University of Southern California Dissertations and Theses

Essays on political economy of privatization

(USC Thesis Other)

Essays on political economy of privatization

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ESSAYS ON POLITICAL ECONOMY OF PRIVATIZATION

by
Shivendu Shivendu

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSTY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(ECONOMICS)

December 2008

Copyright 2008 Shivendu Shivendu

ii
Dedication

To my dearest Kamal and Mridu

iii
Table of Contents

Dedication ii

List of Tables v

List of Figures vii

Abstract viii

Chapter 1: Introduction 1
1.1 Global trends in privatization 6
1.2 Stylized facts about privatization 14
1.2.1 Scale and approach to privatization 14
1.2.2 Differences in privatization strategies and sequencing 14
1.2.3 Difference in Political Settings and Institutional 15
Framework
1.3 Political Economy of Privatization 16
1.4 Determinants of partial privatization 17
1.5 Literature review 18
1.5.1 Theoretical Approaches 20
1.5.2 Empirical Approaches 21
1.5.3 Partial privatization literature 23

Chapter 2: Political Economy of Partial Privatization: 27
Cross Country Analysis
2.1 Literature review 31
2.2 Determinants of Partial Privatization 36
2.2.1 Macroeconomic determinants of privatization 37
2.2.2 Political determinants of privatization 37
2.2.3 Institutional determinants of privatization 38
2.3 Empirical approach, variables and data 40
2.3.1 Empirical approach 40
2.4. Empirical method and model 59
2.4.1 Privatization proportion model 60
2.4.2 Transfer of control model 61
2.5 Results and discussion 62
2.6 Conclusion 79

Chapter 3: Political Economy of Partial Privatization 82
3.1 Introduction 82
3.2 Model 85
3.3 Benchmark case: No private benefit of public control 92
3. 4 Privatization contracts when institutional quality 93
is less than perfect
3.5. A Political economy model of the government 97

iv
3.6. Discussion and conclusion 101

Chapter 4: Political Economy of Privatization in India 102
4.1 Industry Structure in Colonial India 102
4.2 Evolution of Public Sector in India 104
4.3 Performance of public sector enterprises 108
4.4. Privatization as public sector reform strategy 111
4.4.1 Early reform phase 112
4.4.2 Disinvestment phase 113
4.4.3 Privatization phase 116
4.5. Economic performance of firms and privatization 120
4.6 Data 124
4.7 Empirical approach 126
4.8 Results and discussion 132
4.9 Summary and Conclusion 146

Bibliography 148

v
List of Tables
Table 1.1a: Proportion of public sector in GDP in different regions 7
Table 1.1b: Proportion of public sector in GDP in rich and poor countries 7
Table 1.2: Privatization revenue from developed and developing countries 9
Table 1.3: Privatization amount from developing countries (US$ million) 10
Table 1.4: Proportion of SOE assets privatized in selected 11
Table 1.5a: Large Privatizations in Developed Countries 12
Table 1.5b: Large privatizations in developing countries 13
Table 2.1: Share of private sector in GDP 29
Table 2.2: Classification of industries 43
Table 2.3 List of Selected Countries and number of privatizations 56
Table 2.4: Privatization proportion model with full sample 63
Table 2.5: Privatization proportion model for core sector 69
Table 2.6: Privatization proportion model for intermediate sector 73
Table 2.7: Privatization proportion model for competitive sector 75
Table 2.8: Multivariate model for control 77
Table 4.1: Change in Public Sector share in Industrial Production 107
from 1968 to 2000

Table 4.2: Public and private sector share in GDP 108
Table 4.3: Sector-wise performance of SOEs 109
Table 4.4: Privatization proportion and amount by year 121
Table 4.5a: Impact of partial privatization on SOE performance 133
Table 4.5b: Impact of partial privatization on SOE performance with 135
interaction

vi
Table 4.6a: Impact of level of privatization on SOE performance 137
Table 4.6b: Impact of level of privatization on SOE performance 140
Table 4.7: Impact of privatization on employment 141
Table 4.8a: 2SLS estimation of partial privatization impact on firm performance 142
Table 4.8b: 2SLS estimation of partial privatization impact on firm performance 143
Table 4.9: Impact of partial privatization on firm performance 144

Table 4.10: Impact of partial privatization on firm performance without 146
controlling for regulatory conditions with same years and sectors
as in Gupta ( 2006)

vii

List of Figures

Figure 3.1: Timing of contracting under privatization 92
Figure 3.2: Optimal privatization proportion as institutional quality declines 94
Figure 3.3: Optimal privatization proportion 95

viii

Abstract
Empirical observation of partial privatizations across countries is puzzling. This research
finds empirical evidence that privatization programs have not been driven by ideological
or “efficiency” reasons, but rather by the pragmatic cost-benefit tradeoffs made by the
politicians, and the economics of privatization often dominates its politics. Using data
from 43 countries on more than 4700 privatization transactions, I find strong empirical
support for Institutional Quality as consistent and significant determinant of proportion of
partial privatization. Surprisingly, countries having higher corruption tend to have higher
proportion of privatization in competitive sector, but lower privatization in core sector.
Counter to anecdotal evidence, this research finds that political constraints have no
significant impact on partial privatization proportion. Further, fiscal crisis drives
politicians to privatize, but has no significant effect on privatization proportion. These
empirical findings motivate a political economy model of privatization and three results
stand out: first, the distortion in the privatization proportion depends upon the
institutional quality parameter relative to a measure of private sector efficiency and the
distortion increases as institutional quality declines; second, the effort level of private
buyer firm declines as institutional quality declines. And third, under heterogeneous
preferences of citizens, the privatization proportion declines. The third part of this
research I study the impact of the privatization proportion on firm performance in India
and document two important and new findings. First, the privatization proportion has no
significant impact on firm performance when we control for regulatory conditions and
listing on stock exchange. Second, partial privatization has no negative impact on

ix
employment in partially privatized firms. We find significant empirical support for the
argument that the regulatory conditions and listing on the stock exchange are the sources
of improvement in the performance of partially privatized firms and not the privatization
proportion.

1
Chapter 1: Introduction
Privatization is a process that transfers the control of productive assets and
operations from public sector to private sector. It not only includes the selling of a
government owned enterprise to private firms, but also contracting out, leasing, private
sector financing of infrastructure and transfer of managerial control. Some even consider
encouraging the growth of new, small, private enterprises as a form of privatization
(Streeten, 1996, pp44). Though an increasing number of theoretical and empirical studies
in economics conclude that the outcomes of privatization are positive, public perceptions
of privatization are generally negative- and they are getting worse (Birdsall and Nellis,
2003). This makes the political acceptance of privatization as an effective reform strategy
difficult, and in democratic countries like India, the politics of privatization often
dominates its economics.
Although the term “privatization” gained wide popularity after the sale of British
Telecom in 1984, Chile was one of the earliest countries to begin privatizing in earnest.
Though the process of privatization took roots in the developed world and its rapid spread
was dominated by the industrialized nations in 1980s, the collapse of Soviet Union, the
demise of central planning in Eastern European countries and unimpressive performance
of public enterprises in the mixed economies, eventually led the transition economies and
developing countries to embark on widespread privatization
1
programs. Indeed, during
the 1990s, the transition economies led the way, by adopting a variety of mass
privatization techniques, in an attempt to create markets, and in a sense, to privatize
whole economies. For example, the average share of national GDP attributable to the

1
Some countries prefer to use different terminology: “disinvestments”, “restructuring”, “capitalization”,
“corporatization” etc.

2
private sector increased from 20% to more than 50% in Hungary, Poland and the Czech
and Slovak Republics over the three year period 1990-1993 (IFC, 1995).
While many countries, both developed and developing, have taken to privatization
as a core economic policy reform, two very large and fast growing economies, namely,
China and India, have adopted gradual and non-centric approach towards privatization.
China’s economic reforms, under the premise of preserving basic socialist structure, have
not engaged in any substantial degree of privatization. The Communist Party of China
believes that this reform approach can mitigate the impact of inequality that may arise
due to the reallocation of state assets and thereby avoid the conflicts arising from this
inequality ( Lin, 2003). India has also been a slow mover in privatization though its
commitment to privatization has escalated steadily since the economic reforms were
initiated under shadow of a severe balance of payments crisis in 1991. The government
started the privatization program in 1991 as a disinvestment program, (aimed at reducing
the government’s holding by up to 20 percent) in order to raise resources to reduce the
budget deficit. During 2000-2003 periods, the Government gradually moved to strategic
sale of public enterprises, but this process has again been halted after the Indian National
Congress- led coalition came to power with the support of Left oriented parties in May
2004.
This dissertation focuses on the political economy of privatization. Political
constraints have played a significant role in the formulation of privatization policies, the
method of privatization and the speed of privatization programs in both developing and
developed countries. Privatization policies were spurred by the widespread
discontentment due to the poor performance of SOEs and also the public demand to

3
reduce political corruption by limiting the direct role of politicians in running public
firms. Privatization was favored on the grounds that it will depoliticize economic
decision, reduce corruption and reduce the power of unions that come in the way of the
efficient organization of productive assets.
However, a number of economic as well as political difficulties in privatization
programs have also been identified. While the premise that the politicians benefit from
State Owned Enterprises is generally accepted as a true representation of the world of
politics, it is not clear why these same politicians would implement privatization policies
that would appear to be detrimental to their political interests.
In this research, we take a different approach to study privatization. We contend
that the government or policy makers can use alternative means to achieve their goals so
that that one policy can be substituted for another when confronted by a given policy
dilemma. For example, a government faced with large fiscal deficit in an election year
may face a choice between increasing taxes, imposing drastic cuts in social sector
programs and selling SOEs and using the revenue inflows to balance their budgets or
even launching new populist programs. This means that privatization programs may not
be driven by either ideological or “efficiency” reasons, but rather by the pragmatic
tradeoffs made by the politicians. Faced with the alternative of raising taxes or cutting
social programs and thereby facing opposition from large segments of the population, it
may be expedient for the politician to sell off a public firm and probably face opposition
from only a small number of employees. Therefore, in this research we contend that the
politicians’ (or alternatively governments’) willingness to privatize is determined by the

4
perceived cost-benefit ant the margin, that is, the degree to which the expected net
benefits or privatization are preferred to the available alternatives.
This research project is organized as follows. In this chapter we examine global
trends in privatization, closely look at literature in two specific aspects of this
phenomenon, namely, partial privatization and political constraints in the privatization
process. We also motivate the selection of India as a case study for testing the validity of
our empirical findings and theoretical model.
In second chapter we build our empirical model on two stands of literature. The
first strand relates to study of factors that influence privatization. While Bortolotti,
Fantini and Siniscalco (2004) find that privatization is more likely to take place in
wealthy democracies, encumbered by high public debt, but having established stock
markets, Megginson, Nash and Netter (1999) show that populist governments - as
opposed to market oriented governments- privatize for revenue raising purposes rather
than following a policy of under-pricing and targeting the median voter. The second
strand relates to political – institutional factors that influence privatization (Persson and
Tabellini, 2001; Milesi-Ferreti, Perotti and Rostagno, 2001). We use data from 56
countries, comprising more than 6000 privatization transactions over 12 year period from
1988 to 1999 to study the factors that influence the proportion of privatization.
In the third chapter we build a simple model of privatization in which the
politician considers the benefits and costs of privatization as a policy alternative. This
model builds on our empirical results from chapter 2, takes moral hazard
conceptualization from Bolton, Pivetta, and Roland (1997) and embeds political economy
component from Laffont and Martimort (2002). Using this model we attempt to explain

5
privatization programs implemented by different developing countries, especially in Latin
America.
Later in chapter four we undertake a case study of India and find that, contrary to
general belief, privatization programs are neither influenced by the electoral cycles nor
are they determined by the political orientation of the government of the day. It would
seem safe, therefore, to contend that privatization programs are driven by pragmatic
considerations rather than by any ideological commitment or strictly economic
consideration. We study the affect of partial privatization on performance of SOEs in
India and find that regulatory conditions play a significant role in some of the firm
specific financial performance measures. Chapter five concludes.
The objective of this chapter is four-fold. First, it provides an overview of global
trends in privatization, namely the transfer of control of privatization assets and
operations from the public sector to the private sector. Second, it focuses on two
important aspects in the privatization debate – the partial sale of shares of public sector
firms and the role of political constraints in the privatization process. Third, in this
context, it reviews the theoretical and empirical research at the microeconomic and
macroeconomics levels, and develops the research questions that are addressed in
chapters 2 and 3 of this dissertation. Fourth, I motivate the selection of India for the case
study of chapter 4, to validate the political economy dimensions of privatization brought
out in this research project. In this chapter we also take an overview of global
privatization and identify significant commonalities and differences across countries.
Section 1.3 examines the political economy of privatization in the global context and in
section 1.4 we review extant literature on determinants of privatization. Finally, section

6
1.5 provides motivation for study of privatization as an alternative policy instrument
available to the politicians.
1.1 Global trends in privatization: In last two decades increasing numbers of
countries, both developing and developed, have been implementing privatization
programs with varying degrees of success. While transition economies have privatized to
create markets and a new economic system, developing countries have undertaken
privatization programs as an essential part of the economic reform process, especially
where State Owned Enterprises (SOEs) had failed to live up to their developmental
objectives. Ninety-five countries implemented more than 2,600 privatization
programs/transactions between 1988 and 1993, yielding $271 billion (Shafik, 1996).
According to Megginson and Netter (2001), all governments raised more than $400
billion through public offers alone. This figure would be much higher if direct sales were
also taken into account. 1.1a and 1.1b show the changes in public sector share in GDP
between 1990 and 1999 for all the countries of the world, grouped by region and income
levels (as per World Bank classification), respectively.
Though the Margaret Thatcher government in the United Kingdom is usually
considered as the first government in the developed world to launch a large and sustained
deep privatization program, it was the government of Konrad Adenauer, that came to
power in Federal Republic of Germany (FRG) after the 1957 election, launched the first
large scale ‘denationalization’ program of the postwar era. Among developing countries,
General Pinochet’s military government in Chile (1973-1990) was the first to embark
upon large program in 1974 by privatizing all enterprises, except for utilities, that had
been nationalized by the Allende administration.

7
Table 1.1a: Proportion of public sector in GDP in different regions
Year
East Asia
& Pacific
Europe &
Central
Asia
European
Union
Lattin
America
Middle East
& North
Africa
South
Asia World
1990 11.1 17.8 20.0 13.4 19.9 11.7 15.7
1991 11.4 17.0 20.4 12.8 21.6 11.4 16.2
1992 11.5 16.2 20.8 12.4 19.9 11.1 16.2
1993 11.5 18.1 21.3 14.4 19.6 11.1 16.3
1994 11.2 17.2 20.9 14.7 18.5 10.5 16.0
1995 10.5 14.7 20.6 15.5 17.8 10.6 15.7
1996 10.6 14.3 20.7 15.1 18.2 10.5 15.6
1997 10.6 15.3 20.5 14.7 18.7 10.9 15.4
1998 10.9 15.7 20.1 14.6 20.3 11.5 15.4
1999 11.0 15.5 20.1 15.3 19.6 11.3 15.3

Source: World Bank (2001)

Table 1.1b: Proportion of public sector in GDP in rich and poor countries

Year
Heavily
indebted
poor
counrtries
(HIPC)
Least
developed
countries
Low
income
Lower
middle
income
Low &
middle
income
Middle
income
Upper
middle
income
High
income World
1990 14.0 11.3 12.4 13.3 13.9 14.2 14.8 16.2 15.7
1991 13.9 11.9 11.9 13.0 13.8 14.1 14.9 16.9 16.2
1992 13.9 11.9 12.1 12.6 13.5 13.8 14.6 16.9 16.2
1993 14.1 11.7 12.1 13.8 14.4 14.9 15.6 16.8 16.3
1994 12.7 11.1 11.5 13.5 14.1 14.6 15.4 16.5 16.0
1995 11.2 9.5 11.3 12.1 13.7 14.1 15.6 16.3 15.7
1996 11.0 9.0 11.2 12.1 13.5 14.0 15.3 16.1 15.6
1997 10.9 9.0 11.6 12.6 13.6 14.0 15.0 15.9 15.4
1998 11.4 10.0 11.4 13.3 13.9 14.4 15.3 15.8 15.4
1999 11.5 10.0 11.2 13.5 14.1 14.6 15.5 15.7 15.3

Source : World Bank (2001)

It was the British Telecom (BT) equity offering in 1984 that took privatization to
a global scale. That was the largest equity offering in history up to that time. The $4.8
billion issue created 2.25 million new shareholders in the UK, and the overwhelming
response to the trenches offered in New York and Tokyo demonstrated that a global

8
market for privatization existed. The successful privatization of BT removed misgivings
about the global demand for SOE equity, and many other countries followed. France’s
adoption of large scale privatization was dramatic as it signaled a sharp break from with
the dirigiste tradition of government intervention in industrial activities. The conservative
government of Chirac, which came to power in 1986, not only decided in favor of selling
off the industrial and financial groups nationalized by the socialists during 1981-1982,
but also decided to privatize the large banks that were nationalized by Charles De Gaulle
in 1945. Beginning in September, 1986, the Chirac government privatized 22 large
companies in the 15 months that it was in power for more that $12 billion. Though the
Socialists returned to power in 1988 and stopped further sales of SOEs, they did not try to
renationalize those companies. Later, when the conservative government of Balladur
came to power in 1993 and BNP, one of the largest banks, and Rhone-Poulence, one of
the largest pharmaceutical firms, were privatized.
In addition to UK and France, other developed countries also undertook large
privatization programs in the 1980s. The Japanese privatization of Nippon Telephone and
Telegraph (NTT) in an 1987 equity offering, yielded more than $72 billion and was
several times oversubscribed. The only major privatization through equity offering in the
US was the $1.65 billion Conrail issue in 1987, (also the largest equity issue the US till
that time) which though highly politicized, was successfully implemented. After 1987,
privatization programs also gained momentum in the developing countries, particularly in
Latin America. Though most of these programs relied on strategic sales; there were also
significant equity issues in Brazil, Chile, Mexico, and Venezuela. Privatization of
Mexico’s state owned telephone utility, Telephones de Mexico, in 1990 through a

9
combination of equity issue of $1.9 billion and a strategic sale provided an impetus to
subsequent economic reforms and increased foreign direct investment. Table 1.2 gives
global privatization proceeds in the last decade and close to one trillion dollar public
assets have been transferred to the private sector in the world as a whole. However, after
a decade of growth, global privatization proceeds in the year 2000 stood at around $100
billions down to around two third from the level attained in the year 1999 (OECD 2001).
Most of this decline came from the developed countries due to depleting inventory of
assets available for sale as privatization programs in these countries appear to have
matured; proceeds from developing countries have also declined from the peak level of
$66.6 billion in 1997.
Table1.2: Privatization revenue from developed and developing countries

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
Developed countries
Australia 19 1,402 1,893 2,057 2,055 8,089 9,052 16,815 7,146 15,220
France 12,160 5,479 4,136 3,096 10,105 13,596 9,478
Itly 759 3,039 9,077 10,131 11,225 24,536 14,497 25,594
Japan 13,874 2,039 4,544 6,641 15,115
UK 12,906 21,825 604 8,523 1,341 6,691 6,695 30,700
Others 7,635 1,371 4,058 8,772 15,314 3,440 32,236 86,700 47,973 32,404
Total
Developed
Countries 20,506 24,238 7,315 34,551 47,141 48,327 64,343 86,700 89,853 97,811
Developing
Countries 12,658 24,242 26,180 23,663 21,717 21,903 25,400 66,573 49,311 44,075
Global Total 33,164 48,480 33,495 58,214 68,858 70,230 89,743 153,273 139,164 141,886

Precentage
of developing
countries 38.17 50.00 78.16 40.65 31.54 31.19 28.30 43.43 35.43 31.06

Source : Compiled from OCED (2001) and World Bank (2001) (million US$s)

Although Chile was one the first developing countries to start privatizing, during
the 1980s a large number of Latin American countries joined in, and later in 1990s rapid

10
privatization techniques were adopted by the Eastern European and former Soviet Union
countries. Table 1.3
2
shows the revenue raised by developing countries through
privatization from 1990 to 1999. Latin America accounted for around 56% of the
proceeds, followed by Easter Europe and Central Asia, and East Asia and the Pacific,
which accounted for about 19% and 16% of the proceeds, respectively. South Asia and
Sub-Saharan Africa have not shown much enthusiasm in privatizing.
Table 1.3 : Privatization amount from developing countries (US$ million)

Region 1992 1993 1994 1995 1996 1997 1998 1999 Total

East Aisa
and
Pacific 5,161 7,155 5,508 5,410 2,680 10,384 1,091 5,500 44,099
Latin
America
and
Caribbean 15,560 10,488 8,199 4,616 14,142 33,897 37,685 23,614 177,839
Eastern
Europe
and
Central
Asia 3,626 3,988 3,957 9,742 5,466 16,537 8,003 10,334 65,466
Middle
East and
North
Africa 69 417 782 746 1,478 1,612 1,000 2,074 8,197
South
Asia 1,557 974 2,666 916 889 1,794 174 1,859 11,854
Sub-
Saharan
Africa 207 641 605 473 745 2,349 1,358 694 8,267

Total 26,180 23,663 21,717 21,903 25,400 66,573 49,311 44,075 315,722

% Share 8.3 7.5 6.9 6.9 8 21.1 15.6 14

Source : Compiled from World Bank (2001)

2
Derived from World Bank (2001)

11
Table 1.4
3
gives the estimates of proportion of State Owned Enterprise (SOE)
assets privatized in individual developing countries over 1998-98 (Ramamurti, 1996).
This table shows that some countries like Philippines, Peru, Argentina and Bolivia have
privatized much more intensely that others like India, Egypt and South Africa, and some
may have very little left to be privatized. The mere fact that the SOE assets shrunk by
more than 10% in mixed developing economies in a decade is a significant reversal of the
trend in SOE growth in the four decades following World Was II.
Table 1. 4: Proportion of SOE assets privatized in selected

Country
GDP ($
Billion)
Share of SOE in
GDP(%)
Proceeds from
Privatization ($
Billion)
Percentage of
SOE assets
privatized (%)
Phillppines 52.5 2.6 3.4 83.5
Peru 22.1 8.5 4.5 79.1
Argentina 228.8 5.5 18.3 48.3
Bolivia 5.3 11.4 0.8 44.7
Ghana 6.9 9.3 0.6 32.2
Malaysia 57.6 17.0 9.2 31.2
Mexico 329.0 13.1 27.3 21.1
Brazil 360.4 7.6 9.7 11.8
Pakistan 41.9 11.5 1.6 10.9
Sri Lanka 8.8 10.4 0.3 10.6
Turkey 99.7 10.1 3.1 10.2
Venezuela 61.1 19.4 2.5 7.1
Chile 41.2 15.7 1.3 6.7
Indonesia 126.3 15.8 4.0 6.7
Nigeria 29.7 13.9 0.8 6.1
Thailand 110.4 5.4 1.0 5.8
Morocco 28.4 17.7 0.9 5.7
India 214.6 13.5 4.5 5.1
Bangladesh 23.8 2.5 0.0 3.4
Egypt 33.6 27.2 0.7 2.5
Zambia 3.8 29.5 0.1 2.1
S. Africa 103.7 15.5 0.6 1.3
Tunisia 13.9 30.2 0.1 1.0
Total 2003.5 13.3 95.2 11.9

Source: Ramamurti (1999) table 3 (modified)

3
Ramamurti (1996), table 3 (slightly modified)

12
Another significant trend in the 1990s in privatization programs across all
countries has been that all economic sectors have been covered. Table 1.5.a and 1.5.b
give details of large privatizations during 1990-99 for industrialized countries and
developing countries, respectively. Almost all of these large privatizations have been in
‘strategic’ sectors that traditionally have been in the public sector, viz, utilities such as
electricity, telecom; natural resources, including petroleum and coal; railways; steel; and
financial sector, including insurance and banking.
Table1. 5a: Large Privatizations in Developed Countries

Country Sector State Enterprise Year
Porportion of
Privatization
Sale
amount ($
billion)
Australia Power Five distribution companies 1995 100% 6.2

Two power generation
units 1996 100% 3.8
Belgium Telecom Belgacom 1995 49.90% 2.5
Denmark Telecom TeleDanmark 1994 48.30% 3.3
France Petroleum Elf-Aquitaine 1994,1996 100% 8.8
Banking BNP 1993 100% 4.8
Insurance UAP 1994 100% 3.3
Steel Usinor Sacilor 1995 64% 3
Germany Telecom Deutsche Telekon 1996 26% 13.3
Itly Energy ENI 1995-96 31% 9.8
Insurance INA 1994-96 100% 6.1
Japan Railways JR East 1993 62.50% 8.9
JR West 1996 68% 4.5
Tobacco Japan Tobacco 1994,1996 33% 7.7
Netherlands Telecom KPN 1994-95 55% 7.8
Spain Petroleum Repsol 1989-96 90% 5
U.K. Telecom British Telecom 1984-93 100% 22.8
Power
National Power and Power
Gen 1991,1995 100% 10
Scottish Power 1991 100% 3.7
Railways Railltrack 1996 100% 3.2
U.S.A Telecom 102 PCS licenses 1995 100% 7.7
493 PCS licenses 1996 100% 10.2

Source: Guislain (1997), p.4.

13
Table 1.5b: Large privatizations in developing countries

Country Sector Public Enterprise Year
Privatization
Proportion
Revenue
($ billion

Argentina Telecom Telecom Argentina 1990-92 100% 1.8
Telefonica de Argentina 1990-91 100% 1.5
Electricity SEGBA, Hidronor, AEE 1992-96 51%-100% 3.5
Gas Gas del Estado: split into 10 firms 1992-95 70%-100% 2.9
Petroleum YPF 1993 61% 3
Brazil Electricity Light Servicos de Electricidade 1996 51% 2.2
Steel Usiminas 1991-94 90% 2
CSN 1993 81.90% 1.5
Cuba Telecom Ectesa 1994 49% 1.4

Czech Republic Telecom SPT 1995 27% 1.5
Hungary Electricity
Two power plants and six
distribution companies 1995 34%-49% 1.3
Telecom MATAV 1993-95 75% 1.7
Indonesia Telecom Indosat 1994 32% 1
PT Telkom 1995 19% 1.7
Malaysia Electricity Tenga National Berhad 1992 23% 1.2
Mexico Banking Banamex 1991 70.70% 3.2
Bancomer 1991 62.60% 2.5
Telecom Telemex 1990-94 100% 7.5
Railways Northeast Line 1996 80% 1.4
Peru Telecom CPT and Entelperu 1994-96 95% 3.1
Singapore Telecom Singapore Telecom 1993 11% 2.7
Taiwan Steel China Steel 1989-95 53% 3
Venezuela Telecom CANTV 1991-96 100% 2.9

Source: Guislain (1997)

The sectoral distribution of privatization revenue in developing countries during
1990-99 has been heterogeneous across sectors.. The infrastructure sector including
telecommunication and power accounted for almost half of the total proceeds during the
decade, while the primary sector, consisting of oil and gas, accounted for 19% followed
by the manufacturing sector, consisting of chemicals and steel at about 16% and the
financial sector at about 12%. Another noticeable feature during the 1990s was that,
while forty percent of total privatization revenue in all developing counties came from

14
foreign participation, predominantly in the form of direct investment, Latin America had
the highest share of foreign participation at 55%, followed by Eastern and Central Europe
at 24% and Asia and Pacific at 15% (Naib, 2004, pp163).
1.2 Stylized facts about privatization: Privatization of SOEs has been an integral
part of the major economic reforms undertaken by Many developed and developing
countries from the early 1980s onwards. At the end of the 1980s the collapse of Soviet
Union led to an attempt to create markets in Eastern Europe, and, in a sense to privatize
whole economies.
1.2.1 Scale and approach to privatization: The scale and approach to privatization have
varied considerably within and across developed and developing countries. Even within
the OECD, while Sweden, Spain and Holland have gone for only minor experiments with
privatization, Japan, New Zealand and UK have undertaken significant programs. With
regards to approach to privatization there have been significant differences across
developed countries. While the UK privatization program has been driven by the
ideological zeal found in Eastern Europe, most of the developed countries have been
selective and pragmatic in their approach to privatization. On the other hand, Eastern
European countries and Russia rushed towards mass privatization programs after the
collapse of communist regimes in early 1990s. Mass privatization programs were used as
an instrument to create irreversibility in the transition process by giving the domestic
populations stakes in the success of privatizations (Boycko, Shleifer, and Vishny, 1994).
1.2.2 Differences in privatization strategies and sequencing: The privatization
programs of different countries have been implemented by adopting different strategies
such as the following. Public offering of full or partial shares: The government sells all or

15
a large proportion of the stocks of a wholly or partly owned public firm to the general
public. The public offering may be on a fixed price, band price or tender price and may
be open internationally or only domestically. This privatization strategy is usually
adopted for large profitable public firms and is often preceded by restructuring. The
advantage of this strategy is that it allows for widespread shareholding, which is
characterized by transparency and openness. This makes this strategy politically more
acceptable.
1.2.3 Difference in Political Settings and Institutional Framework: Political constraints
are important in any privatization program since there are bound to be both winners and
losers, even if there is agreement that, on balance, aggregate welfare is enhanced. While
these constraints matter ex ante, to convince voters to commit to a privatization program
with a somewhat uncertain outcome, ex post, these constraints may lead to the reversal of
privatization transactions (Roland, 1994).
Political constraints appear to have played a key role in the actual design and
sequencing of privatization programs all over the world. Boycko, Shleifer and Vishny
(1994) hypothesized that the prime justification for giveaway of state enterprises in
Russia was not economic but political in that the political process would have rejected
any other privatization program at that time. Therefore, it’s crucial to emphasize the
centrality of the political process in privatization. No meaningful insights can be gained
by separating the economics and politics. The tendency to take economics as endogenous
but politics as exogenous goes back to Piogovian tradition. Weingast (1997, pp5) refers
to this tradition by noting “an ironic aspect of the economists’ position is that they want
individuals to pursue their self interest in markets but not in politics.”

