 166-7). After
Phnom Penh was named host city for the Third Southeast Asian Games in 1964 a new National
Sports Complex was designed by Vann Molyvann, which included an “Olympic stadium” that
could seat 60,000 people. The stadium was inaugurated in December 1964. The project led to
the rapid economic development of the surrounding area of Tuol Svay Prey (Igout, 16), with
public investments in roads and public markets, and a new water treatment plant.

43
The various construction projects were delivered in stages. Construction of the large hotel (currently the
Cambodiana Hotel) was interrupted by the Khmers Rouges invasion in 1975, and completed in the 1990s.
Source: Helen Grant Ross, ARK exhibit on post-colonial Khmer architecture (Phnom Penh: Cambodiana
Hotel, 2006).

60

For the first time in Cambodian history, housing was planned for low-income segments
of the population, principally lower-level civil servants. The Sangkum declared that “the city’s
beautification should directly benefit the population”
44
, and in this spirit, several medium-rise
complexes of low-cost rental apartments were planned and constructed on the newly reclaimed
Bassac riverfront. These projects included apartment buildings for staff of the National Bank of
Cambodia (currently part of the Russian Embassy compound); an apartment building housing
staff of the Municipality of Phnom Penh (currently known as the “white building” or “Boding”);
and the “Olympic Village Apartments”, originally designed by Vann Molyvann to house
participants in the planned Southeast Asian Games in 1963. When these Games fell through, the
apartments were sold off to civil servants and to the Ministry of Education to accommodate
teachers (Grant Ross and Collins, 18-21).
In addition, the National Bank of Cambodia commissioned housing for its staff outside
the city. Between 1965 and 1967 an ensemble of over 100 low-cost houses were delivered for its
staff as an experiment in “social housing”, also designed by the Khmer architect Vann Molyvann.
The design concept was inspired by the traditional Khmer house: the houses were built from
permanent materials (mostly timber) that could be taken down and moved easily (“Khmer houses
can walk”), but improved using 20
th
century construction techniques. The houses were laid out in
six rows and staggered so that the whole felt less like a grid development and more like a true
village
45
. The houses were rented out for a period of about 20-25 years, after which they were
privately owned.

44
H.M. Norodom Sihanouk, Urbanisme et Tourisme, 16.
45
ARK Research. “Cultures of Independence”, in Papers of the 4
th
Socio-Cultural Research Congress
(Phnom Penh: Royal University of Phnom Penh, 2001).

61

Phnom Penh became known as a “garden city” and as the “pearl of Southeast Asia”,
thanks to numerous city beautification projects. Grand, tree-lined boulevards were planned. A
large public park appeared south of the Royal Palace, and the canal running east-west, just to the
south of this park, became a boulevard bisected with a garden filled with fountains (the present-
day Hun Sen gardens) (Vann Molyvann 2003, 165). An urban refuse collection system was put
in place (Igout, 17), and a municipal Cleanliness and Beautification Committee established.
As part of the beautification campaigns, the new municipal committee issued three public
edicts (prakas) in September 1958 aimed at educating city people (especially the urban poor)
about how to live properly in the city (see Boxes below)
46
. The public edicts were noteworthy,
first of all, for their paternal and patronizing tone, with the “father of national independence” (i.e.,
Norodom Sihanouk) scolding and admonishing his children for failing to maintain appropriate
standards of public hygiene and morality. Second, the edicts demonstrated that, despite the
Sangkum’s identification with the poor (and particularly poor peasants), peasant manners and
peasant dress, as well as the economic activities of the poor, such as vending, were deemed out of
place and inappropriate in the modern, post-independence capital city.

46
All three edicts are reproduced in Bou Saroeun, “Déjà Vu: Plans from the Past”, Phnom Penh Post, 30
March - 13 April 2001: 8-9.

62

This is to inform all brothers and sisters of the nation that … the father of national independence
has criticized the dirty condition of Phnom Penh, of which foreign newspapers have also spread
gossip. Therefore, the King has decided to form a committee to ensure all city streets and
buildings are kept clean and attractive. To assist the committee in its duties, please observe the
following:
Do not litter;
Do not urinate/defecate in public;
Do not build structures for public vending on sidewalks or on land that you have no deed or
rights to;
Those structures built in improper places must be immediately removed.
The successful clean-up of Phnom Penh will only go smoothly with the assistance of the people.
Do not allow your interests to hurt the reputation of the King or the Khmer people. Those who
do not respect this edict will be punished according to the law as a model for others.
Box 1: Public Edict No. 1, Phnom Penh Cleanliness and Beautification Committee, 1958

Cambodia today has achieved independence by our respected King’s hand. Now, the king
understands that if the country is no longer colonized, the people who own the country need to
act with dignity and be well-dressed in order to be seen as “free people”. Therefore, from now
on all people must wear proper clothes in public. When you leave your home at any time
observe the following:
Do not wear kramas;
Do not wear sarongs;
Do not leave your home naked;
Parents must ensure that children are properly dressed before leaving the house.
Those that do not obey this edict will be arrested and educated in understanding the value of
being free people in an independent country.
Box 2: Public Edict No. 2, Phnom Penh Cleanliness and Beautification Committee, 1958

Please ensure the front of your house is attractive and clean:
Plant flowers in the front;
Do not build any unattractive structures on or adjacent to your property;
Do not hang clothes to dry outside your home;
Householders are responsible for the cleanliness of the sidewalk in front of their homes;
No goods may be sold on sidewalks.
Box 3: Public Edict No. 3, Phnom Penh Cleanliness and Beautification Committee, 1958

63

2.7.3 Prelude to Disaster
Even as Phnom Penh was enjoying a prolonged renaissance under the Sangkum regime,
by the late 1960s external and internal troubles were threatening Cambodia’s fragile peace. On
the international front, Sihanouk’s Cambodia was being increasingly drawn into the second
Indochina war, as the United States escalated its war against the Vietcong and threatened to
invade Cambodia in order to halt what it perceived as Sihanouk’s tacit support for the North
Vietnamese.
On the domestic front, Cambodia’s economy had begun to falter. Rural Cambodians
were under severe pressure due to a combination of low crop yields, poor irrigation, excessive
interest charged on loans to farmers, and government indifference (Chandler 2003, 201).
Meanwhile, land holdings were shrinking at the bottom end of the social scale as urban classes,
enriched with foreign capital, bought up more and more rural land from ruined peasants, and as
the disparity in terms of landholdings among peasants themselves continually widened (Kiernan
and Boua, 5). Organized peasant resistance was on the increase. A turning point occurred in
1967, when peasants near the village of Samlaut in Battambang province rebelled against land
seizures by officials and military officers. The rebellion was put down violently by the army:
thousands of villagers had to flee their homes for months, and thousands more were killed. The
Samlaut rebellion manifested “deep political divisions within Cambodian society”, and it led to
further organized peasant resistance across the country that continued to be forcefully repressed
(Kiernan and Boua, 5, 166). For Ben Kiernan and Chanthou Boua (1982), the Samlaut rebellion
demonstrated that a “revolution was on its way”. Popular resistance against Sihanouk’s regime
was growing: hundreds of teachers, students, professors and workers were leaving their urban
positions to join a “mushrooming resistance”. Cambodia’s outlawed and dissident Left, led by

64

the Khmers Rouges, was encouraged in its revolutionary plans for Cambodia’s countryside
(Kiernan and Boua, 5, 181-2).
The end of the Sangkum regime came from the Right of the political spectrum, in the
form of a coup d’état against Sihanouk by his General, Lon Nol, on 18 March 1970. Cambodia
became a republic in October of 1970, and Sihanouk took command of a united front
government, which was allied to North Vietnam and comprised many of the Khmer communists
his regime had been fighting until recently. The new right-wing leadership in Phnom Penh
engaged in two unsuccessful offensives, with the support of the U.S. and South Vietnamese, to
try and rid Cambodia of North Vietnamese forces. This was followed in 1973 by massive U.S.
bombing raids that dropped over a hundred thousand tons of bombs on the Cambodian
countryside in a further effort to dislodge Vietcong troops (Chandler 2003, 205-7).
During the next five years Cambodia found itself embroiled in a civil war featuring, on
one side, a Leftist revolution in the countryside led by the increasingly confident Khmers Rouges
and, on the other side, the increasingly weak government of the new Khmer Republic, which was
reduced to the urban areas, particularly Phnom Penh. From 1973 the Khmers Rouges were
already busy “restructuring Khmer society” in the countryside, organizing the population into
strict cooperatives and production teams or “mutual aid teams” of twenty or more families to
work the land, and organizing youth groups to monitor society (Becker, 148-52).
The growing turmoil resulted in an unprecedented demographic explosion in Phnom
Penh. It is estimated that the population of the city swelled from 900,000 in 1970 to almost two
million inhabitants in early 1975 (Igout, 18; Chandler 2003, 108), as refugees flocked to the city
to escape the spillover of armed conflict from the Vietnam War and the leftist revolution in the
countryside. At first, “refugee villages” were created on the outskirts of the city to accommodate
the influx of desperate people (Igout, 18). But by 1973 the new government had lost control, and

65

migrants began squatting in the city’s public gardens and anywhere they could find space,
including rooftops, riverbanks, sidewalks, and in makeshift shelters. Under the circumstances, all
the Sangkum’s urbanization initiatives in Phnom Penh were halted. Most large urban projects
were stopped. A 1970 Master Plan of Phnom Penh produced by the United Nations Development
Program was never implemented because the city found itself in a dysfunctional state.
Phnom Penh was a city increasingly cut off from the surrounding countryside, both
politically and economically. From 1972 on there were rice shortages in the city. Elizabeth
Becker (1998) argues that the Khmer Republic’s policy toward civilians was one of “criminal
neglect”. At roadblocks set up on the city’s perimeter, army officers and subordinates “fleeced”
refugees from the countryside who tried to enter Phnom Penh (Becker, 152). While in the rural
areas the Khmers Rouges “state” or Angkar attempted to distribute rice as equitably as
circumstances allowed, and managed to instill an ideology that “everyone was in the war
together”, in Phnom Penh by contrast the Khmer Republic government presided over a city
characterized by stark inequalities, and little sense of reality:
A hungry, nearly starving refugee population [lived] next to a fat and rich elite
and…a middle class that stumbled along, trying to stay above the poverty line
while maintaining the pretense that nothing was dramatically wrong, and
carrying on daily affairs with little mention of the ‘war’ (Becker, 152).
By mid-1972, observers in Phnom Penh were convinced that a victory of the Left was
inevitable unless there was thorough reform of the government (Vickery 1982, 110-1). But the
new Khmer Republic was doomed to collapse: it was propped up by U.S. military aid but
internally weakened by leadership disputes and corruption. On the first day of 1975 Khmers
Rouges forces began a coordinated offensive to take Phnom Penh (Ponchaud, 11). A little over
four months later they succeeded in their goal.

66

2.8 The Democratic Kampuchea Regime
By the time Khmers Rouges forces entered Phnom Penh, on 17 April 1975, they were
initially welcomed as liberators by many of the city’s beleaguered residents. In his famous eye-
witness account of the take-over of Phnom Penh by the “men in black”, the missionary François
Ponchaud (2001) describes the palpable sense of relief among the population that the corrupt
Khmer Republic regime was finished, and with it (they assumed) the food shortages, the
blockades of the city, the daily rocket fire, the random killings, and the forced conscription that
had marked residents’ lives during the early 1970s. Peasants who had fled to the city believed
they could finally return to cultivate their fields. The patriarch of the Buddhist community
announced that the war was over, and that people were now “among brothers” (Ponchaud, 16).
But relief soon turned to despair: the newly arrived Khmers Rouges forces ordered all
residents to evacuate Phnom Penh “for two or three days” under the pretext that the American air
force was going to bomb the city. Within a day or two a city whose population had swollen to
almost two million was almost deserted. Other cities in Cambodia were similarly evacuated
during this time (Kiernan 1996, 49-51). In Phnom Penh, water and electricity service were cut
off to make sure that no inhabitants would remain in the city (Ponchaud, 30). It soon became
clear that the evacuations were not temporary, but an essential plank of the revolution. Phnom
Penh was to remain virtually depopulated for the next 3 years and 8 months, as its inhabitants
were sent to the countryside to become a rural proletariat.
2.8.1 A “Clean” and “Pure” Revolution
Two months after the capture of Phnom Penh, Pol Pot
47
(then the chairman of the
communist party’s military committee, later Prime Minister) gave a major address to units of the

47
The nom de guerre of Saloth Sar.

67

revolutionary army of the Communist Party of Kampuchea (CPK). He announced that the
revolution represented an unprecedented victory over imperialism in the 2,000 year history of
Cambodia. The Cambodian nation, the Cambodian people, the Cambodian revolutionary army,
and the CPK had managed to “liberate their own country and people completely, definitively and
cleanly”. The Khmer people would henceforth be independent and self-reliant (Becker, 26-7).
The leadership of the new regime confirmed eight key principles guiding their revolution
(Kiernan 1996, 55). First, the regime had to evacuate the population of all towns. Second, the
regime had to abolish all markets, in order to wipe out all commerce. Third, all money also had
to be destroyed; nobody was to be paid for his or her labor (Becker, 28). Moreover, money and
markets would only risk further exchange with, and dependence on, foreigners. Fourth, all
Buddhist monks had to be defrocked and put to work growing rice; religion and expressions of
religion were to be prohibited. Fifth, all leaders of the previous Lon Nol regime had to be
executed, beginning at the very top. Sixth, high-level cooperatives had to be established
throughout the country, where people ate communally. Seventh, the entire Vietnamese minority
population had to be expelled from Cambodia. And finally, troops had to be dispatched to the
borders, particularly to the Vietnamese border.
For ordinary Cambodians the face of the new regime was a shadowy organization called
the Angkar (the “organization”), which was the name the CPK gave itself. The Angkar enjoyed a
nearly “mystic omnipotency; its word was law, and any attempt to break it was always
discovered” (Becker, 141). It would take several more months before the population began to
obtain a clearer picture of the official structure of the new regime that had taken over. In January
1976 a new constitution was passed, which renamed the country Democratic Kampuchea (DK)
and created a 250-seat National Assembly for which elections were held in March 1976 (O’Kane,
735-6). The DK constitution abolished the monarchy and “reactionary” religion. It proclaimed

68

the existence of only three categories of citizens: workers, peasants and soldiers. All property
became “collectively owned”, and was claimed by the state (Becker, 206). Private property was
abolished; cadastral records and land registers were destroyed. Later in 1976 Pol Pot emerged as
prime minister and Khieu Samphan became president of a new state presidium.
Various rationales have been advanced to explain the Khmers Rouges’ evacuation of the
cities. The DK regime’s own propaganda offered different explanations, one of which was that
city people were exploiters and immoral, in contrast to people of the liberated areas (Kiernan
1996, 62), and therefore they needed to be re-educated. Another official explanation was that city
people had to be moved in order to make sure they would have enough to eat. Some historians
concur that, in a context of widespread food shortages, city dwellers were evacuated so as to
move them closer to food sources in the countryside
48
. For Michael Vickery (1999) there is a
“good deal of truth in the DK contention that only the evacuation could save the city population
from worse starvation than it had already known” (Vickery 1999, 85). For Ben Kiernan (1996)
the evacuations were “devised neither as a punishment for the city dwellers nor as a solution for
their problems”. Instead, they were part of a “calculated, strategic move in an ongoing military
contest”. The Khmers Rouges recognized that they had less support among the urban population
for their class-based revolution, and therefore they determined that city dwellers could not be
trusted. The evacuations (and subsequent killings) got rid of “unapproved racial groups”,
principally the Chinese and Vietnamese, and by dispersing the urban population they made city
dwellers far easier to control (Kiernan 1996, 62-4). For François Ponchaud the rationale for the
urban evacuations was primarily ideological. Through their “clean” and “pure” revolution the
Khmers Rouges aimed to turn the clock in Cambodia back to year zero and create a totally new

48
This theory was advanced by George C. Hildebrand and Gareth Porter, Cambodia: Starvation and
Revolution (New York: Monthly Review Press, 1976).

69

society, one which was to be egalitarian and rural (Ponchaud, 227). In this new society cities had
no place. Even the very idea of the city had to disappear, as Khmer cities (particularly Phnom
Penh) had developed as a result of forces profoundly hostile to the “pure”, “clean” revolution,
such as French colonialism, Chinese commerce, and the statist bureaucracy of the monarchy and
the Khmer Republic. To prove his point that the urban evacuations were primarily ideologically
based, Ponchaud points out that as early as 1972 Khmers Rouges forces were already evacuating
to the forest the inhabitants of villages and provincial towns that they had occupied. Moreover,
they often burned the houses of evacuees, in order to prevent their return (Ponchaud, 34).
Despite the DK regime’s ideology of starting at zero, there were nevertheless echoes of
the past. First, Elizabeth Becker (1998) argues that the DK regime “attempted paradoxically to
take the place of the god-kings” (devarajas) of the ancient Khmer past. The regime also
benefited from the Khmer “cultural heritage of accepting all-powerful rulers” (Becker, 189).
Moreover, the DK regime regularly invoked the greatness of ancient Angkor when referring to its
plans for “building a new Cambodia” (Becker, 188).
Second, Vickery (1999) and Chandler (2003) argue that the power structure of the DK
regime, as in the Angkorian era, was based on an “Asiatic mode of production”, in the Marxist
sense, whereby a “large oppressed class” (in the case of the DK regime, this comprised former
city dwellers and intellectuals) was “regimented for the construction of new infrastructural
installations, mainly water works” (Vickery 1999, 287; Chandler 2003, 54). But whereas for
Vickery the “Asiatic mode of production” in the Angkorian era was a more organic and socially
accepted process, characterized by the “slow accretion of custom seen as the inevitable and right
way for the world to function”, during the DK regime the adoption of this mode of production
was much more brutal—a sudden “overturning of all old customs”, where the “tax or rent does
not visibly bring any rent or return, either psychic or material” (Vickery 1999, 288).

70

2.8.2 Interpreting the Cambodian Revolution
Historians of the DK period have long debated the nature of the regime’s “clean, pure”
revolution: was it a popular revolution or was it a state-led reign of terror? Linked to this debate
are questions about the role played by ordinary Cambodians during the DK period: were they
largely passive victims of the revolution or active participants, or even perpetrators?
What is certain is that the revolution wooed peasants with some seductive political ideas,
including the “myths” of self-reliance and egalitarianism. Peasants appreciated the support of the
Angkar and the material advantages the Khmers Rouges revolution brought to the countryside.
The regime helped farmers build dams, dikes, irrigation channels, ponds, and houses. The
revolution claimed that it would take care of every need of its citizenry (Becker, 28). As part of
this pledge the DK regime committed itself to building houses for everyone, so that people would
be “neither rich nor poor”. The regime ordered the construction of one-room houses on stilts
(Kiernan 1996, 160), although it is unclear how widely this housing policy was implemented. For
the first time in Cambodian history peasants were treated the same as city folk and the elite, for
the simple reason that under the revolution all became equal: “everyone regardless of rank
performed manual chores in the belief this would transform them into workers” (Sarin Ith
49
,
quoted by Becker, 141-2).
For these and other reasons several historians have argued that the DK revolution enjoyed
considerable popular support, at least among the peasantry. Noam Chomsky and Edward Herman
(1979) argued just after the end of the DK regime in 1979 that poorer peasants had given their
support to the DK revolution willingly and rationally
50
. In another interpretation, Michael

49
Sarin Ith, Regrets for the Khmer Soul (Phnom Penh: n.p., 1973).
50
Noam Chomsky and Edward S. Herman, After the Cataclysm: Postwar Indochina and the Reconstruction
of Imperial Ideology (Nottingham, UK: Spokesman, 1979), xi, quoted in Rosemary H.T. O'Kane,

71

Vickery (1999) suggested that the Khmers Rouges revolution was a “peasantist revolution of the
purest sort” (Vickery 1999, 306), driven by an apolitical “peasant populism” that had as its main
features a conservative utopianism, a belief in the sacredness of the soil and those who till it, and
a distrust of state and bureaucracy” (Vickery 1999, 304). Ben Kiernan (1996) argued that
“widespread” peasant support for the Khmers Rouges probably did exist early on in the
revolution, although this is difficult to assess. Kiernan suggests that “far from all being passive
victims, low-moving targets, or terrified onlookers, many peasants did take initiatives of their
own—for and against the revolution, and on separate issues” (Kiernan 1996, 167).
Complicating the question of popular support for the revolution was the fact that so many
Cambodians, particularly the young, joined the Angkar voluntarily or were forced into service.
The recruitment of children was justified early on by the Khmers Rouges’ slogan that “to
establish a new society we need new people”
51
, preferably people without memory. To create this
new society, the Khmers Rouges took over the role of family in raising children. Young
“comrades” were persuaded to cooperate with the revolution because of the psychic rewards and
the belief that they were serving their families. Moreover, there were also concrete rewards: those
who joined militias did not have to work in the fields or on dam projects, and the rations for the
militias were much better than those of ordinary children (Ea and Sim, 14).
On the other side of the debate, several historians have suggested that most ordinary
Cambodians, including the peasantry, were victims rather than active participants, let alone
supporters, in the DK revolution. Kate Frieson (1991) argued that, while it would be wrong to

“Cambodia in the Zero Years: Rudimentary Totalitarianism”, Third World Quarterly 14, no. 4 (1993): 735-
48.
51
Henri Locard, Le ‘Petit Livre Rouge’ de Pol Pot, ou les Paroles de l’Angkar (Paris: L’Harmattan, 1996),
242, quoted in Ea Meng-Try and Sim Sorya, Victims and Perpetrators? Testimony of Young Khmer Rouge
Comrades, Documentation Series No. 1 (Phnom Penh: Documentation Center of Cambodia, 2001), 11.

72

assume that there was no rural support for the Khmers Rouges at all, Cambodian peasants were
on the whole not leading the revolution but instead they were “unwitting participants in a
revolution whose leaders were faceless, whose goals were hidden, whose tactics were terrifying,
and whose strategies seemed to offer little or nothing of benefit to peasants”. Frieson argues that,
in this context, peasant survival strategies consisted of “digging in, bending low and cursing
inwardly”
52
. Ponchaud, too, concluded that the Cambodian people were ultimately victims of the
Angkar: their credulity was manipulated to enable mass migrations and evacuations, to get them
to serve as instruments of production, and to eliminate opponents of the regime. There was no
question of a “popular” revolution: in the end, power rested wholly in the hands of but a few
(Ponchaud, 227), namely, the Angkar.
Elizabeth Becker argued that the DK regime ruled solely through violence and terror, and
had never attempted to cultivate popular support. Invoking Hannah Arendt’s theories on
revolution
53
Becker suggests that, in the logic of terror, the DK regime was afraid of all power
and therefore could not create a new power base after destroying the old one. In an environment
where nobody could be trusted the regime turned not only against its enemies but also against its
supporters. Under such conditions everyone became a victim (Becker, 209-10). Rosemary
O’Kane (1993) suggested that Cambodian society under the Khmers Rouges was controlled
through a system of terror that was deployed to destroy all spontaneity. Using Hannah Arendt’s
definition of totalitarianism as a “secret society”
54
, she argued that the Cambodian revolution was
a “rudimentary” form of totalitarianism, defined as a “stage of totalitarianism before the

52
Kate Frieson, The Impact of Revolution on Cambodian Peasants, 1970-1975 (Ph.D. diss., Monash
University, Australia, 1991), 9, 16, quoted in Ben Kiernan, The Pol Pot Regime (Chiang Mai, Thailand:
Silkworm Books, 1996), 166.
53
See Hannah Arendt, On Revolution (New York: Viking, 1968).
54
See Hannah Arendt, The Origins of Totalitarianism (London: George Allen and Unwin, 1958).

73

[totalitarianism] system has been perfected and fully functional concentration camps and a secret
police system, separate from the army, have been achieved” (O’Kane, 744-6).
2.8.3 The Bitter Legacy of the DK Regime
Irrespective of the nature of the revolution, most observers are clear that for the former
urban population the DK period was an unmitigated disaster. Unlike their rural compatriots, city
dwellers had little to gain from the revolution, and everything to lose. People from the cities were
suspect, and did not enjoy full citizenship. The population was divided into two groups: “full
rights citizens” (neak penh sith), defined as those who had lived in Khmers Rouges (i.e., rural)
zones before 17 April 1975; and “candidates” for such status (neak triem), particularly the newly
deported and evacuated people from the towns (Kiernan 1996, 57). Moreover, urban
qualifications were unnecessary, and urban occupations were abolished, as only peasants, workers
and soldiers were needed for the “clean, pure” Cambodian revolution (Becker, 28). Finally,
urban evacuees died in large numbers. Many of the young, the old and the infirm died of natural
causes, as a result of the long-distance migrations and the hard physical work (O’Kane, 735).
Towards the latter stages of the DK era all Cambodians—whether formerly rural or
urban—suffered at the hands of the Angkar. David Chandler points out that by the middle of
1977 there were rice shortages and, for the first time in Khmer history, rice had virtually
disappeared from the diet (Chandler 2003, 221). There was arbitrary death and punishment for
everyone. There were no laws or rules overseen by courts or judges; people did not know who
was empowered to order an execution. In a growing atmosphere of terror, dread and fear
enveloped the cooperatives. Ben Kiernan argues that DK policies deprived peasants of three of
the pillars of their lives: land, family and religion (Kiernan 1996, 167), although the removal of
the pillars of family and religion must have affected former city dwellers just as hard as peasants.
All Cambodians became “unpaid indentured laborers”, organized into de facto labor gangs.

74

Despite the communal ideal, the Angkar “atomized all individuals to ensure maximum social
control” (Kiernan 1996, 167). The Angkar “monopolized all loyalty and sense of identity”
(Becker, 153). There were reports of periodic acts of rebellion against Khmers Rouges rule:
Margaret Slocomb (2003) explains that “anti-Pol Pot resistance occurred sporadically throughout
the DK years” (Slocomb, 39), but these rebellions were mostly initiated by military elements.
There is little evidence of widespread popular resistance to DK rule. Instead, Becker claims that
ordinary people mostly survived by staying deaf and mute (Becker, 153-4, 191).
By 1977 long-simmering tensions between Democratic Kampuchea and Vietnam on
territorial questions resurfaced, culminating in military offensives. In December 1977 the DK
regime severed diplomatic relations with Vietnam. Both sides were building up their forces
(Chandler 2003, 223). On 25 December 1978 the People’s Army of Vietnam began an invasion
of Democratic Kampuchea, mounting a major offensive on several fronts, with the aim of
capturing Phnom Penh. The Vietnamese officially claimed that “the armed forces and people of
Vietnam were helping the Kampuchean brothers to fight against the genocidal clique in the pay of
Beijing to take back independence and freedom”
55
. But it is more likely that events in Cambodia
were being shaped once again by the Realpolitik of the larger powers, rather than by humanitarian
considerations vis-à-vis the Cambodian people. The United States claimed that the Vietnamese
invasion was an example of imperialism, while several historians argued that the invasion was not

55
Truong Chinh, “A propos du problème Kampuchéen”, official press release (Paris: Service de presse de
l’ambassade de la République Socialiste du Vietnam à Paris, 1980), 21, quoted in Margaret Slocomb, The
People's Republic of Kampuchea 1979-1989: the Revolution after Pol Pot (Chiang Mai, Thailand:
Silkworm Books, 2003), 47.

75

expansionist in terms of imperialism, but to deny the use of Cambodia by a country the
Vietnamese perceived as hostile to their interests, namely, China
56
.
2.9 The People’s Republic of Kampuchea Period
On 7 January 1979 Vietnamese troops entered Phnom Penh and announced the
“liberation” of the capital. For the third time in a decade, Phnom Penh witnessed a radical change
in regimes and political ideologies. A pro-Vietnamese regime was installed, the United Front for
National Salvation (known by its French acronym, FUNSK), composed of purged former Khmers
Rouges operatives including then-Foreign Minister (and later Prime Minister) Hun Sen. On 10
January 1979, the People’s Republic of Kampuchea (PRK) was officially declared. The
Vietnamese invasion and occupation of Cambodia marked the abrupt end of the DK regime, but it
did not yet signal the end of the Khmers Rouges as a political or military force. The capital’s
latest “liberators” were welcomed by a majority of Cambodians, but not by all: as David Chandler
notes, tens of thousands of Cambodians, particularly young people, were still prepared to give
their lives to the Khmers Rouges—“the first organization that had given them power and self-
respect” (Chandler 2003, 225)—and to try and rid Cambodia of the Vietnamese “invader-
occupiers”.
The Vietnamese invasion marked the end of almost four years of mass murder and brutal
civil war in Cambodia, but the suffering of the Cambodian people was not over yet—what was to
follow was “a decade of isolation and poverty” (Vann Molyvann 2003, 17). The isolation and the
poverty were caused and exacerbated by the failure of the new PRK regime to win international
recognition of its right to govern Cambodia, and by ongoing warfare in the country (Slocomb,

56
W.S. Turley, “The Khmer War: Cambodia after Paris”, Survival 32, no. 5 (1990), quoted in David W.
Roberts, Political Transition in Cambodia 1991-99: Power, Elitism and Democracy (Richmond, Surrey,
UK: Curzon, 2001), 8.

76

xiii). The Vietnamese-backed PRK regime was widely condemned as an occupation force in
Cambodia. In 1982 a “Coalition Government of Democratic Kampuchea” was formed that
mounted a counter-insurgency effort based from the northwestern border areas of Cambodia. The
CGDK was an unlikely alliance of Khmers Rouges, republican and royalist forces, which
received active military support from the United States, China and ASEAN countries (Slocomb,
xiii), all of whom feared the close ties between Vietnam and the Soviet Union. With its
combination of Western, Chinese and ASEAN support, the counter-insurgency, led by Khmers
Rouges troops, would continue to destabilize many parts of Cambodia until well into the late
1990s.
2.9.1 A Ruined City Repopulated
After the FUNSK announced on the radio that a new regime had taken power, another
great movement of people took place across Cambodia. People scattered in all directions, looking
for lost relatives, attempting to return to their former homes, or simply trying to stay away from
ongoing fighting between the Vietnamese army and the Khmers Rouges. According to Evan
Gottesman (2004) thousands fled with the Khmers Rouges towards the Thai border, most of them
abducted at gunpoint (Gottesman, 38).
Hundreds of thousands of people attempted to make their way to Phnom Penh. The army
put up roadblocks at key entry points to the city to prevent the population from entering the
capital before a new administration could be set up. During the next several months, people
languished in temporary settlements around Phnom Penh before they could enter the city.
Impromptu informal markets sprang up to cater to the internally displaced people waiting to enter
Phnom Penh. There was no money yet, so all transactions were negotiated through barter (Ty,
54-6). Each of the temporary settlements accommodated between 6,000 and 60,000 people; they
also functioned as screening centers, where the new PRK/FUNSK regime screened individuals

77

and assigned jobs in the civil service to those who could establish their skills and former
residency in Phnom Penh (Slocomb, 54-5).
When the first people returned to Phnom Penh they found a ghost town, where footsteps
echoed eerily in deserted streets, where cars and motorcycles lay abandoned, and where most
apartments and shop houses had been shuttered for years. The city had assumed a rural aspect as
the DK regime had planted coconut palms, banana trees and “useful” crops such as corn in the
city center, to replace “capitalist” and “superfluous” parks and gardens (Khieu 1997, 50, 56).
The people returning were for the most part newcomers to the city, of rural origin; much
of the original population of Phnom Penh did not survive the Khmers Rouges years. The war
years had decimated the people who used to run the city. On a physical level, too, Phnom Penh
after the war was a ruined city. During a ten-year period, basic municipal services and urban
planning had been neglected. Water and electricity connections had been severed. Basic
infrastructure, including roads, riverbanks and canals, had crumbled. The sewage system had
been blocked by the Khmers Rouges. Garbage was everywhere, and it took months to collect it
all.
The governance of Phnom Penh during this time was divided between the central
committee of the Party and the Municipality of Phnom Penh. Monivong Boulevard was the
boundary: everything to the East of the boulevard was the responsibility of the state, and
everything to the West of the boulevard was under the control of the Municipality (Ty, 54). Ty
Yao, a municipal officer in 1979, claims that initially the new regime tried to control the
repopulation of the city by guiding people in groups of 100 families to available housing, but this
system rapidly broke down because it was too slow and because the population pressure was too
high. Instead, during the course of 1979 the new authorities allowed the people to enter Phnom
Penh at will and people ended up resettling themselves, wherever they wanted (Ty, 54-5).

78

The new, post-1979 population in Phnom Penh occupied land and housing on a first-
come, first-served basis, following the criteria of social rank and, for civil servants, the distance
to their places of employment (Balbo, 44). Possession of houses and land was tolerated
(Gottesman, 76). Former residents of Phnom Penh typically returned to the homes where they
lived before the DK period, but many did not end up staying there. According to Ty the raison
d’être of staying there had disappeared as most families had lost loved ones, so these families
sought out new homes (Ty, 59), presumably to escape the bad memories. In other cases families
found that other people were already occupying their former homes.
The PRK authorities instituted a basic civil registration and surveillance system: families
had to register the identity of family members and their place of residence with the local
authorities. This information was recorded in a “family book” that families had to keep with them
(Hughes 2003, 175). The PRK regime declared void all legal claims to property that were valid
before 1979. As during the DK period, all land and housing became the property of the state
(Khemro and Payne, 182), although there are indications that an informal housing “market” was
starting to take shape, with residential land occasionally being unofficially transferred between
people by mutual agreement (Sik, 4).
In the countryside, in another echo of the Khmers Rouges period, the PRK regime
organized Cambodians back into cooperatives, the so-called “solidarity groups” (krom samaki).
Each solidarity group comprised between 10 to 15 families, who shared land, labor and draft
animals (Sik, 4). The official rationale for the solidarity groups was to “get agricultural
production back on its feet with the shortest possible delay
57
”. The new cooperatives were to help
restore rice production, rescue the economy and resolve the food crisis. They also had a

57
Statement by then-Foreign Minister Hun Sen to a Western journalist, March 1979, as quoted by Evan
Gottesman, Cambodia after the Khmer Rouge (Chiang Mai, Thailand: Silkworm Books, 2004), 91.

79

humanitarian function: in the vacuum following the disappearance of the Khmers Rouges, there
was chaos and a free-for-all as people tried to get a hold of land, animals, tools, seeds and carts.
The losers were elderly people, widows and other vulnerable people, who would otherwise have
had no means to fend for themselves. But there were also political motivations: Gottesman
argues that the solidarity groups were also a useful means for the PRK regime to draw people
away from the resistance and keep people under the regime’s control (Gottesman, 91). However,
the new regime did not strictly enforce collectivization: people were encouraged to remain in the
cooperatives, but in practice people soon “voted with their feet” and tens of thousands of families
headed for the cities or the Thai border in search of food and to rebuild their lives (Gottesman,
93). Most of the people on the move went to Phnom Penh, and joined the fast-growing informal
economy there.
2.9.2 The First Post-war Housing Crisis
Much of the city’s housing stock had suffered serious damage during the DK period.
Beng Hong Socheat Khemro (2000) estimated that approximately 82,000 of the 122,000 houses
that existed in Phnom Penh before 1975 had been destroyed or had deteriorated due to neglect.
Leonhardt and Boonyabancha (1993) estimated that 63,700 houses in the city had been destroyed
by the DK regime
58
. Private homes had no electricity connections until 1983. Water connections
were not reestablished until the early 1990s; in the intervening years water vendors did a brisk
business (Ty, 59). For ordinary Cambodians overcrowding was the norm. Even in the nominally
egalitarian PRK the best housing was accorded to preferential groups.

58
Maurice Leonhardt and Somsook Boonyabancha, An Introduction to the Urban Poor in Phnom Penh,
Cambodia (Bangkok: ACHR/Habitat International Coalition, 1993), 39. Estimates by Beng Hong and
Leonhardt and Boonyabancha are quoted in Valérie Clerc and Virginie Rachmuhl, Les marchés foncier et
immobilier des quartiers informels à Phnom Penh (Paris: Groupe de recherches et d’échanges
technologiques, 2005), 17.

80

The larger and less dilapidated houses were generally occupied by Vietnamese or
Cambodian soldiers, Vietnamese advisors, or Cambodian state cadres. Newly
appointed civil servants were assigned housing by their ministries or by
Vietnamese officials. Others took whatever was available. Cambodians tended to
cluster together, often with multiple families to a house and five or six people to
a room (Gottesman, 76).
In addition to occupying existing houses, from the early 1980s newcomers to the city
started occupying and subdividing vacant land and precarious zones, such as boeungs and river
banks. The PRK regime put in place a combination of coercive measures and incentives to try
and prevent urban squatting on a massive scale and channel urban development. In 1981 the
Council of Ministers issued a circular that aimed to send back to the provinces the jobless and
people living with disabilities
59
. In practice, the policy was not effectively implemented and had
little effect. In another initiative the regime established “social housing land reserves” around
Phnom Penh, where the poorest city dwellers were encouraged to settle and obtained building
rights (but not ownership rights). According to Michel Igout, despite the best intentions of the
authorities, the land reserves ending up being “little better than slums built on piles”,
characterized by “wooden and corrugated iron shanties”, over-crowded conditions, and the lack
of public hygiene (Igout, 20). Moreover, the land reserve sites became the objects of speculation,
as private investors sought to buy up the occupation rights of families in order to gain control
over the sites
60
.
In addition to social housing the PRK regime started to plan and subdivide new higher-
end housing areas on the city outskirts. The first such area—in Phnom Penh Tmei, west of the
district of Tuol Kork—was subdivided and equipped with ready-made wooden housing, but it

59
Circular 55, Council of Ministers, 29 December 1981 (Doc. 6-25), as quoted in Valérie Clerc and
Virginie Rachmuhl, Les marchés foncier et immobilier des quartiers informels à Phnom Penh, 179-180.
60
Christiane Blancot, Phnom Penh: urbanisme et génie urbain, Mission report (Paris: Atelier parisien
d’urbanisme, 1992), 8.

81

attracted few takers, perhaps because it was too distant from the city center, or perhaps because it
was poorly serviced. Instead, the area ended up being inhabited by military families
61
.
Around the same time the PRK regime began to distribute and “lend out” parcels of land
to individual families in the countryside, in order to create sufficient incentives to achieve
economic growth, and to encourage former city dwellers to stay in the countryside (Gottesman,
94-5) and thus to limit the inflow of migrants and returnees to Phnom Penh. Cambodians were
also permitted a growing range of economic activities, and to sell their goods in the private
market. In Vietnam the Communist Party was experimenting with similar measures: in December
1987 it had passed Vietnam’s first Land Law, which enabled land to be allocated for free to
households and individuals for agricultural production. All in all these measures represented a
“retreat from collectivization”, but did not (yet) constitute recognition of private ownership
(Gottesman, 95). But in the end, none of the PRK regime’s measures were effective in halting
the (re)migration of the population to Phnom Penh. During the 1980s the city’s population
continued to fill up again, from an estimated 300,000 people by the end of 1979 (Gottesman,
76)
62
to 427,000 in 1985 and 615,000 in 1990 (Igout, 20).
2.9.3 The Restoration of Private Property
In February of 1989, as perestroika was sweeping the Soviet Union, and as it became
clear that collectivization was failing in Cambodia, the PRK regime announced a formal change
in policy and reinstated private property. The regime introduced a new article in the constitution
(Article 15), allowing citizens to “have full rights to hold (kan kap) and use (prae pras) land and

61
Ibid., 10.
62
The population figures of Phnom Penh during the PRK regime are contested. Evan Gottesman quotes
figures from official PRK sources. By contrast Michel Igout (20), quoting unidentified “local sources”,
claims that only 100,000 people had settled in the city by the end of 1979, but this figure seems low
considering the huge numbers of people waiting to (re)enter the city in early 1979.

82

have the right to inheritance of land that the State has granted them to live on and to conduct
business on”. A new sub-decree (Sub-decree 25) recognized all (informal) occupations of land
and buildings from the 7
th
of January 1979 onwards
63
. Land was redistributed to private
households for housing and farming purposes, based on the number of family members and land
availability. After almost 15 years of a nationalized land market under the Khmers Rouges and
under Vietnamese occupation, Cambodians thus regained the right to possess, use, transfer and
inherit land, but the majority of the population occupied land without legal documents.
The state still tried to regulate land use as the amendment prohibited the sale and renting
of land and also prohibited citizens from using farmland and forest land “arbitrarily and for a
different purpose without permission from the competent authorities”. However, the regime
could no longer control the situation: the sale and rental of houses increased dramatically, and
even state cadres were involved. The distinction between the right to hold and use property and
absolute ownership was ignored (Gottesman, 276). The reinstatement of private property rights
was later confirmed in the Land Law of 1992.
As a result of limited public resources at the time, the privatization of land was not
accompanied by the introduction of cadastral mapping or land titling. The only way in which
private property rights were established and articulated was through informal boundaries
separating individual plots and informal agreements between owners (Sophal and Acharya, 4).
Despite the fact that property rights were not clearly defined, most observers conclude that the
privatization of land was very beneficial to ordinary Cambodians. A study by the non-

63
According to Sub-decree No. 25 (1989), “the State will not review and make a new division on land that
is already occupied from 7 January 1979 to the date of this instruction, which shall be applied until there is
a new land law” (Sik, 5).

83

governmental organization Oxfam (1999) claimed that “there was a fair degree of equity in the
distribution of land” and that almost all individuals who were eligible actually received land:
Ownership rights were given for residential plots of sizes not exceeding 2,000
square meters. Possession rights were given for cultivated land of plots not
exceeding five hectares, and concession rights were given in plantation plots
greater than five hectares. In addition, ownership rights were given to citizens
who occupied houses and dwellings in Phnom Penh (Shaun Williams
64
, quoted in
So et al, 11).
But the relatively equitable situation on the land and housing markets following
privatization would not last long. The private economy had been unleashed and inequalities
would grow. Socialist ideology was rapidly making way for private enrichment, particularly
within the regime itself: government cadres were skipping work to conduct private business,
using their “rights of possession” over land and houses as a basis for speculation (Gottesman,
280-2). A new form of patronage was evolving: according to Gottesman, “profits earned by state
and Party cadres in the free market filtered upward to powerful patrons and those with the
authority to permit private business ventures” (Gottesman, 282). But the “end of ideology” also
presented opportunities for the poor. According to Caroline Hughes (2003) the “unsupervised de
facto privatization of public land” in 1989 made areas of illegal housing available cheaply to the
poor” (Hughes 2003, 183) at a time when more and more migrants were coming to Phnom Penh.
The radical economic changes in 1989 would be followed by equally radical political
changes. As a result of its own economic problems, the Soviet Union had begun to adopt a more
pragmatic policy towards the U.S. and China. Mikhail Gorbachev made clear that the Soviet
Union was no longer willing to sustain its sponsorship of Vietnam’s occupation of Cambodia
(Roberts 2001, 16-7). In September 1989 the Vietnamese officially withdrew their army from

64
Shaun Williams, Review of Secondary Sources Relating to Land Tenure and Access Issues (Phnom Penh:
Oxfam Great Britain, Cambodia Land Project, 1999).

84

Cambodia; Vietnamese advisors had already begun returning home. In 1991 the PRK’s socialist
revolution was formally abandoned and socialism was renounced (Slocomb, 261, 268).
International negotiations had begun on the form of a future Cambodian political framework that
could bring together all domestic political factions, under the temporary aegis of the United
Nations. These negotiations would eventually culminate in the Paris Peace Treaty.
2.10 The Paris Peace Treaty and the Transition from War to Peace
The Paris Peace Agreement of 1991 represented the beginning of a new chapter in
Cambodian history—the promise of a return to peace, an end to international isolation, and the
basis for economic recovery. For a period of three years (1991-1993) Cambodia was run by an
international interim administration under the auspices of the United Nations Transitional
Authority for Cambodia (UNTAC), which was then the largest peace-keeping and stabilization
mission in United Nations history.
The early 1990s were a period of major economic, political and legal reforms for
Cambodia. In terms of the struggle for “access” by the urban poor, two legal milestones were the
passage of a new Constitution (1993) and the first Land Law (1992). During this time Cambodia
ratified five international human rights conventions, thus integrating international human rights
norms, including housing rights, into the national normative framework of Cambodia. In addition,
following the withdrawal of UNTAC in 1993, the United Nations Office of the High
Commissioner for Human Rights in Cambodia (OHCHR/Cambodia) and the position of Special
Representative of the Secretary-General for Human Rights in Cambodia (SRSG) were established
at this time
65
. Another milestone was the budding of a local civil society, including housing

65
Since its creation, OHCHR/Cambodia has documented cases of forced evictions in the context of its
protection work. Starting in 1995, the SRSG explicitly considered housing rights and forced evictions as

85

rights organizations: with the signing of the Peace Agreement in 1991 a ban was lifted on the
activities of non-governmental organizations in Cambodia. Since then the number of local and
international NGOs operating in the country increased rapidly.
However, the country’s new legal instruments and civil society organizations were
immediately put to the test: the UNTAC period and the early 1990s witnessed the unleashing of
major economic and political forces that would present a severe challenge to the housing rights of
the urban poor.
2.10.1 The Paris Peace Agreement and the UNTAC Period
On 23 October 1991 four Cambodian factions (the incumbent Vietnamese installed
regime, the Khmers Rouges, the KPNLF, and FUNCINPEC) signed the Paris Peace Agreement,
after years of internationally-sponsored diplomatic activity. The Agreement called for a Supreme
National Council of Cambodia (SNC) to be established as the “unique legitimate body and source
of authority in which, throughout the transitional period, the sovereignty, independence, and unity
of Cambodia are enshrined”. The SNC was to be chaired by Prince Sihanouk and comprise
representatives of each of the four factions. The Agreement paved the way for the arrival of
UNTAC. According to Julio Jeldres (1993) the SNC was to act in partnership with UNTAC,
delegating to the latter all powers necessary for the implementation of the Paris Accords and
providing advice as needed to the U.N. special representative serving as head of UNTAC (Jeldres
1993, 106).
The transitional authority was by all accounts a massive operation. Its foreign
deployment included 15,900 peacekeeping troops, 3,600 civilian police, and approximately 3,000

part of the OHCHR’s mandate. Thus, the SRSG submits regular reports and recommendations to the UN
General Assembly, the UN Commission on Human Rights and the Royal Government of Cambodia on
housing rights issues in the country.

86

administrators and election officials. These were supported by tens of thousands of Cambodians
who were recruited mainly to help organize the election (Lizée, 10). The Paris Peace Agreement
authorized UNTAC to take direct control of five key government ministries (Foreign Affairs,
Defense, Finance, Public Security, and Information); verify the complete withdrawal of foreign
troops from Cambodia; supervise and monitor a cease-fire; canton and disarm the armed forces of
the four Cambodian factions; supervise the detection and removal of land mines; and raise public
awareness about the dangers posed by land mines. UNTAC was also mandated to foster an
environment of respect for human rights; coordinate the repatriation of refugees from the refugee
camps on the Thai border; and organize and conduct free and fair elections for a Constituent
Assembly. The entire operation cost an estimated $2 billion (Jeldres 1993, 106-7).
During its three-year mandate the UNTAC operation managed to improve “access” for
ordinary Cambodians in at least two important ways. First, its refugee repatriation program
enabled many thousands of Cambodians to return in peace to Cambodia from refugee camps on
the Thai border and rebuild their lives, with the help of a basic assistance package from the U.N.
agency for refugees, UNHCR. Second, UNTAC’s successful electoral registration program—and
the relative success of the subsequent national elections in May 1993—enfranchised millions of
Cambodians. But at the level of institution building, many observers conclude that the Paris
Agreement and UNTAC were unable to make a big impact. Most notably, the elections did not
bring true democracy and power sharing in the Western mould, as the United Nations and its
Western backers had (ambitiously) envisaged. This failure was partly due to the nature of the
peace settlement: Jeldres and others have argued that the Paris Peace Agreement was imposed on
the Cambodian factions by their foreign backers (Jeldres 1993, 107), and driven less by the major
powers’ concern for the development and democratization of Cambodia rather than by their
desire to end a prolonged military conflict that was affecting major international interests

87

(Roberts 2001, 30). The result was an “extremely fragile agreement that did not result from any
genuine reconciliation among the four Cambodian factions” (Jeldres 1993, 107). One
consequence of this was that the mutual hostility between the Cambodian factions continued,
albeit through (mostly) political rather than military means.
Pierre Lizée (2000) questioned the very logic of the Paris Peace Agreement and
UNTAC’s goals: according to Lizée the imposition of peace and liberal democratic institutions
was at odds with Cambodian social practices and institutions, on the one hand, and with the
configuration of political power in Cambodia, on the other (Lizée, 134). David Roberts (2001)
also contrasts the international community’s goals with Cambodian institutions and traditions.
He concludes that the transition to democracy in Cambodia in the 1990s foundered on two planks
of tradition. The first is the deeply ingrained system of patronage and clientelism (khsae), with
loyalty passing upwards and gifts passing downwards. This widely spread system of patronage
represents “the main means through which ordinary Cambodians tap into the state apparatus to
guarantee a personal social security net”
66
. The second plank of tradition is the absolutism of the
elite, which tolerates little democratic opposition. Roberts argues that the Paris Peace Agreement
“made no attempt to alter [these two traditional structures of power in Cambodia] to fit
democratization” (Roberts 2001, 32, 205).
UNTAC similarly failed to address the widespread legacy of impunity it encountered in
Cambodia and failed to introduce an “environment of respect for human rights”. An UNTAC
report in 1993 found that Cambodian society lacked the basic institutions and processes upon

66
M. Ward, “Constraints on Re-Building Cambodia’s Economy”, Proceedings of a Conference on
Cambodia’s Economy, University of Washington, Seattle, 1994, 3, quoted in David W. Roberts, Political
Transition in Cambodia 1991-99: Power, Elitism and Democracy (Richmond, Surrey, UK: Curzon, 2001),
33.

88

which respect for human rights depended. It defined the fundamental justice issue as the
complete breakdown in the legal system:
Existing institutions were geared towards rigid political control by an
authoritarian state backed by active military force and unwilling to accommodate
alternative sources of authority. The existing legal and institutional structures
were essentially those which had developed over the previous decade. There
were inadequate or no legal texts, whether on civil law, contracts and property,
criminal law and procedure, rules of court, evidence, or labor law. Institutions
such as the police and courts were not fully organized or properly functioning
67
.
2.10.2 The Rise of Land Speculation and Squatting
In the early 1990s, in the wake of privatization and rapid economic and population
growth, there was unprecedented pressure on land resources throughout Cambodia. Land became
the object of demand by multiple stakeholders with different purposes (Sophal and Acharya, 4).
Local and foreign entrepreneurs acquired land for commercial farming, logging, and non-
agricultural activities. Buyers from Phnom Penh bought land as a source of investment and
speculation. Land was acquired through both formal and informal procedures. In the latter case,
informal transaction letters were drawn up by the parties and signed with the endorsement of
local authorities (Sik, 17). Common property resources, such as forests, rivers, lakes, and
agricultural land not privatized in 1989 came under private control “at an alarming rate”,
according to Sik Boreak (2000) through both legal and illegal mechanisms. Millions of hectares
of forests were granted to private companies on a concession basis, while a considerable amount
of unallocated agricultural land was illegally encroached upon and thus become de facto “private
property”. The privatization of common property resources resulted in reduced access to food
and livelihood sources for millions of rural people (Sik, 21-3).

67
As reproduced in Yash Ghai, Report of the Special Representative of the U.N. Secretary General for
Human Rights in Cambodia to the U.N. General Assembly, E/CN.4/2006/110/Add.1 (New York: United
Nations Economic and Social Council, 2006), 9.

89

Phnom Penh experienced a similar wave of privatization and speculation, as government
officials sold or leased public land, buildings, cinemas, factories, hotels, and other state-owned
properties to private investors, for terms of up to 99 years. No official figures on these
transactions of state property were ever made available to the public (Sik, 18). Many of the city’s
oldest and most characteristic public and historic buildings became the object of speculation and
were threatened with demolition by private investors, who wanted to ride the property boom and
construct hotels, apartments and shop houses as fast as possible. At around the same time the
arrival of 30,000 UNTAC officials, as well as staff of foreign aid organizations and Cambodians
from abroad, sparked an artificial property boom as the new arrivals went in search of offices and
living space in the capital. Many families took advantage of this opportunity to rent out their
houses, often demanding several years’ worth of rent in advance. With this advance they
constructed new houses for rent, thus feeding a speculative spiral
68
.
The city’s land market also experienced increased pressure from lower income groups,
due mainly to steady migration from the rural areas and speculation by a network of informal
brokers. This second large wave of migration to the city since 1979 was a result of both push and
pull factors: extreme rural poverty and growing rural landlessness were pushing people out of
their villages, while the economic upswing in Phnom Penh also had a magnet effect on many
rural migrants. Informal land occupation and “squatting” became increasingly visible in public
spaces all over the city. According to a survey of 187 squatter settlements in the early 1990s,
two-thirds of the city's squatters settled in the city’s southern Chamkarmon district, where they
were concentrated around the banks of the Bassac River. The survey showed that the most
important consideration for squatters was proximity to work, school or family, and not just simply

68
Christiane Blancot, Phnom Penh: urbanisme et génie urbain, 7.

90

availability of land or cost of land and housing
69
. Apart from the river bank, squatters also
settled on many other available public spaces, such as around boeungs and along canals and
railway tracks, and on pavements and rooftops. The Special Representative of the U.N. Secretary
General for Human Rights in Cambodia suggested in 1995 that squatting in Phnom Penh at the
time represented an act of individual desperation, for lack of a better alternative:
Because of the dislocation caused by Cambodia's recent history—and in
particular the forced clearance of cities and towns, the death of many
landholders, the demobilization of the armed forces and the defects and
uncertainties of the land law—thousands of Cambodians have been forced to
seek shelter in squatter communities. Most of these involve housing which is
primitive and makeshift, providing no access to water and sewerage facilities,
regular garbage disposal, education and other public services. Some involve
semi-permanent structures but without assurance of continued right of
occupation
70
.
But while this view emphasized the role of “poor squatters” principally as victims of
larger events, other observers stressed the role of squatters as active negotiators, strategists and
even opportunists. According to Caroline Hughes the poor were able to gain a foothold on the
space of the city, for living or working, through “processes of private accommodation and
negotiation” with “state officials and powerful private citizens who wielded the power either to
recognize or to assault occupancy of space” (Hughes 2003, 177).
For Beng Hong Socheat Khemro (2000), squatting was not only—or not even
predominantly—the result of poverty and rural migration, but a land invasion activity linked to
organized speculation and political interests. In this process of invasion all squatters—whether
poor or better off—were active participants. In his study of urban squatters in Phnom Penh in the
1990s Beng Hong Socheat Khemro found no significant difference between the income levels of

69
Bronwyn Curran, “Fate of Squatters in the Balance”, Phnom Penh Post, 25 February 1994.
70
Michael Kirby, Report of the Special Representative of the U.N. Secretary General for Human Rights in
Cambodia to the U.N. General Assembly, A/50/681 (New York: United Nations Economic and Social
Council, 1995), para. 24.

91

squatters and the majority of the city’s population. Moreover, most squatters were not recent
arrivals from the countryside but migrants who were already more established in the city (Beng
Hong 2000, 269-70). Beng Hong argues that the proliferation of “squatter settlements” in Phnom
Penh during the 1990s was a direct result of land invasions by “squatter landlords” who had good
connections to the authorities, and who in some cases even enjoyed the active cooperation of the
authorities. In other cases, well connected individuals, usually in the military or police force,
would squat on public land, subdivide it, and build rooms or huts which were then sold or rented
out for profit (Beng Hong 2000, 21-39). Squatting was also encouraged by senior politicians
during election campaigns: thus, Prime Minister Hun Sen in 1996 made a public announcement
broadcast on all state media that “squatters should not leave the invaded areas if each family does
not receive thousands of U.S. dollars as compensation” (Beng Hong 2000, 39).
Beng Hong Socheat Khemro’s conclusions were echoed by the French urban planners
Christiane Blancot and Jacques Stévenin (1993), who suggested that the appearance of urban poor
settlements in Phnom Penh during the election campaign of 1993 was neither spontaneous nor
chaotic. During the campaign, Blancot and Stévenin estimated that around 6,000 new make-shift
houses appeared around Boeung Kak lake and in the area along the Bassac riverfront, housing
between 35,000 to 40,000 people. The settlements were planned and subdivided by informal
brokers, and enjoyed the full support of politicians: the settlements appeared after central
government authorities had requested the police and the Municipality to cease controls of public
lands for a six-month period, thus enabling them to be occupied
71
.
The squatters themselves were a diverse group, according to Blancot and Stévenin,
consisting of “true” homeless families but also of less disadvantaged groups, including refugees

71
Christiane Blancot and Jacques Stévenin, “L’Urbanisme de Phnom Penh: pour une mise en place des
projets”, Mission report (Paris: Atelier parisien d’urbanisme, 1993), 5.

92

who had repatriated from the border to rural provinces with U.N. support, and then migrated to
the city; families displaced from one of the city’s “social housing land reserves” established
during the 1980s; and small-scale “speculators”—families who were renting out their houses
elsewhere in the city to benefit from the property boom
72
. Another category of squatters during
the early 1990s were families who profited from the real estate boom by selling their land and
house at a premium, and then occupied public land, which was cheaper
73
. Moreover, the
government claimed that many of the returnees from the refugee camps on the Thai border were
not poor, landless people but speculators who were given land by UNHCR when they left the
camps—people who had land in the provinces but still chose to move to the city
74
.
By the middle of 1993 the artificial property bubble created by the presence of UNTAC
personnel started to burst, as the UNTAC period began to wind down and UNTAC personnel
began to leave Phnom Penh. Hotel vacancy rates rose, and the boom in construction of new
hotels passed. For the many families who had benefited (directly or indirectly) from the
economic development brought about by the UNTAC period, the loss of extra income was
painful. The boom of the early 1990s had drawn many poor families to Phnom Penh from the
rural areas, and when the boom was over, this new population remained in the city, but with
fewer economic opportunities and few alternatives in terms of links to the rural areas
75
.

72
Ibid.
73
Blancot, Christiane, Bruno de la Maisonneuve, François Legrand, Pascale Sorlin, Jacques Stévenin, and
Tep Vatho, “L’Urbanisme de Phnom Penh: du constat aux projets”, Mission report (Paris: Atelier parisien
d’urbanisme, 1993), 27.
74
Bronwyn Curran, “Ugly Squatters Must Make Way for Zoo, says Government”, Phnom Penh Post, 11
March 1994.
75
Christiane Blancot and Jacques Stévenin, “L’Urbanisme de Phnom Penh: pour une mise en place des
projets”, 6.

93

2.10.3 A Policy of Forced Evictions
Even as high-level officials from government, the police and the armed forces (in their
private capacity) were stimulating squatting through their ties with informal brokers, and even as
politicians were encouraging squatting for political gain in the run-up to the national elections in
1993, the Municipality of Phnom Penh and local authorities embarked on a violent forced
eviction campaign.
Between 1990 and 1996 housing rights organizations recorded 29 forced evictions
affecting at least 4,016 families
76
. This figure counts only evictions of larger squatter settlements:
the true number of evicted families is likely to be much higher, as during this time numerous
unrecorded small-scale evictions occurred as part of demolitions in the city center, as many old
properties were transformed into hotels, villas and other urban projects. Many of the properties
being developed housed low-income families who had occupied the structures since their return
to Phnom Penh after the DK period.
The authorities generally offered two main rationales for the large-scale evictions of
squatter settlements. One popular explanation was that the settlements were a “blot on the
landscape of Phnom Penh”. Justifying the decision to bulldoze squatter settlements around the
shore of Boeung Kak lake in 1994, for example, a Secretary of State claimed that the city had to
make itself more attractive to tourists, as it had been “during Sihanouk’s time”
77
. More
controversially, the Municipality claimed that city squatters were “illegal” and labeled them as
“anarchists” (samnang anna thibtai) because they were allegedly occupying public and private

76
Estimates are of recorded evictions only. Sources: Office of the High Commissioner for Human Rights
(2005) and Urban Resource Center, Study on the Relocation of Urban Poor Communities in Phnom Penh
(Phnom Penh: URC, 2002).
77
Bronwyn Curran, “Ugly Squatters Must Make Way for Zoo, says Government”, Phnom Penh Post, 11
March 1994.

94

land “to which other people already held the land titles”
78
—a dubious claim in the early 1990s, as
no new land titling exercises had yet taken place since the PRK regime had declared all previous
land rights invalid.
Phnom Penh Municipality's then-Director of Land Titles, Chhuon Sothy, announced that
the Municipality intended to move the squatters to sroks (rural districts) surrounding the city and
build them new villages, with international funding
79
. But the international funds for the mass
evictions never materialized, so during the next several years urban squatters were forcibly
relocated to empty fields outside of the city, which were typically prone to flooding and lacked
access to services, basic infrastructure, and employment opportunities nearby. From 1991 to
1996, the forced evictions were characterized by a lack of consultation with poor communities
before the issuance of eviction notices; the absence of an adequate notice period before evictions;
the absence of fair and just compensation; the lack of appropriate sites for relocation; the absence
of transportation to the relocation sites; and the lack of building materials for relocated families
80
.
The few organizations that attempted to document the evictions at the time found that only very
few evicted families remained in their new sites, due to the poor conditions there
81
. Most evictees
presumably returned to the city to squat in other locations.
2.10.4 The Emergence of Civil Society Organizations
In the early 1990s the first signs of a local civil society began to emerge, as human rights
and housing rights organizations were established by former political prisoners, by Cambodians

78
Bronwyn Curran, “Fate of Squatters in the Balance”, Phnom Penh Post, 25 February 1994.
79
Ibid.
80
Huy Rumdoul, “Forced Eviction and Housing Rights: Urban Poor in Phnom Penh”, paper presented at
Cambodia Development Research Institute workshop on “Dispute Resolution in Cambodia”, Phnom Penh,
28-30 November 1995, 62.
81
Urban Resource Center, Study on the Relocation of Urban Poor Communities in Phnom Penh (Phnom
Penh: URC, 2002), 7.

95

returning from abroad, and by international networks of non-governmental organizations and
community-based organizations. The new organizations were funded by a range of foreign
donors, whose support helped to ensure a certain degree of political “cover” and protection for
them to operate. The new organizations began organizing the urban poor and putting pressure on
the authorities to halt the wave of forced evictions and improve living conditions for the urban
poor.
The Cambodian Human Rights and Development Association (known by its French
acronym Adhoc) started its activities in December 1991. The association was founded by a group
of former political prisoners who had been detained in the late 1980s for peacefully advocating
human rights and democracy. In 1992 the Cambodian League for the Promotion and Defense of
Human Rights (known by its French acronym LICADHO) was set up as a human rights and
advocacy organization. LICADHO started out by conducting voter education campaigns for the
elections held in 1993 and monitoring the pre-election environment. Subsequently, it began to
address serious human rights abuses occurring in the country, including land confiscations and
evictions. In 1993, eighteen local non-government organizations and eleven international
organizations working with issues affecting the urban poor, particularly those living in squatter
settlements, came together in an informal grouping called the Urban Sector Group (USG) to
support urban low-income populations through building links between NGOs and urban poor
populations, information exchanges, “community organization”, and dialogue and advocacy work
directed at the authorities, and aimed at halting the forced eviction policy of the local authorities.
USG initiated some of the first surveys of urban poor communities and their living conditions.
By the mid-1990s splits began to appear in the various NGOs’ approaches to the urban
poor. Seventeen of the eighteen local partner NGOs broke away from USG to form the Squatter
Urban Poor Federation (SUPF), which was later renamed the Solidarity and Urban Poor

96

Federation. SUPF received structural support from the Asian Coalition for Housing Rights
(ACHR) in Thailand to operate as a local community network at the city and district levels. The
network was to be led by communities themselves rather than by project staff, as it had been
under the USG banner. According to ACHR most of the NGOs and aid agencies in Cambodia
were “operating in the welfare mode, delivering what they felt the poor needed” (ACHR 2001,
61). ACHR offered the urban poor a different development discourse, one that stressed their
political agency and their self-reliance within a collective of urban poor communities:
There was a prevailing assumption that poor people were too weak to organize
themselves, that they couldn’t trust each other or articulate their needs…The
city’s poor were, in fact, very strong but this strength was atomized and therefore
latent. Harnessing this strength to create an organization which poor people
owned and looked after, and building a broad-based support system for that
organization, has been the goal of ACHR’s contribution in Cambodia (ACHR
2001, 61-2).
ACHR’s message of self-reliance and self-mobilization of the urban poor is an essential
component of the philosophy of Shack/Slum Dwellers International (SDI), an international
activist network of slum dweller communities, of which ACHR is an active member. SDI aims to
empower the urban poor by “spreading ideas of savings, housing, citizenship and participation,
‘without borders’ and outside the direct reach of state or market regimes” (Appadurai 2001, 42).
This is achieved through the concept of “federating”, which Arjun Appadurai (2001) describes as
“the idea of individuals and families self-organizing as members of a political collective to pool
resources, organize lobbying, provide mutual risk management devices and, when necessary,
confront opponents” (Appadurai 2001, 32). The central activity of all community-based
federations is the operation of savings and credit schemes. In Phnom Penh ACHR helped SUPF
to build and expand savings and credit schemes in over 50 different urban poor communities. In

97

addition, the mandate of SUPF was to stimulate poor communities to “stand up for themselves”
82

by organizing communities, supporting local environmental improvements and providing
community members with basic vocational training and training in organizational skills. SUPF
and USG worked together to “organize” hundreds of communities across Phnom Penh. This
process represents the first step in the formalization of an urban poor settlement, and usually
entails the election of community leaders, the establishment of a savings group, and the provision
of basic training to community committees
83
. The underlying objective of all these activities was
to stimulate a process of “open and collective learning” to help “communities recover from the
past” and regain “confidence in themselves and in each other” (ACHR 2001, 62).
2.11 A Sharply Contested Urban Space
By the late-1990s Phnom Penh had undergone a decade of economic expansion and the
ruling Cambodian People’s Party had achieved a firm consolidation of its hold on power, based
on a well-organized party system in the rural areas. The economic and political changes in the
city presented the urban poor and their struggle for “access” with both opportunities and threats.
On the one hand, Caroline Hughes (2003) argues that economic expansion resulted in a
broader distribution of economic power in the city, largely in manufacturing (principally the
garment industry) and urban services. In Phnom Penh, in contrast to the rural areas, this has
meant that “economic resources are distributed beyond the monopolizing abilities of the state”,
and non-state actors have an opportunity to “flee from patrons, or to play patrons off against one
another” (Hughes 2003, 183, 219). Furthermore, discontented groups in the city now perceive

82
Presentation by SUPF at a workshop on the Phnom Penh Master Plan 2020, Municipality of Phnom Penh,
15 July 2005.
83
Urban Sector Group, Nurturing and Enhancing Urban Poor Initiatives and Development Process,
Phnom Penh, Cambodia (Phnom Penh: USG, 2004).

98

that a space for public expression of grievances has emerged, and therefore urban protest
movements have grown, most notably (and vocally) by landless groups, trade unions, and market
traders (Hughes 2003, 183). On the other hand, Hughes argues that the CPP is engaged in a
“longer-term strategy to re-appropriate the space of the city, both materially and symbolically, for
the Party and its leaders” (Hughes 2003, 208). This has resulted in a renewed determination to
clamp down on the urban poor and their settlements and livelihoods in the city center. But the
clamp-down on the urban poor now takes different, more sophisticated forms than in the early
1990s, as the CPP attempts to resolve two contradictory impulses—the need to appeal to the
urban poor, including “squatters”, as voters, and the need to appeal to private sector patrons, in an
effort to consolidate the Party’s hold on power.
The result is that Phnom Penh at the turn of the millennium has become a more sharply
contested space for the urban poor, at the symbolic, discursive level, as well as at the material
level. Significant gains for the urban poor during this period include a radical change in
discourse by the authorities, which acknowledges the poor as a disadvantaged group, and no
longer as “anarchists”. Other milestones for the urban poor include the improving relationship
between the Municipality of Phnom Penh and housing rights groups, and further legal advances,
particularly the landmark new Land Law of 2001, which enshrined the concept of possession that
had first been acknowledged in the privatization law passed by the PRK regime in 1989.
However, as part of the re-appropriation of space by the authorities and the CPP, evictions of the
poor continued during this period, as an ever-dynamic land market in Phnom Penh would put
private developers seeking to expand their landholdings in the capital in the direct path of the
poor.

99

2.11.1 A Change in the Official Discourse towards the Urban Poor
By the late-1990s a shift was taking place in the approach of the government towards the
urban poor in Phnom Penh. The first change was in the official discourse: after several years in
which squatters were referred to as “anarchists” and in which authorities proclaimed that the only
solution to the “problem of the urban poor” was the forced eviction of their settlements, the
authorities at both local and national levels appeared to soften their public tone.
In 1998 the Chief of Cabinet of the Municipality declared that from now on, squatters
would no longer be labeled “anarchists” (samnang anna thibtai), but would be referred to simply
as the “urban poor” (neak kray kroh). In 2000, Prime Minister Hun Sen confirmed a shift in
approach when he announced that the term “anarchist” would be officially replaced by another
term, namely “temporary resident” (samnang bandos asonn)
84
. The term “temporary resident”
echoed earlier references by the French colonial planning authorities to the housing areas of the
native quarters as “temporary districts”. As during the French protectorate, the term seemed to
imply that the urban poor did not altogether belong in the city as full citizens. Nevertheless, for
many of the urban poor themselves, the change in language signaled progress.
The change in the official discourse towards the urban poor was accompanied at around
the same time by a shift in government policy. There were two new developments: first, the
Municipality announced that it would henceforth resettle the urban poor in a more planned and
orderly way, instead of by means of forced evictions. And second, the Municipality demonstrated
an official commitment to on-site slum upgrading as an alternative to resettlement of the urban
poor, through its support for a multi-year United Nations technical assistance project based at the
Municipality.

84
Valérie Clerc and Virginie Rachmuhl, Les marchés foncier et immobilier des quartiers informels à
Phnom Penh, 35.

100

2.11.2 An Official Resettlement Policy
For the first time the Royal Government of Cambodia elaborated the equivalent of a
“resettlement policy” for urban poor communities from Phnom Penh. An inter-ministerial
committee was formed to plan the purchase of plots of land on the outskirts of Phnom Penh for
the resettlement of poor communities from the city center. In the new context, the official
justification for resettlement was as a means of reducing poverty by giving the poor new houses.
The new resettlement policy was presented in a patronizing tone as being in the best interests of
the poor. People living in miserable, dangerous and “unstable” conditions—for example, along
riverbanks and railroad tracks, on pavements and in public parks in the city—would be
“encouraged” to move to new areas on the outskirts of Phnom Penh where they would have
complete ownership of land and could “live happily” in newly developing areas of the city (Pen,
6, 12).
Other motivations guided official resettlement policy as well. As part of a new municipal
policy of “promoting rice fields to be town center”—elaborated under the instructions and
guidance of the Prime Minister—the new resettlement sites were supposed to become
“developing regions” on the outskirts of Phnom Penh (Pen, 9), and even part of a series of
“industrial satellite towns” to be planned around the capital
85
. Another justification was security:
resettlement was seen as a way to “develop…territorial integrity when…applied to borders” (Pen,
2). Finally, and not least important, resettlement of poor communities would enable city
beautification projects “for the public benefit” in those sites that were cleared of squatters in
Phnom Penh (Pen, 11).

85
Cambodia Urban Environmental Improvement Project, Interim Report Vol. 3 (Manila: Asian
Development Bank, 1997), A38, quoted in Valérie Clerc and Virginie Rachmuhl, Les marchés foncier et
immobilier des quartiers informels à Phnom Penh (Paris: Groupe de recherches et d’échanges
technologiques, 2005), 36.

101

Within the framework of the new resettlement policy at least 19 resettlement sites were
established on the outskirts of Phnom Penh since 1996, to accommodate poor families resettled
from the city center. Most of the new sites were in the western Phnom Penh district of Dangkor;
the average distance of the sites to central Phnom Penh was between 10 and 15 km. Depending
on the estimates, the resettlements involved between 7,500 and 9,400 families
86
.
One of the first planned land purchases for resettlement under the policy of “promoting
rice fields to be town center” was in Prey Ti Tuy village of Prey Sor commune, Dangkor district,
approximately 17 km to the west of Phnom Penh, where the Municipality bought 122,936
hectares of land from 19 different families. The money for the land purchases came from the
national budget. The land was intended as a resettlement site and “developing region” named
Tuol Rokarkos, and the site was ostensibly chosen for its suitability for agriculture, easy access to
clean water and its proximity to four pagodas, schools and factories (Pen, 9). During the course
of 2000 and 2001, more than 450 families were moved to the site from the settlement of Sambok
Chab (Village 14) along the Bassac riverfront, in what official accounts portray as a voluntary
movement of the people to a new developing region, after being “encouraged” by the authorities
(Pen, 10-11)
87
.
A significant development was that authorities officially acknowledged that the
resettlement sites or “developing regions” needed to have sufficient basic infrastructure, services
and income generation opportunities in place if resettled populations were to stay there. This
meant that, in a departure from the policies of the past, evictions began to involve some form of

86
The URC’s Study on the Relocation of Urban Poor Communities in Phnom Penh (2002) lists 16
resettlement sites housing 7,517 families, in the period up to mid-2002. According to Clerc and Rachmuhl
(2005), there were between 18-21 sites, housing between 8,500 and 9,000 families.
87
According to URC (2002), 469 families moved to Tuol Rokarkos from Sambok Chab; for Pen (2001) the
figure is 459 families.

102

compensation in the form of land, infrastructure and basic services (URC 2002, 8). In Tuol
Rokarkos, the families received individual plots of land of around 120 square meters. The
Municipality acknowledged its responsibility to “develop the new region” to ensure sufficient
roads, wells, health care centers, schools and jobs (Pen, 12). Basic individual family support was
provided—families relocated to Tuol Rokarkos each received a plastic tent, a 200-liter water
container, 20 kg worth of rice, and seedlings for fruit trees (Pen, 11).
2.11.3 A Focus on Slum Upgrading
In 1996 the United Nations Center for Human Settlements (UN-Habitat), the United
Nations Development Program (UNDP) and the Municipality of Phnom Penh launched a multi-
year project that focused on upgrading of informal settlements and support to the Municipality
and local authorities in the development of a “pro-poor approach” to housing issues
88
. The
project had an unusual mix of downstream and upstream objectives, as well as institution-
building targets. In terms of its downstream activities, the project undertook small-scale
infrastructure improvements in slums across the city, and provided emergency support to the poor
during involuntary and unplanned relocations. In terms of its capacity building activities, it
helped to set up new structures to promote dialogue between the Municipality and urban poor
communities, such as the Urban Poverty Reduction Unit (UPRU) within the Municipality, as well
as Community Development Management Committees (CDMCs) at district level.
As part of its upstream work, the project attempted to develop the Municipality’s
commitment towards slum upgrading policies and a voluntary and consultative approach to
relocations of the urban poor as a last resort. By the end of the project, partly as a result of

88
This is the Phnom Penh Urban Poor Communities and Municipality Development Project, known in its
second phase as the Phnom Penh Urban Poverty Reduction Project. The project ran from 1996 to 2004.
The author was hired by the Project from September 2003 to project close in April 2004 to be advisor to the
Governor of Phnom Penh on land and housing issues for the urban poor.

103

project activities and structures, the Municipality and urban poor communities were in regular
dialogue with each other about upgrading and community development issues more generally.
2.11.4 Advocacy by International NGOs and People’s Organizations
The shift in official policy towards slum upgrading and away from forced resettlement
was due in large measure to advocacy efforts by local and international non-governmental
organizations and community-based organizations, as well as some international donors. Perhaps
the most important dialogue with the Municipality was conducted by ACHR and other members
of the SDI network. For ACHR and its partners the relationship with the Municipality formed
part of the SDI network’s “politics of patience”—an approach that emphasizes “slow learning and
cumulative change” and investment in “long-term capacity-building and the gradual gaining of
knowledge” in the government partner and the building of “trust” between ACHR, its community
partners and the Municipality (Appadurai 2001, 30).
The relationship between ACHR/SDI and the government was maintained through
regular meetings at high level between international housing activists, including Jockin Arputham
of the National Slum Dwellers’ Federation in Mumbai and Somsook Boonyabancha of ACHR in
Bangkok, and senior officials in the Municipality and even national government. In addition,
ACHR brought in teams from partner organizations in the SDI network from other countries in
Asia and Africa to help urban poor communities “develop and test community-driven solutions to
problems of housing and poverty in the city”, and to meet with Municipal officials in Phnom
Penh (ACHR 2004, 9). ACHR organized and financed regular “exposure trips” to other countries
in the region for community leaders and Municipal officials.
In a reflection of the growing partnership between the Municipality and civil society, in
1998 the Municipality, SUPF and ACHR created a revolving fund called the Urban Poor
Development Fund (UPDF). The Fund was set up as a joint venture between the three partners to

104

provide affordable credit to poor communities for housing and income generation. It had an
initial capital injection of $60,000, of which the first $5,000 came from people’s own savings
(ACHR 2001, 71).
In order to stimulate the idea of community-led slum upgrading, ACHR, SUPF and
UPDF mounted a vigorous advocacy campaign to promote the idea of community-driven on-site
upgrading of informal settlements in Phnom Penh. Regular private meetings were held with
municipal officials, as well as larger public meetings. Surveys and studies of informal
settlements were undertaken. A pilot upgrading project was unveiled at Ros Reay in early 2003,
which was described as “the first experiment in 100 percent people-planned and people-
constructed comprehensive settlement upgrading” in Phnom Penh
89
. Internal communications
demonstrated that ACHR and its local partners believed that their advocacy campaign had been
successful in expanding the government’s perceived set of options for low-income housing: they
believed that the eyes of municipal officials had been opened to the possibilities of community-
based upgrading, and that the officials were attracted by the low costs involved—and the political
gains
90
.
2.11.5 The 2001 Land Law and “Possession Right”
Another milestone for the poor was the passage of a new Land Law in August 2001. The
new Land Law simplified the concept of “extraordinary acquisitive possession” of immovable
property, which was first introduced in the 1992 Land Law. The legal concept of possession
enables the physical occupation of a piece of property for own use, even though the possessor has
not (yet) been granted official ownership of the land by the responsible authorities. The right was

89
UPDF and SUPF, “Ros Reay”, Community News (Phnom Penh), Issue No. 2 (June 2003): 1.
90
Asian Coalition for Housing Rights, Upgrading 200 Slums in Phnom Penh in Three Years, internal notes
(Bangkok, ACHR, 2003).

105

introduced in order to facilitate private land ownership after nearly two decades of widespread
state ownership (EWMI, 103; 106). But the modern legal concept of “possession” can clearly be
traced back to the past: it echoes the customary principle of “land acquisition by the plough”
enshrined in ancient Khmer law (see section 2.4.4), which recognized de facto individual
possession of land when peasants cultivated and worked the land in a continuous and public
manner. The concept of possession can also be found in the expression of legal principles
adopted in the Civil Code of 1920 during the French protectorate (EWMI, 19-21).
In modern legal terms, the rights of a possessor are “somewhat less” than the rights of a
property owner. But the significance of possession right is that it gives the possessor the right to
convert possession into ownership, through an application process (EWMI, 104; 106). In order
for possession to be transformed into ownership, and in order for possession to be valid, Article
38 of the Land Law stipulates that it has to meet five basic conditions: it has to be
“unambiguous
91
, non-violent, notorious to the public
92
, continuous, and in good faith
93
”. Article
30 of the Land Law adds two other conditions for valid possession: possession must have
occurred for “no less than five years prior to the promulgation of this law”, and it must be
“uncontested”.
Land that can be lawfully possessed includes state land that is categorized as “state
private land”. This includes all land that can be alienated by the state: according to Article 17 of

91
“Unambiguous possession” means that “the possessor has to possess in his capacity as exclusive
possessor acting on purpose for himself but not on the basis of some other rights. If the real possessor
remains hidden behind an ostensible possessor, he cannot claim a title of possession allowing acquisition of
ownership. His possession is null and void” (Article 38, Land Law 2001).
92
Possession that is “notorious to the public” means that the possessor “has to possess without hiding
himself to those who could want to contest his rights on the immovable property and are not able to see him
or to determinate [sic] who he was” (Article 38, Land Law 2001).
93
Taking possession “in good faith” means that “the possessor is not aware of any possible rights of third
parties over the property that the possessor has been possessing” (Article 38, Land Law 2001).

106

the Land Law 2001, state private land may “be the subject of sale, exchange, distribution or
transfer of rights as it is determined by law”. Authorization to occupy state private land can be
transformed into ownership or rights in rem for the benefit of the holder. This is in contrast to
“state public land”, which by law is inalienable, and where possession cannot be transformed into
ownership.
In principle the revised Land Law fundamentally altered the rights of ordinary
Cambodians, as it secured the rights of all individuals to the homes and properties they occupied
on the effective date of the law (31 August 2001) as long as the five basic conditions of
possession were fulfilled
94
. From a housing rights perspective, the new Land Law was
considered an improvement over the previous Land Law of 1992. Whereas under the previous
Land Law people had to submit a written application for possession rights
95
, the 2001 Land Law
made the confirmation of possession rights on state private land automatic, as long as there is no
counter claim, and as long as the possession meets the basic conditions outlined in the law.
Another significant innovation of the 2001 Land Law was that it provided for social land
concessions for the poor, whereby “vacant lands of the State private domain may be distributed to
persons demonstrating need for land for social purposes” (Article 17), in accordance with
conditions set forth by a subsequent sub-decree on social land concessions, dated 2003. These
provisions created the legal basis for future land distributions for the landless poor.
2.11.6 But Evictions and Forced Relocations Continue…
But despite the legal advances, the more inclusive official discourse and the promising
signs of partnership between the authorities and civil society organizations, in practice it became

94
George Cooper, legal advisor, Land Management and Administration Project, Ministry of Land
Management, Urban Planning and Construction, Phnom Penh, interview by author, 23 December 2006.
95
Shaun Williams, Review of Secondary Sources Relating to Land Tenure and Access Issues, 5.

107

clear that progress towards community-led development in Phnom Penh was uneven, and that
there were limits to the Municipality’s new “pro-poor” course.
First of all, the Municipality’s model resettlement project of Tuol Rokarkos turned out to
be the exception rather than the rule. Despite the promise of an official “resettlement policy” and
the policy aim of “developing regions” on the outskirts of Phnom Penh, most subsequent
relocations—and the land purchases necessary to accommodate them—continued to be organized
under emergency conditions after fires and forced evictions, as in the past.
One problem was that there was still little systematic planning, neither of the
resettlements themselves nor of the physical sites. The resettlements of most poor communities
occurred in extreme haste. Partly as a result of the emergency conditions, many relocated
families found their new environments to be little more than empty, flood-prone fields. The
planning and delivery of basic services in the resettlement sites around Phnom Penh was ad-hoc
and poorly coordinated. Multiple donor projects, NGOs, and even private individuals provided
assistance. But despite their best efforts, the lack of basic services, particularly water, electricity,
and garbage collection, remained a big problem in most sites. As a result of deficient public
services, residents had to make do with more expensive private service connections. Most
residents of relocation sites still lacked land titles that were promised to them when they
relocated, though residents rated this as a less urgent problem than the lack of basic services
96
.
Another problem was that dynamic “developing regions” were slow to materialize around
the new resettlement sites. Contrary to the official rhetoric, the rice fields around Phnom Penh
did not become “town centers” but have remained largely isolated and distant from the city’s
economic life, even though a few garment factories and housing developments have started to

96
Relocated residents in Samaki 1-2-3 resettlement sites, Dangkor district, Phnom Penh, interview by
author, Phnom Penh, 23 July 2005.

108

appear in outlying areas of Khan Dangkor during the latter part of the first decade of the 2000s.
The rise in land values in the outlying areas is increasing the attractiveness of the land plots in
relocation sites, but development in surrounding areas is still largely unplanned and slow to
emerge. As a result, the distance to employment opportunities remains the biggest problem for
relocated residents
97
. One study undertaken by URC (2002) reported that residents experienced
an average drop in income of 20 percent after they are relocated to the urban periphery. In some
cases (depending on the distance to the city and personal circumstances) the drop in income has
averaged between 50 to 70 percent, mainly as a result of increased transportation costs to the city
center, the need to accept lower paid work in the new location, or unemployment (URC 2002, 26,
37).
Meanwhile, evictions of informal settlements from the city center continued apace.
Municipal Governor Chea Sophara embarked on a systematic campaign of “city beautification”.
Informal settlements were still considered a “blot on the landscape”, and evictions were ordered
in the name of campaigns to improve public health and environmental conditions, as well as the
need to attract tourists and foreign investors to the city (Hughes 2003, 208-11). The authorities
were prepared to go to great lengths to hide the existence of the informal settlements on the
riverbank: in 2000 Nun Sameth, Deputy Chief of Cabinet of Phnom Penh, ordered a private
company to build permanent billboards to obscure “squatter” houses adjacent to Wat Trei Lak on
the Tonle Sap river
98
. In 2001 suspicious fires destroyed the large informal settlement of Sambok
Chab on the Bassac riverfront, a site the authorities claimed was owned by private entrepreneurs
and which was planned for redevelopment. Residents were resettled to two sites located 17 km

97
Relocated residents in Samaki 1-2-3 resettlement sites, Dangkor district, Phnom Penh, interview by
author, Phnom Penh, 23 July 2005.
98
Bou Saroeun, “Billboards Spell End to Riverside’s Rustic Charms”, Phnom Penh Post, 8 December
2000.

109

and 11 km from the city center, both of which were located on floodplains and lacked all services,
facilities and proper access roads
99
. Residents were forbidden from returning to Sambok Chab,
but most did so anyway given the lack of employment opportunities in the resettlement sites.
Another justification by the Municipality for the continued evictions at this time was the
promotion of “social order” and the need for “greater stability and progress within the city”.
According to Caroline Hughes (2003) these were slogans that complemented Prime Minister Hun
Sen’s security drive after the post-election violence in 1997 (Hughes 2003, 211). In the name of
“improving security” Governor Chea Sophara ordered sweeps of drug addicts, street children and
sex workers
100
, all of whom were considered potential social threats, just like slum dwellers.
Another target were the ethnic Vietnamese, who were quick to be perceived as a social threat by
successive Cambodian governments. In October 1999 hundreds of residents of ethnic
Vietnamese houseboat communities in the city's Meanchay district were evicted. Moreover,
Governor Chea Sophara announced that eligibility for new relocation programs for the urban poor
hinged on proof of Cambodian citizenship, a stipulation that seemed to be aimed at the large
numbers of ethnic Vietnamese who inhabited informal settlements
101
.
Even more significant for Hughes in explaining the official emphasis on “social order”
was the fact that slum dwellers were not tied to any patronage system, unlike the rural poor in the
villages, who were kept under control by traditional patronage structures. In Phnom Penh
political control was undermined by the “opening of the city to the anonymous poor, whose status
was contingent upon cash transactions” and the commoditization of relations rather than political

99
Robert Carmichael and Lon Nara, “Bleak Future for Squatter Settlements”, Phnom Penh Post, 7
December 2001.
100
Bill Bainbridge, “Phnom Penh Governor Chea Sophara on his Poor, his City, and his Ambition”, Phnom
Penh Post, 30 August 2002.
101
Phelim Kyne and Yin Soeum, “…While the Lucky Get to Stay”, Phnom Penh Post, 18 February 2000.

110

loyalties. The informal settlements were the physical manifestation of the “emergence of an
unsurveyed and unruly population” in the cities that was threatening to the CPP and the
authorities (Hughes 2003, 209). The evictions of slum populations to the urban periphery thus
represented an “attempt to reassert government control over public space” (Hughes 2003, 211).

111

CHAPTER 3: THE CASE STUDY SETTLEMENTS
3.1 Overview of Chapter 3
Chapter 2 pointed out how the early 2000s appeared to represent a turning point for the
poor in Phnom Penh, thanks to seminal events including the passage of the Land Law of 2001,
which enshrined the concept of possession rights for land occupants; the budding of a local civil
society; and the start of advocacy campaigns on behalf of the urban poor by foreign and local
NGOs. But the biggest breakthrough yet occurred in May 2003, when Prime Minister Hun Sen
announced a slum upgrading campaign and secure land tenure for most land occupants in the
center of Phnom Penh. The centerpiece of this campaign was a “social land concession”
program, to be carried out through the mechanism of “land sharing”, for four informal settlements
located on some of the most commercially valuable properties in central Phnom Penh. These four
settlements are the case studies for this dissertation.
Chapter 3 examines the political and economic backdrop to the Prime Minister’s slum
upgrading announcement. In section 3.4 I argue that the government’s support for social land
concessions for the urban poor in 2003—after decades of forced evictions of so-called “squatters”
from the city center—was primarily the result of patronage politics in an election year.
However, the new politics of patronage towards the urban poor did not last long, as the profits
from land development that could be reaped following the elections in 2003 were much greater
than any political gains from pro-poor policies. In section 3.5 I attempt to show how the Prime
Minister’s May 2003 land concession pledges were increasingly drowned out by the sound (and
temptations) of commercial land development during the economic boom years in Phnom Penh.

112

The second half of the 2000s was characterized by a rise of “land swaps”, large-scale urban
projects and a reappearance of evictions. Institutional support for the land sharing projects,
within the government and in the Municipality of Phnom Penh, diminished.
In sections 3.6, 3.7 and 3.8 I provide a description of the four case study settlements and I
describe how political and economic realities since 2003 have shaped the development of the land
sharing plans over time.
3.2 A “Housing Breakthrough” for Phnom Penh’s Poor
On 24 May 2003 housing rights groups announced a “housing breakthrough” for Phnom
Penh’s poor
102
. On that day, on the occasion of the fifth anniversary celebration of the Urban
Poor Development Federation, Prime Minister Hun Sen announced that his government agreed to
a proposal from UPDF and the Solidarity for the Urban Poor Federation to support the upgrading
of 100 inner-city settlements in Phnom Penh during the coming year and provide secure land
tenure for all 100 settlements.
The announcement was the result of years of significant advocacy work by the Bangkok-
based housing rights organization Asian Coalition for Housing Rights (ACHR) and other
members of the Slum/Shack Dwellers’ International network, which had been lobbying the
Municipality of Phnom Penh since the late-1990s to promote the strategy of on-site slum
improvement as an alternative to eviction and relocation of the urban poor. The anniversary
event was billed as an opportunity to “showcase the hard work poor communities in Phnom Penh
have been doing around community upgrading” and to “invite the government to support a

102
UPDF and SUPF, “Housing Breakthrough for Phnom Penh’s Poor”, Community News (Phnom Penh),
Issue No. 2 (June 2003): 1.

113

concrete proposal from SUPF and UPDF” to support upgrading of the city’s informal
settlements
103
.
The anniversary event was remarkable for
its scale and symbolism. It was held on the
Bassac riverfront, one of the most contested city
spaces and the scene of frequent evictions of poor
informal settlers. It was attended by over 5,000
residents of poor settlements from around Phnom
Penh as well as ten provincial cities in Cambodia,
by representatives of local NGOs and bilateral
and multilateral aid agencies, by high-level
representatives of local and national government,
and by community leaders and NGO partners
from nine other Asian and African countries in
the SDI network of international housing
activists. The presence of Jockin Arputham of the
National Slum Dwellers Federation in India
prominently at the side of Prime Minister Hun Sen and the new Governor of Phnom Penh, Kep
Chuk Tema (see Figure 4), seemed to constitute powerful evidence of the growing international
clout of SDI and the success of its “politics of patience” strategy in winning housing rights for the
disenfranchised in Cambodia.

103
Ibid.

Jockin Arputham of SDI (left), Prime Minister
Hun Sen (center) and Governor of Phnom Penh,
Kep Chuk Tema (right) under a banner
proclaiming the new upgrading campaign.
Source: PPUPRP, 2003
Figure 4: The Fifth Anniversary Celebration
of UPDF

114

The anniversary event had significant symbolic and concrete outcomes. At the level of
symbolism there was a sudden change in the official discourse towards the poor: instead of being
branded as “anarchists” and banished to the urban periphery, as in the recent past, Phnom Penh’s
poor were for the first time made to feel like an integral part of the city.
Concrete outcomes of the upgrading announcement were not long in coming. On 4 July
2003 the Prime Minister approved a list of four inner-city informal settlements that were to be
pilot projects for the new upgrading program. This was followed by Letter No. 875, dated 8

July
2003, in which the Council of Ministers of Cambodia authorized the Municipality of Phnom Penh
to prepare “social land concession” projects for the four settlements. The letter ordered the
Municipality to “concede land for social purposes to poor communities according to their
demand”.
3.3 Announcing Four Land Sharing Projects
Letter No. 875 announced that four pilot settlements were selected for upgrading: Borei
Keila, a settlement of 1,776 families in the heart of the city; Dey Krahom, a settlement of 1,465
families located near the busy Bassac riverfront; and two settlements along the city’s main
railway line, just south of Boeung Kak lake: Railway A (Santhipeap) and Railway B (Roteh
Ploeung B). Railway A had a population officially estimated at 70 families, and Railway B had a
population officially estimated at 255 families. The authorities estimated the total number of
families in the four pilot settlements to be 3,566, with a total population of 17,348
104
.

104
The population statistics of the four settlements were highly contested. Housing rights agencies
challenged the official population numbers in all four settlements. The figures quoted here were contained
in Letter No. 875, and were the basis for the government’s initial project plans.

115

Settlement Total No. of
Families
Total No. of
Houses
Total Population Settlement Area
(Hectares)
Borei Keila 1,776 1,376 9,979 14.12
Dey Krahom 1,465 1220 5,854 4.70
Railway A 70 60 390 1.32
Railway B 255 233 1,125 10.00
Total 3,566 2,889 17,348 30.14
Source: Municipality of Phnom Penh, 2003

The social land concessions were to take the form of land sharing projects: the slum
dwellers would be given free land and housing, to be financed by private developers through
cross-subsidies from commercial development on a portion of the present site. Land sharing is a
compromise solution between land occupants and private or public landowners or developers,
who claim the land for redevelopment. The compromise consists of dividing up a site so that one
portion of the land is vacated to make way for development, while the land occupants are
resettled on the other (usually smaller) portion, and in the process obtain legal title to their new
plots.
Using this formula, Letter No. 875 determined the following “sharing” solutions for the
four pilot sites
105
:
• In Borei Keila the government would grant residents 30 percent of the total area
of the site, amounting to a concession of 4.98 hectares.
• In Dey Krahom the government would grant residents a concession of 3.7
hectares out of the total of 4.7 hectares, or almost 79 percent of the site.

105
The area sizes (in hectares) cited in Letter No. 875 were later corrected in the case of Borei Keila and
Dey Krahom, after the settlements were re-measured.
Table 2: Official population and area statistics of the four land sharing settlements

116

• In Railway A the government would grant residents 25 percent of the total site,
amounting to a concession of 0.325 hectares.
• In Railway B the government would concede 2.5 hectares to residents, or a
quarter of the total site.
On 12 July 2003 officials from various ministries (including Finance; Education, Youth
and Sports; and Land Management, Urban Planning and Construction), the Municipality of
Phnom Penh, district officials, and representatives of ACHR, SUPF, and UPDF met at the
Municipality to agree the details of the upgrading projects in the four sites. It appeared as if the
first official slum upgrading and land sharing projects in Phnom Penh were about to become
reality.
3.4 The Political Backdrop to Land Sharing
What motivated the government—after decades of forcible evictions of the urban poor—
to proclaim an amnesty for so-called “squatters” in the heart of Phnom Penh?
Years of advocacy work by ACHR, other NGOs and donor agencies at the Municipality
undoubtedly played an important role in steering the government towards a new discourse of
respect for the rights of the poor. But events on the ground were to illustrate that this explanation
was not the whole story, or even a major part of the story. The government’s pro-poor policies
proved to be inconsistent, and even after the Prime Minister’s May 2003 social land concession
announcement they were still not shared by all elements within the Municipality. An illustration
of this is that even after May 2003 the new Phnom Penh “Master Plan 2020” did not reflect any
slum upgrading areas or urban poor areas, and the municipal planners still advocated relocating
most of the urban poor to the urban periphery.

117

Instead, this section will argue that the most important motivation for the land sharing
projects was a return to an old-fashioned politics of patronage. In May 2003, two months before
national elections, the governing party—the Cambodian People’s Party (CPP)—“discovered” the
urban poor and their value as voters and as subjects in a traditional patron-client relationship.
The land concessions to Phnom Penh’s poor made perfect political sense for the CPP, all the
more so since Phnom Penh was traditionally a hotbed of support for the political opposition. In
retrospect, foreign NGOs and their land sharing proposals merely provided an opportunity for the
government to deploy tactics that have been part of a long tradition of patronage to the poor
throughout Cambodian history.
Thus, whereas in the 1990s and early 2000s government policy was to evict the urban
poor by “reasserting control over public space”, as Caroline Hughes (2003) has argued, in May
2003 the government decided (in official discourse at least) to overturn this maxim, even if
temporarily, and patronize the poor by putting its support behind social land concessions in four
urban poor areas of Phnom Penh. An added bonus for the government was that, by introducing
land sharing as the mechanism of social land concession, it was able to muster the support of
private developers, some of whom had already indicated that they were very interested to access
the prime real estate occupied by the four populations of slum dwellers in central Phnom Penh.
The four projects thus appeared to represent a “win-win” solution for all parties concerned.
3.4.1 The Bangkok Model of Land Sharing
The initial idea for land sharing in Phnom Penh came from the Bangkok-based non-
governmental organization ACHR, an active member of the international NGO network
Slum/Shack Dwellers International. During the 1970s and 1980s Thai and foreign housing
professionals—including some future founding members of ACHR—had been actively involved
in helping to realize land sharing agreements in Bangkok between land occupant communities

118

and land owners. The resulting agreements in eight Bangkok low-income settlements were
praised by housing rights organizations and donor agencies as models of compromise: successful
examples of urban redevelopment that were achieved without displacing the urban poor.
Moreover, for public authorities, the attraction of land sharing was that it represented possibly the
“only way in which the urban poor can gain formal access to land and security of tenure within a
city without a substantial [public] subsidy” (Yap, 66).
Ten years later, in the course of its advocacy work at the Municipality of Phnom Penh,
ACHR believed that the land sharing technique might help selected urban communities to access
land in central Phnom Penh, as it had done in Bangkok. But in Phnom Penh the authorities
introduced an innovation: unlike in Bangkok, where residents financed a portion of the cost of
land and housing themselves, through loan schemes, in Phnom Penh residents would obtain land
and new housing entirely for free, as part of “social land concession principles”. In effect, the
land sharing schemes in Phnom Penh were to be financed entirely by cross-subsidies from
commercial development on site. The cross-subsidization principle appealed to the new governor
of Phnom Penh, Kep Chuk Tema, for whom land sharing was a welcome means to finance
upgrading, in an environment where public funding for this purpose was virtually non-existent.
For ACHR and its allies in the SDI network, land sharing appeared to be a pragmatic “political”
strategy in keeping with SDI’s politics of patience.
The land sharing proposals enabled the Municipality to placate both private developers,
who wanted land in the city center to build, as well as civil society and the donor community,
who were pressuring the authorities to put a halt to evictions. Meanwhile, donor agencies
announced that they were prepared to help the Municipality with technical support: shortly after
the May 2003 announcement the United Nations-sponsored Phnom Penh Urban Poverty

119

Reduction Project, which had been based at the Municipality since 1996, hired external experts to
assist the MPP to implement the pilot projects (see section 2.11)
106
.
In sum, in 2003 it seemed that land sharing in Phnom Penh represented a “win-win-win”
solution for the three main parties involved. For the residents of the four concerned pilot
settlements, land sharing held out the prospect of remaining on site, in the city center, in free new
housing to be paid for by private developers. For the Municipal authorities, land sharing offered
a way to help the government fulfill its upgrading campaign pledges, while at the same time
ensuring commercial development in the city. And for the developers, land sharing was a way to
access prime development land in the city center currently occupied by slum dwellers, without
the need to go through time-consuming eviction proceedings
107
.
3.4.2 The Legal Perspective on the Land Sharing Schemes
At the heart of the government’s concept of social land concessions in 2003 was the
premise that the land in the four inner-city settlements belonged to the state—and that the land
was therefore the government’s to give out. The Municipality declared that the land in Borei
Keila was the property of the Ministry of Education, Youth and Sports; that the land in Dey
Krahom belonged to the Municipality itself; and that the land of the two Railway settlements was
the property of the Railway Authority. From this premise flowed an important corollary—if the
land is state land, then the slum dwellers must be “squatters” who are illegally occupying the
land. Interestingly, this central premise of the slum dwellers’ illegality has even been internalized

106
One of these external experts included the author, who was hired to advise the Governor of Phnom Penh
on urban housing and land issues, from September 2003 to the close of the project in April 2004.
107
Paul Rabé, “Land Sharing in Phnom Penh: an Innovative but Insufficient Instrument of Secure Tenure
for the Poor”, paper presented at Expert Group Meeting on Secure Land Tenure: New Legal Frameworks
and Tools, sponsored by the International Federation of Surveyors (Bangkok, Thailand, 8-9 December
2005): 2-3.

120

by a majority—though not all—of the residents of the four settlements themselves (see data
analysis in Chapter 6).
But in actual fact, when viewed within the prism of the revised Land Law of 2001, the
land sharing projects were controversial because they ignored key possession rights held by a
majority of the residents. All four pilot settlements were by definition state private land since
they were slated for redevelopment. The 2001 Land Law considered possession of state private
land valid as long as it complies with five conditions, i.e., that it is “unambiguous, non-violent,
notorious to the public, continuous and in good faith
108
”. These conditions corresponded with the
situation of most of the residents of the four settlements. Moreover, since a vast majority of the
residents of the four settlements had been land occupants of state private land for longer than five
years prior to the promulgation of the Land Law in August 2001, only one conclusion is possible:
the majority of all residents of the four settlements were the legal possessors of the land they
occupied.
From the legal standpoint, therefore, the land sharing projects could be considered less an
act of social “concession” by a charitable government than a straightforward taking of property—
a transgression of the possession rights of thousands of families without their proper input and
without any overriding public benefit. In retrospect, the government’s support for land “sharing”
and the free allocation of alternative housing for the land occupants might be considered as an
implicit recognition of the legal rights of the residents, and an implicit form of compensation. It
certainly appears less compassionate than politically shrewd: in an election year, the CPP knew
perfectly well that people could not be expected to forego compensation for the taking of
properties that were legally theirs in the first place.

108
Article 38 of Land Law 2001. See for more detail the discussion on the Land Law 2001 in section 2.11.

121

3.4.3 The Politics of Patronage
From the very beginning the land sharing pilot projects were identified as social land
“concessions” from the government to the people, for social purposes. In the run-up to the
national elections of July 2003, the slum upgrading campaign and the four land sharing
agreements provided the CPP with a welcome opportunity to dispense patronage to the urban
poor, in the form of promises of land, tenure security, and amnesties. In a modern version of a
long-standing tradition, Prime Minister Hun Sen appeared as father and benefactor, and the poor
slum dwellers were cast in the role of children and grateful recipients of government largesse.
In the case of the land sharing projects, the politics of patronage took an ironic twist—in
a masterful contortion of reality, the government turned what was essentially an illegal taking of
property into an occasion to win loyalty and votes by promising to reward the soon-to-be
displaced land occupants with free replacement housing and property titles. The new housing
was to be given out for free: the Municipality claimed that this was “out of compassion for the
people” who had already suffered so much
109
. Another explanation was that, if the housing was
offered for free, the beneficiaries would be less demanding, and “debate would be closed”
110
.

109
Mann Chhoeurn, Vice-Governor of Phnom Penh (then Chief of Cabinet of Phnom Penh and Head of the
MPP’s Urban Poverty Reduction Unit), interview by author, Phnom Penh, 7 January 2006.
110
Eric Huybrechts, Advisor to the Governor of Phnom Penh, Bureau d’Architecture et d’Urbanisme,
Municipality of Phnom Penh, interview by author, Phnom Penh, 2 February 2005.

122

At community meetings,
residents were shown glossy posters
of the sleek, modern apartment
buildings that were to replace the
shacks and open sewers of the slum
settlements (see Figure 5). Every time
the framed artist’s impression of the
modern apartments of Borei Keila was
rolled out at public meetings, the
audience of assembled residents
would burst into applause.
3.4.4 Two Forms of Paternalism
As it had done for centuries in Cambodia, the politics of patronage went hand in hand
with a pervasive paternalism. The paternalism permeating the land sharing projects took two
distinct forms. The first was a traditional form, common to Norodom Sihanouk’s grand urban
projects in the 1960s (see section 2.7). It supposed that the poor were incapable to act rationally
in their own interests as citizens and agents, and therefore, rulers and policy makers needed to act
on behalf of the poor, for their own well-being, and to protect them from themselves. This form
of paternalism had also guided housing policy for the poor in neighboring countries, as described
by Emiel Wegelin in the 1980s, when commenting on Thai policy makers’ skeptical attitudes to

Artist’s impression of the new Borei Keila community
buildings as presented to residents in 2003 (Source: MPP)
Figure 5: Artist’s impression of the Borei Keila
community buildings

123

self-help housing: policy makers believed that “at best the poor have to be guided, at worst that
the poor have to be prevented from carrying out innately mischievous acts”
111
.
In the land sharing projects, this form of “benevolent” paternalism manifested itself in the
authorities’ identification of slum dwellers as less than wholly moral. At public meetings held for
community residents at the Municipality in 2003 and 2004 the Governor and Municipal Chief of
Cabinet emphasized that the government’s act of “generosity” in giving out land concessions was
conditional on good behavior. They reprimanded residents for gambling and drinking, and for
not saving enough money. The message was clear: the government was being a good parent in
providing for its children, so the children had to be obedient or else they would not be rewarded.
The second form of paternalism was less “benevolent”: it can be described as part of a
deliberate strategy to sideline the poor. While it clearly evolved from the belief that the poor
were subjects who needed to be guided, it manipulated this tradition to impose an agenda that
reflected the interests of dominant elites and ran counter to the interests of the poor.
The less benevolent form of paternalism manifested itself in various ways in the land
sharing process. Most fundamentally, the authorities never approached the slum dwellers as
people with legitimate possession rights in the land sharing sites, but as poor and illegal people
who were being given an amnesty in the form of a land concession. This approach enabled the
authorities to present their land sharing plans to the sum populations on a “take it or leave it”
basis. In the case of Borei Keila, the Municipality of Phnom Penh and Phan Imex Co., the
private company to whom the contract had been awarded, controlled every stage of the planning
and design process of the land sharing projects. The agreement between the company and the

111
Emiel Wegelin, “From Building to Enabling Housing Strategies in Asia: Institutional Problems”, in R. J.
Skinner and M. J. Rodell (eds.), People, Poverty and Shelter (New York: Methuen), 113, quoted in
Ceinwen Giles, “The Autonomy of Thai Housing Policy, 1945-1996”, Habitat International 27(2003): 237.

124

Municipality had been concluded in relative haste and secrecy shortly after the land sharing
announcement, and no amount of community “participation” or technical advice from foreign
donors would alter the nature of the deal. Even though all four proposed projects implicitly
qualified as in-kind compensation for the residents who were to lose their houses, land and
livelihoods in a government taking, in the case of Borei Keila, residents had little real input into
the design of their new housing. Community leaders were invited to meetings but briefed about
upgrading plans rather than consulted. Only minor technical modifications (i.e., those with little
or no cost implications) were considered. Technical advice from the United Nations advisors
based at the Municipality for the PPUPRP was accepted on a selective basis: advice on land
sharing procedures was welcomed, but any advice on the quality or design of the buildings, or on
the nature of the legal contract between the Municipality and the company, was steadfastly
ignored. Meanwhile, residents of the other land sharing settlements, Dey Krahom, Railway A
and Railway B, were invited to present their own design proposals at community meetings held at
the Municipality, which they had produced with the technical assistance of local NGOs,
particularly URC. However, the proposals for low-rise, one or two storey housing were
consistently dismissed as too expensive for private companies to build.
In a second example of less benevolent paternalism, the Municipality drew up its own
criteria to determine which families were “legal” and would be eligible for new housing in the
land sharing projects. These criteria conveniently ignored the Land Law’s pronouncements on
possession rights, by which all families who had occupied land non-violently and continuously
before 2001 were eligible for compensation. Instead, the Municipality drew up four of its own
criteria of eligibility
112
:

112
Meeting at Municipality of Phnom Penh, office of Vice-Governor Chev Kim Heng, Phnom Penh, 25
November 2003.

125

1. Community members with “proper documentation”, including principally the
family book (also known as the “yellow book”) issued by the local sangkat
authorities;
2. Community members without adequate documentation, but known to be long-
term residents in the settlement;
3. House owners not yet organized as community members;
4. “Permanent” renters, defined as those families who had been renting for more
than 3 years prior to the community survey.
Conversely, there were four criteria of non-eligibility:
1. Families who bought land and housing in the settlement but do not stay there on
a permanent basis;
2. Those who have a name on the list of community members, but who have never
lived in the settlement and have no documentation.
3. Short-term renters and renters who do not stay in one permanent place (even if
they have resided in the settlement for over 3 years);
4. Family members without documentation;
Local authorities, comprising district (khan) authorities and commune (sangkat)
authorities, were mandated to enforce these criteria through community surveys and periodic
checks. Surveys were conducted of the residents of the four land sharing settlements in 2003 and
again in later years, in order to collect information about the size of the population, household
composition, tenure status, and the length of time households have resided in the settlement. The
surveys were conducted by a so-called “mixed committee” comprised of officials from the
Municipality, the district of Prampi Makkara, the commune of Veal Vong, district police,
community appointees from the community committee, and representatives of SUPF.

126

3.4.5 Land Sharing Pre-Conditions
In all parties’ enthusiasm to support land sharing, a critical issue was overlooked: several
of the main conditions that had been in place in successful land sharing projects in in Bangkok
and other Asian cities did not exist in Phnom Penh. These included the following six pre-
conditions needed for residents of informal settlements, one the one hand, and landowners and
developers, on the other, to have an incentive to come to the table and negotiate an end to a land
dispute and sign a land sharing agreement (Rabé 2005
113
and Yap 1992).
• Booming property market. During periods of economic boom, commercial development
pressure increases on well-located lands. While evictions of land occupants typically tend
to go up when land values rise, a booming land market may also push landowners to
make concessions with occupants on developable land—provided that a compromise will
enable them to develop right away on a portion of the desired land. Usually, landowners
become amenable to compromise once alternative ways to remove land occupants (both
legal and illegal) from the land have been exhausted. At the same time, development
pressure can also spur land occupants to seek compromise to avoid eviction.
• Well-established communities: The longer a community has been established on a
disputed site, the greater will be its bargaining power vis-à-vis the landowner and
developers. This may be because of legal rights acquired over time, or because of less
tangible factors, such as increased political connections or alliances built up by residents
over the years.
• Community organization and consensus. A strong and cohesive community can resist
eviction by presenting a more unified front to the landowner during negotiations.

113
Paul Rabé, “Land Sharing in Phnom Penh: an Innovative but Insufficient Instrument of Secure Tenure
for the Poor”: 4-5.

127

Conversely, a weak and fragmented community may encourage landowners or
developers to exploit differences among residents and attempt to buy off certain
members, until those resisting eviction are outvoted or otherwise out-maneuvered.
Frequently, community strength will be increased through alliances with people’s
organizations, non-governmental organizations, human rights groups, political parties,
and other types of organizations which may give the slum dwellers’ cause more visibility.
• Third party intermediation. The intermediation of an outside organization with an
interest in an amicable and just outcome to the land conflict is often a critical pre-
requisite of a successful land sharing agreement. Such an intermediary is usually a public
agency, with some political clout. This agency must broker a compromise that is
technically and financially sound, while also meeting sufficiently the interests of all
parties. The intermediary must also ensure that the agreement is enforced on all sides.
• Physical/technical feasibility. A land area that is to be shared must be sufficiently large
to accommodate safely, and in compliance with local regulations, the juxtaposition of
residential and commercial land uses. The new configuration of the shared land area must
be commercially interesting to developers, while at the same time attractive enough for
the re-housed residents. Sometimes, local regulations must be adapted to accommodate
new forms and densities of community housing. In some cases, not all residents can be
accommodated in the new land sharing configuration. In those situations, the community
must be able to agree on who leaves and who can stay—and what the criteria are in each
case.
• Financial feasibility. Each land sharing deal has a unique financial arrangement,
depending on affordability and priorities of residents and developers, and the physical
features of the site. A land sharing agreement is financially viable if residents can afford

128

the new housing and titles, the developer and landowner benefit from the arrangement,
and where relevant, the amount of public subsidy is not excessive.
In Phnom Penh, only three of these pre-conditions were in place when the land sharing
projects were being negotiated starting in 2003—a booming property market, well-established
communities, and financial feasibility. The absence of three other critical pre-conditions, namely
community organization and consensus, third party intermediation, and technical feasibility,
indicated that the land sharing process in Phnom Penh risked developing in a highly distorted
fashion. For while two of the most significant “push” factors were present in the land sharing
equation (the booming property market and the presence of capital indicated that private
developers were ready to develop), two of the most important countervailing forces (as
represented by strong communities and impartial intermediation) were missing.
In the early and mid-2000s, the Phnom Penh economy was booming. Much of this
growth appeared to be driven by construction and land acquisitions. Developers were jostling to
acquire lands and properties—whether in the public or private domain—in strategic locations of
the city
114
. As a result, there was intense interest in strategic lands occupied by poor communities
in the city center, including all four settlements earmarked for land sharing. As the city economy
grew, public authorities (and public officials in their private capacity) became land speculators
and real estate “players” in their own right. The Municipality’s motivations in the land sharing
projects were increasingly being questioned by donors and NGOs. After the U.N.’s Phnom Penh
Urban Poverty Reduction Project officially closed in April 2004, there was no longer any outside
organization that could play the role of impartial intermediary between the residents and the
private developers. By mid-2004, therefore, land sharing in Phnom Penh became a power

114
J. Hamilton, “Building Boom Reshaping City”, Phnom Penh Post, 1-14 July 2005, 8; L. Karner, “Some
Locals Enjoying a Real Estate Boom”, Cambodia Daily, 15 September 2005.

129

struggle between these two parties. As the next section describes, the backdrop of this struggle
was the increasingly “hot” land market in Phnom Penh—and a complete “planning vacuum”,
both for the poor as well as for the entire city.
3.5 The Economic Backdrop to Land Sharing
Even as the government’s new slum upgrading program and the four proposed land
sharing pilot projects were being held up as signs of a new discourse towards the urban poor—a
chance for the poor to remain in the city, as new urban “citizens”—economic realities on the
ground in Phnom Penh were putting the sincerity of the new discourse to the test.
In the early years of the new millennium Phnom Penh was undergoing another era of
rapid expansion. The Cambodian economy as a whole grew at rates between 8 to 10 percent per
annum in the first half of the 2000s. Much of this growth was centered on Phnom Penh, and was
fuelled by land and real estate investments, and by construction. Foreign developers and
commercial real estate service companies were entering the Cambodian market for the first time
since the war years. The construction of Phnom Penh’s first high-rise office towers and privately
planned cities on the urban periphery were emblematic of the new developments. In the context
of this real estate boom, new land market and planning trends emerged that were displacing not
just poor residents of the city, but many other parties who were unable to bid against the high
land values offered by the commercial real estate sector. These trends included a new Master
Plan that completely ignored the needs of Phnom Penh’s low income population, and a rapid rise
in big new urban projects, land swaps, and a return to evictions.