16
1.3 Political Economy of Privatization: Privatization can be viewed as a policy
instrument for reducing the impact of political factors on economic incentives and
performance (Vickers and Yarrow, 1991). More specifically, private ownership reduces
political influence and increases the capital market influence, though the precise impact
of ownership transfer will depend upon the specific characteristics of the relevant
regulatory environment and capital market. The case for privatization is based on the
implicit assumption that there is “government failure” to the extent that the public
ownership is likely to operate in such a way as to impede the efficient functioning of the
markets.
On the other hand, the policy environments in which firms operate are influenced
by the incentive structure faced by policy makers, which depends on a number of
political and institutional factors. These factors include interest groups, the degree of
regulatory discretion, and transparency in public decision-making, enforceability of
contracts, and the institutional framework for oversight, the balance of political forces,
and bureaucratic structure and goals.
The market for regulation
4
is itself imperfect and can produce an inefficient level
of government intervention, and therefore regulatory reform is an integral part of any
successful privatization process. The SOE operating in a competitive market can be
transferred to the private sector fully, cutting off all links between the firm and the
government, and therefore replacing the monitoring activities of the politician and the
bureaucrat by the capital market. But, in case of many public enterprises, the state
influence or intervention may persist in response to perceived market failures and/or to

4
We use the metaphor introduced by Yarrow (1999).

17
politically undesirable redistributive effects that might result as a consequence of
privatization.
Political constraints have played an important role in the formulation and
implementation of privatization policies and processes. In implementing privatization
programs, the fear of the layoffs of employees of privatized SOEs, and the withdrawal or
gradual reduction of subsidies from goods and services produced by the public sector can
be potential sources of political opposition.
In addition to issues of redistributive justice, national sovereignty, strategic
national interest, development goals, and unemployment, politicians forfeit an important
means of generating political support when they privatize and thereby lose the ability to
provide government employment and/or lucrative contracts to the their supporters. This
raises two questions. First, why do politicians take the policy decision to privatize when
they are the biggest losers in terms of loss of control over productive assets? And second,
given the difficulties associated with privatization, what explains significant differences
in privatization policies and approaches across countries? Subsection 1.5 reviews the
literature focused on these two questions. Subsection 1.6 outlines both the empirical
approach and theoretical structure used to address them in chapters 2 and 3, respectively.
1.4 Determinants of partial privatization: Privatization transactions in
developing countries have often been partial as pointed out by Jones et al. (1999) who
analyze 630 privatization transactions from 59 countries during the period 1977-1997 and
report that, of the initial 384 privatization offers, 88.5% were partial sales. Further, of the
11.5% in which the firms were completely sold most of them were in developed
countries. D’Souza and Megginson (1999) compare the pre and post-privatization

18
performance of 78 firms from 25 countries, including 10 developing countries that were
privatized between 1990 and 1994 and report that only 34 firms (44%) were privatized by
more than 50%. In another study, Boubkri and Cosset (1998) look at 79 companies from
21 developing countries that were privatized between 1980 and 1992 through public
offerings. They report that only 20(24%) were fully privatized. More recently, after
initiating a modest privatization program in 1991, only 11(21%) of the 52 firms that India
privatized by 2004, had more than 50% of the equity shares privatized (GOI, 2004). This
indicates that partial privatization has dominated over the full sale of SOEs across
countries in last two decades, and appears to be especially prevalent in developing
countries. In this research, we examine more than 6000 privatization transactions
between 1989 and 1999 from 68 countries and find that more than 75% of such sales
were partial.
1.5 Literature review: Considerable literature in microeconomics has looked at the
relationship between efficiency and firm ownership and whether or not the decision
process of the firm is distorted under government ownership
5
. This can be studied by
looking at the optimization problem: the objective, the constraints and how these are
affected under different ownership structures. Within this literature, it has been
theoretically established that, under conditions of perfect competition, the absence of

5
See, for example Vickers and Yarrow (1988); Stiglitz (1998); Laffont and Tirole (1993,
ch.17); Willing (1993); Galal, et al (1992); Tirole (1994); World Bank (1995); Shleifer
and Vishny (1994, and 1997); Schmitz (1994), Schmidt (1996,2000); Perotti and Guney
(1993); Hart, Shleifer and Vishny (1997); Shleifer (1997).

19
information problems and the existence of complete contracts, ownership should not
matter (Sappington and Stiglitz 1987).
The original arguments in favor of public ownership were justified as a solution to
the failure of the first of these conditions: the market failure justification. The possibility
of exploitation of monopoly power by a private firm in a “natural monopoly” setting
created the need for public ownership. This perspective on government firms has been
called the social view as it takes into consideration social marginal costs.
The other stream of literature studies the asymmetric information issue within
firms and considers the contracting and incentive problems within firms as the key to
increasing efficiency at the macroeconomic level. This perspective is termed the agency
view. Within the agency view there are two perspectives on the reasons for the existence
of poor incentives for efficiency under public ownership. First, the managerial
perspective informs us that, since monitoring is poorer in SOEs, the incentives for
efficiency are low-powered (Vickers, and Yarrow, 1991); the impossibility of complete
contracts plays a fundamental role in explaining why ownership matters in determining
firm level efficiency ( Sappington and Stiglitz, 1987). Second, the political perspective
argues that the source of inefficiency in SOEs lies in the political interference that distorts
the objectives and constraints of public firm managers (Shapiro and Willig, 1990;
Shleifer and Vishny, 1994).
Within these two broad streams of literature, empirical observations
demonstrating the predominance of partial privatization, especially in developing
countries, have not drawn much attention. Two questions are still open. First, why do

20
governments adopt partial privatization policies? And, second, does partial privatization
improve firm level performance across countries?
1.5.1 Theoretical Approaches: Public ownership was initially argued as an answer
to market failure. Production technology characterized by decreasing average costs in the
relevant range does not justify more than one firm on efficiency grounds and public
ownership was justified in these kinds of product markets, termed “natural monopolies”.
Subsequently, the literature has increasingly focused on the relevance of
asymmetric information and contract incompleteness. The market failure argument has
weakened for two reasons; first, the formal analysis of information problems and
incompleteness has shown that efficiency losses in public ownership are non negligible;
and, second, the scope of competition has become larger with the increasing size of the
economy, liberalization of international trade and technical developments. This has led to
the growth of regulation literature positing regulatory mechanisms as second-best
solutions and showing that there is an alternative to public ownership
6
, though Laffont
and Tirole (1993) show that under certain conditions, even public ownership may be
preferred over regulated private firm if the differences in efficiency can be traced to
different institutional, information and incentive structures rather than to the intrinsic
taste differences of mangers between public and private firms.
In the privatization debate, some have argued that the driver of microeconomic
efficiency is competition and, therefore have put ownership at a lower level of
importance in economic policy prescriptions (Stiglitz, 1993; Vernon-Wortzel and
Wortzel, 1989). Though it is true that efficiency gains depend on deregulation policies

6
See Laffont (1994) for an overview of regulatory literature

21
and contestability in the market, the presence of government firms in the market may
deter private firms from entering. Competition not only implies free entry, but also free
exit, if a firm fails. In markets with government firms, private firms may fear that the
government will not allow public firms to fail even if it is inefficient, thus making free
exit a non-credible commitment for public firms.
The theoretical literature studying the macroeconomic effects of privatization is
not as rich as those dealing with microeconomic effects, though it is getting increasingly
more attention in context of transition economies. While there are few models that link
reform policies like privatization with macroeconomic performance
7
, there are a number
of country studies that analyze interaction between macroeconomic performance
indicators and privatization policies
8
. Probably the reason for the limited number of
studies in this area is the fact that it is hard to distinguish the effect of privatization from
other economic policies and other events happening in the economy. We can at best
observe certain trends, but it is difficult to establish causality.

1.5.2 Empirical Approaches: The empirical literature can also be classified into
two categories; those looking at microeconomic efficiency at the firm level and those
analyzing macroeconomic effects of privatization. Within the first category, the literature
can further be grouped as follows.

7
See, Roland and Verdier (1994), Blanchard (1997) and Biais and Perotti (1998).

8
See, Marcel (1989), Lefort and Solimano (1993), Lopez-Desilanes (1993), and World
Bank (1995).

22
(i) Firm-specific case studies that focus on evolution and efficiency before and after
privatization.
(ii) Country-specific, cross-industry studies that examine the performance of firms in
different sectors, before and after privatization.
(iii) Cross country studies for firms that are traded on the stock exchange to compare
financial performance before and after privatization.

Most of the empirical research that compares the efficiency of public and private
ownership has been in the form of cross sectional studies of both types of firms in the
industries in which they coexist. Villalonga (2000) provides an excellent summary of
some 150 studies in different industries relating to different countries and finds the
cumulative evidence to be inconclusive. The diversity in results can be attributed to
differences in both the market structure of each industry in which the studied firms
operate, and the measure of efficiency used in the study. As recently as in 1997, Martin
and Parker (1997) conclude from their survey that “on balance it seems that neither
private nor public sector production is inherently or necessarily more efficient” (their
emphasis; p.93), In contrast to the availability of large numbers of cross sectional studies,
very few longitudinal studies are available. The evidence from these longitudinal studies
about the increase in efficiency of the firms after privatization is also not conclusive.
While studies carried out in the developed world
9
generally indicate improved
performance of the firms after privatization, the conclusions from the studies carried out
in developing world (which according to the World Bank’s classification excludes the

9
See Megginson et. Al (1994), Frydman, Gray, Hessel and Rapaczynski (1997) for empirical studies and
Foreman-peck and Manning (1988) and Galal, Jones, tendon and Vogelsang (1994) for case studies.

23
transition economies of Eastern Europe from this category) are mixed
10
. If private
ownership were the primary source of efficiency, then the state of economic development
and institutional setting of a country would not influence the outcome of the privatization
process. Eliat, Sachs and Zinnes (2001) analyze the economic performance of transition
economies and show (1) that the gains from privatization programs accrue only when
change-of-title reforms are accompanied by key institutional reforms, and (2) initial
conditions matter in determining the economic benefits of privatization. It is by now well
recognized that the existence of well functioning markets supported by institutional and
regulatory capacity is a pre requisite for successful privatization.
1.5.3 Partial Privatization Literature: We observe that most of the
privatization transactions have been only partial in that governments have retained some
equity share in the privatized public firm. Stiglitz (1991) points out that while a major
disadvantage of privatization is the loss of potential rents from the government, a major
advantage is the enhanced commitment of government not to subsidize; that is hardening
of the budget constraint and efficiency gains from private management. An optimal
solution may not necessarily be a “corner solution” (all private or all public) in situations
where the government would balance these advantage and disadvantages and because of
uncertain outcome it would seem reasonable for the government to retain large minority
equity stakes.
Perotti (1995) attempts to explain why there is such overwhelming empirical
evidence that the transfer of ownership occurs very gradually. They motivate this
empirical observation on the basis of investors’ concern about the future interference of

10
See World Bank (1994), John Weiss (19995), Millard (1998), LaPorta and Lopez-DeSilanes (1998) and
Boubakri and Cosset (1998).

24
government and accordingly their model predicts that a government averse to
redistribution retains a passive stake in the firms signaling willingness to bear residual
risk. Cornelli & Li (1997) find that an optimal privatization scheme may require that a
portion of shares be sold to screen-in the most efficient investors since the highest bidder
may not be the most efficient firm and their bid may reflect only that firm’s high private
benefit of control. Since the government’s objective is a combination of efficiency and
revenue and potential buyers are heterogeneous in both these dimensions, these two
objectives conflict across buyers, and therefore using only two instruments, namely price
and the probability of winning, is insufficient to select the most efficient buyer and
government uses the proportion of shares sold as a third instrument to screen the best
buyer.
Lewis & Sappington (2000) examine the case when governments in developing
countries undertake the privatization of public sector units, face domestic buyers whose
wealth is small compared to the value of the enterprise. In their model, when potential
buyers are wealth-constrained, the government optimally retains a significant proportion
of shares in the privatized firm. This is despite the fact that this limits the buyer’s
incentive to operate the firm efficiently, in order to increase revenue from the
privatization transaction. The government uses the proportion of shares sold as an
additional instrument to screen wealth-constrained potential buyers who are
heterogeneous in unobserved ability.
The government may choose to implement a partial privatization program in order
to exercise ongoing control over non-contractible characteristics of the firm’s output or
performance, such as the quality of service (Vicker & Yarrow, 1991; Hart et al., 1997) or

25
to secure some revenue in the short run while waiting until significant uncertainty about
the true value of the public firm has been resolved before finally selling off entire stakes
(Jenkinson & Mayer, 1998; Blanchard et al., 1991). In Partial privatization, the
government may also signal that it has favorable private assessment of the true intrinsic
value of the public firm.
As the privatization decisions are taken by the politicians, it is important to
understand the context in which these decisions are taken. Political constraints have
played an important role in the design of privatization programs, especially in developing
countries. These political constraints may be ex ante or ex post. While ex ante constraints
refer to feasibility constraints refer to feasibility constraints that block implementation of
privatization programs; ex post constraints refer to backlash and reversal after
privatization has been done and the outcome has been observed. Roland (1994) posits
that ex ante and ex post political constraints must be dealt with differently.
Ex ante constraints can be dealt with by compromising on reform plans, providing
credible compensation for losers or implementing privatization programs only gradually.
Ex post political constraints can be addressed by creating irreversibility in the
privatization process. Further, though ex ante and ex post political constraints will be the
same in a deterministic world, they will differ significantly, when the outcomes of the
privatization program are uncertain (Fernandez & Rodrik, 1991). Biais & Perotti (2002,
1997) claim that governments with right wing orientations implement mass privatization
programs in order to enhance their chances of re-election, because, when voters become
stockholders, they oppose redistribution through re-nationalization. Schmidt (2000)

26
argues that mass privatization takes care of ex post political constraints by signaling
commitment against policy reversals, thus generating ex ante support for privatization.
In the next chapter we undertake an empirical analysis to understand the
phenomenon of widespread partial privatization across countries. Later, chapter 3, we
develop a simple theoretical political economy model to explain the empirical
observations of chapter 2.

27
Chapter 2: Political Economy of Partial Privatization: Cross Country
Analysis

The current wave of privatization has come after a rapid expansion of state
ownership in most developed and developing countries in the forty years following
World War II. State owned enterprises (SOEs) were created in response to a number of
political and economic factors. While in the developed countries SOEs were primarily
viewed as a remedy for market failures such as externalities and monopoly, in the
developing countries these justifications were coupled with arguments of social equity,
economic independence and self-reliance. SOEs were also created as a direct instrument
to channel investment in priority sector as perceived by the policy makers, to build
adequate social and economic infrastructure, to implement full employment economic
policies, and to promote balanced regional development. But, at the same time, SOEs
often became an instrument in the hands of politicians to create political constituencies
and were used to reward supporters through inefficient allocation of investment, partisan
selection of managers and over employment.
By the 1970s, the performance of SOEs came under increasing scrutiny due to
widespread inefficiency, mismanagement, corruption and political interference in their
working. Many policy makers, business leaders and economists argued that a transfer of
assets from the state sector to the private sector would unequivocally improve the
economic performance of firms that were once under government ownership. While
economic theorists contended that private ownership was better than state ownership in
pursuing goal of economic growth, others argued that privatization would reduce fiscal
deficits by ending socially costly subsidies, and by reducing corruption. Thus,

28
privatization was spurred in 1980s and 1990s by inefficient functioning of SOEs, and the
widespread public demand to eliminate political corruption.
Other reasons for such a policy reversal in the public utilities sector can be traced
to technological changes and theoretical developments in the area of regulation that made
it possible to introduce competition into areas formerly thought to be natural monopolies
such as telecommunication, and power generation and distribution. Further, globalization
led to growing integration of markets, requiring firms to adopt flexible business
strategies. This resulted in SOEs getting less competitive in global markets due to their
organizational structure and thereby inducing governments to privatize to provide more
flexibility to SOEs. Ideological debate on privatization has also mellowed since the
collapse of Soviet Union and has put an end to confrontation between the socialist and
capitalist models of development. The narrowing of ideological differences has facilitated
a more pragmatic approach to economic policies, of which privatization forms a part. The
pro-private sector orientation of multilateral organizations like World Bank and
International Monetary Fund has also promoted privatization programs in developing and
transition countries.
These factors have led to a wave of privatization that was hard to foresee
only fifteen to twenty years back. More than 100 countries, across every continent have
launched privatization programs and government assets worth more than US$1300bn has
been sold in more than 8,000 transactions up to the end of 2002 (World Bank, 2003). The
extent of privatization has been most dramatic in former Soviet republics and in Eastern
European countries where complete economies have been privatized as shown in Table
2.1 below.

29
Table 2.1 : Share of private sector in GDP

Country 1980 1988 1994 1997
Czech Republic <1.0 <1.0 65 75
Hungary 3.5 7.1 55 75
Poland 15.6* 18.8* 55 65
Romania 4.5 …. 35 60
Russia <1.0 <1.0 50 70
Slovakia <1.0 <1.0 55 75
United States 79.4 79.6 81.1 82
* Almost exclusively from agricultural production (Slay,1993)
Sources : European Bank for Reconstruction and Development, 1997
and Bureau of Economic Analysis, 2000.

Even though the global trend towards privatization has been
overwhelming and appears to follow a uniform pattern, two aspects are increasingly
evident. First, privatization programs have not been homogeneous. The extent of
divestiture, and method, pace and sequencing of privatization have varied significantly
across regions and countries. While Latin American countries have followed more
comprehensive privatization programs, South Asian countries have been particularly slow
and sporadic in their privatization efforts. Within Latin American countries, while Chile
went through the cycle of privatization - re-nationalization - privatization, Brazil has
followed a more cautious and gradual privatization program.
Second, while economists tend to converge on the efficiency benefits of
privatization, the public perception of privatization programs has been generally negative
– and getting worse (Birdsall and Nellis, 2003). While more than two thirds of Russians
surveyed in 2001 perceived themselves as losers from privatization; only 5% perceived to
have gained from it. In Mexico President Vincente Fox has failed to begin privatization
of the energy sector that had been promised during election campaign; and in Ukraine,
the tension between pro-and-anti privatization forces among the government and the

30
parliament stalled the whole program. Thus, there appears to be a widening gap between
the favorable technical assessments of privatization programs by economists in general,
and growing negative public perception about the benefits of these programs.
This brings us to two important questions. First, what factors can explain the
divergence in privatization programs, especially with regards to degree of partial
privatization. Bortolotti and Faccio (2004) report that divestitures in which the majority
of the firm’s capital is sold represent only 17% of major public offerings during 1977-
2003 periods out of 3,813 transactions reported globally. If policy makers undertake
privatization programs in order to maximize social welfare by transferring assets to the
private sector, which is believed to be more efficient, then why do we find most
privatization transactions to be partial?
Second, why are public perceptions of privatization programs so negative and
getting worse? Improved efficiency of SOEs should lead to improved quality of products/
services and lowering of costs, which in turn should benefit population, then, why do vast
majority of people oppose these policies? Thus, privatization programs come with
political costs. This suggests that the governments, even if they believe in efficiency
gains due to private ownership, take privatization decisions based on larger set of
variables, both economic and political, than only efficiency considerations. Since a
policy of privatization is an economic as well as a political issue, we seek answers to
these two questions, first by identifying those macroeconomic, political-economic and
institutional factors that would lead to and affect the extent of partial privatization and,
then using data from 4,766 worldwide transactions to determine the extent to which the

31
country specific differences in these determinants affect government approaches to partial
privatization.
Objectives of this chapter are threefold. First, it provides an overview of literature
related to our research question as outlined above, and identifies the prospective
macroeconomic, political-economic and institutional determinants of partial privatization.
Second, it describes our hypotheses to be tested, data, and empirical approach. Third, it
presents results of our analysis and its theoretical implications that motivate the model in
chapter 3.
This chapter is organized as follows. Section 2.1 presents brief theoretical and
empirical literature review, section 2.2 presents determinants of privatization. Section 2.3
describes data, defines variables, and motivates the empirical approach. Section 2.4
contains empirical results, and its theoretical implications. Section 2.5 concludes.
2.1 Literature review: Perotti (1995) provides one of the first theoretical models
attempting to explain the empirical observation of preponderance of partial privatizations,
contrary to the full privatization predictions based on only efficiency considerations. This
model explains partial privatization by arguing that a government averse to redistribution
retains a passive stake in the firm indicating its willingness to bear residual risk and
thereby signaling its commitment to allay investor’s concern about future interference.
This model suffers from two limitations; First, as admitted by authors themselves, is the
exclusion of any rationale for public ownership in the analysis, Second, it does not
address the observed worldwide phenomenon, in which a majority of privatization
transactions were such that government retained majority control in the firm (Bortolotti &
Faccio, 2004). For example, in the case of India, prior to 1999 the stated objective of the

32
privatization policy
11
was to sell only minority shares in SOEs, without any transfer of
control. As a result, more than 80% of the country’s privatization transactions have been
of minority shares (Department of Disinvestment, Government of India, 2004).
Another theoretical model was provided by Cornelli & Li (1997) who argue that
government may use proportion of shares sold (partial privatization) as a screening
mechanism when bidders have “private benefit of control” and the highest bidder in an
ordinary auction may not be the most efficient. As stated by authors themselves, this
model is motivated by the participation of foreign firms in privatization transactions but
as such does not explain overwhelming evidence of partial privatization across countries,
even in which foreign firms are not involved. Further, this model takes government as
social welfare maximizer and thus completely ignores the political nature of privatization
process.
Lewis & Sappington (2000) show that when buyers are wealth constrained and
are unable to make substantial initial payments for large proportion of shares in the public
firm, then to increase its financial payoff from the privatization, the government
optimally retains a large stake in the public firm, even though this limits the private
buyer’s incentive to run the firm efficiently. Though this model explains some of the
divergence across privatization transactions, it assumes that that even with partial
privatization operational control is transferred to the buyer and government only shares
profits. This does not stand the test of empirical validation as Bortolotti and Faccio
(2004) study the evolution of control structure of 141 privatized firms from OECD
countries over the period from 1996 through 2000 and find that through various cross

11
Referred in government papers as “disinvestment policy”.

33
holding mechanisms, governments control 62.4% of the privatized firms as of end of
2000. Further, this study finds that large government ownership in privatized firms has no
negative impact on either adjusted market value or stock price performance. In another
recent theoretical work Schmitz (2004) adopts the incomplete contract approach, and
show that partial privatization mitigates the disadvantages of public ownership and of
privatization. On one hand, in the case of full government ownership there are no
incentives to improve quality if the manager invests or too strong incentives if the
government invests, and on the other hand in the case of full privatization there are too
strong incentives for the manager to reduce costs. Though this model studies some of the
tradeoffs inherent in privatization under asymmetric information, it does not take political
economy dimensions into consideration.
In addition to this economics literature, finance and political science specialties
have also addressed this phenomenon, though in a different context. In Finance,
literature on the initial public offerings (IPOs) bears resemblance to privatization
transactions. Hayne, Leland and David Pyle (1977) show that a partial sale by an
informed owner is a reliable signal of high value, and Mark Grinblatt and Chuan Yang
Hwang (1989) show that promoters may use under pricing of initial public offering
(IPO) to signal high value. Zingales (1995) shows that initial owner may retain some
fraction of ownership in order to balance the maximization of value of cash flow rights
with maximization of value of control rights. In this model, an initial owner in deciding
whether to undertake an IPO, and also as what fraction of ownership to retain balances
two factors: by selling to dispersed shareholders, he maximizes his proceeds from the sale
of cash flow rights, however by directly bargaining with a potential buyer he maximizes

34
his revenue from the sale of control rights. Though this literature informs about the
technical analysis of SOE privatization, it does not help us in examining the effects of
political dimensions of privatization decision.
Political scientists have also grappled with the privatization wave. Downs (1958)
is one of the early theoretical works that has attempted to outline a theory of government
action in a democracy, combining strains of economics and political science literature.
Later Manzetti (1993) has attempted to develop a framework of analysis that can
incorporate economic as well as political factors that can help or deter a policy of
privatization. While economists have looked at privatization in terms of efficiency,
political scientists have looked at the privatization as a policy shift and have studied it in
the context of complexity of policymaking. Thus, while most of the economics literature
on privatization has been prescriptive or evaluative, political scientists have studied
privatization in terms of shifting powerbase of various interest groups, for example Henig
and Feigenbaum (1997) argue that privatization is based on historical circumstances and
depends on socio-political factors as much as on economic factors.
Empirical research that has looked at factors that determine partial privatization at
firm level is limited as most of the empirical literature has focused on comparing
efficiency before and after privatization. Castro and Uhlenbruck (1997) examine how
country characteristics affect the buying firms’ decision regarding full or partial
acquisition. This study classifies acquiring and target firms’ location into three
categories: developed, less-developed, and former communist countries and finds that
privatization deals in former communist countries are more likely to be total sales than
privatization transactions in less-developed and developed countries. Though

35
informative, this work does not look at the country specific macroeconomic, political and
institutional characteristics lying behind those broad country-type distinctions and
therefore does not address the question of what factors explain partial privatization.
While Clarke and Cull (1997) have done an empirical study of political economy of bank
privatization in Argentina and have demonstrated while political incentives play an
important role in the timing of bank privatization, the macroeconomic variables drive the
decision to privatize. Biglaiser and Brown (2003) look at panel data for 16 Latin
American countries between 1980 and 1997 and find that institutional constraints do not
account for a significant degree of the variance in privatization across the region as a
whole, while only two variables significantly affect privatization: on the economic side,
the rising debt burden within the state sector and on the political side, whether
privatization takes place relatively late in the presidential term.
Two works that have specifically looked at partial privatization and are related to
our work are: First, Gupta (2004) looks at firm level data of partially privatized SOEs
from India and find that even the sale of minority stakes has a positive impact on firm
performance and productivity. Second, Bortolotti and Faccio (2004) study the evolution
of the control structure of 141 privatized firms from OCED countries over the period
1996 through 2000, and find that as of 2000, governments remain the largest shareholders
or retain majority voting control in 62.4% of privatized SOEs. Contrary to accepted
theory, they find that greater government control over privatized SOE does not reduce
market valuation. Our work is different from these works as we do not look at efficiency
or valuation impact of partial privatization but rather look at variations in the extent of
privatization transactions across the world to distil its determinants.

36
In the next subsection we outline the probable macroeconomic, political and
institutional determinants of partial privatization.
2.2 Determinants of Partial Privatization: In the privatization literature,
macroeconomic, and political variables have been used to explain variation in
privatization policies. As privatization is political, we believe that institutional variables
should play a significant role in determining political costs and benefits and to that extent
they need to be taken into account separately as independent factors. North (1994) argues
that institutions not only define the incentive structure of a society for economic agents
but also for the political agents and institutional characteristics play a crucial role in
economic policy formulation. To achieve the objective of economic growth over a period
of time development of a legal system that promotes correct incentives, the creation of
effective and impartial enforcement mechanisms that comprise the legal system and
development of organizations that invest in the skills and knowledge that are essential to
sustain productivity are as much required as transferring assets from the public sector to
the private sector. This implies that privatization programs launched by governments are
affected by the characteristics of the institutions that determine costs and benefits to
politicians and bureaucrats. Zinnes, Eilat and Sachs (2001), using data from transition
economies, find empirical support for the criticality of institutions and show that,
contrary to “Washington Consensus”
12
, economic performance gains come only when
change-of –title reforms are accompanied with key institutional reforms.

12
First decade of transition to market economy was based on a set of tenets that became the mantra of the
donor community centered around Washington D. C. and therefore these tenets are often referred as
“Washington Consensus”. These tenets include fast privatization, immediate macro stabilization, rapid
liberalization and opening of economy to foreign trade and investment.

37
2.2.1 Macroeconomic determinants of privatization:
In every great monarchy in Europe the sale of crown lands would produce a very
large sum of money which, if applied to the payments of the public debts, would
deliver from mortgage a much greater revenue than any which those lands have
ever afforded to the crown….
Adam Smith (1776, p824)

The idea that sale of government assets should be used to improve public finances
is not new. It has been argued that governments, especially in developing countries,
should implement privatization programs as a response to worsening fiscal deficits and
burgeoning public debt (Ramamurti, 1992; Ridley 1994). Alternatively, governments
should sell SOEs to reduce inflation. For example, it is noted that public sector deficits
led to high inflation in the early eighties in several Latin American countries and
subsequently high inflation countries adopted faster and deeper privatization programs
while low inflation Latin American countries were slower to start privatization
programs
13
.
2.2.2 Political determinants of privatization: The Economics literature has focused
primarily on the study of political institutions on reform policies. Alesina and Drazen
(1991) assume that, while the benefit of stabilization accrues to all citizens, costs of
stabilization in terms of higher taxes are shared differently by different interest groups.
Under these assumptions, the process of stabilization is like “war of attrition” between
different interest groups leading to political stalemate until one group gives up. Thus the
stabilization policies are implemented when, at the margin, the incremental cost of
waiting for one group equals the expected benefit of waiting. This leads to the conclusion
that those countries that have more political cohesion, i.e., in which the cost of

13
Bruno and Easterly, 1996 term these reforms in Latin Ameriacn countries in 80s as “Inflation’s
Children”.

38
stabilization is more equally distributed between “winners” and “losers” tend to adopt
stabilization policies faster. Empirical work has found support for this prediction
(Roubini and Sachs, 1989; Alesina and Perotti, 1995). Though this model refers to
stabilization policy, its implications for privatization are significant. Privatization is a
policy that has significant distributional implications and to that extent produces
“winners” and “losers”.
The political institutions and their relative formal power are determined by
electoral rules which in turn characterize political competition. Persson and Tabellini
(2000) compare majoritarian and proportional electoral rules and show that majoritarian
elections foster political competition which in turn reduces rents for politicians. It has
been largely documented that politicians use SOEs as a source of political rent (Vishny
and Shleifer, 1994), and also derive financial gains from direct corruption or have private
benefits of control (Dyck and Zingales, 2002). In our context, these two theoretical
predictions imply that there will be less political opposition to privatization in countries
with majoritarian rule and thus we can expect higher proportion of privatization, as the
political cost of privatization will be lower since politicians derive lower rent from SOEs.
2.2.3 Institutional determinants of privatization: Institutions define the incentive
structure of the politicians as well as of economic agents and firms. As pointed out by
North (1994), “The institutional constraints together with the traditional constraints of
economic theory define the potential wealth maximizing opportunities of entrepreneurs
(political or economic)”. As the economic costs and political benefits of SOEs are
determined by the institutions of governance, the valuation and risk estimates of SOEs by
the prospective buyers is also a function of the institutional set up of the country. A

39
country facing social fragmentation or ethnic conflicts will be considered risky by the
investors and therefore the low demand for privatized SOEs shares may lead to a low
proportion of privatization. Further, in a socially divided country, it may be more difficult
for politicians to negotiate and reach compromises on different aspects of privatization
and may result in low privatization.
Similarly, the quality of bureaucracy and effectiveness of regulatory bodies are
likely to play a significant role in determining the extent of privatization from both the
demand and supply side. While the poor quality of bureaucracy or ineffective regulation
will dissuade investors from buying SOE shares and therefore decrease the demand for
shares, the same factors may also increase the potential for private benefits accruing to
the investors ( such as being able to “rip-off” the SOE), thereby making it more lucrative
to the buyers. So the effect of quality of administrative institutions on demand for
privatized SOE shares is ambiguous.
Further, while good quality governance may prevent politicians from being able
to extract high rents from SOEs and thereby their political loss from privatization would
be expected to be lower. Thus better governance may lead to higher proportion of
privatization. But, on the other hand, good quality of administrative structure may also
mean that SOEs are being managed efficiently and thus privatization is likely to fetch
smaller efficiency gains. Thus better governance may lead to lower proportion of
privatization. Therefore, the effect of the quality of governance on partial privatization is
ambiguous.