130

3.5.1 Mixed Messages to the Poor: the Phnom Penh “Master Plan 2020”
Even after Prime Minister Hun Sen had given his approval to the upgrading of at least
four informal settlements in the city center, Phnom Penh’s planners continued to plan for a city
without the urban poor.
In 1996 the Council of Ministers had mandated a new master plan for Phnom Penh (the
“Master Plan 2020”) in order to guide urban growth in the capital
115
. This was to be the first plan
for Phnom Penh since the “controlling plan” drawn up by the colonial Urban Technical Service in
1950. The Master Plan 2020 set itself several different objectives: to serve as a guide for public
and private investment; to propose the location of the main infrastructure networks, including
roadway systems of more than 20 meters, railroads and canals, as well as lakes, ports, railway
stations, airports, parks and green spaces; and to delimit, at a scale of 1:25,000, the boundaries of
the public domain
116
.
In the area of low income housing, the Master Plan incorporated few of the “pro-poor”
policies that the Municipality had agreed with its partners in civil society and donor organizations
during the past few years. There was no acknowledgment of informal settlements to be upgraded
in the city center. Instead, in a complete contradiction of the Prime Minister’s social land
concession announcement earlier the same year, the Municipality announced that more than
10,000 “squatters” in Phnom Penh would have to be relocated to unspecified sites outside of the

115
The Master Plan 2020 was mandated by Circular No. 2 of the Council of Ministers, adopted 15 January
1996. The Plan was developed during 2002-2005 by the Bureau of Urban Affairs, an office of the
Municipality of Phnom Penh supported by the French government.
116
Bureau des Affaires Urbaines, “Orientations stratégiques: Schéma directeur d’urbanisme de la
Municipalité de Phnom Penh” (Phnom Penh: Assistance à la maîtrise d’ouvrage en gestion de
développement urbain, French Embassy and Municipality of Phnom Penh), 6.

131

city over a period of five years, starting in 2003, as part of the Master Plan’s scheme to
“transform the city with new roads, public housing and improved infrastructure”
117
.
The announcement by the planners pointed to tensions within the Municipality between
the professional staff and the politicians: from the point of view of the planners, the urban poor
clearly did not belong in the city. The “strategic orientations” of the Master Plan, unveiled in
March 2005, claimed that the slum upgrading schemes promising free housing to “squatters” had
“ruined” the efforts of planners during the past years to reinstate a culture of respect for the public
domain and the rule of law: the upgrading schemes sent a signal that “illegal occupation” was
henceforth the preferred means of accessing land in the city center. In an echo of the housing
policies of the PRK regime in the 1980s, the “strategic orientations” of the Master Plan proposed
instead that special social housing zones be made available on the city outskirts to accommodate
the urban poor. These special zones could be planned as conventional sites and service schemes,
i.e., subdivided into lots and equipped with only basic services. Only those informal settlements
that did not infringe on the public domain could remain in the city center
118
.
3.5.2 Symbols of the New Boom: Big Urban Projects
After independence in 1953, Norodom Sihanouk had commissioned grand public
monuments in Phnom Penh to commemorate the glory of Cambodia’s new nationhood (see
section 2.7). Fifty years later, Cambodia celebrated two decades of uninterrupted economic
growth and the longest period of relative political stability since the late-1960s. As a
consequence, the capital was experiencing a second renaissance. As in the post-independence
period, the renaissance was celebrated with grand construction projects. The main difference was

117
Michael Coren, “Mass Resettlement for Squatters,” Phnom Penh Post, 24 October 2003.
118
Bureau des Affaires Urbaines, “Orientations stratégiques: Schéma directeur d’urbanisme de la
Municipalité de Phnom Penh”, 46-7.

132

that, in the early 2000s, the new urban projects were all commissioned and built by the private
sector—in particular, by a small elite of new Cambodian business tycoons (who carry the
honorific title of okhnas) and their families who owned most of the country’s economic assets.
As in the 1960s, one of the main areas of Phnom Penh to undergo substantial
redevelopment during the new boom years was the Bassac riverfront. As a result of large private
landholdings it was being transformed from an area which was home to cultural institutions and
large urban poor settlements into an area characterized almost exclusively by commercial
projects, government buildings and embassies. New construction projects included the massive
Naga Casino and hotel venture; the adjacent new National Assembly building; the Ministry of
State Assembly Relations building; and the new Australian embassy. The largest planned project
was on Koh Pich island, located directly in front of the Bassac riverfront. There, the Overseas
Cambodian Investment Company (OCIC) planned to build a satellite city, billed as a new city
center, featuring commercial buildings, markets, apartments and condominiums, hotels and an
exhibition center, and a 222-meter high tower
119
, to be completed by 2015.
Another area of Phnom Penh that was being transformed into a new city center was
Boeung Kak lake. Located in the heart of the city, the lake serves an important public function as
a recreational area, and as a natural reservoir to absorb excess water during the city’s annual
flooding season. But in February 2007 the government awarded a 99-year lease to Soka Ko In
Company to transform an area of 133 hectares—including the lake, which would be filled in—
into a commercial, cultural, tourist, housing and resort area. The contract was signed between the
Governor of Phnom Penh and the Okhna Lao Meng Khin, President of Soka Ko In Company
120
.

119
Sam Rith, “Koh Pich Hold-outs Yield to City Landing Force”, Phnom Penh Post, 24 February - 9 March
2006, 4.
120
Municipality of Phnom Penh, press release, 6 February 2007.

133

On the outskirts of Phnom Penh, in Roussey Keo district, the Indonesian firm YPL
Ciputra Group had begun building another satellite city in a joint venture with a local firm, the
Grand Phnom Penh International Company. The 260 hectare satellite city would be worth about
$500 million and feature high-end housing, a hospital, a university, a golf course, a business
center and a shopping mall
121
.
The new investment projects were being planned and implemented largely outside of the
framework of the “strategic orientations” of the new Master Plan 2020. There was no public
agency, including the Municipality, which was able and willing to confront powerful private
interests to enforce the strategic directions mapped out in the Master Plan. Instead, the main
criterion guiding the new urban projects seemed to be the availability of land. In fact, the
Municipality appeared to be more than willing to cooperate with developers to make land and
public space available to private interests for development—even if that contravened the Master
Plan’s “strategic orientations”. The case of Boeung Kak lake was a prime example: despite the
fact that the lake is state public land according to the Land Law 2001, and therefore inalienable,
the Municipality was a key backer of the project.
3.5.3 The Rise of “Land Swaps”
One of the most common—and remarkable—means of access to strategically-located
land in the city center for private developers during the construction boom is the “land swap”.
Through this mechanism private developers purchase or swap well-located public properties in
the city center in exchange for land and new construction for public authorities in the urban
periphery, where these companies have large land banks. The land swaps are reminiscent of the
boom years of the early 1990s, when many old public buildings in the city center were

121
Municipality of Phnom Penh, press release, 22 January 2007.

134

demolished for quick profits. Now, the land swaps incorporate at least the principle of in-kind
compensation, although the public benefit is not always clear, and the terms of the exchange are
far from transparent.
Many of the swaps entail the sale and subsequent destruction of historic properties and
the relocation of public institutions. In January 2005, the Phnom Penh municipal police
headquarters was “swapped” for a new building to be built on land located in Roussey Keo
district, 9 km from Phnom Penh. The same month, New Hope Co. Ltd. agreed to pay for and
build a new Ministry of National Assembly and Senate Relations and Inspection on formerly
public land in Phnom Penh’s Tonle Bassac commune. In exchange, the company will receive all
of the land not occupied by the new building. In February, the Ministry of Interior sold
Monivong Hospital to the Royal Group family of companies that controls the Cambodian
Television Network and the Mobitel mobile telephone firm. In March, the Council of Ministers
confirmed the sale of the polytechnic institute to a private company. In February, Mong Reththy
development company confirmed that it had made a deal with the Ministry of Culture and Fine
Arts to build a new arts campus in Roussey Keo district, on the urban periphery, in exchange for
the centrally located campus of the Royal University of Fine Arts (RUFA) near the city’s old
stadium.
The government justifies the land swaps as a necessary way for the public sector to
finance development. But some deals are clearly not serving the public interest. Only a few
months after parts of the new RUFA campus were constructed, for example, the local press
reported that the first RUFA faculty members to move to the new premises were living in a
partially filled-in lake, without access to drinking water, sanitation, or electricity. Students and
faculty alike were complaining of shoddy construction, with the floors subsiding, the ceilings

135

caving in, and cracks in the walls. Moreover, with the greater distance to the city center, fewer
students were attending classes
122
.
The land swaps also affected landmark cultural spaces in Phnom Penh. A prominent case
in point is the Preah Suramarit National Theater (also known as the Bassac Theater), located on
the Bassac riverfront adjacent to the case study settlement of Dey Krahom. In 2005 the theater
was leased to the developer Kith Meng as part of a land swap deal by which the Ministry of
Culture acquired a site on the outskirts of Phnom Penh, officially for the construction of a new
theater. The developer’s plans for the theater and the site were still not clear. Former King
Sihanouk stated publicly on several occasions that he would like to see the original theater
restored again and re-used as a theater.
3.5.4 The Reappearance of Evictions
After a relative lull following the Prime Minister’s announcement of the slum upgrading
program, the number of evictions from urban poor settlements in Phnom Penh rose dramatically
again in 2004. In 2006 the government evicted thousands of families from land they claimed was
owned by private companies or needed for public projects. In many cases, families had lived in
these settlements for more than a decade. In 2007 the Cambodian human rights agency Adhoc
estimated that there were 26 forced evictions of urban poor settlements, involving 5,585
families
123
.
As a result of the rapid growth in private construction and the ever-increasing land prices
in central areas of the city, the lands occupied by the remaining urban poor settlements were
becoming the object of intense speculation. The “new” evictions were predominantly from

122
Chansophea Ung, “L’Université des Beaux-arts dépérit”, Cambodge Soir, 6-8 January 2006, 7.
123
Sokha Cheang, “Evictions Spoil Property Success Story”, Phnom Penh Post, 4 April 2008.

136

private lands, whereas up until several years ago they were mostly from public land. Another
noteworthy new phenomenon is that they were often accompanied by the increased use of
psychological threats and intimidation tactics to split communities and displace residents from
contested land.
But certain core aspects of evictions had not changed much in a few years. The most
common official rationale for the forced evictions was that evictees were so-called “squatters” on
private or public land. The new evictions still took place on the basis of questionable legal
claims, and the authorities showed no hesitation to use excessive use force to enforce these
claims, when needed. Other constants were the lack of adequate advance notice to evictees, and
the relocation of poor families to poorly serviced sites far from the city center.
Among the scores of new evictions, two cases in particular illustrate the authorities’
determination to carry out evictions on behalf of private parties, and the ambiguity surrounding
the legal claims of the companies. In early 2005 the Ministry of Interior signed a 99-year contract
to lease the grounds of the centrally-located Preah Monivong hospital and the adjacent
Department of Drug Control to the Royal Group of companies in a land swap. The contract
ignored the land rights of 168 families who had been living on the site since the late 1980s and
early 1990s—people who therefore qualified for possession right according to the Land Law
2001. The majority of the residents were hospital staff and police—mostly lower-ranking civil
servants who had been invited to live in the area by the hospital chief. The residents had formed
a community (the AB Preah Monivong Community) that had at one point even been awarded a
Certificate of Appreciation by the Municipality of Phnom Penh for being a “model community”.
During a period of one year the families attempted in vain to fight the impending eviction, or to
obtain appropriate compensation. But on 5 June 2006 the General Directorate of the National
Police of the Ministry of Interior issued a notification letter authorizing the Ministry to forcibly

137

remove the residents. Police families who refused to leave the site were threatened with
administrative measures including being removed from their positions. The eviction was carried
out by 200 members of the Ministry of Interior police, armed with tear-gas guns, electric shock
batons and shields. Families of police officers were compensated $1000 and civilian families
$500—paltry sums considering the market value of the site. The families were relocated to Ang
Snuol district in neighboring Kandal province, more than 30 km away from central Phnom Penh,
to a site lacking access to water, sanitation facilities, electricity, and schools. Over 100 of the 168
families immediately returned to live in Phnom Penh with relatives and friends, rather than accept
the poor conditions in the relocation site
124
.
One of the largest single evictions occurred during the same period, in June 2006, when
600 armed military police officers were dispatched to evict over one thousand families living in
Sambok Chab (“Bird’s Nest”; also known as Village 14), the informal settlement located near the
bank of the Bassac river, adjacent to the case study settlement of Dey Krahom. Many hundreds
of families had lived in Sambok Chab since the early 1980s and 1990s: the majority of residents
therefore qualified for possession right under the Land Law 2001. For many others, Sambok
Chab had been a last major refuge area in the city—a place where they had fled to when they
were evicted from other slums, or when they could not afford to live anywhere else. Sambok
Chab had already been the object of previous eviction attempts. In November 2001 a suspicious
fire had destroyed about 2,400 homes in the settlement, but residents had nevertheless returned to
the site. Five years later, the official explanation for clearing the site was that the residents were
squatters on land owned by a company by the name of Suor Srun Enterprises. Human rights

124
Sources: Cambodian Human Rights Action Committee, “Media Statement on the Preah Monivong
Hospital Eviction”, 29 June 2006; Center on Housing Rights and Evictions, Global Survey on Forced
Evictions 2003-2006 (Geneva: COHRE, 2006).

138

agencies, however, claimed that neither the Municipality nor the company could produce any
documentation of the company’s title to the land
125
.
In the run-up to the eviction, the local press reported that the developer resorted to co-
optation tactics to obtain support for the eviction. It rewarded certain villagers with between 10
to 20 plots of land each in the relocation site for opposing critics of the relocation, and thereby
helping the company to relocate residents
126
. In May 2006 police in riot gear ordered residents to
break down their shacks, and removed them by trucks to a one-hectare site located 20 kilometers
from Phnom Penh—an empty field without any housing, running water, sanitation facilities, and
electricity
127
. However, residents returned to Sambok Chab and attempted to rebuild. The
following month, military police arrived at night to evict the returned residents and return them to
the same relocation site. The settlement was subsequently closed off with a fence, and guarded
around the clock by company representatives. Residents were not given any monetary
compensation. With the forced eviction of Sambok Chab the largest single urban poor settlement
had now been cleared from the city center.
3.5.5 The International Response
As the number of evictions increased, international human rights organizations stepped
up their criticism of the Cambodian authorities. In a press release dated 2 August 2006, following
the Sambok Chab and Preah Monivong hospital cases, Human Rights Watch warned about an

125
Human Rights Watch, Cambodia: Phnom Penh’s Poor Face Forced Evictions, Press Release, 2 August
2006.
126
Erik Wasson and Sopheark Chhim, “Some from Tonle Bassac Reject Relocation Site”, Cambodia Daily,
3 May 2006, 17.
127
Human Rights Watch, Cambodia Country Summary Report for 2006 (New York, NY: Human Rights
Watch, 2007), 3.

139

“epidemic of forced evictions” in Phnom Penh, with devastating consequences for the livelihoods
of the poor:
The government is allowing a handful of powerful and well-connected
individuals to line their own pockets while trampling the human rights of
thousands of poor people. The agenda seems to be to rid the city of the poor
while handing over prime real estate to the rich and powerful
128
.
The message was, on the one hand, a clear reference to a perceived culture of impunity
that has flourished in Cambodia in the absence of a strong and independent judiciary. This was a
message that the U.N. and donor governments have raised ever since the UNTAC period. The
Special Representative of the Secretary-General for human rights in Cambodia, Yash Ghai, said
he had received many complaints of executive interference in the work of the judiciary, and many
examples of trials that failed to meet standards of due process
129
. The courts were widely viewed
as corrupt and incompetent, and continued to be used to advance political agendas, silence critics,
and strip people of their land
130
.
On the other hand, in addition to the accusations about impunity, Human Rights Watch
and other human rights organizations pointed to grave procedural flaws with regard to relocation:
the government and the courts were “incapable of offering a fair, transparent, uncorrupt and non-
violent process to implement evictions”. They pointed out that the forced evictions are in
violation of international and Cambodian laws, including
131
:

128
Human Rights Watch, Cambodia: Phnom Penh’s Poor Face Forced Evictions, Press Release, 2 August
2006.
129
Yash Ghai, Report of the Special Representative of the U.N. Secretary General for Human Rights in
Cambodia to the U.N. General Assembly, E/CN.4/2006/110/Add.1 (New York: United Nations Economic
and Social Council, 2006), 11.
130
Human Rights Watch, Cambodia Country Summary Report for 2006 (New York, NY: Human Rights
Watch, 2007), 3.
131
Ibid., 2-3.

140

• The International Covenant on Economic, Social and Cultural Rights, which provides
that Cambodia is legally obligated to respect, protect and fulfill the right to adequate
housing.
• Cambodia's 2001 Land Law, which prohibits deprivation of ownership without due
process, and grants the right to apply for a land ownership title to someone who has had
uncontested possession of private property in a nonviolent, continuous, open, obvious and
good-faith manner for five years.
• Article 44 of the Cambodian Constitution, which states that the government can only
deprive someone of property for “public interest” purposes and requires that the
government pay victims fair and just compensation in advance.
• The government's “Strategy of Land Policy Framework”, which states that the
government should avoid forced evictions if at all possible. If people are being evicted
for public interest purposes, the government must pursue a policy of compensation and
relocation.
• The U.N.’s Comprehensive Human Rights Guidelines on Development-Based
Displacement, which urges governments to conduct social impact assessments and make
arrangements for fair compensation and adequate resettlement conditions in advance of
any eviction.
International human rights organizations called for a moratorium on “massive forced
evictions” until the government comes up with a “comprehensive national housing and
resettlement policy” and a procedure for evictions that is in conformity with U.N. human rights
and standards, covering at least the following provisions:
• Prior to any eviction, a consultation with affected persons of all feasible alternatives to
eviction, with a view to avoiding or at least minimizing the use of force;

141

• Legal remedies available for evictees;
• Access to relocations for human rights and U.N. monitors and journalists.
3.5.6 The Rise of Legal Advocacy
As international human rights organizations attempted to shine a spotlight on the worst
excesses of private sector-led displacement of the urban poor, locally-based legal rights groups
were making quiet—though uneven—progress in trying to instill a rights-based approach to
housing for the poor. The approach of these groups differed from the community-based approach
of members of ACHR and other members of the SDI network: whereas the SDI network prefers
to work outside of the legal system to achieve people-based development, the legal rights groups
attempted to work within the legal system in order to improve the system’s ability to help the
poor access land, housing and legal rights.
In addition to the two main local human rights organizations, Licadho and Adhoc, which
were established in the 1990s, two new institutions were set up that started monitoring housing
and land rights of the urban poor. In January 2004, in response to a violent eviction behind the
Intercontinental hotel in central Phnom Penh, several housing and human rights groups
established a Housing Rights Task Force in order to monitor land conflicts and help prevent
future housing rights violations, especially forced evictions. Two years later, the Task Force
achieved funding from foreign NGOs to establish a permanent secretariat.
In the same year, the Community Legal Education Center (CLEC), a Cambodian NGO
with foreign support, established its Public Interest Legal Advocacy Project (PILAP) to “use the
legal system to assert and protect citizens’ rights, as well as to encourage governmental and
private sector transparency and accountability”
132
. PILAP introduced a new concept of legal

132
Source: CLEC, 2006.

142

empowerment to Cambodia—a confrontational, rights-based approach that reflects CLEC’s
origins as a U.S.-funded legal project
133
. PILAP focused on high-profile cases of land conflict
that affect a large number of families and implicate powerful political interests, in order
ultimately to “generate public debate and demand accountability and respect for legal norms”
134
.
One of PILAP’s groundbreaking cases was Koh Pich (“Diamond Island”), an island measuring 80
hectares located in the Bassac river, directly opposite the Bassac riverfront and the newly built
Naga Casino.
3.5.7 The Precedent of Koh Pich
In 2006 the Municipality of Phnom Penh leased the strategically located island of Koh
Pich to OCIC, the holding company for Cambodia’s largest private bank, Canadia Bank, which is
planning to develop the island as a new satellite city (see section 3.5.2 for the development
plans). In December 2004, in anticipation of the lease agreement, the Municipality had started to
present the more than 300 families living and farming on the island with eviction notices. It
offered to compensate the families less than $2 per square meter and relocate them to
“replacement land” 12 km outside of the city. The compensation offered by the Municipality was
far below the market value of $62.50 per square meter the islanders claimed the company is
paying the Municipality in the lease agreement
135
. Due to a combination of misinformation and
intimidation by military police, however, many families accepted the compensation and agreed to
leave the island.

133
In its early years (1996-2001) CLEC received structural funding from the University of San Francisco
Law School and USAID. PILAP’s current technical advisor is provided through USAID and the American
Bar Association.
134
U.S. Agency for International Development, “Defending Citizens’ Land Rights”. Telling Our Story
[project sheets online: http://stories.usaid.gov] (Washington, DC: USAID, 2006).
135
Sam Rith, “Koh Pich Hold-outs Yield to City Landing Force”, Phnom Penh Post, 24 February – 9 March
2006, 4.

143

At this point PILAP decided to take on the Koh Pich case: it organized several “town
hall” style meetings with the remaining residents, at which it presented the legal case to the
residents and offered to formally represent them. In January 2005 78 families signed on as
PILAP clients. All families were long-term residents of the island and possessed at least a
“family book”, which under the Land Law 2001 is ample proof of possession and should legally
serve to establish ownership. The clients all refused to move to the relocation site offered by the
company, and agreed to attempt a negotiation with the authorities in order to get the highest
amount of monetary compensation possible
136
.
PILAP identified several “classes” of residents, based on the strength of the legal claim to
the land they occupied. Class 1 residents, who comprised 49 families, were those with a
relatively weaker land claim: their land had been the subject of a court injunction in 1999, and
therefore their claim did not qualify as uncontested. Class 2 and 3 residents, comprising 29
families, met all five of the criteria for possession under the Land Law 2001: Class 2 residents
based their ownership claim on a family book only, while Class 3 residents had additional
documentation, including land purchase and sales records.
With the lawyer-client relationship established, PILAP and the United Nations Office of
the High Commission for Human Rights requested negotiations with the company and the
Municipality on a counter-offer of compensation. Meanwhile, PILAP wrote to donors with
business connections to Canadia Bank in search of their support for the case. USAID sent
Canadia Bank a message of concern about the process and level of compensation—a message that
USAID believes was “effectively received”
137
. Negotiations began but quickly ran aground on
legal and technical issues. The Municipality and Canadia Bank claimed that the residents had no

136
Brian Rohan, extracts from internal monthly reports on the Koh Pich case, 2004-2006: 2.
137
Ibid., 5.

144

right to occupy the island because it was state land. They poured scorn on the provisions of the
Land Law 2001 and PILAP’s declaration that the family book is ample proof of possession
138
.
Nevertheless, they were willing to compensate residents in order to hasten their departure. The
compensation, however, would be low to reflect the weak legal case of the residents. Moreover,
compensation was to be based not on the size of the land PILAP’s clients claimed they owned (47
hectares) but on a commune survey that purported to show that these clients owned only 25
hectares on the island.
The Municipality declared that its last offer was a compensation of $2.50 per square
meter for all three classes of clients. This would be topped up by an in-kind contribution from
Canadia Bank in the form of a coupon redeemable for land and improvements at one of several of
the Bank’s properties outside of the city
139
.
But PILAP’s clients rejected this offer as insufficient, and indicated that they wanted an
all-cash settlement instead of a part in-kind compensation. PILAP’s class 1 clients insisted on an
all cash settlement of $5.25 per square meter or better, while the class 2 and 3 residents felt that
they had the strongest legal positions and therefore were not willing to consider anything less
than the appraisal value of the land. PILAP pointed out that any agreement would be subject to
an independent appraisal to verify the value of the properties, and an acceptable resolution of the
land area question. Canadia Bank and the Municipality responded by threatening to shut out
PILAP completely and approach its clients directly through their “agents” on Koh Pich. They

138
Ibid., 8-9.
139
Ibid., 8. The resulting compensation structure was as follows: Class 1 clients: $2.50 in cash from the
Municipality and $2.75 in-kind from Canadia Bank, totaling $5.25 per square meter; Class 2 clients: $2.50
in cash from the Municipality and $3.50 in-kind from Canadia Bank, totaling $6.00 per square meter. Class
3 clients: $2.50 in cash from the Municipality and $4.25 in-kind from Canadia Bank, totaling $6.75 per
square meter.

145

threatened residents that if they did not agree to sell shortly, the Municipality would take the case
to the court, after which they would get nothing.
As the negotiations proceeded in fits and starts, the intimidation tactics increased.
Residents of Koh Pich reported increased intimidation tactics by military police stationed on the
island. A local newspaper, Sangkhum Khmer, printed hostile articles about PILAP and urged
residents to sell their land. It also published vitriolic personal attacks against PILAP’s foreign
technical advisor. The military police forbade a local real estate company hired by PILAP to
conduct a survey of its clients’ land. In mid-2005 approximately 50 to 60 families who had
vacated their land prior to PILAP’s involvement in the case returned to the island. They were a
mix of former renters and owners, some of whom sold out on very unfavorable terms in the face
of serious intimidation. Some housing advisors suspected that their reappearance on the island
was orchestrated by the authorities in order to provide a pretext for the Municipality to evict all
the people remaining on the island. Finally, when a municipal court judge used threatening
language to describe PILAP lawyers, PILAP adopted stricter security measures for the entire
legal team
140
.
In July 2005 the Municipality suddenly agreed to the demands of the class 1 families and
settled with all but two families in this category. The final settlement price was $5.25 per square
meter, with reductions in the length of land plots of no more than 20-40 meters per family—an
outcome that PILAP heralded as a “dramatic success” for its most vulnerable clients, who were
facing an eviction notice less than eight months ago
141
.
In September 2005 the Phnom Penh Municipal Court filed a legal case against the Class 2
and 3 clients, which accused the families of living illegally on the island, and called again for

140
Ibid., 28.
141
Ibid., 10.

146

their immediate eviction. But the class 2 and 3 families refused to be intimidated. Instead, they
sought to maximize publicity on the case. PILAP organized a press conference, at which both
PILAP lawyers and representatives of the class 2 and 3 clients spoke. The press conference was
attended by 11 different media outlets, including local newspapers and radio stations. The
Municipality rejected a demand for compensation of $26 per square meter, and the judge of the
Municipal Court issued an injunction order requiring the remaining clients to leave the island or
be forcibly removed.
International efforts also contributed to keeping the Koh Pich case in the news. In
November 2005 the Cambodian Office of the United Nations High Commissioner for Human
Rights delivered a letter to both the Minister of Interior and the Governor of Phnom Penh. On the
same day the United Nations Special Representative for Housing Rights also sent a diplomatic
note to Cambodia’s representative office in Geneva about the case.
Meanwhile, the land market on Koh Pich was heating up. Some class 3 clients with the
best documentation started selling their land to well-connected people at prices between $20 and
$30 per square meter. Real estate experts noted that other investors were looking at the island,
and could be interested to purchase more of PILAP’s clients’ land if they could be guaranteed
clean title. Market prices were rumored to be climbing as high as $50 per square meter
142
.
In November 2005, a breakthrough occurred. PILAP’s remaining class 2 and 3 clients
were approached by an unidentified representative of OCIC, offering them between $10 and $12
per square meter to settle and leave the island. The Municipality and OCIC structured the deal
such that the families received $6.75 in the form of a check, and the balance in cash. The clients

142
Ibid., 21.

147

each signed a contract stating that they had received $6.75 in compensation
143
. For fear of setting
a precedent, the Municipality and OCIC insisted that PILAP’s clients not tell anyone (particularly
not PILAP) that they had received anything more than $6.75 per square meter.
Despite the best efforts of the authorities to keep the size of the financial deal a secret,
news of the Koh Pich compensation packages spread rapidly throughout poor settlements in
Phnom Penh. The Koh Pich case set an important precedent: it demonstrated to urban poor
residents that those who can prove lawful possession are entitled to fair and just compensation
should they need to be displaced to make way for private development projects
144
. It also
demonstrated that the compensation involved can be substantial, if they manage to hold on long
enough.
For PILAP the Koh Pich case resulted in some important achievements. Its lawyers stood
up to pressure from the courts that were declaring the Koh Pich residents illegal. It managed to
forestall a forced eviction and impose a negotiation with the Municipality and powerful business
interests. It helped its clients to obtain compensation at rates between 2 to 6 times what was
initially offered by the company and the Municipality. Finally, PILAP managed to attract local
newspapers and radio stations to extensively cover the Koh Pich story and to publish detailed
interviews with PILAP lawyers and clients, thus disseminating the lessons of the case throughout
the country
145
.
But the Koh Pich case also highlighted two worrying trends in land conflicts in the
current context in Phnom Penh. The first of these trends was the length to which developers and

143
Ibid., 30. Four families remained on the island and refused to accept the settlement: two Class 1 families
and two Class 2 families vowed to persevere until they received the appraisal value for the land. They were
forcibly evicted in February 2006. The two Class 1 families eventually received special (higher)
compensation amounts on account of the small size of their plots.
144
Brian Rohan, Technical Advisor, PILAP (CLEC), interview with author, Phnom Penh, 19 January 2007.
145
Brian Rohan, extracts from monthly reports on the Koh Pich case, 2004-2006.

148

authorities were willing to go to intimidate the urban poor through psychological tactics and
violent threats, in order to push them off land that they were legally occupying. The fact that, in
the case of Koh Pich, these tactics did not succeed in scaring away all residents was thanks in part
to a combination of PILAP’s tenacious lawyers; successful advocacy work that helped put foreign
pressure on the authorities; and the growing presence of other investors interested in Koh Pich
land, who were willing to outbid the company, thus perhaps putting extra pressure on it to settle.
These favorable circumstances for residents are certainly not present in all land conflict cases in
Phnom Penh.
The second worrying trend that the Koh Pich case exposed is the extent of distrust and
internal divisions among urban poor residents that can be exploited by outsiders. In Koh Pich the
various subgroups of residents never formed a united front. Instead, PILAP’s clients often sought
to undercut each others’ positions, accusing other residents of colluding with the government, or
seeking to demonstrate how certain residents’ legal rights are inferior to their own. Moreover,
many residents also distrusted PILAP’s own intentions: they feared that PILAP was in collusion
with government, or otherwise looking to enrich itself. The distrust was fed by the conflicting
advice the residents received from various parties. Some human rights NGOs and the Sam Rainsy
Party encouraged residents to be even more strident towards the authorities and the company than
PILAP advised them to be
146
.
The developments in Koh Pich provided important lessons for the four land sharing
projects in Phnom Penh that were getting off the ground at around the same time.

146
Ibid., 9-10.

149

3.6 The Case Study Settlement of Borei Keila
The settlement of Borei Keila (“sports complex” in Khmer) is located on a 14-hectare site
in the center of Phnom Penh, in the district of Prampi Makkara. The site is bounded on three
sides by major traffic arteries and city landmarks, including Boulevard Tchécoslovaquie to the
east, Boulevard Kampuchea Krom to the north, and Boulevard Charles de Gaulle and the
Olympic stadium a bit further to the south. As a result of its strategic location, the land values of
the area surrounding the settlement are some of the highest in the city.
Borei Keila is the largest of the four settlements earmarked for land sharing. Official
surveys conducted in 2003 listed the number of families living in the settlement as 1,776.
3.6.1 A Model Athletes’ Village
The original complex of buildings on the site was officially inaugurated as an athletes’
village (“cité sportive”) by Prince Sihanouk on 6 November 1966 for the Games of the New
Economic Forces (GANEFO) of the non-aligned countries
147
, which were held in Phnom Penh in
the National Sports Complex (the current Olympic Stadium) in November and December of the
same year. The athletes’ village and the National Sports Complex, which was inaugurated in
1964, were a central element of Prince Sihanouk’s campaign to develop sports in Cambodia.
Both projects were major public works designed to showcase the modern, post-colonial Phnom
Penh during the Sangkum period.
The athletes’ village complex cost 70 million riel to build, and was a donation to the
Kingdom of Cambodia from the People’s Republic of China. Chinese architects designed the

147
The first GANEFO Games were held in Jakarta in November 1963. After the Phnom Penh event, the
GANEFO Games were discontinued, though Cambodia did organize a domestic version of the GANEFO
Games in 1968.

150

complex and the Chinese also furnished and equipped all the buildings
148
. When it was
inaugurated the athletes’ village comprised eight three-storey apartment blocks that could house
1,000 athletes; a gymnasium capable of seating over 1,000 spectators; an administration building
with a lounge, offices and conference rooms; and two restaurants. The facilities were located on
landscaped grounds featuring an entrance way lined with flagpoles of the participating countries,
and large rectangular pools to enhance the aesthetic effect, and also to facilitate drainage. The
apartment buildings were planned in neat rows, and made of reinforced concrete and raised from
the ground to ensure proper circulation of air (Grant Ross and Collins 2006, 28-9).
3.6.2 From Showcase to Slum
After the GANEFO Games, the apartment blocks of the athletes’ village were most likely
used for private residential purposes. In the early 1970s the athletes’ village became a storage
and training area for government military forces. Reports suggest that, after the Khmers Rouges
took control of Phnom Penh in April 1975, the area first accommodated a medical reception
center (Vickery 1999, 83) and subsequently housed foreign ministry officials of the DK
regime
149
. After the fall of the DK regime in 1979 the complex was taken over by the Ministry of
Interior, which used the apartment buildings to house staff of the National Police Training
Academy. The apartment units were given out for free to police officers, who were later joined
by their families
150
.
After 1991, with the repatriation of Cambodian refugees from the Thai border camps,
hundreds more families moved to Borei Keila; many of them were relatives of the police families

148
Helen Grant Ross, interview by author, Phnom Penh, 20 May 2006.
149
James McCauley, “Report of the Study on Household Income Levels and Business Requirements for
Residents of Borei Keila” (Phnom Penh: UN-Habitat Phnom Penh Urban Poverty Reduction Project, 2004),
2.
150
Y. D., police officer and resident of Borei Keila, interview by author, Phnom Penh, 22 November 2004.

151

already staying there, who had nowhere else to go. In 1994, following the elections, the Borei
Keila site was transferred to the Ministry of Education, Youth and Sports (MoEYS), and the
National Police Training Academy was moved to Tuol Kok district. The original gymnasium was
transformed into a television studio for the TV channel Cambodia 5, while the administration
buildings and conference rooms were taken over temporarily by the Ministry of Women’s
Affairs.

The athletes’ village after it was inaugurated in
1966 (Source: Grant Ross and Collins, “Building
Cambodia”, 2006)

Forty years later, the original apartment blocks
were crumbling and surrounded by shacks
(Author’s photo, January 2006)
Figure 6: Borei Keila, from athletes’ village to crowded slum

Despite the relocation of the National Police Training Academy, many of the police
officers and their families stayed in the apartments they had been assigned in the 1980s, and they
were soon joined by staff of MoEYS. No one was left in charge of day to day control of the site,
and as a result many newcomers moved to the area, mostly from the rural areas. As uncontrolled
settlement into the area increased, makeshift housing sprung up around the original apartment
blocks, as the blocks could no longer accommodate the influx of new arrivals. Two senior police
officers took control and established a committee that was supposed to control settlement in the
area, but the residency criteria were never made clear, and the number of migrants increased even

152

more during this period. The two police officers soon “faded away”, however, and their authority
was taken over in due course by a village chief (phum chief).
As a result of the various waves of migration to the settlement over the years, the
population of Borei Keila today consists of a mix of families of the original police officers and
their families who came to the site in the 1980s, staff of MoEYS and their families, refugees from
the Thai border, rural migrants, and military officers. The various groups are not separated in any
specific residential zones, and tend to live together. The apartment buildings are populated by
families from all different groups.
Today the former athletes’ village is a crowded slum, with almost every square meter of
the original landscaped grounds taken over by migrants and transferred into informal housing.
The settlement is characterized by very poor environmental conditions: forty years after their
construction, the original apartment blocks have suffered bad structural damage due to lack of
maintenance. Most of the original drains and pipes are broken or blocked, and apartment
dwellers have made their own water connections to their units. Large mounds of garbage have
formed between the buildings, as residents throw their garbage from the balconies of the
apartments. Only around 10 percent of apartment dwellers have flush toilets. Those families who
cannot afford to replace the original toilets throw their human waste into the common space
between the buildings, wrapped in plastic bags
151
. Effluent from small-scale enterprises operated
by community residents, such as the dye utilized in small dyeing businesses, is also discharged
freely and collects in open spaces, ponds and footpaths.

151
H.B., Leader of police community in Borei Keila, interview by author, Phnom Penh, 25 January 2006.

153

Apartment dwellers make their own water
connections to their units. Garbage is
thrown from the upper floors (Author’s
photo, 2003)

Shack dwellers living in the open spaces in the
settlement lack proper drainage and footpaths
(Author’s photo, 2003)
Figure 7: Slum conditions in Borei Keila

3.6.3 Eviction Attempts
Ever since the settlement started expanding in the early 1990s, an eviction threat has
hovered over Borei Keila. The first eviction notice came in 1993, in the form of a letter from the
Prime Minister informing residents that they had to vacate the land.
This was followed by a violent eviction attempt in June 1996, which displaced 56
families
152
. Following an inter-ministerial eviction order, police armed with bulldozers and
weapons surrounded the area and attempted to demolish people’s houses, which met with fierce
resistance from the residents. The NGO Urban Sector Group assisted residents to organize and
initiate a media campaign (Ouk, 37). After a period of emergency advocacy, during which Borei
Keila community leaders met with representatives of the National Assembly, the Municipality
and human rights organizations, and demonstrated in front of the houses of the two Prime

152
Figures on the estimated number of evictees from Borei Keila are from the Office of the U.N. High
Commissioner for Human Rights, 2005.

154

Ministers at the time
153
, the Inter-Ministerial Committee put an end to the eviction order, and
transferred the issue of Borei Keila to the Municipality of Phnom Penh.
The Municipality subsequently issued two more eviction notices, in 1998 and 2000, in
both cases informing the residents that they would have to leave the site within 15 days. By this
time the residents were better prepared: they had increased their contacts with NGOs, and USG
and SUPF had started organizing eight communities in Borei Keila (Ouk, 37-8). The residents
organized regular protests and demonstrations on site to oppose the eviction efforts. In 1998
district officials attempted to build a fence around the settlement, but the residents obstructed the
efforts and stole the construction materials for the fence. The authorities’ appetite for a forceful
eviction had waned, and political signals to residents were more mixed as the eviction became
caught up in election politics. Before the 1998 elections the Prime Minister offered each
household $4,000 to vacate the land. Meanwhile, opposition parties promised residents land titles
on site and offered $1,000 to those families that could not be accommodated. After the elections
a stalemate ensued and little change occurred until the Prime Minister’s May 2003 announcement
of a land sharing agreement in the settlement.
3.6.4 Phnom Penh’s First Land Sharing Agreement
By 2003 several factors had come together to make Borei Keila a prime candidate for the
first land sharing project in Phnom Penh, with community housing to be built by developers and
financed through cross-subsidies. Over the years the settlement had developed into one of the
most populous and conspicuous slum areas in central Phnom Penh. Given the site’s large size
and its strategic location, the Municipality faced no difficulties trying to interest private
developers to redevelop the site for commercial use. Even if a developer would have to finance

153
After the 1998 elections Cambodia was ruled by a “co-Prime Ministership”, with Norodom Ranariddh
as first Prime Minister and Hun Sen as second Prime Minister.

155

low-income housing through cross-subsidies from commercial development, and “share” a
portion of the site with low-income apartment blocks, the land values in the area were so high that
the land sharing project still made good business sense.
In the case of Borei Keila, things moved rapidly: in a sub-decree dated 16 September
2003, less than four months after the Prime Minister’s announcement, the Council of Ministers
announced that Phan Imex Construction Company Ltd. had been selected as the private partner in
the land sharing project. The sub-decree confirmed that the government accepted a proposal from
the company
154
to divide the 4.6 hectare land concession into two parts: “community buildings”
for the residents of Borei Keila would be constructed on 2 hectares of the concession, on the
southwestern portion of the settlement, while the remaining 2.6 hectares of the concession, on the
northern perimeter of the settlement, would be granted to the company for commercial
development. The remaining land area, amounting to 9.52 hectares, would revert to the Ministry
of Education, Youth and Sports. No public bidding process to select the developer for the Borei
Keila project was ever held.

Land Category Land area
(hectares)
Borei Keila total land area, of which: 14.12
Land provided to the community, to be used for new housing construction 2.00
Land awarded to the private investor (Phan Imex Co.) 2.60
Land to be returned to the Ministry of Education, Youth and Sports 9.52
Source: UN-Habitat Phnom Penh Urban Poverty Reduction Project, 2003
Table 3: Overview of the Borei Keila land sharing agreement

154
On 22 August 2003 Phan Imex Co. had sent an expression of interest to the Municipality (Letter No. 35-
03 Lor-Sor-Phor-Ar-Mor), containing a proposal for the development of the 4.6 hectare community land
concession (Ouk, 59-60).

156

In its 16 September sub-decree the Council of Ministers ordered the Municipality to
agree implementation of the land sharing project with the community. In this spirit, the
Municipality established a community committee composed of ten community “representatives”
to provide input into the planning process on behalf of Borei Keila residents. The representatives
were leaders of the eight existing communities that had been formed by SUPF and USG in the
late-1990s. In addition, there were leaders of two other communities that were newly established
following the land sharing agreement. One of the two new communities was specifically for
police families, the group that had lived in the settlement the longest. In the meantime the Phnom
Penh Urban Poverty Reduction Project hired a local NGO, the Urban Resource Center (URC)
155
,
to provide technical assistance to the company in the design of the buildings, and to help ensure
that the interests of the residents were represented in the planning process.
On 5 January 2004 a formal contract for the development of the Borei Keila site was
signed by three parties: the Governor of the district of Prampi Makkara, the community
committee, and the Phan Imex Company. The company agreed to invest $7,133,901 for the
construction of ten community buildings of six floors each for a total of 1,776 families. In
addition to the buildings, the company was to construct a bitumen road of 400 meters, to connect
the hew housing to the main road. Each apartment unit would measure 40 square meters; there
would be 29 individual apartment units per floor. The company was responsible for building
temporary housing for those residents displaced by the construction. The contract stipulated that
construction of the community buildings had to be completed within 30 months. Construction
would proceed in two phases. During a first phase of 13 months, the company would complete
the first three community buildings. During a second phase lasting 17 months, the company

155
URC already had experience working with residents in Borei Keila before the land sharing agreement to
produce designs for community housing.

157

would complete the seven remaining buildings. The company was liable to pay a fine to the
Municipality for each month of delay.

Source: Municipality of Phnom Penh, 2003

Figure 8: Map of Borei Keila, showing the location of new community buildings

3.6.5 The Community Response: Co-optation and Internal Tension
The Municipality of Phnom Penh hailed the Borei Keila land sharing agreement as a
victory for the poor, and as an example of slum improvement for the other poor settlements in the
city. A model of the ten apartment buildings was henceforth rolled out at community meetings
across the city, as an example of a modern alternative to existing slums (see Figure 5). In public
meetings Borei Keila residents demonstrated gratitude to the authorities and palpable relief that
their situations would be legalized, the eviction threat hanging over their heads removed, and that
they would receive—for free—new housing and be able to enjoy better living conditions.
Residents of the new apartments would only have to pay for electricity and water supply; the
connections to these public services would be installed by the company. Building maintenance

158

would become the responsibility of newly formed communities to be established in each new
building.
But behind the fine rhetoric, the Borei Keila population was beset with internal tensions
as a result of the land sharing agreement. In meetings with the U.N.-sponsored Phnom Penh
Urban Poverty Reduction Project shortly after news of the land sharing deal with the private
company emerged, residents voiced numerous concerns
156
:
• They questioned how Phan Imex Company could have received the contract to develop
medium-rise housing blocks when the company’s only previous construction experience
was limited to building bus stations.
• They had concerns about the design of the buildings: residents did not want to live in
buildings higher than four floors, and they worried that small informal businesses might
not survive in medium-rise apartments far removed from regular foot traffic.
• They objected to the lack of public consultation: they complained that only a small group
of people from the community had been consulted about the land sharing arrangement
and the building plans for the new community housing. The land sharing agreement had
been reached in a top-down fashion, between the company and senior government
officials, without the input of the beneficiaries. The vast majority of residents had not
received any reliable information about the project.
• They were unsure about who would be included in the project, and who would be
excluded. Residents had heard rumors that project insiders had been promised units in
the new buildings, even though they did not belong to the community.

156
Meetings with Borei Keila community members at PPUPRP project office, Phnom Penh, 19 and 20
November 2003.

159

Despite the fact that the Council of Ministers had ordered the Municipality to seek the
community’s inputs on project implementation, residents quickly understood that the
Municipality and the company were effectively sidelining them. For one thing, residents were
asked to rubber stamp the choice of the private developer. In August 2003, the same month that
Phan Imex Co. had sent a letter of interest to the Municipality to develop the site, it had
convinced a group of 208 residents to submit a thumb-printed letter to the Municipality stating
that they supported the company’s proposal. Moreover, on 18 September 2003, immediately
after the issuing of the Council of Ministers’ sub-decree, the Municipality invited 500 residents to
a community meeting to seek their approval for the private partner for the land sharing project
(Ouk, 60). Two days earlier, however, the developer had already been formally selected, thus
reducing the meeting to a formality.
When it came to the design of the community buildings, residents had just as little real
input. Phan Imex Co. organized three public meetings at the Municipality in November 2003 to
share with residents the company’s plans for the buildings and the individual apartments.
However, the meetings were meant to be informative rather than participatory in nature: residents
who objected to the plans were invited to submit their suggestions to the Municipality, via their
appointed representatives. Fourteen letters were sent to the Municipality and Phan Imex Co.
expressing concern about the issue of apartment sizes, but in the end none were followed up
(Ouk, 74). The building designs ended up being driven mainly by cost pressures on the company.
None of the design recommendations made by URC, the NGO brought in by the U.N. project at
the Municipality, were considered if they implied any additional cost to the overall design.
During the course of 2004 and 2005 tensions within the Borei Keila population mounted
as residents began to accuse the community committee members of “selling out” by undermining
the interests of residents in exchange for short-term personal benefit. It became clear that the

160

community committee members had been co-opted by the company and the Municipality to
withhold information from their members, and to short-circuit consultation procedures within
their communities. Evidence emerged of appointees receiving personal benefits in exchange for
their services, including bribes, jobs with the Phan Imex Co., or jobs as security guards on the
construction site at Borei Keila (Ouk, 61).
NGOs and the Phnom Penh Urban Poverty Reduction Project, which had spent many
years organizing and building the Borei Keila communities in the late-1990s and early 2000s,
criticized the ten community appointees for their lack of transparency, and in the process,
undoing the fragile cohesion that had been built up over the years through the community
organization work. By short-circuiting internal community procedures of consultation, one senior
activist concluded that “ten years of community efforts were destroyed in one minute by bad
community leaders”
157
.
Despite the rising tensions and residents’ substantial concerns about the project,
organized opposition did not materialize. One reason for this was the perverse incentives
generated by the concessionary element of the project: residents were receiving for “free” a
housing asset worth thousands of dollars. Under those circumstances, few residents dared to
complain to officials about the project or the flawed planning process for fear of losing their right
to a new apartment.
3.6.6 Inclusion and Exclusion
The ground-breaking ceremony for the construction of the first three community
buildings was held in February 2004, and was presided over by the Governor of Phnom Penh and
the President of Phan Imex Co., Ms. Suy Sophan. The next “battleground” between residents and

157
Tuy Someth, interview with author, Phnom Penh, 19 November 2003.

161

the authorities was the process to determine residents’ eligibility for apartments in the new
community buildings.
Officially, the new community apartments being built by Phan Imex Co. were reserved
for four categories of residents sanctioned by the Municipality (see section 3.4.4 on the land
sharing eligibility criteria). These were: community members with adequate documentation,
principally the “family book” issued by the commune; community members without the family
book, but recognized to be long-term residents; house owners not yet organized as community
members; and “permanent” renters, who have lived in the settlement in the same house for over
three years. The process to vet community members for eligibility and allocate housing was
carried out by the same committee that was responsible for carrying out the surveys of
community residents—the “mixed committee” of officials from the Municipality; the district of
Prampi Makkara; the commune of Veal Vong; district police; community appointees from the
community committee; and representatives of the people’s organization SUPF.
In practice, however, the process to determine the eligibility of residents was complicated
by widespread irregularities. At the root of the problem was the murky definition of “community
membership”. Local authorities identified community members as long-standing residents in the
settlement who participated in the community savings scheme and possessed savings books. But
this definition was prone to abuse. As soon as the land sharing plans in Borei Keila became
public knowledge, a lively informal market emerged in the buying and selling of community
“names”. Given that the only people eligible for free new housing were those who were
registered as community members before 2003, a name on the list of community members
suddenly became worth thousands of dollars. Under those circumstances, newcomers started
bribing community leaders to get their names on the list of members and to obtain savings

162

books
158
. In the process, community leaders became de facto brokers. Residents reported that
“community leaders’ necklaces became thicker and they started owning new cars”
159
.
The population of Borei Keila increased rapidly during the latter half of 2003. The first
official survey conducted by the mixed committee at the time of the land sharing announcement
in May 2003 counted 1,776 families. After the land sharing announcement the population of
Borei Keila purportedly ballooned to 2,329 families, as outsiders moved in and attempted to get
their names on the list
160
. The informal market in community names took many different forms.
Some families, in need of short-term cash, sold their names immediately to newcomers and either
left the settlement or remained to rent on site. Other families rented out houses and plots of land
to newcomers, or bought the names of other existing families, in the hope of claiming several
community names and therefore several apartments. Another opportunity for residents of Borei
Keila to make quick money was to sell their property (and thus their name) directly to the
developer: informally Phan Imex Co. was paying residents thousands of dollars to move out
before they even moved into the new apartments
161
. Presumably, the empty apartments in the
community buildings could then be sold to outside buyers for higher sums of money.
The buying and selling of properties could not take place without the connivance of the
village chief of Borei Keila, who is responsible for signing off on any transaction in the

158
Meetings with selected Borei Keila community members at PPUPRP project office, Phnom Penh, 19
and 20 November 2003.
159
Heng Chhun Oeurn and Prem Chap, junior researchers and temporary residents in the new buildings of
Borei Keila, interviews by author, Phnom Penh, 12 January and 16 January 2008.
160
Sun Saran, Head of SUPF Office for Prampi Makara district and Borei Keila community leader,
interview by author, Phnom Penh, 15 January 2008.
161
Heng Chhun Oeurn and Prem Chap, junior researchers and temporary residents in the new buildings of
Borei Keila, interviews by author, Phnom Penh, 12 January and 16 January 2008.

163

settlement, and thus plays a role as informal “gatekeeper”. The village chief collected a
commission estimated to be in the range of 3 percent of every transaction
162
.
On the other hand, some established “community members”, previously recognized to be
on the mixed committee’s list of beneficiaries, were later struck from the list. This fate concerned
renters, in particular—even those considered long-term residents by the Municipality’s own
definition. Housing rights advocates claimed that 28 families who were long-term renters in
Borei Keila were removed at the last moment from the list of beneficiaries by the Governor of
Phnom Penh, as the Governor feared that “giving apartments to former renters would send a
signal to all [renters] in Cambodia that they could get free apartments in Phnom Penh”
163
. Human
rights agencies feared that renters were now being consistently targeted and labeled as
professional squatters and opportunists
164
. Another group of people who were struck off the list
of eligible beneficiaries were 31 families living with HIV/Aids. These families suspected that
they were denied eligibility—despite their long-term residence in the settlement—because of
their HIV status
165
.
After conducting a third and final survey, the mixed committee reduced the total number
of eligible beneficiaries of the land sharing scheme to 1,482
166
. The authorities declared that the
original survey by the mixed committee had “double-counted” and “miscounted” hundreds of

162
Ibid.
163
Pierre Fallavier, Advisor to UN-Habitat, “Note on the Situation in Borei Keila, Phnom Penh”, e-mail
memorandum, 17 April 2007.
164
Manfred Hornung, Monitoring Consultant, Licadho, interview by author, Phnom Penh, 9 August 2007.
165
Amnesty International, “Around 32 families living at Borei Keila, Phnom Penh”, press release ASA
23/011/2009, 11 June 2009.
166
Sun Saran, Head of SUPF Office for Prampi Makara district and Borei Keila community leader,
interview by author, Phnom Penh, 15 January 2008.

164

residents
167
. Given the active business in buying and selling of community names, it was unclear
how many of the final beneficiaries were original community members of Borei Keila who had
lived there before 2003, and how many were speculators and newcomers who were on the list
under another name. Housing and human rights organizations had no way to verify this: the final
list of beneficiaries was “jealously guarded” by the mixed committee members and by local
authorities, and it remained inaccessible to outside observers
168
.
3.6.7 The First Residents Move In
The first three community buildings (“buildings A, B and C”) were completed in
February 2007 and unveiled during a televised inauguration ceremony by the Municipal Governor
Kep Chuk Tema on 23 March 2007, several days before local (commune) elections. A lottery
was held at the Municipality to select the first batch of residents: 394 families moved into the
three new buildings; 128 apartments units were left vacant
169
. Municipal Vice-Governor Mann
Chhoeurn used the occasion of the lottery to hand out T-shirts with the logo of the CPP, and
reminded the lottery participants that “the government has always cared about its people”
170
.

167
Chhuon Sothy, Head of Cadastral Department, Municipality of Phnom Penh, interview by author,
Phnom Penh, 25 November 2003.
168
Pierre Fallavier, Advisor to UN-Habitat, “Note on the Situation in Borei Keila, Phnom Penh”, e-mail
memorandum, 17 April 2007.
169
Sun Saran, Head of SUPF Office for Prampi Makara district and Borei Keila community leader,
interview by author, Phnom Penh, 15 January 2008.
170
Chhay Channyda, “Government to Provide Housing for Borei Keila Residents”, Cambodia Daily, 3
March 2007.

165

One of three new community buildings
inaugurated in Borei Keila in early 2007
(Source: Municipality of Phnom Penh, 2007)

View of the interior of an individual apartment
unit in the new community buildings in Borei
Keila (Author’s photo, October 2006)
Figure 9: The new community housing in Borei Keila

On the same day that the first three community buildings were inaugurated, the Governor
held a ceremony to break ground for construction of the next seven buildings. This ceremony
was a great relief to the remaining population waiting to be re-housed. They had watched with
concern how the first three community buildings were delivered two years after the contracted
date, while the company had already finished construction of the commercial units on its own 2.6
hectare plot of land. Some residents had begun to worry that the developer might abandon the
construction of the seven remaining buildings.
In a vain attempt to ensure that the new apartments remained in the hands of the intended
beneficiaries, the Municipality prohibited residents from selling or transferring their new units
during the first five years. New tenants held group titles, and could only obtain individual title to
their units after five years. Building leaders were appointed for each building to monitor any
sales and exchanges taking place, and any new tenants moving in, and they were to send the
Municipality updated reports of property transactions on each floor. But the Municipality’s
measures were no match for market forces in this strategically located area of central Phnom
Penh: the new apartment units became the object of rampant speculation the moment the

166

community buildings were inaugurated. Even without individual title, the units were immediately
worth thousands of dollars. In the six-storey walk-up buildings the market quickly assigned a
premium to apartments located on the lower floors. In October 2007 market prices for apartments
on the first floor reached $12,000. The value of the units was going up fast: by mid-2008 they
had reached $18,000. Apartments on the second and third floors fetched prices of $10,000 and
$8,000, respectively, while apartments on the sixth floor were sold for around $4,000 in 2007 and
between $5,000 and $6,000 in mid-2008
171
.
Few of the original poor families in Borei Keila who did move into the community
buildings have been able to resist the temptations of such huge cash windfalls. By the end of
2007, over 50 percent of the families who had moved into the new apartments (a population
including newcomers as well as original Borei Keila residents) were thought to have already sold
out and left Borei Keila entirely, or else moved to the top floors of the buildings, where the
apartments and rents were cheaper
172
. Many of the buyers were reported to be traders from
nearby Oroussei market, who use the apartments as storage space, and who plan to sell the units
again when the price goes up sufficiently
173
. The building leaders were recording all the property
transactions taking place, and passing the lists of new transactions and tenants over to the
Municipality, as stipulated, but the Municipality either could not—or would not—do anything to
halt the informal market. Therefore, barely nine months after their inauguration, the community
buildings—originally meant as a free asset for urban poor residents of Borei Keila—were already

171
Sources: Heng Chhun Oeurn and Prem Chap, junior researchers and temporary residents in the new
buildings of Borei Keila, interviews by author, Phnom Penh, 12 January and 16 January 2008; Sun Saran,
Head of SUPF Office for Prampi Makara district and Borei Keila community leader, interview by author,
Phnom Penh, 15 January 2008; and Sam Sarin, community leader and building leader of new Building “B”,
interview by author, Phnom Penh, 12 August 2009.
172
Heng Chhun Oeurn and Prem Chap, junior researchers and temporary residents in the new buildings of
Borei Keila, interviews by author, Phnom Penh, 12 January and 16 January 2008.
173
Ibid.

167

largely usurped by higher-income groups from outside the settlement, and a clear segmentation of
income groups had taken place, with the poorest residents living on the top floors and better off
residents (mostly people from outside the community) living in the more desirable lower-floor
apartments.
Meanwhile, further consolidation was also taking place at the highest levels. In July
2007 the government confirmed that the Ministry of Youth, Employment and Sports was selling
its share of the land to Phan Imex Co., thus making the company the landowner of the entire
Borei Keila site, except for the 2 hectare community portion
174
. In return for the land at Borei
Keila, the ministry would obtain land owned by the company outside the city, in neighboring
Kandal province. Thus, Phnom Penh’s one and only land sharing project was gradually starting
to resemble one of the many land swaps taking place all over the city.
3.7 The Case Study Settlement of Dey Krahom
Dey Krahom (“red soil” in Khmer) was the second-largest and most densely populated of
the four settlements earmarked for social land concession and land sharing by Prime Minister
Hun Sen. The settlement comprised two administrative villages (Village 2 covered the northern
area of the settlement, and Village 15 covered the southern portion) and seven communities
located on an L-shaped site in the strategic Bassac riverfront area of Phnom Penh. According to a
2003 survey carried out by the district authorities, the settlement officially housed 1,465 families,
although as in Borei Keila, human rights organizations and many residents challenged the official

174
The sale of the MoEYS land to Phan Imex Co. was made official by Letter No. 1073, dated 31 July
2007.

168

population statistics as far too high (see also section 3.7.5)
175
. The district Office of Land
Management declared that the size of the Dey Krahom settlement area was 3.6 hectares
176
.
The settlement was sandwiched in between the Russian embassy compound to the south,
and on the east and west, by two historic apartment complexes dating from the 1960s—the
“White Building” (also known as Boding) and by the original “Olympic Village Apartments”
complex. The White Building was inaugurated in 1963 to house staff of the Municipality of
Phnom Penh (Grant Ross and Collins, 18-9), but has now been occupied by new arrivals, and is in
a dilapidated condition. The Olympic Village Apartments building was transformed by a
Malaysian developer in the early 2000s and is now a commercial complex called the Phnom Penh
Center, housing a private university and various businesses.
3.7.1 History of the Settlement
The strip of land between the White Building and the Olympic Village Apartments was a
landscaped public space until the first informal settlers occupied the land and built their makeshift
housing there in the late 1980s and early 1990s. Municipal officials claimed that the land was
state land, registered to the Municipality of Phnom Penh and the Ministry of Culture. The first
settlers acknowledged that they knew they were illegally occupying government land but that,
over time, they received family books from the commune, so they felt reasonably confident in
their “ownership”
177
. As in Borei Keila, many of the families who first settled in the area arrived
from the refugee camps on the Thai border, and were drawn to Phnom Penh by the city’s boom

175
New (replacement) community leaders of Dey Krahom claimed that only around 800 families lived in
the settlement in 2003. Source: Meeting with new Dey Krahom community leaders and selected residents
at PILAP office, 20 October 2006.
176
This corrected an earlier survey by the authorities in 2003—reproduced in Letter No. 875 announcing
the social land concession projects—which proclaimed that the size of the area was 4.7 hectares.
177
Meeting with selected Dey Krahom community members at PPUPRP project office, Phnom Penh, 1
December 2003.

169

following the Paris Peace Conference and the arrival of UNTAC forces. The migrants were
joined by groups of artists, who were attracted to the area because of their affiliation with the
nearby Bassac (Preah Suramarit) Theater.
The settlement grew rapidly in the early 1990s, as poor families were drawn to the area’s
strategic location. The settlement’s proximity to the busy Bassac riverfront and surrounding
wealthier neighborhoods enabled residents to make a good living through informal vending and
small trades. Progressively, one and two-storey makeshift housing filled in the entire open space
between the two buildings on either side.

The rooftops of Dey Krahom, in the foreground,
with the “White Building” in the background
(Author’s photo, July 2006)

Densely-packed houses made from wood and zinc
on the main lane through Dey Krahom (Author’s
photo, February 2006)
Figure 10: The settlement of Dey Krahom

In 1994 residents received an official eviction notice. Residents responded by sending a
petition to Norodom Sihanouk, requesting permission to stay in the area permanently, but they
received no reply. No forced eviction was attempted of the Dey Krahom population, although
residents remained on edge as they witnessed how thier neighbors in the settlement of Sambok
Chab were displaced by arson in November 2001 (only to return and be evicted twice again in

170

2006). Consequently, the fear of arson as an eviction tactic gripped residents until Prime
Minister Hun Sen’s land sharing announcement in May 2003
178
.
3.7.2 The Land Sharing Proposal Fails
As part of the Prime Minister’s May 2003 land sharing announcement, the government
“conceded” 1.6 of the site’s 3.6 hectares to Dey Krahom residents, while the remainder of the site
reverted to the Municipality and was earmarked for commercial development. The planned
redevelopment of the Dey Krahom site must be understood within the context of the larger
transformation of the entire Bassac riverfront area from a mainly residential area into an area
characterized by large-scale commercial projects, government buildings, embassies, and a
proposed satellite city on Koh Pich island (see section 3.5.7). As part of this transformation,
urban poor settlements in the area were the first to be uprooted, as the Sambok Chab evictions and
the eviction of Koh Pich residents in 2006 attested.
Initially, a land sharing solution seemed just as feasible in Dey Krahom as in Borei Keila.
The commercial pressure in the Bassac riverfront area was very high, and developers were
certainly interested in the location of the settlement. The Municipality expressed confidence that
developers would be interested to enter into a cross-subsidy arrangement to pay for community
housing, as in Borei Keila. At the suggestion of the U.N.’s Phnom Penh Urban Poverty
Reduction Project, the Municipality organized a public bidding process to invite developers to bid
for the right to develop the site as part of a land sharing framework. The NGO Urban Resource
Center provided technical assistance to residents in the formulation of various housing design
options.

178
Meeting with selected Dey Krahom community members at PPUPRP project office, Phnom Penh, 1
December 2003.

171

But there was one major problem: the narrow shape of the Dey Krahom site did not lend
itself well to accommodating both commercial development and community housing on terms
acceptable to both developers and residents. Developers were only interested to share the site if
residents were re-housed in very dense living conditions, thus leaving more space for commercial
development
179
. But Dey Krahom residents had already made it clear that they were not prepared
to live in high-rise apartments: they had earlier rejected an offer by Phan Imex Company to re-
house residents in an eight-storey apartment tower, with individual units even smaller than the
ones the company was building for residents of Borei Keila
180
. By early 2004 the Dey Krahom
land sharing plans had reached an impasse, and alternatives to land sharing were being
considered. Vice-Governor Chev Kim Heng suggested that, if no developer could be found, then
residents should consider constructing their own housing on their portion of the concession. At a
meeting between Municipal officials, advisors from the PPUPRP project and community leaders,
held at the Municipality on 24 February 2004, community leaders countered that self-build
housing would take too much time, and that the savings of community members were inadequate
for this purpose. Moreover, the sizes of the housing units would likely have to be small, if all
residents had to be accommodated on the community portion of the concession, in a low-density
configuration. Given those circumstances, the community leaders pointed out that “some
residents” preferred instead to receive compensation and leave Dey Krahom altogether. This was
the first indication that community leaders were pursuing alternatives to on-site redevelopment.

179
Based on their discussions with developers, community leaders confirmed that developers required at
least 2 hectares of land to make the land sharing project commercially attractive. Source: Meeting with Dey
Krahom community leaders and Municipal officials, Municipality of Phnom Penh, 24 February 2004.
180
Soeun Yei Ku, Dey Krahom community leader, interview with author, Phnom Penh, 4 December 2004.

172

3.7.3 A “Voluntary” Land Swap?
As the land sharing plans appeared doomed, the leaders of the seven Dey Krahom
communities began meeting directly with developers to consider a land swap whereby, in return
for the right to develop 100 percent of the Dey Krahom site in the city center, a developer would
deliver—for free—housing for the residents on the outskirts of the city, where the company
owned land. The land swap model was being applied all over Phnom Penh during the 2000s, and
the community leaders considered that the model offered several advantages, including the
prospect of larger houses than they could obtain in the city center through land sharing, and at
lower densities. In addition, the community leaders reasoned that land prices in the periphery of
Phnom Penh were rising steadily, so their new houses and plots of land would represent a good
investment
181
.
With the intermediation of members of SUPF and the chiefs of Chamkarmon district and
Tonle Bassac commune, the leaders negotiated with five different developers during the course of
2004, including Mong Reththy Company, Phan Imex Company, Meng Hua Company, Veng Huor
Company, and 7NG Company. The leaders broke off negotiations with the first four companies
because the compensation packages they offered were not considered to be substantial enough:
either the design of the alternative housing offered to residents was unacceptable (multi-storey
instead of one-storey), or the relocation land being offered by the company was too far away from
Phnom Penh, or the boundaries of the relocation land were not sufficiently clearly marked
182
.
The fifth developer, 7NG Company, offered the community leaders a package that the leaders
considered adequate compensation. The company promised to provide each family a one-storey
concrete house of 4 by 10 meters, along with basic utility connections, on a 50 hectare site owned

181
Soeun Yei Ku, Dey Krahom community leader, interview with author, Phnom Penh, 4 December 2004.
182
Ibid.

173

by the company, located 20 km from Phnom Penh, in Damnak Trayeung village, Chom Chao
commune, Dangkor district. The area was still quite remote and far from employment
opportunities, but the developer pledged to construct a paved road connecting the new site to the
main road, and assured the leaders that a garment factory would be built right next to the new site
that would create 500 jobs for community residents
183
. In addition, the company promised to
deliver the following on-site facilities: a market, two pre-schools; a health post, a community
office; a village office; and two buildings to be used as training centers. During the first five
years, residents would obtain temporary land titles issued by the company. After five years, they
could apply for regular land titles with the local authorities. In addition to the 1,465 houses for
the Dey Krahom population, the new site would accommodate hundreds more apartments in more
expensive price categories. Proceeds from the sale of these commercial apartments would be
used to finance the construction of housing for Dey Krahom families.
The land swap deal was agreed between the community leaders and the 7NG Company
“on behalf of” the Dey Krahom population, but without the full knowledge or support of a
majority of community members. In fact, fearing that many residents would object to the
relocation to a new site 20 km from Phnom Penh, community leaders deliberately kept residents
in the dark about the deal until roads and drainage could be installed at the new site, so that by
then “residents would be convinced” that the relocation was going to proceed
184
.
A contract between the 7NG Company and the Dey Krahom “community” was
formalized on 16 January 2005. It was thumb printed by all community leaders and deputy
leaders and signed by the Director of 7NG Company, the chiefs of the district, the commune and

183
Cheang Sokha, “Landlords Keep Evicted Village 15 Renters Homeless,” Phnom Penh Post, 11-24
August 2006.
184
Soeun Yei Ku, Dey Krahom community leader, interview with author, Phnom Penh, 4 December 2004.

174

the two villages of the old site, and by a representative of SUPF. As an alternative for those
families who did not wish to relocate outside of the city, the company offered cash payments for
these families to leave the site. The cash compensation, however, was much lower than the
current market value of land in Tonle Bassac commune
185
. The contract was binding on all
residents currently living in Dey Krahom: it stipulated that by the time 7NG Company had
constructed between 200 and 300 houses in the new site, all community residents would be
expected to vacate the old site in the city center and turn over their properties to the company.
The company’s offer of free housing or cash settlements applied only to house owners—no
provisions were made for families currently renting in the settlement
186
. The company and
community leaders agreed to plan the relocation of residents in stages, between August and
December 2006. A battle now loomed that would split the Dey Krahom population into two
groups: those that would follow their leaders to the new site and those that refused to relocate.
3.7.4 A Population Divided
After the deal was inked, local officials, community leaders and village chiefs organized
a meeting to inform residents, but residents were not given a copy of the contract
187
. Tensions
soon mounted as a group of families declared that they would refuse to accept the deal: they
considered themselves the legitimate owners of their land and homes, and insisted that their land
rights had been violated by “corrupt” community leaders and local authorities by means of a
secret agreement with the developer
188
.

185
Licadho Canada, “Dey Krahom Community Land Case Explained” (Phnom Penh: Licadho 2008), 19.
186
Ibid, 16.
187
Ibid, 19.
188
Meeting with opponents of the land swap at PILAP, Phnom Penh, 20 October 2006.

175

The opponents decided to challenge the legitimacy of the community leaders who had
signed the contract with 7NG Company, as well as the legality of the contract. In February 2005
residents opposed to the land swap fired the leaders who had signed the contract and appointed 23
new community representatives to replace them. The new representatives organized a petition
that was submitted to the Municipality, requesting that the contract with 7NG Company be
annulled because it was signed without the knowledge of the residents. The petition was thumb
printed by 804 families—a majority of the settlement’s population. An NGO, Legal Aid of
Cambodia, helped the opponents to file a request for intervention to the Governor of Phnom
Penh. This was followed by similar appeals to the Prime Minister, the Phnom Penh Municipal
Court, and the district Office of Land Management. None of the appeals elicited any response
189
.
Only lawmakers from the opposition Sam Rainsy Party declared their support for the opponents.
On 15 January 2006 the official inauguration ceremony of the community housing took
place at the new site in Chom Chao commune, attended by representatives of the company and
Municipal Vice-Governors. Residents were transported by bus—at the company’s expense—to
attend the ceremony and to inspect the new houses under construction. For many residents, it was
the first time they had visited the new site. A first lottery was held in August 2006, at which 344
families were assigned their new housing. All the previous community leaders were among this
first group of families to relocate. An earlier lottery planned for July 2006 had to be canceled
because not enough residents signed up to participate
190
.

189
Licadho Canada, “Dey Krahom Community Land Case Explained”, 7.
190
Pin Sisovann and Jason MacBride, “Tonle Bassac Residents’ Lucky Draw Suspended”, Cambodia
Daily, 19 July 2006.

176

In the new site, a family preparing to relocate
inspects the new community housing for the first
time (Author’s photo, January 2006)

In the old site, police armed with machine guns,
pliers and axes confront residents who refuse to
relocate (Source: Licadho, August 2007)
Figure 11: The two faces of Dey Krahom

Just over a week later a group of 200 residents, holding aloft portraits of Prime Minister
Hun Sen, his wife Bun Rany, and King Norodom Sihamoni, held a demonstration on the edge of
the Dey Krahom settlement to protest the land swap and the “corrupted community leadership”
191
.
Opponents of the land swap pleaded with the Prime Minister to intervene and resolve the
standoff, and to implement the social land concession granted to them in 2003, which allowed
them to remain on site.
Meanwhile, as the first relocations were commencing, the 7NG Company and the local
authorities kicked off what human rights agency Licadho claimed was a steady “campaign of
harassment and intimidation” to coerce residents opposed to the land swap to surrender their land
to the company
192
. One of the first tactics was the introduction of a new discourse to brand the
residents resisting relocation as rabble-rousers, “brothel owners”, and speculators—in other

191
Pin Sisovann, “200 Tonle Bassac Residents Protest Imminent Relocation”, Cambodia Daily, 15 August
2006.
192
Licadho, “7NG Company and Phnom Penh Authorities Intent on Inciting Disorder in Cambodia’s
Capital”, press release, 9 January 2008.

177

words, as people behaving illegally and immorally—who were just out to extract more money
from the company before they left the site.
On the night of 16 August 2006 police entered the settlement in the first of many
operations to clear scores of tents inhabited by renters claiming to be newly homeless since the
shacks they were previously renting in Dey Krahom had been destroyed after their owners had
resettled
193
. From that point on 50 police officers were deployed full-time inside the settlement to
prevent former residents and others from resettling plots of land cleared of people
194
.
Another target were vendors operating stalls at the Dey Krahom market. District
authorities issued an instruction in 2006 to dismantle the market stalls, which were a source of
income for many residents. Since then, the authorities made regular attempts to disrupt the local
market, actions which one resident likened to the market closings carried out during the Khmers
Rouges regime
195
. In January 2008 police and company workers put up roadblocks on one of the
access roads to the settlement, in order to prevent supplies from being delivered to the Dey
Krahom market. Later, company workers ignited a truck stationed next to the roadblocks,
apparently in order to foment disorder, which could then be used as a pretext for the police to
move in and evict residents
196
.
The intimidation tactics were punctuated by occasional violent eviction attempts. On 29
August 2007 over 100 heavily armed military police, accompanied by just as many construction
workers hired by 7NG Company, attempted to forcibly evict the opponents. A violent

193
Pin Sisovann, “Destroying Tent Homes, Police Meet Resistance”, Cambodia Daily, 18 August 2006.
194
Pin Sisovann, “50 Police Officers Watch for Village 15 Resettlers”, Cambodia Daily, 21 August 2006.
195
Pin Sisovann, “Angry Villagers Confront Tonle Bassac Removal Force”, Cambodia Daily, 29 August
2006.
196
Licadho “7NG Company and Phnom Penh Authorities Intent on Inciting Disorder in Cambodia’s
Capital”, press release, 9 January 2008.

178

confrontation ensued during which residents were assaulted and several houses destroyed, but the
eviction was unsuccessful. Military police returned the following day, but stayed at a distance, on
the edge of the site. Meanwhile, the 7NG construction workers “persuaded” the hold-outs to
leave the old site
197
.
The strategy of the previous community leaders to present the residents with a fait-
accompli was gradually producing results. The newly appointed community representatives
estimated that 75 percent of Dey Krahom residents had initially rejected the land swap, but they
acknowledged that during 2006 and 2007 more and more families were agreeing after all to
relocate
198
. The original community leaders claimed that this was because residents saw for
themselves that the infrastructure, services and facilities in the new site were so much better than
in the old site
199
. More likely, however, is that a majority of families was reluctantly
acknowledging that the land swap was really going to take place, and that there was no possibility
for them to stay in the old site. The physical and psychological intimidation tactics were also
taking their toll, as opponents were losing the will to fight. In addition, a growing number of
opponents of the relocation were accepting monetary compensation from the developer to leave
the old site.
The final saga in the history of the original Dey Krahom settlement occurred on 24
January 2009. In the early hours of the morning law enforcement agencies blocked off the entire
area and began forcibly evicting the remaining people still occupying the settlement. Housing
rights organizations claimed that these people included approximately 150 families who were
former owners of properties in Dey Krahom, and who were holding out for greater compensation,

197
Licadho, “Developer 7NG Continues to Target Villagers on the Footsteps of Cambodia’s National
Assembly”, press release, 10 October 2007.
198
Meeting with opponents of the land swap at PILAP, Phnom Penh, 20 October 2006.
199
Soeun Yei Ku, Dey Krahom community leader, interview with author, Phnom Penh, 14 January 2008.

179

and over 250 families who included renters and/or market vendors who still lived and slept on the
site
200
. The 7NG Company subsequently transported a majority of the new evictees to the
relocation site at Chom Chao commune where, for lack of better alternatives because they were
ineligible for new housing, they remained camped out at the front of the new site under
emergency plastic sheets provided to them by NGOs.

After the final eviction of residents from Dey
Krahom, the 7NG Company closed off the area for
good with a metal fence (Author’s photo, August
2009)

The last group of residents, market vendors and
various renters from Dey Krahom squats outside
the 7NG relocation site, with no right to new
housing there (Author’s photo, August 2009)
Figure 12: The final clearance of the Dey Krahom site

3.7.5 The Relocation Site Expands
While tensions continued at the old site, the relocation site (which by now had received
the official name of Borei Santepheap or “peace complex”) was gradually expanding, both in
terms of population and services. Most relocated residents had obtained public water and
electricity connections, although as the population of the new site increased the bulk water
connections apparently proved to be inadequate to handle the additional demand, so that by early

200
COHRE, “Legal Analysis of the Forced Eviction of Dey Krahom Community”, 29 January 2009.

180

2009 regular rationing of water was taking place
201
. Some residents were already expanding their
houses by adding a second floor. One pre-school had been set up by a private NGO, but there was
still no health post. The market was crowded with vendors and expanding from year to year. A
Korean-owned garment factory opened next to the relocation site in early 2007. The factory
employed hundreds of families living on site to do fabric cutting work at home, as the company
had promised. For men, the other main source of employment for relocated residents was
construction work on site, for the 7NG Company.

The first families settle in the new site.
(Author’s photo, October 2006)

While some owners left their houses unoccupied,
other owners were already expanding their new
homes (Author’s photo, October 2006)
Figure 13: The Dey Krahom relocation site

But even as the relocation site was gradually coming alive, only about half of the
intended beneficiaries (poor families from Dey Krahom) were actually settling there. As in Borei
Keila, due to the tremendous opportunities to engage in speculation, there was a lucrative
business of buying and selling of community names prior to the relocation. In fact, the human

201
S. Y. and M.P, erstwhile Dey Krahom community leaders, interview with author, Phnom Penh, 12
August 2009.

181

rights agency Licadho Canada claimed that the settlement’s official population statistics were
deliberately inflated in order to enable outsiders, through bribes, to be placed on the list of
beneficiaries who would receive new apartments
202
. The business in so-called “ghost names”
(i.e., the names of people registered as living in the settlement without actually staying there)
benefited outside buyers as well as community leaders and the company. One member of the
people’s organization SUPF who had never lived in Dey Krahom boasted that she obtained two
houses in the new site by paying for ghost names. First, she received one community name for
free on account of her affiliation with SUPF. Subsequently, she paid community committee
members a “commission” in order to buy a house directly from the company for $1,500—an
amount that was significantly less than the official purchase price if she would have come in as an
outside buyer
203
. The value of the land in the original Dey Krahom site would have more than
offset the cost to the company of the subsidized home sales in the new site.
Moreover, in the case of residents who did finally opt to relocate, not all ended up
actually staying in the new site. Former community leaders acknowledged that only about 50
percent of relocated Dey Krahom families were in fact staying in the new houses on a full-time
basis. Approximately 30 percent of families had sold their new units immediately to outsiders
looking for an investment opportunity. Another 20 percent kept their allocated houses padlocked,
and had moved back to the city—perhaps waiting to sell at a future date, when land prices were
even higher. Former community leaders were now active as informal brokers in the new site.
They reported that land prices in the new site were rising rapidly, from an average of $3 per
square meter in mid-2005 to $30 per square meter in early 2008, even before the issuance of
individual land titles. Individual housing units were selling for between $6,000 to $7,000, and

202
Licadho Canada, “Dey Krahom Community Land Case Explained”, 6.
203
M. K., interview with author, Dey Krahom new site, 21 October 2006.

182

around $13,000 if they are located near the market
204
. Under those circumstances, even those
relocated residents who had not sold yet might still be tempted to do so in the future.
3.8 The Case Study Settlements of Railway A and Railway B
The settlements of “Railway A” (Roteh Pleung A, also known as Santipheap) and
“Railway B” (Roteh Pleung B) were located a few hundred meters to the south of Phnom Penh’s
largest lake, Boeung Kak, and the city’s main railway line. The two settlements were located on
either side of Road 132, which was widened and tarred by the Municipality in 2004. Though
adjacent to each other, the settlements were administered by two different districts: Railway B
was located in Tuol Kork district and Railway A was located in the central Daun Penh district.
The area on which the two settlements was located is mostly wetland (boeung or seasonal
lake), and is partially flooded during the rainy season. The State Railway Authority of Cambodia
is the landowner of the entire area, including the adjacent railway tracks and depots.
Railway B was the larger of the two settlements, officially comprising 255 families
205

spread out on a 10 hectare site flanked by Russian Boulevard on the south, Kim Il Sung
Boulevard to the West, Road 132 to the East, and train depots to the north. The population of
Railway B was concentrated mainly in the Eastern portion of the settlement, along the main
access road to the community from Road 132. The Western half of the settlement was more
sparsely populated. The settlement of Railway A was much more compact and densely inhabited:

204
Morm Proeun and Souen Yei Ku, former community leaders of Dey Krahom, interview by author, Dey
Krahom new site, 14 January 2008.
205
The official population count did not include renters: the population of Railway B settlement was 432
families if renters are included.

183

officially it housed 70 families
206
on 2.1 hectares in between the railway line to the north, Road
132 to the West, empty Railway Authority land to the East, and commercial development on the
south.

Footbridge over the boeung in the dense Eastern
half of Railway B (Author’s photo, March 2006)

Parts of Railway B settlement were only sparsely
populated. The company had filled in much of the
wetland already (Author’s photo, March 2006)
Figure 14: The settlement of Railway B

In both settlements, residents constructed their houses from wood, and the houses were
built on stilts. Wooden footbridges over the boeung connected many houses, although these
bridges were almost all in a perilous state. The original bridges were constructed by the residents,
and subsequently repaired or replaced with the help of donor organizations, such as the PPUPRP,
but they had not been repaired in years.
3.8.1 History of the Railway Settlements
People started settling on land owned by the State Railway Authority soon after Phnom
Penh was repopulated following the end of the DK era in 1979. In the areas now known as
Railway A and B, the Railway Authority gave out land for free to some of its employees, so that

206
The official population count did not include renters: the population of Railway A settlement was 85
families if renters are included.

184

they could settle near their place of employment. Outsiders also moved in: initially the first
families to settle in the area simply grabbed land for free, as it appeared to be uncultivated and
unclaimed
207
. By the mid-1980s a land market had started to develop; land originally given out
for free was sub-divided and sold by Railway officials and employees, as well as by outside
brokers and early settlers. Outsiders who wanted to settle in the area now had to pay small fees to
the village chief, to the commune and to the Railway Authority
208
.

Road 132, newly tarred and widened by the
MPP, with Railway A settlement on the right
hand side (Author’s photo, March 2006)

A compact settlement, characterized by wooden
houses and footbridges over the boeung
(Author’s photo, April 2006)
Figure 15: The settlement of Railway A

In the early 1990s, the then-Governor of Phnom Penh, Hok Lundy, announced that
settlers on Railway Authority land would be evicted. This had the desired effect of scaring some
families away, but the majority of families refused to move. In December 1995, in response to
the eviction attempt, the community at Railway B was “organized” by SUPF. SUPF conducted a
survey of the community, which at that time consisted of only 25 families; started a community
savings scheme; elected community leaders; and brought residents together to discuss and resolve

207
May Tonnak, Railway B community leader, interview by author, Phnom Penh, 29 March 2006.
208
Yos Sophat, Railway A community leader, interview by author, Phnom Penh, 4 December 2004.

185

problems. On the other side of Road 132, in the settlement of Railway A, the community was
slower to develop: the first families settled there only in 1987. Railway A was organized by
SUPF in February 2000; at that time the settlement comprised 81 families.
In 1998 the Hassan Cambodia Development Company—one of the largest developers in
Cambodia, owned by CPP lawmaker and adviser to Prime Minister Hun Sen, Okhna Osman
Hassan—signed a 70-year lease contract with the Ministries of Transport and Finance for a 10
hectare area of State Railway Authority land covering the settlements of Railway A and Railway
B
209
. On this site the company planned to erect one of the largest mixed-use developments in
Phnom Penh, comprising—in the first phase—a shopping mall, financial center and
condominiums. The project apparently enjoyed the backing of investors from China, Malaysia,
South Korea, Thailand and Japan
210
. At the time the lease was signed, the combined population
of the two Railway settlements comprised 169 families, according to a survey conducted by the
company. A condition of the contract was that the company would be responsible for finding a
“suitable solution” for the residents of the two informal settlements.
The company put up a fence around the perimeter of its lease area and started dredging
and filling in the wetland where people had not settled. The company started conducting regular
surveys of the population of the Railway A and B settlements, and it found that the number of
families in both settlements kept growing. Three years after the lease agreement, the population
of Railway A had increased to 100 families, and the population of Railway B was 444 families.
Residents recalled that around this time the company tried many “tricks” to scare
residents into leaving the settlements, including threatening arson and the use of bulldozers to

209
The developable land area is only 7.8 hectares, as the lease area includes railway tracks and depots that
remain state property. Okhna Osman Hassan, Director, Hassan Cambodia Development Company,
interview by author, Phnom Penh, 11 March 2004.
210
Lon Nara, “10 Hectare Site for Mall, Hotel Development”, Phnom Penh Post, 7 June 2002.

186

evict them
211
. Moreover, in 2001 the Chief of Tuol Kork district tried to scare residents of
Railway B into accepting a relocation deal that included a meager compensation package of $200
and some building materials, by saying that that was the best offer they were likely to get
212
. But
residents claimed that the scare tactics were counter-productive: they brought the community
closer together and encouraged residents to seek support from NGOs, donors, and human rights
organizations to lobby on their behalf
213
.
3.8.2 Community Land Sharing Proposals
Even before the Prime Minister’s announcement of the government’s support for slum
upgrading, the Railway A and B populations—with the technical assistance of the NGO Urban
Resource Center and SUPF—had proposed a form of land sharing as a way to resolve the
standoff with the Hassan Cambodia Development Company, and as a way to finance new
community housing. At a meeting with the Municipality and the company in April 2003,
community leaders of Railway A and Railway B presented separate plans, each requesting 25
percent of their sites to be reserved for community housing, to be built by the company. The
plans of both communities proposed single-storey houses—reflecting the tradition that families
expand their houses vertically as household members and household incomes expand over time.
In the case of Railway A the requested plots were 4.5 by 15 meters for each family; in the case of
Railway B the plots were 4 by 12 meters per family.
Unlike in Dey Krahom and Borei Keila, in both Railway settlements the land sharing
plans, including the location of the proposed community housing, were elaborated and agreed as

211
Meeting with residents of Railway A and B at PPUPRP office, Phnom Penh, 3 December 2003.
212
May Tonnak and Sok Mom, community leaders of Railway B, interview with author, Phnom Penh, 4
April 2006
213
Meeting with residents of Railway A and B at PPUPRP office, Phnom Penh, 3 December 2003.

187

part of a relatively open community process. In the case of Railway B, the plans were re-
designed with the technical assistance of URC and SUPF as well as the district. Perhaps as a
result of the small size of the populations, and the fact that in each Railway settlement there was
only one official community, the populations of the two Railway settlements were much more
cohesive and better organized than in the two larger pilot settlements.
The Municipality initially reacted positively to the community plans, but the Hassan
Cambodia Development Company was steadfast in its refusal to consider any land concession to
residents of the two settlements on the land covered by its lease agreement with the government.
In addition, the company made it clear that it was only prepared to recognize house “owners” and
not renters living in the Railway A and B settlements, even if renters had been long-term
residents in the area. For this reason, company surveys of the population in the settlements
henceforth counted only owners and not renters. The Municipality accepted the company’s
policy towards renters—as it had in Dey Krahom, where the 7NG Company also refused to
compensate renters. As a result of the company’s policy towards renters over one-third of
Railway B residents (150 renters out of a total of 432 families), and more than one-third of
residents in Railway A (over 30 families out of 81) would be excluded a priori from any future
slum upgrading plan
214
: the poorest of the poor would thus not be able to benefit from slum
upgrading in the Railway settlements.
3.8.3 The First Relocation Attempt
As discussions between the communities, the company and the Municipality proceeded,
the Hassan Cambodia Development Company made residents in both settlements two offers, in

214
Other families not on the list of eligible recipients in Railway B included four “owner” families who
rejected their community’s land sharing proposal: they consider the requested 4.5 by 15 square meter plots
to be too small, and are therefore not interested to be included on the list.

188

an attempt to vacate the land: families could receive a cash settlement of $400 and a plot of land
in a relocation site at Tang Krosang—an expanse of land owned by the company and located near
the airport, approximately 12 km from the city—or they could receive a cash sum of $700 if they
moved out of the settlement altogether, on their own accord. If families chose the first option, the
cost of relocation would be borne by the company. The company had sub-divided the relocation
site into lots of 105 square meters each, and promised residents roads, ponds, water connections
and a generator. Moreover, the site was said to be near employment opportunities in nearby
factories, and the company promised to help the relocated families to find jobs
215
. Relocated
residents were also promised land titles, to be issued by the company.
According to community leaders and community committee members, during the course
of 2002 and 2003, 145 families opted to relocate to Tang Krosang (14 families from Railway A,
and 131 from Railway B) and 51 families accepted the $700 and left to go elsewhere (1 family
from Railway A, and 50 families from Railway B).
It was not long before many residents who chose to relocate to Tang Krosang regretted
their decision, however. The Tang Krosang site was poorly equipped to sustain slum dwellers
relocated from the city: the promised water connections were never made, roads remained
untarred, and the developer’s generator was removed after only one year. Thereafter, residents of
Tang Krosang had to make do with water brought in by private vendors (at 2,000 riel or $0.50 per
barrel) and electricity generated via car batteries. Few relocated residents were able to find work
in the area. Residents felt as if the company had broken its promises to them
216
.

215
Okhna Osman Hassan, Director, Hassan Cambodia Development Company, interview by author, Phnom
Penh, 11 March 2004.
216
Leader of relocated residents at Tang Krosang relocation site, 1 April 2006.

189

Three years on, the Tang Krosang site still had
many vacant lots (Author’s photo, April 2006)

The title cards received by resettled residents
(Author’s photo, April 2006)
Figure 16: The Tang Krosang relocation site for residents of the Railway settlements

As of April 2006, between three to four years after the relocation, only 73 of the 150
original relocated families still lived in the relocation site: 77 families had since returned to
Phnom Penh, while 9 new families from outside of the Railway communities (people with factory
jobs in the area) had moved in to rent plots. Sixty-eight out of the total of 150 lots remained
vacant. Twelve of the 131 families who relocated from Railway B went back to Railway B,
while one of the 14 families from Railway A returned there. The remaining returnees went to
other locations, most of them to the center of Phnom Penh, where there were better opportunities
to make a living.
But many families who relocated never had the intention to stay in Tang Krosang
anyway; instead, speculation was the major motive. Most families who already left Tang
Krosang still kept their plots there, and were waiting for the right time to sell. Residents had
obtained individual land title cards, which were “fully recognized by the cadastral authorities”
217
,
as the company had promised, and the titles would their plots to fetch a higher price. Land values

217
Ibid.

190

were increasing rapidly in the area: whereas in 2002 a single plot in Tang Krosang cost $400,
four years later the average price was $2,700 for a plot on a side road and $5,000 for a plot on the
main road
218
. Those who held on to their titled plots, therefore, stood to gain from the rise in
land values in the area. For the company, meanwhile, the Tang Krosang relocation was a failure:
it had spent money to buy out residents of Railway A and Railway B, but it had not succeeded in
convincing enough families to leave. Worse yet, some families had even taken the company’s
money and returned to the settlements.
3.8.4 Breaking the Deadlock
Three years after the Prime Minister’s social land concession announcement, and ten
years after the company signed the lease for the area, the situation in the settlements of Railway A
and B remained deadlocked. The company’s inability to force through a “solution”, even as
projects in Borei Keila and Dey Krahom were moving full-speed ahead, could be attributed in
part to the Municipality’s support for the population of the Railway settlements. In 2005
Governor Kep Chuk Tema of Phnom Penh had assured residents of the Railway settlements that
they would never be evicted, and that the standoff with the company would have to be resolved in
a peaceful manner.
But another important explanation for why the standoff between the company and the
residents could go on for so long was the strength of the communities. Unlike in Borei Keila and
Dey Krahom, residents in the Railway settlements expressed strong support for their leaders
219
,
and in another significant contrast with the two bigger settlements, they were well aware of their
land rights. Community leaders in Railway B had briefed their residents about the Land Law

218
Ibid.
219
Meeting with residents of Railway A and Railway B, PPUPRP office, Phnom Penh, 3 December 2003.

191

2001 and distributed to all residents a copy of the Council of Ministers’ Letter No. 875 (2003)
confirming the social land concession in the two Railway sites. A portrait of the Prime Minister
inspecting the community’s land sharing plan in 2003 was displayed prominently at the entrance
to Railway B community (see Figure 1, Chapter 1). As an example of Foucauldian reverse
discourse, the portrait’s message was clear: the Prime Minister recognized the slum population,
and therefore they were no longer squatters or “anarchists”.
One other source of community strength was also significant: the residents had good
political connections. One of the two leaders of Railway B summarized their community’s self-
confidence thus:
We can wait: we have the law on our side, and we are strong, for among our
community members are police, army, and military police officers; lawyers;
journalists; NGO workers; district chiefs; and National Assembly representatives,
so they cannot just remove us from here
220
.
After its previous attempts to remove the residents had failed (including the Tang
Krosang relocation offer and the threats and intimidations before that), the company feared that
the longer the present deadlock lasted, the more the current population in the two settlements
would grow, the greater the in-migration would be, and the harder it would be to remove the slum
populations in future. It decided that a “suitable solution” for the slum dwellers now consisted of
identifying a land concession in the immediate vicinity of the Railway A and B settlements, but
outside of its own lease area
221
.
The Municipality assisted the company to identify an alternate site for the relocation of
the two communities in the city, near the present sites. Visits to three nearby locations were held

220
Sok Mom, community leader of Railway B, interview by author, Phnom Penh, 29 March 2006.
221
Okhna Osman Hassan, Director, Hassan Cambodia Development Company, interview by author, Phnom
Penh, 11 March 2004.

192

jointly with community leaders in 2005, but in all cases the existing residents in the alternate sites
refused to accommodate an influx of over 500 families from other settlements.
The company’s next strategy was to push for a “Borei Keila-type” solution: it offered to
build a multi-storey apartment building to re-house both communities (but again, excluding
renters) on the site of Railway B. The company also wanted to re-measure the land area of both
settlements because it considered that the communities were claiming an unfair portion of the
total lease area awarded to the company. On 4 July 2006 the Municipality convened a meeting
with the company and the community leaders of Railway A and B to negotiate a “compromise”
solution. The community leaders rejected the “Borei Keila-type” solution out of hand, because
the residents did not want to live in a high-rise building, and because they feared being bought out
of their new apartments and becoming squatters again, as happened to poor families from Borei
Keila who moved to the community buildings
222
. Moreover, community leaders from Railway B
refused to subject themselves to another measuring exercise, claiming that the original
measurements as contained in the 2003 land sharing agreement remained valid.
By now the Municipality was running out of patience and decided that it wanted the land
sharing standoff resolved. Municipal Vice-Governor Mann Chhoeurn pointed out that the
Municipality had tried its best to implement the social land concession as directed in Letter No.
875, but that the agreement was too “one-sided” in favor of the communities. On the other hand,
with commune elections approaching in April 2007 and national elections in July 2008, the
Municipality realized it could not be seen evicting the poor. In other words, it needed to tread a
fine line between serving business interests and delivering votes
223
. Residents of the Railway

222
May Tonnak and Sok Mom, leaders of Railway B community, interview by author, Phnom Penh, 9
August 2007.
223
Mann Chhoeurn, Vice-Governor of Phnom Penh Municipality, interview by author, Phnom Penh, 23
January 2007.

193

communities detected that a shift was underway: community leaders complained that the
Municipality was starting to pay more attention to the interests of the company than those of the
residents, and that “the parent does not care about its children”
224
.
3.8.5 The Railway Populations Disperse
Another tense stalemate broke out between the company and residents. For the first time
the ongoing uncertainty was starting to take its toll on residents and their determination to stay
united in the face of external pressure. A growing number of residents admitted to “losing hope”:
they doubted that the government would continue to support them, and doubted that a land
sharing solution would ever materialize
225
.
Beneath the surface “calm” a frenzy of secret negotiations was taking place. The Hassan
Cambodia Development Company started to approach individual families and offered them
money to clear the site. But unlike in 2002 and 2003, when the company offered families merely
$700 to leave, this time the company was offering what it called “official market rates”, based on
the size and location of each plot. This strategy had the full support of the Municipality. The
people’s organization SUPF played its familiar role as ally of the Municipality and broker and
agent of developers and tried to convince residents that accepting the money was in their own
long-term interest
226
. The company first approached families in Railway A, as they were

224
May Tonnak and Sok Mom, leaders of Railway B community, interview by author, Phnom Penh, 29
March 2006.
225
May Tonnak and Sok Mom, leaders of Railway B community, interview by author, Phnom Penh, 7
April 2006.
226
Mann Chhoeurn, Vice-Governor of Phnom Penh Municipality, interview by author, Phnom Penh, 23
January 2007.

194

considered to be the most eager to leave the site. By mid-2007 rumors suggested that all families
in Railway A had been bought out for sums between $4,000 and $10,000
227
.
Buying out residents in Railway B settlement was more difficult, however. And here
there was a twist: the Hassan Cambodia Development Company apparently faced competition
from another developer, Phum Pu Company, which claimed that it had bought part of the land on
the social land concession site in 2003
228
. Community leaders indicated that whereas the Hassan
Cambodia Development Company was offering families compensation of around $11,000 to
leave the site, Phum Pu Company was approaching residents with compensation of between
$13,000 and even $20,000. By January 2008, 90 families had apparently settled with the Hassan
Cambodia Development Company and 100 families had settled with Phum Pu Company
229
.
Another 90 families in Railway B still had not settled with either company, including the village
chief, who refused to negotiate with either developer. The “hold-out” families were benefiting
from the competition between the two developers to drive up their price: one family stated that it
was holding out for compensation of at least $40,000, which it felt was the minimum price
necessary to afford decent replacement housing in Phnom Penh
230
.
By the end of 2008—more than five years after the social land concession
announcement—most of the residents had started to disperse. The leaders of Railway A and
Railway B were not around to witness the end of their communities, as they had already moved

227
May Tonnak and Sok Mom, leaders of Railway B community, interview by author, Phnom Penh, 9
August 2007.
228
Some residents questioned whether the two companies were really separate legal entities, and suggested
that the introduction of a fictitious new company was really a ploy by the Hassan Cambodia Development
Company to get residents to move out faster.
229
May Tonnak and Sok Mom, leaders of Railway B community, interview by author, Phnom Penh, 13
January 2008.
230
S. N., resident of Railway B, interview by author, Phnom Penh, 15 January 2008.

195

out. Remaining residents accused their leaders of working as agents for the Hassan Cambodia
Development Company during the last few months, and of betraying their communities. They
claimed that each leader had five names on the list of community members and therefore
pocketed compensation packages worth at least $50,000
231
. Perhaps the Railway A and Railway
B communities might have remained on site if they would have adjusted their housing
preferences according to what the developer was willing to pay. But in the end residents and
community leaders opted for larger pay-offs in the short term. Thus, Railway A and Railway B
joined Dey Krahom on the list of land sharing projects in Phnom Penh that eventually succumbed
to land market forces and speculation.

231
S.N., resident of Railway B, interview by author, Phnom Penh, 31 October 2008.

196

CHAPTER 4: THEORETICAL FRAMEWORK
4.1 Overview of Chapter 4
Chapters 2 and 3 described how, at the start of the 21
st
century, Phnom Penh was a city
being reborn after decades of civil war and political strife. For the city’s poor, the renaissance
provided opportunities as well as threats. On the one hand, the economic boom during the first
decade of the century offered greater prospects for eking out a living somewhere in the city’s
growing informal economy. But at the same time Phnom Penh’s rapid growth was turning the
city into a sharply contested space. Nowhere was this contest more dramatic than in the land and
housing markets: commercial developers were steadily privatizing all available city space and
forcing out economically weaker market actors, by any means possible. The result was
dispossession of the legal rights of tens of thousands of poor urban households and a structural
deficit of low-income housing and living space for Phnom Penh’s poor.
In the case of the city’s slum dwellers, the battle for living space is physical as well as
political, material as well as symbolic and discursive. At one level, the slum dwellers’ struggle is
literally a struggle for a “place in the city”—a quest to retain their homes and livelihoods in an
area where many have lawfully settled for over a decade. But at a more fundamental level their
struggle is about power and “citizenship”, as defined by official recognition and respect, and
translated in concrete terms through the attainment of formal land rights and access to public
services. The slum facing redevelopment pressure becomes a metaphor for this power struggle.
In Chapter 4 I examine theoretical approaches to power in the context of slum dwellers,
in both the literature and in development practice. Specifically, I am interested in theoretical

197

perspectives on the importance of structure and agency in shaping the nature of slums and the
survival strategies of slum dwellers. In doing so, I investigate the slum as a social construct,
shaped by competing discourses. An analysis of these discourses has direct relevance to
development practice, as these discourses contain within them their own sets of assumptions
about the slum and its residents, which in turn determine the kinds of policies and programs that
are aimed at the slum.
Research perspectives, problem description and policy recommendations related to slums
have always been shaped by contrasting political positions (Peattie and Haas, 158). These
positions have traditionally adopted some radically different points of view about the merits of
slums and slum dwellers’ capacity for agency and subjective power to improve their lives.
Development programs and policy have at times blurred these contrasting positions (Peattie and
Haas, 159), but for an understanding of slum dwellers’ power these political and theoretical
positions remain significant.
In this chapter I will examine two extreme perspectives in the literature on slums. I will
label the extreme positions as “slums of despair” and “slums of hope”, using two categories
identified by Charles Stokes (1962) in his theory of slums. Stokes’ original categories referred to
the psychological response of slum dwellers vis-à-vis their social status: slums of hope are places
where the residents indicate an intention to better themselves, and have confidence that they can
do so; by contrast, slums of despair are settlements where residents have a negative sentiment,
including a negative estimate of the probable outcome of their attempt to change their status
(Stokes, 189).
I use the categories of Stokes more broadly, to include also the perception and
expectations of outsiders towards the slum. In section 4.2 I will describe several theories in the
“slums of despair” tradition; I will argue that this tradition reflects the structural determinist view

198

of slum dwellers and the urban poor as victims of a larger system. This is a largely Marxist
perspective, which ultimately denies the poor the capacity to deploy their agency to better
themselves. By contrast, the “slums of hope” tradition provides a much more optimistic view of
slum residents’ capacity for agency. In section 4.3 I describe several theories in the “slums of
hope” tradition; I suggest that this tradition reflects the pluralist perspective in sociology, which
tends to underestimate the structural obstacles facing the urban poor in the modern slum in their
effort to get ahead.
In section 4.4 I present a “synthesis view” of the importance of both structure and agency
in the battle for living space and citizenship waged by slum dwellers. I argue that slums are
neither places of despair nor places of hope, but are better understood as “spheres of force
relations”, where power relations are dynamic and where an analysis of residents’ actions needs
to take account of the importance of individual agency as well as structural impediments. This
synthesis view is informed by the French philosopher Michel Foucault and his pronouncements
on “capillary power”. In section 4.5 I will argue that Foucault’s views on capillary power have
great importance for an understanding of the process and outcome of the battle around the four
land sharing settlements in Phnom Penh—and for wider development policy aimed at slums and
slum dwellers.
4.2 Slums of Despair
Despair over slums and the conditions of their residents has traditionally come in two
different varieties in the theoretical literature and in development practice. On one side, slums are
viewed as a source of pathological conditions. On the other side, slum dwellers are viewed as
victims of wider social and political forces beyond their control. Both perspectives are to varying

199

degrees structurally determinist, in the sense that they consider slum dwellers’ subjective
power—the ability of the urban poor to improve their condition—to be limited.
In the first category of pathology I will describe two perspectives: the first is the
discourse of urban planners, which at its genesis in the late 19
th
and early 20
th
centuries advocated
a modern, rational city which had no place for the poor on account of their perceived
estrangement from modern urban life. The second perspective on pathology invokes the
perceived cultural characteristics of poverty; it holds that there is a “culture of poverty” that is
ingrained in slum dwellers and which presents an obstacle to their advancement.
In the second category of slums and slum dwellers as victims I describe two perspectives
from the left based on traditional Marxist elements of power that are still current in neo-Marxist
analyses of the urban poor. The first perspective is a version of the Marxist class struggle: the
urban poor are powerless in the face of socially dominant interests that are aligned with capitalist
accumulation. In the second perspective, the urban poor are not only caught in an economic and
political web of dominant interests, but they are also the victims of an ideological hegemony
imposed by these same interests.
4.2.1 Planning Discourse and the Modern, Rational City
Slums have been a source of despair and shame for city authorities and politicians ever
since the beginning of modern planning. The genesis of urban planning, during the late 19
th
and
early 20
th
centuries in Europe, North America and parts of Latin America, brought with it its own
discourse—a “science of city planning”—in pursuit of the modern, rational city. The city was
portrayed as a modern and rational utopia (Fishman, 1982).
One of the standard-bearers of this new city-making was Baron Georges-Eugène
Haussmann, whose transformation of Paris as the city’s prefect during the second half of the 19
th

century would leave a lasting mark on the physical template of the city. In an effort to

200

“modernize” Paris, Haussmann “aligned, embellished and monumentalized the city” but most
importantly he “opened up” wide swathes of the historic core of Paris by cutting new streets and
boulevards—a process Émile Zola described as “radical surgery accomplished by saber” (Jordan,
89-90). The dense urban fabric of medieval and classical Paris made way for a “severe
rectilinearity of a transformed city” through “obsessively straight streets” and the striking
regularity of building façades (Jordan, 89). The “Haussmannisation” of the city left little room
for the poor: the opening up of Paris for thoroughfares and government buildings caused the poor
to flea the city as Haussmann’s grandes avenues replaced many “wretched quarters”, but no
provisions were made for the lower-income classes displaced by the building process (Bennett,
279).
With the establishment of the French protectorate over Cambodia in 1865, the rational
urban planning discourse of the period was exported to Cambodia, and particularly to Phnom
Penh, which the French wanted to transform into a “modern” capital city. The straw and wooden
huts and dirt lanes that lined the banks of the Tonle Sap and Bassac rivers made way for an
“urbanism of engineers”, characterized by new public administration buildings, public spaces,
and a uniformity of space marked by straight, tree-lined boulevards and streets, mandated at 20
meters wide, directly inspired by Haussmann
232
.
All over the world, one of the main components of the discourse of city planners in this
period was the disciplining of society through social reform. The aim was to modify the daily
attitudes and behaviors of the population through the introduction of certain rules and patterns: a
project to discipline society through the city. Planners “invented a social question” that arose in

232
Christiane Blancot and Aline Hetreau-Pottier, “1863-1953: Une ville nouvelle dans un site d’occupation
ancienne”, quoted in Atelier parisien d’urbanisme, Phnom Penh: développement urbain et patrimoine, ed.
N. Starkman (Paris: Ministry of Culture, 1997), 32-3.

201

the cities and “built a representation of daily life called ‘urban problems’” (Outtes, 146-8). This
discourse was applied most enthusiastically in the area of housing. Planners discovered the
dwellings of the urban poor and even the working class to be dirty and dangerous (Outtes, 148).
Just as Foucault described how prisons were born in the 19
th
century to transform criminals and
eliminate delinquency as part of the state’s discourse of “criminology”, planners, as part of the
discourse of a “scientific, rational city”, depicted slum dwellers as degenerate. This provided a
justification for intervention in blighted areas (Outtes, 159). When medicine became social
medicine, the city emerged as an object of hygienic interest. Slums and tenement housing were
considered dangerous places because of overcrowding and poor basic services (Outtes, 151).
But slums were found to be not just places of physical squalor but of moral degeneration
as well. Slum clearance, therefore, represented a battle to improve not just physical health but a
way to raise morality, too, as this presentation by a Brazilian engineer and founding member of
the national institute of architects in that country typified in the early 1930s:
Visiting the slums of the federal capital is sufficient to give a clear view of this
problem. From them, one can say, come all moral and material miseries and all
vices. In the slums there is tuberculosis and alcoholism. Low instincts are
developed there. Fighting against slums is taking part in a battle for raising the
morality and the physical health of the race…They must be demolished.
233

There was little room for the poor in the modern, rational city. Policy measures aimed at
the urban poor in the planning discourse ranged from segregation measures, on one end of the
spectrum, to outright slum removal, on the other end of the spectrum. In the United States,
federal aid was made available to help cities attack slums and blight on a large scale, instead of

233
M.T.C. Mendonça, “Casas Populares, Cidades Jardins”, Annaes do 1º Congresso de Habitação (São
Paolo: Instituto Engenharia, 1931), 141, quoted by Joel Outtes, “Disciplining Society through the City: the
Genesis of City Planning in Brazil and Argentina (1894-1945)”, Bulletin of Latin American Research 22,
no. 2 (2003): 140.

202

merely through isolated housing projects. The 1949 Housing Act authorized funds for the
elimination of slums, while the 1954 Housing Act expanded the original Act to support slum
clearance as part of massive urban renewal programs. For municipal authorities and taxpayers,
urban renewal was billed as a promising way of replacing “outmoded, worn-out and blighted
areas with sound, well-planned development geared to modern needs” (Wedge, 55). Housing
professionals in that era characterized slums as a burden on cities; removing this burden would be
good for business:
The removal of slums and blight leads to a sharp, sometimes amazing, reduction
in crime, disease, and the need for municipal services, and a tremendous increase
in tax returns from these areas when they have been restored to good use. Most
observers of the American urban scene are agreed, as a matter of fact, that sound
urban renewal will not only pay for itself but also will be profitable to the
community in the long run. Urban renewal, therefore, is good business (Wedge,
56).
In the developing world, measures aimed at segregating the poor from the rest of the city
have deep historical roots, particularly in the colonial and post-colonial setting in many parts of
the developing world. In British colonial cities of Southern and Eastern Africa, as in many
European colonies, the native population was denied access to permanent urban residence and
land ownership, and in many cities, access was controlled through pass laws (Davis, 51). In Latin
America, cities were long viewed as the enclaves of Europeanized culture—citadels of the elites,
and highly homogenous in class composition—and new arrivals were looked down on as
unwanted intruders (Perlman, 92, 246). In Cambodia the French colonial administrators similarly
divided up Phnom Penh into ethnic enclaves, and the indigenous Cambodian population formed a
minority in the city until after independence. Another aspect of the planning discourse that was
exported to Phnom Penh was the emphasis on public hygiene. In post-independence Cambodia,
as Sihanouk’s prakas in the 1950s made clear (see section 2.7), the poor could live in the city, but
only if they adhered to certain codes of conduct.

203

4.2.2 A “Culture of Poverty”
In the 1960s, the anthropologist Oscar Lewis caused a stir by suggesting that there
existed a “culture of poverty” that reflected itself in universal social-psychological characteristics
among the urban and rural poor in capitalist societies. Lewis based his findings partly on two
ethnographies
234
of urban slum families, the first of a low-income Mexican family in Mexico
City, and the second of a Puerto Rican family in San Juan and New York. In the spirit of
ethnographic realism, Lewis aimed to let his research subjects speak for themselves: over the
course of several years, Lewis tape recorded how various family members described their lives,
and wrote down their stories.
Lewis described the culture of poverty as a “way of life, remarkably stable and persistent,
passed down from generation to generation along family lines” (Lewis 1961, xxiv).
[It is] both an adaptation and a reaction of the poor to their marginal position in a
class-stratified, highly individuated, capitalistic society. It represents an effort to
cope with feelings of hopelessness and despair which develop from the
realization of the improbability of achieving success in terms of the values and
goals of the larger society. Indeed, many of the traits of the culture of poverty can
be viewed as attempts at local solutions for problems not met by existing
institutions and agencies because the people are not eligible for them, cannot
afford them, or are ignorant or suspicious of them. (Lewis 1966, xliv).
The culture of poverty comes into being in a variety of historical contexts. Most
commonly, “it develops when a stratified social and economic system is breaking down or is
being replaced by another, as in the case of the transition from feudalism to capitalism or during
the Industrial Revolution” (Lewis 1961, xxv). The people most likely to be affected by a culture
of poverty are those who come from the lower strata of a rapidly changing society and are
partially alienated from it (Lewis 1966, xlv):

234
“The Children of Sánchez” (1961) and “La Vida: A Puerto Rican Family in the Culture of Poverty—San
Juan and New York” (1966).