40
The next section outlines our empirical approach and defines our macroeconomic,
political and institutional variables and discusses what directionality in the relationship is
expected on the basis of existing theoretical and empirical literature.
2.3 Empirical approach, variables and data: In this research I view privatization
as a policy choice made by politicians on the basis of political cost-benefit analysis.
While the costs include the loss of their political support base, the loss of possible rent
seeking in lucrative contracts, and the loss of opportunities to provide partisan
employment, the potential benefits include lowering of fiscal deficit and thereby reducing
inflation, increased funding for social programs, economy wide efficiency gains and
better provision of public services. Further, while most of the costs are immediate,
benefits accrue over a time period and are often uncertain. As these political costs and
benefits are not directly observed, we hypothesize that the politician’s use partial
privatization to do the balancing act, and that the partial privatization proportion is
determined by the tradeoffs made by the politicians between expected political costs and
benefits.
2.3.1 Empirical approach: Our approach here is based on Rodrik (1994) and Banerjee
and Munger (2004). Rodrik (1994) posits that the benefits arising from the economic
efficiency of trade reforms must be weighed against the political cost of redistributing
income and defines the ratio of redistribution generated by reform to its efficiency
benefits as political-cost-benefit ratio (PCBR). Since it’s not possible to operationalize
this ratio, we use the proportion of privatization as a proxy for measuring the tradeoff
between political costs and benefits. Banerjee and Munger (2004) follow a similar

41
approach and define ‘net political benefits’ (NPB) as the dependent variable and use the
timing, pace and intensity as its proxies.
We use a set of macroeconomic, political and institutional variables as proxies for
determinants of privatization. The rationale for using these variables and expected
directionality is based on the political economy of reform literature in general and the
privatization literature in particular. . We use the proportion of privatization as the
dependent variable and explain its variation across countries in terms of macroeconomic,
political and institutional variables while controlling for regional groupings, income
levels, trade openness and industry type. Later, we will examine the relationship between
partial privatization and its determinants for each type of industry as the political costs
and benefits are likely to be different for different industry types. Politicians are likely to
view the privatization of a power utility employing 15,000 persons differently than
privatizing a car manufacturing unit having the same employee strength. While only the
employees of the car manufacturing public unit may feel threatened by privatization, in
case of power utility large number of consumers who get electricity at subsidized rates
may also join employees in opposing privatization. To that extent, it is likely to be
insightful to study the determinants of partial privatization for each industry grouping.
We group SOEs that were privatized into three categories, namely core sector,
intermediate sector and competitive sector. Core sector consists of firms operating in
infrastructure sector or those that can be defined as natural monopolies, like electricity,
telecommunication, railways and roads. Intermediary sector consists of those firms that
are likely to be large in size and therefore to have large employment like steel, and
petrochemicals, or those that are considered critical for country’s economic activities like

42
banking, insurance or financial services. All other firms which are either small or are in
sectors where large number of firms are either operating or have potential to operate are
grouped as competitive sector. Table 2.2 provides the details.
Table 2.2 : Classification of industries

Type of sector Type of industries/SOEs
Core Sector
Roads, Telecommunication, Railroad, Power utility, Water, Defence, Electricity,
Power distribution, Postal service, Telephone services, Metro system,
Hydropower,
Intermediate Sector
Airlines, Petroleum, Petrochemical, Shipping, Shipbuilding, Financial Services,
Banking, Gas, Steel, Coal, Water and sewerage, Airport, Rail/bus equipment
manufacturing, Fertilizer, Gas distribution, Insurance, Minning,
Competetive Sector
Chemicals, Television station*, Detergent, Food stuff, Paper and pulp, Tourism,
Agribusiness, Livestock, Port facilities, Race track, Electrical/Electronics,
Manufacturing, Real estate, Transport, Airport services

Macroeconomic variables:
Share of agricultural sector (SAGRI): This variable measured as the proportion of the
agriculture sector in Gross Domestic Product (GDP) captures the structure of the
economy. If the share of agriculture in the economy is large, that will indicate the country
does not have a large manufacturing and service sector, which in turn will mean that the
country has a relatively small private sector. Thus for a country with a high share of
agriculture in the economy, the cost of privatization is likely to be high as market
institutions are less likely to be developed to facilitate the efficient functioning of the
capital market. Therefore, we expect that those countries with higher share of agriculture
in economy will have lower proportion of partial privatization.
Market Capitalization as proportion of GDP (MARKTCAP): This variable
measures the size of capital market relative to the economy. Market capitalization (also
known as market value) is the share price times the number of shares outstanding of

43
domestically incorporated companies listed on the country's stock exchanges at the end of
the year. A higher market capitalization to GDP ratio indicates well developed capital
markets, which in turn means a lower cost of raising capital and thus a lower cost of
privatization. We hypothesize that a higher MARKCAP variable will mean higher
privatization proportion and thus, we expect a positive relationship.
Fiscal deficit as proportion of GDP (DEFGDP): We postulate that policy makers
implement privatization programs to alleviate fiscal problems. They face the tradeoff
between cutting funding of populist social sector programs, on one hand, and selling
SOEs to raise resources, on the other hand. If the deficit to GDP ratio is large, then the
benefit of privatization (by using privatization proceeds to fund populist programs) is
likely to overweigh the cost of privatization (in terms of losing the partial support base
amongst workers). Therefore we expect that DEFGDP and the partial privatization
proportion will be a positively related.
Literature in study of reforms has generally concluded that reforms take place
when the political cost of maintaining the status quo becomes higher than that of
undertaking reform. A Crisis like a high fiscal deficit increases the cost of delaying
privatization and thus increases the net benefit from privatization. This appears more
applicable in the context of developing countries as Rodrik (1994) has noted that “no
significant case of trade reform in a developing county in the 1980s took place outside
the context of a serious economic crisis”.
GDP growth rate (GDPGR): We postulate that the affect of GDP growth rate on of
privatization will be positive. If the economy is growing at a fast rate, then excess
employees who may be laid off due to privatization may be able to find alternate

44
employment in private sector easily and to that extent the political costs of privatization
will be lower. Further, if the growth rate is high, then the government may be in a better
position to compensate those who are adversely affected by privatization, and to that
extent lower its political costs.
Foreign firm’s participation (FORBUY): This variable represents the proportion of
foreign firms’ share in the consortium of buyers
14
.As discussed earlier, the literature has
pointed out two possible reasons for governments engaging in partial privatization
contracts: (1) due to the tradeoff between efficiency gains from privatization and the
private benefits of control accruing to foreign buyers (about which government has
asymmetric information) and (2) due to domestic buyers are wealth-constrained. If indeed
partial privatization is motivated by the first reason, we should observe the privatization
proportion to be high when only domestic firms are buyers. It is possible that government
allows foreign firms to participate in privatization transactions because domestic firms
are wealth-constraint, if so, the share of foreign firms in the buyers consortium will
reflect the degree to which domestic buyers are wealth-constrained. Therefore, for the
second reason also, we expect any increase in FORBUY to lower the proportion of
privatization. Thus we expect to get a negative relationship between proportion of
privatization and FORBUY.
Political variables:
Political constraints: The political economy literature has increasingly focused
on the consequences of the political conflicts between politicians on economic policy.
Biais and Perotti (2002) use a partisan politics model, where right wing political party

14
If only one firm is buyer, then it’s zero when that firm is domestic and 1 otherwise.

45
privatizes to gain support from the share owners of newly privatized SOEs and shows
that by allocating a substantial amount of shares of privatized SOEs to the middle class,
the rightwing politicians make median voters averse to redistributive policies of the left-
oriented parties. We take a different approach and postulate that privatization programs
involve a policy shift and due to the presence of checks and balances in a democratic
polity, any political party, whether right wing or left oriented, is constrained in making
these policy shifts. The political orientation of a political party reflects its preferences in
terms of policy choices and signals to the electorate as to what “it will like to do”, but
what “it can do” depends on the political structure of the government. For example, it
may be far more difficult for a right wing party to pursue its privatization program in a
country with parliamentary democracy than in a country with presidential form of
government.
Bortolotti and Pinotti (2003) have studied the effect of political institutions and
political orientation of the executive on privatization programs. They have constructed a
political-institutional index, which they call POLINST based on three components: (a)
Gallagher index
15
(1991); (b) the effective number of parties; (c) a dummy variable to
indicate the type of executive. As pointed out by Lijphart (1999), though these three
components refer to specific feature of political institutional setting, they are likely to be
strongly correlated and may be jointly determined. To take care of potential
multicollinearity problem for POLINST, they standardize the quantitative measure of
these three components and then take their mean. Though this is an innovative approach,

15
The Gallagher Index (or least squares) can be used to measure the disproportionality of electoral systems,
in particular majoritarian systems. The index involves taking the square root of half the sum of the squares
of the difference between percent of vote and percent of seats for each of the political parties. The lower the
index value of a country is the lower the amount of disproportionality is in this country and vice versa.

46
it fails to capture the measure of political constraints that may impede policy changes
required for privatization as it does not adequately take into account the political power
of players with veto in the policy agenda setting.
The political science literature has also studied the role of political constraints in
implementing policy changes. They have used more precise measure of political
constraints than economists who tend to use dummies or discrete variables and time
invariant measures. Measures developed by political scientists are also superior because
they use simple mathematical formulas to model electoral outcomes and thus limit
researcher’s discretion. In this research, we use measures developed in the political
science literature to take advantage of their well defined characterization, time variance
and neutrality to researcher’s assumptions.
POLCON (Political constraint): Our source of data for political constraints is
POLCONIII developed by Henisz (2002). This measure of political constraints estimates
the feasibility of policy change to the extent to which a change in preferences of any one
actor may lead to change in government policy. The measure is constructed by first
identifying the number of independent branches of government, like executive, lower and
upper legislative chambers, that have veto power over policy change. The preferences of
each of these branches and the status quo policy are then assumed to be independently
drawn from a uniform, one-dimensional policy space. This initial measure is then
modified by taking into account the extent of alignment across branches of government
using data on the party composition of the executive and legislative branches. The
feasibility of policy change increases with the increasing alignment. The measure is then
further modified to capture the extent of preference heterogeneity within each legislative

47
branch which increases (decreases) decision costs of overturning policy for aligned
(opposed) executive branches.
Legislative fractionalization (LEGRAL): The political economy literature has
extensively used categorical variables to distinguish between left and right wing
governments since the seminal work by Hibbs (1997). This classification is good only for
those countries where the political competition leads to clearly marked winners and the
political-institutional set up is either monolithic or the different institutions providing the
checks and balances are controlled by the same political party. However, in most
countries where the party system is fragmented, electoral alliances keep on shifting. As a
result, different political institutions are controlled by different political groupings at the
same time, suggesting that the use of dummy variables to capture the orientation of the
government is not satisfactory.
For example, India has a bicameral parliamentary system. Members of the Lower
house, which has a term of five years, are elected by direct voting. More often then not,
the governments are not able to maintain a majority in the Lower house due to shifting
coalition politics, and mid term elections are held
16
. Members of the Upper house are
elected by indirect voting by an electoral college consisting of members of state Lower
and Upper houses, with a term of 6 years and elections for one third seats are held every
two years without any changes in the cycle. Most legislation, other than Finance Bills,
requires approval by majorities in both houses, or by a joint sitting of the two houses.
Though, the Congress Party has maintained a clear majority in the Upper house, without
any interruption in the last 25 years, it has been able to gain majority in the Lower house

16
In last 25 years India has held 6 parliamentary elections, of which 3 were mid term elections.

48
only the 2004 elections, that too with a supporting coalition of more than 10 regional and
left-oriented parties. Due to the political-institutional arrangement, the orientation of the
dominant political party in coalition does matter in policy shifts, but is constrained.
On the other hand, the privatization policy of the previous government of
Bhartiya Janta Party (BJP) was challenged in the Supreme Court of India in 2002 on the
grounds that government has no executive or legislative power to change the “basic
structure of the constitution” and privatization goes against the goal of “socialistic
development” as enshrined in the Directive Principles of the constitution. All
administrative and legislative work was put on hold for almost a year and could only
proceed after the courts clarified that government has a legal and constitutional right to
implement privatization programs that are initiated for cogent economic reasons and not
for “political” gains or “ideological” reasons.
In order to address the continuum of political orientation of different parties and
to reduce the abruptness of dummy variable approach in measuring legislative
fractionalization, we again borrow from political science literature that has relied on
precise mathematical expressions, removing researcher’s bias and making them
comparable over time period . Using Heinz (2002), legislative fractionalization is
approximately the probability that two random draws from the lower (upper) legislative
chamber will be from the same party. The formula includes a modest adjustment to
reflect the difficulty of maintaining a coalition as the number of parties in that coalition
increases. The exact formula is:

]
1 - N
N
n
1) -
n
(
[ - 1
i
i n
=1 i
Σ

49
where n = the number of parties, n
i
= seats held by nth party and N = total seats. Note that
this measure is different from that used by Bortolotti and Pinotti (2003) who used an
index based on Huber and Inglehart (1995) expert survey
Institutional variables: Study of role of institutions in economic growth is not new.
The role of institutions that constrain government powers and its ability to interfere with
economic activities was stressed by Montesquieu, as far back as, 1748. The new
institutional economics literature has supported this view with theoretical and empirical
work in recent years. While DeLong and Shleifer (1993) use data on urbanization of
European regions during 20
th
century to show that a “constrained” local government
leads to faster city growth, Mauro (1995) finds positive relationship between economic
growth and good institutions. Recent work, including Hall and Jones (1999), Acemoglu
et al. (2001,2002), Dollar and Kraay (2003) and Rodrik et al. (2004), appear to go further
converging to the premise that institutions determine economic growth. In the context
privatization, as institutions determine the relative costs and benefits of state ownership
to the politicians, what is privatized and how much is privatized will be influenced by the
same institutional factors, but surprisingly their role in determining privatization policies
has not been studied in literature except for a recent work by Banerjee and Munger, 2004.
Governance Indicators: The literature has made use of three different datasets for
measuring institutions or institutional effectiveness. The first data set from International
Country Risk Guide (ICRG) are survey indicators of institutional quality collected over
1980s and 1990s. Though this data set has been extensively used
17
in the recent past, we
do not use it in this research for two reasons. First, it is based on subjective assessments

17
Knack and Keefer (1995); Hall and Jones (1999) and Accmoglu et al. (2001).

50
of risk to international investors along various different dimensions that are relevant only
to buyer’s perspective on privatization which is not the focus of this study. Second, the
index is created by aggregating arbitrarily assigned points to different dimensions of three
major risks, i.e., political, economic and financial risk. For example its not clear as why
12 points are assigned to socioeconomic conditions and only 4 points are assigned to
ethnic tension in calculating political risk. Further these measures tend to capture what
happened in a country rather than some permanent rules of the game and to that extent
indicate the choices made by the leaders of different countries at a particular time and not
the quality of institutions or governance in these countries. For example, in 1984, the top
ten countries with the lowest expropriation risk include Singapore and the U.S.S.R ,
which reflect the choices of dictators and not political constraints (Glaeser, LaPorta,
Lopez-De-Silanes, and Shleifer (2004).
The second dataset, called Polity IV has been collected by political scientists
(Jaggers and Marshall, 2000) and attempts to measure constraints on the executive power
of the government. This data set is very rich and has been widely used in political
economy literature. As pointed out by the authors, this measure looks at the checks and
balances in the decision making process and thus, rich democracies like India get a
perfect score of 7, while dictatorship like China gets the worst score of zero. The problem
in using this measure is that it is calculated by using outcome of the elections rather than
some permanent rules of the game, and therefore tends to have wild swings. For example,
Haiti gets the worst score of 1 during 1960-1989, rises to 6 when Aristide is elected,
goes back to 1 during 1991-1993 when he is removed from power and jumps back to 6
when Aristide comes back in 1994 and later rises to 7. Thus Polity IV provides a time

51
variant assessment of electoral outcomes, not a measure of actual political constraints on
government or durable rules. More so this measure again fails to capture the tradeoffs
made by the politicians due to administrative inadequacies or opportunities for rent
seeking, and thus may not be appropriate for this research.
We use the third dataset that includes aggregate indicators of six dimensions of
governance from Kaufmann, Kraay and Mastruzzi (2003) which have been developed
using a broad definition of governance as traditions and institutions through which
authority is exercised. This includes the process of selection, monitoring and replacement
of government, freedom and flexibility in policy formulation exercised by government
and respect of citizens and government for the institutions of economic and social
interactions. The six governance indicators are measured in units ranging from about -2.5
to 2.5, with higher values corresponding to better governance outcomes. The governance
indicators in this dataset reflect the statistical compilation of responses on the quality of
governance given by a large number of enterprises, citizens and expert survey
respondents in industrial and developing countries, as reported by a number of survey
institutes, think tanks, non-governmental organizations, and international organizations.
Aggregation of multiple surveys and opinions ensures that the measure reflects a balance
of competing viewpoints and is not disproportionately influenced by particular actors
such as investors or potential buyers of SOEs.
Voice and Accountability (VOACC): This variable includes a number of indicators
measuring various aspects of political process, civil liberties, independence of media and
political rights and measures the extent to which the citizens of a country are able to
participate in the selection of the government. If institutions are such that the government

52
is held accountable for its actions and independent media ensure that different view
points are brought out in public policy debates, one would expect that government would
be required to put in much more effort in building up public support. Thus, we would
expect that the effect of VOACC on the proportion of privatization would be negative.
Government Effectiveness (GOVEFF): This variable measures the ability of
the government to formulate and implement policies and aggregates responses on the
quality of public service provision, the quality of the bureaucracy, the quality of civil
servants, independence of the civil service from political pressure, and the credibility of
government ‘s commitment to policy implementation. We expect that the higher is
government effectiveness; the lower would be the rent seeking opportunities from public
ownership. This may mean that efficiency benefits of privatization are likely to outweigh
political costs and we predict that the relationship between proportion of privatization and
GOVEFF should be positive.
Regulatory Quality (REGQUA): This variable measures the strength and expertise of
regulatory authorities and bureaucrats in formulating and implementing policies
independently and consistently even if government should change. Poor quality of
regulation will imply that the bureaucrats are likely to be “captured” by the firms that
they are supposed to regulate (Laffont and Tirole, 1993). Thus greater is the
independence of regulatory authorities, greater is the quality of regulation, and less
attractive is public ownership to the politicians ( Shleifer and Vishny, 1994). On the other
hand, high quality of regulation will also allay popular fear that after privatization private
business would be able to exploit the monopolistic power, causing the cost of services
like electricity and telecommunication to increase sharply (Campos, 2000). Therefore, we

53
hypothesize that quality of regulatory institutions will positively affect the privatization
process.
Rule of Law (RULAW): This variable measures the extent to which agents have
confidence in and abide by the rules of the society and includes perceptions of the
incidence of crime, the effectiveness and predictability of the judiciary, and
enforceability of contracts. High value of this variable reflects the success of a society in
developing institutions and governance environment in which fair and predictable rules
and norms form the basis for economic and social interactions and property rights are
protected. Literature has pointed out that if rules by which society’s interacts are
transparent and protect property rights, then returns to investment will improve (Isham
and Kaufman, 1999); and on the other hand low security of property retards innovation,
stunts investment and slows down flow of foreign capital (Mauro, 1995). Higher RULAW
would imply higher economic benefits and lower political costs of privatization, and thus
we hypothesize that the privatization proportion would be positively affected by RULAW.
Control of Corruption (CONCOR): This variable measures perceptions of corruption,
defined as the exercise of public power for private gain. This measure aggregates
different sources of corruption such as frequency of “additional payments to get things
done”, “grand corruption” in political arena or tendency of interest groups to indulge in
“state capture”. Some have argued that corruption can facilitate economic development as
“speed money” provides incentive to bureaucrats to work efficiently (Huntington, 1968).
As a result, the effect of corruption on the extent of privatization may be ambiguous
because while corrupt politicians may view privatization as a way to expropriate state
property and implement higher proportion of privatization, private investors may shy

54
away from participating in privatization program, thereby reducing the demand for SOE
shares. Though recent literature points to negative effects of corruption on economic
growth and private investment (Shleifer and Vishny, 1993), its effect on partial
privatization are not clear
Control variables: In order to isolate the effects of economic, political and institutional
determinants under study from other influences on partial privatization, we control for
other determinants identified by literature. First, we control for income levels across
countries by using GDP (GDPCAP) per capita as a proxy. This allows us to isolate the
effects of development and the capacity of countries to offset any adverse distributional
impacts of privatization. This is because developed countries may have not only better
institutional structure but also strong market mechanisms with which to facilitate reforms.
Second, we use for trade openness (OPEN) to control for the country’s global integration.
This variable is calculated as proportion of gross domestic product (GDP) to the sum of
exports and imports of goods and services and also captures endogenous rents accruing to
politicians in the economy (Banerjee and Munger, 2004). The lower is the trade openness
(OPEN), the higher would be protectionism and the higher would be potential for rent in
the economy.
Third, we control for type of industry sector by using indicator variable CORE for
infrastructure sector and INTER for Intermediate sector. This is important as we expect
that governments will pursue varying privatization policies for for different industry
segments. Table 2.2 provides the classification of industries in these categories.
Finally, we use regional dummy variables for Latin America (LATAM), Central
America (CENAM), Caribbean countries (CARIB), South Asia (SOUAS), Eastern Asia

55
(EASAS), Estern Europe (EASEU) and North Africa (NORAF) to control for regional
variations across different parts of the world.
Data: We have used World Bank’s Privatization Group’s transaction database for
information on privatization transactions. This data base has information on 6587
privatization transactions carried out in 73 countries from 1989 to 1999. For each
transaction, there is information on the name of the SOE, the business sector, the
proportion of privatization, the sale amount, the foreign exchange earned on privatization
transaction, the method of privatization, and whether the buyer was a local firm, a foreign
firm or a consortium of domestic and/or foreign firms. For some transactions information
on the name of the buyer firm and/or composition of domestic/foreign firm consortium
with percentage share is also provided. We drop those transactions that have either zero
privatization proportion or no information on this variable and also those countries that
had fewer than 5 privatization transactions in the 12 year period. This leaves us with
4,760 privatization transactions from 43 countries.
A summary of the privatizations transactions for these countries is given in Table
2.3. Note that although approximately two thirds privatization transactions were partial
18
,
they accounted for more than ninety percent privatization revenue. Surprisingly, the
proportion, both in terms of the number of transactions and privatization revenue, has
been very high for the partial privatized firms for Eastern European countries and Russia.
This contrast with the popular perception is that everything has been totally privatized in
those countries.

18
For this analysis, any privatization transaction that transfers 90% or more ownership to private sector is
classified as full privatization.

56
Table 2.3 List of Selected Countries and number of privatizations

S.No. Country No. of transactions Region
1 Argentina 166 Latin America
2 Brazil 164 Latin America
3 Chile 30 Latin America
4 Venezuela 62 Latin America
5 Colombia 28 Latin America
6 Bolivia 92 Latin America
7 Mexico 263 Central America
8 Nicaragua 77 Central America
9 Honduras 39 Central America
10 Trindad & Tobago 17 Caribbean
11 Jamaica 46 Caribbean
12 Bangladesh 28 South Asia
13 India 77 South Asia
14 Pakistan 119 South Asia
15 Sri Lanka 95 South Asia
16 Thailand 19 East and South East Asia
17 Philippine 99 East and South East Asia
18 Malaysia 47 East and South East Asia
19 Indonesia 23 East and South East Asia
20 China 143 East and South East Asia
21 Slovk Repbulic 351 Central/Easten Europe and Russia
22 Russia 69 Central/Easten Europe and Russia
23 Romania 1186 Central/Easten Europe and Russia
24 Poland 445 Central/Easten Europe and Russia
25 Lituania 110 Central/Easten Europe and Russia
26 Lativa 34 Central/Easten Europe and Russia
27 Hungary 999 Central/Easten Europe and Russia
28 Czech Republc 47 Central/Easten Europe and Russia
29 Bulgaria 353 Central/Easten Europe and Russia
30 Armenia 113 Central/Easten Europe and Russia
31 Egypt 104 North Africa
32 Morocco 77 North Africa
33 Tunisia 71 North Africa
34 Zambia 86 Sub-Saharan Africa
35 Uganda 62 Sub-Saharan Africa
36 Tanzania 104 Sub-Saharan Africa
37 South Africa 17 Sub-Saharan Africa
38 Nigeria 58 Sub-Saharan Africa
39 Mozambique 164 Sub-Saharan Africa
40 Malawi 36 Sub-Saharan Africa
41 Kenya 102 Sub-Saharan Africa
42 Ghana 103 Sub-Saharan Africa
43 Cote d'lvorie 52 Sub-Saharan Africa
TOTAL Privatization Transactions = 6377
Source: Investment Promotion Network
http://www.ipanet.net/infores/drilldown.cfm?menu_chosen=3001&type=16&family=360,405
&child=405&pcount=225

57
We construct a pooled cross section/time-series data set with 4,760 observations
from 43 countries that represent low-income or middle-income countries from the
transition economies and developing world. These countries come from Latin America,
Central America, Caribbean, South Asia, South East Asia, Eastern Europe and Russia,
North Africa, and Sub-Saharan Africa. For this research, pooled data set is the best
approach to study the partial privatization process over time as it captures both cross-
sectional and temporal dynamics of privatization process and enables the use of a number
of specifications that control for heterogeneity bias.
Out of 4,770 privatization transactions only 236 (approximately 5% of all ) were
of SOE s operating in the core sector though they yielded approximately US$95 billion in
privatization revenue that was approximately 40% of the total proceeds. The partial
privatization proportion in the “core sector” was on average approximately 45%. The
“intermediate” sector accounted for 519 (approximately 11% of all) transactions and
yielded approximately US$74 billion in revenue to the governments which contributed
approximately 31% of total privatization revenue. The remaining 4,015 privatized SOEs
were operating in the “competitive sector” and yielded approximately US$67 billion in
revenue or 28% of total receipts. Partial privatization proportion for intermediate sector
and competitive sector were 53% and 68%, respectively.
Foreign firms accounted for 815 transactions (approximately 17% of all), joint
ventures for only 247 cases (approximately 5% of all) leaving more than three fourths of
all transactions only to be done by domestic firms. Surprisingly, contrary to our
expectation, we find that out of the 236 core sector privatization transactions in our
sample, only 71 transactions (approximately 30%) SOEs were sold to foreign firms, and

58
47 transactions (approximately 20%) to joint ventures. The remaining 118 core sector
transactions involved only domestic firms. In the intermediate sector 113 transactions
involved foreign firms, 85 privatization sales were with joint venture firms and remaining
321 buyers were domestic firms. Approximately 77% of the privatization transactions
that involved only foreign firms and 45% of the joint venture privatization buyers were in
competitive sector. Average privatization proportion in transactions involving foreign
firms, joint ventures and domestic firms was approximately 60%, 60% and 67%
respectively.
Form this pattern of global privatization three specific generalizations emerge: 1.
Most of the privatization transactions involving exclusively foreign firms or their joint
venture with domestic firms have been in “competitive sector”. 2. On average the
privatization proportion is lowest for “core sector” SOEs and highest for competitive
sector SOEs. 3. The privatization proportion does not vary much across foreign, joint
venture or domestic buyer firms.
First, we use our entire dataset to analyze the effects of various macroeconomic,
political and institutional variables in determining privatization proportion. Then we look
at “core sector”, “intermediate sector” and “competitive sector” separately to study how
the role of explanatory variables changes across these industry categories. Later we
modify our dependent variable as binary to capture ‘transfer of control’ characteristic of
privatization; Control takes value 1 if privatization proportion was 50% or more, and 0
otherwise
19
. We use this modified dependent variable to study the dynamics of ‘transfer

19
We also test the affect of “Control” variable by taking cutoff point as 35%, i.e. Control = 1, if
privatization proportion was 35% or more.