204

The culture of poverty would apply only to those people who are at the very
bottom of the socio-economic scale, the poorest workers, the poorest peasants,
plantation laborers, and that large heterogeneous mass of small artisans and
tradesmen usually referred to as the lumpen proletariat (Lewis 1961, xxv).
The main significance of the culture of poverty theory for the struggle for access of the
urban poor is perhaps Lewis’s suggestion that a crucial characteristic of this culture is the lack of
effective participation and integration of the poor in the major institutions of the larger society
(Lewis 1966, xlv). The culture of poverty affects participation of the poor in the larger national
culture:
By the time slum children are age six or seven they have usually absorbed the
basic values and attitudes of their subculture and are not psychologically geared
to take full advantage of changing conditions or increased opportunities which
may occur in their lifetime (Lewis 1966, xlv).
At the community level, Lewis remarked that slum dwellers in the culture of poverty
have only a “minimum of organization beyond the level of the nuclear and extended family”:
It is the low level of organization which gives the culture of poverty its marginal
and anachronistic quality in our highly complex, specialized, organized society.
Most primitive peoples have achieved a higher level of socio-cultural
organization than our modern slum dwellers (Lewis 1966, xlvii).
At the level of the individual, people with a culture of poverty are “provincial and locally
oriented and have very little sense of history. They know only their own troubles, their own local
conditions, their own neighborhood, their own way of life” (Lewis 1966, xlviii).
Lewis’s culture of poverty theory has attracted much debate. It was roundly rejected by a
large number of scholars, particularly on the left. One of the most vehement refutations of Lewis’
work was provided by Charles Valentine (1968), which appeared during the civil rights protests
in the United States. Valentine claimed that Lewis’ interpretation of poverty “blamed the victim”
and belonged to the same “patronizing and moralizing” tradition of black family studies that was
causing an uproar in the United States at the time, sparked particularly by Patrick J. Moynihan’s

205

critical study of black family life entitled, “The Negro Family: the Case for National Action”,
which had appeared in 1965 (Harvey and Reed, 469-70). Another influential critique from the left
at the time was provided by Eleanor Leacock (1971), who argued among other things that Lewis
over-emphasized psychological factors and neglected the importance of underlying social and
economic structure in furthering poverty. Moreover, the culture of poverty theory focused on a
“negative, distorted and truncated view of a cultural whole”, and represented an ethnocentric
tendency to treat behavioral patterns among lower-class people as “shortcomings from some
presumed ‘middle-class’ ideal” (Leacock, 12, 17, 35).
Several decades later, however, Lewis’s culture of poverty is being reevaluated. David
Harvey and Michael Reed (1996) have called the culture of poverty a “splendid tool for
understanding poverty’s cultural superstructure” (Harvey and Reed, 486). They suggest that, far
from representing a conservative tradition blaming the poor for their poverty, Lewis was in fact a
Marxist humanist, and that the culture of poverty theory fits within a critical theory of capitalist
production. The culture of poverty reproduces subjectively from below what is objectively
imposed from above, as part of the capitalist mode of production (Harvey and Reed, 483, 485-6).
Within this perspective, Lewis believed that people in the culture of poverty are politically
powerless and seldom have the wherewithal to alter on their own the social relations that
constrain them. But this does not mean that the poor according to this theory are entirely passive.
Within the constraints of their social relations, the culture of poverty celebrates the
“resourcefulness of the poor”: the poor construct collective responses to their poverty and “shape
poverty’s life space” so as to ease the pain of living poor. The culture of poverty can be regarded
as a “positive social construction”, the result of a process by which the poor “pragmatically
winnow what works from what does not, and pass it on to their children” (Harvey and Reed, 482).
Examples of such resourcefulness include the use of informal credit schemes without interest in

206

the absence of credit from banks; working in a range of unskilled occupations for lack of wage
labor; and relying on herbs and home remedies for lack of access to hospitals and formal care
(Lewis 1961, xxvi-i).
Lewis concluded that real change for the poor can come only through systemic change:
By creating basic structural changes in society, by redistributing wealth, by
organizing the poor, and giving them a sense of belonging, of power and of
leadership, revolutions frequently succeed in abolishing some of the basic
characteristics of the culture of poverty even when they do not succeed in
abolishing poverty itself (Lewis 1966, lii).
Lewis believed that the poor could overcome their provincial worldview and rise above
their cultural conservatism, if they were swept up in a revolutionary cause. Under those
circumstances, “once the poor begin to identify with larger groups and larger causes…they
rapidly begin to lose some of the crucial aspects of the culture of poverty
235
. There was therefore
a way out of the culture of poverty, but it seems that the circumstances under which this would
occur would most likely have to be created by outsiders and not by the poorest themselves.
4.2.3 An Expression of Socially Dominant Interests
To Friedrich Engels in the mid 19
th
century, the slums of English cities during the
industrial revolution resembled “industrial ghettoes”, and they represented an “awful paradigm of
what a free capitalist society could produce” (Roberts 1971, 14). For neo-Marxists such as
Manuel Castells, writing a century later, the slum is the product of an urban crisis “engendered by
uneven development and the new international division of labor in the world economy”, forcing
“millions and millions of people to live in physical and social conditions that are reaching the
point of ecological disaster” (Castells 1983, 175). Both observations are steeped in a Marxist,

235
Oscar Lewis (1962), quoted in Susan M. Rigdon, The Culture Façade: Art, Science and Politics in the
Work of Oscar Lewis (Urbana, Illinois: University of Illinois Press, 1988), 231, quoted by David L. Harvey
and Michael H. Reed, “The Culture of Poverty: an Ideological Analysis”, Sociological Perspectives 39, no.
4 (1996): 483.

207

class-based perspective that considers power to be systemic and conditioned by the reproduction
of exploitative social relations. The actual and potential power of the “subordinated class” is
constrained by strong sovereign power and the class struggle.
In conventional Marxist thinking, power is in the hands of the main institutions in
society, such as religion and capital—but above all, the state. The state in a capitalist system is
almost always in the hands of the strongest economic class; its institutions exercise a hostile
domination over the individual, because these institutions are themselves dominated by a class of
men who are alien and hostile to him (Ollman, 219). To understand how the state is able to
dominate the population it is necessary to understand the condition of human beings in a capitalist
system. In capitalist societies, ordinary men live in a condition of alienation. Marx defined
alienation as “all that is under the sway of inhuman power” (Ollman, 132, ref German Ideology
whole paragraph). Man’s inherent human nature and social mode of existence are distorted by
capitalism: he is separated from his work (which represents his life activity), from his own
products, and from his fellow men. Communal interest is lost behind competing particular
interests: individuals lose sight of the communal interest. People experience others only by
struggling against them (Ollman, 218). In this condition of weakness, man becomes an object of
manipulation by others and by his own products, in every sphere of his life: “as a producer, he is
told when, where and how to work. As a consumer he is told what to buy and how to use it”
(Ollman, 244-5).

208

In this context of domination and alienation, human agency is limited by structural
determinism. The systemic view of power was summarized by Marx as follows:
Men make their own history but they do not make it just as they please; they do
not make it under circumstances chosen by themselves, but under circumstances
directly encountered, given and transmitted from the past (Lukes, 22, citing Marx
1962
236
).
The state is “an organism independent of society and above it,” and classes are acting on
the state from the outside (Poulantzas 1976, 81). But just how much autonomy human agents do
have within this pre-determined structure is still a matter of debate within Marxist thought
237
.
According to the determinist Althusser school of Marxism, power is redefined as structural
determinism. In this view, power is “the capacity of a social class to realize its specific objective
interests,” and class relations are at “every level relations of power” (Poulantzas 1973, 101-4).
For less deterministic Marxists, however, the “historical subject” does have a “crucial and
ineradicable explanatory role” in the exercise of power (Lukes, 52).
But for all Marxists the ability of the working class to enforce its potential power is
anyhow limited. To express it in modern political sociology terms, those at the bottom of society
have very little “agency”, for several reasons. First of all, individuals cannot free themselves
from their mode of production, which is their way of life: “what [individuals] are… coincides
with their production, both with what they produce and with how they produce. The nature of
individuals thus depends on the material conditions determining their production” (Marx 2001,
62). Second, proletarians as individuals have little power. In order to assert themselves as

236
Karl Marx and Friedrich Engels, Selected Works, Volume 1 (Moscow: Foreign Languages Publishing
House, 1962), 247.
237
The debate about determinism in Marxist thought pits structural Marxists such as Althusser and later
thinkers such as Poulantzas against the so-called “humanist” or “subjectivist” interpretations of
philosophers such as Sartre and earlier thinkers such as Lucàks, Korsch and Hegel (Lukes, 52).

209

individuals, proletarians must overthrow the state; short of that, they are doomed to remain stuck
in their class:
For the proletarians…the condition of their existence, labour, and with it all the
conditions of existence governing modern society, have become something
accidental, something over which they, as separate individuals, have no control,
and over which no social organisation can give them control. The contradiction
between the individuality of each separate proletarian and labour, the condition
of life forced upon him, becomes evident to him himself, for he is sacrificed from
youth upwards and, within his own class, has no chance of arriving at the
conditions which would place him in the other class (Marx 2001, 120-1).
Third, capitalist development both unifies and fragments the working class. On the one
hand, the capitalist system enables workers to organize nationally, because of the creation of
national capital. But on the other hand, because capital accumulation occurs unevenly, the
working class is heterogeneous in terms of its workers’ conditions, skills, unemployment histories
and income. This causes internal political divisions and ultimately leads to the disorganization of
labor (Alford and Friedland, 282-3).
In current neo-Marxist urban analysis, two structurally determinist Marxist assumptions
guide the struggle of the urban poor
238
. The first assumption is that the poor are still involved in
some form of “class struggle”, even though they are now no longer a “proletariat” but simply a
so-called “marginal sector” (Castells 1983, 179). Power in the urban structure is defined through
class relations. But whereas for classical Marxists the class struggle pitted dominant capitalist
interests against the working class proletariat as the “dominated class”, in neo-Marxist urban
analysis, the new class struggle pits dominant corporate and state interests against “urban social
movements”, a term which Castells defines as a “conscious collective practice originating in
urban issues, able to produce qualitative changes in the urban system, local culture, and political

238
There is a third assumption, which is that urban social movements are the major actors in urban social
change. This is a pluralist theory, and I will discuss it in the section on “Slums of Hope” (see section 4.3).

210

institutions in contradiction to the dominant social interests institutionalized…at the societal
level” (Castells 1983, 278). In this analysis, urban social movements comprise a wide range of
different local organizations, from neighborhood associations to student groups, mounting
community struggles around higher levels of “collective consumption” and better conditions of
life in the community, including housing, transportation, and education. These community issues
have replaced struggles over the work process and the wage rate as the most significant focus of
conflict in the modern class struggle
239
.
The second core assumption in neo-Marxist urban analysis is that the marginal sector
wages its class struggle against “socially dominant interests” in the city, which resist change.
These interests are never well defined, but are generally referred to as “multinational corporations
dominating uneven economic growth” (Castells 1983, 175). In current neo-Marxist urban
analysis, urban structure, processes and problems as well as social relationships in capitalist
societies are shaped by, and rooted in, the capital accumulation process and the imbalances and
contradictions associated with it (Jaret, 503-4). The city is not only a spatial form but expresses
the social organization of the process of reproduction (Castells 1979, 431). Social organization is
characterized by the power of dominant interests, which are mostly corporate interests. The way
that the dominant class expresses its power—the manner in which it achieves “structural
reproduction of the dominant mode of production”—is through the interventions of the politico-
juridical apparatus of the state, and specifically through urban planning (Castells 1979, 432).
Castells speaks of planning agencies submitting completely to a “dominant social and political
force” (Castells 1978, 84). The way in which urban planners achieve social control is subtle. By

239
M. Dear and A.J. Scott. eds., Urbanization and Urban Planning in Capitalist Society (London:
Methuen, 1981), 91-121, quoted in Charles Jaret, “Recent Neo-Marxist Urban Analysis”, Annual Review of
Sociology 9 (1983):499-525.

211

presenting themselves as professionals and ‘scientists’ above conflict, planners can carry out the
interests of the central administration much more efficiently than if they present themselves as
mouthpieces of the dominant faction: “it is in this way that planning is part of the process of
social regulation, rather than that of direct domination on the part of the class in power” (Castells
1978, 85).
The first and second neo-Marxist assumptions both contain a measure of “despair” about
the prospects of the marginal sector in the capitalist city because they assume that the market
economy and state politics, on their own, would never be able to provide adequate shelter and
urban services to an increasing proportion of city dwellers, including the majority of regularly
employed salaried workers, as well as practically all people making a living in the informal sector
of the economy (Castells 1983, 185). Based on her research in the slums (favelas) of Rio de
Janeiro, Janice Perlman (1967) concluded that, in their socio-economic conditions, the favelados
were indeed “marginals”, not on account of their pathology, but because of their powerlessness.
Perlman concluded that the favelados “have the aspirations of the bourgeoisie, the perseverance
of pioneers, and the values of patriots”. But “what they do not have is the opportunity to fulfill
their aspirations” (Perlman, 243). Perlman viewed the favelados in Rio de Janeiro as a “powerless
group subject to a good deal of coercion and doing very little coercing” (Perlman, 245). Perlman
suggested that there were “fundamental antagonisms between the interests of the privileged and
non-privileged” in society, and suggested that there was an “inherent asymmetry of the
relationship” (Perlman, 245). She concludes, in determinist fashion, that “the favela is a necessity
of the Brazilian social structure. It demands relations of economic dependence which result in the
temporary or permanent misery of the dependent element for the benefits of society (Perlman,
245)”.

212

The neo-Marxist (and Marxist) structural approach to urban issues offers a unified
approach to urban phenomena and analyses that transcend disciplinary boundaries, where
conventional urban social science has offered only disconnected middle-range theories (Jaret,
521-2). At a policy level, too, many observers of housing policy in developing countries now
accept that there is a need to tackle the structural problems facing the urban poor, instead of
adopting piecemeal and purely technical measures such as the housing improvements and the
physical and environmental improvement efforts that have been the focus of so many programs in
slums (UN-Habitat, xxvii). As Stella Lowder (1982) points out, the urban poor do face
substantial structural impediments in resolving their housing crisis: the appearance of illegal
settlements all over the world is a response to city housing markets priced so as to exclude a
significant proportion of the population (Lowder, 115). And the lesson from decades of policy
interventions is that remedial policies directed only at slums are doomed to fail: many Western
governments have spent years devising spatially oriented policies before heeding advice
underlining the structural forces in the greater economy in the generation of inner-city deprivation
(Lowder, 114).
But at the same time, the Marxist and neo-Marxist emphasis on “socially dominant
interests” offers only a limited understanding of the full range of power in the urban environment;
the one-sided description of power in the slum is of course exactly what the Foucauldian analysis
of capillary power rejects. The assumption of a sovereign power, as in the case of the “socially
dominant interests” in the neo-Marxist perspective is contentious in the fluid political and
institutional context of most developing countries.

213

4.2.4 Ideological Hegemony and the “Myth of Marginality”
Another cause of pessimism on the Left about the capacity for agency among the urban
poor has to do with their ideological indoctrination by socially dominant interests. This idea is
associated with the Marxist concept of ideological domination, which holds that apart from
repression and direct domination through capital accumulation, the state and ruling classes also
further their domination indirectly, through ideas and symbols:
The ideas of the ruling class are in every epoch the ruling ideas… The class
which has the means of material production at its disposal has control at the same
time over the means of mental production, so that thereby, generally speaking,
the ideas of those who lack the means of mental production are subject to it. The
ruling ideas are nothing more than the ideal expression of the dominant material
relationships grasped as ideas; hence, of the relationships which make the one
class the ruling one, therefore, the ideas of its dominance (Marx 2001, 92).
The masses assume a passive and fatalistic acceptance of this dominant social order of
the elites, an attitude that Marxists identify as “mystification” or “false consciousness”. This
causes workers to fail to identify with their own objective class interests (Alford and Friedland,
283). For Antonio Gramsci (1971), the dominance of this discourse is so powerful and so
pervasive that the masses are more enslaved at the level of ideas than at the level of behavior (ref
by Gramsci)(Scott 1985, 39). Gramsci referred to ideological domination as “hegemony”, which
the dominant group exercises throughout society through intellectuals, who are the dominant
group’s “deputies”. Hegemony is made possible because:
The great masses of the population give their spontaneous consent…to the
general direction imposed on social life by the dominant fundamental group; this
consent is ‘historically’ caused by the prestige (and consequent confidence)
which the dominant group enjoys because of its position and function in the
world of production (Gramsci, 145).
When spontaneous consent fails, the apparatus of state coercive power resorts to direct
domination through the state and “juridical government”, which “legally” enforces discipline on

214

those groups who do not consent either actively or passively. This apparatus is constituted for the
whole of society in anticipation of moments of crisis of command and direction (Gramsci, 145).
A number of development traditions have accentuated the influence that ideological
power of socially dominant interests has on the poor. The most prominent example is perhaps the
“classic marginality theory of the 1960s”, famously captured by Janice Perlman (1967) in her
work The Myth of Marginality on the favelas (slums) of Rio de Janeiro. Perlman found that the
myth of marginality consisted of two sets of negative stereotypes about the social, cultural and
economic marginality of favelados, widely held among the general public, but also among
urbanists, planners and policy makers. The first and prevailing assumption is that favelas are
“pathological agglomerations”, populated by “marginal elements”, such as “unemployed loafers,
abandoned women and children, thieves, drunks, and prostitutes” (Perlman, 14-5). Residents of
the favelas manifest “all the symptoms of social disorganization”, including family breakdown,
anomie, and mutual distrust, and which makes them prone to rampant crime, violence and
promiscuity. Moreover, favelas are a drain on the city, both economically and socially—a
parasite, demanding high expenditures for public services and offering little in return. At the
level of public health, the favelas are “filthy and disease-ridden shantytowns”. Embedded in this
myth are several implicit threats that the favelas and their residents pose to the rest of the city at
various levels. Favelados are perceived as “invading rural peasants”, arriving “lonely and
rootless from the countryside, unprepared and unable to adapt fully to urban life”. At the political
level, favelados are perceived as a “seething, frustrated mass” prone to fall easy prey to radical
rhetoric (Perlman, 1-3).
A second set of assumptions is that favelas represent “inevitable blight”. This is a more
moderate view, which sees the favela as a “natural, though unfortunate, consequence of rapid
urban growth”, a by-product of rapid urban-rural migration and the existence of widespread

215

unemployment or underemployment and inadequate infrastructure. Favelados are useful in
providing cheap labor and easily bought votes, but still regarded as unproductive economically,
naïve politically and rather undesirable socially (Perlman, 16-7). This perspective is highly
paternalistic: favelados are like children: they need to be guided, taught and educated, so that the
more worthy favelados will overcome their many obstacles and become integrated into society
(Perlman, 17).
While the myths of marginality described by Perlman do not represent a coherent
“ideology”, they certainly reflect class bias and, according to Perlman, fulfill the “ideological-
political function of preserving the social order which generated [it]” (Perlman, 246). Even
though Perlman rejected the myths about the favelados, she hinted at a certain amount of “false
consciousness” among the favelados, as the myths affect the self-perception of the slum dwellers.
Perlman points to findings of Brazilian sociologists suggesting that, by using the favela as a
“scapegoat for many embarrassing social problems that are unresolved…these definitions of the
favela are so deeply ingrained…that many favelados themselves, principally the poorest strata,
are convinced of their own inadequacies” (Perlman, 16). The ideology of domination and
legitimation, underpinned by the favelados’ own poor self-image, have acted as to justify the
authorities’ policies towards the favelas. Perlman points out that “favela removal was permitted
as part of the restructuring of the urban system precisely because the favelados were considered
marginal and therefore dispensable” (Perlman, 248). Because the favelados were considered
“marginal”, they had no rights or claims on the system and were therefore easier to manipulate”
(Perlman, 249). For those policy makers for whom the favelas represented “inevitable blight”
rather than pathological danger, policies are paternalistic rather than aimed at slum removal.
They include “palliative types of assistance” such as distribution of food and clothing and the
organization of services for the favelados (Perlman, 17).

216

In the intervening years the classic marginality theory of the 1960s has made way for a
new construction of poverty referred to simply as “the new poverty”. In this new construction,
the concept of marginality has made a comeback, but in a different guise. According to Loïc
Wacquant (1996), one of the most forceful proponents of a new “advanced marginality”, the
urban poor in the United States and Western Europe form part of a new post-industrial
underclass, trapped in “sub-proletarian housing projects” by economic uncertainty and
precariousness, a fragmented labor market, the “casualization of employment”, “class
decomposition” and “social dispersion and fragmentation” (Wacquant, 123-4; 127-8). Under
conditions of advanced marginality, the slums of the poor no longer provide protection from the
outside, nor do they provide internal solidarity. Instead, the ghetto has become an area of
“territorial stigmatization”,
a vector of intra-communal division and an instrument for the virtual
imprisonment of the urban sub-proletariat of color, a dreaded and hated territory
from which, as one informant from Chicago's South Side tersely put it,
‘everybody's tryin' to get out’ (Wacquant, 126).
For many proponents of the new marginality, the socio-economic trends are also
replicated in developing countries. In the Latin American context, for example, Peter Ward
(2004) argues that as a result of changing economic conditions, the “contemporary urban scene”
is characterized by rising unemployment, social exclusion, rising violence and insecurity and
declining opportunities, even in the informal sector
240
. Thirty-five years after her original
research, Perlman undertook an intensive re-study of her original subjects in Rio de Janeiro.
Perlman found that the new marginality only partly matched the social reality she rediscovered in
the favelas. While levels of unemployment were even higher compared to three decades before,

240
Peter A. Ward, “Introduction and Overview: Marginality Then and Now”, in “From the Marginality of
the 1960s to the ‘New Poverty’ of Today: a LARR Research Forum”, Latin American Research Review 39,
no. 1: 185-6.

217

social inequalities had not markedly increased. The most remarkable finding was that there was
little evidence that favelas were prisons of the urban poor, as theorists of the new marginality
would have it. On the contrary, the favelas Perlman studied remained racially, socially,
culturally, and economically as heterogeneous as before, and there was considerable migration
out of the favelas, including to higher class neighborhoods, thus suggesting “upward mobility”
(Perlman et al, 2004, 191-2).
4.3 Slums of Hope
Proponents of the “slums of hope” perspective come from the left as well as the right, but
all can be described as adherents to some degree of a pluralist, rather than a deterministic,
worldview within political sociology.
In the pluralist perspective, individuals are the level of analysis. It involves a focus on
the behavior of individuals (and groups) in decision-making on issues over which there is an
observable conflict of interests, seen as express policy preferences, revealed by political
participation (Lukes, 15). In this perspective, power is situational, and is ultimately about
influence (Alford and Friedland, 7). In contrast to a class-based or Marxist perspective on power,
the pluralist tradition is one-dimensional because it is the study of overt, actual behavior: it
“cannot reveal the less visible ways in which a pluralist system may be biased in favor of certain
groups and against others” (Lukes, 21, 37). Pluralist theorists reject the notion of a dominant,
monolithic, hierarchical and centralized structure called the “state”. Instead, the pluralist
perspective is centered on multiple sources of authority and institutions differentiated according
to function (rule making, rule enforcing, rule adjudication, and rule breaking). The pluralist
worldview makes a few fundamental assumptions about the nature of the state and modern
society. These include the existence of an “open, accessible government”; the fact that the

218

government is an institution chosen by individuals and that it is functional for society; and the
premise that modern society is self-regulating, but based on shared values (Alford and Friedland,
41).
The fundamental assumptions of the pluralist worldview have certain clear implications
for power. The first is that individuals have agency. Organizations and societies are constructed
and shaped by individuals, and not the other way around as in the systemic perspective of the
class-based and Marxist tradition. The second implication is that power is based on seeking
representation and influence. Social movements are instrumental in creating collective goods for
movement beneficiaries. The right of political participation is an essential aspect of the
democratic experience in a pluralist state (Alford and Friedland, 49). The pluralist perspective
assumes a political culture that encourages responsiveness by political leaders to legitimate
demands (Alford and Friedland, 52). Third, governance is a critical challenge in pluralist society,
as there is constantly a threat of excessive participation and a breakdown in societal consensus.
I argue that some of the basic assumptions held by the pluralist perspective are ill-suited
to an analysis of power forces in a developing country environment. For instance, there can be no
assumption of an open, accessible government, common values in society, or policy continuity in
government in an environment where democratic structures are new and imperfect. Modern
political sociology, which can be understood as an expansion of the (older) pluralist world view,
provides an alternative and improved framework, one which recognizes the relative lack of
importance of laws and the lack of continuity of urban policies, and concurrently recognizes the
greater importance of negotiation in urban development outcomes. Within political sociology the
“general perspective” on power resources holds that “a web of complex networks of political,
economic, and cultural interdependencies has to be analyzed if the actual potential for power by
different participants in these networks is to be deciphered” (Fox Piven and Cloward, 39). This

219

perspective holds that even weaker groups in society do have some power, at least some of the
time, and they do challenge domination in many different ways. A critical insight here is that
“defiance does not have to be bold” in order to be understood as a power strategy (Fox Piven and
Cloward, 41, 50).
In this section I will describe five perspectives of slums and slum dwellers that
acknowledge the merits and power of individual agency in the slum. Each of these perspectives
holds out the prospect that the urban poor and individual slum dwellers can improve their
environment in some way. In the early 1960s Jane Jacobs’ categorization of “slumming and
unslumming slums” was one of the first to make a distinction between different types of slums.
She argued that if only more slums would be allowed to become “unslumming slums”, these
areas might become more like the rest of the city in time. Around the same time, housing
professionals such as John F. C. Turner advocated the concept of self-help housing, arguing that
individual slum dwellers were better able to build their housing than government planners were.
From the right, Hernando De Soto argued that the only thing slum dwellers in fact required to
improve their condition was the recognition of property title: in one fell swoop, “informals”
would become formal residents of the city, and capital would be injected into the informal
economy. From the left comes the (newly-pluralist) perspective of Manuel Castells, who argues
that urban social movements, including grassroots activity in the slums, have the capacity to
achieve social action and thereby obtain the collective goods that they require. A new discourse
of “empowerment” is starting to influence development assistance programs. The discourse is
based on pluralist assumptions, but is ultimately paternalistic: it does not recognize the many
power strategies of the poor, and ultimately belongs to the same paternalistic tradition of many
state programs criticized by Jane Jacobs and others, that hold that “power” has to be bestowed
onto the poor (and thus, by extension, that the poor lack their own forms of power). In my final

220

example of a pluralist approach to the slum, I describe a grassroots international movement,
Slum/Shack Dwellers International, which organizes people’s movements in slums in countries
across three continents, including Cambodia. SDI’s approach to community development using
local people’s movements is unique for involving the poor themselves.
4.3.1 Unslumming Slums and Perpetual Slums
Starting in the 1960s a changing discourse about slums started to make its way felt in
housing and urban policy circles. Several writers argued that slums were not just a problem but
also a “solution” to the challenge of low-income housing. They forced a reassessment of slums
and their place in the city, and attributed a sense of agency in the lives and actions of slum
dwellers.
One of the first of these was Jane Jacobs, who, in her classic work The Death and Life of
Great American Cities (1961), attacked urban planning and redevelopment policies in the United
States for destroying the lifeblood of large cities, namely, the “intricate and close-grained
diversity of uses that give each other mutual support, both economically and socially” (Jacobs,
14). Jacobs contrasted the lively “chaos” and diversity to be found in organic neighborhoods in
cities with the “great blight of dullness” in urban renewal areas. Jacobs explained that slums
were an inherent part of the fabric of the diversity of the city. Jacobs distinguished between two
kinds of slums: “unslumming slums” and “perpetual slums”. Unslumming slums are areas which
are upgraded spontaneously by their residents as their economic and social conditions improved
over time. The key to an unslumming slum is that a significant percentage of the population
decides to stay in the area. While the unslumming process might appear to planners to generate
“social untidiness and economic confusion”, Jacobs concluded that unslumming and the “self-
diversification” that accompanies it are possibly the greatest regenerative forces in American
cities (Jacobs, 290). By contrast, in perpetual slums, the more energetic, affluent and ambitious

221

residents desert the area in too rapid a space of time. The result is that the neighborhood becomes
a dull place, the slum stagnates, and the area shows no or only few signs of social or economic
improvement over time (Jacobs, 272-3).
Jane Jacobs’ contribution to the debate about slums was two-fold. First, she introduced a
more nuanced vision of slums and slum dwellers, emphasizing positive aspects. Second, Jacobs
highlighted the inadequate policies towards slums. Jacobs criticized conventional planning
approaches to slums and slum dwellers for being “thoroughly paternalistic” (Jacobs, 271). This
dual contribution by Jane Jacobs was picked up on by later observers writing about slums in
developing countries, such as Wegelin et al. (1983), who pointed out the paternalism of planners
in the Thai context. Perlman pointed out the positive aspects of the favela in Rio de Janeiro:
Beneath the apparent squalor [of a favela] is a community characterized by
careful planning in the use of limited housing space and innovative construction
techniques… Dotting the area are permanent brick structures that represent the
accumulated savings of families who have been building them little by little,
brick by brick (Perlman, 1. See also, 242-3).
Moreover, Perlman found that some people chose to live in a slum as a matter of lifestyle,
or as an economic preference, “even though they have stable employment as workers in the
capitalist sector” (Perlman, 244). The solution, Jacobs argued, was putting more faith in the
actions of the residents themselves:
To overcome slums, we must regard slum dwellers as people capable of
understanding and acting upon their own self-interests (Jacobs, 271).
4.3.2 The Promise of Self-Help Housing
Another early and well-known defender of slums and people’s housing in a developing
country context was the architect John F. C. Turner, who worked as a housing professional in the
informal settlements of Peru during the late 1950s and early 1960s. Turner (1972) pointed out that
slums arise out of necessity: the poor build their own housing under circumstances where neither

222

the organized private sector nor the public sector can provide new housing for a huge and rapidly
growing market (Turner 1972, 171).
In this respect, as John F. C. Turner pointed out as far back as the 1960s
241
, slums are
well adapted to the needs and circumstances of their residents (Harris, 257). In the (few) cases
where public agencies do have a budget for subsidized low-income housing, Turner critiqued the
disastrous effects of public sector planned housing. He pointed to the high and “humanly
exacting” administrative costs of the kind of aided self-help housing projects that were being
planned and financed by government departments and international agencies (Turner 1972, 144).
Turner questioned the feasibility and the desirability of direct public sector authority over housing
(Turner 1972, 145).
Instead, Turner was a great supporter of the self-help housing principle in informal
settlements. Turner defined self-help not just as do-it-yourself building but as housing where the
resident himself is in control of the housing process through participation in housing design,
construction and management (Turner 1977, 134). Self-help housing unleashes “initiative,
ingenuity, perseverance, and hope” (Turner 1972, 145). This is contrasted with aided self-help or
other cases where public sector authorities have direct corporate control over housing, a situation
Turner labeled the “impersonality of authoritarianism”:
Building carried out by a large and hierarchically organized agency, whether
public or commercial, provides little room for dialogue between people. All
decisions are vertical and all operations are carried out by more-or-less
unchallengeable order. When people are building for themselves, on the other
hand, or when the builders are building for the users, there is plenty of room for
genuine relationships between the people brought together by the activity, and
therefore, for creativity, pride, and satisfaction from the work itself (Turner 1972,
145).

241
In the 1960s, the architect and housing professional John F.C. Turner propagated the (then) radical idea
that slums were a “solution” for the housing needs of the poor because they are well adapted to the needs
and circumstances of their residents and because they tend to be improved over time by the residents
themselves (Harris, 257).

223

Turner pointed out the reasons for slums. The rigidity of building codes contributes to a
shortage of safe and sanitary housing. Housing standards, such as subdivision codes and rent
freezes, to the extent that they are enforceable, “price the great majority of would-be home-
builders out of the market, and…inhibit legitimate, inspected, taxable private investments in low-
income housing, whether for rent or for sale” (Turner 1972, 149-50). Hence “it is not surprising
to find” that two-thirds of all dwellings built in Lima since the early 1940s, and over 90 percent in
the poorer provincial city of Arequipa, were put up by squatters or buyers of lots in clandestine
subdivisions (Turner 1972, 149). Turner described the situation in the squatter settlements of
Peru in 1960s, but since then little has changed. Slums today owe their existence to the same
structural factors as those pointed out by Turner.
Turner also pointed out that squatters have stamina and are highly pragmatic. The typical
Peruvian urban squatter must be highly pragmatic in order to capitalize on his very small savings
margins and relatively low skills. There are many thousands of families with very low incomes
who have built themselves homes that have potential or even actual market values amounting to
five years’ income or more. The typical Peruvian squatter has also proved that he is highly
resistant to the exercise of police power (Turner 1972, 162).
But the promise that self-help housing held out for the poor as a way to escape slum
conditions was relatively short-lived. Turner himself acknowledged the structural impediments to
self-help housing, including the political obstacles: self-help policies have never been popular
with governments, who “cannot accept that what poor people do for themselves can be right or
proper” (Turner 1977, xxxiv). Stella Lowder also challenged the viability of self-help housing as
a solution in Latin American cities: the dream of self-help as a means of providing tolerable
homes in Lima has faded as a result of economic decline and the scarcity of remaining terrain.
There is now a general rejection of self-help ideals (Lowder, 119, 121-2).

224

4.3.3 De Soto and the Promise of Property Rights
In a special Global Report on Human Settlements devoted to the challenge of slums
(2003), the United Nations Human Settlements Program (UN-Habitat) emphasized the need for
secure tenure and “involving the poor in housing development”. The Peruvian economist
Hernando De Soto is perhaps most famous for advocating the benefits of property rights as a way
of liberating millions of the world’s poor from informality. In his ground-breaking account of the
informal sector in Peru, The Other Path, De Soto acknowledged the entrepreneurship of
“informals”, who are responsible for making available most of the housing, markets and public
transportation in Lima when the state itself could not provide, to the extent that by the late 1980s,
informal financial investment in housing in Lima dwarfed state investment by a scale of 60 to 1
(De Soto 1989, 237). De Soto acknowledges the sphere of force relations involved in this
informal expansion: the development of informal settlements in Lima proceeded in stages, but
advanced steadily through a “succession of unexpected actions, mass movements, political
intrigues, and exchanges of favors” (De Soto 1989, 33). But this informal activity is not carried
out in chaos or anarchy; on the contrary: “informals have clear and specific interests” and a
“previously unsuspected level of organization governed by rules that they have spontaneously
developed to replace those which the state has failed to provide” (De Soto 1989, xxii).
De Soto concluded that the rulemaking of informal organizations is potentially the source
of many of the solutions to the crisis facing poor countries (De Soto 1989, xxii). The informals
are staging a revolution against a mercantilist state—a regime which blocks access to private
enterprise, an excessive and obstructive legal system, massive public and private bureaucracies,
and a state that intervenes in all areas of activity. The massive investment in informal housing in
cities like Lima is part of a “long march towards private property”, and the informals are
“subjugating the state and formal society as they go” (De Soto 1989, 57). The challenge for the

225

formal legal and institutional system is to regain social relevance: it needs to allow the informal
economy to function in an orderly fashion, to produce with security, and to transfer to private
individuals those responsibilities and initiatives which the state has thus far monopolized
unsuccessfully (De Soto 1989, 245).
De Soto expanded his argument several years later in the Mystery of Capital. He claims
that the poor have enormous assets in the form of their land and housing, but that these assets are
“dead capital” because they are stuck in informal systems. Because the poor have no tenure
security and lack legally enforceable property rights that are needed to use their properties as
collateral, they cannot translate their capital into investment. Contrary to what Marx claimed, the
poor are not oppressed proletarians but oppressed extra-legal small entrepreneurs with a sizeable
amount of assets (De Soto 2000, 216). Titling programs are needed to make the poor part of the
property system. This would instantly create massive equity—“trillions of dollars worth”. Part
of this new wealth would supply capital to credit-starved micro-entrepreneurs, and ultimately
slums would be transformed into “acres of diamonds” (De Soto 2000, see ref 301-31).
But De Soto’s proposals have been criticized as unrealistic, and even utopian. If it were
that easy to regularize tenure for the urban poor, and if secure tenure were indeed a win-win
situation for governments, the formal sector, and the poor alike, why has it not been achieved
anywhere on any meaningful scale? Clearly some deep-seated structural factors are in the way.
One of these is the difficulty of implementing successful titling programs. Political reform may
be required in cases where land ownership is disputed, or where criminal interests are involved
(Krueckeberg, 3).
But even if all structural obstacles to massive titling schemes were removed, many
question the underlying assumptions of a link between property rights and greater wealth and
well-being for the poor. There is the likelihood that for many poor families, even with a title,

226

collateral at the bank is not assured. A title on its own has little meaning; “a whole complex of
necessities and amenities gives value to property”, such as local infrastructure and services. If
those are not present, a property’s value may not be sufficient to make the poor bankable
(Krueckeberg, 3). There is also the question of “property bias” (Krueckeberg, 17). Not all
population groups desire a formal title, and may rather prefer to deal with more familiar
customary institutions to achieve de facto (even if not de jure) tenure security. Many urban poor
may simply not want to deal with the mainstream banking system, because they do not desire or
qualify to get bank financing, or because they can continue to borrow more easily from family
sources. Finally, there are equity considerations: titling may accelerate social differentiation in a
slum. Many poor families may simply not be able to bear the brunt of sudden payments for
property taxes and municipal services that come due after a property has been legalized, thus
leaving the owner with an unencumbered title or ultimately putting their newfound ownership at
risk (Krueckeberg, 3). And titling does nothing to aid renters, who are the majority of residents in
many slums (Davis, 80). Evidence from titling in selected Rio de Janeiro favelas indicates that
after tenure regularization the real estate market in the informal settlement consolidates, and with
the soaring land and housing values there is an emergence of a “slum within the favela” as the
poorest of the poor resort to renting single rooms (Davis, 81).
More broadly, critics have accused “neoliberals” such as De Soto of presenting a “semi-
utopian view of the informal sector” (Davis, 179). While De Soto might wish to believe that the
informals of the developing world’s growing cities are a “frenzied beehive of proto-capitalists
yearning for formal property rights and unregulated competitive space” (Davis, 179), the reality is
that the overly simplistic argument for property rights disregards some very complex structural
problems facing the urban poor. Contrary to the stereotype of informals as “heroically self-
employed” entrepreneurs, a growing number of informals are actually former salaried employees

227

in the formal sector who are now forced entrepreneurs due to the decline in formal sector
employment
242
. Most informals in fact work directly or indirectly for someone else, and are
captured in “myriad invisible networks of exploitation”. This exploitation typically comes down
hardest on the weakest members, i.e. women and children. The informal sector generates jobs by
fragmenting existing labor and thus subdividing incomes, which is made possible by a vast
surplus of labor and an intense competition for work. This increasing competition depletes social
capital and dissolves self-help networks in the context of “absolute immiseration”. In the context,
neoliberal prescriptions for making labor even more flexible are catastrophic (Davis, 180-5).
While De Soto has correctly recognized many aspects of the sphere of force relations that
informals in the developing world’s cities find themselves, his policy prescriptions have tended to
overemphasize the agency and capacity of individual informals to deploy their power, even armed
with assets such as property rights, and underestimated the structural power mechanisms in place
that will present obstacles to individual power.
4.3.4 Castells and the Grassroots
Perhaps one of the most idealistic adherents of the power of collective action for the
urban poor is Manuel Castells. In The City and the Grassroots (1983), Manuel Castells
introduced a theory of grassroots action that veered away from the pure neo-Marxist structural
determinism and class and party-based social action of his earlier works, and which was
influenced by humanist (pluralist) criticism of agentless structures (Zukin, 460). For Castells, the
major actors in this theory of urban social change are urban social movements (for a definition,
see section 3.4.1 on socially dominant interests). It is because the socially dominant interests in

242
Alejandro Portes and Kelly Hoffman, “Latin American Class Structures: their Composition and Change
during the Neoliberal Era”, Latin American Research Review 38, no. 1 (2003): 55, quoted by Mike Davis,
Planet of Slums (New York: Verso Books, 2006), 180.

228

the city are resistant to change that any transformations in the urban structure are necessarily the
outcome of these movements (Castells 1983, xviii). Castells uses the example of the Citizen
Movement in Madrid to show that, through neighborhood mobilization and protest, urban social
movements are able to accomplish social change. These movements are not agents of structural
social change, but “symptoms of resistance to social domination, even if, in their effort to resist,
they have major effects on cities and societies” (Castells 1983, 329):
To accomplish this social innovation, neighborhood protest had, in its practice, to
articulate urban demands, the search for community, and political goals
challenging the state’s institutions. It also had to define its own mobilization as
part of a broader social movement led by citizens, and had to be related to society
through several operators, requiring at least the support of the media,
professionals and political parties (Castells 1983, 278).
The struggle of these movements goes beyond the class struggle between capital and
labor, because it brings together a diverse range of social groups, not just the working class and
the urban poor (Castells 1983, 276). Urban social movements are active in slums, too. Castells
expresses optimism that, “contrary to those who belief in the myth of marginality…social
organization seems to be stronger than social deviance in [squatter] communities, and political
conformism seems to outweigh the tendencies towards popular upheavals” (Castells 1983, 175).
Castells bases his optimism about the potential for grassroots activity in slums on “the local self-
organization of squatter settlements and its particular connection to the state and the political
system though urban populism” around the delivery of land, housing and public services. There
are new forms of interaction between the state and the “popular masses” and between the city and
the grassroots “in the Third World in general, and in Latin America in particular”, which are the
result of a growing political courtship of slum populations (“populism”) as well as a “growing
social concern for squatter settlements” (Castells 1983, 175).
Castells’ later work on grassroots urban movements appears to share a major premise of
the Foucauldian concept of capillary power—the willingness to critique determinism and

229

mechanistic doctrines of neo-Marxist urban political economy and “bring people back into
history, not merely as quaintly suffering folk but as an active force transforming social
environments” (Molotch, 141). Moreover, Castells is close to the Foucauldian notion of the
battlefield when he says that the theoretical challenge is to “understand the production of cities
and urban meaning as a conflictive and permanently open process between historic social actors”
(Castells 1983, 72). But Castells’ notion of “community” reflects perhaps an overly pluralist faith
in the possibility of collective action, particularly given the wide range of different citizens’
groups in cities. Are these citizens’ groups really as “homogenous” as Castells suggests (Castells
1983, 328)? In the slum context, Castells’ faith in the “local self-organization” among slum
dwellers underestimates the fragmentation inherent in slum populations and interests. And it is
ironic that Harvey Molotch accuses Castells, a neo-Marxist, of “wishful romanticism” for
focusing on societies and groups and local community as “Great Actors of History” without
paying sufficient attention to “structural contingencies”, history and larger contexts (Molotch,
141-2) that form the backdrop to these local-level struggles.
4.3.5 The New Empowerment Discourse
In a sign of the new policy discourse on slums and poverty, a self-styled “war against
poverty” has taken center stage in international development. For over a decade the multilateral
development banks, donor agencies and national governments have spent billions of dollars
trying to tackle a problem that is still poorly understood. Whereas ten years ago poverty was
measured principally by economic indicators such as low income and consumption, today it is
considered a broader phenomenon that encompasses all dimensions of human deprivation.
One area that is attracting growing interest in the fight against poverty is intrinsically
political—the concept of empowerment. In its 2000/2001 World Development Report, the World
Bank introduced empowerment as one of its three pillars of poverty reduction, along with

230

opportunity and security. “Powerlessness” and “voicelessness” were recognized to be key
characteristics of the experience of poverty
243
. Empowerment was adopted as a key strategy in
attacking poverty, not only to address the sense of exclusion of the poor, but also to contribute to
better targeting of services and programs to meet their needs (World Bank 2001, 15-16), and to
help build a climate for investments, jobs and growth
244
.
In its broadest sense, the World Bank defines empowerment as the expansion of freedom
of choice and action (Narayan 2002, 14). The empowerment strategy draws explicitly on
Amartya Sen, who, in Development as Freedom (2000) claims that freedom is both the primary
end and the principal means of development (Streeten, 154), and in that context, argues that
human deprivations severely restrict the substantive freedoms and capabilities that a person
enjoys to lead the kind of life he or she values (Sen 2000, 87).
The empowerment approach represents a radical shift in development policy. After
decades in which politically-neutral, technical interventions have largely failed to eradicate
poverty, the new approach offers the prospect that development agencies and national
governments will finally confront the deeply ingrained social values, entrenched interests, and
political will that many believe are at the root of poverty. The approach also appears to
acknowledge the dignity and capabilities of the poor themselves. Can the millions of poor,
through their own actions and by using their own capabilities, achieve for themselves the
development outcomes that outside technical assistance has been unable to deliver for them—if
only they receive the necessary political support?

243
Outcome of the Voices of the Poor study (Deepa Narayan ed., 2000), based on discussions with 40,000
poor people in 50 countries, conducted as background for the 2000/2001 World Development Report.
244
In its Strategic Framework Paper (FY2002-FY2004), the World Bank identified empowerment of poor
people as one of two priority areas for support to client governments to increase development effectiveness.

231

A critical question for development practice is: how to measure characteristics of poverty
such as voicelessness and powerlessness and, conversely, voice and power? The 2000/2001
World Development Report predicted that:
“measuring these [new] dimensions of poverty in an accurate, robust and
consistent way so that comparisons can be made across countries and over time
will require considerable additional efforts on both the methodological and data-
gathering fronts” (World Bank 2001, 19).
Since the announcement of its new poverty reduction strategy, the World Bank has been
at the forefront of efforts to come up with an analytical framework for empowerment in
development assistance. The development of instruments and indicators to evaluate
empowerment processes and outcomes is still at an early stage (Alsop and Heinsohn, 5), and there
is no uniform approach. Nevertheless, the 2000/2001 World Development Report and a
subsequent empowerment and poverty reduction sourcebook (2002) provide the outlines of an
initial empowerment framework and a basis for development interventions
245
.
A fundamental premise underlying the World Bank framework is that the poor are
located outside of power arrangements and possess little power themselves. The Bank’s Voices
of the Poor study concludes that the poor lack access to resources and opportunities.
Powerlessness results from a dependency on others (Narayan 2000, 11, 266). In addition, the poor
are “confronted with unequal power relations” and “unable to influence or negotiate better terms
for themselves with traders, financiers, governments, and civil society”, and are “dependent on
others for their survival” (Narayan 2002, 13). It is assumed that this powerlessness is embedded
in social norms and the nature of institutional relations. Poor people’s choices are limited both by
their lack of assets and by their powerlessness to negotiate better terms for themselves with a

245
The sourcebook on empowerment and poverty reduction (Narayan, 2002) presents a framework and
strategic actions to be supported by the World Bank.

232

range of institutions that control their lives—both formal and informal. From this flows the
Bank’s initial working definition of empowerment in the context of poverty reduction, as:
“the expansion of assets and capabilities of poor people to participate in,
negotiate with, influence, control, and hold accountable institutions that affect
their lives” (Narayan 2002, 14).
Action is recommended at two levels. The first level is the reform of institutions,
particularly state institutions, to make them more accountable and responsive to poor people
(World Bank 2001, vi). The second level is building the assets and capabilities of the poor
themselves. The sourcebook identifies four empowerment approaches (“elements of
empowerment”) that support institutional reform at these two levels. At the first level, state
institutions and governments must provide timely access to information for citizens, and create
accountability mechanisms—politically (through elections), in the overall administration (through
internal mechanisms), and socially (to citizens). State institutions should also support
mechanisms for inclusion and participation in planning, priority setting and policymaking. At the
second level, the local organizational capacity of the poor themselves needs to be supported, i.e.,
the ability of the poor to work together, organize themselves, and mobilize themselves to solve
problems of common interest (Narayan 2002, 18-22). But the premises of the Bank’s
empowerment framework are problematic in the context of informal settlements and land
conflicts, where much evidence suggests that the poor are not powerless all the time and that
there are different manifestations of power, depending on the context.
Another problematic assumption underlying the empowerment framework is that the poor
possess “high ‘bonding’ social capital” and enjoy “close ties and high levels of trust with others
like themselves” (Narayan 2002, 15). This assumption is questionable, certainly in the urban
informal settlements in Phnom Penh, which are in fact very heterogeneous and socially quite
fragmented. While it may indeed be the case that “organized communities are more likely to

233

have their voices heard and their demands met than communities with little organization”
(Narayan 2002, 22), the capacity and desire for participation and collective action on the part of
the poor—key pillars of the World Bank’s framework for action—should not be overestimated, at
least not in urban areas.
4.3.6 Grassroots Power and the Case of Slum/Shack Dwellers International
One of the most concrete examples of collective action and “people’s power” in slums
today is provided by the emerging transnational network of housing activists, given official form
in 1996 as Slum/Shack Dwellers International, a network of federations of the poor at grassroots
level in scores of countries on four continents. Its model is explicitly one of self-reliance: the
network helps the urban poor to organize themselves and improve their own conditions and
rejects the paternalism and project-driven approach of outsiders, including international experts
and development agencies.
Arjun Appadurai (2002) analyzed one prominent example of this grassroots network—an
urban activist movement in Mumbai that is in many ways a standard bearer for the international
movement of urban poor federations, whose leader, Jockin Arputham of the National Slum
Dwellers Federation (NSDF) in India, has been one of the earliest and most significant forces
behind urban poor activism in Phnom Penh. The Mumbai network is an Alliance of three very
diverse organizations: the NSDF, a community-based organization; SPARC, a non-governmental
organization working on urban poverty, with strong ties to global funding sources and networking
opportunities; and Mahila Milan, an organization of poor women with a network of branches
throughout India. The Alliance is dedicated to helping slum dwellers in Mumbai access secure
land tenure, adequate housing, and basic infrastructure, not through the familiar models of “social
work, welfarism, and community organization”, but through a form of pro-poor activism that
recognizes and builds upon the agency of the poor themselves.

234

In the face of the “tyranny of emergency” that characterizes the lives of the poor
246
, the
Alliance’s long-term political strategy is a “politics of patience” that is aimed at long-term asset
building rather than short-term solutions (Appadurai 2002, 30). In the slums this strategy refers
to the slow process of mobilization, teaching, and learning that build on what the poor already
know and understand. As a political approach it refers to a strategy of accommodation,
negotiation, consensus building and long-term pressure aimed at whoever is in power at the
national, state, municipal and local levels, rather than a policy of direct confrontation or threats of
political reprisal (Appadurai 2002, 28-9).
The work of the Alliance is based around three central concepts: federating, savings, and
precedent-setting. The concept of “federation” encompasses the idea of political union among
preexisting collectives as a way for the poor themselves to enact change in the arrangements that
disempower them. Within each collective individuals and families organize themselves, as
members, to engage in advocacy, provide mutual risk-management devices, and confront
opponents, when necessary (Appadurai 2002, 32). The significance of federating is that it
provides the basis for collective action in the slum:
The idea of federation is a constant reminder that groups (even at the level of families)
that have a claim to political agency on their own have chosen to combine their political and
material power. The primacy of the principle of federation also serves to remind all
members…that the power of the Alliance lies not in its donors, its technical expertise, or its
administration, but in the will to federate among poor families and communities. At another
level, the image of the federation asserts the primacy of the poor in driving their own politics,
however much others may help them to do so (Appadurai 2002, 32-3).

246
Appadurai cites a term coined by Jérôme Bindé, in “Towards an Ethic of the Future”, Public Culture 12
(2000): 51-72.

235

The bedrock of all federation activities are informal savings groups among the poor.
Savings has been used as a principal tool for mobilization in India and as an entry point to
relationship building with international urban poor groups, including in Cambodia and Thailand.
Savings is framed as a “moral discipline” and as a technique for improving “financial citizenship”
for the urban and rural poor:
The practice builds a certain kind of political fortitude and commitment to the
collective good and creates persons who can manage their affairs in many other
ways as well. Daily savings, which do not generate large resources quickly, can
therefore form the moral core of a politics of patience (Appadurai 2002, 33-4).
The third concept of “precedent setting” is a technical strategy resembling a “show and
tell” exercise, whereby “the poor claim, refine, and define certain ways of doing things in spaces
they already control and then use these practices to show donors, city officials, and other activists
that their “precedents” are good ones and encourage such actors to invest further in them”
(Appadurai 2002, 34). But behind this seemingly innocuous technique lies a more radical
strategy that enables the poor to claim their own space, at a physical and legal level as well as in
discourse:
By invoking the concept of precedent as enshrined in English common law, the
linguistic device shifts the burden for municipal officials and other experts away
from a dubious whitewashing of illegal activities to a building on “legitimate”
precedents. The linguistic strategy of precedent-setting thus turns the survival
tactics and experiments of the poor into sites for policy innovations by the state,
the city, donor agencies, and other activist organizations. It is a strategy that
moves the poor into the horizon of legality on their own terms (Appadurai 2002,
34).
Precedent-setting is a classic example of Foucauldian counter-hegemonic discourse,
where slum dwellers aim to set precedents through their actions that counter the dominant
discourse of their illegality, and invite authorities and other outsiders to buy into the new
(counter-hegemonic) discourse by offering them new forms of partnership.

236

There are several other forms of counter-hegemonic discourse that the Alliance adopts in
its efforts to improve tenure security for its federation’s members—strategies that Appadurai calls
instruments of “counter-governmentality” from below, which have as their goal to challenge and
reverse the structural bias of existing knowledge processes about the urban poor emanating from
outside “experts” or the authorities. One example is the strategy of self-surveying and self-
enumeration: Alliance members are taught a variety of methods of gathering reliable and
complete data about households and families in their own communities (Appadurai 2002, 36),
rather than relying on official surveys and censuses undertaken by the authorities. The census is a
well-known instrument of power that facilitates the creation of a dominant discourse, a process
that Foucault identified as the “examination” or “surveillance” through the modern disciplines:
The examination as the fixing, at once ritual and ‘scientific’, of individual
differences, as the pinning down of each individual in his particularity…clearly
indicates the appearance of a new modality of power in which each individual
receives as his status his own individuality, and in which he is linked by his
status to the features, the measurements, the ‘gaps’, the ‘marks’ that characterize
him and make him a case (Foucault 1995, 192).
Population censuses have emerged as a flashpoint in the slums of Phnom Penh, as
authorities have used their “official” surveys of poor communities as a basis to declare individual
households and sometimes even whole settlements to be illegal occupants, and justify eviction.
For this reason, many community leaders have responded by organizing informal surveys of their
own.
Another instrument of “counter-governmentality” from below are housing exhibitions,
which the Alliance organizes in slums by and for the poor to demonstrate and discuss designs for
housing that suit slum dwellers’ own needs. The emphasis on people’s housing might seem like a
throw-back to the self-help housing movement of the 1960s and 1970s, but the significance of
these exhibitions for Appadurai is symbolic and discursive: they are an example of the “creative
hijacking of an upper-class form— historically developed for the display of consumer goods and

237

high-end industrial products—for the purposes of the poor” (Appadurai 2002, 37). Seen in this
manner the housing exhibitions take on the form of another counter-hegemonic discourse,
reversing conventional forms of expert knowledge about housing design and processes.
For Appadurai, the achievements of the Alliance in Mumbai represent “deep democracy”:
a “fluid and quixotic” process aimed at bringing improvement to slum dwellers’ lives and
conditions through inclusion, participation, transparency and accountability “as articulated within
an activist formation” (Appadurai 2002, 45). Deep democracy extends far beyond the boundaries
of the nation state and its traditional organs of governance—the SDI network represents
globalization from below, a “transformation in the nature of global governance” characterized by
an explosive growth of NGOs of all scales and varieties. This is an “emergent political horizon”,
a new “strategy of urban governmentality”, global in its scope, that presents a “post-Marxist and
post-developmentalist vision of how the global and the local can become reciprocal instruments
in the deepening of democracy” (Appadurai 2002, 25, 30).
Appadurai’s vision of deep democracy has yet to take root in the slums of Phnom Penh.
Despite over a decade of advocacy work with the Municipality of Phnom Penh and organization
of slum dwellers’ groups by various member organizations of the SDI global network, a true
people’s movement does not yet exist among the city’s slum dwellers. Evidence of this are the
disintegration of SUPF, which was set up by the SDI network in the mid-1990s to organize the
urban poor in Phnom Penh, and the failure of savings schemes in many of the city’s slums,
including the four pilot settlements. The failure of collective action thus far among the urban
poor, despite the best efforts of the new “strategy of grassroots urban governmentality”, to use
Appadurai’s words, points to the existence of tenacious structural forces in the city’s landscape of
power.

238

4.4 A Synthesis View: Foucault’s Concept of Capillary Power
In contrast to the extreme views of “slums of despair” (emphasizing the dominance of
structure) and “slums of hope” (emphasizing the dominance of agency), I argue that a more
relevant, synthesis view that can be applied to power relations in slums is provided by Michel
Foucault. Foucault is regarded as one of the pre-eminent modern philosophers, and power was at
the center of his work. Foucault rejected traditional analyses of power to be found in the social
sciences, which he argued did not account sufficiently for the “fecundity” of power (Thiele, 243).
Foucault’s work analyzed the modern subject and his relation to power and knowledge, and
investigated different forms of rationality as they are applied by human subjects to themselves
(Kelly 1994a, 372).
Within sociological and political science theory, Foucault contributed an unorthodox
understanding of power, based on his central premise that power is diffuse and omni-present
(“capillary”), with all actors in the role of subjects as well as objects of power. On the one hand
Foucault describes a web of power in which all individuals are caught, a structure where power
relations are “rooted deep in the social nexus, not reconstituted above society as a supplementary
structure” (Foucault 1982, 791), and where human subjects are re-written in ways which
challenge pluralist (humanist) understandings of self (Fox, 428). On the other hand, Foucault’s
concept of power is less monolithic than that of either Marx or Weber, leaving room for
individuals to deploy their subjective power successfully as part of manifold acts of resistance.
Foucault does not attempt to resolve the structure/agency dichotomy; instead, he conceives of
man as simultaneously subject of action and object of structure (Smart, 137). In so doing,
Foucault offers a post-modern, fragmentary social theory, accepting of paradox and contradiction,
and committed to opening up new readings rather than seeking truth in a definitive account (Fox,
429).

239

Foucault never produced a formal theory or methodology of power as such, but rather
proposed a series of self-styled “ideas” (Foucault 1982, 777). In my theoretical framework I
organize these ideas according to two main premises. The first is that power is a multiplicity of
force relations. The second is that, within this multiplicity of force relations, individuals are both
subjects and objects of power
247
.
4.4.1 Power as a Multiplicity of Force Relations
The Death of Law and Sovereignty
Foucault considered that political thought and analysis were still under the spell of an
outmoded representation of power he called the “sovereign”, “monarchical” or “juridical” modes
of power. In this perspective, power is a unitary and centralized construct that is primarily
repressive in nature (Fox, 416). In the juridical perspective, moreover, power always takes on a
negative expression, including “refusal, limitation, obstruction, and censorship”—in sum, the
“kind of power that says no” (Foucault 1980, 139). This representation of power has survived for
so long in political analysis because it is a heritage of Western society, which has, since the
Middle Ages, conceived of and exercised power in terms of law and sovereignty (Thiele, 247).
The sovereign mode of power attaches importance to the “problem of right and violence, law and
illegality, freedom and will, and especially the state and sovereignty”, questions that for Foucault

247
Foucault stated that the multiplicity of force relations can be coded either in the form of “war” or
“politics” (Foucault 1978, 93). In the 1970s, with his work on human sexuality in the series on The History
of Sexuality, the metaphor Foucault used for power was that of the battlefield, but in the early 1980s
Foucault’s emphasis shifted towards describing power as “government”, from the Latin term gubernare, to
steer (Thiele, 258). The analytic of power that I use for my theoretical framework dates mainly from
Foucault’s work on human sexuality in the 1970s. This is because, firstly, Thiele argues that “Foucault’s
new field of investigation [in the 1980s] did not dissociate itself from his previous studies of
power/knowledge” (Thiele, 259) and therefore there is no substantial break in Foucault’s thinking on power
between the 1970s and 1980s. And secondly, I argue that the idea of “force relations” is the most
applicable metaphor to use in the context of redevelopment in a developing country urban context, and the
“struggle” for secure tenure on the part of the poor.

240

are “characteristic yet transitory” because “while many of [these] forms have persisted to the
present, [they] have gradually been penetrated by quite new mechanisms of power that are
probably irreducible to the representation of law” (Foucault 1978, 88-9).
The new mechanisms of power, for Foucault, are:
utterly incongruous with the new methods of power whose operation is not
ensured by right but by technique, not by law but by normalization, not by
punishment but by control, methods that are employed on all levels and in forms
that go beyond the state and its apparatus (Foucault 1978, 89).
The “theoretical privilege of law and sovereignty” is no longer capable of “coding
power” and of “serving as its system of representation” (Foucault 1978, 89-90). In its place
Foucault presents a construct of power that must be understood as a “multiplicity of force
relations”, which “must not be sought in the primary existence of a central point, in a unique
source of sovereignty from which secondary and descendent forms would emanate”, but which is
a “moving substrate of force relations which, by virtue of their inequality, constantly engender
states of power, but the latter are always local and unstable” (Foucault 1978, 92-3). No one
person or no one group, be it the sovereign, the bourgeoisie or the state, can intentionally be the
author of all power or possess it absolutely:
One should not assume a massive and primal condition of domination, a binary structure
with ‘dominators’ on one side and ‘dominated’ on the other, but rather a multiform production of
relations of domination which are partially susceptible of integration into overall strategies
(Foucault 1980, 142).
Neither the caste which governs, nor the groups which control the state apparatus, nor
those who make the most important economic decisions direct the entire network of power that
functions in a society (and makes it function)(Foucault 1978, 95).
Power in the multiplicity of force relations has other important characteristics as well. In
terms of its form, power is first and foremost an exercise rather than something that can be

241

possessed. It is part of a chain of forces which are not only negative but also productive. In
terms of its location, power is ubiquitous, and it comes from below.
Power’s Form: Relational and Productive
In the juridical perspective, because power is monolithic, it can be possessed; it
resembles “something that is acquired, seized or shared, something that one holds on to or allows
to slip away” (Foucault 1978, 94). But in a multiplicity of force relations power is exercised.
The exercise of power is a set of actions which acts upon the actions of others:
It is a total structure of actions brought to bear upon possible actions; it incites; it
induces; it seduces; it makes easier or more difficult; in the extreme it constrains
or forbids absolutely; it is nevertheless always a way of acting upon an acting
subject or acting subjects by virtue of their acting or being capable of action
(Foucault 1982, 789).
Moreover, power cannot be essentialized, for this would be to undermine its strategic
function (Fox, 419), and it would also underestimate its dynamic nature. The metaphor Foucault
used to express the constant strategic nature of power is that of the battlefield:
[P]ower exercised on the body is conceived not as a property, but as a
strategy…its effects of domination are attributed not to “appropriation”, but to
dispositions, manoeuvres, tactics, techniques, functionings;…one should
decipher in it a network of relations, constantly in tension, in activity, rather than
a privilege that one might possess; …one should take as its model a perpetual
battle rather than a contract regulating a transaction or the conquest of a territory
(Foucault 1995, 26).
Implicit in these interpretations of power as an exercise and strategic is the relational
nature of power. In the multiplicity of force relations, forces are connected as in a chain or
system, where “force relations find support in one another”, or alternatively, experience
“disjunctions or contradictions which isolate them from one another” (Foucault 1978, 92-94).
Relations of power do not operate outside of other relations (economic, sexual, or knowledge
relations) but are immanent within these.

242

A critical aspect of power’s form that distinguishes Foucault’s perspective from other
perspectives within sociological and political science theory is that power is productive, rather
than (solely) negative. While Foucault does not deny that power can occasionally manifest itself
in a top-down way, as subjugation, he finds that the repressive model is too narrow (Allen, 133):
If power were never anything but repressive, if it never did anything but to say
no, do you really think one would be brought to obey it? What makes power hold
good, what makes it accepted, is simply the fact that it doesn’t only weigh on us
as a force that says no, but that it traverses and produces things, it induces
pleasure, forms knowledge, produces discourse. It needs to be considered as a
productive network which runs through the whole social body, much more than
as a negative instance whose function is repression (Foucault 1980, 119).
One of the main things power produces is discourse and “effects of truth”. In the
multiplicity of force relations these are important elements of structural power (in the form of
hegemonic discourses) as well as opportunities for agency (in the form of counter-hegemonic
discourses). Discourses and effects of truth are “in themselves…neither true nor false” (Foucault
1980, 118); instead, they have strategic value as weapons of power. For Foucault, therefore,
power lacks a moral dimension. Because power is not simply negative or repressive but enjoys
strategic value, it is not a priori an evil. Power is defined simply as the capacity to modify the
actions of others; as such it is inherently free of value judgments (Heller, 110n). If the moral
dimension enters at all, it may come in only at the end point of an analysis of power, when it
becomes clear how power is used: whether it is used to increase the freedom of others, or to
capture them in relations of domination (Heller, 103-4).
Power’s Location: Ubiquitous and Micro-level
As a result of its location within a multiplicity of force relations, “power is everywhere”:
[This is] not because it has the privilege of consolidating everything under its
invincible unity, but because it is produced from one moment to the next, at
every point, or rather in every relation from one point to another” (Foucault
1978, 93).

243

It is precisely because power is ubiquitous that Foucault claims that “power comes from
below”:
[The] manifold relations of force that take shape and come into play in the
machinery of production, in families, limited groups, and institutions, are the
basis for wide-ranging effects of cleavage that run through the social body as a
whole (Foucault 1978, 94).
Foucault emphasizes the “local and intimate operations of power” or “micro-powers” that
are exercised at the level of daily life, rather than focusing exclusively on the supreme power of
the state. Power is exercised from innumerable points: it is “diffuse and capillary, omnipresent,
and both productive and repressive”, enabling and constraining. Because power is not localized
in the state apparatus, “nothing in society will be changed if the mechanisms of power that
function outside, below and alongside the state apparatuses, on a much more minute and everyday
level, are not also changed” (Foucault 1980, 60).
The sovereign or juridical perspective on power assumes that wherever individuals are
out of reach of the sovereign they are free from power (Allen, 133). But Foucault rejects this,
arguing that the analysis of power and relations of subjugation and domination should not
concern themselves principally with the “regulated and legitimate forms of power in their central
locations”, but on the contrary with power “at its extremities, in its ultimate destinations, with
those points where it becomes capillary, that is, in its more regional and local forms and
institutions” (Foucault 1980, 96).
Overall Foucault rejects a “binary and all-encompassing opposition between rulers and
ruled at the foot of power relations”, claiming instead that “more often one is dealing with mobile
and transitory points of resistance, producing cleavages in a society that shift about, fracturing
unities and effecting regroupings, furrowing across individuals themselves…”. The result is a
“dense web that passes through apparatuses and institutions, without being exactly localized in
them” (Foucault 1978, 96).

244

4.4.2 Combining Structure and Agency
The dualism between structure and agency is a persistent theme in the social sciences and
in social theory, and when structure and agency are treated as two essentially different realms,
populated by two separate orders of being, the dualism becomes unresolvable (Fuchs, 24).
Foucault has attracted much interest by “reformulating” the structure/agency dualism, not by
pretending to resolve or synthesize a dualism which may remain insoluble within the terms of the
human sciences (Smart, 121, 137), but by conceiving of man as both subject and object, and
therefore suggesting how these seemingly paradoxical extremes might co-exist within the same
time and space.
Foucault is well-known for his “seemingly paradoxical statement” (Heller, 79) that
“power relations are both intentional and non-subjective” (Foucault 1978, 94-5). Foucault breaks
with structuralism by denying that power is merely coercive and by presenting power as a
“productive process, creating human subjects and their capacity to act”
248
. But Foucault also
breaks with pure theories of agency (humanism), by de-centering the individual as the prior agent
in creating the social world, rejecting subjectivity as something essential, prior to discourse,
which power acts against (Fox, 417). In the Foucaldian approach to power, individuals are
subjected to the multiple, shifting relations of power in their social field (repressive power), while
at the same time they play the role of subject in and through those relations (productive power)
(Allen, 135).
In his works Foucault described several forms of structural and agency power. The most
prominent of these are summarized in Table 4 and described in more detail in this section.

248
J. Butler, Gender Trouble (London: Routledge, 1990), 139, quoted in Nick J. Fox, “Foucault,
Foucauldians, and Sociology”, The British Journal of Sociology 49, no. 3 (1998): 417.

245

Structure/Object
Sources of structural power
Agency/Subject
Sources of subjective power
Power diagram Resistance
Unintentional consequences of action: strategies Intentional consequences of action: tactics
Hegemonic discourse Reverse discourse
Table 4: A rudimentary classification of Foucauldian forms of structure and agency power

On one side there are several forms of structural power which may be said to constrain
individuals and their agency. The most fundamental and all-encompassing constraint is the
“power diagram”. The second constraint is the unintended outcome of action, or strategy. The
third constraint is hegemonic discourse. On the other side are several forms of power that give
individuals the possibility and the space to become agents by assuming “counter-hegemonic
subject-positions”. The most fundamental of these is what Foucault calls the “inevitability of
resistance”. The second form of subjective power can be considered another form of resistance:
these are the intentional consequences of action, or tactics. The third of these forms of power is
the adoption of counter-hegemonic discourse. The end result of these competing forms of
structure and agent power is an ongoing power struggle—a “war”, according to Foucault—where
the power exercised by subject-position X will always be opposed by the power exercised by
subject-position Y (Y
1
, Y
2
…Y
n
) (Heller, 99). It is against this backdrop of continuous force
relations that the forms of agency permit individuals to engage in a form of struggle for social
justice, despite the presence of a dominant hegemony.
The Power Diagram
The notion of the “power diagram” illustrates Foucault’s understanding of the
relationship between structure and agency. The exercise of power always depends on a pre-
existing system of power relations, the power diagram, which represents the totality of the

246

structurally determined mechanisms of power. While the decision to exercise power is always
intentional, the mechanisms of power that individuals use to exercise that power are inherently
non-subjective because they do not depend on the existence of those individuals for their own
existence. There is a dialectical relationship between the mechanisms of power and the exercise
of power: the exercise of power continually transforms a diagram’s mechanisms of power, yet it
is only possible through the utilization of those same pre-existing mechanisms (Heller, 85).
Hence, “every relationship of power puts into operation differentiations [mechanisms of power]
which are at the same time its conditions and its results (Foucault 1982, 792).
As a result of the existence of the power diagram, all individuals are equally trapped
within a system of power relations that is beyond their complete control. No individual or group
has the power to control the entire power diagram, and no one has absolute power:
One doesn’t have here a power which is wholly in the hands of one person who can
exercise it alone and totally over the others. [The power diagram] is a machine in which
everyone is caught… Power is no longer substantially identified with an individual who possesses
or exercises it by right of birth; it becomes a machinery that no one owns (Foucault 1980, 156).
But this does not mean that power is equally distributed—certain structural positions
within the diagram enable certain individuals and groups to control more of a diagram’s
mechanisms of power than others (Heller, 86): “certainly, everyone doesn’t occupy the same
position. Certain positions preponderate and permit an effect of supremacy to be produced”
(Foucault 1980, 156).
Unintended Consequences of Action
A second form of non-subjective power has to do with an individual’s use of power: there
is an inevitable disjunction between an action’s intention and its actual effect. The exercise of

247

power is always potentially non-subjective because of the unintended consequences of action
(Heller, 87). Foucault pointed out that “people frequently know why they do and what they do;
but what they don’t know is what they do does”
249
.
Foucault makes a distinction between “tactics” and “strategies”. Tactics are the
intentional actions carried out in determinate political contexts by individuals and groups, while
strategies are the unintentional effects produced by the non-subjective articulation of individual
and group tactics (Heller, 87-8). Both tactics and strategies involve power, because both create
social change, but only tactics involve intentional and therefore subjective power. Heller (1996)
claims that differentiating tactics and strategies enables Foucault to create not only a taxonomy of
power but a taxonomy of institutions as well. Foucault distinguished between institutions that
function primarily through tactics, in which the real social function of the institution is
consciously recognized by those involved in its operation, and those that function primarily
through strategies, in which the real social function of the institution is only dimly perceived, if at
all, by the people involved in its operation (Heller, 88). For Foucault, a prime example of the
latter institution was the modern prison, which instead of rehabilitating the individual (the explicit
tactic of those operating the prison) has ended up producing more delinquency as its real social
function.
Hegemonic Discourse
The third restriction on an individual’s subjective power is in the form of discourse. The
intentionality of subjects is never completely their own because their choice of tactics is shaped
and limited by institutionally-determined linguistic and historical traditions over which they have
little, if any, control. As in the case of the power diagram, discourses temporally and

249
Michel Foucault, personal communication to Hubert Dreyfus and Paul Rabinow, quoted in Heller, 87.

248

ontologically precede all subject-positions. Subjects can only choose tactics that they are able to
discursively formulate (Heller, 91, 94).
As mentioned earlier in the discussion on the form of power, discourses do not have
natural meanings: they produce “effects of truth” that become meaningful only in their strategic
context. It is in this sense that power is “productive” for Foucault. Foucault described the
association of productive power with the fabrication of effects of truth as “power/knowledge”.
This is where “power assumes a relationship based on some knowledge which creates and
sustains it; conversely, power establishes a particular regime of truth in which certain knowledges
become admissible or possible”
250
.
It is this characteristic of power as “productive” that distinguishes Foucault’s perspective
on power from both radical and Marxist views of power (Philp, 35). The Marxist and radical
emphasis on the negative conceptions of power as repressive, constraining and distorting
implicitly presuppose the possibility of social relations not marked by the effects of power.
Foucault denies this possibility (Philp, 35-6), because “truth is not outside of power, or lacking in
power”:
[T]ruth isn’t the reward of free spirits, the child of protracted solitude, nor the
privilege of those who have succeeded in liberating themselves. Truth is a thing
of this world: it is produced only by virtue of multiple forms of constraint. And
it induces regular effects of power. Each society has its régime of truth, its
general politics of truth: that is, the types of discourse which it accepts and makes
function as true (Foucault 1980, 131).
One of the techniques of power/knowledge, and therefore discourse, is the “gaze”, which
can be likened to the act of surveillance and information gathering to inform and create a
discourse on its subject matter. By informing science and the professions about their subject

250
D. Armstrong, The Political Anatomy of the Body (Cambridge: Cambridge University Press, 1983),
quoted in quoted in Nick J. Fox, “Foucault, Foucauldians, and Sociology”, The British Journal of Sociology
49, no. 3 (1998): 416.

249

matter, the “gaze” makes possible psychological, medical, and penal knowledge and practices
(Kelly 1994a, 376-7), but it also turns people into objects of professions or disciplines. The power
and authority of these disciplines is played out in the everyday encounters between professionals
(or other privileged others) and their subject matter (Fox, 416). The notion of the “gaze” refers to
the process whereby subjects make objects of themselves in the course of establishing certain
practices and institutions, and establishing new sciences. It does not have solely a negative role,
but a positive and productive one as well by making psychological, medical, and penal
knowledge and practices possible (Kelly 1994a, 376-7).
Foucault’s fundamental rule about discourse is that there is no position outside discourse
or power/knowledge. Subjects are absorbed into knowledge and power (Strozier, 57):
It seems to me that power is ‘always already there’, that one is never outside it,
that there are no ‘margins’ for those who break with the system to gambol
in…Power is co-extensive with the social body; there are no spaces of primal
liberty between the meshes of its network (Foucault 1980, 141-2).
But despite producing limitations on intentionality, discourse for Foucault—just like
power itself—is neither monolithic nor all-encompassing, and for this reason there are openings
for subject positions and agency. There is a “complex and unstable process”, whereby discourse
can be both an “instrument and an effect of power, but also a hindrance, a stumbling-block, a
point of resistance and a starting point of an opposing strategy” (Foucault 1978, 101). Moreover,
there is not one discourse, but an inevitable multiplicity of discourses, which produce hegemonic
and counter-hegemonic subject-positions (Heller, 94). The world is not divided between
“accepted discourse” and “excluded discourse”, but there is a “multiplicity of discursive elements
that come into play in various strategies” (Foucault 1978, 100). And it is this creation of counter-
hegemonic discourses which is one of the main forms of agency that individuals have at their
disposal.

250

The Inevitability of Resistance
The “seeming paradox” in Foucault’s propositions on power is that despite the apparently
crushing burden on individual intentionality presented by the power diagram, hegemonic
discourse, and the unintended consequences of action, subjects can still escape their situation of
determinism and deploy their own power.
In Foucault’s classification of forms of power, the most important mechanism of agency
for every subject is his own resistance. Resistance to power is part of the exercise of power, and
ultimately just another form of power. As with power, resistance has the capacity to create social
change. Points of resistance are present everywhere in the power network. A “swarm of points of
resistance traverses social stratifications and individual unities” (Foucault 1978, 96):
There are no relations of power without resistances; the latter are all the more
real and effective because they are formed right at the point where relations of
power are exercised; resistance to power does not have to come from elsewhere
to be real, nor is it inexorably frustrated through being the compatriot of power.
It exists all the more by being in the same place as power; hence, like power,
resistance is multiple and can be integrated in global strategies (Foucault 1980,
142).
Power and resistance continuously affect and shape each other. On the one hand, “there is
no power without resistance”, and on the other hand, since resistance to power is part of the
exercise of power, resistance invites further power and resistance. There is thus a dynamic cycle
of power and resistance, illustrated by the simple diagram below:

251

Power Resistance

Figure 17: The basis for agency: “there is no power without resistance”

But what is the nature of resistance? Foucault points out that there are various forms of
resistance. Rather than “one source of all rebellions”, there is a “plurality of resistances, each one
of them a special case” (Foucault 1978, 94). The different forms of resistance will be the subject
of the micro-level theoretical framework, described in the EVLN model in Chapter 5 and
evaluated in Chapter 7.
Counter-Hegemonic or Reverse Discourse
A distinct form of resistance that represents a second form of agency is the counter-
hegemonic discourse, or reverse discourse. In the same manner that power breeds resistance, so
discourse naturally engenders counter-discourse. Because no individual or group can control all
mechanisms of power, all social formations produce both hegemonic and counter-hegemonic
subject-positions, and all mechanisms of power are potentially capable of counter-hegemonic re-
appropriation (Heller, 102).

252

In his studies of the prison system (in the 1960s) and human sexuality (in the 1970s),
Foucault suggested many examples of counter-hegemonic discourse. In Discipline and Punish,
Foucault presents the bourgeoisie and the proletariat in 18
th
and 19
th
century Europe as two
structurally opposed subject positions, each with their own discourses. While bourgeois
broadsheets vilified proletarian “criminals” as individual psychopaths, in the popular discourse of
justice criminality became a political act, where condemned men found themselves transformed
into heroes, appearing to wage a collective struggle with which the proletariat all too easily
identified (Heller, 95-6). In The History of Sexuality, Foucault illustrated how two different and
opposed discourses existed around homosexuality in 19
th
century Europe. On the one hand there
was a hegemonic discourse where identification and classification of “the species and subspecies
of homosexuality, inversion, pederasty, and ‘psychic hermaphrodism’ made possible a strong
advance of social controls into the area of ‘perversity’” (Foucault 1978, 101). This set in motion
an anti-perversity discourse, which, like its anti-crime counterpart, having been given a name,
“began to assert its own value” (Heller, 98):
[The anti-perversity discourse] made possible the formation of a “reverse
discourse”: homosexuality began to speak on its own behalf, to demand that its
legitimacy or “naturality” be acknowledged, often in the same vocabulary, using
the same categories by which it was medically disqualified (Foucault 1978, 101).
Because discourse is potentially capable of being utilized in a counter-hegemonic way, it
is a “reversible” power mechanism, as are all power mechanisms (Heller, 102). Since discourses
do not have natural meanings, their meanings are no less contested than the institutions they
attempt to legitimize, and they can be used to support tactics that their “creators” never intended
(Heller, 102). No hegemonic group can ever exercise complete control over every counter-
hegemonic group; those groups will always have access to some mechanisms of power that they
can use to resist their domination. No group is therefore ever completely powerless (Heller, 101).

253

A Model of Force Relations
The main elements of my macro-level model of “capillary power” are now in place. The
model is of a multiplicity of force relations, a “perpetual battle” to use Foucault’s metaphor of
war, characterized by an “innumerable points of confrontation, focuses of instability, each of
which has its own risks of conflict, of struggles, and of an at least temporary inversion of the
power relations” (Foucault 1995, 27). Power relations go “right down into the depths of society”
(Foucault 1995, 27): the principal forces of my model at local level are the various forms of
structure and agency power. The main structural forces in the Foucauldian approach to capillary
power are the underlying power diagram, hegemonic discourses and strategies. The main sources
of power and agency available to individuals are multiple forms of resistances and tactics as well
as counter-hegemonic discourses. These structural and agency forces confront each other in a
constant cycle of action and reaction, power and resistance.

HEGEMONIC
DISCOURSE
Power diagram
Power diagram
Hegemonic
Discourse
Hegemonic
Discourse
Resistance
Resistance Resistance Resistance Resistance
Strategies
Strategies Strategies Strategies Strategies
Strategies
Tactics
Tactics
Strategies
C ounter-hegemonic
discourse
Counter-hegemonic
discourse
Resistance
Agency power
Agency power
Structural power
Structural power

Figure 18: Imagining the Foucauldian sphere of force relations

254

The resulting sphere of force relations might resemble something like the illustration in
Figure 18. The forces of individual agency are depicted in the inner circle, surrounded by
structural forces in the outer circle, reflecting the fact that both the power diagram and discourse
are pre-existing systems that “trap” all individuals. Nevertheless, because the power diagrams
and hegemonic discourses are multiple, and never entirely monolithic themselves, and because
power always engenders resistance, power never flows in only one direction, and individuals
have opportunities to deploy their subjective power.
4.5 The Slum as Sphere of Force Relations
Foucault’s “ideas” on capillary power provide a significant insight for an analysis of the
shape and outcome of slum redevelopment: the slum needs to be considered as a sphere of force
relations, where power does not merely flow neatly from top to bottom—from the “sovereign” to
the “oppressed”—but instead circulates in all directions, thereby making the outcome of the
redevelopment battle difficult to control or predict.
For residents of the four land sharing settlements in Phnom Penh, I argue that the sphere
of force relations is characterized simultaneously by structural forces and opportunities for
agency, in several significant ways. First, the major players in the slum redevelopment battle
(the residents, local authorities and developers) are not locked into fixed battle positions, but are
engaged in shifting alliances. Second, slum dwellers are not a monolithic group, but are divided
by internal fissures, so that within their ranks, there is a secondary sphere of force relations.
Third, the chaotic land tenure situation in Cambodia provides slum dwellers with opportunities
for “power from below” (in Foucauldian terms) even as it also subjects them to arbitrary policies
by the authorities. Fourth, the battle for living space in Phnom Penh is being waged just as
intensely at the level of discourse as it is at the physical level. In this discursive battle, slum

255

dwellers and their allies are subject to the hegemonic discourse of the authorities, but they also
resort to reverse discourses of their own.
This ever-changing sphere of force relations represents a very fluid situation, where
“property relations, use, and rights are defined through strategies of control, contestation,
negotiation and renegotiation. Such strategies span the spheres of market and non-market, legality
and illegality” (Razzaz 1993b, 345). In this situation, I argue, the outcome of the struggle for
contested space in the slum is not determined by the law, by bureaucratic diktat, or by any other
form of technical intervention, but by each actor’s relative power at a given moment.
4.5.1 Shifting Alliances
The struggle of slum dwellers to fight against eviction and remain in the city does not
merely resemble a straightforward battle between the residents, the authorities, and private
capital. None of the actors enjoys “sovereign” power—each is divided internally, and is at times
at odds with other internal elements as well as with other external actors. The authorities at
central and local levels have different agendas. There are internal divisions among the residents—
some residents want to stay in the city center at all costs, whereas others are willing to leave the
city if they are adequately compensated. Private developers compete with each other by using
their connections with public authorities to buy public land and development contracts, or bypass
the authorities altogether to negotiate directly with residents. Alliances between and within each
of these three parties are fragile and subject to change. Ideological factors are often much less
important in cementing alliances than straightforward practical considerations, such as money,
favors, and information.

256

4.5.2 Fragmented Slum Populations
In contrast to the optimistic pronouncements of the new empowerment school that the
poor enjoy “high ‘bonding’ social capital”, the populations of the four land sharing sites in
Phnom Penh are hardly monolithic, and even highly fragmented; internal divisions appear in
relation to income status, tenure status, ethnic identity, and along other social lines. First, there
are relatively better-off residents within each settlement and there are the very poor. Second,
there are big differences in social status and tenure security between house owners and renters.
Third, some social or ethnic groups stand out in the slum environment, and are vulnerable to
discrimination: these include the ethnic Vietnamese minority, women headed households, and (in
Borei Keila) households with members who have HIV/Aids. Finally, there is a clear gap in status
and authority between community members and “ordinary” residents. Power relations take place
between residents within the same slum in the same way that they take place between residents
and developers and local authorities outside the settlement.
It follows that slum populations cannot automatically be considered as “communities” in
the sociological sense of Gemeinschaft, i.e., as a group of people who have shared interests and
(therefore) are assumed to have a bond with one another
251
. Instead, in the Phnom Penh context,
a sense of solidarity cannot be taken for granted. Slum dwellers are still poorly organized. Most
of the existing “community groups” have been set up with the sanction of the Municipality of
Phnom Penh, and are kept in place by the local authorities at district and commune level, and
therefore these groups do not represent true residents’ associations per se rather than mechanisms
of control of the authorities. An emerging network of non-governmental organizations from the

251
My use of the word “community” in this study follows the very specific official usage of the word in the
Phnom Penh informal settlement context, as a collection of (usually) between 50-100 families within a
settlement, officially sanctioned by the local authorities and headed by a “community leader” appointed by
the other leaders, with the approval of the local authorities. This form of the term “community” has a
geographical and administrative connotation rather than a sociological one.

257

SDI network is springing up to assist the slum dwellers improve their housing conditions, form
savings groups, and protect their housing rights in the face of urban redevelopment. At times,
this network of “allies” provides very real support to residents, in the form of savings groups,
community networks and advocacy, while at other times, some of these “allies”, such as SUPF,
are co-opted by the authorities and private capital and become hidden adversaries. The
Foucauldian skepticism of collective action appears to better describe the internal fissures in the
slums in Phnom Penh.
4.5.3 Power from Below: Tenure Confusion and Opportunities for the Urban
Poor
Fourth, the authorities and developers—supposedly the most “powerful” parties based on
their powers of contract, capital and connections—cannot always impose their will over the slum
dwellers. The residents sometimes prevail, often in less than obvious ways. Slum dwellers are
not merely passive victims, but actors in their own right, whose decisions have multiple impacts
on the outcome of urban development schemes and policies. Moreover, because the authorities
and developers are internally divided, competition among them is as much a feature of the urban
redevelopment process as is their effort to negotiate with, and displace, the slum dwellers, and the
residents’ own divisions.
One of the main sources of opportunity for “power from below” in the Phnom Penh
context has been the unclear legal and property rights regime in Cambodia. As a result of the
legacy of the Democratic Kampuchea (Khmers Rouges) regime, until recently less than one
percent of all properties had an official legal title. As a result of a series of sub-decrees in the late
1980s and a new Land Law in 2001 that aimed to regularize the land market again after the
Khmer Rouge period, residents were given some form of possession right to their properties,
provided that their occupation was peaceful and the land involved is not “state public land”.

258

Today, “title” is in most cases legitimized de facto and informally by a bewildering array of
receipts and certificates, sometimes with official stamps from the local commune authorities. The
result is “tenure confusion” on a large scale. The urban poor benefit from this legal confusion (to
grab land) as much as they are victims of the authorities’ periodic arbitrary enforcement of the
law. Most slum dwellers consider themselves to be legitimate “owners” of their properties on
account of their long-time residence in the area, but their status is often disputed by the
authorities.
4.5.4 The Discursive War for the Slum in Phnom Penh
Finally, one of the most significant fronts in the battle for living space in Phnom Penh is
at the level of discourse. The slum redevelopment process in Phnom Penh is an ongoing war of
competing discourses and symbols. On one side, the authorities’ push to evict city-center slum
populations is guided by the official discourse that the state is merely trying to enforce public
order by removing dirty and unruly city-center slums, which are nothing more than “illegal”
settlements populated by “anarchists”. This discourse conveniently ignores the fact that,
according to the Cambodian Land Law of 2001, most of the residents have a legal possession
right to the land on account of their peaceful occupation of this land before the Land Law’s
administrative cut-off date of August 2001.
On the other side, slum dwellers and their allies in the political opposition and non-
governmental organizations accuse the authorities of acting like the “new Khmers Rouges”,
pushing out people from the city who are rightful “owners” of lands that most of them have
occupied peacefully for decades, many of them since the collapse of the Democratic Kampuchea
regime in 1979, when they came to Phnom Penh to rebuild their lives after years of war and
hardship. The residents’ reverse discourse conveniently overlooks the fact that, for many
residents, full “ownership” is more a state of mind than a foregone legal conclusion. According

259

to the 2001 Land Law possession must be officially transferred into ownership title by the
residents, yet very few possessors have bothered to do so over the years. The result is that, while
many slum dwellers are legal possessors of their properties, they continue to exist in a situation of
legal uncertainty that the authorities are able to exploit.
In this context of legal fuzziness, discourse and reverse discourse have emerged as
critical weapons, with each side using symbols (including the expressions “anarchy”, “order”,
“illegality” and “Khmers Rouges”; see Chapter 2) that are loaded with meaning in the Cambodian
historical context. In the spirit of Foucauldian power/knowledge and the fabrication of effects of
truth through discourse, slums should be analyzed first and foremost as a social construct:
discourse and symbolism are the main instruments of power used to “create” the slum in the first
place, and they are subsequently the main forms of power wielded in the struggle to dispossess, or
alternatively, to enfranchise slum dwellers. The war of symbolism and discourse vis-à-vis the
slum shapes how outsiders justify their policies towards slums, and determines how slum
dwellers perceive themselves.

260

CHAPTER 5: RESEARCH DESIGN AND METHODOLOGY
5.1 Overview of Chapter 5
In the first four chapters, this study has argued that slum redevelopment represents a
power struggle between residents of the slum and the forces of redevelopment, including private
developers and public authorities. According to the principle of “capillary power” of Michel
Foucault, the slum under redevelopment pressure resembles a sphere of force relations, where
power is ubiquitous, being imposed from above by so-called dominant forces as well as
emanating from below by the slum dwellers, through micro-power resistance strategies.
In this chapter I detail the research design and methodology that I use to analyze the
shape and outcome of the power struggle of slum dwellers in the four urban poor settlements that
were announced as pilots for the Government’s slum upgrading program through the technique of
land sharing. In the first part of this chapter I operationalize the concepts of empowerment and
citizenship in the context of the land sharing settlements in Phnom Penh. I will apply an expanded
version of Albert Hirschman’s original framework of exit, voice and loyalty, which was
originally introduced to analyze the responses of consumers and citizens to conditions of decline
in products and services and public organizations. In section 5.2 I provide the theoretical context
for the model. In section 5.3 I explain how I will operationalize the concepts of citizenship and
empowerment using an adapted version of the EVLN model of Lyons and Lowery, and in section
5.4 I present the main theoretical propositions flowing from my adapted EVLN model.

261

In the second part of this chapter (section 5.5) I describe my approach to data collection,
my questionnaire design, and the techniques I attempted to ensure a representative sample of
respondents for the structured interviews. Finally, I highlight the possible bias inherent in the
research methodology.
5.2 The EVLN Model in Theoretical Perspective
Since its introduction, Hirschman’s original typology of exit, voice and loyalty has been
applied to analyze individual behaviors in a wide range of situations in different disciplines and
across the social sciences where there is deterioration of some form, and where there is some
level of dissatisfaction as a result. Among others, it has been applied to analyze health policy
(Davies, 1973; Klein, 1980); romantic relationships (Rusbult, Zembrodt and Gunn, 1982; Rusbult
and Zembrodt, 1983); state formation and nation building in European history (Rokkan, 1974);
many studies of employment, job satisfaction and labor issues (Rusbult, Farrell, Rogers and
Mainous, 1988; Withey and Cooper, 1989); urban policy and citizen responses (Orbell and Uno,
1972; Lyons and Lowery, 1986); and urban policy more broadly (Young, 1972; Dei Ottati, 2000).
Hirschman used the framework himself to analyze the collapse of the former German Democratic
Republic (1993). The literature generated by Albert Hirschman’s original work, Exit, Voice and
Loyalty, has been aimed principally at producing suggestions for strengthening the opportunities
for voice by individual consumers (Klein, 423).
I argue that the exit, voice and loyalty framework is well-suited to an analysis of
individual behavior within a Foucauldian context of “capillary power”. Its vantage point is “from
below” (Hirschman 1980, 440), from the point of view of individual customers or members
during periods of dissatisfaction. The framework emphasizes the relational nature of power
which is so central to Foucauldian capillary power. The individual customer or member is not an

262

island on his own; his strategies have an impact on others, which in turn will impact him—these
various impacts are all part of his calculations. Moreover, the model is interested in the
situational determinants of participation whereas the traditional participation literature has tended
to focus on individual predispositions, such as demographic and psychological factors, that cause
individuals to participate (Lyons and Lowery 1989, 852). The implication is that as individuals’
situations change, their levels of participation and behaviors will also change. This fits well with
the fluid nature of power relations in the capillary model of power.
But perhaps most important, the exit, voice, and loyalty framework is recognized as a
“pioneering endeavor to bring together two models of man within the same framework of
analysis”: the two disciplines of economics and politics (Klein, 417). On one side is exit,
representing the market response, neat and impersonal (Hirschman 1970: 15, 120), and on the
other is voice, representing “political action par excellence” (Hirschman 1970, 16). Hirschman
bridges the “fundamental schism” between economics and politics, by acknowledging the
usefulness and full reciprocity of both economic and political concepts in the model; by
acknowledging that exit and voice were never intended to be mutually exclusive (Hirschman
1970, 15); by recognizing that the two behaviors affect each other in a multitude of different
ways; and by concluding that some institutions “need both exit and voice to be maintained in
good health” (Hirschman 1970, 126). The result is
an exceptional opportunity to observe how a typical market mechanism and a
typical nonmarket, political mechanism work side by side, possibly in harmony
and mutual support, possibly also in such a fashion that one gets into the other’s
way and undercuts its effectiveness (Hirschman 1970, 18).
The interplay of exit and voice, and of economics and politics—sometimes supportive of
each other, sometimes undercutting and battling each other—thus resembles the capillary nature
of power described by Foucault. In much the same way, I argue that Hirschman’s conceptual
framework bridges the traditional schism between structure and agency, and determinism and

263

subjective power. Hirschman’s original exit, voice, loyalty framework has been called a
“utilitarian paradigm”, with the customer or member in the role of rational agent, weighing,
calculating and evaluating the benefits and costs of his actions, and choosing the course of action
that is the most gratifying, easy and efficient to him (Bajoit, 327). The individual can actively
resist or protest (voice) or leave a deteriorating situation (exit), or he can choose to passively
follow events and hope for the best (loyalty). While Hirschman, as an economist, certainly based
his model on the complete intentionality of the main actors involved, he also warned against “the
assumption of undeviating rationality” (Hirschman 1970, 11): neither firms, organizations and
states, nor customers and members are fully rational agents. Firms normally aim at no more than
a “satisfactory”, rather than an optimal, rate of profits. Moreover, firms and other organizations
are “conceived to be permanently and randomly subject to decline and decay”, and are subject to
a “gradual loss of rationality, efficiency and surplus-producing energy, no matter how well the
institutional framework within which they function is designed”. In time, this process of decline
activates its own counter forces (Hirschman 1970, 11, 15). Customers and members, meanwhile,
display their own form of irrationality. For example, citizens do not normally use up more than a
fraction of their political resources, and do not participate to their fullest extent, due to a degree of
apathy (Hirschman 1970, 14). Ultimately, the assumption of less-than-full rationality pervades
the exit, voice, and loyalty model: individuals approach their choices strategically, with less-than-
full information, and, when none of the options open to them are satisfying, sometimes indeed
with apathy. In this way, Hirschman’s framework is a true conjunction of economic and political
forces.
With the expansion of the exit, voice, and loyalty framework to include a fourth response,
neglect, the EVLN model becomes more evenly balanced between rationality and “irrationality”,
agency and structure. In the 1980s several users of Hirschman’s framework expanded the

264

original model so that it could encompass a broader range of attitudinal and behavioral
alternatives. Lyons and Lowery indicated that considerable emphasis was traditionally placed on
active forms of individual responses to dissatisfaction with local political decisions, but only
limited attention was given to more passive types of responses (Lyons and Lowery 1986, 324).
Therefore, the EVLN model is an attempt to correct the bias towards active behaviors and
contribute a greater awareness of the important dimension of passive behaviors. The sociologist
Guy Bajoit (1988) suggested that the original, utilitarian exit, voice and loyalty paradigm missed
an important aspect of human behavior by not explicitly acknowledging “apathy” as one of the
main behaviors. Bajoit pointed out that apathy was already implicit in the original typology, but
that Hirschman incorrectly categorized it as part of loyalty, even though apathy is non-
cooperative while loyalty is inherently cooperative (Bajoit, 328-30).
Lyons and Lowery (1986 and 1989) made a similar point, but they labeled their
additional response as “neglect”, which nonetheless displays many of the same passive, fatalistic
traits as apathy. In both cases, the inclusion of a fourth response with these passive features
enables a typology of behaviors along two critical dimensions of social (and thus power)
relations—constructiveness/destructiveness and activity/passivity. Voice and loyalty are two
cooperative or constructive responses, while exit and neglect are non-cooperative; voice and exit
are active, while loyalty and neglect are passive behaviors
252
. The significance of the addition of
the fourth response, neglect, is that it introduces the dimension of determinism in the conceptual
framework, alongside agency: even if the customer or member does nothing at all and expects the

252
For Bajoit, the two comparable dimensions are cooperation and social control. Voice and loyalty are
cooperative responses; exit and apathy are non-cooperative. Voice and exit challenge social control,
whereas loyalty and neglect reproduce social control. I have chosen to adopt the typology, terminology and
dimensions of Lyons and Lowery over that of Bajoit because the Lyons and Lowery model comes from an
“urban” perspective, and its focus on participation modes is closer to my area of study than the sociological
model of Bajoit.

265

worst through his apathy, he still makes a “choice”, although one that is perhaps neither rational
nor utilitarian but is more a reflection of a lack of better alternatives. Thus, the newly expanded
EVLN framework is able to accommodate within its dimensions both subjective (agency) and
structural behavior and other essentially different realms in the social sciences, such as macro and
micro, subject and object, and action and passivity. The range of options available to these
customers and members at micro level thus mirrors the competing mechanisms of structure and
agency, determinism and subjective power, swirling around in the sphere of force relations at
macro level.
5.2.1 Translating the EVLN Model to the Phnom Penh Context
William Lyons and David Lowery designed their EVLN model to examine the nature of
individual responses to dissatisfaction with local government services in the United States. The
model requires a certain amount of “reframing” to suit the Cambodian context.
In translating the EVLN model to the slum redevelopment situation in Phnom Penh I
make four basic assumptions. First, in my EVLN model, I assume that the slum—and not the
local government, as in the Lyons and Lowery model—is the appropriate geographical
“jurisdiction” for the analysis of residents’ responses to dissatisfaction. Second, I assume that the
nature of dissatisfaction of slum residents is not just related to a perceived decline in the quality
of the “tax/service package”, as in the American model, but to other problems specific to the land
sharing projects in Phnom Penh. Third, I assume that dissatisfaction among slum dwellers is
directed not (only) at the local government, as it is in the Lyons and Lowery model, but at the
whole range of different actors that are involved in the redevelopment of the slum. Fourth, I
assume that the individual household is the most appropriate unit of analysis to understand
agency at the micro level, as households are the most common decision-making units in
Cambodia.

266

The Slum as Physical and Social “Jurisdiction”
In the Lyons and Lowery model, the relevant “jurisdiction” for the analysis of EVLN
responses is the local government as a “spatially defined social world”, where it is assumed that
families with similar resources, beliefs and habits tend to cluster in urban space (Lyons and
Lowery 1986, 321-3).
In the EVLN model adapted to slums, the appropriate jurisdiction for the analysis of
EVLN responses by dissatisfied residents is the slum itself, which, as in the definition adopted by
Lyons and Lowery, should also be defined as both a physical space and a social world. Just as
local government jurisdictions in the United States offer different (physical) “tax/service
packages” and (social) “packages of lifestyle maintaining characteristics" so, I argue, does each
slum in Phnom Penh. As a physical space, the slum settlement is an area that people can exit or
enter in search of a satisfactory and cost-effective package of land, housing, and basic services.
As a social space, the slum offers a “social community” of sorts—albeit one which I will argue
cannot be assumed to offer solidarity and a uniform set of shared values. When slum dwellers
deliberate about exiting, voicing or remaining loyal in the face of their dissatisfaction in the slum,
they do so by weighing the costs and benefits of living there—which include the cost of land,
rents, and basic services, the benefits of the location and distance to employment opportunities,
schools, markets, health centers and pagodas, and good relations with neighbors, community
leaders and other residents—relative to the costs and availability of those services and the social
infrastructure outside the slum, in comparable low-income areas in the city or even outside the
city.
An important consequence of defining the slum as both a physical and social world is that
residents can belong to the same slum even when the slum changes its physical location. Such a
situation can occur when an entire slum population is moved from one location to another, as in a

267

voluntary relocation or an involuntary eviction: in those circumstances, residents who follow the
community to its new location cannot be considered “exiters” because they still have not left the
social world of the particular slum.
The Nature of Dissatisfaction in the Land Sharing Settlements
The Lyons and Lowery EVLN model is a theory of citizen response to dissatisfaction in
urban communities. In the U.S. context of the original theory, residents experience dissatisfaction
with local government policies and with the declining quality of the “tax-service package” and
the “package of lifestyle maintaining characteristics” in their local government jurisdictions.
In the context of the four informal settlements in Phnom Penh where the land sharing
projects are being planned, I argue that there are three main sources of dissatisfaction for
residents. The first is dissatisfaction with the poor living conditions and services in the slum
environment. This source of dissatisfaction most closely approximates the dissatisfaction with
the “tax-service package” of the Lyons and Lowery model, except that in the Phnom Penh case,
dissatisfaction is caused not necessarily by declining public services but by the absence of many
public services.
The second source of dissatisfaction is the growing sense of uncertainty about the
outcome of the redevelopment projects in all four settlements. In Borei Keila, where a land
sharing deal has been agreed and the first community buildings have been constructed, progress
on the remaining buildings is slow, leading some residents to question whether the developer will
ever finish the projects. In Dey Krahom and the two Railway settlements, the land sharing plans
are not favored by developers. Despite the high level official support for land sharing, residents
question whether the Municipality and central government would impose a land sharing solution
over the objections of powerful private developers. The nagging uncertainty feeds fears of forced

268

eviction—an outcome the residents see all around them in the Phnom Penh in the first decade of
the new millennium.
A third source of dissatisfaction of many slum dwellers is the fraud and corruption that is
rumored to be enveloping the four land sharing projects. At a “macro level” residents worry
about the close ties between the developers and the Municipality, and they fear that their interests
will be squeezed out by these two dominant parties. At a “micro level” residents are aware of
(even as many may participate in) the widespread buying and selling of community names
perpetrated by residents, outside speculators and developers. Individual families fear that they
will receive fewer rights than other families as well as outside speculators, and even risk losing
their legitimate rights to new housing entirely through the practices of fraud.
Dissatisfaction with a Broad Range of Parties
Whereas in the Lyons and Lowery model the citizen’s dissatisfaction is directed
principally at the local government, which is responsible for providing the “tax/service package”
in his jurisdiction, in the Phnom Penh land sharing settlements dissatisfaction is directed more
broadly at all parties with an interest in redevelopment of the slum. Slum dwellers do not always
have a clear understanding of who those parties are, but in the public perception, three parties are
most closely associated with, and complicit in, the redevelopment plans: the “authorities”, private
developers, and community leaders.
The “authorities” comprise three levels of government: national-level governmental
interests, including central government Ministries and the Prime Minister; the Municipality of
Phnom Penh; and local authorities, including commune and district-level government. Private
developers are perceived to be the main authors and financiers behind redevelopment in the city,
together with high-level central government politicians, and assisted by municipal, district and

269

commune officials. At the ground level, some community leaders are perceived to be in the
pocket of the developers, and therefore these leaders are not universally trusted. In some
settlements, residents also perceive certain representatives of community-based organizations to
be acting as informal representatives of the developers. Dissatisfaction is therefore aimed at a
large and varied group of people, even if it is rarely vented directly.
The dissatisfaction of most residents did not manifest itself immediately. Initially,
following the Prime Minister’s upgrading announcement and the announcement of the land
sharing pilot projects in May 2003, most residents felt relief and satisfaction with the prospect of
obtaining free new housing, legal tenure, and official recognition as newfound “citizens”. But the
initial “honeymoon period” did not last long: by the end of 2003 and early 2004 the problems
facing the land sharing schemes become clear. It was during this second period that large-scale
(but often silent) dissatisfaction with the land sharing projects made itself known.
EVLN as Household Based Responses
The Lyons and Lowery model focuses on the individual as the main unit of analysis. This
individual is a resident of a local government jurisdiction, and is engaged in various degrees of
political participation. In the EVLN model applied to the slum I will focus on the individual
household as my main unit of analysis, as the household is the most common decision-making
unit in Cambodia, and therefore the most appropriate form of agency at the micro level. Such an
approach, focused on individual households rather than the amorphous notion of “community” is
consistent with the individual-level propositions advanced by Hirschman and Tiebout, and is also
consistent with the Foucauldian skepticism of “community”. I argue that the focus on individual
households also reflects a reality in Phnom Penh slums, where slum populations are quite diverse
in terms of their residents’ socio-economic status and are therefore likely to have diverse interests

270

and display a diverse range of different responses to dissatisfaction that cannot be captured by an
analysis of slum dwellers as an imagined collective.
5.3 Presenting an Adapted EVLN Model
The EVLN model, as structured and tested by Lyons and Lowery (1986 and 1989),
consists of three components: individual responses to dissatisfaction (various forms of exit, voice,
loyalty and neglect); two dimensions along which the individual responses can be arrayed; and a
set of conditions hypothesized to affect the likelihood of an individual engaging in behaviors or
attitudes associated with the responses to dissatisfaction (Lyons and Lowery 1989, 842).
In this section I will describe the original components of the EVLN model of Lyons and
Lowery, and I will explain how I adapt the responses to dissatisfaction of exit, voice, loyalty and
neglect to act as indicators of empowerment/disempowerment and active citizenship/passive
citizenship in the context of informal settlements in Phnom Penh. I propose two adaptations to
the original EVLN model of Lyons and Lowery: I replace their exit-privatization response type
with “symbolic exit”, and I propose an expanded sequential choice model to incorporate two
forms of loyalty and satisfaction (detailed in section 5.4.2).
5.3.1 Two Dimensions of Empowerment and Citizenship
In the Lyons and Lowery model the responses of exit, voice, loyalty and neglect are
arrayed along “active-passive” and “constructive-destructive” dimensions of behaviors regarding
local political participation as a response to dissatisfaction (Lyons and Lowery 1986, 324).
The active-passive dimension indicates whether or not residents act on their
dissatisfaction with their jurisdictions by taking some form of initiative. Active behavior (voice
and exit) indicates personal initiative in an effort to change the current situation, whereas passive
behavior (loyalty and neglect) signals the absence of such initiative. The constructive-destructive

271

dimension refers to the level of commitment of residents towards their jurisdictions.
Constructiveness (voice and loyalty) indicates a desire to remain in the local government
jurisdiction and engage in some way with the local government in response to dissatisfaction,
while destructiveness (exit and neglect) indicates a desire to break off all ties with the jurisdiction
and the local government. Figure 19 illustrates the two dimensions, response types, and
prototypical behaviors associated with each response type, as suggested by Lyons and Lowery for
dissatisfied residents in U.S. local government jurisdictions.
Constructive / Participatory
Destructive / Non-participatory
Active / Empowered
Passive / Disempowered
NEGLECT LOYALTY
VOICE EXIT
•Contacting officials
•Discussing political issues
•Campaign work
•Campaign contributions
•Participation in
neighborhood groups or
demonstrations
•Leaving or contemplating
leaving the jurisdiction (on
permanent or part-time
basis)
•Voting
•Speaking well of the
community
•Showing support for the
community by attending
public functions
•Non-voting
•Feeling that fighting city
hall has no impact
•Distrust of city officials

Figure 19: Prototypical response types in the original EVLN model (Lyons and Lowery 1986
and 1989)

In my EVLN model adapted to slum dwellers’ quest for citizenship in Phnom Penh, the
constructive-destructive dimension of behaviors serves as a measure of “good” citizenship (on
one end of the axis) and absent citizenship (on the other end), because on the side of
constructiveness it describes the behaviors of residents who participate in the activities and affairs
of their slum jurisdiction (voice and loyalty), whereas on the other end it describes the behavior

272

of people who refuse to participate because for various reasons they feel they have no stake in the
slum’s future (exit and neglect).
The active-passive dimension of behaviors provides me with a measure of empowerment,
in the sense that the two active responses to dissatisfaction (voice and exit) represent the agency
of residents who act from a position of strength because, in the Lyons and Lowery EVLN model,
they have good alternatives outside the jurisdiction (see section 5.4.1). Conversely, the two
passive responses (loyalty and neglect) are behaviors of people whose bargaining position is
weak, because they do not have better alternatives outside the jurisdiction. But voice and exit
represent two very different expressions of power. Voicers seek to improve conditions not only
for the individual but also for the community as a whole, whereas exit is motivated solely by
individual interest. Thus, the two forms of power have vastly different implications for
communities as a whole: successful voice leaves the community better off, through collective
action in the interest of all residents, whereas successful exit impoverishes the community by
leading to a loss of members.
Thus, in my adapted EVLN model, the highest ideal of “active citizenship” is represented
through the expression of collective power through voice. The “passive citizenship” of
individuals who participate but who lack the collective power to enforce their claims is
represented by loyalty. The “threat to citizenship” of empowered individuals who act solely in
their own interest, to the detriment of the community as a whole, is represented by exit. The
“absent citizenship” of residents who have no stake in the future of the slum community because
they have nothing more to lose, is represented by neglect. My adaptation of the Lyons and
Lowery EVLN model is illustrated in Figure 20. The four EVLN response types in my adapted
model are described in further detail in the next section.

273

Constructive / Participatory
Destructive / Non-participatory
Active / Empowered
Passive / Disempowered
NEGLECT LOYALTY
VOICE EXIT
•Active citizenship
•Individual or
collective power
•Threat to
citizenship
•Individual power
•Passive citizenship
•Collective lack of
power
•Non-citizenship
•Individual lack of
power

Figure 20: EVLN responses in my adapted Lyons and Lowery model

5.3.2 Voice: the Act of Articulation
When an individual resorts to voice, he makes an attempt to change the practices, policies
and outputs of the firm from which he buys, or the organization of which he is a member.
Hirschman defines voice as any attempt to change an objectionable state of affairs in an attempt
to improve quality or service, rather than simply going over to the competition, as in exit
(Hirschman 1970, 30). There are several ways in which customers or members can “kick up a
fuss”:
through individual or collective petition to the management directly in charge;
through appeal to a higher authority with the intention of forcing a change in
management; or through various types of actions and protests, including those
that are meant to mobilize public opinion (Hirschman 1970, 30).

274

Hirschman identified voice as “political action par excellence” (Hirschman 1970, 16).
By its very nature, voice is associated with seeking to achieve greater political clout, including
access to decision makers, and seeking to obtain a stake in the decision-making process.
There is a certain degree of inefficiency in the workings of the voice mechanism, just as
there is in exit. Just as the link between dissatisfaction, exit and quality restoration is not always
so efficient in less-than-perfectly competitive markets, so it is in politics. There is a theoretical
assumption that a democratic system requires a “maximally alert, active, and vocal public”, but in
practice there is considerable political apathy among large segments of the public. Hirschman
concludes that, as in the case of exit, a mixture of alert and inert citizens may serve democracy
better than either total, permanent activism or total apathy. Voice has the function of alerting a
firm of organization to its failings, but it must then give management, old or new, some time to
respond to the pressures that have been brought to bear on it” (Hirschman 1970, 32-3).
Voice is costly compared to exit, in terms of time and effort, and more likely to fail. Its
success is conditioned on the influence and bargaining power customers and members can bring
to bear from within their firm or organization (Hirschman 1970, 40). Moreover, as with exit,
there are equity concerns: the ability to use voice may not be equally distributed among
consumers or members. Evidence from political science shows that willingness to engage in
political activities or civic activities is associated with education, income or age (Klein, 426).
In the context of the redevelopment struggle in the slum, I suggest that voicers aim to
improve conditions in their settlement by means of different strategies of direct action, ranging
from active community organization to protest. More broadly, voice is associated with seeking to
achieve greater political clout, including access to decision makers, and seeking to obtain a stake
in the decision-making process. As with exit, voice is empowering because voicers aim to
improve their situation, but in contrast to exit, voicers seek to make all residents better off, not

275

only themselves and their own household. In contrast to exit, voicers are motivated by prior
satisfaction with their settlement or/and their community. For Hirschman, voice is far more
complex and unpredictable than exit, but its merit is that it transcends the self-regarding instinct
of exit. While prompted initially by the concerns of the individual, it can become an “enjoyable,
exhilarating experience when it is also in the public interest” (Hirschman 1976, 387).
Voice as “Active Citizenship”
Voice is frequently held up as an active and “ideal” form of citizenship because it
represents political commitment and collective action, although it can also be expressed in
individual forms. The identification of voice with active citizenship has its roots in Hirschman’s
original exit, voice and loyalty typology. Hirschman referred to voice as “interest articulation”—
a “basic portion and function of any political system” (Hirschman 1970, 30-1). In the expanded
EVLN model, Lyons and Lowery claimed that “much of what traditionally falls under the topic of
political participation can be identified as ‘voice’” (Lyons and Lowery 1986, 332).
In the development literature, voice is presented as a critical form of citizenship for the
poor as it is linked to many other sorts of resources that improve their agency and subjective
power. One such resource is the “capacity to aspire”, conceived by Arjun Appadurai (2004) as a
cultural resource—a “navigational capacity” to “explore the future more frequently and more
realistically, and to share this knowledge with one another…routinely” (Appadurai 2004, 69).
For Appadurai, the posture of voice carries with it the “capacity to debate, contest, inquire and
participate critically”, which is at the heart of the capacity to aspire. It is also the only way in
which the poor might find “locally plausible ways to alter…the terms of recognition in any
particular cultural regime” (Appadurai 2004, 66-70).

276

For Amartya Sen (2000) voice is linked with the concepts of “capability” and “freedom”,
which refer to the way in which human beings can engage their own futures and achieve various
lifestyles, and which are based on the “agency role of the individual as a member of the public
and as a participant in economic, social, and political actions” (Sen 2000, 19, 75).
In the Cambodian context, Kim Sedara and Joakim Öjendal (2007) claim that voice has
become the instrument for civil society’s enhanced demands on the local state (Kim and Öjendal,
6). One of the manifestations of this voice in present-day Phnom Penh takes the form of what
Caroline Hughes calls “urban protest movements” of the landless poor, market traders and trade
union members, who seek to “press their own, primarily economic, grievances onto an elite
political agenda” (Hughes 2003, 212).
5.3.3 Exit: the Act of Desertion
In Hirschman’s original framework, exit resembles the classic market response to a
situation of quality decline. In a competitive market situation where a firm has competitors, if
consumers are dissatisfied with a product they will stop buying it and switch to another product
that gives them greater satisfaction (Hirschman 1970, 15). For Lyons and Lowery, residents’ exit
response ruptures the relationship with the local government that is the source of their
dissatisfaction (Lyons and Lowery 1986, 332). For Hirschman, exit is held to be the most
powerful response to decline because it is, in principle, so straightforward: it ends the relationship
between customer and management, or between member and organization. Exit is the “act of
desertion” (Hirschman 1970, 31): customers and members do not make an effort to repair the
cause of the dissatisfaction themselves; they leave that to management.
In a perfectly competitive market, exit is considered one of the principal mechanisms of
the free enterprise system: by exiting, dissatisfied customers or members inflict revenue losses on
delinquent management, which leads management to undertake to repair its failings, and restore

277

quality. This mechanism works as a “discontinuous three-value function”: no corrective action
occurs with a small drop in revenue; full recovery follows upon a drop of “intermediate size”; but
if the revenue decline exceeds a certain large percentage of normal sales volume, the losses will
weaken a firm so badly bankruptcy will occur before any remedial measures can take effect
(Hirschman 1970, 23-4)
253
.
In the EVLN model, residents rupture their relationship with the local government by
physically leaving a local government jurisdiction (which Lyons and Lowery label “physical exit”
or “exit-move”). Exit resembles the behavior of “consumer voters” who, as described by Charles
Tiebout (1956), will “vote with their feet” when dissatisfaction arises with poorly performing
local government services and move to another jurisdiction with a better tax service package.
Residents who opt for exit-move generally have few ties that bind them to their current
jurisdiction.
Exit as “Anti-Citizenship”
In contrast to the “ideal citizenship” of voice, exit is typically identified as a “threat to
citizenship” (Ådnanes, 795) because exiters are guided by self-interest at the expense of the
broader community interest, and because they have broken off all ties with the local government
and the slum community.
One problem for citizenship is that exit may not lead to correction of the problems that
give rise to dissatisfaction. In a less-than-perfectly competitive market, exit does not necessarily
lead to effective “quality recuperation”, to use Hirschman’s terms (Hirschman 1970, 25). Exit is

253
The actual process is more complex than this: Hirschman points out that for exit to work as a
mechanism of recuperation from lapses in quality, it is “generally best for a firm to have a mixture of alert
and inert customers. The alert customers provide the firm with a feedback mechanism which starts the
effort at recuperation, while the inert customers provide it with the time and dollar cushion needed for this
effort to come to fruition” (Hirschman 1970, 24).