59
of control’ as a sub theme with in privatization landscape. In the next section we describe
our empirical approach and model.
2.4. Empirical method and model: Based on the analysis in section 2, we develop an
empirical framework in this section to test the hypotheses. As mentioned earlier we use
pooled cross-section time series design for our empirical analysis. The country provides
the cross-section variation while year provides the time series variation, and therefore,
this analysis can identify both cross-country and temporal dynamics of privatization.
The pooled cross-section time-series design allows the use of a number of
specifications that control for heterogeneity bias. It enables us to control for those time-
invariant country-specific effects that may be omitted in the regression model. We are
including many country specific control variables in our model to mitigate the problem of
omitted variable. The country-specific effects in the model are estimated by the country
specific intercept and the same may be fixed or random in a panel dataset. (Nielsen and
Grady, 2000). In datasets that are cross section dominant and in cases where there are
significant unit specific effects, random effects model is most useful. The advantage of
the random effects model is that, unlike the fixed effects specification, it removes only a
portion of the country specific means and thus permits the use of time-invariant
repressors. In effect, it does not include any ‘between countries’ variation in the sample
(Banerjee and Munger, 2004). But for understanding the dynamics of partial privatization
with time-variant regressors, ‘between countries’ variation is important.
In our setting the random effects model is most appropriate as the sample group of
countries is drawn from a large population of countries where country specific intercept
terms are randomly distributed across elements. Further, random effects model is

60
asymptotically efficient compared to fixed effects model (Tuma and Hannan, 1984). One
of the drawbacks of random effects model is the potential problem of autocorrelation.
Our dataset is less likely to suffer from the biases of auto-correlated errors as our data
matrix is ‘wide’ (cross section dominant) rather than tall (Stimson, 1985). If the error
terms are autocorrelated then the estimated coefficients of the repressors and being
cognizant of this issue, we perform Durbin Watson test for the estimated models to rule
out this possibility. Further, we test the consistency of the specifications using Hausman
test and the insignificant values of Chi-square statistic suggest that random effect’s is an
appropriate specification.
2.4.1 Privatization proportion model : Privatization proportion of SOE i in country
j at time t is determined by firm characteristics vector X , country’s time invariant
characteristics vector Z and country’s time variant economic characteristics vector E ,
political constraint vector P and institutional characteristics vector I .
ijt i j E jt P jt I jt ijt
PP X Z E P I α ε = +Π +Β +Δ +Δ +Δ +
SOE characteristic vector X consists of categorical industry type variable CORE and
INTER where the reference category is “competitive sector”, and categorical variable
characterizing the composition of buying firm/s, i.e., FORBUY and JOINBUY where
reference category is domestic buyer firms. Country-specific time-invariant characteristic
vector Z consists of regional grouping categorical variable LATAM, CENAM, CARRIB,
SOUAS, EASAS, EASEU and NORAF with Sub-Saharan African countries as reference
category. Country-specific time-variant economic variables vector
E
Δ consist of SAGRI,
MARKTCAP, DEFGDP, and GDPGR, time-variant political constraint vector
P
Δ consists

61
of POLCON and LEGRAL variables, while time variant institutional variable vector
I
Δ
has VOACC, GOVEFF, REGQUA, RULAW, and CONCOR variables.
Independent variables in our model include both time-varying and time-invariant
covariates. It is important to retain this characteristic in our approach as it is likely to lead
to insights into the dynamics of political-institutional nature of privatization process.
Further, we use a model similar to the equation (1) on previous page to study
effect of macroeconomic, political and institutional variables on privatization proportion
for core, intermediate and competitive sector separately. We use two other models to
study the determinants of ‘transfer of control’ in privatization process, where dependent
variable Control is dichotomous. The implicit assumption here is that if the privatization
proportion is more than 50%, then the operational control is transferred to the buyer.
Ideally, we would have liked to have information on this variable directly from the terms
of the privatization deal as it is possible that ‘control’ may be transferred even if less than
50% shares are sold and more than 50% sale of equity may not necessarily mean transfer
of ‘control’ all the time Bortolotti and Faccio (2004). In the absence of data, this is the
best assumption that we can make and it also seems reasonable to assume that 50% or
more sale of shares to private sector signals a government’s commitment to transfer the
‘control’.
2.4.2 Transfer of control model: In the first case the dependent variable is binary and
we fit a random effects model. Control of SOE i in country j at time t is transferred to
private buyer is determined by firm characteristics vector X , country’s time invariant
characteristics vector Z and country’s time variant economic characteristics vector E ,

62
political constraint vector P and institutional characteristics vector I and unobservable
SOE, country and buyer characteristics denoted by error term
ijt
ε .

ijt i j E jt P jt I jt ijt
CONTROL X Z E P I α ε = +Π +Β +Δ +Δ +Δ +
2.5 Results and discussion: First to rule out potential pitfall of multicollinearity in our
model estimates, we check for correlation amongst all the quantitative explanatory
variables. We find that GOVEFF and RULAW variables are highly correlated with
REQUA and CONCOR. Using the standard methodology to address this issue, we find
that GOVEFF and RULAW, have lower correlation with our dependent variable PRIVAT
PROP. Further, we do sensitivity analysis by using different specifications keeping
different combinations of GOVEFF, RULAW, REQUA and CONCOR and find that
models have higher explanatory power when REQUA and CONCOR are kept in the
model and in their presence GOVEFF, and RULAW are never significant. Thus we drop
GOVEFF and RULAW variables from further analysis.
Table 2.4 contains coefficient estimates using information from all 4770
transactions for different specifications with privatization proportion as dependent
variable.
We used share of agricultural (SAGRI) sector in GDP as a proxy for size and
diversity of manufacturing base of domestic economy. Countries with large
manufacturing structure usually have developed capital markets and also have sufficient
managerial skills available in the private sector. Therefore, we expected to see higher
proportion of privatization in those countries that have lower share of agriculture in GDP.
Though in all our models the coefficient is negative, it’s not significant in any one of

63
Table 2.4: Privatization proportion model with full sample

Dependent variable: Privatization proportion
Explanatory variable I II III IV
SAGRI -0.211 -0.141 -0.122
0.264 0.428 0.489
MARKTCAP 0.038 0.040 0.033 0.036
0.221 0.191 0.274 0.245
DEFGDP 0.536 0.539 0.512 0.490
-0.050 0.049 0.061 0.071
GDPGR -0.393 -0.404 -0.402 -0.397
-0.033 0.025 0.026 0.028
FORBUY -6.197 -6.280 -6.596 -6.685
0.000 0.000 0.000 0.000
JOINBUY -3.419 -3.385 -3.286 -3.224
0.170 0.174 0.187 0.195
POLCON -7.771 -7.873
0.205 0.175
LEGRAL 1.288
0.816
VOACC -10.379 -11.095 -11.775 -12.652
0.001 0.000 0.000 0.000
REGQUA 31.209 32.067 31.237 32.234
0.000 0.000 0.000 0.000
CONCOR -24.661 -24.104 -22.281 -22.170
0.000 0.000 0.000 0.000
GDPCAP 0.001 0.001 0.001 0.001
0.152 0.132 0.193 0.034
OPEN -0.070 -0.072 -0.081 -0.085
0.029 0.023 0.008 0.005
CORE -26.764 -26.704 -26.713 -26.695
0.000 0.000 0.000 0.000
INTER -15.637 -15.523 -15.302 -15.183
0.000 0.000 0.000 0.000
LATAM 2.140 2.552 2.084 1.914
0.472 0.379 0.468 0.504
CENAM 5.111 5.930 4.873 6.532
0.533 0.467 0.549 0.401
SOUAS -6.133 -6.724 -8.947 -9.734
0.097 0.062 0.006 0.001
EASAS -11.680 -13.037 -12.773 -13.302
0.001 0.000 0.000 0.000
EASEU 10.772 11.453 10.342 11.424
0.003 0.001 0.004 0.000
NORAF -0.023 -0.858 -0.818 -0.892
0.995 0.831 0.839 0.825

Values in italics are p values for the estimated coefficients.
R-Sq (adj) 20.600 20.600 20.400 20.400

64
them. Similarly, for variable MARTCAP we expected to see a significant positive
coefficient, as a country with large market capitalization is expected to have more
demand or absorption capacity for SOE shares and thus a higher privatization proportion,
but we find that this variable is not significant in any regression, though it has a positive
sign. There are two possible explanations. First, as pointed out by Banerjee and Munger
(2004), agrarian economies tend to embark upon divestiture programs early on but face
problems in their implementation and thus the impact of this variable is not visible in
final outcome variable. Second, privatization proportion may be lower in agrarian
economies for large privatization programs, rather than across all transactions. Since our
data has more than 80% competitive sector privatization transactions in it, it is possible
that the effect of this variable is too noisy for all sectors. If the second reason has validity,
then we expect to see a negative significant coefficient of this variable in at least core
sector analysis.
Economic crisis is an important determinant of privatization programs and may
also have significant role in determination of proportion of shares to be sold in to be
privatized public firm. A country facing large budget deficits is likely to go for higher
privatization proportion to generate more revenue to meet immediate needs. Our results
show a positive and significant relationship
20
between DEFGDP and privatization
proportion. This result supports the previous research that finds that reforms are driven by
crisis (Bruno and Easterly, 1996). Governments faced with difficult choice of cutting
government expenditure in order to manage deficits, take up privatization programs as a

20
In some models the coefficient is significant at 10% level.

65
less difficult short term solution. This result confirms the notion that politician’s cost-
benefit tradeoff in undertaking policy reforms drives the partial privatization.
Our basic premise that partial privatization is driven by political-economy rather
than technical efficiency considerations, is further confirmed by negative significant
coefficient of GDPGR in all our models. If the economy is growing at a higher rate, then
a lower proportion of shares is sold as the politicians do not want to give away the control
of SOEs. Although this result is at variance with the finding of Dollar and Swenson
(1998) who find that reforms are more difficult to implement in a recession, it is
consistent with in Bortolotti, Siniscalco and Fantini (2000) who find that the volume of
privatization is negatively correlated to growth rates of GDP.
FORBUY has, as expected, a significant negative coefficient in all our models.
We find that the governments, on an average, lower the proportion of privatization by
anywhere between 6.2% to 9.6%, if the buyer is a foreign, all else being equal. This is in
agreement with the theoretical model of Cornelli & Li (1997), who predict that
government use partial privatization when foreign buyers have private information about
value of ‘control’. We will look at this more carefully when we fit regression model to
explain dependent variable ‘control’. Further, non-significant coefficient of JOINBUYI,
confirm that governments do not distinguish between domestic firms and joint venture of
domestic and foreign firms.
The significant result of these models is that surprisingly, political constraints do
not appear to play a role in determining privatization proportion, and rather it’s the
institutional variables that are most important. Bortolotti and Pinotti (2003) find that
political institutions are key to explaining probability of privatization They use data

66
from developed OECD countries and study the role of political institutions in determining
the probability of privatization Our approach is different in four ways: first, our sample
excludes developed OECD countries and is representative of developing and transition
economies; second, our dependent variable is proportion of privatization and politicians
may be determining this variable after the decision to privatize has been taken; third our
political constraint variable POLCON and LEGRAL capture the possibility or difficulty in
policy changes due to interplay of different political agents with veto power, unlike
POLINST and PARTISAN variables of their model that capture the characteristics of
political institutions, and fourth, we also have institutional variables in our model.
Though we expected to see a significant negative impact of political constraints on
proportion of privatization, it may be that in presence of institutional variables, their
explanatory power is reduced if we examine data from all sectors. We will revert to this
issue when we look at each industry sector separately.
We find that institutional variables are significant determinants of proportion of
privatization. We find that countries that have higher VOACC implement lower partial
privatizations. The coefficient is significant at 1% in all the models and on an average 1
unit increase in this variable leads to decrease in privatization proportion anywhere from
10.3% to 13.8%. I think this is a direct result of negative perceptions about privatization
and its negative distributional impacts. Countries like India, with liberal democracies
trade off the political costs of privatization by lowering the privatization proportion and
thereby smoothing the negative public perception. As expected we find that REGQUA is
significantly and positively correlated to partial privatization. If the country has robust
regulatory infrastructure in terms of institutions and legal framework, then, on the one

67
hand politicians will derive less rent from SOEs and so their political cost of privatization
will be low, and on the other hand perceived negative distributional impact of
privatization will be low as people will believe that regulatory institutions will stand up to
the private interests.
With regards to the control of corruption variable CONCOR, we find that its
coefficient is negative and significant in all models. Countries with high levels of
corruption (Given that negative value of CONCOR indicates high level of corruption)
tend to have higher privatization proportions. This does not support the findings that good
governance makes the privatization process more acceptable to people (Nellis, 1999).
Among the control variables, as expected, geographical location matters. The
sector of operation of SOE is critical in determining the extent of partial privatization.
Countries with higher GDP per capita tend to privatize more. Surprisingly, countries
which have higher trade to GDP ratio have lower proportion of privatization. The
privatization proportion is on average lower in core sector by approximately 26% and in
intermediate sector by 15.5% compared to competitive sector. This also indicates that
determinants of partial privatization of each of these sectors may be different and lead us
to analyze each sector separately. Macroeconomic, political and institutional factors
that determine partial privatization are likely to have different impacts in different sectors
of the economy. While market capitalizations may be an important determinant in
privatization proportion of a telecommunication SOE, it may play relatively insignificant
role in privatization of state owned hotels or consumer goods stores. Similarly,
privatization of electricity sector may not only require generating adequate support
amongst coalition partners, but may require amendments in existing laws for government

68
to allow sector to participate in distribution of electricity. For example in India,
privatization of State Electricity Boards remained on hold for more than three years as
federal government was not able to muster sufficient majority in both houses of the
parliament to get the amendments to Electricity Act , 1948 passed. Thus it is important to
look at determinants of partial privatization for each sector.
Partial Privatization in core sector: Table 2.5 contains estimates of different OLS
models with privatization proportion as the dependent variable. These results contain
some surprises but provide some insights into the political process of determining the
proportion.
First, we find that the FORBUY coefficient is not significant. This empirical result
does not support the motivation of Cornelli & Li (1997), which posits that governments
use partial privatization to screen foreign buyers who may have ‘private benefits’ of
control. Privatization programs are often criticized for selling national assets to foreigners
and to that extent one will expect that politicians will be lowering the privatization
proportion when foreign buyers participate in the deal. This suggests that in large
privatization programs relating to infrastructure, governments are not driven by any
‘foreign phobia’ but take a pragmatic approach in deciding what proportion of shares to
be sold.
Second, we find that while REGQUA does not matter, the coefficient of CONCOR
is positive. This result stands out in relation to full sample where both these variables
were significant and CONCOR had a negative coefficient. One of the possible
explanations for

69
Table 2.5 : Privatization proportion model for core sector

Dependent variable: Privatization proportion for core sector transactions
Explanatory
variable I II III IV
SAGRI 0.339 0.381 0.325 0.295
0.701 0.625 0.659 0.561
MARKTCAP -0.267 -0.284 -0.289 -0.271
0.053 0.036 0.030 0.032
DEFGDP -0.580 -0.282 -0.493
0.753 0.867 0.730
GDPGR -0.692 -0.806 -0.807 -0.719
0.498 0.420 0.418 0.307
FORBUY -11.490 -10.986 -11.433 -8.230
0.090 0.101 0.075 0.132
JOINBUY -5.709 -5.461 -5.323 -4.246
0.468 0.481 0.490 0.462
POLCON 49.090 43.740 43.240 47.970
0.092 0.108 0.110 0.005
LEGRAL -4.100
0.841
VOACC -28.500 -28.470 -29.090 -30.580
0.067 0.047 0.039 0.019
REGQUA -5.260 -4.100
0.763 0.811
CONCOR 22.280 22.550 21.170 21.340
0.132 0.122 0.112 0.072
GDPCAP 0.003 0.003 0.003 0.002
0.236 0.207 0.210 0.153
OPEN -0.089 -0.085 -0.086 -0.116
0.532 0.515 0.511 0.305
LATAM 14.185 14.314 13.934 16.040
0.108 0.102 0.104 0.044
CENAM -12.730 -12.550 -12.390 -13.540
0.500 0.503 0.510 0.311
SOUAS -22.960 -22.550 -21.900 18.450
0.033 0.034 0.033 0.056
EASAS -12.695 -12.464 -12.169 -10.548
0.140 0.140 0.144 0.169
EASEU -10.689 -10.598 -10.269 -4.505
0.167 0.166 0.171 0.506
NORAF 32.270 32.450 32.710 12.710
0.212 0.207 0.202 0.479

Values in italics are p values for the estimated coefficients.
R-Sq (adj) 26.200 26.500 26.900 24.900
n 184 185 185 225

70

REGQUA to be insignificant may be that large infrastructure privatization programs
receive wide publicity and generate debates in public forums and to that extent the effect
of quality of regulation is getting captured by VOACC variable which has a very large
negative impact on privatization proportion. While for the full sample the coefficient of
VOACC is -13.855, for the core sector it is -27.07. Thus irrespective of the quality of
regulation, privatization is lower in those countries that allow for free expression and
political debates, a least for the infrastructure sector. The governments are able to
privatize a higher proportion shares in SOEs in the core sector in those countries that
have higher control of corruption. This again hints to the pragmatic approach of the
politicians. Low control of corruption may lead to higher rents to the politician from the
transaction itself, but will also increase the political cost for not being able to deliver on
the promise of improved efficiency of the SOE.
The third surprising observation is that, while for the full sample the coefficient of
DEFGDP was positive and significant and coefficient of MARTCAP was insignificant,
for the core sector these are insignificant and negative and significant at the10% level,
respectively. Again this result is not in conformity with the widely accepted notion that
governments are forced into privatization when they run high deficits. This points out that
governments take a long term strategic view regarding the infrastructure sector and its
management and proportion of SOE share sales is not determined by immediate crisis of
budget deficits. We were expecting to see a positive significant coefficient for the
MARTCAP for the core sector, as a higher market capitalization would mean higher

71
demand for SOE shares and thus higher privatization, especially in this sector that
requires very large investment.
Recent works have shown a positive relation ship between stock market
development and privatization ( Demirguc-Kunt and Levine, 1994; Banerjee and Munger,
2004) and have hypothesized that higher market capitalization indicates that adequate
resources and appropriate infrastructure exists for private sector to the buy the assets of
public sector enterprises and thus makes it easier for governments to implement
privatization programs and to that extent it may appear that our results are not in
conformity with previous findings. There is one critical difference. These studies have
taken total privatization proceeds as dependent variable, while in this work privatization
proportion of each transaction is dependent variable. While it may be true that countries
with developed capital markets may be able to generate more privatization revenue by
selling more SOEs in any year, they may also sell lower proportion of shares of any one
particular SOE in any one year to decrease the supply of shares of that SOE and thus to
sell at higher price. A developed market capital market will mean low cost of accessing
the market for sale of SOE shares and thus the government may sell SOE shares in small
tranche in many transactions over a period of time.
Privatization programs in the developed and in developing countries provide
support for this pattern. U.K. and India both have well developed capital markets and
robust financial institutions. British Petroleum’s government holdings of 66% were
completely sold off in four phases in 1977, 1979, 1983, and 1988, by selling 15%, 6%,
14.6% and 31.4% respectively. In India, the government disinvested 74% of its holding
in Videsh Sanchar Nigam Ltd. (VSNL) in five transactions in the years 1991, 1996, 1998,

72
1999 and 2001. Therefore, it seems plausible that well developed capital markets provide
flexibility to the governments in determining the proportion of shares to be privatized in
each transaction. Politicians use this flexibility to lower the political cost by lowering the
privatization proportion.
Coefficients of the geographical location of the countries indicate that Latin
American countries have privatized higher proportion of their core sector SOEs, while it
has been rather low in the South Asian countries. Though one will tend to believe that
whole economies have been privatized in transient economies , surprisingly, privatization
of core sector public sector has been lower in these economies compared to our base
group of Sub-Saharan African countries. This result finds support in the recent work of
Bortolotti and Faccio (2004), who use data from OECD countries to show that
privatization process has been carried out reluctantly and governments are either the
largest shareholders or have substantial veto powers in almost two thirds of formerly
state-owned firms.
Partial privatization in intermediate sector: Table 2.6 has the results. These are
consistent with the results for the core sector except for three significant differences.
First, coefficient for SAGRI is significant and negative. Industries that may affect
the agriculture sector most, like fertilizer, petrochemicals, agricultural machinery
manufacturing and banking have been classified in the INTER category and thus we will
expect that an economy with higher share of agricultural in its GDP will be relatively
more sensitive to privatization of industries in this sector and may also face opposition
from storm farm lobbies. Second, the coefficient for CONCOR is negative and
significant. This indicates that politicians take different view of corruption in determining

73
Table 2.6 : Privatization proportion model for intermediate sector

Dependent variable: Privatization proportion for intermediate sector transactions
Explanatory variable I II III IV
SAGRI -0.259 -0.266 -0.515 -0.3853
0.511 0.497 0.082 0.163
MARKTCAP -0.111 -0.176 -0.222 -0.2175
0.005 0.005 0.000 0
DEFGDP 0.157 0.145
0.784 0.798
GDPGR -0.306 -0.311 -0.354 -0.4072
0.402 0.390 0.265 0.197
FORBUY -3.626 -3.616 -4.576 -4.888
0.352 0.353 0.173 0.143
JOINBUY -6.494 -6.433 -5.375 -5.149
0.095 0.096 0.134 0.152
POLCON -1.840
0.887
LEGRAL 8.123 7.902 -4.987
0.413 0.420 0.694
VOACC -6.308 -6.497 -4.987 -6.778
0.301 0.275 0.370 0.205
REGQUA 42.708 42.241 44.319 42.676
0.000 0.000 0.000 0
CONCOR -22.329 -22.001 -16.014 -14.14
0.001 0.001 0.003 0.007
GDPCAP 0.001 0.001 0.000 0.0008
0.455 0.463 0.736 0.556
OPEN 0.185 0.181 0.175 0.18612
0.012 0.008 0.003 0.002
LATAM -10.470 -10.471 -12.324 -11.827
0.078 0.077 0.027 0.031
CENAM 32.440 32.220 -9.180 -9.56
0.129 0.131 0.489 0.459
SOUAS -19.035 -19.581 -13.733 -15.779
0.044 0.023 0.037 0.011
EASAS -17.132 -17.232 -13.768 -17.49
0.010 0.009 0.017 0.002
EASEU -19.445 -19.556 -21.419 -21.63
0.034 0.032 0.005 0.004
NORAF -7.740 -7.782 -14.519 -17.496
0.420 0.417 0.053 0.013

Values in italics are p values for the estimated coefficients.

74
the proportion of privatization of SOEs in different sectors and privatize more with high
prevalence of corruption either to get kick-backs or, if we take a benevolent view,
perceive privatization as a strategy to control corruption. But if we take benevolent view,
then its hard to explain why they do not apply the same principle to the core sector
industries. Thus kick-back hypothesis looks more plausible, though we can not test that in
this study. Third, VOACC is negative as expected but no more significant while,
REGQUA is both positive and significant.
Partial privatization in competitive sector: Table 2.7 has the coefficient estimates. As
expected we find that all the three institutional variables are significant. VOACC estimate
is negative but much smaller than core sector estimate, indicating that comparatively
lower proportion of core sector is privatized in more pluralistic countries with open
governance institutions than competitive sector. As expected we find that privatization
proportion is lower when buyers are foreign firms. Countries that have higher GDP
growth rates tend to have higher privatization in competitive sector.
Looking at the control variables, two observations standout. The coefficient of
OPEN is positive and significant indicating that countries that have large foreign trade
tend to privatize more. This result is in conformity with Banerjee and Munger (2004). All
geographical locations coefficient that are significant, are negative. Thus, for the
competitive sector, highest proportion of privatization has taken place in Sub-Saharan
countries, while it has been lowest in the South Asian and Eastern South Asian countries.
Privatization and transfer of control: Webster dictionary gives the meaning of ‘privatize’
as ‘to make private; especially: to change (as a business or industry) from public to

75
Table 2.7 : Privatization proportion model for competitive sector

Dependent variable: Privatization proportion for competitive sector transactions
Explanatory variable I II III IV
SAGRI 0.399 0.586 0.585 0.330
0.049 0.000 0.000 0.000
MARKTCAP 0.044 -0.049 -0.049
0.263 0.121 0.119
DEFGDP -0.121
0.717
GDPGR 0.154 0.656 0.655 0.942
0.473 0.000 0.000 0.000
FORBUY -4.643 -5.447 -5.494 -5.659
0.025 0.000 0.000 0.000
JOINBUY 6.609 1.223
0.092 0.680
POLCON -29.140 -17.754 -17.848 -7.344
0.000 0.000 0.000 0.079
LEGRAL 35.102 26.400 26.441 20.789
0.000 0.000 0.000 0.000
VOACC -11.680 -14.480 -14.220 -5.716
0.001 0.000 0.000 0.007
REGQUA 32.572 30.951 30.942 18.810
0.000 0.000 0.000 0
CONCOR -30.491 -20.096 -20.108 -22.417
0.000 0.000 0.000 0.000
GDPCAP 0.005 0.005 0.005 0.003
0.000 0.000 0.000 0.000
OPEN -0.038 0.082 0.082 0.119
0.305 0.001 0.001 0.000
LATAM -7.508 -6.537 -6.501 -7.240
0.061 0.041 0.043 0.014
CENAM 2.853 12.792 12.799 1.364
0.755 0.010 0.009 0.722
SOUAS -10.393 -12.013 -11.975 -18.332
0.011 0.000 0.000 0.000
EASAS -8.148 -16.424 -16.392 -19.890
0.054 0.000 0.000 0.000
EASEU 8.184 5.168 5.109 -2.043
0.048 0.060 0.062 0.313
NORAF 0.954 -4.890 -4.858 -5.107
0.844 0.150 0.153 0.112
Values in italics are p values for the estimated coefficients.

76
private control or ownership’. In economic literature, as well as in economic policy
debates, privatization is referred to as the transfer of ownership and control of State-
owned enterprise (SOE) to the private firms. But our data and many other studies
(Bortolotti, Fantini, and Siniscalco, 2000; Bortolotti and Faccio, 2004) indicate that only
few governments have transferred full ownership and control of SOEs to the private
sector. In our dataset of 4760 privatization transactions in 43 countries, more than two
third transactions were partial accounting for more than 90% privatization revenue
indicating that full privatization is even less in case of large privatization programs.
In order to get in insight into microeconomic, political and institutional
determinants of privatization that leads to transfer of control we create a dependent
variable CONTROL. Ideally we will like to have direct information if the privatization
transaction stipulated transfer of control to private firm, in absence of this information,
we make a reasonable assumption that any transaction that has more than 50%
privatization proportion leads to transfer of control. Thus CONTROL is a binary variable
that takes value 1 if PRIVTPRO is greater than 0.5 and 0 otherwise.
Table 2.8 describes estimates of coefficients of microeconomic, political and
institutional variables. Higher budget deficits lead to increase in transfer of control while
high GDP growth rate leads to lowering the transfer of control. This finding supports the
hypothesis that governments transfer control of SOEs when faced with crisis. As
expected coefficient of FORBUY is negative and significant, indicating that in
privatization programs governments are reluctant to transfer control to foreign buyers,
while do not differentiate between domestic and joint ventures of domestic firms with
foreign firms as JOINBUY coefficient is insignificant. POLCON has high positive and

77
Table 2.8: Multivariate model for control

Dependent variable: Control
Explanatory variable I II
SAGRI 0.0051
0.0600
MARKTCAP -0.0002
0.6850
DEFGDP 0.0158 0.0123
0.0000 0.0000
GDPGR -0.0014 -0.0041
0.5850 0.0860
FORBUY -0.0538 -0.0885
0.0360 0.0000
JOINBUY 0.0030
0.9340
POLCON 0.0348 0.1622
0.6970 0.0270
LEGRAL 0.0466
0.5630
VOACC -0.1096 -0.1730
0.0120 0.0000
REGQUA 0.3153 0.2514
0.0000 0.0000
CONCOR -0.3153 -0.2266
0.0000 0.0000
GDPCAP 0.0000 0.0000
0.0960 0.0270
OPEN 0.0004 0.0000
0.3690 0.8160
CORE -0.3491 -0.3313
0.0000 0.0000
INTER -0.1643 -0.1576
0.0000 0.0000
LATAM 0.0689 0.0645
0.1110 0.0910
CENAM -0.0900 -0.0704
0.4500 0.5030
SOUAS 0.0850 -0.0835
0.1140 0.0480
EASAS -0.1657 -0.2603
0.0010 0.0000
EASEU 0.1900 0.1885
0.0000 0.0000
NORAF 0.1070 -0.0301
0.0700 0.5660
Values in italics are p values for the estimated coefficients.

R-Sq (adj) 16.5000 17.3000

78
significant coefficient. This means that those countries that have lower checks and
balances over the decision power of the chief executive are likely to have higher
proportion of SOEs whose control is transferred to the private firms after privatization.
Thus we have democratic countries like India that have no privatization transaction in
which control has been transferred, and most of the Sub-Saharan countries who have
more autocratic regimes to have more than 50% privatization transactions leading to
transfer of control.
Coefficients of institutional variables provide an insight into the decision process
of politicians. While coefficient of REGQUA is positive and significant, coefficients of
VOACC and CONCOR are negative and significant. This reflects that politicians appear
to take into account the popular negative perceptions on privatization leading to transfer
of control. Good quality regulatory framework and bureaucracy will increase the
potential benefits of privatization while safeguarding against negative distributional
impacts and thus we expected to have a positive relationship. It’s not clear as why more
corruption leads to higher transfer of control. As discussed earlier, corruption may
provide more rent seeking opportunities to the politicians in the privatization and they
want to sell off full units to get maximum ‘bribes’, while on the other hand politicians
may privatize full units to isolate them from corrupt officials and government
interference.
Governments are least enthusiastic in transferring control of SOEs in the
infrastructure sector, while trade openness of a country has insignificant influence on
privatization policies. This also hints at the heterogeneity in reform policies across
governments in different spheres. Coefficient of GDPCAP is positive and significant

79
indicating that relatively richer countries tend to have higher proportion of transactions
leading to transfer of control. As expected there are significant regional variations. Latin
American countries and transition economies are the only regional grouping that has
transferred higher proportion of SOEs control to the private firms as compared to Sub-
Saharan countries, while it has been much lower in South Asian and East Asian countries.
2.6 Conclusion: Quality of institutions, which has been captured by three distinct
variables of voice and accountability, (VOACC), quality of regulation (REGQUA) and
control of corruption (CONCOR) , has emerged as most consistent and significant
determinant of proportion of partial privatization, but the effect is not uniform across
three broad categorization of industries. We find indirect support for the contention of
Birdsall and Nellis (2004) observation that public perceptions of privatization programs
are negative. Politicians in countries with high level of public debate and freedom of
expression try to reduce the negative political fallout of privatization programs by
lowering the proportion of privatization and thus the negative impact of this variable is
most pronounced in core sector.
Results with regards to control of corruption are both surprising and insightful.
When we look at overall transactions, we find that countries that have low level of
control of corruption tend to have higher proportion of privatization. Though this result is
insightful in itself, it’s not clear if it’s because politicians are able to get away with
corrupt privatization deals (with high kick-backs) or because politicians perceive high
economic benefits to economy as public sector is corrupt and public resources are being
used for private benefit by public managers including some politicians. The sectoral
results do provide some insight into politicians’ motives. While countries with low

80
control of corruption have high privatization proportion in the intermediary and
competitive sector, privatization proportion is low for core sector. Rent seeking
opportunities are large in core sector due to very high value. If politicians’ were
motivated by private benefit, we will expect to see high privatization proportion in core
sector also in high corruption countries. Thus in our analysis politicians do not emerge as
the ‘villain’, but rather appear to have ‘good’ intentions in pursuing privatization
programs.
Political variable appear to play a role only in determining partial
privatization proportion in the core sector and whether control will be transferred or not.
We do not find either political constraint or political fractionalization to play any
significant role in partial privatization, though studies ( Bortolotti and Pinotti, 2003;
Banerjee and Munger, 2004) have found a strong relationship between privatization and
political factors. There are two possible reasons. First, both of these studies have taken
either total number of privatizations or total privatization proceeds as dependent variable.
Aggregating all privatization transactions, either by count or by revenue, one loses the
information contained in each individual privatization transaction. While the objective of
this study is to find out the determinants of privatization proportion of each firm, those
studies have focused on macro level country wide privatization programs. To our
knowledge, this is the first work that attempts to look at determinants of partial
privatization. Second, both those studies do not look at the sectoral distribution of
privatized companies as they view privatization of a hotel or a steel manufacturing unit or
a power generating unit being influenced by the same set of factors in the same manner.
Our study looks at the sectoral distribution and we find overwhelming support for our

81
hypothesis that political constraints play a role in core sector and transfer of control only.
One other possible reason for the divergence of our results from others is that we are
looking at privatization proportion and not at privatize or not. The privatization
proportion is more likely to be determined by economic and institutional factors while the
preceding decision of privatize or not is likely be more political.
Regarding microeconomic determinants, our results provide insights into decision
making process of politicians. DEFGDP which captures budget crisis to is not significant
in any sectoral model, but is positive and significant in determining transfer of control
only. Crisis may be driving the decision to privatize or not, but not the exact proportion
of privatization. Other surprising result is that whether buyer firm is domestic or foreign
does not play any role in determining privatization proportion in the core sector, though
its negative in determining transfer of control. Similarly we find that market
capitalization has a negative impact only for the core sector.
We do find support for regional variations in partial privatization across different
regions, both in different sectors and in transfer of control. Our surprising result is that
privatization proportion is relatively very high for the poorest region, i.e., Sub-Saharan
countries.
In the next chapter we will analyze the theoretical implications of these empirical
results and will identify issues that need to be addressed. Thereafter, we will provide a
simple model that takes into account institutional setting in which politicians operate and
will study the tradeoffs made by them in balancing between political benefits and
economic costs.