278

poor in information: in complex services, it provides little or no guidance for management as to
the nature of the dissatisfaction (Klein, 421), thus making recuperation of the quality of the
product or service difficult. And there is an equity factor: there is an assumption that exit has a
built-in bias towards the better off, who can afford to “buy out” (Klein, 426).
Another problem is that exit inhibits voice. Hirschman suggested that rapid exit of highly
quality-conscious customers can “paralyze voice by depriving it of its principal agents”
(Hirschman 1970, 51). Jeffrey Herbst (1990) described this phenomenon in an African context,
when he argued that in pre-colonial Africa migration was the traditional manifestation of
discontent with a political community. This “ultimate form of exit from a political community”
has long acted to inhibit organized political resistance activity. Conversely, the closing off of the
migration option in many African states in the modern era has increased the threat of violent
popular protest against African governments, as would-be migrants now flock to the cities in their
own countries instead of leaving the polity altogether (Herbst, 186, 199). For all these reasons I
suggest that exit represents “anti-citizenship” in that it undermines the conditions for the
emergence of an active citizenship.
5.3.4 Loyalty: the Act of Keeping Faith
The classic definition of loyalty, as “passive support of the system” (Lyons and Lowery
1986, 332) indicates two of the main characteristics of loyalists in the Lyons and Lowery
model—their cooperative behavior vis-à-vis local government, and their reticence to engage in
direct political action. In the slum environment I suggest that loyalists participate in community
activities, such as community meetings and savings schemes, and in general, follow their
community leaders.
In their respecified EVLN model Lyons and Lowery distinguished between “two very
different forms of loyalty”. The first is a “ritual loyalty” of satisfied residents. This is a default

279

form of behavior separate from the four responses to dissatisfaction (Lyons and Lowery 1989,
853). The second form is “loyalty as a response to dissatisfaction”, which is the behavior of
individuals who are dissatisfied with conditions of decline in their jurisdiction, but whose lack of
better alternatives causes them to passively and hopefully wait for conditions to improve (Lyons
and Lowery 1989, 860). Both varieties of loyal behavior are found in the slum environment.
“Ritual Loyalty” as Default Condition of the Model
Ritual loyalty is an indicator describing the passive support of satisfied residents. For
Lyons and Lowery ritual loyalty is the initial condition of the model—a default situation before
dissatisfaction sets in, a point when relative dissatisfaction with local government is still low.
In the four land sharing settlements I assume that—as in the Lyons and Lowery model—
ritual loyalty is the default situation for a majority of the residents in an early phase of the land
sharing process. At this stage, most residents are still satisfied with the land sharing plans in their
settlements, for three main reasons. First, I have argued that the projects fit within a long
tradition of patronage in Cambodia (see Chapter 3, section 3.4.3), which rewards ordinary people
with benefits in exchange for their political support. Second, the slum upgrading campaign in
Phnom Penh was announced by Prime Minister Hun Sen, so residents had reason to believe that
the land sharing projects would materialize. And third, the land sharing schemes promised
substantial and unprecedented benefits for registered residents, including free new housing on
site, and tenure security. At an early stage, before land values appreciated significantly in the
four land sharing settlements, most residents would not have been able to obtain a better package
of benefits elsewhere through their own initiative.

280

Loyalty as a Response to Dissatisfaction
In the face of (growing) dissatisfaction, loyalists passively but optimistically wait for
conditions in their jurisdiction to improve, and in the meantime they cooperate with the local
authorities. Loyalty as a response to dissatisfaction is generally adopted by residents who have a
high degree of prior satisfaction with their package of goods and services, as well as a high level
of investment, but few good alternatives outside the jurisdiction. In Hirschman’s typology, voice
is not an attractive option for loyalists because “the availability of exit is the threat necessary to
make voice credible” (Hirschman 1970)
254
. Without the threat of exit, therefore, loyalists would
have little credibility were they to raise their voice.
Hirschman offered several suggestions as to why loyalists remain members or customers
despite their dissatisfaction with a declining service or product. First, they might not be able to
exit, or they might dare to exit, particularly if there is a form of penalty imposed against exit; this
penalty might be directly imposed, or internalized. Second, it could be that some members do not
want to leave: particularly in the case of public goods and organizations, loyalists—especially the
more influential members—might be driven by a genuine commitment to their organizations, and
they might fear that if they left, their organizations would go from bad to worse (Hirschman 1970,
98-9).
Loyalty is a response Hirschman describes as “less rational” than exit or voice, “though
far from wholly irrational” (Hirschman 1970, 38). Hirschman recognized that remaining loyal
has a number of psychological benefits for loyalists. The reason loyalty is not irrational is because
it is based on “an enormous dose of reasoned calculation” (Hirschman 1970, 79): loyalists have

254
Hirschman’s theory of loyalty remained somewhat ill-defined, as loyalty was treated as a residual
category for many unexplained, passive behaviors (Bajoit, 328). This problem has been rectified somewhat
with the introduction, more than a decade later, of the neglect behavior, which has filled in much of the
void as far as “explaining” apathy.

281

the expectation that “someone will act or something will happen to improve matters” (Hirschman
1970, 78). This expectation makes enduring the situation of dissatisfaction bearable. Loyalty
also has benefits for the organization as a whole because it has a stabilizing function: it saves
firms and organizations from the dangers of excessive or premature exit, and strengthens voice
(Hirschman 1970, 92).
In the land sharing environment, I argue that loyalists will patiently wait, hoping that the
“system” in which they place their faith (comprising the Municipality, local authorities, the
developers and the community leaders) will eventually deliver more optimal living conditions for
them. The expression of loyalty, therefore, is a passive expression of good faith that deteriorating
conditions in the slum will improve. In the best case, loyalists will be rewarded for their loyalty
to the slum redevelopment projects by being first in line for the new housing promised to them by
the authorities. In the worst case, Hirschman suggests that loyalists can be trapped by their own
good faith:
The ultimate in unhappiness and paradoxical loyalist behavior occurs when the
public evil produced by the organization promises to accelerate or to reach some
intolerable level as the organization deteriorates; then…the decision to exit will
become ever more difficult the longer one fails to exit. The conviction that one
has to stay on to prevent the worst grows stronger all the time (Hirschman 1970,
103).
Loyalty as “Passive Citizenship”
If voice is identified with active citizenship, then loyalty’s characteristics—a “passive
support of the system with little direct action to redress the problem giving rise to dissatisfaction”
(Lyons and Lowery 1986, 332)—have commonly been associated with passive citizenship. Arjun
Appadurai suggests that loyalty signifies “consensus” as opposed to the “dissensus” of voice
(Appadurai 2004, 65).

282

In the Cambodian context, I argue that the passive citizenship of loyalists resembles the
behavior of people who have few other options except to cooperate with dominant classes in
order to maximize their own security. Traditionally, poor Cambodians have achieved this
through involvement in patronage structures, as suggested by Jan Ovesen, Ing-Britt Trankell and
Joakim Öjendal (Ovesen, Trankell and Öjendal, 86). In stereotypes and myths about ordinary
Cambodians, the passive citizenship of loyalists has long been celebrated as stereotypical
behavior of poorer Cambodians, and even as a social good. In his writings about the poor (his
“little people”), Norodom Sihanouk speaks of the traditional loyalty of the poor in glowing terms
and bemoans its presumed disappearance with development:
I have noted with dismay that the more Cambodians rise in the social hierarchy,
the less they cultivate the principle of loyalty. The poorest Khmers elevate
themselves by refusing to become weathercocks, turning with any wind. They
deliberately turn away from any opportunism (Jeldres, 41).
Passive citizenship can ultimately be detrimental to loyalists: according to Appadurai, the
consensus of loyalists can take the form of “compliance” and false consciousness in the Marxist
sense—“not mere surface compliance, but fairly deep moral attachment to norms and beliefs that
directly support their own degradation” (Appadurai 2004, 65).
5.3.5 Neglect: the Act of Withdrawal
Of the four EVLN responses, neglect is perhaps the most mysterious and the least
understood, partly because it is a behavior not included in Hirschman’s original typology, and
partly because of the very nature of neglect itself—a passive behavior by people who withdraw
from the social and organizational life of a polity and are therefore relatively more hidden from
public view. Lyons and Lowery describe neglect as the standard political behavior of

283

disadvantaged citizens of the urban core, for whom “exit is not possible and voice is unlikely to
be effective” because their views are likely to be treated as “non-issues”
255
.
Lyons and Lowery (1986) added the neglect response to the original Hirschman
framework in order to expand the range of passive behaviors that members or customers might
deploy in the face of dissatisfaction. The result is an additional passive behavior that is much
closer to apathy than loyalty. For Lyons and Lowery neglect is a passive behavior characterized
by a withdrawal from participation in local government and politics. Neglect behavior differs
from loyalty in that it is not directed at recovery of the relationship with local government
(Withey and Cooper, 522): citizens displaying neglect behaviors have little expectation that
conditions will improve, and therefore they do not consider it worthwhile to engage in activism.
As in the case of loyalists, neglecters have no good alternatives outside the settlement, and
therefore no exit option either.
In the land sharing context, I hypothesize that neglect is a behavior that might be
commonly found among those in the slum that have no stake in the land sharing process—in the
four Phnom Penh cases these include those residents not eligible for new housing, such as renters
and the poorest of the poor, who may have already sold their rights to new housing early on in the
process because of financial necessity, but who have nowhere else to go. Neglect is perhaps the
worst condition possible for slum dwellers. Neglecters have no stake in the future of the slum
redevelopment process because they are not invested in the slum. In contrast to loyalists, these
residents have little or no prior satisfaction from—and commitment to—the slum, yet at the same
time they have nowhere else to go as they lack alternatives outside the slum. Citizens displaying

255
P. Bachrach and M.S. Baratz, Power and Poverty: Theory and Practice (New York: Oxford University
Press, 1970), quoted in William E. Lyons and David Lowery, “The Organization of Political Space and
Citizen Responses to Dissatisfaction in Urban Communities: an Integrative Model”, The Journal of Politics
48, no. 2 (1986): 336.

284

neglect behaviors have little expectation that things will improve. The result is a mix of
behaviors that are destructive, because they are not oriented toward activism or turning things
around. These behaviors may include lack of participation in community structures, mistrust of
local authorities or any other representative of the local polity, and silent frustration and
resentment. Apathy is their main coping strategy, and staying out of public view their main
survival strategy. Because of their physical and symbolic withdrawal from the life of the slum, I
suggest that neglecters are the “invisible people” of the slum.
Neglect as “Non-Citizenship”
I will categorize neglect as a second form of “non-citizenship”, alongside exit, because,
just like exiters, neglecters have broken off all ties with the local government and the slum
community, and are not oriented towards addressing the sources of dissatisfaction governing their
lives. Unlike exiters, however, most neglecters do not withdraw from participation in local
government and politics by choice. Nevertheless, their apathy undermines efforts to build an
active citizenship in the slum community. Figure 21 illustrates the EVLN categories in my
adapted EVLN model.

285

Constructive / Participatory
Destructive / Non-participatory
Active / Empowered
Passive / Disempowered
NEGLECT LOYALTY
•Keeping faith in successful
project outcome
•Passive participation in
community activities
•Following community leaders
•Consensus behavior
•Withdrawal
•Lack of participation in
community structures
•Silent resentment
•Mistrust of authorities &
community leaders
VOICE
•Articulation of interests
•Active participation in
community activities
•Community organization
•Protest
EXIT
•Physically leaving the
slum (on permanent or
part-time basis)
•Actively making plans to
physically leave the slum

Figure 21: Selected anticipated EVLN responses by slum dwellers in Phnom Penh

5.4 Theoretical Propositions
The Lyons and Lowery EVLN model provides me with a number of theoretical
propositions with which to identify citizenship and empowerment behavior among slum dwellers,
through the proxy of the four EVLN responses.
5.4.1 Determinants of EVLN Behavior
In the respecified Lyons and Lowery model (1989), three determinants are advanced as
independent variables for individual EVLN behaviors of dissatisfied residents: relative
dissatisfaction, investment level, and the availability of good alternatives.

286

Relative Dissatisfaction
The EVLN model is designed to measure responses to dissatisfaction among residents of
urban jurisdictions, but it leaves open the possibility that not all residents are dissatisfied, even in
the face of mounting problems. As Lyons and Lowery remark, “individuals must perceive
problems or be dissatisfied with the current state of affairs before they are likely to expend the
time, energy, and resources to engage in…responses to dissatisfaction” (Lyons and Lowery 1989,
854). To deal with this eventuality, Lyons and Lowery created the variable of relative
dissatisfaction
256
. Lyons and Lowery conceded that it may be difficult to establish a baseline
from which the intensification effects of current dissatisfaction can be measured (Lyons and
Lowery 1989, 845). In my adapted EVLN model, I dealt with this problem by creating a relative
dissatisfaction variable that gauged respondents’ feelings about current services and conditions in
the slum, compared to “before”. In the case of current residents of the slum, the baseline period
“before” referred to the situation in the settlement before the social land concession
announcement by the Prime Minister in May 2003. In the case of former residents of the slum,
the baseline period “before” referred to the time when the respondents still lived in their former
settlement, before they moved out.
Investment Level
The variable of investment level measures the extent to which residents have significant
tangible and intangible (socio-psychological) investments in their settlements. This variable
works in a similar way to relative dissatisfaction. Residents who are highly invested in their
jurisdictions should in theory have more to lose from instability in the settlement than individuals

256
Relative dissatisfaction was added as a determinant in the respecified EVLN model of Lyons and
Lowery (1989); it was not included in the original (1986) model.

287

with fewer investments. As a result, they should be more likely to adopt constructive behaviors,
either to try and address the source of dissatisfaction (through voice), or to maintain their support
of the local government (through loyalty), in the hope that this will help to stem the source of
decline. Those with fewer investments should feel little commitment: if they have alternatives
elsewhere they would be tempted to exit; if they have no alternatives they would allow the
situation to decay further (Lyons and Lowery 1986, 333).
In the U.S. context Lyons and Lowery cited important tangible investments such as
homeownership and children enrolled in local public schools, and intangible investments
including long term residence and attachment to the area, for various reasons. In the Phnom Penh
slum context similar tangible and intangible investments play a large role. The “ownership” of a
home and a plot of land (whether it was negotiated through an informal or formal transaction) is
likely to be the most important investment for most families in informal settlements in Phnom
Penh. Other significant tangible investments can be the financial participation in community
savings schemes and investments in a small business (especially if the business is not mobile and
based in the settlement). Intangible investments include length of residence, participation in
community structures (community based organizations, women’s groups, savings groups, etc.), as
well as a host of other investments of psychological value to the individual and their family.
Good Alternatives
For Lyons and Lowery a good alternative is defined as the availability of another
jurisdiction that offers a tax-service package or package of lifestyle maintaining characteristics
that is closer to the dissatisfied resident’s optimal desired level (Lyons and Lowery 1986, 333).
Dissatisfied residents who have good alternatives should be more likely to adopt active behaviors,
i.e., voice or exit, to address their dissatisfaction, while those with poor alternatives would be

288

expected to adopt passive responses, as they cannot afford to “kick up a fuss”, through voice, or
move elsewhere, through exit.
In the land sharing context in Phnom Penh, most respondents who were registered as
community members (and who therefore qualified for free housing as part of the land sharing
program) defined a “good alternative” as a place that is affordable, which offers better tenure
security than their current location, which has access to formal public services, and which is
based on an official “plan”
257
. In addition, the alternative location must not be too far away from
their present location, or have access to other employment opportunities. For those residents not
registered as community members (and with no right to free replacement housing) the concept of
a good alternative was defined more pragmatically, as they enjoyed fewer options. In their case,
a “good alternative” constituted any other settlement that was affordable, and where they would
be able to stay, even if only temporarily.
5.4.2 Decision Tree Model of Responses to Dissatisfaction
The hypothesized relationships in my adapted EVLN model, as described in the previous
section, can be summarized graphically in a decision tree depicting responses to dissatisfaction in
the land sharing slum under redevelopment (Figure 22).

257
In the Phnom Penh context, slum dwellers frequently refer to “formal” (non-slum) areas as areas that
have a “plan” (plang in Khmer), as distinct from urban poor settlements (sahakum neak kraikraw), which
are unplanned because they grow organically as residents arrive and build their own structures and install
their own services, over time.

289

Relative Dissatisfaction
Low High
“Ritual
Loyalty”
(Satisfied)
Investment Level
High Low
Good Alternatives Good Alternatives
Voice Loyalty
Yes No
Current Residents
Exit Neglect
Yes No

Figure 22: Decision tree model of responses to dissatisfaction among slum dwellers
(Model adapted from Lyons and Lowery, 1989)

My decision tree model differs from that of Lyons and Lowery (1989) in that it
incorporates two forms of loyalty (ritual loyalty and loyalty as a response to dissatisfaction),
whereas the Lyons and Lowery model incorporates only one (ritual loyalty)
258
. The decision to
include both forms of loyalty is based on the need to distinguish between these two very different,
and very important, behaviors in the Phnom Phnom slum context.

258
In my adapted decision tree, loyalty as a response to dissatisfaction replaces the “privatization” response
of the Lyons and Lowery (1989) decision tree model. According to Lyons and Lowery, the inclusion of
loyalty as a response to dissatisfaction in the model (instead of privatization) is “theoretically viable”,
(Lyons and Lowery 1989, 863).

290

The decision tree model identifies the EVLN responses of residents to conditions in their
living environment and the slum redevelopment process based on three different behaviors and
attributes, corresponding with the determinants of the model. The first important distinction in
the model is between residents who are dissatisfied with conditions in their settlements are those
who are “ritually loyal”, i.e. those who are wholly satisfied with all aspects of their lives in the
settlement. Residents who are ritually loyal are removed from the model as, by definition, they
cannot deploy responses to dissatisfaction (the basis of the EVLN model) if they are not
dissatisfied in the first place. Residents who demonstrate some degree of dissatisfaction with any
aspect of their lives in their settlement are retained in the model.
In the second stage, the model distinguishes between dissatisfied residents who are well-
invested in their settlements and those who are not. In the third stage, the model makes two
additional distinctions, on the basis of good alternatives. On the one hand it distinguishes
between dissatisfied and poorly invested residents who have good alternatives outside of the slum
versus those that do not; the former are identified as exiters, while the latter are identified as
neglecters. On the other hand the model distinguishes between dissatisfied, highly invested
residents who have good alternatives elsewhere versus those that do not; the former are identified
as voicers, and the latter are identified as loyalists.
5.4.3 Four Propositions about Residents’ EVLN Behavior
The sequence of responses illustrated in the decision tree model provides me with four
basic propositions about slum dwellers’ EVLN behavior. According to the respecified model of
Lyons and Lowery, all four responses are characterized by relative dissatisfaction; what sets the
response categories apart are “varying patterns of investments and alternatives” (Lyons and
Lowery 1989, 864). Table 5 summarizes the determinants and their hypothesized relationship to
the EVLN behaviors.

291

• Exit: Dissatisfied residents who are not well invested in their settlement and who have
good alternatives outside the settlement, will physically leave—or make plans to
physically leave—the settlement on a full or part-time basis.
• Voice: Dissatisfied residents who have good alternatives elsewhere, but who are well
invested in their settlement, will use voice to try to improve conditions in their settlement.
• Loyalty
259
: Dissatisfied residents who are well invested in their settlement, but who have
few good alternatives elsewhere, will remain loyal to the authorities and will display little
(outward) dissatisfaction; they will wait patiently for redevelopment prospects in the
slum to improve.
• Neglect: Dissatisfied residents who are not well invested in the settlement, and who also
have no good alternatives outside the settlement, will withdraw from any organized
community life and display neglect behavior, for lack of better options.

Determinants of response Responses to dissatisfaction
Exit

Voice Loyalty* Neglect
Relative dissatisfaction + + ─ +
Investments ─ + + ─
Good alternatives + + ─ ─
+ indicates positive relationship; ─ indicates negative relationship
*Loyalty refers to “loyalty as a response to dissatisfaction”.
Table 5: Hypothesized relationships between determinants responses to dissatisfaction
(Adapted from the Lyons and Lowery EVLN model, 1989)

259
This refers to loyalty as a response to dissatisfaction, as opposed to ritual loyalty.

292

In the second part of this chapter I will describe the data collection methods and
techniques I utilized for operationalizing these propositions during the structured interviews.
5.5 Data Collection
5.5.1 Research Timeframe
There were two distinct phases in the data collection for this study. During the first
phase, in December 2002 and from May to July 2003, I conducted a background study of past
land sharing settlements in Bangkok, Thailand. Together with a small team of research assistants
affiliated with Chulalongkorn University in Bangkok, I documented the land sharing experience
in seven inner-city settlements where land sharing was first instituted during the 1980s and early
1990s, through a series of site visits and interviews with community leaders and other key
informants involved with the process at the time. The research phase in Bangkok enabled me to
acquire a first-hand understanding of the mechanisms and impact of land sharing in the city
widely acknowledged to be the birthplace of organized land sharing as a technique of slum
upgrading and redevelopment.
The second research phase took place in Phnom Penh, Cambodia, from September 2003
to January 2009. This period covers the start of official planning for the four land sharing
projects (in September 2003) and the point at which (in January 2009) the final outcome of the
redevelopment processes in the four settlements became clear for all the residents concerned
260
.
The research period in Phnom Penh covered several phases. Between September 2003 and April

260
The cut-off point of January 2009 for this research is somewhat arbitrary, as redevelopment of the four
settlements (including new construction of apartment buildings in Borei Keila, and the commercial
development in the original Dey Krahom and the Railway sites) is still ongoing. But I argue that 24 January
2009 is a significant milestone because it was on this day that the final eviction took place of the remaining
residents in the original Dey Krahom site. On this day, therefore, the final redevelopment outcome for the
residents of all four settlements became clear.

293

2004 I followed the four land sharing projects first-hand in my capacity as advisor to the
Governor of Phnom Penh on urban poor land and housing issues for the Phnom Penh Urban
Poverty Reduction Project under UN-Habitat. After completion of the PPUPR project in April
2004 I remained in Phnom Penh and followed developments in the four land sharing case study
settlements first-hand, as an independent researcher. I conducted regular interviews with key
informants involved with the slum redevelopment process from the four settlements, the
Municipality of Phnom Penh, local authorities, private developers, and civil society
organizations. Between January and April 2006, following approval of my research plan by the
University of Southern California
261
, I conducted interviews with current and former residents of
the four land sharing settlements, with the assistance of two senior local research associates.
From May 2006 to January 2009, I once again followed developments in the four land case study
settlements in my capacity as an independent researcher.
5.5.2 Data Collection Methods
The study employed a variety of data collection methods. Information about the EVLN
behaviors of residents in the case study settlements and their levels of prior satisfaction,
investment and availability of alternatives (the three hypothesized determinants of EVLN
responses in the Lyons and Lowery model) was obtained through in-depth, structured interviews
with a sample of heads of household or household decision-makers in the four settlements, using
a questionnaire. Information about the land sharing projects was supplemented by both direct and
participant observation by the author and the local research associates.
Additional information about the land sharing projects and wider development trends in
Phnom Penh and Cambodia was obtained through open-ended and focused interviews with key

261
The University of Southern California’s University Park Institutional Review Board formally approved
my research study on 13 December 2005 (USC UPIRB #UP-05-00241).

294

informants from central government, the Municipality of Phnom Penh, local authorities, private
developers, donor agencies and civil society organizations. The study also made use of official
documents from the Municipality of Phnom Penh, the Royal Government of Cambodia and donor
agencies, as well as archival records on historical events in Phnom Penh from the National
Archives of Cambodia. The author’s professional connection to the Municipality of Phnom Penh
in 2003 and 2004 and, subsequently, to the Ministry of Land Management, Urban Planning and
Construction in 2004 and 2005, facilitated access to information about the land sharing projects
(in Phnom Penh) and the Royal Government of Cambodia’s social land concession program
(nationwide).
5.5.3 Interviews with Residents
Between January and April 2006 in-depth, structured interviews using a questionnaire
were held with 183 current and former residents of the four case study settlements. Target
interviewees were either heads of household or decision-makers in their household. Heads of
household are generally considered to be male in Cambodia, while decision-makers can be both
women and men. Thus, aiming the research study explicitly at both categories enabled the
interviewers to reach men as well as women.
Current and Former Residents
Of the total of 183 interviews, 152 interviews were conducted with current residents of
the four case study settlements and 31 with former residents who had left one of the four case
study settlements and were now living elsewhere in Phnom Penh. Table 6 provides a breakdown
of the number of current and former residents interviewed in each settlement.

295

Category of resident Total Name of Settlement
Borei Keila Dey Krahom Railway A Railway B
Current residents 152 65 52 11 24
Former residents 31 6 10 4 11
Total respondents 183 71 62 15 35
Table 6: Number of former and current residents interviewed by settlement

Former residents could not be included in the main EVLN sample because, as they no
longer resided in one of the four settlements, their actions were not relevant for a determination of
EVLN behaviors as a response to dissatisfaction. Rather, the rationale for interviewing former
residents was to obtain a better background understanding of the motivations for their exit
behavior. Specifically, I sought to find out whether, in the Phnom Penh slum context, the exit of
these respondents resembled the empowering strategy of “voting with one’s feet” as posited by
Charles Tiebout (1956) and as adopted by Hirschman, Lyons and Lowery and others in the EVLN
literature. An important objective of the interviews with former residents was therefore to
establish the reason why they left their former settlements. Residents who left their settlement
were presumed Tiebout exiters, provided that they left the settlement of their own free will.
However, those families who left involuntarily, for example because they were evicted or because
a family emergency forced them to sell out so that they could get access to cash, are not acting
out of a position of power and therefore cannot be considered true exiters. The former residents
of greatest interest to the study were those who moved to other areas in Phnom Penh, for it is with
these people that a direct “market test” of their satisfaction with the old settlement can be
established. By contrast, residents who moved out of one of the four settlements to go to another
city or to the rural areas might be leaving for reasons other than dissatisfaction with the
settlement.

296

Even though current and former residents were both asked similar questions about their
prior satisfaction, investment level and availability of alternatives, slightly different
questionnaires were developed for both groups because questions for former residents needed to
refer to events in the past (their life in the old settlement), and because former residents were
asked to compare conditions in the old settlement with the present conditions in their new living
environment.
5.5.4 Questionnaire Design
The questionnaires for current and former residents were designed to solicit qualitative
information from respondents about their housing strategies and their lives in the slum (or former
lives in the slum) during the redevelopment process. The questionnaires consisted of open-ended
as well as multiple choice questions (see questionnaires in Appendices 2 and 3). Multiple choice
questions recorded basic statistical information about the household as well as “measurable”
categories of behaviors that could be used for subsequent statistical analysis. The open-ended
questions were designed to record more analytical responses, in language used by the
respondents. In order to minimize the possibility of bias, the questionnaires sought to obtain
information in several different ways—both directly and indirectly—in order to “test” and
confirm respondents’ answers.
Measuring Investment Level
Part 1 of the questionnaires measured respondents’ level of investment in their
settlements. Interviews started with this determinant because questions about objective tangible
and intangible investments were considered to be an easier way to “break the ice” in an interview
than questions about more subjective topics covered by the other two determinants (satisfaction
and good alternatives).

297

Households invest in their informal settlements in several different ways. Tangible
investments comprise financial and time investments. The most obvious form of time investment
made by a household is the number of years the household has lived in the settlement. In
addition, an important measure of time investment is whether the household lives in the
settlement on a full-time basis, or only part-time. If some household members live in the
settlement only part of the time, and somewhere else the rest of the time, this may indicate exit
behavior and the ready availability of alternatives. The biggest financial investment by a
household is usually the cost of the plot of land and house over the years (from purchase of land
and house through to subsequent renovations and expansions of the house). Other tangible items
for the household include: the financial and time investment to obtain legal documentation for the
land and house, including title deeds or receipts for application of title, if any; the financial and
time investment spent on community projects, such as infrastructure improvement projects, for
which residents typically contribute money and in-kind labor; and money contributed to the
community savings scheme. An indirect measure of investment has to do with the location of the
work that the respondent or the household members does: those people with jobs that are far away
from the settlement are considered to be more likely to consider living somewhere else, outside
the settlement, and are therefore less invested in the settlement.
In terms of intangible and psychological factors, the most important investment a
household makes is in the community. The degree of practical commitment to the community can
be measured in various ways, including whether or not household members are community
leaders or community members and are regularly involved in community organization, or
regularly participate in community groups and meetings. The higher the participation level, the
higher a respondent’s “jurisdiction-specific commitment” (Lyons and Lowery 1986, 333).
Another measure of a household’s investment in the community may be how much effort

298

household members make to keep informed about the development plans in the settlement.
Presumably, the more extensive the information network of the household is and the better
informed it is, the more it cares about development in the settlement, and thus, the greater its
investment level.
Respondents’ investment level is highly likely to change over time, as a result of both
external forces (including pressures from outside or perceived better alternatives outside of the
settlement) and internal forces (such as changing fortunes in the household). This in turn is likely
to lead to changes in EVLN behavior. To capture any changes in investment level over time,
respondents were asked whether their behavior had changed compared to “before” with respect to
several investment indicators, including joining community groups, attending community
meetings, actively trying to get information about development plans in the settlement; and
getting directly involved in activities to help the community. The time period “before” was
deliberately left unspecified to allow respondents to determine for themselves whether there was
an earlier period when their behavior or conditions in the settlement were different compared to
now.
Measuring Satisfaction Level
Part 2 of the questionnaires measured respondents’ current and prior satisfaction levels
with their lives in their settlement. As noted in section 5.2, the concepts of a “tax/service
package” and a “package of lifestyle maintaining characteristics” do not translate in exactly the
same way from the U.S. context to the environment of a developing country slum. In the informal
settlements of Phnom Penh, residents do not pay official taxes. On the other hand, the rates and
fees that residents pay for almost all services are higher than those paid by residents in the formal
neighborhoods of the city, as these services are delivered for the most part by private vendors.

299

Therefore it can be said that, unofficially at least, slum dwellers are paying a tax for living in an
informal area. Moreover, the package of lifestyle maintaining characteristics for slum dwellers is
likely to differ from that of suburban U.S. families, with access to basic services such as water,
health services and food markets forming a higher priority than those factors that typically drive
locational choice in the U.S., such as the quality of schools.
Nevertheless, the basic principle of a “calculus of making and maintaining a locational
choice” (Lyons and Lowery 1986, 323) is common to both contexts: just as suburban residents in
the U.S. are consumers of public goods who shop around for the optimal package of public goods
at the price that is right for them, so I argue that slum dwellers in Phnom Penh are highly mobile
and make decisions about where to live and how to manage their lives, based on similar practical
constraints. Many choose to live in a slum because its location offers convenient access to
employment and social facilities, at a price they can afford. In this respect, as John F. C. Turner
pointed out as far back as the 1960s, slums are well adapted to the needs and circumstances of
their residents (see Chapter 4, section 4.3.2).
In the interviews, respondents’ satisfaction with living in the settlement was assessed in
several different ways. Respondents’ overall satisfaction level with the settlement was evaluated
by asking first whether respondents were satisfied to live in the settlement in general. Next,
respondents’ satisfaction with different aspects of the settlement was assessed, to order to isolate
which specific factors were most important to understanding the satisfaction or dissatisfaction
level of the household. The specific factors included the physical location of the settlement; the
quality of services, including water, electricity, roads and footpaths, drainage, sanitation, health
services (nearby), and garbage collection; physical safety in the area; the quality of the
community and community support; and the legal status of the area and the household’s security
of tenure in the settlement.

300

As in the case of investments, respondents’ satisfaction is highly likely to change over
time, as a result of a combination of external forces and internal forces. To understand changes in
satisfaction level, current residents were asked to compare their current level of satisfaction with
their satisfaction “before” with regard to their overall satisfaction level and their analysis of the
situation with regard to physical safety; conflict resolution; their personal commitment toward the
community; and their perception of their tenure status. Here again, the time period “before” was
deliberately left unspecified to allow respondents to determine for themselves whether there was
an earlier period when their behavior or the situation in the settlement were different from now.
Measuring Good Alternatives
Part 3 of the questionnaires assessed whether the respondents have good alternatives to
living in their present settlement. In the case of former residents, the questionnaire sought to
confirm whether respondents considered their present living arrangement outside their former
slum as a “good alternative”.
In the interviews, the presence or absence of alternatives was assessed both directly and
indirectly. Direct assessment consisted of asking respondents whether they were “currently”
making plans to leave the settlement, or had made plans to leave the settlement “before”, on their
own initiative (and not as part of the planned relocation of the community, as in the case of the
Dey Krahom). In addition, respondents were asked whether they believed they had the possibility
to leave the settlement, either “currently” or “before”, should they want to leave. Possibility was
expressed both in terms of their financial ability (affordability) and in terms of another location
where they could live, i.e., the interviewers sought to find out whether respondents had another
house or plot of land available somewhere else.

301

In addition to direct questions about alternatives, it was also deemed necessary to
indirectly assess alternatives because of the likelihood that many respondents were unlikely to
provide an honest answer about the true extent of their alternatives. One of the most important
indirect measures of alternatives was thought to be income level and assets: the higher a
household’s income level and the greater its assets, the greater the possibility it can leave the
settlement and live somewhere else if need be. Therefore, the questionnaires sought to obtain a
“picture” of the household’s overall income level. The household’s wealth was assessed by
asking the respondents directly about their household’s income level, expenditure and assets,
including land and houses elsewhere; the variety of other income sources; and the number of
household members that earn an income. Assets such as land, cars, motorcycles and (to a lesser
extent) televisions can be sold for cash, and so they were considered as “income”.
Other indirect measures of good alternatives can be gleaned from the kind of work that
the respondent and his or her household does. Households with members who own small
businesses or who have employment in the formal sector, for example, can be expected to have
greater earnings and job stability, and therefore better alternatives than those who do not.
Finally, respondents were asked about their views of the development plan in the
settlement and their future plans. The aim of these (open-ended) questions was to assess whether
the respondent and his or her household felt strongly enough about the situation in their
settlement that they would consider leaving.
5.5.5 Sub-Groups of Respondents
As a result of the various waves of migration to Phnom Penh since the end of the DK
period, the four land sharing settlements (and most other slums in Phnom Penh) house residents
with a range of income levels and socio-economic backgrounds (see also section 6.2 for a general

302

profile of respondents). The result is that residents have diverse interests and display a range of
different responses to dissatisfaction.
In order to obtain a sample of respondents that was representative of all main socio-
economic backgrounds in the slums, I identified in advance seven principal “sub-groups” of
residents, defined as people who form easily identifiable categories within the four settlements,
and who were expected to share certain characteristics that make them more or less likely to
demonstrate certain EVLN behaviors as a result of group traits that are assumed to be different
from those of the majority of residents. The seven sub-groups were community leaders; renters;
families of army officers; families of police officers; ethnic Vietnamese residents; women-headed
families; and owners of small businesses in the community.
The sub-groups do not represent exclusive categories—respondents can belong to one or
more sub-groups at the same time. Only one sub-group (the ethnic Vietnamese minority)
describes a permanent status; all the other sub-groups reflect a group-based association or a socio-
economic status that may change over time. As such, respondents may belong to one or more
sub-groups at different stages in their life in the settlement. For the purpose of the interviews,
respondents’ sub-group labels reflect their current situation at the time of the interview (in the
case of “current residents”) or their situation at the time when they left the settlement (in the case
of “former residents”).
Table 7 lists each of the special population sub-groups and summarizes their assumed
group characteristics as well as their assumed EVLN behaviors.

303

Sub-group Assumed group traits Assumed EVLN response
Community leaders Personal commitment to the community;
Well-invested in the community;
Well-connected to the authorities.
Most likely to be voicers
Renters Lower than average household income;
Insecure tenure due to absence of written rental
contracts;
Not sure of eligibility for community housing
due to inconsistent policy of authorities;
Not well invested in the slum, because less
likely to be “community members”.
Most likely to be neglecters
Police officers (Borei
Keila)
Members of a tight-knit community;
Earliest settlers in Borei Keila, therefore likely
to be well invested and committed to staying in
the area.
Most likely to be voicers
Army officers (Borei
Keila)
Members of a tight-knit community;
Well invested in Borei Keila;
High level of commitment to staying in the area.
Most likely to be voicers
Vietnamese minority Members of an ethnic minority that is disliked,
and therefore likely to be more vulnerable and
isolated than other residents;
Not well invested in the slum, because less
likely to be “community members”;
Not sure of eligibility for community housing;
Most likely to be neglecters
Small business owners Higher than average household income;
Well-invested in the slum due to business
interests;
Likely to possess alternatives outside the slum,
but reluctant to leave because of high level of
investment in or near the slum area.
Most likely to be voicers
Woman-headed
households
Weaker than average economic position and
vulnerability due to weak economic situation;
Fewer options outside the slum;
Eligibility for community housing if community
members.
Most likely to be loyalists
Table 7: Sub-groups of respondents and their assumed EVLN responses

304

Community Leaders
The sub-group of community leaders refers to the heads or deputy heads of a so-called
“community” (i.e., sub-community) in a settlement. The term “community” in Phnom Penh has a
very specific official connotation: after an informal settlement is officially recognized and
“organized” by non-governmental organizations and local authorities, the settlement is sub-
divided into a number of separate “communities”, headed by a “community leader”. The size of a
community depends on the settlement, and can vary from several dozen to several hundred
families. The settlement of Borei Keila is sub-divided into ten communities, while Dey Krahom
has seven communities. Both Railway A and B comprise only one community each. In
principle, community leaders are nominated and elected by community members, although the
nature and frequency of these “elections” varies somewhat according to each community.
During the slum redevelopment process, the community leaders of the four land sharing
settlements negotiated on behalf of their members in all official meetings with the Municipality
and developers. In the context of my adapted EVLN model, community leaders would be
expected to display active citizenship and voicing behavior, on account of their official mandate
and responsibility to their members. Community leaders’ contacts with municipal officials, local
authorities and developers should in principle give them the political clout to pursue their
mandates.
Renters
Most renters tend to be poorer than the average slum dweller. Their economic situation
often precludes them from being able to buy a plot of land or a house in an informal settlement in
the city. Hence, rental is often the only option for people who depend on livelihoods in the city’s
informal market, but cannot afford to buy into the city. Renters are also vulnerable for another

305

reason: renters suffer from more tenure insecurity than property “owners” because all rental
agreements in the slums of Phnom Penh are negotiated in the absence of written contracts, which
leaves renters vulnerable to the whims of landlords.
In the land sharing context in Phnom Penh, a distinction must be made between long-
term and short-term renters in a settlement. In 2003 the Municipality agreed to grant community
membership status to people who had been renting for over three years in the settlements, but it
declined such status to “short-term renters”, defined as people who had been living in one of the
settlements for less than three years. Consequently, short-term renters are ineligible for new
housing as part of a land sharing deal. In practice, however, the status of long-term renters
remained just as insecure as short-term renters. In 2007 the Governor of Phnom Penh seemed to
withdraw the Municipality’s earlier pledge to long-term renters when he cancelled the rights of
long-term renters in Borei Keila to free housing in the new community buildings
262
.
In my adapted EVLN model, I assume that, as a result of their poor economic situation
and their insecure tenure status and poor prospects of sharing in the land sharing proceeds, most
renters will feel less commitment to the community than other residents. Given that they have
few good alternatives elsewhere, most renters will display neglect behavior, and they will be
treated as “non-citizens” by the authorities and by their own community leaders.
Army and Police Families in Borei Keila
As a result of the location of the former police training academy in Borei Keila, the
settlement of Borei Keila houses a substantial number of families of police officers. The police
sub-group is the longest resident population in the settlement, and most families have remained

262
Pierre Fallavier, Advisor to UN-Habitat, “Note on the Situation in Borei Keila, Phnom Penh”, e-mail
memorandum, 17 April 2007.

306

there even after the police training academy moved to another location. In addition to police
families, Borei Keila also accommodates a large number of families of army officers. Both
police and army families have formed their own “communities” in Borei Keila, headed by their
own community leaders.
Police families in Borei Keila, most of whom live in the original apartment blocks built in
the 1960s, have viewed with great apprehension the influx of other migrants into the settlement
since the 1980s, which has transformed the area into a dense low-rise slum characterized by rising
crime and insecurity. As a consequence, they are anxious to avoid further deterioration of the
area, and anxious to prevent the new community buildings from turning into instant new slums
263
.
As a result of their high level of investment, their internal organization and their concerns about
problems in the settlement, I assume that police families in Borei Keila can be categorized as
active citizens and voicers. I assume that army families, while less well-invested in the
settlement than the police community because they are more recent arrivals, nevertheless share
many of the same concerns and are equally well-organized and vocal in attempting to counter
negative developments in the settlement.
Small Business Owners
In the four case study settlements I defined small business owners as those who owned a
business in the settlement that employs more than one person, i.e. full-time businesses that cannot
be defined as “micro businesses”. This sub-group therefore does not include most hawkers or
motorcycle taxi (motodop) drivers, which are two of the most common livelihood activities for
urban poor Cambodians. This more restrictive definition of informal businesses enables me to
access residents who, in principle, should have an above average stake in the future of the slum,

263
Meetings with members of the police community, Borei Keila, Phnom Penh, 19 November 2003.

307

and therefore might represent a barometer of the state of active citizenship in the slum
environment. I expect that the larger their business in the settlement and the more money and
time invested in the business, the more invested small business owners will be in the settlement.
This is because small entrepreneurs will feel that the success of the business depends in part on
the health of the community, and therefore they have a stake in the community.
Given their level of investment in the slum and the surrounding neighborhood, through
their business, and given the likelihood that they earn above average incomes in the settlement, I
expect small business owners to demonstrate voice behavior in response to declining satisfaction
in the slum.
Ethnic Vietnamese Residents
Small ethnic Vietnamese populations can be found in two of the four case study areas—
in Dey Krahom and Borei Keila. This is a sub-group that has traditionally encountered much
suspicion and discrimination by the majority Khmer population, and which as a result tends to
stay within its own sub-community, preferring not to mix with the larger community. Their
participation in community events is likely to be less than the average resident. Ethnic
Vietnamese residents would probably be less well integrated and invested in the slum
environment than most other residents: in all likelihood most members of this sub-group would
not be official community members, and therefore their chances of qualifying for new community
housing under the land sharing scheme would be small, as in the case of renters.
Based on these traditional stereotypes of the Vietnamese minority population in
Cambodia, I assume that this sub-group would display behavior typical of neglecters or exiters as
a response to dissatisfaction in the slum, because their stake in the future of the community is
relatively small. Those with good alternatives outside the slum would move out, while those

308

without good alternatives, at least in the short term, would stay put and withdraw further into their
own community, without expecting a major improvement in their situation.
Woman-Headed Households
Another sizeable sub-group of community residents that is relatively more vulnerable
than the majority population is woman-headed households. Female headed households are
typically more economically vulnerable than male headed households on account of their reduced
income earning capacity, and they are at greater risk of social isolation as well. Their alternatives
outside the slum are likely to be limited. For these reasons I assume that, in the face of
dissatisfaction, woman headed households will have recourse to mostly passive behaviors, i.e.
loyalty and neglect. They will tend to opt for loyalty if they have experienced prior satisfaction
living in the slum, and are relatively well invested there, through social contacts or otherwise.
Conversely, if they have no such prior satisfaction and few investments, they will fall into neglect
behavior.
5.5.6 Selection and Recruitment of Respondents
The main principle guiding selection and recruitment of respondents for the study was to
obtain a representative, though not random, sample of respondents. In order to ensure that the
interview sample included members of all seven population sub-groups in the settlements, I
selected respondents through non-probability sampling, using a proportional quota sample
method. This method gives interviewers the possibility to choose the sample in such a way that it
includes specific percentages of various groups (Wright, 10). The specific percentages of
residents in each sub-group were determined, in three out of four settlements, in rough proportion

309

to the size of the sub-groups in each settlement, based on population figures obtained in advance
from community leaders and village chiefs in the settlement
264
.
The quota sample method was used also in the selection of “former residents”: figures
from the village chiefs were used to determine how many families had left each settlement since
the social land concession announcement in 2003
265
. Village chief estimates of the proportion of
former residents to leave the settlements were verified by asking interviewees how many former
neighbors or community members they knew that had departed in the last few years
(questionnaire items C3.8 and C3.9). The interviewers then attempted to obtain a roughly
proportionate ratio of former to current residents in the interview sample as exist in the
population of the settlements. Table 8 details the number of respondents from each sub-group
interviewed in each settlement, and indicates how many respondents were current residents and
how many were former residents, i.e., people who belonged to one or more of these sub-groups at
the time when they left the settlement.
Within each sub-group the recruitment of respondents was done in several different ways.
As a general rule, interviews were conducted in all communities of a settlement, to ensure a
geographical spread of respondents. A majority of community leaders was interviewed in each
settlement, and in some cases, vice-leaders were also interviewed. For respondents in certain
sub-groups, especially renters, women-headed households and ethnic Vietnamese—all of whom

264
This approach proved feasible in Borei Keila and the two Railway settlements, where community
leaders shared with the interview team—in advance of the interview process—the numbers of residents in
each community belonging to each sub-group. The approach was not feasible in Dey Krahom, where
community leaders were much more reluctant to share information on residents with the interviewers. In
the case of Dey Krahom, therefore, the survey team had to make its own estimate of the number of
residents belonging to each sub-group, based on informal discussions with the leaders and with residents.
265
The village chief officially records each new entrant and each departure from the settlement. Even
though some residents depart the settlement without informing the village chief, the village chief
nevertheless has the most complete and up to date information on population movements in and out of the
settlement.

310

are difficult to identify in the community on the basis of outward characteristics—the support of
community leaders was necessary to locate these people in the first place. Within all sub-groups
the recruitment of (additional) respondents was based on the snowball sampling approach,
whereby interviewers asked respondents to recommend other people to interview who met the
criteria for the study. The snowball sampling approach carries with it well-known risks of bias,
including that the sample depends on the first people interviewed and that the probability of
choosing other individuals depends on their social networks (Wright, 10). But the interviewers
had little choice but to accept these risks, given the disadvantages of approaching respondents in
other ways. The interviewers attempted to reduce the risk of bias by creating multiple snowball
samples in each community, by relying on more than one “first” respondents’ social networks.
The identification of “former residents” proved to be the most difficult recruitment
challenge, as most people who had departed the settlement could no longer be tracked down, and
few people in the old settlement (often including the village leaders and neighbors) still had their
contact details. Moreover, even if former residents could be traced, many of those who left a long
time ago were not interested to talk about the situation in their old settlement. In general,
departed residents who could be traced, through snowball sampling, and who were interested to
meet with the interviewers were those who left recently and who lived nearby
266
.

266
This included a sizeable number of families from Dey Krahom who now lived in the neighboring
settlement of Sambok Chab because they did not want to relocate with the rest of the community outside of
Phnom Penh.

311

Sub-group of respondents Total Name of Settlement
Borei Keila Dey Krahom Railway A Railway B
Community leaders 20CR/4FR 6CR/2FR 11CR/1FR 2CR/1FR 1CR/0FR
Renters 44CR/3FR 19CR/1FR 14CR/2FR 3CR/0FR 8CR/0FR
Police families 13CR/4FR 10CR/2FR 2CR/2FR 1CR/0FR 0CR/0FR
Army families 10CR/0FR 2CR/0FR 3CR/0FR 2CR/0FR 3CR/0FR
Vietnamese minority 5CR/0FR 2CR/0FR 3CR/0FR 0CR/0FR 0CR/0FR
Small business owners 20CR/4FR 7CR/2FR 9CR/1FR 2CR/0FR 2CR/1FR
Woman-headed household 32CR/5FR 14CR/0FR 12CR/2FR 0CR/1FR 6CR/2FR
Other /No special sub-group 39CR/18FR 19CR/2FR 10CR/5FR 9CR/2FR 1CR/9FR
CR denotes “current resident”; FR denotes “former resident”.
Totals for all respondents (CR and FR) add up to more than 100 percent since some respondents belong to
more than one sub-group.
Table 8: Total number of persons interviewed by sub-group, category and settlement

5.5.7 Potential Limitations of the Methodology
My data collection methodology was vulnerable to several forms of bias. The potential
for bias was anticipated before the interviews with respondents, so that as much as possible it
could be eliminated through the design of the questionnaires. Nevertheless, several forms of
structural bias remained, and had to be compensated for by exercising caution in data analysis.
Selection Bias
In the recruitment and selection of respondents there was a danger of inherent selection
bias as residents demonstrating “constructive” behaviors (voice and loyalty) were more likely to
be interviewed than residents demonstrating “destructive” behaviors (exit and neglect). This was
because voicers and loyalists are more visible than exiters, who by definition spend physically
less time in the slum, and neglecters, who physically withdraw from the slum community.

312

Moreover, residents demonstrating “constructive” behaviors tend to be less reluctant to be
interviewed. Thus, residents most likely to be interviewed included community leaders and
residents pushed forward by community leaders and/or local authorities to be interviewed because
they cooperated with the authorities as part of the “system”. The interviewers sought to
overcome the problem of selection bias by actively seeking out sub-groups assumed to be prone
to neglect or exit (such as renters, for example) and by seeking out respondents by themselves as
much as possible, rather than relying on the “helpfulness” of community leaders.
The Difficulty of Locating Former Residents
A specific risk of selection bias existed with relation to the small sample of former
residents interviewed for this study. Because locating respondents proved to be so difficult once
they had left the settlements, selection was driven largely by who could be located rather than by
obtaining a representative sample within pre-identified quotas of respondents, as was done with
current residents. As a result, the sample of former residents cannot be considered representative
of the total population of people who left the case study settlements. Of the small sample of 31
former residents, 15 respondents were originally from Railway A and B, 10 were from Dey
Krahom and 6 were from Borei Keila. The relative predominance of former residents from the
two Railway settlements and Dey Krahom was due to the fact that these people could be more
easily located than those from Borei Keila. Whereas families who left Borei Keila tended to
scatter throughout Phnom Penh, those from the two Railway sites overwhelmingly went to the
relocation site of Tang Krosang, where the developer offered residents willing to relocate a plot
of land, a legal title and basic services. Many families who left Dey Krahom, meanwhile, went to
the neighboring settlement of Sambok Chab, until that settlement was itself evicted in mid-2006.

313

The Risk of “Satisfaction Bias”
In measuring respondents’ relative dissatisfaction, satisfaction and “ritual loyalty” with
the status quo in the land sharing settlements, it is probable that respondents exaggerated their
satisfaction with the redevelopment plans, and downplayed their concerns. Despite the best
efforts of the interviewers to carefully explain the purpose of the study to respondents in advance
of each interview
267
, it is likely that some respondents nevertheless distrusted the impartiality of
the interviewers and suspected that the Municipality of Phnom Penh may be associated with the
research. Consequently, these interviewees might not wish to advertise their dissatisfaction with
the Municipality-approved redevelopment process for fear of reprisals against them. Respondents
who were community members (and therefore eligible for free replacement housing as part of the
land sharing arrangements) might feel that speaking badly about the Municipality, the developers
or their community leaders would lead to these parties striking their names off the list of
beneficiaries for free new housing. On the other hand, respondents who were currently not
eligible for new housing might exaggerate their satisfaction with the redevelopment process,
thinking that by doing so they might curry favor with the community leaders and that this would
improve their chances of qualifying for free housing.
Another possible reason for satisfaction bias is that in early 2006, the time period of the
interviews, the final outcome of the redevelopment projects was not yet known in two of the four
case study settlements, i.e., Railway A and Railway B. Consequently, in these two settlements,
the interviewers could ask residents of Railway A and B only in general terms about the land

267
Prior to each interview, the interviewers (who comprised the author, accompanied by two Khmer senior
research assistants cum translators) handed out an information sheet (in Khmer) to each respondent, which
explained in detail the purpose of the study and stated clearly that the interviewers had no affiliation at all
with the Municipality of Phnom Penh, the developer, or the land sharing project. In addition to handing out
a copy of the information sheet, the interviewers also verbally described the information sheet prior to each
interview and made clear to the respondents what the purpose was of the interviews.

314

sharing plans. The open-ended questions therefore did not capture the views of the residents
when it finally became clear that these plans were starting to collapse. Therefore, a possible
major source of dissatisfaction (focused on relocation and the compensation packages offered by
the developer to residents) was not reflected in the responses of residents from the two Railway
settlements.
The Assistance of Community Leaders
In several critical stages of the research, the interviewers relied on the assistance of
village chiefs and community leaders in each of the (sub) communities in the four case study
settlements. First of all, as is customary in Cambodia, the interviewers had to make contact with
village chiefs and community leaders in order to introduce the research and to seek (informal)
permission to conduct interviews in their jurisdictions. Second, the researchers needed the
community leaders and village chiefs in order to get updated information about the settlement
redevelopment plans, and (in the case of community leaders) to interview them as respondents in
their own right. Third, the researchers often relied on the assistance of community leaders to
access respondents in their communities as, of necessity, they were often the first point of contact
in the researchers’ snowball sampling approach.
The interviewers’ ties with the community leaders proved to be both a blessing and a
curse during the data collection phase. On the one hand, it was not possible to enter the
communities to conduct interviews without the permission (and often accompaniment) of the
leaders, and the assistance of the leaders in identifying various members of sub-groups of
residents proved to be critical to the success of the interview phase. On the other hand, it was
clear that some (but by no means all) residents viewed the assistance of their community leaders
to the interviewers as proof of close ties. As suspicion of the community leaders began to

315

increase during the redevelopment process (particularly in Borei Keila and Dey Krahom), the
assistance of the leaders at times proved to be as much of a hindrance to the interviewers’ ability
to access ordinary residents as a benefit. The researchers attempted to deal with this problem by
requesting community leaders (politely) not to be present during the interviews, so as to grant
respondents a more private environment to encourage them to speak their mind. In addition, in
other cases, after the necessary introductions by community leaders, the interviewers attempted to
persuade the leaders that their help and “protection” during the interview process would not
always be needed, and that they could find the necessary respondents through their own efforts.
Suspicion of the Interviewers
Despite the best intentions of the interviewers to win the trust of residents of the four land
sharing settlements as neutral researchers with no interest in the outcome of the slum
redevelopment process, it is likely that some residents nevertheless believed that the interviewers
were too close to their community leaders—and even that the interviewers might be undercover
informants for the Municipality of Phnom Penh. This is because residents had first met the
interviewers (the author as well as his two senior associates) through the U.N.-supported Phnom
Penh Urban Poverty Reduction Project based at the Municipality. Even though the interviewers
had then played a role as technical advisors to the communities, it is possible that in some
quarters, residents continued to associate the interviewers with their former roles at the
Municipality. In order to win the trust of ordinary community members (and to get beyond the
community leaders), the interviewers spent many hours talking to ordinary residents in between
interviews, and giving advice when asked.

316

Making a Good Impression
Another risk of bias resulted from the flipside of the suspicion described above, i.e., the
risk of skewed results as a result of respondents concealing their true intentions in order to try and
make a good impression vis-à-vis the interviewers, in the hope that this would bring them benefit
in the social land concession process. I assumed that this risk applied particularly to residents
with good alternatives outside the slum, particularly likely exiters. These categories of
respondents would be less likely to talk about their good alternatives (a key component of exiting
behavior) because the local authorities in Phnom Penh have announced that only those families
with no alternatives (which includes another house somewhere or excess material assets which
would qualify them as “too rich”) would be eligible for a new apartment in the land sharing
projects. Interviewees might therefore tend to exaggerate their constructive behaviors as well as
their lack of alternatives, which may lead to a perceived over-representation of loyalists at the
expense particularly of exiters, as the latter will have an incentive to hide their assets as well as
other options. The challenge, therefore, was two-fold: first, to win the interviewees’ trust so that
they felt they could talk openly to the interviewers without fearing that they would be reported to
the authorities, and second, to design the questionnaires such that the interviewers would be able
to pick up on a family’s entire range of assets and range of alternatives in both direct and indirect
ways—without being perceived as asking solely about “income”.

317

CHAPTER 6: DATA ANALYSIS
6.1 Overview of Chapter 6
In Chapter 6 I analyze data from structured interviews conducted between January and
April of 2006 with a representative sample of 152 “current residents” of the four land sharing
settlements of Borei Keila, Dey Krahom, Railway A and Railway B. The objective of the
interviews was to determine whether respondents displayed some measure of dissatisfaction with
the conditions in their settlements, and if so, to examine whether they deployed exit, voice,
loyalty or neglect behaviors—or a combination of those—to address their dissatisfaction. In
addition to the interviews with current residents, I conducted separate interviews with 31 “former
residents”—those respondents who had already left one of the four case study settlements and
were now living elsewhere in Phnom Penh. The purpose of this separate analysis was to discover
whether or not former residents could be characterized as “true exiters” according to the EVLN
literature, i.e. whether former residents resembled the proverbial “consumer voters” as defined by
Charles Tiebout (1956) who, in response to dissatisfaction in their jursidiction, move (voluntarily)
to another community which best satisfies their preference pattern for public goods.
The chapter is divided into three major parts. In the first part (section 6.2), I present an
overall profile of respondents, as a basis for the data analysis. The profile describes the
background of residents interviewed for the study, their socio-economic status, their perceptions
of their tenure status and the reasons for their satisfaction and dissatisfaction in the four
settlements.

318

The second part of the chapter contains the data analysis and results for current residents
(sections 6.4 through 6.5). Data analysis was carried out in four steps. The first step was to
identify the prevalence of relative dissatisfaction among current residents (section 6.4).
Dissatisfied respondents constituted the main sample for the analysis, because as Lyons and
Lowery point out, “individuals must perceive problems or be dissatisfied with the current state of
affairs before they are likely to expend the time, energy, and resources to engage in…responses to
dissatisfaction” (Lyons and Lowery 1989, 854).
In the second step (section 6.4) I identify the EVLN strategies of current residents
through an “independent assessment” based on three indicators of respondents’ behavior:
participation level, response to the redevelopment plans, and commitment to the settlement. In
the third step (section 6.5) I test the strength of the Lyons and Lowery decision-tree model (1989)
as a predictor of the independently assessed EVLN strategies among current residents.
In the third part of this chapter (section 6.6) I measure the extent of “Tiebout exit”
behavior among former residents, following a similar four-step approach as in the case of current
residents.
6.2 General Profile of Respondents
The general profile serves as a background for the analysis of slum dwellers’ behaviors
for the independent assessment and the Lyons and Lowery decision-tree model. The profile
covers both current and former respondents.
6.2.1 Background of Respondents
Of the total number of current and former residents interviewed, 60 percent were female
heads of household (N = 94) and 40 percent were male heads of household (N = 58). The
discrepancy in gender representation is mainly due to the fact that women tend to work or stay in

319

or around the home, and were therefore easier to recruit for interviews, whereas most men work
outside the settlement. The average age of the interviewees was 44. The average household size
was 5.8 members
268
.
From the point of view of the length of residence of most of their inhabitants, the four
case study slums can be considered as “settled communities”. Among current residents, the vast
majority (71 percent; N = 108) of respondents have lived in their settlement for over ten years.
Among former residents, too, almost three-quarters of respondents (74 percent; N = 23) had lived
in their settlements for over ten years before they departed. Most respondents arrived in Phnom
Penh as part of a wave of migrants seeking better economic opportunities in the capital in the
early 1990s, after the Paris Peace Accords were signed in 1991. Ten respondents belonged to the
very first wave of refugees who arrived in Phnom Penh shortly after the fall of the DK regime in
1979. Two of these first arrivals (both families of police officers) settled in Borei Keila; the eight
other families settled on Railway company land, in what later became known as Railway B
settlement. The number of newcomers who arrived in the settlements after the social land
concession announcement by Prime Minister Hun Sen in 2003 was small, at 5 percent of
respondents (7 respondents).
Fifty-six percent of all respondents came to the settlement more or less directly from
areas outside Phnom Penh. Of these respondents, 85 percent came from other provinces in
Cambodia; 8 percent came from the refugee camps on the Thai border; and 7 percent from
southern Vietnam, of whom 3 households were ethnic Vietnamese and 4 households were ethnic
Khmer (Khmer Krom). Forty-four percent of all respondents came to live in the settlements after
first living in other locations in Phnom Penh.

268
For the purpose of the study, “households” were defined as related persons living under the same roof at
the time of the interview.

320

6.2.2 Occupations and Income
Respondents had a wide variety of occupations, from waste pickers, at one end of the
economic spectrum, to government employees and owners of formal businesses such as furniture
stores, on the other. The most common occupation of women respondents was food vendor: for
nearly a quarter of all female interviewers this was the main income generation activity. This was
followed by operating a small grocery business attached to the house (16 percent of cases);
working as a seamstress (7 percent); working in the hospitality industry (restaurants, massage
parlors, karaoke bars and nightclubs) (7 percent); and working as a police or army officer (4
percent). Eighteen percent of women respondents were currently staying at home full-time, either
because they were sick, or to “take care of the house” or to look after the household’s children.
Twelve percent of women were staying at home occasionally, and working on an intermittent
basis
269
. In two-thirds of cases, women’s occupations and income generation activities took place
in the settlement or within walking distance of the settlement.
For men, the principal occupation was motodop driver: 35 percent of male respondents in
the settlements relied on motodop driving for at least part of their income. The second most
common activity was food vendor (12 percent), followed by construction work (7 percent) and
government employment (7 percent). In Borei Keila, as a result of the settlement’s previous links
to the Police Training Academy, there is still a substantial population of police officers and their
families: 27 percent of men interviewed in that settlement worked for the police force. For 85
percent of all male respondents, income generation activities took place across Phnom Penh, and
were not confined to the immediate proximity of the settlement.

269
For women’s and men’s occupations, totals do not add up to 100 percent as many respondents reported
more than one income generation activity.

321

On the basis of self-reported information
270
, the median household income for current
residents in the four settlements was established at 16,000 riel per day ($4). Reported incomes
ranged widely, from 1,700 to 200,000 riel per household per day. The median income per
household member, based on the number of household members and incomes reported for each
household, was 2,857 riel ($0.71) per person per day. This figure is slightly higher than the
government’s overall poverty line (OPL) of 2,470 riel ($0.62) per person per day for Phnom
Penh, as applied by the National Poverty Reduction Strategy for the period 2003-2005
271
. More
than one-third (36 percent) of all households who reported income data earned less than the
government’s OPL of 2,470 riel per household member per day. The largest contingent of these
poorest households was made up of renters, women-headed households and families of police
officers. Thirty-four percent of renter households who reported their incomes earned less than the
OPL; the corresponding figure for women-headed households was 45 percent and for police
families 90 percent.
Women-headed households could also be found at the other end of the income spectrum
in the four settlements: 28 percent of female headed households reported incomes that were
higher than 5,000 riel per household member per day, or over twice the OPL level. This was true
of 64 percent of households with small and medium sized businesses, and two out of three ethnic
Vietnamese respondents.

270
Income information is based on information from 115 households of current residents. Given that self-
reported income figures are notoriously unreliable—as earnings fluctuate and as respondents may be
reluctant to share information with the interviewers—I “checked” self-reported income data (C5.7A) with
data on household expenditures (C5.7B), to make sure these were consistent with one another.
271
The Royal Government of Cambodia’s National Poverty Reduction Strategy (2003-2005) establishes an
overall poverty line for Phnom Penh of 2,470 riel per person per day. The OPL represents “an adequate
income for a person to consume a food basket that provides at least 2100 calories of energy per day with a
small allowance for non-food items such as shelter, and clothing” (RGC 2002, 31).

322

Former residents (those families that had left one of the four case study settlements)
reported a median income of 15,000 riel per household per day ($3.75), and a median income per
household member per day of 2,853 riel ($0.71)
272
. Just under a third (32 percent) of households
of former residents who reported income data earned less than the government’s OPL of 2,470
riel per household member per day. These income indicators suggest that the economic profile of
former residents is very similar to that of current residents of the case study settlements.
6.2.3 Perceptions of Tenure Security and Tenure Status
Almost three-quarters of all respondents (73 percent; N = 132) claimed to be the
legitimate “owners” of their properties in the settlements. The great majority of owner occupants
(73 percent) had purchased the land on which they settled through informal transactions. In most
cases these transactions were negotiated with other, already established residents in the rapidly
growing settlements. In some cases these informal transactions were concluded with police or
other authorities, such as staff of the State Railway Company.
Over a quarter of owner occupants (27 percent) mentioned that they had obtained the land
they occupied for free. Most of these people were early arrivals from the 1980s and early 1990s
who occupied still vacant land before land was parceled out and sold by residents, but a small
number of respondents (8 families) obtained free properties from the authorities. In Borei Keila,
four families of police officers received their apartment units for free as in-kind compensation
from the Police Training Academy. In Railway B settlement two families of State Railway
Company employees obtained free land from the State Railway Company, while more recently,
two families in Railway A received a free house and plot from the Municipality of Phnom Penh

272
Income indicators for former residents referred to their present situation, after they had moved out of the
four case study settlements.

323

as compensation for losing their original homes to road widening works undertaken by the
Municipality in the early 2000s.
No current or former residents interviewed for the study possessed formal titles for their
properties, and no respondents had ever applied for a formal title. The most common documents
used by respondents to “prove their residence” in the settlement were different forms of
identification issued by the local authorities (chiefly, the sangkat). These include voting cards
(held by 80 percent of respondents); family books (73 percent); and identification cards (51
percent). In addition to these official documents residents used two informal documents to prove
their residence in the settlement. These included saving books used for the community savings
schemes (held by 63 percent of respondents), which are issued by community leaders to
households registered as community members in their community of the settlement. The second
type of informal documents was receipts for land purchases negotiated between buyers and sellers
in (informal) land transactions (held by 18 percent of respondents).
Despite residents’ lack of formal titles the majority of respondents (52 percent) from all
four case study settlements nevertheless felt confident that their properties enjoyed “secure legal
status”. Thirty-four percent of respondents felt that they did not enjoy secure legal status, while
12 percent were not sure. The main reason (in 76 percent of cases) for respondents’ perception of
security was the belief that they had been accorded greater recognition by the authorities. Sixty
percent of these respondents felt that the social land concession announcement by Prime Minister
Hun Sen in 2003 was the most important act of “official recognition” by the national authorities;
this recognition gave them the legal right to remain on site. Twelve percent of respondents felt
that the most important factor explaining their secure tenure was the various acts of recognition
given to them by local authorities, including the issuance of family books by the sangkat, or the
permission to construct in the settlement by the village (phum) chief, or the recognition of

324

community leaders. Six percent of respondents mentioned that the absence of recent evictions by
the authorities was the main factor underpinning their secure legal status. Interestingly, far fewer
respondents (16 percent) indicated that their perception of secure tenure was related to their own
legal possession right, which in turn was the result of their long residence on site. Even fewer
respondents (3.7 percent) felt that their secure tenure was due to stronger community organization
or to the contribution of NGOs.

We have ‘secure legal status’ (N = 108) We do not have ‘secure legal status’ (N = 69)
Main reasons cited by respondents N Main reasons cited by respondents N
Social land concession is proof of
recognition by authorities
65 We have no legal right to stay here
because this is public land
37
We have stayed in this settlement for a
long time, so we have a legal right to stay
17 We are not eligible because we are renters
or because we are not community
members
11
Local authorities gave us recognition by
issuing family books and granting
permission to construct
13 We feel ‘temporary’ here because the
government needs this land for
development
7
There have been no evictions for a long
time
7 Our houses are not in ‘good order’ 5
We trust community efforts and get
support from NGOs
4 We do not trust the Municipality and the
developer
3
Various other reasons 2 Various other reasons 6
Table presents number of responses (N) for each category. Number of responses is greater than number of
respondents because some respondents cited more than one reason.
Table 9: Rationale for Respondents’ Perceptions of Tenure Security or Insecurity

On the other hand, respondents who felt that they did not enjoy secure legal status cited
as their main reason (in 54 percent of cases) the belief that they had no legal right to occupy their
plots because their settlement was on state land. Other reasons behind the perception of tenure
insecurity were the fact that some respondents thought that they themselves were ineligible for
new community housing (in 16 percent of cases), or the belief (in 10 percent of cases) that

325

residents “felt temporary” because the government needed the land for development.
Respondents’ main reasons for perceptions of tenure (in)security are summarized in Table 9.
In an apparent contradiction with their views about secure legal status, a substantial
majority of current residents (78 percent) as well as former residents (87 percent) indicated that
they felt like “squatters” (samnang ana thibtay) in the four settlements. The question referred to
the Khmer term samnang ana thibtay, which translates roughly as “anarchist”—the pejorative
term long used by the authorities to refer to squatters, which conjures up associations of illegality,
disorder, and threats to the wider public. Even a majority (66 percent) of current residents who
felt they enjoyed “secure legal status” said they still felt like squatters.
The apparent contradiction can be better understood when it becomes clear that, for the
majority of current and former residents (63 percent of cases), “feeling like a squatter” is not
based primarily on a legal definition of what it means to be a squatter. Rather, the main non-legal
reason for “feeling like a squatter” (in 36 percent of cases) has to do with physical disadvantages
characterizing informal settlements, such as “houses and streets not being in good order”; the
poor living conditions of residents; the poor quality (and high cost) of basic services; the
proliferation of shacks; and the concerns about fire and arson. Fifteen percent of respondents
indicated that “feeling like a squatter” has to do with the poor treatment they receive from the
authorities, including the poor management of the area; the refusal to allow further construction;
or the threats of eviction. Finally, 10 percent of respondents mentioned that it is social stigma and
social disadvantages that make them “feel like a squatter”, including feeling shame about living
in a poor and dirty area; hearing from others that their settlement is a “squatter place”; and having
to share the settlement with a “mixed group of people”, including community members who have
no respect for others.

326

For the minority of respondents who indicated that they did “not feel like a squatter” (18
percent of cases), the most commonly cited reason for this (in 65 percent of cases) was the
recognition these respondents felt they obtained from the local and national authorities as a result
of the social land concession announcement by Prime Minister Hun Sen. Far fewer respondents
(27 percent) mentioned their possession right, as stipulated in the Land Law 2001, as a reason for
“not feeling like a squatter”, and only two respondents (5 percent of cases) mentioned community
organization and support by NGOs as important factors in this context.

I feel ‘like a squatter’ (ana thibtay) here (N=195) I do not feel ‘like a squatter (ana thibtay)’ (N=37)
Main reasons cited by respondents N Main reasons cited by respondents N
Legal status of settlement is unresolved:
we live on public land; we don’t have real
titles; we are not recognized by the
cadaster, etc.
72 We have recognition from the
Municipality of Phnom Penh and national
authorities as a result of the social land
concession announcement
15
Physical aspects of settlement are sub-
standard: houses are ‘not in good order’;
poor living conditions; basic services are
inadequate; access roads are narrow, etc.
70 We have stayed in this settlement for a
long time, so we enjoy legal rights
according to the Land Law 2001
10
Local authorities neglect the settlement:
local authorities don’t care about slum
dwellers; management of the area is poor;
there are eviction rumors, etc.
29 We have recognition from the local
authorities (sangkat and village chief)
9
Social stigma around the settlement: other
people say this is a squatter settlement;
we are ashamed to live here; residents
don’t respect each other, etc.
20 The community is organized and we enjoy
the support of NGOs
2
Various other reasons 4 Various other reasons 1
Table presents number of responses (N) for each category. Number of responses exceeds number of
respondents as some respondents cited more than one reason.
Table 10: Rationale for respondents’ perceptions of tenure security or insecurity

These results testify to respondents’ low level of awareness of their legal rights, and
signal respondents’ continued dependence on the goodwill of the authorities (in the form of the
recognition by Prime Minister Hun Sen and the municipal authorities) to determine their tenure

327

status. The principal reasons given for “feeling like a squatter” or “not feeling like a squatter” are
summarized in Table 10.
6.2.4 Satisfaction and Dissatisfaction among Slum Residents
Satisfaction with Central Location and Community Solidarity
A vast majority of current residents interviewed (N = 120; 79 percent) claimed that they
were satisfied with their lives in the settlement (C2.1). Only 22 respondents said that they were
dissatisfied, and another seven respondents said that they were ‘sometimes’ dissatisfied. The
main reason for the satisfaction of respondents was the central location of the settlements: 85
percent of current residents who were satisfied with their lives in the slum (102 out of 120)
mentioned their settlement’s proximity to income generation sources in the city and services
(schools, hospitals and pagodas) as the main reason for their satisfaction. Many residents
mentioned that their settlements were like “rice fields” to them due to the good income earning
opportunities in their inner city locations.
Location was also the most important factor for former residents, 70 percent of whom (14
out of 20) cited their former settlement’s proximity to income generation sources, markets and
schools as the main reason for their satisfaction while they were living there. Another source of
satisfaction with life in the slum included community solidarity. A vast majority of current
residents (134 out of 152, or 88 percent) indicated that residents help each other and trust each
other (C2.5). Mutual assistance and support were extended particularly during emergencies, or
during periods of family crisis.
Dissatisfaction with Basic Services and Security
In all four case study areas, the main source of dissatisfaction of respondents was the
poor living environment. Current residents of the four land sharing settlements were dissatisfied

328

with key components of the (informal) “tax-service package” of the slum, including electricity,
water, roads and footpaths, drainage and safety and security.
Over 80 percent of respondents in all four settlements (123 out of 152) experienced some
or substantial dissatisfaction with their electricity service, principally because of the high cost of
these services. Informal settlements are not serviced by the state electricity company, so as a
result residents have to arrange private hook-ups, which are invariably more expensive than even
the highest rate charged to consumers by the public provider. Six percent of respondents could
not afford any electricity connection at all and made do with petrol lamps, candles, or street
lighting. A majority of current residents interviewed in all settlements (58 percent, or 88 out of
152 respondents) also expressed dissatisfaction with their water service, for the same reason as
the electricity service, namely, the high cost charged by private vendors.
The poor state of roads and footpaths was cited by almost three-quarters of current
residents in all four settlements as another major source of dissatisfaction. Roads and footpaths
are generally narrow and unpaved, and therefore muddy and washed out in the rainy season. In
the two Railway settlements, poorly maintained footbridges over the boeung have become rotten
and unstable over time due to lack of maintenance. The lack of maintenance, in turn, is a result of
the insecure tenure that a majority of residents experience: residents are disinclined to invest
money and time into the upkeep of footbridges as long as it remains unclear whether or not they
can remain on site. Over half of current residents in all settlements (53 percent) also expressed
some or substantial dissatisfaction with the inadequate drainage in their settlements. In the two
Railway settlements natural drainage was provided by the boeung, which produced problems
during heavy rains when the area flooded. In Borei Keila and Dey Krahom blocked or poorly
maintained drains were blamed for regular flooding and public health concerns.

329

Finally, among current residents, over 90 percent of respondents in all settlements were
dissatisfied with the safety and security in their settlements: 16 percent said they do not feel safe
at all, while 75 percent said they ‘sometimes’ do not feel safe. A majority of respondents
indicated that security had improved compared to before, thanks to increased interventions
sanctioned by Prime Minister Hun Sen and carried out by police and local authorities targeted at
“drug users” and gangs (bong thom). Nevertheless, burglary, gang activity and prostitution were
cited as persistent problems.
Broad Satisfaction with the Land Sharing Plans
At the time of the interviews during the first half of 2006, the redevelopment plans in
three of the four case study settlements still called for on-site upgrading through land sharing. In
these three settlements (Borei Keila, Railway A and Railway B), the land sharing plans were
welcomed with high levels of satisfaction by a large majority of current residents. In their open-
ended comments about the redevelopment plans (C3.5), 77 percent of current residents
interviewed in Borei Keila (50 out of 65 interviewees) expressed full support for the land sharing
project, although there was some concern about living and doing business in high rise buildings.
Some older residents, residents with disabilities, and people running small businesses from home,
in particular, expressed concern about being allocated housing units on the upper floors of the
walk-up apartment blocks.
In the Railway settlements almost all current residents interviewed expressed their
satisfaction with the land sharing plans, with 82 percent support in Railway A (9 out of 11
interviewees) and 92 percent support (22 out of 24 households interviewed) in Railway B. But
the satisfaction levels in the Railway settlements must be interpreted with some caution. At the
time of the interviews in early 2006 the land sharing plans for the two Railway settlements

330

existed only on paper, so the interviewees were not able to comment on an actual redevelopment
project, unlike residents of Borei Keila, where the land sharing construction process was already
underway
273
.
The high levels of satisfaction with the land sharing plans in Borei Keila and the Railway
sites can be explained by several factors. First, the plans offered the prospect of formal (i.e., state
approved) housing, utility connections and tenure to residents experiencing profound
dissatisfaction with the high cost of informal services and the poor security in their present slum
environment. Second, the plans promised to re-house the residents on site, thereby ensuring that
they could continue residing in the city center, which is the single most important reason that
attracted most residents to all four settlements in the first place. A third reason was the fact that
the new housing would be provided for “free” to community members
274
.
In Borei Keila, while there was little opposition to the land sharing project as such, more
than a third of all respondents (34 percent; 22 out of 65 interviewees
275
) nevertheless expressed
some dissatisfaction with the project, not out of opposition to the redevelopment plan as much as
the untransparent process led by the Municipality and the developer. There were three main areas
of concern about the land sharing process. First, respondents worried that construction of the
“community buildings” was too slow, and that the project would never be completed. Second,

273
In early 2006, the time period of the structured interviews, residents of Railway A and B were still
unaware that the land sharing plans in their settlements would ultimately collapse and that they, too, would
have to relocate.
274
According to the Land Law 2001 most residents would have enjoyed possession right to their existing
properties in the four settlements. Therefore, in legal terms, it would perhaps be more accurate to refer to
the new community housing as a form of compensation by the developer to the residents for the right to
develop on site. However, the authorities never accepted this legal argument, and most residents, too,
seemed unaware of the Land Law 2001 and its implications for their tenure rights. As a result, most
residents tended to regard the new housing as “free” and as a gift from Prime Minister Hun Sen and the
Municipality of Phnom Penh.
275
Responses on satisfaction and dissatisfaction are greater than total number of respondents, as some
respondents who supported the land sharing plan had objections to the land sharing process.

331

respondents mentioned that they did not trust the developer to invest in adequate construction
materials for the community buildings. And third, respondents voiced their disagreement with
the untransparent procedures for allocation of housing units to community members. Long-
established residents (those who lived in the original apartment buildings), including particularly
families of the original police community, resented the inflow of more recent arrivals in Borei
Keila and objected to the equal treatment of all families in the lottery process for allocation of
new apartment units.
In addition to respondents’ concerns about the redevelopment process, at least a third of
all interviewees in Borei Keila
276
(22 households) felt unsure whether they would be eligible for a
new apartment at all. This group of people included 19 renter households, who were left in doubt
about their eligibility as a result of the Municipality’s shifting policies towards all renters. But
this group also included at least three families of house owners, who were worried about not
getting a new house because of the untransparent procedures, two of whom were long-term
residents in Borei Keila and one of whom was a recent arrival.
In the two Railway settlements, where a slum redevelopment project had not yet been
agreed with the developer, few respondents mentioned their concerns with the redevelopment
planning process in their open-ended answers. In Railway A, only 2 out of 11 interviewees
expressed doubt that a land sharing project was ever going to materialize; in Railway B, this
concern was uttered by only one out of 24 respondents. While this result is to be expected given
that redevelopment was not yet underway, the lack of dissatisfaction with process is somewhat
surprising in light of the collapse of the land sharing agreement just two years later.

276
Some respondents in this group also belonged to the group voicing concern about the land sharing
process.

332

Dissatisfaction with Relocation in Dey Krahom
In Dey Krahom, the situation was very different from that in the other three case study
areas. In this settlement, where the slum redevelopment plans at the time of interviews in early
2006 called for relocation outside Phnom Penh instead of land sharing on site, support for the
redevelopment plans was decidedly weaker than in the other three slum areas. Instead of ritual
loyalty, there was widespread dissatisfaction: only 21 respondents (40 percent of current residents
interviewed) expressed satisfaction with the relocation plans. A majority of current residents
interviewed in Dey Krahom expressed serious misgivings and dissatisfaction with the relocation
proposal of the developer. Nineteen out of 52 current residents interviewed (37 percent)
expressed outright dissatisfaction or concern with the project; another ten respondents (19 percent
of interviewees) said their households did not support the project but that they had no choice but
to follow the community to the relocation site as they had nowhere else to go. Another two
respondents claimed to have no idea about the development plan at all.
The most important reason for the dissatisfaction with redevelopment plans in Dey
Krahom was residents’ reluctance to relocate 15 km outside the city center, where livelihood
opportunities, markets, schools and health clinics were still scarce, and from where the cost of
transportation back and forth to the city would be high. For those respondents who did support
the relocation project, the main reasons given were the secure land title, the larger and more
durable housing units and the prospect of living in a better environment, with more open space
and (perceived) better security. A number of interviewees also mentioned their reluctance to live
in a high-rise building, as in Borei Keila, which would have been the alternative if land sharing
had been carried out in Dey Krahom. Two current residents in Dey Krahom commented on the
process by which the relocation project had been negotiated and planned between the developer,

333

community leaders and local authorities. These two respondents described the process as
“untransparent” and “immoral”
277
, and referred to some community leaders as “corrupted”
278
.
6.3 A Common Measure of Relative Dissatisfaction
The first step in both my independent approach to identifying EVLN as well as the Lyons
and Lowery (1989) decision-tree model was to measure the extent of relative dissatisfaction
among current residents and to exclude from the sample those respondents who are found to have
no dissatisfaction whatsoever with their lives in the settlements (the so-called “ritual loyalists”). I
utilized a common measure of relative dissatisfaction for both models, based on the same pool of
dissatisfied (non-ritually loyal) respondents. This enabled me to directly compare the results of
both approaches.
6.3.1 Current versus Prior Satisfaction
In order to capture the essence of a “baseline” notion of satisfaction or dissatisfaction
among slum residents, I created an ordinal measure of relative dissatisfaction composed of five
items assessing respondents’ current levels of satisfaction with conditions in their settlements or
the redevelopment plans/process relative to their prior satisfaction with these items. I identified
four sets of respondents, using the same ordinal scale as employed by Lyons and Lowery (1989).
At one end of the scale are respondents who indicated that they had always been satisfied, both at
the current moment as well as in the past; these respondents were identified as ritual loyalists. At
the other end of the scale are respondents who had been satisfied in the past but who were
currently dissatisfied. The dissatisfaction of these residents is treated as more extreme than the

277
H.L., current resident of Dey Krahom, interview with author, Phnom Penh, 16 February 2006.
278
C.V., current resident of Dey Krahom, interview with author, Phnom Penh, 16 February 2006.

334

dissatisfaction of chronically dissatisfied people
279
, given that it is less familiar and more
unpredictable. In the middle of the scale are respondents identified as merely “satisfied”, who
demonstrated dissatisfaction in the past but who indicated that they currently experienced greater
satisfaction, and respondents who were chronically dissatisfied, both in the past as well as the
present.
The measure of relative dissatisfaction was comprised of five items assessing satisfaction
or dissatisfaction with different aspects of life in the slum. Responses to the five items were re-
coded from both multiple choice as well as open questions. The variety of different measures of
satisfaction—and the different ways of asking residents about satisfaction—helped to obtain a
multi-faceted picture of satisfaction or dissatisfaction in the settlements. The relative satisfaction
measure has a reliability coefficient of α = 0.73
280
. The first item was a measure of satisfaction
with general conditions in the settlement as an overarching measure of relative satisfaction. This
item was based on asking respondents directly about how satisfied they were to live in the
settlement. The four additional items were indirect or proxy measures assumed to be highly
correlated with satisfaction. The second item concerned security of tenure: respondents were
asked whether they perceived their tenure in the settlement to be legally secure or not. Those
who answered that they did were considered satisfied; those who indicated that they felt insecure
or were unsure about their status were categorized as dissatisfied. The third item asked
respondents whether they had ever (in the past or at the present time) made plans to leave the
settlement; those who answered that they had were considered dissatisfied, as they were

279
The treatment of more recent dissatisfaction as “extreme dissatisfaction” follows the methodology of
Lyons and Lowery (Lyons and Lowery 1989, 853).
280
The reliability of all multi-item measures in the study was tested using Cronbach’s alpha (α), which is a
model of a measure’s internal consistency based on the average inter-item correlation. A reliability of α =
0.70 is commonly considered to be a minimum threshold indicating reliability of the measure.

335

presumed to have a good reason for wanting to leave the settlement. The fourth and fifth items
assessed respondents’ level of satisfaction with the redevelopment plans and process,
respectively, for their settlements. Satisfaction with either the plan or the process (or both) was
assessed by analyzing and coding respondents’ answers to three open-ended questions. The first
question (C3.5) asked respondents directly what they thought of the redevelopment plans or
process. The second question (C3.6) followed up on this by asking respondents about their future
plans, specifically, whether they intended to stay in their community after redevelopment. The
third question (C4.2) gave respondents an opportunity to add any of their own thoughts about the
redevelopment plans or process
281
.
6.3.2 Dissatisfaction in the Four Case Study Settlements
Using the multi-item measure described above, just over 90 percent of respondents (N =
137) indicated at least some degree of relative dissatisfaction with their lives in the settlement or
with the redevelopment plans or process slated for their settlement. Slightly fewer than 10
percent of respondents could be considered ritually loyal, or fully satisfied both in the past and at
the present with all items included in the measure.
The results of the measure are summarized in Table 11. The multi-item measure of
satisfaction uncovered a fair amount of hidden dissatisfaction: when asked directly about their
satisfaction with overall conditions in the settlement (the first item), “only” 37 percent of
respondents acknowledged feeling dissatisfied or highly dissatisfied, but when other aspects were
taken into account (such as respondents’ perceived insecurity of tenure, their past or current
desire to leave the settlement and their thoughts about the redevelopment of their settlement) a

281
Even though residents were known to have serious misgivings about other aspects of the quality of life
in the settlements, such as quality of services and crime, these items could not be included in the measure
of relative (dis)satisfaction as they did not correlate highly with overall dissatisfaction and would thus have
rendered the overall measure unreliable.

336

more complete picture of dissatisfaction was revealed. The analysis of responses to
dissatisfaction in this study (the basis for EVLN behaviors) is based on both of these forms of
dissatisfaction, that is, the openly expressed dissatisfaction as well as the more hidden incidences
of dissatisfaction revealed by this measure.

Individual items
measuring relative
dissatisfaction
(Satisfaction = ‘yes’)
Highly Satisfied Satisfied Dissatisfied Highly
Dissatisfied
Current SAT
Prior SAT
Current SAT
Prior DISSAT
Current DISSAT
Prior DISSAT
Current DISSAT
Prior SAT
Satisfied with overall
conditions in the slum
N = 76 N = 20 N = 12 N = 44
Perceive tenure in the
slum to be secure
N = 37 N = 48 N = 65 N = 2
Never made plans to
leave the slum
N = 103 N = 15 N = 22 N = 12
Support redevelopment
plan
N = 75 N = 16 N = 13 N = 48
Support redevelopment
process
N = 71 N = 18 N = 18 N = 45
Fully satisfied in all
items (“ritual loyalty”)
N = 15 n/a
Total sample N = 152; SAT = ‘Satisfied; DISSAT = ‘Dissatisfied’.
Ritual loyalists (N = 15) are eliminated from the sample of dissatisfied respondents.
Reliability coefficient of the relative dissatisfaction measure: α = 0.73.
Table 11: Component items of relative dissatisfaction and results for each item
(Ordinal scale adapted from Lyons and Lowery, 1989)

6.4 EVLN Measures in the ‘Independent Assessment’
My independent approach to identifying EVLN behavior among slum residents is based
on placing interviewees in EVLN categories according to their responses to three key indicators.
The first independent indicator is the degree to which respondents participated in organized

337

events and meetings in their communities. The second indicator examined the reaction of
respondents to the redevelopment plans and process in their settlements. The third indicator
looked at whether respondents are currently considering leaving their settlements and whether or
not they have the possibility to do so. The combination of the three indicators has a reliability of
α = 0.82. The sample for the independent model included only those residents (N = 137) who
expressed some degree of dissatisfaction with their past or present condition in the slum, as
identified by the measure of relative dissatisfaction in section 6.3.
The categorization of respondents according to EVLN behaviors in the independent
assessment is based on identifying respondents by their “dominant” EVLN behavior
282
. Thus,
respondents were categorized a certain EVLN behavior if they “scored” positively for that
behavior on at least two out of three of the key indicators. If respondents showed no dominant
EVLN strategy (i.e., if respondents did not score positively on at least two indicators of any
strategy), then they were considered to demonstrate a mixture of EVLN behaviors.
Within each indicator I assessed respondents’ EVLN behaviors on the basis of secondary
multi-item measures consisting of various “sub-indicators”, which were also tested for reliability.
Respondents’ behaviors were categorized according to an ordinal scale measuring degrees of
collective action that were associated with EVLN behaviors, as illustrated in Figure 23. In
categorizing respondents according to EVLN behaviors, I followed a similar coding system as
with the key indicators, based on identification of a dominant EVLN strategy: thus, respondents
were categorized as a certain response type if they scored positively for that behavior on a

282
The EVLN literature acknowledges that individuals (whether consumers in Hirschman’s theory or
residents in the model of Lyons and Lowery) display a mixture of EVLN behavior at any one given time.
For the purposes of coding in this study, it was the dominant EVLN behavior that was measured in
interviewees’ responses.

338

majority of the sub-indicators. Those who showed no dominant strategy among all the sub-
indicators were considered to demonstrate a mixture of EVLN behaviors.

More Less
Actively engaged:
“articulating” for
community benefit
Seeking to improve
redevelopment
plans or process
Passively engaged:
“joining and
following” most
activities
Unconditional
acceptance of
redevelopment
plans or process
Not (very) engaged:
“too busy with other
things”
Planning to leave
and pursue own
agenda,
irrespective of plans
Not (very) engaged:
“don’t care” or
“don’t know”
Resigned
acceptance of
redevelopment
plans or process
Independent indicator 1: Participation in community activities
Independent indicator 2: Reaction to redevelopment plans & process
Possibility but no
intention to leave
the settlement
No intention and no
possibility to leave
the settlement
Planning to leave
the settlement and
possibility to do so
Desire to leave the
settlement but no
possibility to do so
Independent indicator 3: Leaving or staying
Neglect Exit Loyalty Voice
Collective Action

Figure 23: Key indicators in the independent approach to assessing EVLN

6.4.1 Indicator 1: Participation in Community Activities
The first indicator of the independent approach measured the extent to which respondents
participated actively in community activities as a way to express their dissatisfaction with the
state of affairs in their settlement. The kinds of activities to which this indicator refers include:
seeking contact with community leaders, the private developer, local authorities or the
Municipality; organizing or attending community meetings; participation in community groups
such as women’s groups, environmental groups and savings groups; participation in training
workshops and small-scale infrastructure upgrading activities organized by NGOs; and

339

participation in committees to monitor construction of the new community housing (as in the case
of Borei Keila and the relocation site for Dey Krahom).
Analysis was undertaken on the basis of responses to three sub-indicators of
participation. The first sub-indicator asked respondents whether they felt a personal commitment
to help their community (C2.7). The second examined whether they had joined any community
groups (C1.18), while the third sub-indicator asked respondents whether they were directly
engaged in activities to help their community (C1.22). The three sub-indicators had a certain
degree of intentional overlap but they also tested different aspects of participation, including
stated commitment level (personal commitment), formal inclusion (joining community groups)
and actual participation (direct engagement). The combined sub-indicators had a reliability
coefficient of α = 0.81.
The four EVLN behaviors were associated with different levels of participation, guided
by the EVLN literature and the participation literature focused on slum dwellers. Voice was
associated with active engagement to seek change. This is in line with Hirschman’s description
of voicers as “kicking up a fuss” in order to “force improved quality or service upon delinquent
management” (Hirschman 1970, 30). In the context of informal settlements Sheela Patel and
Diana Mitlin (2004) suggest that this may be translated as participation “because you feel it is
worth it”—a form of participation motivated by residents’ urge to “stick their necks out” to help
improve their communities. It is adopted by people who become (informal or formal) catalysts
and leaders in the community (Patel and Mitlin, 234)
283
. In the Phnom Penh case study context
respondents were considered actively engaged and “sticking their necks out” if they could point

283
Sheela Patel and Diana Mitlin (2004) present a ranking of different forms of participation in informal
settlements in the Mumbai, India context based on “community perspectives on participation”. I borrow
selectively from these community perspectives in describing voice and loyalty behavior (the first two levels
in Patel and Mitlin’s ranking) in the Phnom Penh informal settlement context.

340

to concrete activities which they had organized, or in which they were actively involved. Such
efforts included community organizing, helping to prepare community action plans, monitoring
of redevelopment plans and construction as part of agreements between the community leaders
and the Municipality, and involvement in deliberations about the redevelopment plans, jointly
with the authorities, developers or NGOs.
The second level of participation is what Patel and Mitlin describe as participation
“because someone else persuaded you to jump in”. This is a passive form of participation that I
equate with loyalty: it is deployed by residents who do not initiate activities themselves but who
“agree to go along” and “follow someone else’s leadership” because these others have told them
what to do or because they can imagine benefiting from participating (Patel and Mitlin, 234). In
the EVLN literature the corresponding loyalty behavior is described as “passive support of the
system with little direct action to redress the problem giving rise to dissatisfaction” (Lyons and
Lowery 1986, 332). In the case study context this passive form of participation includes support
to community leaders, attending meetings about the redevelopment plans, contributing to small-
scale infrastructure improvement projects in the community, and site visits to inspect the new
community housing (in the case of Dey Krahom and Borei Keila). Participation in community
savings schemes was also included as loyalty behavior, although this categorization of individual
savers is more controversial (see Box 4).

341

When informal settlements in Phnom Penh are “organized” (and thus formalized) by NGOs, one of the first
activities to be established are savings groups—community run collectives that stimulate residents to
regularly save small amounts of money, and which make collective decisions about how to use the money.
Community savings groups in Phnom Penh were first set up in informal settlements by the people’s
organization SUPF in the mid-1990s, and later by UPDF, with structural support from the Thai housing
rights organization ACHR and its sister organizations within the worldwide Slum/Shack Dwellers
International (SDI) alliance.
Housing rights organizations claim that the importance of savings to poor urban communities goes well
beyond imparting financial discipline and pooling money: they help build collective management skills,
trust and self-reliance. Arjun Appadurai has argued that, in the ideology of the SDI alliance, savings
schemes have an “ideological, even salvational status” as the bedrock of urban poor federations, and as a
principal tool of self-mobilization and relationship building among people’s groups around the world
(Appadurai 2001, 32-3).
Viewed from this perspective, respondents’ participation in the savings schemes in their communities
would seem to be an act of voice par excellence—an example of their political participation and their
contribution to the mobilization of their communities.
But in practice, in the four land sharing settlements, the performance of the savings schemes has often been
lackluster. The schemes are run mostly by community leaders, and schemes’ success has depended on the
leadership and commitment of each individual leader. Many respondents claim that after the social land
concession announcement in 2003 the savings groups in their communities collapsed, as residents believed
there was no more incentive to save since they would be getting new housing units for free anyway.
Because of the low level of activity of most savings schemes in the case study settlements, and because of
the fact that most residents who participate in the schemes do not control or manage the schemes
themselves, but are dependent on community leaders to do so, I have categorized most participation in
savings schemes as loyalist behavior rather than as voice behavior, unless it demonstrated active behavior
in other community groups.
Lately, participation in savings groups in the case study settlements has become important for another
reason: as a proof of community membership and as a “ticket into the community”. The Municipality of
Phnom Penh determined that membership of a savings group was one of its proofs of community
membership, and a criterion for eligibility for new community housing. As a result, since then, savings
books have increased in value and are traded for several thousands of dollars between residents and
between residents and outsiders seeking to move into one of the social land concession settlements. For a
small minority of residents, the savings book has thus evolved from being an instrument of collective
empowerment to a passport for self-enrichment.
Box 4: Is participation in community savings schemes an expression of loyalty or voice?

Finally, I distinguished two forms of minimally active or non-active participation, which
I associate with neglect and exit. Neglecters are those who are dissatisfied with their lives in the
settlement but who see no options to improve these conditions. As a consequence, they passively
withdraw from the community around them and allow “conditions to deteriorate through reduced
interest or effort” (Rusbult, Farrell, Rogers and Mainous 1988, 601). In addition, some neglecters

342

may be excluded from the community as a whole, for whatever reason, or perceive themselves to
be excluded. As a result of their “alienation, cynicism and distrust” (Lyons and Lowery 1986,
332), neglecters were expected to have only a minimal propensity to participate and to
demonstrate only minimal personal commitment to help their communities. In the case study
context neglecters were expected to have little knowledge of community events and little interest
in participating in them. If they participated at all, neglecters were expected to do so irregularly,
not out of personal conviction but because of peer pressure or pressure from community leaders.
The inactive participation or non-participation of exiters differs from that of neglecters in
that it is due to their active engagement elsewhere, which means they are “too busy” to
participate, or to their lack of interest. Exiters by definition have little incentive to engage in
participation in the community because their behavior is not aimed at recovery of the situation
leading to their dissatisfaction; rather, they are scheming to leave the settlement altogether and to
pursue their own individual agenda. In the case study context exit behavior for this sub-indicator
is characterized by non-participation in collective structures except in cases where there is some
prospect of individual benefit (as with contributions to infrastructure improvements, for instance).
6.4.2 Indicator 2: Response to the Redevelopment Plans and Process
The second indicator of the independent assessment analyzed respondents’ strategies to
cope with the redevelopment plan and process in their settlement. This item went beyond merely
assessing a respondent’s level of satisfaction with the plans and process (as in items 4 and 5 of the
relative satisfaction measure; see section 6.3): rather, it sought to examine what—if any—were
the strategies of respondents to mitigate their dissatisfaction with the plans or process (in the case
of opposition to these plans) or to support or improve these plans (in case of acceptance).
Analysis was undertaken on the basis of responses to three separate, open-ended
questions that gauged attitudes toward the redevelopment plans in three different ways. The first

343

question (C3.5) asked respondents in general terms what they thought of the redevelopment plans
in their settlement. The second question (C3.6) asked respondents what their future plans were;
this item functioned as a practical test of respondents’ more general thoughts about the
redevelopment plans in the previous item. The third question (C4.2) gave respondents an
opportunity to add any of their own thoughts about the redevelopment plan and process and/or
about their household’s housing strategy. The third question was posed at the end of the
interview process and sought to obtain information about respondents’ views and plans in a
different (less direct) way than the first two questions, by letting respondents talk. The combined
measure had a reliability coefficient of α = 0.76.
I distinguished between four types of response to the plans, corresponding with the four
EVLN behaviors. Voicers (those “sticking their necks out”, to use the language of Patel and
Mitlin) were actively involved in trying to shape the redevelopment plans and process for their
community’s benefit, regardless of whether they were supporters or opponents of the plans. This
matches a critical aspect of voice in the theoretical literature, i.e., respondents’ determination to
“change, rather than escape from, an objectionable state of affairs” (Hirschman 1970, 30),
whether through individual or collective effort. In the case study context, respondents were
considered to be voicing if—as a result of their dissatisfaction—they articulated constructive
critiques of the redevelopment process, in an effort to improve, alter or overturn the process
.
. To
be considered as voicing, it had to be clear that the critiques were articulated to an external
audience (not just the interviewers), such as neighbors or other residents, community leaders, the
local authorities or the Municipality, or to “anybody who cares to listen” (Hirschman 1970, 4).
Respondents who passively supported the redevelopment plans and process “as is”,
without any preconditions or suggestions for improvement, and without undertaking any active
interventions, were identified as loyalists. I argue that this unconditional and passive acceptance

344

is symptomatic of the loyalist’s tendency to “stick with it” in the face of dissatisfaction, with the
expectation that “someone will act or something will happen to improve matters” (original italics)
(Hirschman 1970, 78).
Residents who showed neglect behavior, i.e., those who “dropped out” of community life
and felt no stake in the future of the settlement, were expected to have little motivation to try and
influence or support the settlement’s future redevelopment plan or process. For this reason,
neglecters feel resigned to accept the plan, not out of loyalty but because of a perceived lack of
better alternatives. Neglecters were expected to say that they “do not know” or “do not care”
about the redevelopment plan, and that (for those who were eligible for the new community
housing) they would go along with the plans but were unsure about their future plans.
Exit behavior in response to the redevelopment plans and process can be described as a
fairly predictable “market response” to dissatisfaction (Hirschman 1970, 19): in the exiters’
calculation the individual (self) interest weighs larger than the collective interest. Consequently,
exiters’ strategy with regard to the redevelopment plan is expected to be pragmatic: if they
perceive the redevelopment plan to be a “good deal” for them, they will stay part of the
community, but if by contrast the plan is considered unattractive and/or the redevelopment
process is beset by problems, then they will “vote with their feet” and make plans to leave the
settlement and pursue better options outside.
6.4.3 Indicator 3: Intention, Ability and Desire to Leave the Settlement
The third independent indicator examined the degree to which respondents were
intending to leave the settlement on their own initiative (not as part of an organized resettlement,
as in Dey Krahom, for example) and whether they had the ability to act on their intention, i.e.,
whether they had the financial means to leave as well as another location to move to. ‘Intention
to leave’ in this indicator was assessed in response to general conditions in the settlement (living

345

conditions, services, community networks and future prospects in the settlement and the vicinity),
not just respondents’ attitudes to the redevelopment plans (as in the previous indicator).
As an indicator of exit proneness, ‘intention to leave’ has some obvious problems. As
John Orbell and Toru Uno (1972) pointed out in their study of EVLN in the context of
neighborhood problem solving in the United States, intention to move is not necessarily the same
thing as actually making a move, as intentions may not be carried out. However, Orbell and Uno
cite studies in the U.S. context to show that, despite this limitation, “a high proportion of people
who indicate such intentions do act on them” (Orbell and Uno, 477). I assumed that in the Phnom
Penh informal settlement context, this finding would also hold true. However, even if not all
intentions are followed through, I argue that this indicator is nonetheless significant for
highlighting a respondent’s possible lack of long-term commitment to the settlement or to the
community—which is a key characteristic of proneness to exit. In addition, to compensate for the
possible gap in practice between intention to exit and actual exit, this indicator has two built-in
reality checks: first, ‘intention to leave’ is offset by ‘ability to leave’, and second, I make a
distinction between stated and unstated intention and ability to leave.
This measure consisted of five items, which assessed intention to leave and ability to
leave both directly (based on respondents’ answers to the direct questions) and indirectly (based
on facts about the respondents gleaned from other or indirect questions). Because of the potential
discrepancy between what some respondents claimed were their intentions and their actual
intentions, I made a distinction between “stated” intention and ability and “unstated” intention
and ability. The former was based directly on what respondents told the interviewers, while the
latter also took into account other known facts about the respondents as divulged directly or

346

indirectly by them in other parts of the interview
284
. Thus, one sub-indicator assessed intention to
leave based on a direct question about respondents’ current and former plans to leave (C3.1 and
C3.2), while a second sub-indicator assessed intention to leave based on additional indirect
information from open-ended questions (C3.6; C3.7; C4.2). Similarly, one sub-indicator assessed
respondents’ ability to leave based on a direct question about their financial means and other
residences (C3.3 and C3.4), while a second sub-indicator assessed ability to leave based on
indirect information from open-ended questions (C3.6; C3.7; C4.2). The “stated” and “unstated”
sub-indicators were strongly—but not perfectly—correlated with each other, thus demonstrating
that it is useful to make a distinction between the two kinds of information: data from stated and
unstated intention to leave had an inter-item correlation of 0.60 (p = 0.00) while data from stated
and unstated ability to leave enjoyed a correlation of 0.83 (p = 0.00). Finally, in order to enhance
the stability of the overall measure and to cross-check information from the four other sub-
indicators I added a fifth sub-indicator that assessed respondents’ stated desire to leave the
settlement (irrespective of whether they had the means to do so or not). The overall combined
measure produced a reliability of α = 0.79.
The categorization of respondents according to EVLN behaviors was undertaken on the
basis of all five items, including stated and unstated responses. In case of a contradiction between
stated and unstated intentions and ability, the unstated responses were given greater weight. On
this basis, respondents who made clear that they had the (stated as well as unstated) intention to
leave the settlement as well as the (stated as well as unstated) ability and desire to do so were

284
For example, when asked directly about their plans (C3.1), some respondents claimed that they had no
intention whatsoever to leave the settlement on their own initiative. However, in another part of the
interview, when they were asked about their future plans (C3.6), these same respondents admitted that they
were considering leaving if they could not pursue their business once they were living in the new
community housing, and that they might sell their new apartment units in order to raise money to buy or
rent other housing elsewhere. To account for this contradiction, these respondents were categorized as
“intending to stay” for ‘stated intention’ and as “prone to exit” for ‘actual intention’.

347

coded as exiters. Those who indicated that they were able to leave the settlement if they wanted
to, but did not have the intention or desire to leave, were coded as voicers. Respondents who
mentioned that they had neither the intention nor the desire or the ability to exit the settlement
were coded as loyalists, while those who indicated a desire to leave the settlement but no ability
or any active plan to act on this desire were categorized as neglecters.
6.4.4 Results of the Independent Assessment of EVLN
On the basis of the three indicators described above, the independent assessment of
dominant EVLN strategies among dissatisfied (non-ritually loyal) interviewees in the four case
study settlements
285
yielded the overall results presented in Table 12 and as illustrated in the
accompanying bar charts. In two cases respondents showed an even mix of EVLN responses and
it was not possible to determine a dominant EVLN strategy. These “non dominant strategies”
were treated as missing cases for the subsequent data analysis.
Three general trends stand out. First, loyalty was the dominant EVLN strategy for a
majority of respondents in all settlements. Second, active responses to dissatisfaction (exit and
voice) were less frequent than passive responses (loyalty and neglect). And third, taken as a
whole, voice was the least utilized behavioral response to dissatisfaction. Within these overall
trends the mix of dominant EVLN responses varied somewhat in each case study settlement; this
variation reflected the different redevelopment circumstances of each settlement.

285
For this and other statistical measures, data for Railway A and Railway B settlements were combined to
reduce the possibility of skewed results due to the relatively small sample size of Railway A.

348

Dominant response to
dissatisfaction*
All settlements Borei Keila Dey Krahom Railway A + B
N % N % N % N %
Exit 27 19.7 9 15.8 9 17.6 9 31.0
Voice 15 10.9 5 8.8 8 15.7 2 6.9
Loyalty 63 46.0 34 59.6 18 35.3 11 37.9
Neglect 30 21.9 8 14.0 16 31.4 6 20.7
No dominant strategy 2 1.5 1 1.8 0 0.0 1 3.4
Total 137 100.0 57 100.0 51 100.0 29 100.0
Stacked bars show % response per settlement; Data for Railway A & B settlements are combined.
*‘Dominant strategies’ were the most prevalent responses in at least two out of three indicators. ’No
dominant strategy’ comprises respondents who showed no dominant EVLN response.
Total sample of dissatisfied residents N = 137; Missing cases N = 0.
Table 12: Frequencies of dominant EVLN strategies for dissatisfied respondents

Voice and neglect were relatively more dominant in Dey Krahom than in the other three
settlements, reflecting the contested nature of the planned relocation of the population outside
Phnom Penh, which was opposed by many residents. By contrast, in Borei Keila, where
redevelopment took the form of land sharing on site—an outcome much preferred by residents to
relocation—loyalty was relatively more dominant there than in the other settlements. In the two

349

Railway settlements exit was a relatively more frequent response to dissatisfaction than in Borei
Keila and Dey Krahom (at 32 percent of cases for Railway A and B combined), mainly because
of the efforts of the developer over several years to entice residents there to relocate as the
stalemate in the two settlements continued.
Assumptions about EVLN responses by population sub-group (see section 5.5.5) were
tested using Pearson’s Chi square and cross-tabulation statistics. The results are presented in
Table 13.

Sub-group Assumed dominant response Strength of association
Community leaders Voice 0.74*
Renters Neglect 0.39*
Police families Voice 0.026
ns

Army families Voice 0.17**
Vietnamese minority Neglect 0.085
ns

Small business owners Voice 0.061
ns

Woman-headed households Loyalty 0.096
ns

Strength of association measured with Cramér’s V: 0 (no association) ≤ V ≤ 1 (perfect association)
Confidence intervals employed: * = p ≤ .01; ** = p ≤ .05; *** = p ≤ .1;
ns
= not significant.
Total sample of dissatisfied residents N = 137.
Table 13: Testing EVLN hypotheses by population sub-group

Only three hypotheses held as assumed and were statistically significant. As expected,
there was a very strong and statistically significant positive association between community
leadership and voice, and a reasonably strong and statistically significant positive association
between renters and neglect behavior. There was a statistically significant positive relationship
between army families and voice, as assumed, but this association is very weak. The assumptions
about dominant EVLN behavior for the other sub-groups could not be confirmed in the data, but

350

neither could they be comprehensively rejected due to the statistical insignificance of the data for
these variables.
The statistically strong results for community leaders and renters point to additional
noteworthy patterns in the data. One is that voice is heavily dominated by community leaders: 13
out of 15 respondents who were dominant voicers in the overall sample of dissatisfied residents
were community leaders. Conversely, this means that only 2 out of 15 dissatisfied respondents
who demonstrated dominant voice behavior were “ordinary residents” from other sub-groups.
In the case of renters, a surprising finding is that loyalty is the second most frequent
EVLN response type after neglect (at 35 percent of responses). This appears to contradict the
assumption of renters as primarily “withdrawing” and prone to neglect given their precarious
economic status and their insecurity of tenure. The prevalence of loyalty is reflected in a
statistically significant (though weak) positive association between renting and loyalty, with a
Cramér’s V coefficient of 0.15 and significance at the 90 percent confidence interval.
Table 14, with accompanying bar charts, presents the final tally of dominant EVLN
behaviors among dissatisfied respondents as measured by the three indicators of the independent
assessment of EVLN.

351

Indicator 1: Participation in
community events
Indicator 2: Response to
redevelopment
Indicator 3: Intention and
ability to leave the settlement
Frequency of response (N) Frequency of response (N)* Frequency of response (N)
Exit 14 Exit 27 Exit 27
Voice 19 Voice 23 Voice 3
Loyalty 72 Loyalty 61 Loyalty 61
Neglect 32 Neglect 25 Neglect 46
Bar chart shows frequencies of dominant EVLN responses for dissatisfied respondents;
Total sample of dissatisfied residents N = 137; *1 Missing case; Total missing cases N = 1
Table 14: Frequencies of dominant EVLN strategies as measured by three indicators of the
independent assessment of EVLN

In the following sections I will analyze these findings in more detail for the various sub-
indicators.

352

Results for Indicator 1: Participation in Community Activities
Response Item 1:
Commitment to
help community
Item 2: Join
community groups
Item 3: Directly
engaged to help
community
Dominant EVLN
responses for
Indicator 1*
N % N % N % N %
Exit 4 2.9 15 10.9 19 13.9 14 10.2
Voice 20 14.6 15 10.9 13 9.5 19 13.9
Loyalty 89 65.0 79 57.7 64 46.7 72 52.6
Neglect 24 17.5 28 20.4 41 29.9 32 23.4
Total sample of dissatisfied residents N = 137; Missing cases N = 0.
* Dominant EVLN strategies for respondents were identified in relation to other two indicators.
Table 15: Frequency results for component items of Indicator 1

Around two-thirds of dissatisfied respondents claimed that they participated in
community events in their settlements, as identified by the three items making up this measure
(commitment to help the community, joining community groups, and direct engagement to help
the community) listed in Table 15. But for most residents this participation and commitment was
limited to passive and constructive support that helped to preserve the status quo in the
community, and which I categorized as loyalty behavior. Three kinds of activities made up the
bulk of loyal participation: participation in community savings groups, joining occasional
community meetings and supporting occasional small scale infrastructure upgrading schemes, by
providing labor or small monetary contributions. Underlying these forms of participation were
passive but constructive attitudes towards the community, as typified by the most frequent
expressions, “We are waiting to help” or “We are always ready to help, but currently there are
few activities”, attitudes which Sheela Patel and Diana Mitlin (2004), in their classification of

353

levels of community participation in the Mumbai slum context, might describe as “following
someone else’s leadership”.
By contrast, only around 14 percent of all dissatisfied respondents indicated that they
were actively engaged in community activities that sought to improve the state of affairs in their
settlement—the kind of active participation associated with voice, which Patel and Mitlin might
characterize as “sticking one’s neck out”. These active forms of participation include community
organizing, rallying residents around issues of concern to the community, and establishing
contacts with the authorities or with other outside groups. With only a few notable exceptions,
active participation in community affairs was undertaken by community leaders, not by “ordinary
residents”. Seventeen of the 20 respondents (85 percent) who indicated that they participated
actively to help improve conditions in their communities were community leaders. The notable
exceptions included residents who actively opposed the relocation plan in the settlement of Dey
Krahom (see the discussion on the next item of this measure, following section).
Approximately one third of dissatisfied respondents indicated that they were not—or
rarely—involved in community events. Of these non-participating respondents, 70 percent
demonstrated neglect responses, including feeling that they were not invited to participate or join
community groups; not knowing or not caring about what happens in the community; or lacking
trust in the authorities or community leaders, but feeling helpless to do anything about it. Thirty
percent of non-participating respondents indicated that they were not very involved in their
communities because they were too busy with other things or because they were making plans to
leave the community in the future—attitudes that correspond with exit behavior.

354

Results for Indicator 2: Response to Redevelopment Plans
Response Item 1: General
perception of
redevelopment
Item 2: Future
plans
Item 3: Comments
on household
strategy
Dominant EVLN
responses for
Indicator 2*
N % N % N % N %
Exit 12 8.8 28 20.4 22 16.1 27 19.7
Voice 52 38.0 14 10.2 23 16.8 23 16.8
Loyalty 55 40.1 69 50.4 64 46.7 61 44.5
Neglect 18 13.1 26 19.0 28 20.4 25 18.2
Total sample of dissatisfied residents N = 137; Missing cases N = 0.
* Dominant EVLN strategies for respondents were identified in relation to other two indicators.
Table 16: Frequency results for component items of Indicator 2

Data for the second indicator show that respondents were considerably divided in their
perceptions of the redevelopment plans for their settlements. The first point of division
concerned respondents’ support or l