82
Chapter 3: Political Economy of Partial Privatization
Introduction: As described in chapter 2, there is a great variation in the extent of partial
privatization across countries. Data suggest that this heterogeneity exists regardless of
income levels as measured by per capita income or capital market development, as
measured by market capitalization as percentage of GDP. Not only do many developing
as well as developed countries continue to maintain large public sectors, but also most of
the privatization transactions have been partial. Bortolotti and Faccio (2004) study 141
privatized firms from OECD countries over the period 1996 through 2000 and find that as
of 2000, governments were still the largest shareholder and retained voting control in
62.4% of privatized firms. This is puzzling in the face of growing consensus among
economists over the benefits of privatization. While there is a rich theoretical and
empirical literature on the relative merits and performance of public and private firms,
little is known about why governments prefer partial privatization and what determines
the proportion of privatization.
While the economic benefits of privatization are more widely accepted, at least in
theory, its political acceptance is still suspect. The widespread presence of corruption in
public spheres is a critical factor leading to government decisions to implement
privatization programs. This is especially so in democratic polities, as people fear that the
politicians and bureaucrats may enter into side contracts with private firms for private
benefits and undersell the public enterprises, which are regarded as national wealth. The
experience of Eastern European countries has shown that these fears are not unfounded.
The Czech Republic was exceptional in the “success” of its privatization program,
implemented through a voucher program (Kotrba and Svejnar, 1994). The share of

83
private sector in GDP increased from 3% in 1990 to 90% in 1996
21
, and even by 1993-
1994, Czechs were the largest per capita shareholders of any country, including the
United States (Schwartz, 1997). But this “success” started looking muddy by 1997. It was
then that the Czech public learned that dishonest operators had systematically
appropriated assets of many of the country’s best companies, municipalities and banks,
primarily through corrupt collusion or by the benign neglect of the state (Schwartz,
1997). This confirmed the fears that inappropriately designed privatization programs, that
fail to put effective regulatory/legal framework in place in time, might result in the
private appropriation of national wealth (Turnovec, 1999). Since the privatization
decision is taken by the politicians, the key to understanding partial privatization lies in
asking if politicians in office represent the interest of the electorate or that of themselves
alone or some combination of the two. Put in another way: what is the objective function
of politicians in implementing privatization programs?
The literature in the incomplete contracting framework has looked at the question
of partial privatization from two perspectives. One perspective focuses on shortcomings
in the government’s ability to commit to market-friendly tax and regulatory policies. This
inhibits private firm’s investment in privatization programs and necessitates government
partial ownership in production as a substitute (Weingast, 1995; Levy and Spiller, 1996;
Che, 2001). This view predicts that those factors that raise the opportunity cost of public
funds to politicians should increase the size of the public sector and therefore should
lower the partial privatization proportion, all other things being equal. This prediction is
based on the argument that such factors make it hard for the politicians to convince

21
Report of Ministry of Finance of the Czech Republic on the Law on Capital Market Regulation, 1997.

84
private firms that they would refrain from using regulatory policies or manipulating tax
policies to extract quasi-rents from private firms (Esfahani, 2000).
The second perspective focuses on the role of incomplete contracting over public
firm’s inputs or outputs that may be important to the politicians, but cannot be directly
contracted or influenced by them unless the government has direct control over some key
aspects of the firm (Hart, Shleifer and Vishny, 1997; Rajan and Zingales, 1998). The
prediction of this perspective is opposite to that of the first. If the objective of public
ownership is to control some aspect of the production process, then the partial
privatization proportion is likely to be high as the cost of public funds rises (Shleifer and
Vishny, 1994. 1998; Boyko etal., 1996). Thus, it’s not very clear from the theoretical
literature as to what drives the partial privatization proportion.
Privatization has also been viewed as a policy instrument for reducing the
impact of political factors on economic incentives and performance (Vickers and Yarrow,
1991). More specifically, private ownership reduces political influence and increases the
capital market influence, though the precise impact of ownership transfer would depend
upon the specific characteristics of the relevant regulatory environment and the capital
market. The case for privatization is based on the implicit assumption that there is
“government failure” to the extent that public policy is likely to operate in such a way as
to impede the efficient functioning of markets. On the other hand, the policy
environments in which firms operate are influenced by the incentive structure faced by
the policy makers, which depends on a number of political and institutional factors.
These factors include interest groups, the degree of regulatory discretion, transparency in
public decision-making, enforceability of contracts, and the institutional framework of

85
oversight, balance of political forces, and bureaucratic structure and goals. The market
for regulation
22
itself is imperfect, and can produce inefficient levels of government
intervention. But, for many public enterprises the state’s influence or intervention may
persist in response to perceived market failures and/or politically undesirable
redistributive effects that might result from privatization.
Recognizing the political nature of the privatization process, and keeping in view
the empirical results of chapter 2, in this chapter we construct a political economy model
that informs about optimal privatization contract. The model takes into account
simultaneously the inefficiencies of the political system, the institutional structure and the
tradeoffs made by politicians between efficiency benefits and political costs. Section 2
develops the model, section 3 discusses contracts when institutional quality is not perfect
and section 4 discusses the political economy model. Section 5 concludes.
3.2 Model: In this section we construct a simple model of political economy of
privatization. We do not conceptualize government as a benevolent social welfare
maximizer. Instead, our government is controlled by politicians who may derive private
benefits from public control of the firm. In our conceptualization, private benefit does not
necessarily mean direct corruption as it may just represent political power, prestige or
ideological commitment.
We consider privatization of a large public firm that requires specific specialized
managerial and technical expertise to manage its operations efficiently. We assume that
there are but few other private firms that produce similar types of goods and services, or

22
We use the metaphor introduced by Yarrow (1999).

86
have specific expertise to manage the operations of the public firm, and so in our model
there are few prospective buyers of the public firm.
Prospective Buyers: One of the rationales for privatization is that the private firms,
which are prospective buyers of the public firm, may bring better managerial skills,
technology and access to financial markets, which may increase the value of the
privatized firm. The prospective buyers differ in their ability to manage and operate the
public firm. The valuation of the public firm by the prospective buyer depends upon its
own technical and managerial competence. Buyers participating in the privatization
process have different technical and managerial capabilities, denoted by the parameterβ .
The efficient prospective buyer is characterized with a highβ , which means that she has
superior technology to reduce cost and/or improve the managerial efficiency of
operations compared to the prospective buyer with a lowβ . To keep the model simple,
without loss of generability, we normalize the valuation of the public firm under
government control to unity.
While the valuation of the public firm depends on the capability parameterβ of
the prospective buyer, it also depends on the effort e exerted by the buyer. It is
reasonable to postulate that the net realized value of the privatized firm will be
determined by a combination of the technical and managerial competence of the private
management and the effort put in by that management in either restructuring or day-to-
day operations. Following Bolton, Pivetta, and Roland (1997), we assume that ability
interacts multiplicatively with effort to influence the net valuation of the privatized firm.
This assumption seems reasonable since a private firm with high technical competence
will need to put in effort in reorganizing the privatized firm before its value could be

87
realized. Even a firm with low technical competence may be able to realize high value if
it puts in sufficiently high effort in restructuring and operating the privatized firm. While
in their model the interaction between ability and effort determines the probability of
producing high output, in our model the multiplicative interaction between ability and
effort determines the net realized valuation of the public firm. Thus in our model, a
prospective buyer of type β who puts in effort e realizes the privatized firm’s value
as e β . Further we assume that effort e is a continuous variable R
+
∈ . High effort ( ) e
leads to higher valuation but effort is costly. Following the literature, we assume the cost
of effort to be convex in effort and given by
2
( )
2
e
c e = . Further, we assume (1) that the
prospective buyer firm’s share of the total value of the privatized firm is the same as the
proportion of privatization, and (2) that the government neither puts in any effort nor
does its partial ownership in the privatized firm affect the functioning of the privatized
firm. This assumption implies that, even after partial privatization, operational control is
transferred to the private firm and no political interference takes place thereafter.
Thus, the utility function of the prospective buyer firm is given by:

2
2
B
e
U e p αβ = − − (1.1)
where the prospective buyer firm obtains α proportion of shares (partial privatization
proportion is given by α ) of the public firm by paying price p to the government.
Government: We assume the government’s privatization policy is determined by
politicians who derive private benefits from public control over the firm. Thus there are
three components in the government’s objective function, namely, efficiency benefits, the

88
social cost of public funds and loss of private benefits to the politicians. This approach is
based upon the empirical observation in chapter two, where we find that privatization
proportion does not appear to be driven by technical considerations of efficiency alone
and institutional variables appear to be most consistent and robust determinants of the
proportion of privatization announced by the government. The efficiency gains of
privatization realized by the government (but shared with the private firm) are
proportional to the shares held by the government. Since the government holds the
proportion of shares1 α − , the benefits due to efficiency gain are given by( ) 1 e α β − ,
where the private buyer firm is type β and puts in effort e in restructuring and/or
operations.
There is a social cost of the public funds as the government can raise revenue only
through distortionary taxation (excise tax, e.g.). The social cost of one unit of revenue
raised by the government is ( ) 1 λ + ,whereλ is the dead-weight loss in the economy as
the government revenue is raised through distortonary taxes
23
. The social cost factor λ is
generally high for developing countries compared to developed countries (Laffont,
1987)
24
. This factor captures the efficiency of tax collection by the government. The
value of λ is estimated to be 0.3 for developed countries (Laffont, 1996), while in a
World Bank study, Jones, Tandon, and Vogelsang (1990) estimate λ to be as high as
2.48 for Philippines. The value of λ will depend on the stage of development of a
country, the effectiveness of government, and the efficacy of public institutions in

23
See Browing (1976), Rosen (1978) and Ballard, Shoven and Whalley (1975), for various approaches to
estimate the cost of public funds due to distortive taxes.
24
See Laffont (1987), The Economic Journal,Vol.97, pp 17-31.

89
curbing corruption and improving efficiency of tax collection. As one unit payment made
out by the government has ( ) 1 λ + unit social cost, and similarly a unit of revenue
received directly in the treasury has ( ) 1 λ + unit social value to the society. Therefore, the
government values the revenue p received from the sale of α proportion of the public
firm as ( ) 1 p λ + .
In our conceptualization, government is not an ideal social welfare maximizer. As
the goal of this research is neither prescriptive nor normative, it aims to model the
observed pattern of privatization programs across countries and to that extent is driven by
observed patterns. We model the politician as one who derives private benefit from
public control. This private benefit may be in terms of ability to influence production,
investment allocation, employment decisions or perhaps to obtain kick-backs in public
firm’s commercial contracts. Private benefit may even be political wherein implementing
privatization or not implementing privatization is construed as an ideological signal to
support base. Private benefit to politicians of public control will depend upon economic,
regulatory, political and judicial institutions. An efficient, responsive and effective
institutional setting will mean limited opportunities available to politicians for deriving
private benefits and therefore institutions play a critical role in privatization policies.
Some of the features of LDCs that further heighten the importance of institutions in these
countries are as follows:
a) High auditing costs: LDCs often lack detailed accounting systems, transparent
administrative procedures and trained auditing staffs. The inefficiency of auditing

90
makes corruption more widespread and provides a relatively easy venue for
private benefits to politicians.
b) Lack of constitutional control of the government: LDCs often lack checks and
balances provided by well-functioning democracies, higher courts, and
government auditing bodies that counter politicians’ executive powers. This
makes it easier for the interest groups to capture government and also for
politicians to implement partisan programs that excessively favor the voters
supporting them.
c) Lack of political democracy: The lack of well functioning political institutions
increases the uncertainty of future regulations and makes it difficult for the
government and regulatory institutions to make credible commitments to long run
policies. Thus the economic policies of developing countries are more sensitive to
ratchet effects and renegotiations constraints.
d) Poor financial institutions: Developing countries have poorly developed capital
markets which increases the cost of funds. More so, sheer lack of wealth makes
limited liability constraints more binding.
In our model parameter θ captures opportunities available to politicians to derive private
benefit from public control. A low θ denotes high quality of institutions, i.e., good
quality of regulation, low level of corruption, high level of accountability, and high score
on rule of law, which means a relatively lower possibility of rent seeking from public
control. Countries with low θ are likely to have a fair judicial court system and other
institutions like an independent election commission, a general audit office and a free
press. Thus it is reasonable to model the rent accruing to politicians from public

91
ownership to be proportional to the quality of institutions parameterθ . Further, we
assume that the politician’s political cost or loss of ability to extract rent from a public
firm is a convex function of the privatization proportion. For simplicity, we take the
political cost function to be continuous and it’s given as
2
2
α
θ . This political cost
function can also be seen as the politicians’ loss of private benefits due to the
privatization of the proportion α of the public firm. Thus we can write the politicians
objective function as:
( ) ( )
2
1 1
2
P
U e p
α
α β λ θ = − + + − (1.2)
where the first term is the efficiency gain due to privatization, the second term is the
social value of privatization proceeds and the third term is the political cost to the
politicians due to loss of control over the public firm.
Timing of contracting: We consider the case in which the government or
politicians first observe the prospective buyer’s type β and the institutional quality
parameterθ . Then, the government offers a “take it or leave it” contract consisting of two
dimensions, a proportion of shares α and a price p to the prospective buyer. The private
firm accepts or refuses the contract. If accepted, then the private firm determines the level
of effort and the contract is executed. Figure 3.3 gives the timing of the privatization
contract.

92
Government
observes
Buyer type and
Institutional
quality
Buyer firm
accepts or
rejects the
contract
Buyer firm
determines
effort level
Government
offers
contract
( ) , p α
e
Contract
is
executed
0 t =
1 t = 2 t = 3 t = 4 t =

Figure 3.1: Timing of contracting under privatization
3.3. Benchmark case: No private benefit of public control: We first consider the case
when the quality of institutions is very good and politicians have virtually no
opportunities for engaging in rent seeking activities in the public firm. In this case,
parameter 0 θ = , and privatization does not entail any political cost to the politicians.
Lemma 1: When politicians derive no private benefit from controlling the public firm, the
government will offer a contact
( )
( )
2
* *
2
1
, ,
1
2 1
p
β
α
λ
λ
 
≡ 
 
−
−
 
; and the buyer firm will
participate and put in the optimal effort level
( )
*
1
e
β
λ
=
−
.
Proof: The government knows the utility function of the buyer firm and given any
contract ( ) , p α offered by the government the buyer firm will put in effort such that
B
U
given by equation (1.1) is maximized. Thus we have,
*
0 0
B
U
e e
e
αβ αβ
∂
= ⇒ − = ⇒ =
∂
.
The government will offer a contract in such a manner that the buyer firm’s utility is
pinned down to its reservation utility which is assumed to be zero. Thus the price offered
by the government given the privatization proportionα
,
is given by

93
2 2
* * *2 *
0
2
e p e p
α β
αβ − − = ⇒ = . The politicians in the government will determine
the privatization proportion such that their utility given by equation (1.2) is maximized.
Substituting
*
e and
*
p in equation (1.2) and maximizing with respect to α gives
us
( )
*
1
1
α
λ
=
−
. Substituting
*
α , in
*
e and
*
p gives Lemma 1.
We can clearly see that, when politicians derive no private benefit from the public
firm and thus the institutional quality is perfect, the government will engage in full
privatization.
3. 4 Privatization contracts when institutional quality is less than perfect: Now we
consider the more realistic situation in which politicians derive some private benefits
from public control of production in the economy. In this case, the politicians have to
make a tradeoff between (a) the efficiency benefits that may accrue due to private
production, (b) productive and socially useful investment of the privatization proceeds,
(c) any lowering of taxes as the privatization proceeds allow privatization revenue to
Lemma 2: When politicians derive private benefits from controlling public firm, the
government will offer a contact( )
2
2
2
2
1
, ,
1
2 1
r r
p
β
α
θ
θ
λ
λ
β
β
 
 
 
≡
 
 
   
 
+ −
+ −  
   
 
 
     
; and the
buyer firm will participate and put in the optimal effort level
*
2
1
e
β
θ
λ
β
=
 
+ −
 
 
.

94
substitute for taxes, and (d) the loss of rent seeking opportunities or other political benefit
that they get from public control of the firm.
From Lemma 2, we can see that privatization proportion decreases with increasing θ
for a given level of social cost of public fund parameterλ . In our model, high θ indicates
poor institutional quality and thus higher opportunities for rent seeking to the politicians.
From Figure 3.4, we can see that full privatization takes place as long as the institutional
quality parameter is less than
2
λβ , which we call the threshold institutional quality
parameter for partial privatization. This threshold will shift to right if β increases, i.e.,
the managerial/ technical capabilities of the private sector improve. The economic
intuition behind this is that the politicians’ perceive greater efficiency gains if private
sector is efficient and thus are willing to give up more ‘rent seeking benefits’ of public
control by privatizing more.

2
λβ
1
Institutional quality parameter
θ
Privatization
proportion
α

Figure 3.2: Optimal privatization proportion as institutional quality declines

95
The institutional quality parameter will depend upon the quality of regulation, the
efficiency and effectiveness of the judiciary, political and administrative accountability
and checks and balances imposed on executive and legislature by the constitutional
design. Thus, in a very simple setting, our model predicts that the relationship between
privatization proportion and institutional quality parameter will be negative, that is, as
institutions will become less and less effective
( )
θ ↑ , there will be less privatization. This
theoretical prediction is in conformity with our empirical observation from chapter 2,
where we find that as the quality of regulation gets poorer, the privatization proportion
declines.
Figure 3.5 shows the relationship between the privatization proportion α and the
social cost of public funds parameterλ . The first interesting observation is that even
when tax administration is very efficient and there is no corruption, i.e., the social cost of
public funds parameter λ is zero; full privatization may not take place if politicians
2
θβ
−
1
Social cost of public funds parameter
Privatization
proportion
α
λ
( )
2
2
β
θ β +

Figure 3.3: Optimal privatization proportion

96
derive private benefits from public firms. The distortion in the privatization proportion
will depend upon the institutional quality parameter θ relative to a measure of private
sector efficiency
2
β . The distortion will increase as institutional quality declines relative
to efficiency of the private firm. Second, we observe that as social cost of public funds
increases, the privatization proportion increases. The intuition behind this result is that, as
the social cost of public funds increases, politicians value privatization proceeds more
and thus choose higher privatization, even if that entails losing out on some rent seeking
opportunities. The third observation is that, given a level of institutional quality and
efficiency and effectiveness of tax collection administration, as private firm’s capabilities
increase, more privatization takes place. This may explain why privatization has been
common even in countries with very different institutional settings and corruption levels.
One interesting result is that the effort level of private buyer firm also declines as
institutional quality declines. The intuition behind this result is that lower institutional
quality leads to a lower privatization proportion and, since the private firm gets only a
partial share in efficiency gains brought about by its effort, its incentive to put in effort
declines and thus the optimal effort put in by the private firm declines.
Thus far we assumed the structure of the government to be homogeneous and
assumed that the institutional quality parameter θ directly measures the rent seeking
opportunities of homogeneous politicians. In the next section we consider a simple
political economy model with two types of politicians, representing two different voter
segments, who have heterogeneous valuations of rent seeking opportunities.

97
3.5. A Political economy model of the government: We know that, in a world of
complete contracting, a benevolent government owned firm can always do as well as a
privately owned firm, i.e., ownership does not matter (Williamson 1985; Grossman and
Hart 1986; Sappington and Stiglitz 1987). This means that whatever is achieved by
privatization in efficiency gains can be achieved by writing an appropriate contract
25
.
Why, then, are we so often told that the performance of public enterprises will improve
by privatization? We think that the answer to this question primarily lies in the sphere of
political economy and not so much in the domain of conventional economics. The factors
that influence government decision-making will determine the efficiency of operations of
the public enterprise to a large extent. Therefore, the issue of privatization needs to be
analyzed in an incomplete contracting setting. The introduction of this incompleteness
will allow us to analyze how the private objectives of politicians affect the outcomes,
since they cannot be perfectly controlled by appropriate contracts. Next, we introduce
political economy issues in our setting through a simple majority rule model.
In particular, we consider two types of politicians representing two different
consumer types. One group of voter/citizens values the public control more than the other
group because they consider themselves to be beneficiaries of the rent seeking
opportunities. We represent these two groups of voter/citizens as types 1 and 2 and
assume that they appear in proportions μ and( ) 1 μ − , respectively, in the population.
These two consumer types are represented by two types of politicians who we denote as
types 1 and 2 politicians, respectively. Type-1 politicians derive more rent from public

25
This argument is valid even for economies that have decentralized information, moral hazard, adverse
selection, and non-verifiability issues. Sometimes the optimal contract can be implemented by appropriate
choice of property rights (See Tirole 1994).

98
control and thus for them the political cost of privatization is given by
2
2
α
ηθ , where
1 η > . For simplicity we assume that type-2 politicians’ political cost of privatization
continues to be
2
2
α
θ . This introduces heterogeneity in politicians’ behavior which we
will exploit in developing a simple political economy model of the government to study
privatization contracts under different majority rules.
In this simple model we can think of type-1 politicians as those having socialist
leanings and belonging to the political parties to the left, while type-2 politicians
represent political parties having rightist orientations. This conceptualization is highly
stylized but is sufficient for the purpose of bringing out differences in privatization
policies under different situations with the same majority rules.
Privatization under a non-partisan government: As a benchmark, we consider the
case in which the government adopts a non-partisan privatization policy and thus
maximizes the expected utility of both type of politicians. As the politicians represent
their constituents truthfully, this in turn maximizes the expected utility of both types of
voters. Each type of politician’s utility is weighted by the proportion of voters/citizens
represented by that type to get the expected utility and the same is given by
( ) ( ) ( ) ( )
2
1 1 1 1
2
NP
U e p
α
α β λ μ η θ = − + + − + − (1.3)

99
Lemma 3: Under a non-partisan government that maximizes expected social welfare, the
government will offer a
contact( )
( )
( )
2
2
2
2
1
, ,
(1 1
(1 1
1
2 1
r r
p
β
α
θ μ η
θ μ η
λ
λ
β
β
 
 
 
≡
 
+ −  
  + −  
 
+ −
  + −
   
 
 
 
     
; and the buyer
firm will participate and put in the optimal effort level
( )
*
2
(1 1
1
e
β
θ μ η
λ
β
=
+ −  
+ −
 
 
.
Comparing Lemma 2 with Lemma 3, we see that under heterogeneous preferences
of citizens, the privatization proportion has declined. This is due to the fact that the
expected value of the institutional quality parameter θ has gone up to
( ) 1 , 1 θ θμ η η + − > , indicating a decline in institutional quality.
Privatization under partisan government: Until now we have assumed that institutional
quality parameter θ and the social cost of public funds parameter λ are observed by the
government and both types of politicians. Since these parameters will change with time
and with the changes in institutional environment, the optimal privatization contract will
remain contingent on the values of these parameters. But these parameters are dependent
on non-verifiable business conditions, and therefore the government must delegate the
design of the privatization contract to another institution that will be ex post informed
about the value of the parameters θ and λ . Politicians are the best candidates to take the
responsibility of residual decision-making.

100
In order to analyze this delegated privatization mechanism, we will construct a
simple random majority political model. The proportion μ of type-1 citizens fluctuates
randomly each period. With probability
1
2
, μ equals
1
*
2
μ > ; indicating that type-1 has
the majority. With probability
1
2
, μ equals
1
1 *
2
μ − < ; implying type-2 has the majority.
The task of choosing the optimal privatization contract design is delegated by each
majority to the politician, who has no personal agenda, and he faithfully represents the
interest of the majority that has chosen him. The politicians elected by type-1 citizens are
denoted as type-1 politicians and type-2 citizens elect type-2 politician. For the sake of
simplicity, we assume that privatization contacts are such that the participation
constraints of both type of politicians are met.
Lemma 4: When type-1 politicians are in the majority, then the optimal
privatization contract is given by( )
2
2
2
2
1
, ,
1
2 1
r r
p
β
α
ηθ
ηθ
λ
λ
β
β
 
 
 
≡
 
 
   
 
+ −
+ −  
   
 
 
     
; and the
buyer firm will participate and put in the optimal effort level
*
2
1
e
β
ηθ
λ
β
=
 
+ −
 
 
. When
type-2 politicians are in the majority, then there is no distortion in the privatization
contract and the same is given by Lemma 2.
From Lemma 4, we can see that when type-1 politicians are in power, they are
averse to privatization and may undertake too little privatization. Similarly when type-2
politicians are in power, they indulge in too much privatization. The privatization

101
proportion in Lemma 3 is optimal from the social welfare point of view as it maximizes
the expected utilities of both types of politicians and through them the citizens whom
they represent.
The simple political economy model indicates that the privatization proportion
will change since the majority rule shifts from one type of politicians to another, though
it continues to be determined by economy wide factors that persist even if the political
composition of the government changes.
3.6. Discussion and conclusion: Privatization has been a controversial economic issue in
recent times. In a world of complete contracting with a benevolent constitution it should
not matter (Williamson 1985; Grossman and Hart 1986; Sappington and Stiglitz 1987).
We introduce an incomplete contract in order to better understand a more realistic
situation. The main consequence of this incomplete contracting through a simple majority
rule model is that now the different political objectives of politicians will affect the
outcomes. Since the politicians who have been delegated the power to offer privatization
contracts cannot be perfectly controlled by appropriate contracts, the privatization
proportion will depend on institutional quality and the social cost of public funds.

102
Chapter 4: Political Economy of Privatization in India
The post-independence economic development of India can be characterized as a
half-full or half-empty glass. On the one hand India is a Nuclear power with a large,
diversified, and technologically advanced industrial sector, most of which has been
developed in public sector since the 1940s; on the other hand, given India’s natural
resources, manpower skills, entrepreneurship and developed financial markets, it has
fallen far behind a number of other developing countries with regard to the pace of
industrialization and economic growth. While a vast literature in economics has
addressed the subject of economic development in India over last fifty years, an extensive
literature on politics of economic development in India also exits. But these two streams
of literature have seldom intersected. This chapter reflects on the interconnections
between the process of industrial development in general and the public sector in
particular in the wider context of political democracy to understand the interaction
between economics and politics in independent India.
4.1 Industry Structure in Colonial India: Before the First World War, the only major
industries in colonial India were cotton and jute textiles, for which the country had
natural advantage. The Colonial government adopted more progressive industrial and
fiscal policy in the nineteen twenties, including the policy of discriminatory protection in
1922. These policies yielded dividends and between 1922 and 1939, the production of
steel ingots rose 8 times and the country was able to produce its entire steel requirement
by 1936. The production of other goods like matches, glass, hydrogenated refined oil and
several engineering products including electrical equipment also recorded significant
increases in the inter-war period.

103
The Second World War created conditions for increased demand for
manufactured goods and in response, industrial production recorded high growth rates.
Although several industries such as ferroalloys, diesel engines, pumps, bicycles,
chemicals like soda ash, super phosphate, and certain types of machine tools and simple
machinery were started on a modest scale during this period, the major impact of the war
was felt in the sector of medium and small-scale industries, such as light engineering,
pharmaceuticals, medicines and drugs. Immediately after the war, there was considerable
new investment in industries like rayon, automobiles, machine parts, locomotives,
fertilizers, and cement as the war and post-war industrial development was influenced by
the prevailing inflationary conditions and scarcities.
The major emphasis in industrial development in colonial India had been on
consumer goods industries; basic capital goods industries lagged behind. The production
of domestic iron and steel industry was only around 50 percent of the post-war volume of
demand. Even with the entire industrial production boost during World War II, the
industrial sector’s share in the overall economy in colonial India remained very low. In
1948-49, factory establishments accounted for only 6.6 percent of total national income,
while the total labor force engaged in such establishments was only about 2.4 million, or
0.8 percent of the total work force.
One interesting characteristic of the Indian economy at the time of the end of
British rule in 1947 was that the relative share of the public sector in the ownership of
productive capital assets was already large. The book value of gross fixed assets owned
by the Central and State governments, amounted to over Rs. 12 billion at the end of 1950-
51, and its distribution was roughly as follows: railway- Rs. 8.37 billion, irrigation

104
works– Rs. 2.3 billion and communication, electricity, industrial, ports and civil aviation
Rs. 1.23 billion. The value of productive assets in the private sector in 1950 was not
perhaps more than Rs. 15 billion. According to the census of Manufactures for 1949, the
net capital employed in twenty-nine groups of factory industries amounted to only about
Rs. 5.1 billion.
4.2 Evolution of Public Sector in India: India adopted an inward-oriented
development strategy immediately after gaining independence in 1947. The first Prime
Minister, Pandit Jawaharlal Nehru, outlined his vision of an industrially strong India
while addressing the 7
th
Annual Session of the All India Manufacturers’ Conference in
1947 as “Essentially, no good can come without political freedom and all other freedoms
flow out of political freedom. Now that we are on the verge of political freedom, we shall
run and bounce towards economic freedom”.
Post independence industrial policy accorded a dominant role to the state in the
articulation and implementation of economic development strategy in general and
industrialization in particular. The reason for this strategy can be traced to two factors.
First, leaders who led the independence movement identified the colonial economic
system with capitalism and perceived its liberal foreign trade policy as the colonial
government’s goal of exploiting the natural resources of the home country on one hand
and allowing unhindered access of foreign companies to domestic markets for export on
the other hand. Second, many of the leaders after independence were deeply influenced
by Fabian socialism and that contributed to their distrust of markets and to their emphasis
on a large role for the state in economy (Srinivasan, 2003).

105
The distrust of the private sector in the minds of political-economic policy makers
in the early 1950s is very clearly reflected in the First Five Year Plan approach
paper(1951) that explains the relation between the public and private sectors as “The
distinction between the public and the private sector is, it will be observed, one of relative
emphasis: private enterprise should have a public purpose and there is no such thing
under present conditions as completely unregulated and free private enterprise.”
By the time the First Five Year Plan was formulated in 1952, the preeminent role
of the public sector was fortified as a cornerstone of the development strategy of post
independent India as it recorded:
Whether one thinks of the problem of capital formation or of the introduction of
new techniques or of the extension of social services or of the over-all re-
alignment of the productive forces and class relationships within society, one
comes inevitably to the conclusion that a rapid expansion of the economic and
social responsibilities of the State will alone be capable of satisfying the
legitimate expectations of the people……..It does mean, however, a progressive
widening of the public sector and a re-orientation of the private sector to the needs
of a planned economy.

Thus, the first five year plan clearly stated that nationalization of existing private
enterprises was a low priority and the purpose of transfer of ownership can be achieved
by judicious regulation. The cornerstones of economic policies of the planned economy
as articulated in the First Five Year Plan (1952) were:
1. The role of private enterprise is to work within the framework of national policy
of industrial development
26

2. The competing demands on the available capital resources are so high that the
government must control capital issues and also regulate the uses of accumulated

26
Industries (Development and Regulation) Act, 1951 and its predecessor Industries (Development and
Control) Bill, 1949, provided detailed procedures for licensing of private investment.

106
funds. Prior permission (license) of the Government before establishment of new
industrial units as well as the substantial additions to existing ones was conceived
of as a direct way of regulating capital investment in the private sector.
3. In order to promote the flow of capital investment in the preferred sectors,
specific incentives will be provided to certain firms/industries.
4. Foreign investment should be channeled into fields of high priorities as set out in
the Five Year Plans. Foreign investment should be permitted in spheres where
new lines of production are to be developed, special types of experience and
technical skill are required or where the volume of domestic production is small
in relation to demand and there is no reasonable expectation that the indigenous
industry can expand at a sufficiently rapid pace.
The government’s control over industrial units and private investment was made
complete with the Industrial (Development and Regulation) Act, 1956, that, provided that
since the competing demands on the available capital resources are very high, the
government must control capital issues and also regulate the uses of accumulated.
Further, the Industrial Policy Resolution of 1956 stated:
The adoption of the socialist pattern of society as the national objective as well as
the need for planned and rapid development, require that all industries of basic
and strategic importance, or in the nature of public utility services should be in the
public sector. Other industries, which are essential and require investment on a
scale that only the State, in the present circumstances, could provide, have also to
be in the public sector. The State has, therefore, to assume direct responsibility for
future development of industries over wider areas.

The 2
nd
Five Year Plan document further reinforced the dominance of public
sector as it clearly stated “the public sector has to expand rapidly. It has to not only
initiate developments which the private sector is either unwilling or unable to undertake,

107
it has to play the dominant role in shaping the entire pattern of investment in the
economy”. Successive five year plans put increasingly high reliance on the public sector
in national development strategy. As shown in Table 4.1, the nationalization of
commercial banks in 1969 by the then Prime Minister Mrs. Indira Gandhi, added to the
dominance of the public sector in industrial production, public utilities and financial
sectors.
The public sector in India can broadly be classified into three categories (Goyal,
2001): (1) Departmentally owned and administered undertakings like the posts and
telegraph, railways, and irrigation projects; (2) Banks, financial institutions and
insurance companies; and (3) and units established under the Companies act, 1956 and
under special statutes passed by the parliament. In 2007, there were 247 public sector
units (PSUs) owned by the Central government while 986 units were owned by various
State governments.
Table 4.1: Change in Public Sector share in Industrial Production from 1968 to
2000

National Production Share of SOEs
1968 1990 1995 2000 1968 1990 1995 2000
Fuel
Coal 71400 201350 247450 309630 17.66% 96.23% 96.89% 97.05%
Lignite 3980 8420 12540 22950 100% 100% 100% 100%
Crude Oil 6060 19380 23500 32430 50.83% 85.38% 89.56% 87.39%
Refined Crude 16550 52600 71560 103440 48.88% 72.21% 78.50% 68.61%

Non-ferrous Metals
Aluminium 125.30 424.65 545.80 642.84 0% 45.35% 48.45% 49.34%
Primary Lead 1.90 34.45 41.70 45.50 100% 75.45% 74.72% 73.62%
Zinc 17.00 80.50 108.40 178.01 80.60% 84.35% 85.60% 83.28%

Source: Public Enterprises Survey, 2000-2001, Ministry of Heavey Industries and
Public Enterprises, Department of Public Enterprises, Government of India.

108
Though central government PSUs were only about 20% of all PSUs, they
accounted for more than 85% of the capital and assets of all PSUs. Though both public
and private sectors have co-existed in India, the share of the public sector in GDP has
grown consistently but reached a plateau of around 25% by 1990, as shown in Table 4.2.
Table 4.2: Public and private sector share in GDP

Year At current price in billion US$ % Share in GDP
Public Sector Private Sector Total Public Private
1970 5802 36420 42222 13.74 86.26
1980 25525 104651 130176 19.61 80.39
1990 128398 382556 510954 25.13 74.87
1991 153632 435454 589086 26.08 73.92
1992 175320 497901 673221 26.04 73.96
1993 202512 578833 781345 25.92 74.08
1994 234395 682663 917058 25.96 74.44
1995 270129 803142 1073271 25.17 74.83
1996 297443 946103 1243546 23.92 76.08
1997 352518 1037524 1390042 25.36 74.64
1998 405185 1210848 1616033 25.07 74.93

Source : National Income Statistics, Center for Monitoring Indian Economy (CMIE),
December, 2001

4.3 Performance of public sector enterprises: Table 4.3 gives the trends of
profitability of public sector enterprises from 1998 to 2005. It is evident that the return on
capital has continued to remain very low, though the performance of profit-making
companies has improved, the gains have been offset by deteriorating financial
performance of loss-making units. In the year 2005, 122 PSUs made profit while 112
units made losses. Within this group, the top 5 firms earned more than 53% of the total
profits earned by all profit making firms. The Oil and Natural Gas Corporation (ONGC)
alone accounted for 22.23% of total profits, while National Thermal Power Corporation
(NTPC), Videsh Sanchar Nigam Limited (VSNL), Indian Oil Corporation (IOC), and

109
Mahanagar Telephone Nigam Limited (MTNL) accounted for 12.24%, 8.45%, 6.12%
and 4.16% of total profits, respectively.
Table 4.3: Sector-wise performance of SOEs
Net profit in million$s
Sector Firms 1998 1999 2000 2001 2002 2003 2004 2005
Petroleum 12 74.2 86.1 95.7 117.3 220.4 228.9 240.2 278.5
Power 4 27.4 35.4 39.4 51.2 68.9 64.9 104.0 157.3
Telecommunication 2 21.0 26.2 19.3 33.2 42.8 23.2 71.0 154.4
TOTAL 18 122.6 147.7 154.4 201.7 332.1 317.0 415.2 590.2
Percentage of profit
of all profit making
SOEs 0.6 0.7 0.6 0.7 0.7 0.7 0.8 0.8

Source: Public Enterprise Survey, 1999, 2004, and 2007 Ministry of Industry,
Government of India

The next 5 firms in terms of total net profits accounted for more than 21% of all
profits earned by the profit making PSUs and these firms were Gas Authority of India
Limited (GAIL), Hindustan Petroleum & Chemicals Limited (HPCL), Nuclear Power
Corporation of India Limited (NPC), Bharat Petroleum & Chemicals Limited (BPCL)
and Power Grid Corporation of India Limited (PGCIL). It is evident that all the top 10
profit making PSUs were operating in regulated environments with little competition
from private enterprises. While ONGC, IOC, GAIL, HPCL and BPCL are in the oil
sector, VSNL and MTNL are in the telecommunication sector, and NTPC, NPC, and
PGCIL are operating in the power sector. Sector-wise analysis in Table 4.3 reveals that
SOEs operating in highly regulated sectors have been making the major contribution to
total profits and their share has been rising over the last five years.
Amongst the 112 loss-making SOEs in 2005, the lowest 10 enterprises accounted
for more than 64% of all losses. While the Hindustan Fertilizer Corporation (HFC)
accounted for one fourth of all losses, three firms of the coal sector, Bharat Coking Coal

110
Limited (BCCL), Eastern Coal Fields Limited (ECL) and Central Coal Field Limited
(CCL), together accounted for more than 34% of all losses. It is noteworthy that amongst
the 10 worst performing SOEs in terms of profits, three are from the coal sector and two
from the fertilizer sector (Hindustan Fertilizer Limited and Fertilizer Corporation of
India). Sector-wise performance of SOEs in India has not been uniform. Only a few
sectors have contributed a major share in profits. Table 4.3, shows that the state
enterprises in three sectors, namely, petroleum, power and telecommunication account
for more than two thirds of all the profits earned by all profit making enterprises. This
table also shows changes in the financial performance of public sector units’ since
1990s. Although deregulation and globalization have been occuring since 1991, increased
competition in the economy and the withdrawal of budgetary support adversely affected
the profitability of some of SOEs, the growth of profits has more than compensated
increased losses. This does not mean that the operating efficiency has improved as most
of the increase in profits has been contributed by firms operating in near monopoly
markets with price control and barriers to entry. This becomes clear as we look at the
performance of SOEs with respect to the labor productivity. As per National Council for
Applied Economic Research (NCAER) Report (2000), the per capita salary plus benefits
of SOE employees have risen at a compounded rate of 3.95% per annum in real terms in
last two decades, while the real value of sales has risen only at a rate of 3.51%. This
indicates that SOE labor force has appropriated disproportionately higher share in factor
payments

111
4.4. Privatization as public sector reform strategy: Policy makers and economists in
India have viewed privatization as one of the strategies for State-owned enterprise (SOE)
reforms, together with other strategies like hard budget constraints, institutionalizing the
relationship between public sector managers and administrative ministries through
Memorandum of Understanding
27
(MOU), and increasing competition in product and
factor markets. India’s policy in the area of privatization reflects the same gradualism that
characterizes reforms in other public policy areas including economic policy. This
gradualism has been criticized for not only diluting the flow of benefits, but also for
allowing time for the opposition to harden. The one-step at a time approach also
obfuscates policy intent and can lead to sub-optimal policy choices at each step. On the
other hand, the advantage of this approach is that it makes it easier to evolve a gradually
moving political convergence in what is otherwise a highly fragmented, pluralist and
cacophonic democracy. It appears that political parties have indeed converged on the
privatization issue with recent statements of Communist Party of India (CPI) and
Communist Party Marxists (CPM) that accept privatization as a desired economic goal
and emphasize their differences with other political parties only on methodology and
sequencing. Therefore, successive governments, led by different ends of the political
spectrum, having opposed privatization when out of the governing coalition nevertheless
ended up following the privatization path step by step.
India’s policy towards public sector reform can be put in three distinct phases:

27
This was similar to the system of “contracts” prevailing in France. Memorandum of Understanding
(MOU) was a kind of contact between the government and the SOE management which laid down
quantified performance goals for the enterprise.

112
I. From the mid 1980s to 1990, “reform” phase, it was recognized that public
sector performance was inadequate and extensive reform was needed. But
economic policy makers and political leaders including Rajeev Gandhi, the
then Prime Minister, felt that public sector efficiency could be improved
through institutional arrangements between the SOEs and the government
without privatization.
II. During the 1991 to 1997 “disinvestment” phase, the government sought to
raise revenue by selling minority holdings in central public sector
enterprises, but with management control remaining firmly in Government
hands.
III. The third phase, began in 1998 under Bhartiya Janata Party rule at the
federal level with Atal Bihari Vajpayee as Prime Minister, This phase can
be called the phase of “privatization”. In this phase government explicitly
stated that it would withdraw from managing public sector enterprises and
hand over management to private sector owners through strategic sales of
SOEs.
4.4.1 Early reform phase: Phase I of the reform strategy focused on improving the
efficiency of SOEs without changing the ownership structure. It was based on a two
pronged approach, (a) to create suitable managerial autonomy for viable SOEs, and (b) to
create an exit mechanism for chronically unviable ones. The government promised a
White Paper dealing with these approaches and efforts were made in the late 1980s to
prepare such a document. However, the issue of exit policy for unviable public sector
units proved politically too sensitive and the White Paper, though drafted, was never

113
approved by the Cabinet. Yet, the other leg of the reform strategy—giving managerial
autonomy to public sector units—did not present ideological problems and some steps
were taken in that direction. The Government of India approach was based on the
system of public sector management in France which involved entering into “contracts”
with public sector managers laying down the specific goals and also indicating the
specific nature of the support that would be provided by the government and, after
signing the contract, giving full managerial autonomy to the SOE managers. Indian
incarnation of the system of contracts took the form of a Memorandum of Understanding
(MOU) between the Government and the management which laid down quantified
performance goals for the enterprise. Although several such MOUs were finalized, the
exercise proved to be a complete failure, illustrating the difficulty in transplanting
systems from one institutional environment to another (Ahluwalia, 2002).
4.4.2 Disinvestment phase: The beginning of the second phase of public sector reform
can be traced to the balance of payments crisis of 1991. Since the crisis was widely seen
to have been precipitated by fiscal profligacy in the course of the second half of the
1980s, a reduction in the fiscal deficit was immediately seen as the needed response for
restoring macroeconomic stability. It was also a necessary condition for obtaining IMF
support, which was critical to managing the crisis. One of the measures which could help
to reduce the fiscal deficit was the sale of equity in public sector enterprises and was first
explored in early 1991 by the short-lived coalition government headed by the Prime
Minister Mr. Chandrasekhar. But his government fell even before it could present the
Budget. This was followed by a general election in 1991 which led to the formation of a
Congress led government, under the Prime Minister Mr. Narasimha Rao.

114
The new government moved quickly to address the crisis through a bold program
of economic stabilization which involved basic changes in many aspects of economic
policy, including industrial and trade policies. Since the reduction of the fiscal deficit was
a core objective of the policies, it included the sale of up to 20% of the equity of public
sector enterprises in order to mobilize resources for the budget. Forty seven profit making
public sector enterprises were included in the disinvestment program and their shares
were sold by auction in bundled lots, each bundle consisting of shares of different Public
Sector Enterprises (PSEs) were drawn from three different categories classified on the
basis of financial performance. Instead of selling the shares of each enterprise separately,
they were sold as a bundle by auction. The publicly stated reason for selling shares in
bundled lots was to avoid a situation whereby shares of the better performing PSEs might
get sold, while the others might be left unsold except at prices that may be perceived to
be too low. Notably, the auctions were limited to public financial institutions including
United Trust of India (UTI), Life Insurance Corporation (LIC) and mutual funds run by
public sector banks. This was done to avoid criticism that public sector equity was being
cornered by rich private individuals who might thus obtain undue power or influence
over the public sector units. Although the direct sale of shares to institutions for possible
resale to individuals implied that the new private shareholders would acquire voting
rights, the government went out of its way to emphasize that there would be no effective
reduction in government control since the “disinvestment” was limited to only 20% of the
equity. The government also emphasized that the emergence of non-government
shareholders with a minority holding would make public sector management more
sensitive to profitability performance as stock prices would reflect financial performance.

115
The government also reiterated its commitment to “strengthen” the public sector,
by reviving those public sector units that were potentially viable to make the policy of
disinvestment politically feasible. It announced the creation of a National Reconstruction
Fund to deal with the financial needs of restructuring viable public sector units, including
the provision of terminal/retirement payments to workers being laid off. While these
features did avoid criticism from ideological supporters of the public sector, there was
criticism of disinvestment on the grounds of the transparency of the procedure and of
whether or not fair value was realized for the equity being sold. Since the bundling of
shares was immediately criticized as nontransparent (indeed it was so), it was abandoned
after the first year in favor of auctions of shares of individual companies subject to
reserve prices fixed by government on the basis of advice of merchant bankers. The
government policy to limit the sales to financial institutions was also criticized as
inconsistent with competitive pricing and the government responded in 1994-95 by
opening up the process to all institutions and individuals qualified to hold equity shares in
India. In 1995, the government decided to expand the set of potential buyers by allowing
shares to be sold internationally through the mechanism of Global Depository Receipts
(GDRs)
28
.
In retrospect, it appears that the policy of disinvesting a minority share of equity
while retaining majority control was fundamentally flawed from the perspective of SOEs’
operating efficiency. The policy was a compromise between the goal of raising resources
to lower budget deficits and the objective of retaining control over the public sector. But

28
This had the advantage that the GDRs were tradable in international markets and capital gains arising
from these sales were not taxable in India. Holders of GDRs also could not exercise the voting rights
associated with the shares unless they converted the GDRs into actual holdings of underlying shares

116
these goals were not compatible. Private investors could not be attracted to hold a
minority stake in a public sector enterprise unless there was some clear signal that the
enterprise would be as well managed as other private enterprises in the field. In the
absence of a transfer of control rights to the private investors, this signal could only be
generated if there were credible reforms in public sector management. A new approach
was clearly needed and this came after the general election of 1996. United Front which
was a coalition of left of center parties came to power and against all conventional
wisdom, it was this government that took the privatization agenda forward in a strategic
direction, albeit in a gradual manner.
4.4.3 Privatization phase: The United Front government announced that it would
“carefully consider the case” for withdrawal of the government from the “non-core, non-
strategic areas” and would set up a Disinvestment Commission to advise it on these
matters
29
. This was a text book example of “one step at a time” gradualism as the
government did not announce that it would actually withdraw from non-core non-
strategic areas, nor did it clarify how extensive the category of “non-core non-strategic”
would be. The Disinvestment Commission was established in 1996. The Commission did
not propose any general principles for determining which type of SOEs should be
privatized, though its recommendations clearly endorsed privatization in a number of
public sector enterprises. The Commission recommended privatization of all SOEs which
were making losses or were marginally profitable and operating in competitive areas but
it appeared to accept that profitable PSUs which were doing well, even if operating in
non-strategic areas, should be retained in the public sector. As it transpired, the United

29
The announcement was made in the Budget Speech of the Finance Minister Shri P. Chidambaram.

117
Front government did not implement any of the recommendations for privatization as it
did not survive long enough to do so. Though in its two years existence, (1996 and 1997),
the United Front government implemented the first successful sale of shares to
international institutional investors (VSNL and MTNL through GDRs), and also did
make yet another effort to introduce greater functional autonomy for public sector units
by designating a handful of well performing companies as “Navratnas”
30
.
After the United Front government lost its majority in Parliament in 1998, general
elections for Parliament were held in early 1998 and brought in the National Democratic
Alliance (NDA) government led by the Bharatiya Janata Party (BJP). The new
government formally accepted privatization as a desirable objective and outlined a new
policy. The main elements of this policy can be summarized as follows:
a) Majority government holding will be retained only in SOEs that are deemed to be
strategic that is related to defense production, atomic energy and railway
transportation. This meant that all other SOEs would be privatized either through
gradual sale or “strategic sale”
b) Potentially viable PSUs would be restructured and revived and those that could
not be revived would be closed down
c) Receipts from disinvestments and privatization would be used for meeting
expenditure in social sectors, restructuring of PSUs, creating a safety net for
workers, and retiring public debt

30
These units, identified on the basis of demonstrated management efficiency and profitability, were to be
granted greater flexibility of operation and freedom from government control. “Navratna” literally means
nine jewels and was traditionally used to refer to the nine most luminary courtiers in the Court of the Mogul
Emperor Akbar in the 16th century.

118
In order to provide strategic policy direction, a separate Ministry of Disinvestment
was created, with a Cabinet Minister in charge. The Ministry was made responsible for
the actual conduct of the sale, including negotiations with merchant bankers and strategic
partners. The most important achievement of the new policy and associated institutional
framework was that a number of public sector units were successfully privatized.
o The first public sector company to be successfully privatized was Modern Foods
India Ltd., a bakery with 13 plants all over the country, employing more than
2000 persons. Seventy four percent of the equity was sold in 2001 to Hindustan
Lever, a subsidiary of the Anglo-Dutch multinational, Unilever.
o Bharat Aluminium Company Ltd. (BALCO), a fully integrated aluminum
producing company employing 7000 people was privatized in 2001 by selling
51% of the equity to Sterlite Industries. Subsequently, Hindustan Zinc, a zinc
producing company, was also sold to the same company.
o Lagan Jute Machinery Company Ltd., a small machinery manufacturer employing
400 workers, was privatized by selling 74% of the equity to a private strategic
partner.
o Paradeep Phosphates Ltd. (PPL), a small chronically loss-making fertilizer unit,
was sold to M/s. Zuari Industries.
o Computer Maintenance Corporation (CMC), a software development company,
was sold (51%) to Tata Consultancy Services, one of the leading private sector
companies in the field.
o VSNL was the monopoly international telecommunications company, in which
43% had earlier been disinvested. The monopoly was ended in 2001 and the unit

119
has been privatized by selling 25% out of the government’s 57% stake to the Tata
Group as a strategic partner with full management control.
o Nine hotels belonging to the India Tourism Development Corporation (ITDC) and
three to the Hotel Corporation of India were sold to private buyers, including
foreign investors in some cases.
o India Petrochemicals Ltd. (IPCL) was privatized by selling 24% of its equity to
Reliance Industries, one of India’s largest private sector companies on the basis of
a competitive bid in which Reliance outbid the Indian Oil Corporation, a public
sector company, and Nirma Industries, a well known dynamic private sector
company.
o Maruti Udyog, India’s largest automobile producer, a joint venture of the
Government of India and the Suzuki Motor Corporation of Japan, was privatized
by surrendering the government’s portion of a rights issue in favor of the private
sector partner at a negotiated price reflecting the premium on full management
control.
In General Elections in 2004, National Democratic Alliance was defeated and
Congress led United Progressive Alliance came to power with “out side”
31
support of Left
parties. This slowed down the process of privatization and as a symbolic gesture;
Government of India dissolved the Ministry of Disinvestment and created a Department
of Disinvestment in the Finance Ministry. As Table 4.4 will show, no new public sector
firm has been taken up for partial or full privatization and only 3 moderate privatization
transactions have taken place in already partially privatized PSUs in the last three years.

31
This implies that the Left parties in not have their Ministers, but will support Government in the
Parliament.

120
4.5. Economic performance of firms and privatization: Large privatization programs
undertaken by the governments across the globe in last two decades have been studied
widely. Although there are number of empirical studies on privatization, we are still far
from having achieved any definitive conclusions about the effect of privatization on firm
performance. Shirley and Walsh (2000) point out that from 1971 to 1999, 32 studies out
of 52 show the superior performance of private or privatized firms, 15 studies show no
significant relationship between ownership structure and performance, and five studies
conclude that the performance of publicly-owned firms are significantly superior to
private firms. Galal, et al (1994), La Porta and Lopez-de-Silanes (1999), Dewenter and
Malatesta (2000), Bhaskar, Gupta and Khan (2006) are some of the most persuasive
studies that have examined the relationship between firm level performance and
privatization in non-transition economies.
Galal, Jones, Tandon, and Vogelsang (1994) compare the actual post-privatization
performance of 12 large firms in Britain, Chile, Malaysia, and Mexico to the predicted
performance had these firms not been privatized. Using this counter-factual approach,
this paper reports net social welfare gains, which were, on average, 26 percent of pre-
privatization sales, in 11 of the 12 cases. Authors do not find any case in which workers
were worse off after privatization. La Porta and Lopez-de-Silanes examine almost the
entire population of Mexican privatizations and compare performance changes to
industry matched private firms. They find that privatized firms rapidly closed a large
performance gap with industry-matched private firms that were there prior to
privatization. This is one of the most positive studies of privatization and finds that
output increased by 54.3 percent, in spite of a reduced level of investment spending, and

121
Table 4.4: Privatization proportion and amount by year

Year
Number
of SOEs
privatized
Average
percentage
of
privatization
Total amount
of privatization
proceeds
(million$s)
Maximum
proportion of
privatization
Method of
privatization
1992 30 12.50% 690.5 0.2
Bundled sales to
financial
institutions
1993 23 4.10% 434.5 0.1
Auctioning with
minimum reserve
price
1995 15 5.85% 1100.6 0.02
Auctioning with
minimum reserve
price and NRIs
and OCBs
permitted to
participate
1996 6 1.00% 82.3 0.02
Auction, FII
permitted to
participate
1997 1 5% 86.3 **
Depository
Receipts
1998 1 7% 205.1 **
Global Depository
Receipts
1999 5 12% 1220.7 0.25
Domestic cum
GDR and Cross
holding sale in
which part of
government
holding were sold
to other SOEs
2000 5 19% 357.5 0.74
GDR, Domestic,
Cross holding and
Straegic sale
2001 3 67% 424.5 1
Cross holding and
Strategic sale
2002 6 58% 711.4 1
Strategic sale and
auction
2003 6 21% 742.1 0.49
IPO and Strategic
sale
2004 3 18% 658.1 26%
IPO and sale to
employees
2005 1 14% 348.7 21%
Sale to public
sector financial
institutions and
employees
2006 0 0% 0.0 0% _

122
sales per worker almost doubled. The privatized firms reduced employment by almost
half, but the average worker salary increased significantly. Authors attribute most of the
post-privatization performance gains to productivity gains from better incentives and
lower labor costs.
While many empirical studies tend to find positive effects of privatization on firm
performance, some indicate the serious selection bias problem. In some cases, soon-to-
be-privatized state-owned firms began introducing reforms and were already performing
better than those not scheduled to be privatized, even before privatization (e.g. Pinto,
Belka and Krajewski (1993)). Dewenter and Malatesta (2000) have studied the effects of
government ownership and privatization, using a sample of firms from three time periods,
namely, 1975, 1985, and 1995. After controlling for firm size, location, industry
category, and the business cycle, they find that net income-based performance measure
increased significantly, while the operating income-based measures did not. A recent
work reports that efficiency gains from privatization depend on regulatory environment
and privatized firms with monopolistic characteristics operating under minimalist
regulation improve profitability due to market power and not due to efficiency
improvements (Domney, Wilson and Chen, 2005).
Though, privatization has received wide spread attention in business and
economics literature, the phenomenon of partial privatization has not been studied much.
This is surprising, given the fact that most of the privatizations in developing, transition
and developed economies have been partial (Bortolotti and Fassio, 2004 and Table 2.3,
Chapter 2). Understanding partial privatization and its impact on firm level performance
is not only important from political economy and public policy view points; it is also of

123
theoretical value. As such, it may offer insight into the debate on the comparative
performance of private vs. state-owned firms. It has been argued by some that partial
privatization leads to increase in the yardstick competition in the regulation of public firm
because heterogeneous ownership undermines collusion and also makes excess
employment more transparent to the general public (Bhaskar, Gupta and Khan, 2006).
The political view argues that principal-agent problems are the source of inefficiency of
SOEs, as the manager may run the firm to achieve political objectives (such as higher
employment) rather than the economic objective of profit maximization. This view
argues that privatization transfers ownership and control to outside investors and thus,
after privatization, greater emphasis is placed on efficiency (Shleifer and Vishny, 1994).
But this view does not address the question as to why politicians, who are supposed to be
the beneficiaries of SOEs, themselves, decide to pursue policies detrimental to their self
interest. As pointed out in chapter 3, partial privatization may be a mechanism through
which politicians’ trade off political benefits with economic costs.
Further, outside the theoretical models of markets and competition, in the real
world, private ownership also comes with costs. In their seminal work, Berle and Means
(1932) point out that in the absence of suitable internal and external governance
mechanisms, diffuse ownership may expose minority shareholders to the risk of
expropriation by managers. While large shareholders may reduce the agency costs of
managerial control (Jensen and Mecking, 1976; Vishny, 1986), ownership concentration
may come with the risk of extraction of private benefits by controlling shareholders at the
expense of the firm efficiency or outside investors (Claessens et al., 2002; Johnson et al.,

124
2000; La Porta, Lopez-de-Silanes, and Shleifer, 2002). Thus, the benefits of government
control may outweigh the costs of political interference.
India is a good place to look for empirical evidence to address these research
questions about the impact of partial privatization. As mentioned earlier, India’s
privatization program was dominated by the sale of minority shares of SOEs from 1991
to 1997, the strategic sale of pubic sector firms starting only in 1999. Some SOEs have
been given partial privatization treatments multiple times during our sample period.
Hence, we have a robust panel data for 13 years on most of the federal government SOEs,
including privatized, partially privatized and not at all privatized firms. The privatization
program was also accompanied by significant changes in regulatory and competitive
environment at around the same time. Thus, by controlling for the effect of regulatory
environment, and its interaction with privatization, we may get an insight in the source of
efficiency gain, if any, in partially privatized SOEs.
4.6 Data: This research uses data from the official records of the Comptroller and
Auditor General, Government of India. The accounting and financial firm level data for
the State-Owned-Enterprises (SOEs) is taken from the annual report on the accounts of
the central government companies and corporations, titled ‘Review of Accounts’, issued
by the Comptroller and Auditor General (CAG) . As the provincial government owned
enterprises are set up under state government statutes and may face different institutional
and political setting, in order to be consistent, we include only 254 Central Government
enterprises in our data set. The CAG is the statutory authority for certifying the accounts
of SOEs in the form of balance sheets and profit and loss statements, and thus this
appears to be the most reliable source of firm level data.

125
We observe the privatization proportion, industry type, and the range of firm level
financial data for 238 public sector firms over a period from 1991 to 2004. These firms
belong to the agriculture, manufacturing, financial and service sectors and are fully or
partially owned by the central government. Unlike Gupta (2005), we do not use data from
firms owned by the state governments for two reasons; first, the regulatory conditions for
state government firms may vary from state to state, and second, the data for many years
is missing for many state firms. We compile data regarding the privatization proceeds,
and other aspects from the official documents of the then Ministry of Disinvestment,
Government of India (Department of Disinvestment, Ministry of Finance, and
Government of India since June 2004). We also use the Public Enterprise Surveys, a
yearly publication of Department of Public Enterprises, Ministry of Industry, and
Government of India. These surveys are mandated by the Parliament and contain a macro
appraisal of performance of SOEs, a sectoral analysis of SOEs performance and firm
specific financial data for each of the central government-owned firms. The data on
reform policies, price indices, regulatory conditions, listing in the stock exchanges and
competitive environment are compiled from yearly Economic Surveys from the years
1990 to 2004. These surveys are published every year by the Ministry of Finance,
Government of India, a week before the presentation of the budget in the Lok Sabha
(lower house of parliament). We also referred to the documents available on the official
web site of Reserve Bank of India to get historical information on public and private
sector investment. Further, we use Election Commission of India official publications to
compile data regarding the party composition of Governments in power at the Federal as
well as State levels.

126
4.7 Empirical approach: This research work parallels that of Gupta (2005), wherein
the effect of partial privatization on firm performance has been studied using data from
India for the period 1990 to 2000. Their firm level data includes 247 non-financial
companies owned by the central government, and 92 firms owned by various state
governments. This study does not use the accounting data after the year 2000 as in that
year the Government of India switched to strategic sales of some public sector companies
together with partial privatization in some others. She finds, contrary to widespread
belief (that, since partial privatization does not transfer the control of the firm to the
private owner it has little impact on performance), that partial privatization has positive
and highly significant impacts on firm sales, profits, and labor productivity.
Though it is a very competently executed study, it fails to adequately take into
account firm-specific controls and the listing of the partially privatized firms on the stock
exchange (it assumes that all partially privatized firms are listed and traded on the stock
exchange). Further, Gupta (2005) takes government borrowing to total borrowing as an
exogenous variable, but this may be an endogenous variable in as much as the
government borrowings to total borrowings ratio is likely to be determined by the
privatization proportion itself. The role of regulatory conditions has also not been
addressed adequately as her study captures such changes in competitive environment
only in the reform year (1991) and not subsequently. This study finds “that partial
privatization, in which minority shares of state-owned firms become available on stock
markets, has a positive and highly statistically significant impact on the operating
performance of firms”. It’s not clear if the improvement is driven by listing on the
exchange or due to proportion of partial privatization Further, if the gains in the firm

127
were attributed to the managerial view, why should it depend on the proportion of partial
privatization? Thus it calls for further research to study the role of regulatory conditions
and also distilling the impact of listing in the stock exchange on the partially privatized
firm performance.
Apart from the obvious advantage of our 13 years panel data as compared to 10
years panel data in her study, our dataset consists of both ‘disinvestment phase’ (1991-
1997) and ‘privatization phase’ (1998-2003) data points. To control for the selection bias
in the firms that were partially privatized, we control for ‘historical performance’ of the
SOE by controlling for net worth to capital ratio (NWCAPR). Further, we control for the
dynamic regulatory conditions and the type of industry to take out industry specific
dynamic effects. In order to study the source of performance gains in privatized firms we
look at the interaction terms.
In this research our empirical strategy is to study whether the operational
performance of partially privatized firms depends on the proportion of equity sold when
we control for other factors that may also affect performance and may be changing at the
same time. We have strong reasons to believe that the privatization in India has not
followed any definite strategic path and has been generally incoherent (The Economic
Times, 2001). The disinvestment policy of Government of India has not followed any
strategic direction and has been evolving over last two decades as mentioned in the White
Paper on Disinvestment of Central Public Sector Enterprises, 31
st
July, 2007 “The policy
on disinvestment has evolved largely through statements of Finance Ministers in their
budget speeches.” So we start our empirical investigation taking privatization proportion
as an exogenous variable. But we know from our analysis of determinants of partial

128
privatization with cross country data set and also from literature (Gupta 2005) there may
be potential endogeneity issues in estimation and we use several approaches to address
the same.
First, in order to reduce the possibility of simultaneity between privatization
proportion and operational performance we examine the impact of the previous year
share of private ownership on current year’s firm performance. This also permits us to
account for the managerial actions that may affect performance due to partial
privatization as they are likely to appear with a lag. Second, in order to address the
potential issue of selection bias, (for example, more profitable or firms with large capital
base are selected for privatization) we estimate a firm fixed effects specification.
Third, our empirical model specifications include firm specific controls and year
dummies to control for contemporaneous macroeconomic shocks. Fourth, later we relax
the assumption of strict erogeneity in the fixed effects model and estimate the 2 stage
least square model using instrumental variables. The privatization proportion may itself
be correlated with the regulatory conditions or firm’s profitability. In order to take care of
this problem we need to use valid instruments that are correlated with privatization
proportion but are not correlated with firm level performance based in part on our
empirical work in Chapter 2. Fifth, we also use an alternative strategy to 2SLS approach
with LAGPRIVAT as our key regressor. This is weakly exogenous rather than strictly
exogenous. We take the time difference to eliminate fixed effects and then use further
lagged variables as Instrumental Variables. The justification of choice of IVs and formal
tests are described in subsequent analysis.

129
Thus, first we want to study the impact of privatization proportion on firm
specific performance. Our empirical model has the following firm fixed effects
specification:

1 it i t it it it
Y PRIVTPRO X α α β γ μ
−
= + + + + 4.1
where
it
Y is the performance measure,
i
α is the firm fixed effect,
t
α is a year effect
captured by dummy variables for each year, and
it
X is a vector of firm-specific as well as
political economy factors that affect firm performance, the is dependent variable Y . The
firm specific fixed effect ( )
i
α captures fixed differences across firms that are constant
but are unobserved over time, while the year dummies( )
t
α reflect contemporaneous
correlation. While the variable
1 it
PRIVTPRO
−
32
reflects proportion of privatized equity in
the year 1 t− for the firm i ,
it
μ is the unobserved shock affecting the performance of
firms. In our analysis throughout this paper for ease of exposition we denote
1 it
PRIVTPRO
−
by LAGPRIVAT .
In this research we use four measures of firm level performance: return to capital
( ) RETCAP which is profit before tax per unit of capital employed; profit before tax per
unit of sale( ) PROFIT ; return to labor ( ) RETLAB which is profit before tax (Rupees in
Crores
33
) per employee; and labor productivity ( ) LABPROD which is sales (Rupees in
Crores) per employee. In our analysis RETCAP and ( ) PROFIT capture firm
performance in terms of overall operational efficiency, while RETLAB and

32
We study the impact of lagged privatization proportion on current performance to avoid potential
simultaneity problems.
33
Crore is a unit used in Indian system. 1 Crore = 10million.

130
LABPROD are measure of productivity. While both these measures of labor productivity
have been used by Gupta (2005), RETLAB has been used by LaPorta and Lopez-di-
Silanes (1999) among others.
As pointed out in the chapter 2, regulatory institutions and political economy
factors may play an important role in performance of public sector firms. Thus, in order
to distil the impact of partial privatization proportion on performance indicators we need
to control for these influences. We construct three variables to capture the impact of these
factors: REG , PARTY and FRACTION . REG is a dummy variable that takes value 1 if
the industry had been reserved for the public sector (either for monopoly under schedule
A or for dominance under the schedule B of the Industrial Policy Resolution, 1956) and
operates in a fully regulated environment, and takes value 0 if the industry has been
dereserved and operates in regulation- free environment. PARTY is again a dummy
variable and takes value 1 if the Federal Government is headed by centrist party (either
Congress or Janata Dal) and takes value 0 if its headed by rightist party (Bhartiya Janata
Dal). Variable FRACTION captures fractionalization of the national polity. This
variable is defined as the percentage of States that have the same party or coalition
Government as at the Federal level. For example in the Year 2002, there were 12 States
that had National Democratic Alliance (NDA) Governments and the Federal Government
was also headed by NDA , and as India has 28 States, FRACTION takes value 42.86 for
the year 2002. Other than these variables, firm specific controls in vector X
are CORE , INTER , NWCAPR and LIST . Type of industry sector in which the public
sector unit is operating, is captured by using indicator variable CORE for infrastructure
sector and INTER for Intermediate sector, where 0 CORE = and 0 INTER= indicates

131
that the firm operates in competitive sector. NWCAPR is ratio of net worth to capital
employed of a firm and captures historic performance of the firm. LIST is again a
dummy variable and captures whether a partially privatized public firm has been listed in
Bombay Stock Exchange (BSE) or/and in National Stock Exchange (NSE). The variable
LIST=1 if the firm is listed in either of the exchange in that year, and =0, otherwise.
In order to get further insight into the impact of partial privatization on firm level
performance, we include interaction between
1 it
PRIVTPRO
−
and core sector CORE and
regulatory condition REG in our empirical model given by:
1 1 1 2 1
* *
it i t it it it it it
Y PRIVTPRO X PRIVTPRO REG PRIVTPRO CORE α α β γ δ δ μ
− − −
= + + + + + +
4.2
There may be threshold levels of privatization that may be required for any
significant impact of privatization on firm performance and in order to study the impact
of some threshold level of privatization on firm performance we construct dummy
variables: HPVIT and first we estimate the fixed effects empirical model without
interaction terms given as
1 2 it i t it it
Y HPVIT MPVIT X α α β β γ μ = + + + + + 4.3
The model with interaction terms is given as
1 2 1 2 it i t it it
Y HPVIT MPVIT REGHPVIT REGMPVIT X α α β β δ δ γ μ = + + + + + + + 4.4
In order to study the impact of partial privatization on employment in the firm we
estimate following empirical models where dependent variable is log (total number of
employees), without and with interaction terms
1 it i t it it it
LEMP PRIVTPRO X α α β γ μ
−
= + + + + 4.5
The model with interaction terms is given as

132
1 2 1 2 it i t it it
LEMP HPVIT MPVIT INTERCORE INTERREG X α α β β δ δ γ μ = + + + + + + + 4.6
As pointed out earlier and also brought out by the analysis in the chapter 2,
privatization proportion may not be strictly exogenous in our model specifications in
equations 4.1 to 4.6. Quality of institutions which was captured by the variable quality of
regulation (REGQUA) and control of corruption (CONCOR) , had emerged as most
consistent and significant determinant of proportion of partial privatization, other than the
type of industry ( CORE and INTER ). Further, as the privatization policy in India has
been influenced by political parties in power ( Goyal 2001), PARTY may also be
impacting privatization proportion. In order to take care of this potential endogeniety, we
use 2stage least square model, where we instrument privatization proportion in the first
stage and use , , , CORE INTER PARTY REGQUA and CONCOR as instruments. The
empirical model that we estimate at the first stage is given as
1 2 3 4 5 it i t i i t t t it
LAGPRIVAT CORE INTER PARTY REGQUA CONCOR α α β β β β β μ = + + + + + + + 4.7
Further as an alternative to 2SLS, as our main regressor (privatization proportion)
is not strictly exogenous, we take the time difference of our performance variables to
eliminate fixed effects and estimate the following empirical model:
1 1
y t y X it
it t it it it it
PRIVAT α β γ δ ε
− −
Δ = + Δ + Δ + Δ + Δ 4.8
4.8 Results and discussion: We start our empirical investigation by first narrowing our
data set to be comparable to Gupta, 2005 data set by years and sectors and use same
performance indicators, independent variables and instruments
34
. Our results are
presented in Table 4.6a. Not surprisingly, we also find that lagged privatization

34
There may be some variation in Gupta (2005) and our data as she has some firms owned by the state
governments also in the dataset, while our data set has only those firms that are owned by the Federal
Government.

133
proportion has significant and positive impact on firm performance. But as indicated in
our empirical strategy, we want to investigate whether or not regulatory conditions and
listing on the stock exchange have impact on the performance of partially privatized
firms. This calls for further investigation by estimating the models given by equations
4.1 to 4.8.
Table 4.5a: Impact of partial privatization on SOE performance

RETCAP PROFIT RETLAB LABPROD

LAGPRIVAT 15.483 6.916 1.348 1.319

CORE 223.521 -214.936 65.841 169.892

INTER 1134.212 524.648 53.649 46.937

NWCAPR 671.112** 361.239* 184.262** 110.771**

LIST 63.461** 14.748** 30.778* 1.319*

REG 340.173** 89.859** 56.393** 14.172**

PARTY 23.057 -5.779 -24.557 6.003

FRACTION -37.342 -13.706 -7.004 2.744

Firm fixed effects Yes Yes Yes Yes

Year dummies Yes Yes Yes Yes

Total obsevations 2368 2302 2356 2356

* significant at 5% level, ** significant at 1% level

Table 4.5a presents the results from estimating equation 4.1. Surprisingly, we find
that the proportion of privatization in the previous year ( ) LAGPRIVAT has no
statistically significant impact on any of the performance measures of overall operational
efficiency or labor productivity. The impact of privatization proportion is indeed

134
negative, though not statistically significant, for the performance measures, profit per unit
of capital employed ( ) RETCAP and sales per employee( ) LABPROD . At the outset, our
results are at variance with Gupta (2005) who finds that proportion of equity held by
private investors has a positive and significant impact on all measures of firm
performance even when regulatory conditions in which firms operate is not controlled.
There are three apparent differences between Gupta (2005) and our study: first and
probably foremost, we control for regulatory conditions together with political economy
variables as well as listing of the firm on the stock exchange, second, our dataset
includes manufacturing as well as other SOEs information over a period of 12 years
(1991 to 2002) as against Gupta’s dataset that had information only for manufacturing
firms over a period of 8 years (1991 to 1998), and third, the definition of SALES and
PROFIT in this study are different from theirs.
35

Table 4.5b presents the results from estimating model with interaction terms
given in equation 4.2. We find that the both interaction terms are positive though are not
significant. This implies that a firm that higher proportion of privatization for any firm
operating in highly regulated environment leads to improvement in performance.
Similarly, there is some positive synergy between partial privatization and core sector.
But these positive impacts are not significant and so we can not draw any robust
conclusion.

35
This study has taken Gross Sales and Gross Profit Before Tax as reported by the Comptroller and Auditor
General. Gupta (2005) define SALES as ‘Annual sales generated by the enterprise from its main business
activity measured by charges to customers for goods supplied and services rendered. Excludes income from
activities not related to main business, such as dividends, interest, and rents in the case of industrial firms,
as well as non-recurring income’, and PROFIT as ‘Annual excess of income over all expenditure except
tax, depreciation, interest payments, rent, and extra ordinary expenditures’.e

135
Table 4.5b: Impact of partial privatization on SOE performance with interaction

RETCAP PROFIT RETLAB LABPROD

LAGPRIVAT -13.254 6.759 4.248 -1.672

CORE 233.326 -201.006 55.427 163.331

INTER 1032.016 559.681 30.663 46.183

NWCAPR 60.102** 301.945** 138.062* 97.445**

LIST 60.701** 4.866** 27.403** 0.903**

REG 237.631** 80.639** 49.493** 23.002**

PARTY -30.509 -11.444 2.318 7.224

FRACTION 18.655 1.77 -16.659 16.593

INTERREG 33.428 11.721 21.557 28.396

INTERCORE 41.768 36.882 23.374 56.667

Firm fixed effects Yes Yes Yes Yes

Year dummies Yes Yes Yes Yes

Total obsevations 2368 2302 2356 2356

* significant at 5% level, ** significant at 1% level

Two key results emerge. First, that efficiency as well as productivity gains in
partial privatized firms are explained by the regulatory environment and whether the firm
has been listed on the stock exchange or not. As expected, listing of the firm leads to
monitoring by the market and its impact is positive and significant for all four firm level
performance measures. In our data set, out of the total 245 SOEs, 58 firms have been
partially privatized and out of these 29 firms are listed in one or both stock exchanges.
All the public sector banks, power and oil companies that have been privatized are listed,

136
while small firms and manufacturing sector firms are rarely listed. This is an interesting
insight as partially privatized firms continue to have Ministerial control over the
management as majority shareholding remains with the Government. Thus, we can say
that the efficiency gains in partially privatized firms come from stock market monitoring,
and not from partial privatization. Further, the regulatory policy framework in which the
public sector firms operate is an important determinant of firm profitability and
productivity as REG is significant in all four models. As expected REG has a positive
impact on firm level performance because firms operating in regulated environment are
able to thwart competition and are also able to influence policy in such a way to benefit
them. This is contrary to general belief that public sector firms suffer as many of them
operate in highly regulated environment.
The second interesting result is that when we control for regulatory conditions,
then neither the sector of operation of the public firms nor political economy variables
PARTY and FRACTION have any significant impact on firm performance. This
confirms our qualitative analysis in section 4.2, that there has been little difference in the
economic policies or approach towards public sector firms amongst centrist parties led by
Indian National Congress (INC) or Janata Dal (JD) and rightist parties led by Bhartiya
Janata Party (BJP).
The results from Tables 4.5a and 4.5b bring out a critical, but not much discussed,
aspect of regulatory environment on relationship between privatization and performance
of firms in India. Under the socialistic pattern of planned economy, the economic
regulation was undertaken by the government directly and was effected primarily through
licensing and price controls. While licensing created barriers to entry, price controls did

137
not allow market to determine demand and supply. For example, public sector firms
operating in regulated oil market have gone in red in last six months after showing high
profits in the year 2006-2007, not because of any deterioration in their operational
efficiency, but due to the fact that the government has refused to increase oil prices
though the crude oil prices have almost doubled. To that extent we find that the impact of
partial privatization is not statistically significant on operational efficiency. This is at
variance with Gupta (2005) that finds that partial privatization has a positive impact on
firm performance across all regulatory environment and industry types.
Table 4.6a: Impact of level of privatization on SOE performance

RETCAP PROFIT RETLAB LABPROD

HPVIT -9.727 12.871 1.963 -0.834

MPVIT -6.832 9.675 2.208 -1.292

CORE 180.994 132.855 64.643 48.9972

INTER 683.055 455.638 84.977 34.346

NWCAPR 477.73** 326.563* 148.547* 77.609*

LIST 49.641* 39.683** 48.884* 3.558*

REG 179.302** 132.671** 45.593** 14.509*

PARTY -14.658 -8.446 6.347 1.408

FRACTION 12.683 6.759 11.503 1.443

Firm fixed effects Yes Yes Yes Yes

Year dummies Yes Yes Yes Yes

Total obsevations 2368 2302 2356 2356

* significant at 5% level, ** significant at 1% level

138
One of the possible explanations for the positive relationship between stock
market listing of partially privatized firm and its performance, though the managerial
control of the firm remains with the government, is that manager’s incentive to work hard
increase as good results of the firm and thus higher share prices, signal manager’s
competence which they can leverage to seek higher compensation when they move to
private sector (Gupta 2005). This does not appear to be a likely situation in regulated
sectors as are few very large private sector companies in these sectors like Power and
Steel. Therefore, it is more likely that partial privatization may have positive impact on
those firm that are in regulated environment from the point of view of possibility of
manager’s seeking more lucrative opportunities in the private sector.
It may be that there are some threshold levels of privatization that impact firm
level performance. MPVIT
.
We have experimented with four different levels for
HPVIT (high proportion of privatization), and MPVIT (Medium proportion of
privatization) in the estimating model given by equations 4.3 and 4.4. These levels are:
a) HPVIT =1, if proportion of privatization is greater than 50%, 0 otherwise,
MPVIT =1, if proportion of privatization is less than or equal to 50% but greater
than 25%, 0 otherwise.
b) HPVIT =1, if proportion of privatization is greater than 60%, 0 otherwise,
MPVIT =1, if proportion of privatization is less than or equal to 60% but greater
than 30%, 0 otherwise.
c) HPVIT =1, if proportion of privatization is greater than 50% , 0 otherwise,
MPVIT =1, if proportion of privatization is less than or equal to 50% but greater
than 25%, 0 otherwise.

139
d) HPVIT =1, if proportion of privatization is greater than 50%, 0 otherwise,
MPVIT =1, if proportion of privatization is less than or equal to 50% but greater
than 20%, 0 otherwise.
We find that our results are robust to these specifications and since specifications given in
(d) appear most reasonable cutoffs, we use this throughout our analysis.
Table 4.6a presents the results for impact of privatization on firm performance
measures with these different threshold levels. Again we find that there is no statistically
significant impact of level of privatization on firm performance. Table 4.6b presents the
estimated fixed effects model with interaction terms. Again, we find that the effect of
privatization level is not statistically significant. Further, we find that estimates given in
Table 4.8a and Table 4.8b are in harmony with the results when the privatization
proportion is taken as a continuous variable. Thus we can draw the inference that in
partial privatization, there are no threshold levels that lead to efficiency gains in public
firms.
Impact of partial privatization on number of employees is presented in Table 4.7.
We find that partial privatization has no statistically significant impact on the number of
employees. These results are in conformity with Gupta (2005) in that they have also
found no significant impact of partial privatization on employee size. Though prior to
1998 ( Gupta’s study period) , no strategic sales were made, while in this study’s period
number of strategic sales have been made by the government including sale of some loss
making units. But it appears that it has not impacted number of employees on an average.
In many of these firms, new management has infused more capital and operations have
been expanded. For example, Computer Maintenance Corporation (CMC) had around

140
Table 4.6b: Impact of level of privatization on SOE performance with interaction

RETCAP PROFIT RETLAB LABPROD
HPVIT -10.238 13.282 2.347 -0.981

MPVIT -7.453 10.926 2.019 -1.174

CORE 210.772 161.902 32.394 58.006

INTER 787.317 319.565 71.438 29.662

NWCAPR 510.739** 360.682* 135.667* 159.073*

LIST 60.238** 28.377* 31.408** 1.985**

REG 218.372** 92.783** 58.083* 17.236**

PARTY -18.913 -14.555 1.938 5.207

FRACTION 18.603 3.216 -13.226 9.398

REGHPVIT 3.475 6.324 3.567 1.216

REGMPVIT 5.791 12.668 5.504 0.572

Firm fixed effects Yes Yes Yes Yes

Year dummies Yes Yes Yes Yes

Total obsevations 2368 2302 2356 2356

* significant at 5% level, ** significant at 1% level

3200 employees when it was sold to a strategic partner in 2001. Within the first year of
its operation under private management, more than 600 new employees were hired as
operations expanded
36
. But it appears that such cases are far too few and in many partial
privatized firms the number of employees has declined.
In order to address the potential problem of selection bias in the firms that were
selected for privatization or the endogenity of privatization proportion, we use an

36
http://www.divest.nic.in/pvtpsu/cmc.htm

141
instrumental variables approach. As discussed earlier we use industry type ( CORE
and INTER ), quality of regulation ( REGQUA ) and control of corruption (CORCON ) as
instruments in the first stage in our 2SLS estimation. Thus we estimate the model as
given in equation 4.7, stage 1 and use the estimated privatization proportion as regressor
in model in 4.1. The results for performance indicators RETCAP and PROFIT are
given in Table 4.8a while those for RETLAB and LABPROD are given in Table 4.8b.

Table 4.7: Impact of privatization on employment

LEMP
LAGPRIVAT 0.0026

CORE 1.3981**

INTER 0.7183

NWCAPR 1.0113

LIST 2.1397

REG 0.3889

PARTY 0.0146

FRACTION 0.7365

Firm fixed effects Yes

Year dummies Yes

Total obsevations 2356

* significant at 5% level, ** significant at 1% level

The Sargen test statistics are insignificant, indicating that our model is not over
identified. Again, we find that the privatization proportion has no significant impact on

142
Table 4.8a: 2SLS estimation of partial privatization impact on frim
performance

RETCAP PROFIT

1st stage
Privat.
proportion
2nd stage
Return on
capital
1st stage
Privat.
proportion
2nd stage
Profitability
LAGPRIVAT 6.352 4.778

CORE 1.472** 112.541* 1.372** 89.547**

INTER 0.895* 863.428 0.875* 450.747

NWCAPR 671.342 657.658

LIST 56.783** 43.794**

REG 12.758** 8.579**

PARTY 0.672 5.766 0.605 4.635

FRACTION 16.734 12.668

REGQUA 3.462** 3.276**

CONCOR 2.237 1.994

Firm fixed effects Yes Yes Yes Yes

Year dummies Yes Yes Yes Yes

Total obsevations 2368 2368 2302 2302

Rsq 0.163 0.160

Sargan Test 0.226 0.187

The null hypothesis H
0
for Sargan Test is that instrumental variables are
uncorrelated to the set of residuals, and therefore the instruments are valid
and acceptable. The test statistic has χ
2
distribution and values given in the
Table are pvalues.

* significant at 5% level, ** significant at 1% level

143
Table 8b: 2SLS estimation of partial privatization impact on firm performance

RETLAB LABPROD

1st stage
Privat.
proportion
2nd stage
Return on labor
1st stage
Privat.
proportion
2nd stage labor
productivity
LAGPRIVAT 3.769 0.894

CORE 1.489** 34.675* 1.489** 11.768**

INTER 0.898* 65.752 0.898* 21.542

NWCAPR 53.338 17.851

LIST 19.793** 2.886**

REG 8.953* 0.775*

PARTY 0.682 13.657 0.682 7.852

FRACTION 32.337 6.338

REGQUA 3.352** 3.352**

CONCOR 2.207 2.207

Firm fixed effects Yes Yes Yes Yes

Year dummies Yes Yes Yes Yes

Total obsevations 2356 2356 2356 2356

Rsq 0.168 0.168

Sargan Test 0.120 0.328

The null hypothesis H
0
for Sargan Test is that instrumental variables are uncorrelated
to the set of residuals, and therefore the instruments are valid and acceptable. The
test statistic has χ
2
distribution and values given in the Table are pvalues.

* significant at 5% level, ** significant at 1% level

144
any of the firm level performance indicators. Though the instruments used in our analysis
are good as the first stage Rsquared is 0.16, the results are in harmony with our earlier
analysis. This indicates that it’s very unlikely that privatization proportion is endogenous.
When we use the time differencing to eliminate the fixed effects and estimate the model
given by equation 4.8, the results are given in Table 4.9. Here again we find that lagged
privatization proportion difference has no significant impact on any of the firm level
performance measures.
Table 4.9: Impact of partial privatization on firm performance

∆RETCAP ∆PROFIT ∆RETLAB ∆LABPROD

∆LAGPRIVAT 3.386 2.776 1.604 0.963

∆LAGRETCAP 0.886**

∆LAGPROFIT 1.148**

∆LAGRETLAB 0.563**

∆LAGLABPROD .803**

∆NWCAPR 87.854 68.769 76.337 3.865

∆FRACTION 23.758 17.763 9.864 1.436

Year dummies Yes Yes Yes Yes

Total obsevations 2128 2082 2116 2166

Rsq 0.446 0.389 0.372 0.216

Sargan Test 0.273 0.183 0.332 0.148

The null hypothesis H
0
for Sargan Test is that instrumental variables are
uncorrelated to the set of residuals, and therefore the instruments are valid
and acceptable. The test statistic has χ
2
distribution and values given in the
Table are pvalues.

145
Hence, our results suggest that the partial privatization proportion does not appear
to be explaining firm performance. These results are at variance from Gupta (2005) and a
possible explanation lies in lack of controlling for regulatory conditions in her work. In
order to get further insight and also to contrast our results, we narrow our data set to be
comparable to her data set by years and sectors and use same performance indicators,
independent variables and instruments
37
. In this case, we find that the lagged privatization
proportion has a significant and positive impact on firm performance (Table 4.10). But,
when we control for regulatory conditions and listing on the stock exchange as in Table
4.6a, the lagged privatization proportion no longer remains significant. This further
strengthens our argument that it’s the regulatory conditions and listing on the stock
exchange that impact the partially privatized firm’s performance and not the proportion
of equity held by the private sector.
Even from the qualitative analysis, it is less likely that partial privatization plan in
India suffers from any dynamic selection bias as there does not appear to be any long
term coherent privatization plan or strategy. This sentiment is also reflected in the
comments of Montek Singh Ahluwalia, one of the architects of economic reforms of
1990s and currently an advisor in the Planning Commission in an article presented at
Center for the Advanced Study of India, University of Pennsylvania, “However, there
was also a different kind of gradualism in policy evolved in a "one step at a time process"
in which the full extent of change envisaged was not clearly spelt out initially. Indeed
even the intention to take further steps was at times denied. At times, it was also a

37
There may be some variation in Gupta (2005) and our data as she has some firms owned by the state
governments also in the dataset, while our data set has only those firms that are owned by the Federal
Government.

146
reflection of political expediency, i.e. policy makers may have intended to go further in
due course, but it was felt politically inconvenient to spell out all the stages at the
beginning. At times, this reflected genuine uncertainty about the extent of change
envisaged. Indian leaders have often talked of the "middle path" implying a willingness
to go thus far and no further”. (Ahluwalia, 2002).
Table 4.10: Impact of partial privatization on firm performance without
controlling for regulatory conditions with same years and sectors as
in Gupta (2006)

∆Sales
t
∆Profit
t
∆RETLAB
t
∆Labor
t

∆LAGPRIVAT 0.103* 0.052** 0.006** 0.011

∆Y
i,t-1
0.546** 0.067 0.746* 0.306

∆ GOVT LOAN/TOT
BORR
i,t-1

0.814 0.328 0.011 0.009

∆ASSETS
i,t-1
0.157 -0.078 -0.234 0.008

Year Dummies Yes Yes Yes Yes

Number of
observations
1388 1388 1344 1344

Sargan Test 0.358 0.556 0.274 0.305

The null hypothesis H
0
for Sargan Test is that instrumental variables are
uncorrelated to the set of residuals, and therefore the instruments are valid
and acceptable. The test statistic has χ
2
distribution and values given in the
Table are pvalues.

* significant at 5% level, ** significant at 1% level
4.9 Summary and Conclusion: First we trace the evolution of public sector in post-
independent India and point out that in the first decade since independence there was
political consensus on industrialization policy that created an eminent position for the
public sector. Further, we point out that the licensing regime provided tacit support for

147
prominent industrial conglomerates as it was deftly used to create entry barriers and
isolate domestic firms from international competition at the cost of general consumers.
Further, we study the impact of the privatization proportion on firm
performance variables. We document two important and new findings. First we show that
the privatization proportion has no significant impact on firm performance when we
control for regulatory conditions and listing on stock exchange. Second, we find that
partial privatization has no negative impact on employment for firms operating in
regulated as well as unregulated sectors. The first result is at variance with Gupta (2005)
and would seem to be due to the failure of Gupta to control for regulatory conditions and
listing on stock exchange. This conclusion derives from the fact that when we narrow our
data set to be comparable to her data set by years and sectors and use same performance
indicators, independent variables and instruments we replicate her result in that the
lagged privatization proportion has significant and positive impacts on firm performance.
But as we see from the results in the Table 4.8a, when we control for regulatory
conditions and listing on the stock exchange, lagged privatization proportion no longer
remains significant. This further strengthens our argument that it’s the regulatory
conditions and listing on the stock exchange that allow the partially privatized firm’s
performance to improve and not the proportion of equity held by private sector by itself.
However, our data and analysis cannot identify the particular source that leads to
improvement of performance of partially privatized firms operating in the regulated
environment, As the partially privatized firms remain under government control, more
often than not, we can safely conclude that source of this improvement cannot be a
reduction in political interference but it could be due to monitoring by the stock market.

148
Bibliography

Abu-Shair, Osama. 1997. Privatization and Development. New York: St. Martins Press.

Adam, Smith. 1776. An Inquiry into Nature and Causes of the Wealth of Nations.
London: Methusen and Co., Ltd.

Ademola, Ariyo, and Jerome Afeikhena. 1999. "Privatization in Africa: An Appraisal."
World Development 27:201–13.

Aghion, Philippe, and Patrick Bolton. 1990. "Government Domestic Debt and Risk of
Default. A Political Economic Model of the Strategic Role of Debt." Pp. 315–45 in R.
Dornbusch and M. Draghi, eds., Public Debt Management: Theory and History.
Cambridge: Cambridge University Press.

Ahmad, Muzaffer. 1982. "Political Economy of Public Enterprises." Pp. 49–64 in L. P.
Jones, R. D. Mallon, E. S. Mason, P. N. Rodan-Rosenstein, and R. Vernon, eds., Public
Enterprises in Less Developed Countries. Cambridge: Cambridge University Press.

Alchian, Armen. 1965. "Some Economics of Property Rights." In W. L. Megginson, ed.,
The Financial Economics of Privatization. Oxford: Oxford University Press Inc.

Alesina, A. and A. Drazen (1991). Why are stabilizations delayed? American Economic
Review, 81, 1170-88.

Alesina, A. and P. Perotti (1995). The political economy of budget deficits. IMF Staff
Papers, March, 1-37.

Alesina, A., N. Roubini and G.D. Cohen (1997). Political cycles and the macro
economy.Cambridge (MA), The MIT Press.

Allio, Lorene, Mariusz Mark Dobek, Nikolai Mikhailov David L. Weimer (1997). Post-
Communist Privatization as a Test of Theories of Institutional Change, in: David L.
Weimer The Political Economy of Property Rights. Cambridge : Cambridge University
Press: 319–348.

Al-Obaidan, A. M. 2002. "Efficiency Effect of Privatization in Developing Countries."
Applied Economics 34:111–18.

Alt, James E. Robert C. Lowry (1994). Divided Government, Fiscal Institutions, and
Budget Deficits. Evidence from the States, American Political Science Review. 88: 811–
828.

Anastassopoulos, Jean-Pierre C., “The French Experience: Conflicts with Government,”
in Raymond Vernon and Yair Aharoni, eds., State-Owned Enterprise in the Western
Economies. London: Croom Helm, 1981. pp. 99-116.

149

Asfaw, K. 1996. "The Political Economy of Privatization in Sub-Saharan Africa."
International Review of Administrative Sciences 62:75–87.

Avishur, Arieh. 2000. "A Positive Theory of Privatization." Journal of Applied
Economics 3 (1):247–68.

Banerjee, Ghosh Sudeshna, and Michael C. Munger. 2004. "Move to Markets? An
Empirical Analysis of Privatization in Developing Countries." Journal of International
Development 16:213–40.

Banks A.S., Day T.C., Muller W.R. (2002), Political Handbook of the World 2000-2002 ,
CSA publications, State University of New York.

Banks, Jeffrey and R. Sundaram, “Long-lived Principles, Short-lived Agents,” University
of Rochester, Rochester Center for Economic Research Working Paper 353, July 1993.

Barnett, S. 2000. Evidence on the Fiscal and Macroeconomic Impact of Privatization.
IMF Working Paper. Washington, DC: IMF.

Bates, Robert, Markets and States in Tropical Africa: The Political Basis of Agricultural
Policies. Berkeley: Univ. of California Press, 1981.

Beck, T., A. Demirgüç-Kunt, and R. Levine (1999) A New Database on Financial
Development and Structure (updated 2003), World Bank Policy Research Working Paper
n. 2146.

Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer, Patrick Walsh (2001). New
Tools in Comparative Political Economy: The Database of Political Institutions, World
Bank Economic Review 15(1): 165–176.

Becker, Gary S. (1983). A Theory of Competition Among Pressure Groups for Political
Influence, Quarterly Journal of Economics Vol. 98(3): 371–400.

Besley, Timothy and Anne Case, “Does Electoral Accountability Affect Economic Policy
Choices? Evidence From Gubernatorial Term Limits,” Quarterly Journal of Economics,
August 1995, pp.769-798.

Biais, Bruno, and Enrico Perotti. 2002. "Machiavellian Privatization." American
Economic Review 92:240–58.

Bienen, H., and John Waterbury. 1989. "The Political Economy in Developing
Countries." World Development 17 (5):617–32.

Biglaiser, Glen, and David S. Brown. 2003. "The Determinants of Privatization in Latin
America." Political Research Quarterly 56 (1):77–89.

150

Boix, Charles (1998). Political Parties, Growth and Equality. Conservative and Social
Democratic Economic Strategies in the World Economy, Cambridge University Press,
Cambridge.

Borner, K. 2004. The Political Economy of Privatization: Why Do Governments Want
Reforms? The Fondazione Eni Enrico Mattei Note di Lavoro 106. Rome: Fondzaione Eni
Enrico Mattei.

Bortolotti, B., Fantini, M. and Siniscalco, M. (2003). Privatization around the world:
evidence from panel data, Journal of Public Economics, 88, 305-32.

Boubakari, Narjess, and Jean-Claude. Cosset. 1999. "Does Privatization Meet the
Expectations? Evidence from African Countries." Presented at the Plenary on
Privatization and Corporate Governance: Nairobi, Kenya.

Boycko, Maxim, Andrei Shleifer Robert Vishny (1995). Privatizing Russia. Cambridge :
MIT Press. Brada, Josef C. (1996). Privatization is Transition – Or is it, Journal of
Economic Perspectives 10(2): 67–86.

Brune, Nancy, Bruce Kogut, and G. Garrett. 2004. "The International Monetary Fund and
the Global Spread of Privatization." IMF Staff Papers 51 (2):195–218.

Buchanan, James M., Robert D. Tollison Gordon Tullock, ( (1980). Toward a Theory of
the Rent-Seeking Society. Texas : College Station.

Buchanan, James. 1972. The Theory of Public Choice. Ann Arbor, MI: University of
Michigan Press.

Campbell, White, and Anita Bhatia. 1998. Privatization in Africa: The Directions in
Development Series. Washington, DC: World Bank.

Caprio, Gerard and Daniela Klingebiel, “Bank Insolvency: Bad Luck, Bad Policy, or Bad
Banking?” paper presented at the World Bank’s Annual Conference on Development
Economics, 1996. Central Bank of the Republic of Argentina, “Main Features of the
Regulatory Framework of the Argentina Financial System,” mimeo, September 1996.

Caprio, Gerard, “Bank Regulation: The Case of the Missing Model” World Bank Policy
Research Working Paper 1574, January 1996.

Clarke, G. K., and R. Cull. 2002. "Political and Economic Determinants of the
Likelihood of Privatizing Argentine Banks." Journal of Law and Economics 45:165–98.

Clarke, George R.G. and Robert Cull “Why Privatize? The Case of Argentina’s Public
Provincial Banks,” World Bank mimeo, 1997.

151
Clarke, George Robert Cull (1998). The Political Economy of Privatization. An
Empirical Analysis of Bank Privatization in Argentina. World Bank Policy Research
Working Paper No. 1962.

Cook, Paul, and Colin Kirkpatrick, eds. 1988. Privatization in Less Developed Countries:
An Overview. Sussex: Wheatsheaf Books.

Cook, Paul, and Yuichiro Uchida. 2003. "Privatization and Economic Growth in
Developing Countries." Journal of Development Studies 39 (6):121–54.

Cox, D. R. (1972). Regression Models and Life Tables, Journal of the Royal Statistical
Society, Series B 24, 187-220.

Cox, Gary W. and Matthew D. McCubbins, “Structure and Policy: The Institutional
Determinants of Policy Outcomes,” mimeo, October 1996.

Cuervo, Alvaro, and Belen Villalonga. 2000. "Explaining the Variance in the
Performance Effects of Privatization." Academy of Management Review 25:581–94.

Cui, Zhiyuan. 1997. "Privatization and the Consolidation of Democratic Regimes: An
analysis and an Alternative." Journal of International Affairs 50:675–92.

Cull, Robert. “Financial Sector adjustment Lending: A Mid-Course Analysis,” World
Bank mimeo, July, 1997.

Demsetz, Harold (1967). Towards a Theory of Property Rights, American Economic
Review. Papers and Proceedings 57(2): 347–359.

Dinavo, J. 1995. Privatization in Developing Countries. Its Impact on Economic
Development and Democracy. Westport, CT: Praeger Publishers.

Djankov, Simeon Peter Murrell (2002). Enterprise Restructuring in Transition: A
Quantitative Survey, Journal of Economic Literature 40(3): 739–793.

Donahue, John D., The Privatization Decision. New York: Basic Books, 1989.

Drazen, Allan Vittorio Grilli (1993). The Benefit of Crises for Economic Reforms,
American Economic Review 83(3): 598–607.

Dumontier, P. and C. Laurin (2002). The financial impact of the French government
nationalization/privatization strategy, mimeo.

Ellul, A., and M. Pagano (2002). IPO undepricing and after-market liquidity, CSEF
University of Salerno, mimeo.

152
Fernandez, Raquel Dani Rodrik (1991). Resistance to Reform. Status Quo Bias in the
Presence of Individual-Specific Uncertainty, American Economic Review 81: 1146–55.

Filipovic, A. 2005. "Impact of Privatization on Economic Growth." Issues in Political
Economy 14:1–22.

Ford, Neil. 2004. "Ghana: Crunch Time Over Privatization." African Business 296:50–
51.

Galiani, S., P.Frey, Bruno Friedrich Schneider (1978). An Empirical Study of Politico-
Economic Interaction in the United States, The Review of Economics and Statistics
60(2): 174–183.

Galal, Ahmed, Leroy Jones, Pankaj Tandon, and Ingo Vogelsang, Welfare Consequences
of Selling Public Enterprises. Oxford: Oxford University Press, 1994.

Gallagher, M. (1991). Proportionality, disproportionality and electoral systems. Electoral
Studies, 10, 33-51.

Gavin, Michael Dani Rodrik (1995). The World Bank in Historical Perspective,
American Economic Review Papers and Proceedings. 85(2): 329–34.

Gertler, and E. Schargrodsky. 2005. "Water for Life. The Impact of Privatization of
Water Services on Child Mortality." Journal of Political Economy 113 (1):83–120.

Global Development Network Growth Database. Available at (http://www.nyu.edu).

Globerman, Stephen, and Daniel Shapiro. 2003. "Governance Infrastructure and US
Foreign Direct Investment." Journal of International Business Studies 34:1–19.

Goodman, Timothy John. 1998 Privatization and Democratic Governance: Political
Impact and Consequences of State Enterprises in Mexico and Argentina, 1988–1999.
Ph.D. Dissertation. Georgetown University.

Greene, W. Econometric Analysis 3rd ed. New York: Macmillan, 1997.

Greenwald, Bruce, and Joseph Stiglitz. "Externalities in Economies with Incomplete
Information and Incomplete Markets." American Economic Review 45, no. 1 (1955): 1–
25.

Hallerberg, Mark Basinger Scott (1998). Internationalization and Changes in Tax Policy
in OECD Countries: The Importance of Domestic Veto Players, Comparative Political
Studies. 31: 321–352.

Haltinwanger, J. And M. Singh (1999). Cross-country evidence on public sector
retrenchment, The World Bank Economic Review, 13:23-66.

153

Hanke, Stephen. 1987. Privatization and Development. Institute for Contemporary
Studies Press.

Hanke, Steve H., ed. Privatization and Development. San Francisco: International Center
for Economic Growth, 1987.

Harberger, Arnold. 1987. "Growth of the Public Sector in Our Time." Presented at the
Annual Meeting of Argentine Bankers Association. Buenos Aires.

Harris, John, Janet Hunter, and Colin Lewis, eds. The New Institutional Economics and
Third World Development. New York: Routledge, 1995.

Hausman, J. “Specification Tests in Econometrics.” Econometrica, 1978 , pp. 1251-1271.

Hellman, Joel S. Mark Schankerman (2000). Intervention, Corruption and Capture: The
Nexus Between Enterprises and the State, Economics of Transition 8(3): 545–576.

Hemming, Ethan, and Prabha Unnithan. 1996. "Determinants of Privatization in
Developing Countries." Social Science Quarterly 77 (2):432–44.

Henisz, Witold Bennet A. Zelner (2003). Legitimacy, Interest Group Pressures and
Institutional Change: The Case of Foreign Investors and Host Country Governments.
William Davidson Institute Working Paper No. 589.

Hibbs, D. (1977). Political parties and macroeconomic policy. American Political Science
Review, 71, 1467-78.

Hirschman, Albert. 1985. "Reflections on the Latin American Experience." American
Economic Review 83 (3):598–607.

Hirschman, Albert. 1987. "The Political Economy of Latin American Development:
Seven Exercises in Retrospection." Latin American Research Review 22 (3):7–36.

Huber, J. and R. Inglehart (1995). Expert interpretations of party space and party
locations in 42 societies, Party Politics, 1, 73-111.

Jones, Leroy P. (1985). Public Enterprise for Whom: Perverse Distributional
Consequences of Public Operational Decisions, Economic Development and Cultural
Change 33: 333–348.

Jones, S., W.L. Megginson, R. Nash, J. Netter (1999) Share issue privatizations as
financial means to political and economic ends, Journal of Financial Economics, 53, 217-
53.

154
Kalbfleisch, J and Ross L. Prentice, The Statistical Analysis of Failure Time Data. New
York: John Wiley and Sons, 1980.

Kaufmann, Daniel, Aart Kraay, and Mastruzzi Massimo 2005. Governance Matters 1V:
Governance Indicators for 1996–2004. World Bank Policy Research Working Paper
Series No. 3630. Washington, DC: World Bank.

Kayizzi-Mugerwa, K. S. 2002 Privatization in Sub-Saharan Africa: On Factors Affecting
Implementation. World Institute of Development Economic Research Discussion Paper
No. 2002/12. Helsinki.

Kayizzi-Mugerwa, K. S., ed. 2003. Reforming Africa's Institutions: Ownership,
Incentives and Capabilities. New York: United Nations University Press.

Keefer, Philip Stephen Knack (1997). Why Don't Poor Countries Catch Up? A Cross-
national Test of an Institutional Explanation, Economic Inquiry 35(3): 590–602.

Kiefer, N.M. (1988) Economic duration data and hazard functions, Journal of Economic
Literature, 26, 646--679.

Kikeri, Sunita, and Aishetu Kolo. 2005. Privatization: Trends and Recent Development.
World Bank Policy Research Working Paper 3765. Washington, DC: World Bank.

Kikeri, Sunita, John Nellis, and Mary Shirley, Privatization: The Lessons of Experience.
Washington, DC: The World Bank, 1992.

Knight, Jack Douglass C. North (1997). Explaining the Complexity of Institutional
Change, in: David L. Weimer ( The Political Economy of Property Rights. Cambridge :
Cambridge University Press: 349–354.

Kobeissi, Nada. 2005. "Impact of Governance and Economic Freedom on Foreign Direct
Investment." Journal of Comparative International Management 8:20–41.

Kornai, J. 1979. "Resource-Constrained Versus Demand-Constrained Systems."
Econometrica 47 (4):801–20.

Krueger, Anne O. (1998). Whither the World Bank and the IMF, Journal of Economic
Literature. 36(4): 1983–2020.

Kumar, Nagesh, and Jaya Prakash Pradhan. 2002. Foreign Direct Investment,
Externalities, and Economic Growth in Developing Countries: Some Empirical
Explorations and Implications for WTO Negotiations on Investment. RIS Discussion
Paper No. 27/2002. New Delhi, India: RIS.

La Porta, Rafael and Florencio López-de-Silanes, “The Benefits of Privatization -
Evidence from Mexico,” The World Bank, Viewpoint, note 117, 1997.

155

Laakso, M. and R. Taagepera (1979). Effective number of parties: a measure with
application to West Europe Comparative Political Studies 12: 3-27.

Laffont, J. J., and J. Tirole. 1991. "Privatization and Incentives." Journal of Law and
Economics 7:84–105.

Laffont, J. J., and M. Meleu. 1999. "A Positive Theory of Privatization for Sub-Saharan
Africa." Journal of African Economies 8:30–67.

Laffont, Jean-Jacques and Jean Tirole, “Privatization and Incentives,” Journal of Law,
Economics, and Organization, VII, September, 1991, pp. 84-105.

Laver, M. and K.A. Shepsle (1996). Making and breaking governments: cabinets and
legislatures in parliamentary democracies. New York, Cambridge University Press.

Lijphart, A. (1999). Patterns of democracy. New Haven and London, Yale University
Press.

Levy, B., and P. T. Spiller, eds. 1996. Regulations, Institutions, and Commitment:
Comparative Studies of elecommunications, Political Economy of Institutions and
Decision Series. Cambridge: Cambridge University Press.

Li, W., and L. C. Xu. 2002. "The Political Economy of Privatization and Competition:
Cross-Country Evidence from the Telecommunications Sector." Journal of Comparative
Economics 30:439–62.

Li, Wei Xiaonian Colin Xu (2002). The Political Economy of Privatization and
Competition: Cross-Country Evidence from the Telecommunications Sector, Journal of
Comparative Economics 30: 1–25.

Libecap, Gary (1989). Contracting for Property Rights. Cambridge University Press,
Cambridge.Lin, Justin Yifu (1989). An Economic Theory of Institutional Change:
Induced and Imposed Change, Cato Journal 9(1): 1–33.

Lindenberg, Marc, and Noel Ramirez, eds. Managing Adjustment in Developing
Countries. San Francisco: International Center for Economic Growth, 1989.
London: Croom Helm, 1981, pp. 85-98.

López-de-Silanes, Florencio, Andrei Shleifer, and Robert W. Vishny, “Privatization in
the United States,” RAND Journal of Economics, Autumn, 1997, pp. 447-71.

Megginson, W. L., Nash, R. C., van Randenborgh, M. (1994) The financial and operating
performance of newly privatized firms: An international empirical analysis, Journal of
Finance, 49, 403-452.

156
Megginson, W.L., J. Netter, (2001). From State to market: a survey of empirical studies
on privatization, Journal of Economic Literature, 39, 321-389.

Megginson, W.L., Nash, R.C., Netter, J.M., and Poulsen, A.B., (2004). The choice
between private and public markets: evidence from privatizations, Journal of Finance, 59,
2835-2870.

Megginson, William L. 2005. The Financial Economics of Privatization. Oxford: Oxford
University Press Inc.

Megginson, William L., and J. M. Netter. 2001. "From State to Market: A Survey of
Empirical Studies on Privatization of New Privatized Firms: An International Empirical
Analysis." Journal of Finance 49 (2):403–52.

Megginson, William L., Robert C. Nash, and Mathias van Randenborgh, “The Financial
and Operating Performance of Newly Privatized Firms: An International Empirical
Analysis,” Journal of Finance, XLIX, 1994, pp.403-52.

Mueller, Dennis C., Public Choice II. Cambridge, UK: Cambridge University Press,
1989.

Murillo, Victoria M. (2002). Political Bias in Policy Convergence. Privatization Choices
in Latin America, World Politics 54: 462–493.

Murphy, David (2003). Old Volkswagen Chases New China, Far Eastern Economic
Review. March 06: 26–28.Nordhaus, William (1975). The Political Business Cycle,
Review of Economic Studies 42: 169–190.

Nellis, John. 2003. Privatization in Africa: What Has happened? What is to be Done?
Center for Global Development, Working Paper Series No. 25. Washington, DC: Center
for Global Development.

Noorbakhsh, Fahrad, Alberto Paloni, and Ali Youssef. 2001. "Human Capital and FDI
Inflows to Developing Countries: New Evidence." World Development 29:1593–1610.

North, Douglass C. (1971). Institutional Change and Economic Growth,The Journal of
Economic History 31(1): 118–125.

North, Douglass C. (1981). Structure and Change in Economic History. New York :
W.W. Norton & Company.

North, Douglass C. (1994). Economic Performance through Time, American Economic
Review 84: 359–368.

157
North, Douglass C. Barry Weingast (1989). Constitutions and Commitment. The
Evolution of Institutions Governing Public Choice in Seventeenth Century England,
Journal of Economic History 49(4): 808–32.

Olson, Mancur (1965). The Logic of Collective Action: Public Goods and the Theory of
Groups. Cambridge : Harvard University Press.

Olson, Mancur (1982). The Rise and Decline of Nations. Economic Growth, Stagflation
and Social Rigidities. New Haven : Yale University Press.

Opper Sonja, Sonia M.L. Wong Ruyin Hu (2002). Party Power, Market and Private
Power: CCP Persistence in China's Listed Companies, Research in Social Stratification
and Mobility 19: 103–136.

Opper, Sonja. 2004. "The Political Economy of Privatization. Evidence from Transition
Economies." Kyklos 57 (4):559–83.

Perotti, R., Y. Kontopoulos (2002) Fragmented Fiscal Policy, Journal of Public
Economics, 82, pp. 191-222.

Persson, T. and G. Tabellini (2003). The Economic Effects of Constitutions, Cambridge
(MA), The MIT Press.

Persson, Torsten, Gerard Roland Guido Tabellini (1997). Separation of Powers and
Political Accountability, Quarterly Journal of Economics 112(4): 1163–1202.

Plane, Patrick. 1997. "Privatization and Economic Growth: An Empirical Investigation
from a Sample of Developing Market Economies." Applied Economics 29:161–79.

Raiser, Martin, Maria L. Di Tommaso Melwyn Weeks (2000): The Measurement and
Determinants of Institutional Change, Evidence from Transition Economies. EBRD
Working Paper No. 60.

Ramamurti, R. 1999. "Why Haven't Developing Countries Privatized Deeper and
Faster?" World Development 27 (1):137–55.

Ramamurti, R. 2000. "A Multilevel Model of Privatization in Emerging Economies."
Academy of Management 25:525–50.

Ramamurti, Ravi (1992). The Impact of Privatization on the Latin American Debt
Problem, Journal of International Business Studies. 23: 93–125.

Rodrik, Dani (1993). The Positive Economics of Policy Reform,The American Economic
Review, Papers and Proceedings. 83(2): 356–361.

158
Roland, Gerard (1994). On the Speed and Sequencing of Privatisation and Restructuring,
The Economic Journal 104(426): 1158–1168.

Roubini, N., and J, Sachs (1989) Political and economic determinants of budget deficits
in the industrial democracies, European Economic Review, 33, 903-33.

Roubini, Nouriel Jeffrey Sachs (1989). Political and Economic Determinants of Budget
Deficits in the Industrial Democracies, European Economic Review 33: 903–938.

Sappington, David E. Joseph E. Stiglitz (1987). Privatization, Information, and
Incentives. Journal of Policy Analysis & Management 6: 567–581.

Sartori, G. (1976). Parties and Party System: a Framework for Analysis, Cambridge,
Cambridge University Press.

Schipke, Alfred (2001). Why Do Governments Divest?— The Macroeconomics of
Privatization, Berlin : Springer Publishing.

Schultz, Kenneth A. (1995). The Politics of the Political Business Cycle, British Journal
of Political Science 25(1): 79–99.

Scully, Gerald W. 1988. "The Institutional Framework and Economic Development."
Journal of Political Economy 96:652–62.

Seo, Myeong-Gu W.E. Douglas Creed (2002). Institutional Contradictions, Praxis and
Institutional Change: A Dialectical Perspective, Academy of Management Review 27(2):
222–247.

Shapiro, Carl Robert Willig (1990). Economic Rationales for the Scope of Privatization,
Olin Program Discussion Paper 41. Princeton University.Shleifer, Andrei Robert Vishny
(1994). Politicians and Firms, Quarterly Journal of Economics. 109: 995–1025.

Shleifer, Andrei and Robert Vishny, “Politicians and Firms,” Quarterly Journal of
Smith, R.J., and R.W. Blundell (1986). An Exogeneity Test for a Simultaneous Equation
Tobit Model with an Application to Labor Supply, Econometrica, 54, 679-86.

Spolaore, E. (2004). Adjustments in Different Government Systems, Economics and
Politics, 16(2). Tsebelis, G. (2001).Veto Players Data.
(http://www.polisci.ucla.edu/tsebelis/vpdata.html)

Starr, Paul. 1988. "The Meaning of Privatization." Yale Law and Policy Review 6:6–41.
Stiglitz, Joseph. Globalization and its Discontents. New York: Penguin Books, 2002.

Tommasi, M and Velasco, A. (1996), “Where are we in the political economy of
reform?” Policy Reform, vol. 1, no. 2.

159
Van den Berg , G.J. (2001). Duration models: Specification, identification, and multiple
durations, in J.J. Heckman and E. Leamer, editors, Handbook of Econometrics, Volume
V, North Holland, Amsterdam.

Van der Hoeven, Rolph, and Gyorgy Sziraczki, eds. Lessons from Privatization: Labour
Issues in Developing and Transitional Countries. Geneva: International Labour Office,
1997.

Vella, F., and M. Verbeek (1999). Two-step estimation of panel data models with
censored endogenous variables and selection bias, Journal of Econometrics, 90, 239-63.

Vickers, J., and Yarrow, G. (1988). Privatization: an economic analysis, Cambridge
(MA), MIT Press.

Vernon, Raymond, and Yair Aharoni, eds., State-Owned Enterprise in Western
Economies. London: Croom Helm, 1981.

Vernon, Raymond, ed. 1988. The Promise of Privatization: A Challenge for US
Economic Policy. New York: Council on Foreign Relations.

Vickers, John George Yarrow (1991). Economic Perspectives on Privatization, Journal of
Economic Perspectives 5(2): 111–132.

Vining, Aiden and Anthony Boardman, “Ownership vs. Competition: Efficiency in
Public Enterprise,” Public Choice, LXXIII, March, 1992, pp. 205-39.

Wallsten, Scott J. 2001. "An Econometric Analysis of Telecom Competition,
Privatization and Regulation in Africa and Latin America." Journal of Industrial
Economics 59:1–19.

Winiecki, Jan (1990). Why Economic Reforms Fail in the Soviet System – A Property
Rights Based Approach, Economic Inquiry 28: 195–211.

Woldendrop, J., H. Keman and I Budge (1998). Party government in 20 democracies: an
update European Journal of Political Research, 33, 125-164.

Wong, Sonia M.L., Sonja Opper Ruyin Hu (2004). Shareholding Structure, De-
politicization and Enterprise Performance: Lessons from China's Listed Companies,
Economics of Transition 12(1): 29–66.

World Bank (1995). Bureaucrats in business, Oxford, Oxford University Press.

Woolridge (2002). Econometric Analysis of Cross Section and Panel Data, Cambridge
(MA), The MIT Press.

World Development Indicators. Washington, DC: World Bank.

160

Yarrow, George (1998). A Theory of Privatization, or Why Bureaucrats are Still in
Business, World Development 27(1): 157–168.

Abstract (if available)

Abstract Empirical observation of partial privatizations across countries is puzzling. This research finds empirical evidence that privatization programs have not been driven by ideological or “efficiency” reasons, but rather by the pragmatic cost-benefit tradeoffs made by the politicians, and the economics of privatization often dominates its politics. Using data from 43 countries on more than 4700 privatization transactions, I find strong empirical support for Institutional Quality as consistent and significant determinant of proportion of partial privatization. Surprisingly, countries having higher corruption tend to have higher proportion of privatization in competitive sector, but lower privatization in core sector. Counter to anecdotal evidence, this research finds that political constraints have no significant impact on partial privatization proportion. Further, fiscal crisis drives politicians to privatize, but has no significant effect on privatization proportion. These empirical findings motivate a political economy model of privatization and three results stand out: first, the distortion in the privatization proportion depends upon the institutional quality parameter relative to a measure of private sector efficiency and the distortion increases as institutional quality declines

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Political determinants and economic effects of HIV/AIDS: a push for the multisectoral approach

PDF

Essays in political economy and mechanism design

PDF

Political economy of private commercial arbitration in Colombia

PDF

The political economy of R&D collaboration: micro- and macro-level implications

PDF

Essays on political economy and corruption

PDF

The political economy of Brazilian trade policy: domestic determinants, world and regional strategies

PDF

Essays on sovereign debt

PDF

An empirical analysis of the quality of primary education across countries and over time

PDF

The political economy of implementation: intellectual property rights protection across the world

PDF

The political economy of oil, terrorism and institutional development in Iran and its impact on the Middle East

PDF

Essays on economics of education and private tutoring

PDF

Essays on inflation, stock market and borrowing constraints

PDF

Financial crises and trade policy in developing countries

PDF

Essays in international economics

PDF

Essays on interest rate determination in open economies

PDF

Against the wind: labor force participation of women in Iran

PDF

Essays on human capital accumulation -- health and education

PDF

The impact of social capital on economic performance: lessons from small and medium size enterprises (SMEs)

PDF

Three essays on the microeconometric analysis of the labor market

PDF

Creating destruction: the political economy of zombie firms

Asset Metadata

Creator Shivendu, Shivendu (author)

Core Title Essays on political economy of privatization

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Economics

Publication Date 12/09/2008

Defense Date 05/09/2005

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag determinants of partial privatization,firm performance,OAI-PMH Harvest,Political Economy,privatization contracts under assymetric information,privatization program in India

Place Name India (countries)

Language English

Advisor Nugent, Jeffrey B. (committee chair), Chellappa, Ramnath (committee member), Dekle, Robert (committee member), Moon, Hyungsik Roger (committee member)

Creator Email shivendu@usc.edu,sshivendu@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1901

Unique identifier UC175189

Identifier etd-shivendu-2334 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-142968 (legacy record id),usctheses-m1901 (legacy record id)

Legacy Identifier etd-shivendu-2334.pdf

Dmrecord 142968

Document Type Dissertation

Rights Shivendu, Shivendu

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu