University of Southern California Dissertations and Theses

Social policy and presidential ideology in Latin America: the political economy of social spending and anti-poverty programs

(USC Thesis Other)

Social policy and presidential ideology in Latin America: the political economy of social spending and anti-poverty programs

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Social Policy and Presidential Ideology in Latin America:
The Political Economy of Social Spending and
Anti-Poverty Programs

By Fabián A. Borges

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
POLITICAL SCIENCE AND INTERNATIONAL RELATIONS

December 2014
Social Policy and Presidential Ideology in Latin America
ii

Abstract:

Social Policy and Presidential Ideology in Latin America:
The Political Economy of Social Spending and Anti-Poverty Programs

By Fabián A. Borges

Doctor in Philosophy in Political Science

University of Southern California

Professor Gerardo L. Munck, Chair

This study explains the relationship between arguably the two most significant developments in
Latin American politics over the last 15 years: the widespread election of left-leaning presidents
and the implementation of large-scale anti-poverty programs. In doing so, it sheds light on one of
the major questions in the broader comparative political economy literature: what has been the
effect of ideology and left-leaning governments on social policy during the era of globalization?

Starting in 1998, a wave of left-leaning candidates campaigning on promises of higher social
spending and income redistribution came to power in most Latin American countries. As this
political transformation was taking place, social policy in the region underwent a transformation of
its own. Historically, social policy in the region has benefited the middle class and politically
organized groups while largely neglecting the rural poor and urban informal sector that make up
the majority of the population. However, starting in 1997, governments throughout the region
began launching ambitious social policy initiatives targeting the poor. The most notable of these
programs have been conditional cash transfer (CCT) programs, which award stipends to families
conditional on their children remaining in school. Today, these programs cover nearly a quarter of
the region’s population.

Through a combination of cross-national statistical analyses on a sample of 18 countries and in-
depth case studies based on field research in Argentina, Bolivia and Costa Rica, this study finds
that the effect of presidential ideology on social policy has been varied and complex. In Latin
America, as in industrialized countries, the effect of ideology on social spending has been weak
during the era of globalization. However, spending masks the substantial autonomy Latin American
presidents have to shape social policy. Although left presidents are no more likely to adopt CCTs
than other presidents, levels of CCT coverage relative to the size of a country’s population tend to
be substantially higher under left presidents than under centrist or right-wing presidents.

A deeper look at the political process that culminated in CCT adoption in Argentina, Bolivia and
Costa Rica reveals that politicians from the right and center – not the left – were responsible for
introducing CCTs into the political agenda of those countries. The left’s initial position on CCTs
in those countries ranged from apathetic (Costa Rica and Bolivia) to outright opposition
(Argentina). The left ultimately embraced these policies once they had been incorporated into the
political agenda and had proven to be both effective at reducing poverty and politically popular at
the local level and abroad.
Social Policy and Presidential Ideology in Latin America
iii

The effect of ideology on CCT design extends beyond coverage levels to the core mission of these
programs. CCTs enacted by left-leaning presidents in Argentina and Bolivia emphasize the goals
of poverty reduction and achieving high levels of coverage over the goals human capital
accumulation and conditionality enforcement. The opposite was true of the CCT adopted by a
centrist president in Costa Rica: it targets a very narrow population with the explicit goal of
improving high school attendance.

Thus, true to their ideological roots and despite constraints on their ability to increase spending,
left-leaning presidents moved further toward establishing a basic income floor for all citizens and
in the process engaged in greater redistribution toward the poor. Although CCTs were not the left’s
idea, it eventually ran with them, expanding the coverage of existing programs and emphasizing
poverty reduction in the design of new programs. The expansion of CCTs, which was heavily
influenced by presidential ideology, helps to explain decreasing inequality in Latin America at a
time when it has been increasing in both industrialized countries and the largest emerging markets.

Social Policy and Presidential Ideology in Latin America
iv

To Gaby:
You’ll never know how hard this was…

To Rebecca:
You know exactly how hard this was…
Thank you!
Social Policy and Presidential Ideology in Latin America
v

Table of Contents

List of Tables vi

List of Figures vii

Acknowledgements viii

Chapter 1 Recent Transformations in Latin American Politics:
The Left Turn and the Rise of Conditional Cash
Transfers
1

Chapter 2 Social Policy in Industrialized Countries and Latin
America: Past and Present
11

Chapter 3 Presidential Ideology or Income Distribution?:
Determinants of Social Spending, 1990-2010
62

Chapter 4 Does the Left Actually Redistribute More?: Adoption
and Scope of Conditional Cash Transfer Programs
1996-2010
89

Chapter 5 Ideology and Varieties of CCTs: Adoption and Design
of CCTs in Three Latin American Countries
122

Chapter 6 Conclusion and Pending Research Agenda 182

References 190

End Notes 231

Social Policy and Presidential Ideology in Latin America
vi

List of Tables

Table 1-1 Poverty and per Capita GDP (2009) in 18 Latin American Countries 4

Table 1-2 Average Social Spending (2004-2008) in 18 Countries 5

Table 1-3 CCTs in Latin America by Target Population, Size and Cost 10

Table 2-1 Social Spending Literature: Effects of Politics 28

Table 2-2 Determinants of Social Spending: Effects of Globalization 29

Table 2-3 Effects of Social Spending on Poverty and Inequality 30

Table 2-4 Impact of CCTs on Poverty in 13 Latin American Countries 43

Table 2-5 Cost of CCTs in 16 Latin American Countries 44

Table 2-6 Political Effects of CCTs 50

Table 2-7 Leakages in CCTs in 12 Latin American Countries
(2010 or Most Recent Year Available)
60

Table 3-1 Key Explanatory Variables 71

Table 3-2 Total Social Spending to GDP (Baseline Models) 76

Table 3-3 Total Social Spending to GDP (Full Model) 78

Table 3-4 Total Social Spending (Robustness Checks) 81

Table 3-5 Disaggregated Spending to GDP 83

Table 3-6 Findings on the Determinants of Social Spending 84

Table 4-1 Presidential Ideology and CCT Adoption 95

Table 4-2 Determinants of CCT Adoption 1996-2010 101

Table 4-3 Findings on the Determinants of CCT Adoption 104

Table 4-4 Determinants of CCT Coverage as a % of Total Population (1996-2010) 109

Table 4-5 Determinants of CCT Coverage as a % of Poor Population (1996-2010) 112

Table 4-6 Findings on the Determinants of CCT Coverage 114

Table 5-1 Case Selection Criteria 131

Social Policy and Presidential Ideology in Latin America
vii

List of Figures

Figure 1-1 Latin America’s Left Turn 1998-2014 2

Figure 1-2 Latin American Countries with CCTs: 1997 (Left) and 2014 (Right) 6

Figure 2-1 Number of CCT Beneficiaries by Year in Latin America, Brazil and Mexico
(2001-2010)
40

Figure 2-2 CCT Beneficiaries as a Share of Latin America’s Population (2001-2010) 41

Figure 2-3 Impact of CCTs on Poverty Headcount ($2.5 US PPP Poverty Line) in
Selected Latin American Countries
42

Figure 2-4 CCT Benefits as a Share of Average Incomes ($2.5 US PPP Poverty Line)
in Selected Latin American Countries
43

Figure 3-1 Estimated Effects of Income Distribution Skew 79

Figure 4-1 Number of CCTs Operating in Latin America (1996-2010) 97

Figure 4-2 Estimated Effect of Time, Diffusion and Income Inequality on CCT
Adoption (Model 1-1)
103

Figure 5-1 Cash Transfer in Argentina 132

Figure 5-2 Primary School Children (Ages 6 to 12) Enrolled in AUH by Income Level 144

Figure 5-3 Secondary School Children (Ages 13 to 17) Enrolled in AUH by Income
Level
144

Figure 5-4 Percent of Argentine Minors (0 to 17) Covered by Social Security Regimes
by Income Quintile (2010)
145

Figure 5-5 Consumer Price Inflation and the Purchasing Power of AUH (2009-2014) 146

Figure 5-6 Bolivia, Bono Juancito Pinto, Number of Beneficiaries and Grades Covered
(2005-2014)
147

Figure 5-7 Bolivian Cash Transfer Programs Share of Beneficiaries by Income (2009) 154

Figure 5-8 Bolivian Cash Transfer Programs Coverage by Beneficiary Income (2009) 155

Figure 5-9 Consumer Price Inflation and the Purchasing Power of Bolivian Cash
Transfer Programs (2006-2014)
156

Figure 5-10 Costa Rica, Avancemos, Number of Beneficiaries and Beneficiaries
Relative to the Total Public High School Population Covered (2005-2013)
157

Figure 5-11 Costa Rica, Avancemos Beneficiaries by Proxy-Means Test Score (2006-
2011)
165

Figure 5-12 Costa Rica, Consumer Price Inflation and Purchasing Power of Avancemos
(2006-2014)
166

Social Policy and Presidential Ideology in Latin America
viii

Acknowledgements

This project has been a long time in the making and would not have been possible without
the many institutions, colleagues, friends, and family members who helped me on this journey.
First and foremost, I owe a large debt of gratitude to my dissertation advisor, Professor
Gerardo Munck, who helped transform a vague desire to study social policy in Latin America into
a full research agenda that spanned four years. Gerry proved accessible and good humored, yet
uncompromising in his commitment to quality scholarship – everything I could want from an
advisor. His tireless work ethic pushed me to work harder every day and not settle. I am proud to
call Gerry a mentor, a role model, and a friend.
I am also grateful to Professor Carol Wise, who has been a supportive mentor and a tireless
advocate on my behalf since the day I first set foot on USC’s campus. She encouraged me to submit
my work to journals early on, an experience that, although initially frustrating and discouraging,
toughened me up early in my academic career. And I would like to personally acknowledge the
decisive role Professor Jefferey Sellers played in helping me develop this research agenda. I was
first exposed to research on social policy and the welfare state in his Comparative Politics seminar,
where he encouraged me to think about how this literature applied to developing countries. Later,
he agreed to sponsor a directed research course where I explored the topic further and ultimately
concluded that it was worth spending the next four years on.
This study would not have been possible without the generous funding and continued
support provided by USC’s Political Science and International Relations Doctoral Program, the
School of International Relations and the Center for International Studies. I am especially grateful
to Professor Patrick James, director of the Center for International Studies, and the CIS staff,
particularly Indira Persad and Marisela Schaffer, for the support they provided during my time as
Social Policy and Presidential Ideology in Latin America
ix

a CIS Dissertation Fellow. It was comforting to be part of a congenial office environment during a
writing process that would have otherwise been the loneliest time of my professional career. I
would also like to thank Professor Saori Katada and Veridiana Chavarin of the POIR program and
Linda Cole of the School of International Relations for all their help during these years. Portions
of this study were made possible by the generous funding received from USC through the Judith
Grayson Manning Memorial Fellowship, Gold Family Fellowship, Zilpha R. and Joan A. Main
Dissertation Completion Fellowship, and the Dornsife College of Letters, Arts and Sciences Field
Research Grant.
Despite the occasional bout of altitude sickness, field research in Latin American proved to
be a much more valuable and rewarding experience than I could have ever imagined. I was
consistently humbled by the willingness of busy people to share their valuable time with me. In
Argentina, Cintia Quiliconi, a close friend and fellow Trojan, helped me find housing in Buenos
Aires on short notice and provided me with office space at FLACSO-Argentina’s International
Relations Department. Valeria Loira and Ezequiel lo Valvo provided me with full access to ANSES
and with good humor answered my obscure questions regarding the functioning of Asignación
Universal por Hijo. Mercedes Botto was kind enough to put me into contact with several former
policymakers. I am indebted to the World Bank’s Rafael Rofman, who exposed me to the idea that
the goals of CCTs in Argentina and Bolivia might be different from those of CCTs in Mexico and
Costa Rica. That insight proved central to the argument made in chapter 5. In Bolivia, I was lucky
to count on the support of Tatiana Vargas, Costa Rica’s chargé d'affaires, who, in addition to being
a gracious host, allowed me to work out of the Costa Rican Embassy in La Paz. I am also indebted
to Orlando Murillo for providing me documents on Bono Juancito Pinto that I would not have been
able to access otherwise given much of the government’s reluctance to talk to a student from a U.S.
Social Policy and Presidential Ideology in Latin America
x

university. In Costa Rica, Katherine Stanley-Obando put me in touch with former President Oscar
Arias and Education Minister Leonardo Garnier. Over drinks, Carla Valverde, a dear friend and
former Avancemos staffer, provided a first-hand account of the political intrigue that surrounded
that program during its first years in operation. Olga Sonia Vargas and the staff of Avacemos
showed great interest in my research and did everything in their power to help me see it to fruition.
I came out of this experience thoroughly convinced that, far from an outdated practice, field
research remains an invaluable asset in an era when cross-national data can be easily downloaded.
Back at USC, the great friends I made were perhaps the best part of the graduate school
experience. The companionship and support of dear friends played a crucial role in getting me
through this long process. I am particularly grateful to Justin Berry, Mariano Bertucci, Juve Cortés,
Nic de Zamaroczy, Parker Hevron, Deniz Kuru, Jess Liao, Jillian Medeiros, Mike Perez, Cintia
Quiliconi, Simon Radford, Mariana Rangel, and Christina Wagner Faegri. In addition, I thank
Mauricio Rivero Celestino and Bárbara Zárate Tenorio, whose sharp comments greatly improved
early drafts of chapters 3 and 4. Beyond academia, Steve Barry and Carolyn and Colin Coyne have
been great friends to me and my family and helped make Los Angeles a home to us.
Finally, I would like to thank my family for instilling in me the desire to be a researcher. I
thank my mother, Iris, and father, David, for their love and support in every phase of my life, this
project being no exception. I thank my aunts Pilar and Libia, uncle Fernando and cousin Marcelo
for providing me with a home in Costa Rica even though I have not lived there in almost a decade.
Finally, words cannot convey the gratitude I owe Rebecca, my wife, part-time editor and full-time
cheerleader. She encouraged me to come to the U.S. for graduate school and supported my plan
pursue a doctorate. She worked tirelessly to build a home for us in Los Angeles, bent over
backwards to cheer me up when I occasionally “lost my smile,” and was a pillar of support during
Social Policy and Presidential Ideology in Latin America
xi

a period of uncertainty surrounding my health. I also thank her for bringing into my life Gaby, my
daughter. The afternoons of “daddy daycare” we spent together during the last year and a half of
this project helped clear my head and taught me to keep things in perspective.
As always, all mistakes are my own.

Social Policy and Presidential Ideology in Latin America
1

Chapter 1.
Recent Transformations in Latin American Politics:
The Left Turn and the Rise of Conditional Cash Transfers

1. Introduction
Latin America has undergone two significant transformations over the past 15 years: the
widespread election of left-leaning presidents and the implementation of large-scale anti-poverty
programs. The goal of this study is to explain the relationship between these two transformations.
In doing so, it sheds light on one of the major questions in the broader comparative political
economy literature: what has been the effect of ideology and left-leaning governments on social
policy during the era of globalization?
In the years following Latin America’s transition to democracy, governments were
overwhelmingly from the center and right and practically every government implemented market-
oriented reforms – a right-wing economic project (Weyland 2010, 1). This all began to change
with the election of Hugo Chávez as president of Venezuela in 1998. The region’s politics have
since undergone an undeniable “left turn” (see figure 1). Left-leaning candidates campaigning on
promises of higher social spending and income redistribution came to power in Chile (2000),
Brazil (2001) and Argentina (2003).
1
With time Chávez’ political project became increasingly
radical (Corrales and Penfold 2011) and radical leftists were also elected in Bolivia (2005) and
Ecuador (2006). This prompted researchers and political commentators to distinguishing between
a moderate social-democratic left and a radical populist one (Castañeda 2006; Weyland, Madrid
and Hunter 2010).
2
As this chapter was being written, the left remained in power in all of the
countries mentioned above. And, with the election of left-leaning leaders in Nicaragua (2006), El
Social Policy and Presidential Ideology in Latin America
2

Salvador (2009) and Costa Rica (2014), the left turn spread to Central America, a region
traditionally considered the backyard of the United States (Pastor 2001; Smith 2012).
3

Figure 1-1: Latin America’s Left Turn 1998-2014

As this political transformation was taking place, social policy in the region underwent a
transformation of its own. Historically, Latin America has pursued what have been aptly described
Social Policy and Presidential Ideology in Latin America
3

as “reverse Robin Hood” social policies (Lindert, Skoufias, and Shapiro 2006), which primarily
benefit the middle class and politically organized groups (Mesa-Lago 1978) while largely
neglecting the rural poor and urban informal sector. Social policy was traditionally centered on
dependable and, in some cases, quite generous contributory social insurance for formal-sector
workers and, at best, minimal safety nets for those outside the formal labor market. Given the
region’s high levels of informality (Perry et al. 2007), this approach to social policy resulted in
very limited (and in some cases even negative) income redistribution (Breceda, Rigolini and
Saavedra 2009; Goñi, López and Servén 2011; Lustig et al. 2012), despite what were, by
developing country standards, high levels of social spending (Haggard and Kaufman 2009).
The sustainability of this approach to social policy came into question in the aftermath of
the 1980s Debt Crisis and the decade-long economic slump that followed. This crisis led to sharp
reductions in all forms of social spending at a time when poverty and inequality had risen sharply
(Huber 1996; Huber and Stephens 2005). However, starting in 1997 with the adoption of Progresa
(later Oportunidades) in Mexico, governments throughout the region began launching ambitious
social policy initiatives targeting these traditionally excluded groups. The most notable of these
programs have been conditional cash transfer (CCT) programs, which award stipends to families
conditional on their children remaining in school. In the short-term, CCTs relieve poverty by
supplementing the regular incomes of needy families. In the long-run, they aim to build human
capital with the goal of improving the future employment prospects of beneficiaries and ultimately
preventing the inter-generational transmission of poverty. These programs, which have been called
“the most important innovation in social policy of the last fifteen years” (Lustig 2011, 19) now
cover nearly a quarter of the region’s population (Stampini and Tornarolli 2012).

Social Policy and Presidential Ideology in Latin America
4

1.2. Case Selection: Why Latin America?
Beyond Latin America, this study speaks to a broader debate within comparative political
economy: the effect of government ideology on social policy under globalization. Latin America
offers an ideal setting in which to study this topic for two reasons. First, Latin America is a
developing region, studying it serves to counterbalance the heavy focus on Western Europe in the
welfare state literature (Carnes and Mares 2007). And, second, Latin America stands out because
it is the most democratic developing region and, thus, complies with the scope condition for the
main explanations of social policy: the power resource theory of the welfare state (Esping-
Andersen 1985; Huber and Stephens 2001; Korpi 1983; Stephens 1979) and explanations of
redistribution based on income distribution (Meltzer and Richard 1981).
Table 1-1: Poverty and per Capita GDP (2009)
Country Year Poverty Indigence
GDP
per Capita
(PPP$ 2009)
Argentina 2009 11.3 3.8 13,202
Bolivia 2007 54.0 31.2 4,013
Brazil 2009 24.9 7.0 9,455
Chile 2009 11.5 3.6 13,057
Colombia 2009 45.7 16.5 8,136
Costa Rica 2009 18.9 6.9 10,085
Dominican Republic 2009 41.1 21.0 7,658
Ecuador 2009 40.2 15.5 7,508
El Salvador 2009 47.9 17.3 6,020
Guatemala 2006 54.8 29.1 4,286
Honduras 2007 68.9 45.6 3,488
Mexico 2008 34.8 11.2 12,429
Nicaragua 2005 61.9 31.9 2,398
Panama 2009 26.4 11.1 11,857
Paraguay 2009 56.0 30.4 4,107
Peru 2009 34.8 11.5 7,836
Uruguay 2009 10.7 2.0 11,977
Venezuela 2008 27.6 9.9 11,190
Sources: Poverty data from the Economic Commission for Latin America
and the Caribbean’s 2010 Social Panorama Report (ECLAC 2010). GDP
data from the World Bank’s Worldwide Development Indicators (World
Bank 2011).

Social Policy and Presidential Ideology in Latin America
5

The region has also exhibited substantial variation during the period studied in terms of the
key explanatory variable – presidential ideology – and dependent variables – levels of social
spending and CCT design. As noted, after electing primarily centrist and right-leaning
governments during the 1990s, most of the region’s countries elected left-leaning presidents during
the 2000s (Castañeda 2006; Flores-Macías 2012; Levitsky and Roberts 2011; Weyland, Madrid
and Hunter 2010).
The region presents substantial variation in terms of poverty (see table 1-1) and social
spending levels (see table 1-2), as well as in the overall development of social policy. In the 1920s,
Southern Cone countries were the first developing countries to establish comprehensive pension
systems (Mesa-Lago 1989) and by 1980 several of the region’s countries had full-fledged welfare
states (Huber and Stephens 2005, 554). However, many of the region’s poorer countries, primarily
in Central America and the Andean region, failed to develop broad-based and institutionalized
social policy (Filgueira 2005; Martínez Franzoni 2008; Mesa-Lago 1978; Pribble 2011).
Table 1-2: Average Social Spending (2004-2008) in 18 Countries
Education Healthcare Social Security Housing Total
Argentina 4.9% 4.7% 9.6% 1.7% 20.9%
Bolivia 6.5% 3.2% 5.1% 2.1% 16.9%
Brazil 5.0% 4.6% 12.7% 1.7% 23.9%
Chile 3.5% 2.9% 6.2% 0.3% 13.0%
Colombia 3.0% 1.9% 6.4% 0.6% 11.8%
Costa Rica 5.4% 5.1% 5.3% 1.9% 17.8%
Dominican Republic 2.0% 1.3% 2.2% 1.9% 7.5%
Ecuador 2.6% 1.2% 2.2% 0.2% 6.2%
El Salvador 3.2% 3.6% 2.5% 2.2% 11.5%
Guatemala 2.9% 1.2% 1.1% 2.1% 7.3%
Honduras 7.7% 3.5% 0.3% 0.1% 11.5%
México 3.9% 2.7% 2.5% 1.9% 11.0%
Nicaragua 5.0% 3.5% NA NA NA
Panama 3.9% 2.2% 1.3% 1.2% 8.6%
Paraguay 4.2% 1.7% 3.3% 0.3% 9.4%
Peru 2.7% 1.3% 3.9% 0.6% 8.5%
Uruguay 3.8% 3.7% 11.8% 1.6% 20.9%
Venezuela 5.3% 1.7% 4.3% 1.2% 12.5%
Source: CEPAL Social Spending Database

Social Policy and Presidential Ideology in Latin America
6

Within-region variation in spending is jarring. According to the United Nations Economic
Commission for Latin America (CEPAL), average total social spending as a share of GDP during 2005-
2009 ranged from almost 25 percent in Brazil to less than 5 percent in Ecuador (see table 1-2). Whereas
countries in the Southern Cone spend roughly $1,200 per person per year, the poorer Andean and Central
American countries spend less than $200. In brief, the variation in social spending in Latin America makes
it well-suited for testing and refining theories developed in Western Europe.
Figure 1-2: Latin American Countries with CCTs: 1997 (Left) and 2014 (Right)
With regard to CCTs, although most countries in the region now possess such a program, it took
more than a decade for these programs to spread throughout the region (see figure 1-2). Whereas in 1998
only Mexico had a national-level CCT (Brazil had local-level programs), by 2008 nearly every country in
the region had adopted a CCT program with the bulk of these adoptions occurring after 2005. Furthermore,
there is substantial variation in the scope of such programs: whereas nearly 60 percent of Bolivians benefit
from a CCT program, only 4 percent of Costa Ricans benefit (see table 1-3).

Social Policy and Presidential Ideology in Latin America
7

3. Research Design and Findings
To explain the relationship between these two transformations, this study analyzes the
effects of presidential ideology on two facets of social policy in Latin America: levels of social
spending and CCTs. This is done through a combination of cross-national statistical analyses on a
sample of 18 countries and in-depth case studies based on field research in Argentina, Bolivia and
Costa Rica.
This study finds that the effect of presidential ideology on social policy has been varied
and complex. First, in Latin America, as in industrialized countries, the effect of ideology on social
spending has been weak during the era of globalization. However, spending masks the substantial
autonomy Latin America’s presidents have to shape social policy. Although left presidents are no
more likely to adopt CCTs than other presidents, levels of CCT coverage relative to the size of a
country’s population tend to be substantially higher under left presidents than under centrist or
right-wing presidents.
A deeper look at the political process that culminated in CCT adoption in Argentina,
Bolivia and Costa Rica reveals that politicians from the right and center were responsible for
introducing CCTs into the political agenda of those countries. The left’s initial position on CCTs
in those countries ranged from apathetic (Costa Rica and Bolivia) to outright opposition
(Argentina). The left did ultimately embrace these policies once they had been incorporated into
the political agenda and had proven to be both effective and politically popular at the local level
and abroad. However, when the left adopted a CCT, it tended to emphasize the goals of poverty
reduction and increased coverage over the goals of human capital development and conditionality
enforcement.
Social Policy and Presidential Ideology in Latin America
8

Thus, true to their ideological roots and despite constraints on their ability to increase
spending, left-leaning presidents moved further toward establishing a basic income floor for all
citizens and in the process engaged in more redistribution in favor of the poor. Although CCTs
were not the left’s idea, it eventually ran with the idea, expanding the coverage of existing
programs and emphasizing poverty reduction in the design of new programs. The expansion of
CCTs, which was heavily influenced by ideology, helps to explain decreasing inequality in Latin
America at a time when it has been increasing in both industrialized countries and the largest
emerging markets (Lustig, López-Calva and Ortíz-Juárez 2013; Soares et al. 2009; World Bank
2014, 27).

4. Organization of this Study
The following chapter reviews the existing literature on the determinants of social policy
in industrialized countries and Latin America before assessing the growing body of research on
CCT programs. This is followed by three empirical chapters. The study then concludes by
summarizing the results and presenting unresolved puzzles that emerge from the empirical
analysis.
Through a time-series cross-sectional analysis of 18 Latin American countries, chapter 3
demonstrates that income distribution is a more important determinant of social spending levels
than is ideology. The factor that determines spending levels is not, as the median voter hypothesis
(Meltzer and Richard 1981) predicts, how unequal a country’s income distribution is but, rather,
how skewed it is in favor of the rich (Lupu and Pontusson 2011). As the gap between the shares
of a country’s wealth controlled by the rich and middle class expands, the middle class becomes
more likely to align with the poor in favor of increased spending. While having a center-left
Social Policy and Presidential Ideology in Latin America
9

president has no effect on spending levels, there is evidence that spending is higher under far-left
presidents. This effect operates primarily through education spending.
Chapters 4 and 5 focus on the politics of CCT adoption and design. Chapter 4 conducts
event history and times-series cross-national analyses covering the same 18 countries. These
analyses demonstrate that, although ideology has no effect on the probability of adopting a CCT,
it does affect the scope of programs once adopted. Specifically, CCT coverage, both as a share of
the population and relative to the share of a country’s population living in poverty, tends to be
higher in countries governed by the left. For example, while non-left presidents enacted CCTs in
Brazil and Ecuador, these programs remained relatively small until the left came to power and
expanded them.
Chapter 5 complements the preceding chapter by conducting in-depth country case studies
on the adoption and design of CCTs in Argentina, Bolivia and Costa Rica based on more than 50
elite interviews conducted during field research in those countries. The cases demonstrate that
politicians of the center and right – and not the left – were responsible for introducing CCTs in the
political agendas of those countries. Presidential ideology did, however, play a decisive role in
shaping the design of CCTs once the decision had been made to adopt them at the national level.
CCTs enacted by presidents of the left in Argentina and Bolivia emphasize poverty reduction and
achieving high levels of coverage over human capital accumulation and conditionality
enforcement. The opposite was true of the CCT adopted by a centrist president in Costa Rica: it
targets a very narrow population with the explicit goal of improving high school attendance.
Chapter 6 concludes by analyzing the results of the preceding three empirical chapters as
a whole and evaluating their significance for the study of Latin American politics and the study of
comparative social policy.
Social Policy and Presidential Ideology in Latin America

Page 10 of 245

Table 1-3: CCTs in Latin America by Target Population, Size and Cost
Country Program Name Launch Target Population
Recipients
in Millions
(2010)
Recipients/
Population
(2010)
Recipients/
Poor
($2.50)
Recipients/
Poor
($4.00)
Argentina Plan Familias/AUH 2009
Children under 18 from families
without formal workers
11.79 29.0% 47.4% 42.8%
Bolivia Bono Juancito Pinto 2006 Public school students 5.69 57.0% NA NA
Brazil Bolsa Familia 2001 Poor families 52.39 27.0% 55.1% 46.8%
Chile Chile Solidario 2002 Families in extreme poverty 1.30 8.0% 32.7% 29.1%
Colombia Familias en Acción 2001 Families in extreme poverty 11.69 25.0% 53.4% 49.9%
Costa Rica Avancemos 2006
Poor families with high school
students (12 to 25)
0.19 4.0% 23.9% 24.3%
Dominican
Republic
Solidaridad 2005 Poor families with children 2.98 30.0% NA NA
Ecuador
Bono de Desarrollo
Humano
2002
Poor families with children,
elderly or disabled members
6.13 42.0% 64.5% 58.0%
El Salvador Red Solidaria 2005
Families in extreme poverty
living in rural areas
0.57 9.0% NA NA
Guatemala Mi Familia Progresa 2008
Families in extreme poverty
living in vulnerable areas
3.25 23.0% 48.4% 42.5%
Honduras
Programa
Asignación Familiar
1999
Poor families with children and
pregnant women
1.07 14.0% NA NA
Mexico Oportunidades 1997 Families in extreme poverty 0.83 31.0% 53.4% 23.3%
Panama Red Oportunidades 2006 Families in extreme poverty 27.25 24.0% 31.6% 23.3%
Paraguay Tekopora 2005
Families in extreme poverty
living in rural areas
0.36 10.0% 14.4% 11.2%
Peru Juntos 2005 Poor families with children 0.55 9.0% 37.4% 28.5%
Uruguay
Asignaciones
Familiares
2008
Children under 18 from poor
families
2.59 9.0% 84.3% 77.6%
Aggregate 128.63
2
25.96%
3
51.37%
3
39.04%
3

Sources: Gasparini and Cruces (2012, 56) and Stampini and Tornarolli (2012, 10 and 17).
Notes:
1
. The poverty lines are calculated in 2005 U.S. dollars at purchasing power parity.
2.
Aggregate sum.
3
. Average weighted by population.
Social Policy and Presidential Ideology in Latin America
11
Chapter 2.
Social Policy in Industrialized Countries and Latin America:
Past and Present

1. Introduction
Social policy has a direct impact on the lives of most members of a society, particularly
those living in poverty. Government programs targeted on income can push a family above the
poverty line. Public health policies mark the difference between sickness and health, life and death.
Access to and quality of education directly affects an individual’s economic prospects. Pensions
and social insurance protect workers from workplace- and old age-related risks. Social policy can
also be analyzed through the lens of democratic quality. Unless their most basic needs are met,
individuals cannot fully exert their rights as citizens. Surveys show that the majority of people in
Latin America – the region on which this study focuses – equate democracy with the existence of
a minimum income floor. Prominent schools of social policy in Latin America go so far as to assert
that declining support for democracy in the region is caused by the failure of democratic regimes
to successfully address “questions of people’s welfare, from outright poverty to poor quality
education and health services” (Huber and Stephens 2010, 155).
Given its substantive importance on people’s wellbeing, social policy has become one of
the most developed research agendas within comparative politics. Pierson (2000, 791) posits that
“political scientists are interested in the welfare state in part because it so clearly reveals the
significance of policy choices. Variations in public social provision have big effects on social life.”
Amenta (2003, 114) notes that, relative to other research agendas within comparative politics,
research on social policy has been very successful at generating cumulative knowledge. He
Social Policy and Presidential Ideology in Latin America
12
attributes this success to the existence of a “relative agreement on what was to be explained, while
there was disagreement over possible explanations.” This common ground has made possible
lively debates in which scholars from different theoretical and methodological traditions are able
to communicate, engage and challenge one another.
However, the substantial cumulative knowledge that has been generated on this topic is
derived primarily from a small group of industrialized countries. Meanwhile, research on social
policy in developing countries remains in its infancy. In a recent review of this literature, Carnes
and Mares (2007, 868-69) go as far as to assert “that the most exciting research opportunities in
the study of welfare states lie in examining the variation in policies of social protection in
developing economies.”
Studying the developing world offers an opportunity to correct this bias. Beyond helping
to explain the causes and consequences of social policy in poor countries, research on developing
countries can further our understanding of social policy in industrialized countries. Developing
countries provide new cases on which to test and ultimately refine existing theories and allow for
testing the effects of variables that have been excluded from the developed-country literature.
4
By
assessing how presidential ideology influences social policy in Latin America, this study seeks to
contribute to the growing and fast-evolving body of research on social policy in developing
countries and thus help to address the geographic bias in the social policy literature.

2. The Politics of Social Policy in Industrialized Countries
This section discusses five leading explanations of welfare states, social policy and
redistribution in advanced industrialized economies: structural explanations based on the
consequences of industrialization and economic openness; power resource theory which attributes
Social Policy and Presidential Ideology in Latin America
13
welfare state emergence and development is a function of the relative power of leftist parties and
labor unions; the combined role inequality and democracy in shaping demand for income
redistribution; state-centered explanations that emphasize the role of political institutions in
facilitating and obstructing social policy expansion; and the small but growing literature on the
role of business – as part of cross-class coalitions – in shaping social policy. This section concludes
by surveying research on the effects of globalization on advanced welfare states.
2.1. Structural Explanations
Structural explanations attribute the development of the welfare state to economic factors
that cannot be easily altered in the short run. Logic of industrialism arguments attribute welfare
states to modernization and the social and demographic changes brought about by industrial
development. Industrialization creates the need for a more educated workforce and policies that
protect workers in old age. Governments respond to these needs by expanding social policy. In a
pioneering book, Wilensky (1975) concludes that economic development is the single most
important determinant of social spending levels. Collier and Messick (1975) concur with regard to
the adoption of social security, but find that late-adopters enacted policies at much lower levels of
development than pioneers, a sign of policy diffusion. More recent work analyzing longer time
periods finds weak or insignificant support for the logic of industrialism argument (Adsera and
Boix 2002). For Carnes and Mares (2007), the main weaknesses associated with this approach are
its failure to specify the precise mechanisms though which industrialization translates into social
policy and its inability to predict the types of social policies that countries will adopt.
Industrialization cannot explain differences between Anglo-Saxon and Nordic welfare states or
why postwar East Asia focused on increasing human capital for its entire population while Latin
America focused resources on social security for urban workers.
Social Policy and Presidential Ideology in Latin America
14
A second group of structural explanations seek to explain the robust correlation that has
existed in industrialized countries between economic openness and the size and social spending
generosity of the public sector. In the path-breaking article that launched this research agenda,
Cameron (1978) hypothesized that governments in open economies enacted generous social policy
to compensate workers increasingly exposed to international competition. Katzenstein (1985)
framed this relationship as a bargain between workers and the government that made possible the
pursuit of liberal economic policies.
The onset of globalization gave new salience to this puzzle. Garrett (2001) obtained mixed
results when contrasting this “compensation hypothesis” with the “efficiency hypothesis” – the
popular view that the pressures of globalization are unleashing a social policy race to the bottom.
Rodrik (1997, 1998) finds evidence of compensation, but argues that exposure to economic
volatility, not openness, explains demand for larger governments. In an influential paper, Iversen
and Cusack (2000) find that relationship to be spurious and argue instead that de-industrialization
– the shift toward a service-based economy – fueled demand for compensation and, thus, social
spending.
Most of the recent work on the effects of globalization on developing countries finds
support for the efficiency hypothesis (see Chapter 3; Kaufman and Segura Ubiergo 2002; Rudra
2002; Wibbels 2006).
5

2.2. Power Resource Theory
In contrast to the approaches discussed above, the dominant explanation of welfare state
development, power resource theory (PRT), emphasizes the role of politics. Rooted in Marxist
class analysis, this perspective attributes the expansion of and cross-national variation in social
policy to the relative strength of labor unions and social democratic parties (Esping-Andersen
Social Policy and Presidential Ideology in Latin America
15
1985; Korpi 1983; Stephens 1979). Skocpol and Amenta (1986, 140) put it succinctly: “the earlier
and more fully workers become organized into centralized unions and a social democratic party,
and the more consistently over time the social democratic party controls the state, the earlier and
more ‘completely’ a modern welfare state develops.”
Given its initial focus on advanced industrialized countries, PRT takes for granted the
existence of a democratic government, a premise that cannot necessarily be assumed for
developing countries. Though seldom explicitly acknowledged in the literature of advanced
industrialized countries, democracy constitutes a necessary pre-condition for power resource
explanations, In line with this reasoning, democratic countries should, therefore, be expected to
have more developed social policies than non-democratic ones.
Esping-Andersen (1990) argues that, in addition to explaining size, power resources
explain the overall structure of welfare states, what he calls ‘welfare regimes.’ He contends that
universalistic social policies tend to prevail in states where social democrats have been politically
dominant; in contrast, states dominated by right-leaning parties tend to emphasize residualistic and
means-tested policies.
Countries ruled by Christian Democratic parties are associated with generous welfare states
in which benefits are linked to employment status. As will be discussed in section 3, Latin
American countries attempted to emulate this welfare regime. Empirical work by Huber and
Stephens (2001) strongly supports power resource explanation of both the size and structure of
advanced welfare states.
Despite its intuitive appeal and popularity, PRT possesses several key weaknesses. For
starters, it cannot explain why authoritarian Germany and Austria were Europe’s welfare state
pioneers (Hicks and Esping-Andersen 2005, 509). Carnes and Mares (2007) also critique it for
Social Policy and Presidential Ideology in Latin America
16
treating labor and social democratic parties as monolithic entities with single and stable interests.
Indeed, PRT largely ignores the possibility of disagreements within the labor movement and
between unions and social democratic parties. Whereas a union’s objective is to secure improved
conditions for its members, a party’s goal is to earn votes and win elections. These goals can clash,
particularly when the majority of the population is not part of the working class, as in Latin
America.
2.3. Median Voter Explanations
An alternative approach posits that income distribution within a democratic society is a key
factor in shaping demand for redistribution (Meltzer and Richard 1981). The median voter theory
(MVT), which constitutes the backbone of much of the political economy literature, posits that the
more unequal a society, the greater the demand for redistribution. This occurs because the more
unequal the society, the lower pivotal median voter’s income will be relative to the mean income
and, therefore, the more the median voter stands to gain from redistribution through taxes and
transfers. Thus, the model predicts the median voter’s decisive role in the political process will
lead to increased redistribution in unequal societies.
However, there is little empirical support for the model’s predictions. A cursory look at
income distribution and social spending among OECD countries reveals what Lindert (2004) has
termed the Robin Hood paradox: the most egalitarian countries are also the ones that engage in the
most redistribution (Alesina and Glaeser 2004; Bartels 2008; Iversen and Soskice 2006;
Kenworthy and McCall 2008; Kenworthy and Pontusson 2005; Moene and Wallersteim 2003).
Cross-national studies by Perotti (1996) and Bassett et al. (1999) – including both developed and
developing countries find no relationship between inequality and redistribution. Relying on better
data and looking at fewer countries, Milanovic (2000) finds some support for Meltzer-Richard, but
Social Policy and Presidential Ideology in Latin America
17
these results fall apart when pension spending, which does not redistribute income, is excluded.
6

Looking at individual-level demand for redistribution, Lübker (2007) as well as Haggard,
Kaufman and Long (2013) also fail to find support for the model’s predictions.
Cross-national comparisons may represent a misspecification of the MVT, which was only
meant to explain redistribution variation over-time within a given country. Within-country
analyses by Kenworthy and Pontusson (2005) find that OECD countries increased redistribution
in response to rising inequality, but not through the mechanisms outlined by Meltzer-Richard.
Rather, redistribution increased through automatic transfers and, to a smaller extent, through
compensatory taxes. In contrast, Kenworthy and McCall (2008) find little within-country support
for the Meltzer-Richard model among OECD countries.
In a ground-breaking recent paper, Lupu and Pontusson (2011) argue that it is not the level,
but rather the structure of inequality that drives the demand for redistribution. They cite evidence
among OECD countries that middle class voters seek redistribution when the distance between
their share of national income and that of the poor is small relative to the distance between their
share of the national income and that of the rich. Thus, voters near the median income align with
the poor in support of social spending when the income distribution is skewed towards the rich.
Chapter 3 will test to determine if this finding “travels” to Latin America.
In its quest for parsimony, the Meltzer-Richard model leaves out several factors that have
been shown to influence inequality. Electoral turnout varies greatly across countries and
systematically within countries in relation to income and education. Kenworthy and Pontusson
(2005) provocatively argue that voter turnout may be the missing link that connects PRT with the
MVT and could thus account for the Robin Hood paradox. If lower income citizens, the Left’s
natural constituency, do not vote, their preferences will not be taken into account and less
Social Policy and Presidential Ideology in Latin America
18
redistribution will take place. Mahler (2008) finds turnout is positively associated with government
redistribution and that, once turnout is controlled for, inequality is positively associated with
redistribution.
People do not exist in a cultural vacuum: there are important cross-national differences in
perceptions of what constitute acceptable levels of inequality and the government’s role in
addressing inequality (Alesina and Glaeser 2004; Ardanaz 2009; Lübker 2007). Alesina and
Glaeser (2004) find that, compared to Americans, Europeans are about twice as likely to believe
the poor are trapped in poverty and that luck determines income; they also find that European are
approximately half as likely as Americans to believe the poor are lazy. Ardanaz (2009) shows that
Latin Americans possess views similar to those of Europeans. Haggard, Kaufman and Long (2013)
find that demand for redistribution in developing countries varies greatly depending on an
individuals’ occupation and whether they live in an urban area.
2.4. State-Centered Explanations
The Meltzer-Richard model also takes for granted the existence of majoritarian electoral
institutions. There is substantial evidence that democracies with proportional representation (PR)
spend and redistribute more than majoritarian ones (Alesina and Glaeser 2004; Austen-Smith
2000; Persson and Tabellini 2003).
7
In line with power resources explanations, Iversen and Soskice
(2006) find that left governments redistribute more. However, they also find that partisan
dominance is a function of the electoral system with PR favoring the left and majoritarian systems
favoring the right. Similarly, there is strong evidence that an increased number of veto players is
associated with having a smaller and less redistributive welfare state (Alesina and Glaeser 2004;
Huber and Stephens 2001). In contrast, research on Latin America finds a positive effect of veto
players on education and social security spending (Huber and Stephens 2012). Alesina and Glaeser
Social Policy and Presidential Ideology in Latin America
19
(2004) attribute much of the difference in welfare state development between Europe and the
United States to majoritarianism, federalism and checks and balances, institutions purposely
established by the Founding Fathers of the United States to prevent the expropriation of wealth.
2.5. Business-Centered/Cross-Class Coalitions
Carnes and Mares (2007) are particularly critical of PRT for assuming a zero-sum game
between labor and capital. A new and growing strand of the literature challenges these assumptions
by closely examining the role of firms in promoting and shaping social policy (Estevez-Abe,
Iversen and Soskice 2001; Hacker and Pierson 2002; Mares 2003; Iversen 2005; Swenson 1991,
2002). “Bringing capital back in” does not imply that employers were the sole architects of social
policy, but rather that they were key members in cross-class alliances that led to the adoption of
social policy. Past research has found that a firm’s characteristics shape its social policy
preferences. Estevez-Abe, Iversen and Soskice (2001) argue that firms dedicated to gradual
innovation, such as those in coordinated market economies, require workers with highly specific
and thus non-transferrable skills. The acquisition of such skills constitutes a major risk for workers,
one that they will only assume if there is strong employment protection and generous
unemployment insurance. Thus, firms in sectors with gradual innovation will support social
policies that make workers willing to invest in acquiring non-transferable skills. Mares (2003)
finds that a firm’s size, skill intensity and incidence of risk define the type of social insurance it
will support. For Swenson (2002), a firm’s preferences toward social policy are a function of the
relative scarcity of labor and the level of domestic competition it faces.
In a critique of this literature, Korpi (2006) argues that these explanations confuse business
consenting to social policy with it playing a proactive role in promoting it. He argues that PRT can
only be challenged by evidence that employer involvement in drafting social policy went beyond
Social Policy and Presidential Ideology in Latin America
20
responding to and modifying proposals put on the table by other groups. Mares (2009) argues that
this literature needs to better specify the channels through which employers influence social policy
and more systematically analyze the effects of political variables such as partisanship on
bargaining and the formation of cross-class alliances. So far, this approach has been used to explain
social policy in only a handful of industrialized countries. Developing countries constitute an
important avenue for future research on this topic.
2.6. Globalization and the Welfare State
A major debate within the advanced-country welfare state literature has centered on the
effects of globalization, increased unemployment and weakening labor unions on existing welfare
states. Pierson (1994; 1996) insightfully argues against historical symmetry: that the presence of a
given factor led to an outcome does not necessarily mean that its disappearance will lead to a return
to the prior status quo. Specifically, that the strength of social democratic parties and unions was
decisive in the formation of welfare states does not mean that their weakening post-1980 will lead
to welfare state retrenchment. Looking at the hard-cases of the U.S. under Ronald Reagan and the
U.K. under Margaret Thatcher, he finds only limited evidence of retrenchment. He argues that the
expansion of social policies created organized constituencies in favor of those policies, for
example pensioners associations, that proved willing to mobilize to defend their benefits.
Politicians of all ideological stripes have been unwilling to challenge such powerful interest
groups. As a result, radical retrenchment failed to take place.
Looking at spending levels, Huber and Stephens (2001) do find some evidence of
retrenchment post-1980, particularly among countries with few veto players governed by
conservatives. However, they attribute this retrenchment not to globalization but rather to the
difficulty of financing entitlements under a context of high unemployment. Similarly, Korpi and
Social Policy and Presidential Ideology in Latin America
21
Palme (2003) and Allan and Scruggs (2004), independently find evidence of considerable
retrenchment and party influence. These two works are notable in that rather than looking at
spending, they assess social policy replacement rates – the percent of a recipient’s income replaced
by a program – adjusted for ease of access and duration of benefits. Regrettably, such detailed
social policy measures are not available for developing countries.
Pierson (1994, 1996) argues that, in the current era, the ideological differences between
political parties have little effect on social policy outcomes. Similarly, Huber and Stephens (2001)
find a significant reduction in the influence of partisan ideology over social policy since 1980.
Other scholars have found strong evidence that traditional power resource factors helped countries
stave off retrenchment. Garrett (1998) finds that countries with social democratic corporatism,
defined as having dominant left parties as well as broad and encompassing labor unions, have
responded to globalization by enlarging the size of governments without sacrificing growth and
employment. Swank (2002) finds that the welfare state weathered the effects of increased capital
mobility in countries with proportional representation, corporatist interest representation,
centralized political authority, and universal social insurance.
Despite these disagreements, scholars of advanced welfare states tend to agree that welfare
states are sticky, slow-moving and path dependent variables that are unlikely to exhibit dramatic
changes over the course of a few years. In sharp contrast, the social policy arrangements in
developing countries have proven to be much more fluid, changing dramatically during the current
era of rapid globalization. As will be discussed in the following section, the very nature of Latin
American social policy changed following the 1980s Debt Crisis and the subsequent wave of
market reforms that swept the region during the 1990s (Huber and Stephens 2005).

Social Policy and Presidential Ideology in Latin America
22
3. Social Policy in Latin America
Latin America has been a social policy pioneer among developing regions. While there
continue to be dramatic cross-national differences with regard to the level and composition of
social spending, and the effects social policy on levels of poverty and inequality, it is possible to
speak of a “Latin American Model” of social policy (Haggard and Kaufman 2008). This model
has been traditionally centered on dependable and, in some cases, generous contributory social
insurance for formal-sector workers and, at best, minimal safety nets for urban informal workers
and the rural poor. The latter two have tended to constitute the majority of the region’s population.
Applied to Latin America, a region with high levels of informality, this approach to social
policy leads to very limited income redistribution (Breceda, Rigolini and Saavedra 2009; Goñi,
López and Servén 2011; Lustig et al. 2012). In fact, the region’s social insurance policies have
been described as “reverse Robin Hood” because they may, in some instances, redistribute wealth
from the poor to the rich (Lindert, Skoufias and Shapiro 2006). Benefits paid out by the
government regularly exceed contributions. As a result, every citizen, even those excluded from
social insurance, by virtue of paying taxes, contributes to funding social insurance (Ferreira and
Robalino 2011).
Beyond social insurance, the other components of the region’s welfare regime have also
failed to target the neediest. Availability of healthcare varies greatly depending on geographic
region and employment status often determines service quality. Furthermore, primary education
has generally been of low quality and exhibited a serious urban bias. In addition, education
spending has tended to be disproportionately biased in favor of universities to the detriment of
basic education. With regard to labor market regulation, governments attempted to compensate for
the absence of state-run unemployment insurance through labor codes that offered extensive
Social Policy and Presidential Ideology in Latin America
23
protection for formal workers. The high cost of firing non-productive and/or unneeded employees
disincentivized job creation, further accentuating the region’s labor market dualism (Haggard and
Kaufman 2008, 5).The Latin American model is in essence Esping-Andersen’s (1990)
conservative/Christian-Democratic welfare regime with the crucial caveat that most of the
population is excluded. As a result, Barrientos (2004; 2009) classifies the region’s welfare regimes
prior to the 1990s as conservative/informal welfare regimes.
8
It is conservative in that social
insurance is linked to the employment of an assumed “male breadwinner” in much the same way
as in the welfare regimes of Southern Europe (Esping-Andersen 1990; Huber 1996). However, it
is informal in that it excludes the substantial portion of the population employed in the informal
sector. This approach to social policy was adopted under the assumption that as the region’s
countries modernized, agricultural and informal workers would be absorbed into industry and,
therefore qualify for the full range of contributory social security programs (Lo Vuolo 2008).
However, the import substitution industrialization (ISI) policies of the 1930s-1970s failed to create
the number of jobs needed to make social policy coverage universal. And, with the model’s
abandonment in the aftermath of the 1980s Debt Crisis, the number of excluded workers
substantially increased.
The region’s governments increased their focus on targeted social assistance in response
to rising poverty and inequality in the aftermath of the debt crisis and market reform. And, starting
in the late 1990s, they began launching large-scale non-contributory pension schemes and targeted
anti-poverty programs such as conditional cash transfer programs, which have been successful at
reducing poverty (Stampini and Tornarolli 2012) and inequality (López-Calva and Lustig 2010).
Spending on traditional contributory social insurance, however, continues to dwarf spending on
these newer, more progressive and far-less generous non-contributory policies (Lindert, Skoufias
Social Policy and Presidential Ideology in Latin America
24
and Shapiro 2006). Thus, at the start of the 21
st
century Latin America possesses a bifurcated social
policy.
3.1. Explaining the “Latin American Model”
A small body of literature has sought to explain the adoption of social policy in the region
during the early to mid-20th century. The first generation of such research focused on the most
salient feature of Latin American welfare regimes: social security schemes for workers in the
formal sector. In a landmark study, Mesa-Lago (1978) finds that the gradual expansion of social
security reflected the power of organized groups, most notably the military, bureaucrats and
organized labor (teachers, workers in charge of key infrastructure, mining, and manufacturing),
which succeeded in extracting benefits from governments; and, reciprocally, the governments
themselves, which used benefits to co-opt the interest groups mentioned above.
9
In summation,
the piecemeal expansion of social policy helps explain the regressive welfare regimes that persist
to this day.
More recent research has sought to apply PRT. Implicitly, these works aim to explain why
Latin America’s welfare regimes are less robust than Europe’s. As Huber and Stephens (2005,
555) assert, “compared to Europe, then, the much lower size and generosity of Latin American
systems of social protection can be attributed to a combination of lower levels of industrialization,
the scarcity of democratic periods, and the weaker position of reformist parties and organized
labor.”
Relying on a wide range of methods, scholars have agreed on the importance of democracy
and organized labor. Evidence of an effect of left-leaning parties is more tentative. This
triangulation of results strengthens these authors’ claims and helps to demonstrate that power
resources “travels” outside Western Europe. Comparative case studies on the evolution of social
Social Policy and Presidential Ideology in Latin America
25
policy in nine countries in the region during 1945-1980 by Haggard and Kaufman (2008) reveal
that democracies and, to a lesser extent, semi-democracies and competitive authoritarian regimes,
expanded coverage in social security, healthcare and, to a lesser degree, education. Organized labor
was also key to expanding social policy. However, because Latin American unions represented
only a small segment of the population, they pushed social security policies that, as discussed
above, were regressive.
Similarly, cross-national quantitative research on social spending by Huber and Stephens
(2012) reveals that the size of the urban working class and the number of years a country had been
democratic since 1945 are the strongest determinants of social welfare spending in 1980. Having
had a left government was also conducive to increased social spending. However, there were few
left governments during the second half of the 20
th
century and these tended to be followed by
repressive authoritarian regimes that rolled back reforms.
10
Segura-Ubiergo (2007) reaches similar
conclusions through qualitative comparative analysis, finding that the development of welfare
states in the region was contingent on possessing two out of the following three: favorable
economic conditions (defined as relatively high development and a closed economy), strong left-
leaning unions, and a consistently democratic regime.
11

Thus, PRT constitutes the dominant explanation for the emergence of Latin American
welfare regimes. However, there are important differences in how power resources played out in
Europe versus Latin America. Carnes and Mares (2007) insightfully note that a major weakness
of the power resources approach is that it treats labor unions and social democratic parties
interchangeably, when in fact they have different goals. In their quest for power, parties competing
in presidential systems will propose policies that are likely to be supported by a majority of the
population. Unions, on the other hand, will be content with policies that benefit their members
Social Policy and Presidential Ideology in Latin America
26
even if those exclude other groups.
12
This is precisely what happened in Latin America, where
unions were strong but social democratic parties were weak or non-existent.
13
Rather than leading
to universalistic policies, the mix of strong unions and a weak left produced truncated welfare
regimes.
Huber (1996) cites Costa Rica as the only example of a social democratic welfare regime
in Latin America. It is no coincidence that a social democratic party, Partido Liberación Nacional,
dominated the country’s politics from 1950 until the debt crisis. Uruguay, which was not governed
by the center-left until recently, comes closest in the region to following the Costa Rican example.
By regional standards, Uruguay has had remarkably low levels of informality (Huber and Stephens
2010) and high levels of industrial employment (Huber and Stephens 2005; World Bank 1982),
which helped it achieve the region’s highest levels of social security coverage pre-1980 (Mesa-
Lago 1994, 22). Under such circumstance, the Latin American model can provide near-universal
coverage.
Moving away from PRT, an innovative paper by Wibbels and Ahlquist (2011) offers a
cross-class coalition explanation of Latin American social policy. While scholars have long posited
the existence of a link between ISI and the Latin American model of social policy (Haggard and
Kaufman 2008; Huber 1996; Huber and Stephens 2005; Pribble 2011; Segura-Ubiergo 2007),
these authors are the first to explicitly theorize and test such a mechanism.
14
They argue that, in
larger countries, the combination of ISI and a relatively scarce labor supply explains the adoption
of social security-centered welfare regimes.
15
Scarcity allowed labor to overcome collective action
problems and organize to obtain social protection. Employers could afford social insurance
because trade barriers allowed them to pass costs onto consumers. As an added benefit for
Social Policy and Presidential Ideology in Latin America
27
employers, by increasing the incomes of workers, social policy boosted demand for consumer
goods. They find that democracy has had no effect on social spending.
3.2. Determinants of Social Spending
A growing body of work relies on regression analysis to assess the recent determinants of
social spending. This line of research has focused on two types of factors: political variables and
globalization. A smaller subset tests the effect of spending on social outcomes such as poverty
(Pribble, Huber and Stephens 2009) and inequality (Huber at al. 2006; Rudra 2004; Yi 2013).
While initial research focused on total social spending, more recent research sacrifices sample size
for more disaggregated data on social security and human capital expenditures.
16
Tables 2.1-2.3
summarize the main findings from this literature.
Political factors have tangible effects on social spending levels in Latin America and the
developing world more broadly. Haggard and Kaufman (2008) posit that democracy should have
a positive effect on social policy through two channels: electoral competition, which incentivizes
politicians to offer popular social programs, and greater interest group freedom, which allows such
groups to directly lobby and, if necessary, protest for social programs. However, in contrast to
most research on the topic, they find no effect of democracy either on total social spending or on
human capital spending in Latin America during1980-2000.
17

Social Policy and Presidential Ideology in Latin America
28
Table 2-1: Social Spending Literature: Effects of Politics
Hypothesis Positive Negative Other

Democracy on
social spending
Brown & Hunter (1999),
Rudra (2002) on total
spending.

Avelino, Brown & Hunter
(2005); Kaufman & Segura-
Ubiergo (2001); Segura-
Ubiergo (2007); Stasavage
(2005); Zarate Tenorio
(Forthcoming) finds positive
effect on human capital
spending.

Huber, Mustillo and Stephens
(2008) find cumulative years
of democracy have positive
effect on human capital and
social security spending.

Brown & Hunter (2004);
Stasavage (2005) find it leads
to basic education being
prioritized. Brown (1999);
Eterovic & Sweet (2014) find
positive effect on enrollment;
Altman & Castiglioni (2009);
Gerring, Thacker & Alfaro
(2012) find positive effect on
human development; quality
of democracy improves
infant mortality and life
expectancy
Kaufman & Segura-Ubiergo
(2001) find a negative effect
on social security. Segura-
Ubiergo (2007) finds same,
but less consistent.
Wibbels (2006) finds
inconsistent effects. Kaufman
& Segura-Ubiergo (2001);
Segura-Ubiergo (2007) find
no effect on total spending.

Brown & Hunter (1999);
Haggard & Rudra (2005) find
democracies less likely to cut
social spending during crises.

Pribble, Huber & Stephens
(2009) find negative effect on
poverty. Huber et al. (2006);
Rudra (2004) find social
security spending has a
positive effect on inequality.
.
Government
ideology on
social spending
Kaufman & Segura-Ubiergo
(2001) find popularly-based
governments have positive
effect on social security.
Segura-Ubiergo (2007) finds
less consistent, but positive,
results.
Kaufman & Segura-Ubiergo
(2001); Segura-Ubiergo
(2007) find a generally
negative effect of popularly-
based governments on human
capital.
Huber, Mustillo & Stephens
(2008) find no effect of
executive or legislative
partisanship. Kaufman &
Segura-Ubiergo (2001);
Segura-Ubiergo (2007) find
no effect of popularly-based
governments on total
spending.
.
Labor power on
social spending
Niedzweicki (2010); Rudra
(2002); Wibbels (2006) find
positive effect of potential
labor power.
Wibbels (2006) finds a
negative effect of union
density.
------------
.
Political protests
on social
spending
Zarate Tenorio
(Forthcoming) finds a
positive effect of strikes on
social security spending.
------------ Zarate Tenorio
(Forthcoming) finds no effect
of strikes/protests on human
capital spending. In
democracy protests decrease
likelihood of cuts.

Social Policy and Presidential Ideology in Latin America
29
Table 2-2: Determinants of Social Spending: Effects of Globalization
Hypothesis Positive Negative Other

Openness to trade
on total spending
Avelino, Brown & Hunter
(2005) using a measure based
on purchasing power parity.
Kaufman & Segura-Ubiergo
(2001); Segura-Ubiergo
(2007); Wibbels (2006).
Huber, Mustillo & Stephens
(2006) find no effect.

Openness to trade
on human capital
spending
Avelino, Brown & Hunter;
(2005).
------------ Huber, Mustillo & Stephens
(2006); Kaufman & Segura-
Ubiergo (2001); Segura-
Ubiergo (2007) find no
effect.

Openness to trade
on social security
spending
Avelino, Brown & Hunter
(2005).
Kaufman & Segura-Ubiergo
(2001); Segura-Ubiergo
(2007).

Rudra (2002) finds positive
effect but a negative effect of
openness interacted with
PLP.
Huber, Mustillo & Stephens
(2006); Huber & Stephens
(2012) find no effect.

Openness to
capital on
spending
Wibbels (2006), but effect
is inconsistent.
Kaufman & Segura-Ubiergo
(2001) only when interacted
with trade openness; Segura-
Ubiergo (2007) under high
fiscal deficits.

Avelino, Brown & Hunter
(2005) find no effect.

Rudra (2002) finds positive
effect but a negative effect of
openness interacted with
PLP.

Foreign direct
investment on
spending
Niedzweicki (2010) finds a
positive effect on social
security and human capital
spending.
Zarate Tenorio (Fortcoming)
finds a negative effect on
social security spending.
Huber, Mustillo & Stephens
(2008) find no effect.

Economic crisis
on spending
------------ Wibbels (2006) finds a
negative effect on human
capital spending.
Wibbels (2006) finds no
effect on social security
spending.

Social Policy and Presidential Ideology in Latin America
30
Table 2-3: Effects of Social Spending on Poverty and Inequality
Hypothesis Positive Negative Other

Social spending
on inequality
Rudra (2004) finds negative
social security spending
under openness has a positive
effect.
Huber et al. (2006) find
social security spending
under democracy has a
negative effect.

Rudra (2004) finds negative
education spending under
openness has a negative
effect. Healthcare spending
under openness is marginally
negative.
Huber et al. (2006) find
human capital spending has
no effect.

Social spending
on poverty
------------ Pribble, Huber & Stephens
(2009) find social security
has a negative effect. Huber
& Stephens (2012) find
health spending has a
negative effect
Pribble, Huber & Stephens
(2009) find human capital
spending has no effect.
Huber & Stephens (2012)
find education spending has
no effect.

Other research has generated substantial empirical evidence that democratic regime has a
positive effect on total social spending (Avelino, Brown and Hunter 2005), as well as on social
security (Rudra 2002) and human capital spending (Avelino, Brown and Hunter 2005; Brown and
Hunter 2004; Kaufman and Segura-Ubiergo 2001; Wibbels 2006; Zarate Tenorio Forthcoming).
Similarly, cumulative years of democracy since 1945 has been associated with increased social
spending (Huber, Mustillo and Stephens 2008). Furthermore democracies tend to place greater
emphasis on preschool and primary education relative to other levels of education (Brown and
Hunter 2004). And there is some evidence that democracies spend more on human capital relative
to social security (Kaufman and Segura-Ubiergo 2001). Finally, in contrast with authoritarian
regimes, democracies appear to increase social spending during economic crises (Brown and
Hunter 1999; Rudra and Haggard 2005).
Democracy is also associated with a wide array of desirable policy outcomes: it is
negatively associated with poverty rates (Pribble, Huber and Stephens 2009) and inequality among
Latin American countries (Huber et al. 2006) and developing countries more broadly (Rudra
Social Policy and Presidential Ideology in Latin America
31
2004).
18
Worldwide, being democratic during a given year is robustly correlated with a variety of
positive education and public health outcomes (Altman and Castiglioni 2009; Lake and Baum
2001) as well as higher average daily caloric consumption (Blaydes and Kayser 2011). Cumulative
years (Gerring, Thacker and Alfaro 2012; McGuire 2010) and quality (Altman and Castiglioni
2009) of democracy are associated with reduced infant mortality. Furthermore, through its effect
on human capital, democracy has a positive effect on economic growth (Baum and Lake 2003).
A large body of research has sought to untangle the relationship between left-leaning
parties and social spending.
19
Kaufman and Segura-Ubiergo (2001) find that popularly-based
governments, defined as those from programmatic parties and/or closely linked to unions, tended
to protect social security transfers while cutting more progressive human capital spending. Huber
and Stephens (2012) find no relationship between executive or legislative partisanship and social
spending.
20
They do, however, find that partisan balance has a negative effect on both poverty and
inequality. These findings lead them to conclude that the region’s left-leaning governments, while
unable to increase social spending, have affected its composition and made it more progressive.
Case study evidence from the region’s most developed welfare regimes supports their claims.
While the effects of electoral competition and ideology on social spending have been
studied in great detail, less attention has been paid to the role of protests. In the only paper to date
to assess this, Zarate Tenorio (Forthcoming) finds that labor strikes, which are associated with
organized labor, positively affect social security spending. In contrast, anti-government
demonstrations and riots, which are associated with labor market outsiders, have no effect on
human capital spending. However, in conjunction with democracy, they make cutbacks to human
capital spending less likely.
Social Policy and Presidential Ideology in Latin America
32
Research on social spending confirms the ineffectiveness and even regressiveness of Latin
American social policy. Huber et al. (2006) find evidence that social security spending does reduce
inequality, though only under democracy. Looking at developing countries more broadly, Rudra
(2004) finds that only education spending- and not health or social security spending- reduces
inequality. Pribble, Huber and Stephens (2009) find no effect of health and education spending on
poverty in Latin America. They find inconsistent evidence that social security spending has a
negative effect on poverty, though only when they fail to control for political variables.
Research aimed at assessing the applicability of the efficiency and compensation
hypotheses (Garrett 2001) to Latin America and the developing world has found robust evidence
of a negative relationship between openness to trade and social policy (Kaufman and Segura-
Ubiergo 2001; Wibbels 2006).
21
The one exception is Avelino, Brown and Hunter (2005) use an
alternate measure of trade openness based on purchasing power parity, find a positive relationship.
There is evidence that globalization’s negative effect occurs primarily through cuts in social
security (Kaufman and Segura-Ubiergo 2001). This is attributed to the fact that, unlike human
capital investment, which is funded through regular taxation and delivers productivity gains over
the long-term, social security is funded through payroll taxes that directly affect firm
competitiveness.
3.2.1. What We Still Don’t Know About Spending
While the literature on Latin American social policy has advanced greatly in recent years,
there is still a great deal that is unknown. In this regard, further research is needed, on the impact
of social spending of income inequality, levels of taxation, ethnic diversity, and the region’s left
turn on social spending. These factors will be central to the analysis of the determinants of social
spending levels that will be conducted in chapter 3.
Social Policy and Presidential Ideology in Latin America
33
Research testing the effects on social spending of presidential ideology, the key variable in
PRT, has so far been inconclusive. Meanwhile, countries throughout the region have in recent
years elected left-leaning presidents who campaigned on promises of increasing social spending
and redistributing income. This “left turn” is arguably the most important political development in
the region over the past decade. Recent research on this region wide phenomena has distinguished
between a moderate social-democratic left and a radical populist one (Castañeda 2006; Weyland,
Madrid and Hunter 2010),
22
but there has been scant research contrasting the actual social policies
of these “two lefts.” The only published research on this topic, a preliminary analysis at the end of
a paper on inequality by Birdsall, Lustig and McLeod (2012), concludes that social democrats have
outspent the populists.
23
Further research is needed to specifically ascertain whether: radical left
governments are more generous than moderate ones and/or whether left governments have a
greater likelihood of emphasizing more redistributive expenditures in education and healthcare
over more regressive social security expenditures. Chapter 3 addresses these questions.
Income inequality and level of taxation are two key economic variables that have so far
been completely absent from research on social policy in Latin America. As discussed, MVT is
one of the most influential explanations of income redistribution; yet, to date, there are no
empirical tests of its applicability to the region. Do the more unequal countries in Latin America,
in contrast to the experience of advanced capitalist countries, actually spend more on social policy
than more egalitarian countries? Additionally, does the structure of inequality (Lupu and
Pontusson 2011) affect social spending in Latin America?
Timmons (2005) argues that, in ignoring taxation, past work on social expenditure has
missed a crucial part of the story. He finds robust evidence among a sample of 90 countries that
income taxes, which fall primarily on the rich, lead to the types of policies desired by the rich,
Social Policy and Presidential Ideology in Latin America
34
namely secure property rights. Consumption taxes, which fall primarily on the poor, on the other
hand, lead to policies desired by the poor, namely increased social spending. Applied to Latin
America, the region’s shift toward greater reliance on consumption taxes during the 1980s and
1990s should have translated into increased social spending (Mahon 2012; Tanzi 2000). By
looking only at Latin American countries, Timmons’ findings can be re-tested using disaggregated
social spending data. Do consumption taxes have different effects on the various forms of social
spending? More specifically, do consumption taxes increase spending on education and healthcare
relative to more regressive social security spending?
Relevant to the question of how ethnic diversity and the presence of large indigenous
populations affect spending levels social spending, a study by Luttmer (2001) and Alesina and
Glaeser (2004) has demonstrated that ethnic diversity has been an obstacle to redistributive social
policy. Their conclusions are consistent with the empirical evidence showing that Latin American
countries with large indigenous populations (Guatemala and the Andean countries) have among
the least developed social policies in the region, while those countries with the most developed
welfare states tend to be quite ethnically homogenous (the Southern Cone and Costa Rica).
Similarly, Ardanaz (2009) finds a negative bivariate correlation between ethnic diversity and
support for redistribution in the region. Among Latin American countries, there is evidence of
ethnic diversity having a negative effect on healthcare spending (Huber and Stephens 2012) and a
positive impact on poverty (Pribble, Huber and Stephens 2009) and inequality (Huber et al. 2006).
This debate is far from resolved and there is room for more research on the topic.
3.2.2. Critiques of Spending Research
Research on social spending in Latin America has several important weaknesses that need
to be taken into account. The most obvious problem is the so-called “dependent variable problem”
Social Policy and Presidential Ideology in Latin America
35
(Green-Pedersen 2008), which stems from the use of social spending as an outcome variable.
24
As
Esping-Andersen (1990, 19) famously remarked: “expenditures are epiphenomenal to the
theoretical substance of welfare states.” Expenditure data says little about who social programs
target or how effective these policies actually are. At the same time, even the most disaggregated
data used in cross-national research remains highly aggregated. For example, in most Latin
American countries, the rubric of social security spending includes programs such as conditional
cash transfers that benefit labor-market outsiders, thus reduce poverty and inequality. However,
such spending also includes programs such as contributive pensions, which benefit insiders and
are quite regressive. The same can be said about education budgets, which include progressive
spending on primary education and regressive spending on universities. While more disaggregated
data are available from governments, these are not comparable across countries. Furthermore,
spending data start in the 1970s, greatly complicating quantitative work on the origins and early
development of welfare states.
25
Garay (2012) notes that national-level data also tend to
underestimate spending by sub-national governments, which, particularly since the 1980s, is
responsible for an increasingly important portion of the education and healthcare spending in
federal countries. Given these limitations, in addition to analyzing the determinants of social
spending, chapters 4 and 5 will focus on explaining the adoption and cross-national variation in
the design of a specific type of policy across the region: conditional cash transfer anti-poverty
programs.
3.3. Latin America’s Evolving Welfare States
Beginning in the 1990s, the region’s welfare regimes began to undergo a substantial
transformation. In contrast to Europe, where, as discussed above, welfare regimes proved resistant
to change, Latin America’s welfare regimes underwent major transformations. The most notable
Social Policy and Presidential Ideology in Latin America
36
reforms have been in the areas of pensions and poverty relief. Following the Chilean example,
several countries replaced pay-as-you-go social security schemes with individual and often times
private retirement accounts (Brooks 2002; 2005; 2007; 2008; Madrid 2005; Mesa-Lago 2007;
Weyland 2005; 2006). However, as the private sector was beginning to play a greater role in the
administration of pensions for the middle class and the rich, governments throughout the region
launched basic non-contributory pensions for the elderly (Barrientos and Lloyd-Sherlock 2002;
Carnes and Mares 2014; FIAFP 2011; Sojo 2006). On the poverty front, most of the region’s
governments have adopted innovative conditional cash transfer (CCT) anti-poverty programs such
as Brazil’s Bolsa Escola/Bolsa Familia and Mexico’s Progresa/Oportunidades. These means-tested
programs award poor families regular stipends on the condition that they keep their children in
school and subject them to regular medical checkups. In the short-term, CCTs seek to relieve
poverty by supplementing the regular incomes of needy families. In the long-run, they seek to
build human capital with the goal of improving the future employment prospects of beneficiaries
and ultimately preventing the inter-generational transmission of poverty.
These changes were a direct consequence of the 1980s Debt Crisis and subsequent wave
of market reform that swept the region during the 1990s. The crisis brought about sharp reductions
in all forms of social spending as well as the deterioration of public services at a time when poverty
and inequality rose sharply (Huber 1996; Huber and Stephens 2005). The combination of slow
growth, large deficits and the inability of governments to borrow from abroad challenged the
financial sustainability of the region’s social insurance systems. At the same time, the number of
union members and formally employed workers decreased throughout the region. Through the
downsizing of public employees and privatization of state-owned enterprises, liberalization
weakened public-sector labor unions. Trade liberalization and the dismantling of ISI exposed many
Social Policy and Presidential Ideology in Latin America
37
of the region’s previously protected firms to international competition. A significant number of
these firms failed and those that survived became leaner and more efficient either by switching
workers to more flexible arrangements with reduced benefits or by altogether eliminating their
positions. Many of these downsized workers were cast-off into the informal sector and thus lost
access to social security (Huber 2005; 2009). The expansion of informality further heightened the
dual nature of the region’s welfare regimes and made the differences between labor market insiders
and outsiders politically unsustainable.
It would be misguided to dismiss the social policy innovations cited above simply as social
policy retrenchment or the withdrawal of the state from social policy. While representing a
fundamental shift in the logic of the region’s welfare regimes, these reforms have not implied an
across-the-board reduction in social spending but, rather, a rethinking and re-prioritization of the
state’s role. The move toward greater private-sector involvement in the pension system implies a
shift in the state’s role from direct provider to regulator and facilitator. Basic pensions and CCTs
actually constitute an increase in the state’s involvement in combating poverty. Borrowing and
adapting Esping-Andersen’s (1990) seminal typology, Barrientos (2004, 2009) provocatively
argues that these reforms represent a shift in the nature of the region’s welfare regime from a
conservative/informal regime toward a variant of the liberal welfare regimes found in Anglo-
Saxon countries. Through labor market liberalization, individual savings accounts and this new
generation of anti-poverty programs, the market’s role in social policy provision has increased
while the state’s role as a direct provider has been refocused on the poorest members of society.
Still, the magnitude of the transformation should not be overstated. Contributory social
insurance programs continue to constitute the bulk of social spending among Latin American
governments. Largely regressive contributory social security policies still make up a third of what
Social Policy and Presidential Ideology in Latin America
38
the region spends each year on social policy. In the Southern Cone, in particular, social security
expenditures are higher than education and healthcare expenditures combined. And, as noted
earlier, the region’s contributory social insurance programs run deficits covered by all tax payers
including those excluded from their benefits. Thus, although Latin American social policy may be
undergoing, as Ferreira and Robalino (2011) claim, a “quiet revolution,” new and old policies
coexist.
From a normative perspective, this transformation is positive as scarce government
resources are now destined towards those who most need them. However, this can also be
interpreted as a move away from the aspiration of developing full-fledged universalistic welfare
states. The Latin American model was not intended to be exclusionary. The region was expected
to grow its way to full-employment and universal social insurance. These new policies are based
on the acceptance of informality as a permanent feature of Latin America’s economies. In a
permanent context of austerity, the region’s governments will focus increasingly on the neediest,
leaving the middle class and the well-off to rely on the market.
An overarching goal of this study is to develop and test a theory that explains the rapid
diffusion and cross-national variation in the design of CCT programs in the region. In the following
section, I go into depth about what we currently know about this new type of anti-poverty
programs.

4. Conditional Cash Transfer Programs
Conditional cash transfer (CCT) programs aim to simultaneously reduce poverty and
increase human capital by awarding families regular monetary stipends conditional on their
children remaining in school and, in some cases, attending regular medical checkups. In the
Social Policy and Presidential Ideology in Latin America
39
immediate term, transfers relieve poverty by boosting the incomes of the poor and economically
vulnerable. In the long-term, conditionality seeks to increase the human capital of beneficiary
children, which in turn is expected to help them achieve higher future earnings and thus prevent
the inter-generational transmission of poverty.
Proposed and designed by Latin American intellectuals and first adopted at the local level
in Brazil and at the national level in Mexico in the late 1990s, CCTs constitute a uniquely regional
policy innovation. Since then, these types programs have spread rapidly across the region. Whereas
in 1998 only Brazil and Mexico had CCTs, by 2008 nearly every country in the region had adopted
a CCT program.
26
By 2010, approximately 130 million people – nearly a quarter of Latin
America’s population – benefitted from a CCT (Stampini and Tornarolli 2012) (see figures 2-1
and 2-2). As result of their success at reducing poverty in Latin America, similar policies have also
been adopted in Africa, Asia and Eastern Europe, as well as in New York City (Fiszbein and
Schady 2009; Hanlon, Barrientos and Hulme 2010).
Fiszbein and Schady (2009) define CCTs as having five main characteristics. First, they
are targeted toward a specific population, generally families with school-age children. Second,
they have simple benefit structures and eligibility criteria; the latter generally includes being poor
and having children enrolled in school. Third, benefits are conditional on certain requirements,
most notably school attendance, although some programs also have medical and nutritional
requirements. Fourth, implementation requires coordination across different areas (education,
health and finance) and levels (national and local) of government. Fifth, they incorporate an
explicit monitoring and evaluation framework to assess their effectiveness. All of the region’s
CCTs meet the first four criteria; however, the fifth criteria, which was one of the most innovative
aspects of early CCTs, is largely absent in some of the more recent programs (see chapter 5).
Social Policy and Presidential Ideology in Latin America
40
As noted, CCTs and other recently adopted non-contributive programs mark an important
break with the Latin American model of social policy (Barrientos and Santibáñez 2009). CCT
benefits are entirely de-linked from employment, formal or otherwise. Furthermore, these
programs utilize sophisticated means-testing to target the poor. In contrast to past social assistance
based on subsidies and in-kind transfers, CCTs provide monetary transfers. The underlying
philosophy behind this approach is to empower households by letting them choose how best to
spend government assistance. In most CCTs, transfers are awarded to mothers, who are believed
to be more likely than fathers to spend the additional income on the children and family (Fiszbein
and Schady 2009, 9). This contrasts with traditional social policy centered on the male
breadwinner.

Figure 2-1: Number of CCT Beneficiaries by Year in Latin America, Brazil
and Mexico (2001-2010)

Source: Stampini and Tornarolli (2012) based on SEDLAC data.
0
20
40
60
80
100
120
140
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Total Brazil Mexico Others
Millionsof People
Social Policy and Presidential Ideology in Latin America
41

Overall, CCTs have been successful at achieving their twin goals of increasing enrollment
and reducing poverty. It is estimated that the region’s poverty headcount would be fully 13%
higher on average in the absence of CCTs, a figure that amounts to more than 15.5 million people
(Stampini and Tornarolli 2013, 26) (see table 2-4).
27
In 2010, CCTs boosted the incomes of
beneficiaries by an average of 20 percent and by as much as 50 percent in Panama. While it is still
too early to measure the effects of CCTs on inter-generational poverty transmission, several studies
cite them as a reason for the region wide drop in inequality over the last decade (López-Calva and
Lustig 2010; Lustig, López-Calva and Ortíz-Juárez 2013; Soares et al. 2009). With regard to
human capital development, there is strong evidence that CCTs have increased school attendance
and use of medical facilities, raised the average number of years of school attended, and improved
child nutrition (Adato and Hoddinott 2010; Baird et al., 2013; Fiszbein and Schady 2009; Rawlings
and Rubio 2005; Saavedra & García, 2012; Valencia Lomelí 2008; Villatoro 2005).
28
There is also
evidence, albeit weaker, that CCTs have, through increased school enrollment, reduced child labor
Figure 2-2: CCT Beneficiaries as a Share of Latin America’s Population
(2001-2010)

Source: Stampini and Tornarolli (2012) based on SEDLAC data and own
calculations based on WDI.
8.0%
9.4%
13.4%
15.0%
16.3%
17.8%
18.7%
19.1%
23.4%
24.1%
0%
5%
10%
15%
20%
25%
30%
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
% of Population Receiving CCTs
Social Policy and Presidential Ideology in Latin America
42
in the region (Barrientos 2013; de Hoop and Rosati 2014).
29
Furthermore, there is evidence from
Brazil and Mexico that CCTs reduce violent crime (Lance 2014).
These results stands in sharp contrast with those of the region’s traditional social security-
centered social policies, which largely neglected the poor and, at best, kept inequality constant
and, at worst, deepened it. Among a sample of 56 transfer programs in eight Latin American
countries, CCTs stand out as the most redistributive (Lindert, Skoufias and Shapiro 2006). Soares
et al. (2009) find that, despite representing expenditures of less than 1% of GDP, CCTs were
responsible for about 21% of the decline in inequality in Brazil and Mexico between 1990 and
2000. A less generous CCT in Chile was responsible for 15% of the decline in inequality in that
country during that period.
Figure 2-3: Impact of CCTs on Poverty Headcount ($2.5 US PPP Poverty
Line) in Selected Latin American Countries

Source: Stampini and Tornarolli (2012) based on SEDLAC data and own
calculations based on WDI.
Note: Average is weighted by population size.

0
5
10
15
20
25
30
35
40
45
ARG
2010
BRA
2009
CHI
2009
COL
2010
CRI
2010
ECU
2010
GUA
2011
MEX
2010
PAN
2008
PER
2008
PRY
2010
URU
2010
AVG
Estimated Poverty Headcount without CCT Observed Poverty Heacount
% of Population in Poverty
Social Policy and Presidential Ideology in Latin America
43
Figure 2-4: CCT Benefits as a Share of Average Incomes ($2.5 US PPP
Poverty Line) in Selected Latin American Countries

Source: Stampini and Tornarolli (2012) based on SEDLAC data and own
calculations based on WDI.
Note: Average is weighted by population size.

Table 2-4: Impact of CCTs on Poverty in 13 Latin American Countries
Country Year
Poverty
Headcount
(Millions)
Poverty Count
Without CCT
(Millions)
Difference
(Millions)
Difference
(Percent)
.
Argentina 2010 6.14 7.49 1.35 18.0%
Brazil 2009 15.21 16.90 1.69 10.0%
Chile 2009 4.29 4.59 0.30 6.5%
Colombia 2010 12.13 13.44 1.31 9.7%
Costa Rica 2010 4.57 5.06 0.49 9.7%
Ecuador 2010 15.91 19.24 3.33 17.3%
Guatemala 2011 41.84 42.72 0.88 2.1%
Jamaica 2010 9.50 9.95 0.45 4.5%
Mexico 2010 12.63 14.35 1.72 12.0%
Panama 2008 17.88 18.62 0.74 4.0%
Paraguay 2010 18.31 19.62 1.31 6.7%
Peru 2010 18.45 18.68 0.23 1.2%
Uruguay 2010 2.75 4.39 1.64 37.4%
Total 179.61 195.05 15.44 7.9%
.
Source: Stampini and Tornarolli (2012, 26) based on SEDLAC and own calculations.
Notes: Poverty is defined at as $2.5 US per day at PPP.

The success of these types of programs, Lindert, Skoufias and Shapiro (2006) argue, stems
not from the size of the transfers but, rather, from the efficiency with which they are targeted. By
0%
10%
20%
30%
40%
50%
ARG
2010
BRA
2009
CHI
2009
COL
2010
CRI
2010
ECU
2010
GUA
2011
MEX
2010
PAN
2008
PER
2008
PRY
2010
URU
2010
AVG
All Beneficiaries Poor Beneficiaries
% of of Group's Average Income
Social Policy and Presidential Ideology in Latin America
44
focusing on the poorest, CCTs have succeeded in reducing poverty and inequality at a relatively
small cost. With nearly 75 percent of their resources reaching the poorest 40 percent, CCTs
constitute one of the most redistributive tools available to Latin American policymakers (Goñi,
López & Servén, 2008, p. 21). With the notable exception of Ecuador, which includes the elderly
and disabled in its CCT, the cost of the programs in the region averages 0.5 percent of GDP or
about three percent of social spending (See table 2-5). And all this appears to be happening without
significantly disincentivizing work among beneficiaries (Alzúa, Cruces and Ripani, 2013;
Barrientos 2013; Cecchini and Madariaga 2011).
30

Table 2-5: Cost of CCTs in 16 Latin American Countries
Country Program
CCT Costs
as a Share
of GDP
Social
Spending
as a Share
of GDP
1

CCTs a
Share of
Social
Spending
.
Argentina Asignación Universal por Hijo 0.20% 27.78% 0.7%
Bolivia Bono Juancito Pinto 0.33% 18.42% 1.8%
Brazil Bolsa Familia 0.47% 27.06% 1.7%
Chile Chile Solidario 0.11% 15.64% 0.7%
Colombia Familias en Acción 0.39% 13.62% 2.9%
Costa Rica Avancemos 0.39% 22.93% 1.7%
Dominican Republic Solidaridad 0.51% 7.27% 7.0%
Ecuador
2
Bono de Desarrollo Humano 1.17% 9.84% 11.9%
El Salvador Red Solidaria 0.02% 13.01% 0.2%
Guatemala Mi Familia Progresa 0.32% 8.07% 4.0%
Honduras Programa Asignación Familiar 0.24% 12.04% 2.0%
Mexico Oportunidades 0.51% 11.29% 4.5%
Panama Red Oportunidades 0.22% 10.53% 2.1%
Paraguay Tekopora 0.36% 11.01% 3.3%
Peru Juntos 0.14% 9.95% 1.4%
Uruguay Asignaciones Familiares 0.45% 23.32% 1.9%
Weighted Average 0.42% 18.81% 2.78%
.
Sources: Cruces and Gasparini (2013, 55) and own calculations based on ECLAC and WDI.
Notes:
1.
Data covers the most recent year available.
2.
Program also covers the elderly and disabled.

Social Policy and Presidential Ideology in Latin America
45
4.1 The Intellectual Origins of CCTs
A series of intellectual trends that took hold during the 1990s help explain the emergence
of CCTs as a region-wide model for combating poverty. In the aftermath of market reform, a broad-
based consensus emerged behind the idea that growth alone was insufficient for reducing poverty
and that permanent targeted anti-poverty programs – rather than short-term safety nets – were
necessary (Birdsall and de la Torre 2001; Birdsall and Székely 2003).
31
There was also substantial
agreement that investment in human capital offered the best hope for breaking the inter-
generational transmission of poverty.
For Santiago Levy (2006, 10), who is widely credited as the father of CCTs, these programs
represent the culmination of advances in scholarly understanding of poverty made during the 1980s
and 1990s. Scholars adopted a holistic understanding of the causes of poverty that emphasized the
links between food consumption, nutrition, health, and education. Providing any of these benefits
in isolation, it was argued, would likely be insufficient to pull families out of poverty. In Levy's
(2006, 12) words:
“To break the vicious cycle of poverty it was essential not only to subsidize food
(that is, to transfer purchasing power) but simultaneously to give poor households
more certainty about the availability of income to buy food or other necessities; to
deliver more information on hygiene and reproductive health; to provide alternative
sources of income for present consumption in order to replace income earned by
children, allowing them to go to school more regularly, better fed, and for longer
periods of time; to modify households’ structure of health risks by inducing families
to have more frequent and systematic contact with health service providers; to
Social Policy and Presidential Ideology in Latin America
46
reduce their perception of vulnerability so that they can bear more risk and perhaps
save and invest in new projects."
Thus, CCTs were originally envisioned as comprehensive programs that included the
provision of multiple services other than education. Education, however, quickly became the most
visible aspect of early CCTs to the point that most of the recent programs condition benefits only
on school attendance.
CCTs were compatible with three broader “economic-ideological” trends in the social
policy community at the time: the shift away from a rights-based view of social entitlements
toward an obligations-based conception, the move toward market and incentive-based anti-poverty
policies, and the increased acceptance of narrow poverty targeting (Adato and Hoddinnott 2010,
13-14). Additionally, by addressing concerns such as female empowerment, human capital and
community participation, these programs fit well with the “current mainstream discourse on
poverty reduction” (Feitosa de Britto 2008, 185).
As Levy was developing what became Mexico’s Progresa (later renamed Oportunidades)
CCT, a slightly different approach to combating poverty was gaining ground in Brazil. In 1991,
the same year that Levy (1991) published a World Bank research paper detailing many of the ideas
that came to be part of Progresa, Eduardo Suplicy, a Senator from the left-wing Partido dos
Trabalhadores (PT), proposed an unconditional basic income guarantee of approximately two
minimum wages for all Brazilians over 25. The proposal was approved unanimously by the Senate
but languished in the House of Representatives. In 1993, economist Jose Marcio Camargo retook
the proposal and refocused it toward preventing the inter-generational transmission of poverty.
Instead of awarding transfers to adults, he proposed targeting families with school-age children
and conditioning benefits on school attendance (Bandeira Coelho 2012; Sugiyama 2013). These
Social Policy and Presidential Ideology in Latin America
47
ideas would form the basis for a series of local and state-level programs that culminated in the
creation of a national CCT at the turn of the century.
Starting from two different intellectual traditions, the two largest countries in the region
arrived at similar solutions. Mexico rolled out Progresa in 1997 and, while this was the first
national CCT, the first sub-national CCTs were launched in the Brazilian cities of Brasilia and
Campinas two years before. It was not until 2001 that Brazil adopted its own national CCT in
Bolsa Escola, which was later expanded and renamed Bolsa Familia in 2003.
32

While it is naïve to ignore the political calculus behind the adoption of CCTs, these
programs would not have been possible without a clear change in the way some leaders envisioned
social policy. In much the same way as their predecessors approached market reform, Presidents
Zedillo in Mexico (Zedillo 2009) and Cardoso in Brazil (Sugiyama 2013, 83) shielded social policy
technocrats from political pressures, thus freeing them to design and implement policies based on
best practices. It is hard to downplay how path-breaking it was that in Mexico, under the Partido
Revolucionario Institucional (PRI) of all parties, a cadre of policy-motivated technocrats led by
Levy was allowed to develop a large-scale anti-poverty program largely free of political
interference. The emergence of CCTs was equally surprising in Brazil, which has had a long
history of federal funds being used for clientelistic purposes at the municipal and state levels
(Hunter and Sugiyama 2009).
Given the initial success of the Brazilian and Mexican programs, CCTs were quickly
embraced by other Latin American governments, academics, and international financial
institutions (IFI), most notably the World Bank and Inter-American Development Bank (IDB).
Having become increasingly focused both on combating poverty and improving education in the
aftermath of the Washington Consensus, these institutions, were particularly receptive to CCTs.
Social Policy and Presidential Ideology in Latin America
48
By the mid-2000s, these institutions were dedicating an important part of their lending portfolios
to assisting countries in setting up and expanding these types of programs (Ancelovici & Jenson,
2013; Lana & Evans, 2004; Martínez Franzoni & Voorend, 2011; Sugiyama, 2011).
Although the Bolsa Escola/ Bolsa Familia came to more closely resemble Mexico’s
Progresa/Oportunidades than a minimum basic income policy, its approach to conditionality was
still greatly influenced by the idea of providing citizens with a basic minimum income. Following
Cecchini and Martínez (2011), Chapter 5 argues that these two pioneering programs actually
constitute two distinct varieties of CCTs with regard to how they balance the twin objectives of
reducing poverty and increasing human capital.
4.2 What’s New about CCTs?
CCTs are a Latin American innovation. These programs represent a clear break not only
from the region’s traditional social policy model centered on contributive social insurance for the
formally employed, but also from earlier attempts at providing social assistance through
consumption subsidies, in-kind transfers and emergency social funds. The key features that
differentiate CCTs from the policies that preceded them are their scope and generosity, degree of
progressivity, multi-dimensional and inter-generational understanding of poverty, and the non-
discretionary manner in which benefits are awarded (Barrientos and Santibáñez 2009).
Prior to the 1980s, social assistance in Latin America was minimal and almost exclusively
centered on consumption subsidies and, to a lesser extent, in-kind transfers of food (Ferreira and
Robalino 2011). These types of policies represented a top-down approach to poverty relief: a
planner decided what goods the poor needed and either arranged for them to be delivered to them
or used regulation to ensure that they were priced affordably. In addition to being expensive and
market-distorting, these policies did little to relieve poverty. Blanket subsidies, even those on
Social Policy and Presidential Ideology in Latin America
49
staples of the poor’s diet, ultimately benefit the middle class as much or more than the poor, while
in-kind transfers often failed to reach the most remote and poorest communities (Levy 2006).
In contrast, the social investment funds launched in the early 1990s as a response to the
debt crisis and structural adjustment, were bottom-up in that they allowed communities to propose
their own economic development projects, which were paid for by the government. Funds targeted
poor communities, but ultimately failed to reach the poorest communities. As Siri (2000, 12) notes:
“the demand-driven approach, which by definition relies on the requests of potential beneficiaries,
tends to bypass those who have little capacity to formulate projects or articulate demands.”
Furthermore, in targeting communities rather than individuals, these programs funded the priorities
of local elites. They did little to empower under-represented groups such as women, who had
limited say in the selection and formulation of projects and obtained less than 15 percent of the
jobs created by them (Goodman et al. 1997; Siri 2000, 14).
In contrast to the policies described above, CCTs primarily benefit the poor and vulnerable
(Lindert, Skoufias and Shapiro 2006). They also represent a qualitative leap in terms of their
ambition. Never before had the region’s countries attempted to provide an income boost of this
size to such a large number of people. In 2010, CCTs benefited roughly half of the region’s poor,
boosting their incomes by a full third on average (see figure 2-4).
33
In contrast with previous
policies, which focused only on relieving current poverty, CCTs seeks to prevent future poverty
by linking transfers to education and healthcare. Proponents of CCTs envision conditionality as a
means of integrating a wide range of interventions aimed at developing human capital and
empowering the poor.

Social Policy and Presidential Ideology in Latin America
50
4.2.1. ‘Silver Bullet’ Against Clientelism?
The selection of beneficiaries through transparent and well-publicized rules and the actual
application of said rules – two hallmarks of programmatic policy (Kitschelt and Wilkinson 2007;
Stokes et al. 2013) – are arguably the most revolutionary aspects of CCTs. Social policy in Latin
America and the rest of the developing world has traditionally been clientelistic: powerful actors
(patrons operating through brokers utilized government handouts to buy the poor’s political
support. At the cost of their political autonomy, the poor, for whom a small government handout
can make a huge difference, have served as a captive constituencies for political machines. Once
beneficiary selection ceases to be discretionary and becomes rules-based and de-politicized,
officeholders lose the ability to use social policy to reward supporters and punish opponents. This
in turn sets the stage for major improvements in the quality and accountability of democracy.
Beyond rules-based targeting, CCTs also incorporate further innovations that reduce the
likelihood of political manipulation. Fearing interference at the local level, the technocrats who
designed the pioneer CCTs opted to create programs that minimized the discretionary authority of
sub-national politicians (Díaz-Cayeros, Magaloni and Estevez Forthcoming, 191; Fenwick 2009;
Sugiyama and Hunter 2013), an approach that was emulated by late adopters. In the most
developed countries in the region, CCT stipends are deposited directly to ATM cards, reducing the
opportunities for funds to be diverted and minimizing the role of local officials in the disbursement
of funds (Sugiyama and Hunter 2013).
That CCTs are programmatic and shielded from clientelism does not mean that they lack
political effects (see table 2-6). To the contrary, in a clear sign of retrospective economic voting,
there is substantial evidence that CCTs benefit incumbent governments – even when those
governments did not initiate the programs in question (Díaz-Cayeros, Estevez and Magaloni 2009;
Social Policy and Presidential Ideology in Latin America
51
Zucco 2013). Utilizing a variety of statistical and experimental techniques, scholars have found
robust evidence showing that beneficiaries were more likely to support incumbents in presidential
elections in Brazil (Hunter and Power 2007; Zucco 2008; Zucco 2013; Zucco and Power 2013)
and Mexico (Cornelius 2004; Díaz-Cayeros, Estevez and Magaloni 2009; de la O 2013).
34

Similarly, there is cross-national evidence (Layton and Smith 2011) as well as evidence from
Colombia (Baez et al. 2011l; Camargo and Raga 2011; Nupia 2011) and Uruguay (Manacorda,
Miguel and Vigorito 2011) that beneficiaries would be more likely to back incumbents in
hypothetical elections,.even during an economic crisis (Layton and Smith 2011). CCTs may even
tip the balance in a close race. Looking at Mexico’s highly contested 2006 presidential elections,
Díaz-Cayeros, Estevez and Magaloni (2009) conclude that incumbent bias among Oportunidades
beneficiaries boosted support for Felipe Calderón of the incumbent right-wing Partido Acción
Nacional (PAN) enough to allow him to pull ahead of Andrés Manuel López Obrador of the left-
wing Partido de la Revolución Democrática (PRD).
35

For better or worse, CCTs are closely associated with the incumbent president but appear
to have no effect on incumbent parties in legislative (Hunter and Power 2007; Zucco 2013) or sub-
national elections (Montero 2010).
36
There is also evidence that CCTs increase political
participation and turnout (Baez et al. 2012; de la O 2013; Hunter and Power 2007) and that this
effect may be larger among women, which constitute the overwhelming majority of beneficiaries
(Baez et al. 2012). CCTs may also improve governance: corruption tends to be lower in Mexican
states where a large share of the population is enrolled in Oportunidades (Grimes and Wängneru
2010).
Social Policy and Presidential Ideology in Latin America
52
Table 2-6: Political Effects of CCTs
Authors Case
Level of
Analysis
Dependent
Variable
Operationalization
of CCT
Method Outcome
Cross-National Research

Leyton and
Smith (2011)
9 Latin
American
countries (2010)
Survey data
Vote for incumbent
president in
hypothetical
election, evaluation
of president
Someone in
household receives
CCT
Logit
Beneficiaries give president higher
evaluation and more likely to reelect.
Among those who perceive crisis, CCT
boosts vote intention 9 points
compared to 4 points for those who do
not perceive crisis. Among those who
perceive crisis, CCT leads to 10 point
drop in vote intention, compared to 15
points for those without CCT.
.
Mexico

Cornelius (2004)
Progresa (2000
elections)
Survey data
Vote for incumbent
party candidate
Responded receives
Progresa
Multinomial
logit
Beneficiaries, particularly women and
elderly, more likely to support
incumbent. Effect is small.

De la O (2013)
Progresa (2000
elections)
Electoral
data (precinct
level)
Voter turnout and
vote for incumbent
party candidate
Community was
randomly selected to
participate in CCT
Natural
experiment
Turnout and support for incumbent
increased by 7% and 9%, respectively.
No effect on support for other parties.

Díaz-Cayeros,
Estevez and
Magaloni (2009)
Oportunidades
(2006 elections)
Survey data
(exit polls)
Voting intention,
opinion of
candidates, partisan
identification,
presidential
approval, and
pocketbook
evaluation.
Responded receives
Oportunidades
Matching and
Probit
Beneficiaries more likely to support
incumbent. This allowed him to win
election. Beneficiaries less likely to
support main challenger; more likely to
have positive opinion of incumbent,
identify with incumbent party, approve
of president and give positive
pocketbook evaluation.

Green (2005)
Progresa (2000
elections)
Electoral
data (locality
level)
Party vote shares
and turnout
Community was
randomly selected to
participate in CCT
Regression
discontinuity
No effect on incumbent party vote
share, opposition party vote share or
voter participation.

Grimes and
Wängnerud
(2010)
Oportunidades
(2005)
State-level Level of corruption
Percent of
households in state
with Oportunidades
OLS
States with larger proportion of
housholds receiving CCT benefits have
less corruption.

Social Policy and Presidential Ideology in Latin America
53

Table 2-6: Political Effects of CCTs (Continued)
Authors Case
Level of
Analysis
Dependent
Variable
Operationalization
of CCT
Method Outcome
Brazil

Bohn (2011)
Bolsa Familia
(2002 and 2006
elections)
Survey data
Likelihood of
voting for Lula da
Silva
CCT recipient Probit
No effect in 2002. Negative effect in
2006.

Montero (2010)
Bolsa Familia
(2006 elections
in three
northeastern
states)
Electoral
data
(municipal
level)
Vote for right-wing
incumbent
governors and left-
wing opposition
Percentage of
families receiving
CCT
OLS and
geographically
weighted
regression
using GIS
No effect.

Zucco (2008)
Bolsa Familia
(2006 elections)
Electoral
data
(municipal
level)
Vote for incumbent
president
Percentage of
families receiving
CCT
OLS
Incumbent benefited in every model
specification tested.

Zucco (2013)
Bolsa
Escola/Familia
(2002, 2006 and
2010 elections)
Electoral
data
(municipal
level)
Vote for candidate
from incumbent
party in presidential
and legislative
elections
Percentage of
families receiving
CCT and CCT
expenditures
Matching and
OLS
Incumbent benefited in every election
regardless of party. Expenditure effect
declined from 15 points in 2002 to 6.5
points in 2010. Scope effect remained
constant. No effect on legislative
elections.

Bolsa Familia
(2006 and 2010
elections)
Survey data
Likelihood of
voting for
incumbent
Percentage of
families receiving
CCT and CCT
expenditures
Matching and
logit
Incumbent benefited in both elections.
Effect declined between elections.

Zucco and Power
(2013)
Bolsa Familia
(2002 and 2006
elections)
Survey data
(replication
of Bohn
2011)
Likelihood of
voting for Lula da
Silva
CCT recipient Probit Incumbent benefited.

Bolsa Familia
(2006 elections)
Survey data
Likelihood of
voting for Lula da
Silva
CCT recipient
Matching and
probit
Beneficiaries more likely to vote for
Lula.

Social Policy and Presidential Ideology in Latin America
54
Table 2-6: Political Effects of CCTs (Continued)
Authors Case
Level of
Analysis
Dependent
Variable
Operationalization
of CCT
Method Outcome
Colombia

Baez et al.
(2012)
Familias en
Acción (2010
elections)
Individual-
level
Political support for
incumbent
Percent beneficiaries
in municipality
Regression
Discontinuity
A 10% increase in beneficiaries raises
incumbent vote share 1.5 points.

Voting
precinct
People registered to
vote and voter
turnout
Percent beneficiaries
in precinct
Regression
Discontinuity
Increased probability of registering to
vote and voting, particularly among
women.

Camargo and
Rodríguez-Raga
(2011)
Familias en
Acción (2010)
Survey data
Vote for incumbent
president in
hypothetical
election
Someone in
household receives
CCT
Logit
Beneficiaries more likely to vote for
incumbent by 1.6-2.5 points and vote
by 7-9 points.

Nupia (2011)
Familias en
Acción (2002,
2006 and 2010
elections)
Electoral
data
(municipal
level)
Vote for
presidential
candidate from
incumbent party
Number of
beneficiary families
per 1000 inhabitants
Instrumental
Variable
Regression
A 1% increase in beneficiaries
increases votes for incumbent by 0.5
points. This effect was larger among
municipalities that had not supported
incumbent in prior election.

Other Countries

Linos (2013)
Honduras:
PRAF (2001 and
2005 elections)
Electoral
data
(municipal
level)
Vote for incumbent
mayor
CCT beneficiary
Experimental,
Difference in
differences
Re-election probabilities increase by
39%.

Vote for
presidential
candidate from
incumbent party
CCT beneficiary
Experimental,
Difference in
differences
No effect.

Manacorda,
Miguel and
Vigorito (2011)
Uruguay:
PANES (2007)
Survey data
Opinion of current
government relative
to previous one
Received CCT
Regression
Discontinuity
Beneficiaries thought current
government was better than
predecessor by 11-13 points. One year
after program ended, beneficiaries
thought government was better than
predecessor by 8-12 points.

Social Policy and Presidential Ideology in Latin America
55
That CCTs target beneficiaries based on objective criteria does not rule out the possibility
of them being manipulated for electoral purposes. Among the region’s nationwide CCTs, only four
utilize pre-determined formulas to adjust stipends (Cecchini and Madaraiga 2011, 61).
37
This
opens the possibility that incumbents will adjust them at politically opportune moments. This
threat is compounded by the fact that only three programs are enshrined in law while the rest
operate through presidential decrees (Cecchini and Madaraiga 2011, 161-65).
38
The reliance on
decrees rather than laws is an example of the limited institutionalization prevalent among social
assistance programs in developing countries (Barrentos 2013).
Having summarized the main features and consequences of CCTs, the next section
discusses the main critiques that have been levied against them.
4.3 Limitations of CCTs
Despite their very promising results, CCTs are far from a perfect tool for combating
poverty and have been subject to a number of criticisms. The main critiques relate to limitations
inherent to CCTs, conditionality and their use of means testing and targeting.
By their very design, CCTs are subject to important limitations which prevent them from
serving as a country’s only anti-poverty policy. As policies geared toward tackling long-term
structural poverty, CCTs are ill-equipped to deal with the “transient” poor: lower and lower-middle
class families who might be temporarily pushed below the poverty line as a result of job loss
(Fiszbein and Schady 2009). Program eligibility is determined using long-term indicators and the
information-gathering systems used are not designed to incorporate new beneficiaries on short
notice (Handa and Davis 2006). This presents a particularly serious challenge in light of the
region’s current economic slowdown. Furthermore, as noted by Fiszbein and Shady (2009, 195),
“because of their focus on building the human capital of poor children, CCTs are not feasible for
Social Policy and Presidential Ideology in Latin America
56
the elderly poor, poor households without children, or households with children outside the age
covered range.”
CCTs have also been criticized for distracting policymakers from the crucial task of
improving education since there is no evidence that they have improved in-class learning (Draibe
2006; Levy 2006; Ponce 2006; Valencia Lomelí 2008) and increased enrollment cannot be equated
with increased learning, particularly when the poorest students attend sub-standard schools.
Furthermore, there is concern that by increasing enrollment without increasing education funding,
CCTs may be placing an excessive strain on schools and teachers, which will ultimately worsen
the quality of education. Estevan (2013), however, does find that large-scale CCTs, those whose
coverage extends to the median voter, are associated with increased education spending.
Even ignoring education quality, which falls outside the scope of CCTs, the validity of the
central premise of conditioning transfers to school attendance – transforming today’s student into
tomorrow’s more productive (and thus better remunerated) worker – is far from certain. This
assumes “a socioeconomic context in which abilities can be realized and skills can be translated
into higher earnings” (Valencia Lomelí 2008, 486), a policy challenge that extends well beyond
the realm of social policy.
39
Unless this crucial link is in place, CCTs will have only limited effects
on long-term poverty. The one study of the long-term effects of CCTs participation did not yield
encouraging results. In what is so far the only study of its kind, Yaschine (2012) finds no statistical
difference in terms of the quality of the jobs worked by a sample of 3,500 Mexicans who were
enrolled in Progresa/Oportunidades for 7-10 years relative to a comparable sample that did not
benefit from the program.
In response to this critique, Fiszbein and Shady (2009, 143) note that even in the absence
of increased learning, CCTs may improve wages through ‘the diploma effect’ – the very tangible
Social Policy and Presidential Ideology in Latin America
57
wage differential that exists between those who complete high school versus those who do not.
Furthermore, they argue that staying in school longer could help students develop “important non-
cognitive skills, like discipline, responsibility and motivation, which may be rewarded by the labor
market.”
4.3.1. Critiques of Conditionality
The conditional nature of CCTs has been criticized from multiple angles. Freeland (2007,
77) and proponents of basic citizen’s income (Lo Vuolo 2013a; Standing 2008) criticize
conditionality, the logic behind CCTs. For them, it is contradictory for governments to claim that
social protection is a human right, while at the same time depriving their poorest citizens of that
right if they fail to comply with program requirements. Conditionality enforcement is most likely
to hurt the people most in need of government assistance as they are the ones who face the greatest
opportunity cost from having their children in a classroom instead of working (Bastagli 2010, 12).
Conditionality can also be criticized from a cost-benefit perspective. Enforcing
conditionality is bound to increase administrative costs compared to unconditional programs.
However, an analysis of three early CCTs finds that the administrative costs of these programs are
comparable, if not lower, than that of more traditional anti-poverty programs in the region (Caldés,
Coady and Maluccio 2006).
Others have argued that education conditionalities are unnecessary, at least at the
elementary level.
40
Hanlon, Barrientos and Hulme (2010) maintain that poor parents are not any
less likely to want their children to go to school than other parents, but do face much larger
opportunity costs. Under this logic, unconditional transfers should work just as well as conditional
ones as long as the stipend is generous enough to compensate for a child’s forgone earnings. Funds
saved through lower administrative costs could then be used to increase stipends. In contrast to
Social Policy and Presidential Ideology in Latin America
58
this view, a series of recent studies utilizing randomized designs have found robust evidence of a
positive effect of conditionality on enrollment (Baird, McIntosh and Özler 2011; de Brauw and
Hoddinott 2011; Schady and Araujo 2008).
From a political economy perspective, Fiszbein and Schady (2009) argue that conditioning
transfers on “good behavior” may make CCTs easier to “sell” to tax payers. Without the good
behavior hook, policies that benefit small and politically weak groups will be politically
vulnerable. Taxpayers may be more open to transferring funds to poor people that are perceived
as helping themselves by investing in their children. Related to this, they speculate that programs
centered on building human capital among children will be more acceptable to voters because
children cannot be blamed for being poor (Fiszbein and Schady 2009, 63). Indeed, survey research
shows that Latin Americans are supportive of conditional transfers and opposed, by large margins,
to unconditional minimum income policies (Lo Vuolo 2013b; Waltenberg 2013).
From the perspective of gender, it has been argued that conditionalities constitute an
additional burden on women’s already hectic schedules (Molyneux 2006).
41
These added
responsibilities may make it more difficult for women to work outside the home. Furthermore, by
putting women in charge of a significant source of income, CCTs may upset existing family
dynamics and create tension between spouses that may culminate in domestic violence (Bradshaw
and Quirós Víquez 2008). While it is undeniable that these programs create additional work for
women, there is also evidence that the added responsibilities and power provided to them by CCTs
“increases their sense of empowerment, their standing in their local communities, their access to
health and nutritional information, and, in some cases, their decision making power within their
own households” (Benderly 2011, 1). Some women have used the additional income provided by
Social Policy and Presidential Ideology in Latin America
59
CCTs to start their own businesses and used CCT-related meetings to organize politically (Escobar
Latapí and González de la Rocha 2008).
4.3.2. Critiques of Means-Testing and Targeting
While generally praised for reaching those who most need government assistance, CCTs
have also been criticized for their reliance on means-testing. There are two main critiques of means
testing: that it is prone to administrative problems and that by restricting beneficiaries to the poor,
who are politically weak, it limits the generosity of CCT programs and threatens their long-term
sustainability.
On administrative grounds, critics argue that means-tested programs are more costly to
administrate, disincentivize work among people with incomes near the inclusion threshold, and
are prone to errors of inclusion of the non-poor and, perhaps more worryingly, exclusion of the
poor (McGuire 2013). While it is true that means-testing carries a cost, CCTs, despite their scope,
are very affordable, particularly compared with the social security programs that dominate the
region’s social spending budgets. Additionally, their administrative costs are comparable to those
of other social programs (Caldés, Coady and Maluccio 2006). There is also little evidence that they
disincentivize work (Alzúa, Cruces and Ripani, 2013; Barrientos 2013; Cecchini and Madariaga
2011).
Focusing exclusively on the poor may threaten the long-term sustainability of CCTs. From
a political economy perspective, programs whose benefits are restricted to politically weak groups
such as the poor are highly vulnerable to budget cuts and even cancellation, particularly during
tough economic times. Relaxing means tests to target a broader cross-section of society would
increase the size of the coalition supporting these policies (Huber 2009). Furthermore, universality
would potentially give wealthier and more politically influential groups a stake in the continuation
Social Policy and Presidential Ideology in Latin America
60
and expansion of CCTs (McGuire 2013). Given a fixed amount of funding, means-testing focuses
benefits on those who need them the most. However, budgets are not fixed and policies supported
by the rich and middle class are more likely to see their budgets increase over time. Under this
logic, it not unreasonable to expect that universal transfers will be more generous than well-
targeted ones and, thus, more beneficial for the poor (Gelbach and Pritchett 2002).
While still very much focused on the poor, CCT coverage has expanded dramatically over
the last decade. The coalition of beneficiaries that could potentially mobilize to defend CCTs in
the face of cutbacks now spans nearly a quarter of the region’s population. Since 2006, the number
of CCT beneficiaries has surpassed the number of poor (measured at a standardized poverty line
of $2.50 per day at purchasing power parity). Increased coverage, however, has brought with it
increasing leakage levels (see table 2-7).
Table 2-7: Leakages in CCTs in 12 Latin American Countries
(2010 or Most Recent Year Available)
Country Program Name
Leakage as
share of
beneficiaries
($2.50)
Leakage as
share of
beneficiaries
($4.00)
.
Argentina Asignación Universal por Hijo 75.7% 54.8%
Brazil Bolsa Familia 50.0% 28.1%
Chile Chile Solidario 87.1% 69.6%
Colombia Familias en Acción 71.4% 49.4%
Costa Rica Avancemos 89.4% 70.9%
Ecuador Bono de Desarrollo Humano 65.1% 40.7%
Guatemala Mi Familia Progresa 32.2% 10.4%
Mexico Oportunidades 61.4% 40.5%
Panama Red Oportunidades 21.6% 7.9%
Paraguay Tekopora 41.2% 19.2%
Peru Juntos 33.1% 11.3%
Uruguay Asignaciones Familiares 84.4% 57.0%
Weighted Average 57.7% 36.3%
.
Source: Stampini and Tornarolli (2012, 17) based on SEDLAC.
Notes: Coverage is defined as the percentage of the poor who benefit from the programs.
Leakage is the percentage of beneficiaries that are not poor. Average is weighted by
population.
Social Policy and Presidential Ideology in Latin America
61

For now, CCTs do not appear to be in danger of disappearing since most Latin Americans
are highly supportive of redistributive social policies and by nearly a two to one margin consider
poverty a consequence of societal injustice rather than laziness according to a 2006 survey (Lindert
and Vicensini 2010, 11). Similarly, more than half of Latin Americans surveyed in 2004 believed
that “the poor are trapped in poverty” (62%) and that “luck determines income” (55%) (Ardanaz
2009, 23).
42
On the other hand, while CCT stipends have not been cut in the face of slowing
economies, they have not increased either. Stipends in the pioneer CCTs of Brazil and Mexico
have largely tracked inflation. However, as will be analyzed in chapter 5, two of the second
generation CCTs studied, Costa Rica’s Avancemos and Bolivia’s Bono Juancito Pinto, have not
adjusted their stipends since their launch in 2006. The third program studied, Argentina’s
Asignación Universal por Hijo, has adjusted stipends erratically, but for the most part has kept up
with inflation.
Having surveyed existing research on social policy in this chapter, I now proceed to the
empirical portion of this study.
Social Policy and Presidential Ideology in Latin America
62
Chapter 3.
Presidential Ideology or Income Distribution?:
Determinants of Social Spending, 1990-2010

1. Introduction
Social policy has a substantive impact on the majority of a country’s population and
particularly on the poor. Means-tested programs can push a family above the poverty line. Public
healthcare can mark the difference between sickness and health and ultimately between life and
death. Education directly affects an individual’s economic prospects. Unless their most basic needs
are met, individuals cannot fully exert their rights as citizens in a democracy. Support for
democracy is also linked to people’s perceptions of their material well-being and the quality of
basic social services such as education and healthcare.
A large body of research has analyzed the many facets of social policy and succeeded in
generating substantial cumulative knowledge. However, compared to the vast and well-established
empirical literature on industrialized countries (Esping-Andersen 1990; Huber and Stephens 2001;
Pierson 1995), research on the politics of social policy in developing countries remains
underdeveloped (Carnes and Mares 2007). Nonetheless, in response to this gap in our knowledge
of social policy, in recent years there has been an upsurge of research on the determinants of social
spending in Latin America (Avelino, Brown and Hunter 2005; Brown and Hunter 1999; Huber,
Mustillo and Stephens 2008; Huber and Stephens 2012; Kaufman and Segura-Ubiergo 2001;
Segura-Ubiergo 2007; Wibbels 2006; Zarate Tenorio, forthcoming) and developing countries
more generally (Stasavage 2005; Rudra 2002), as well as comparisons across developing regions
Social Policy and Presidential Ideology in Latin America
63
(Haggard and Kaufman 2008). In brief, research on the developing world has started to counteract
the overreliance on research focused exclusively on industrialized countries.
Seeking to contribute to this new research agenda on social policy in developing countries,
this chapter addresses the following three questions. First, why do some developing countries
spend more on social policy than others? Second, why do developing countries prioritize some
forms of social spending and others? Third, and more broadly, how well do existing theoretical
explanations of social policy, which were derived primarily from the experiences of Western
Europe and the United States, apply to the developing world?
This chapter contrasts the applicability to Latin America of two leading theories on the
determinants of social spending and redistribution: power resource theory (PRT), which predicts
that governments with left-leaning ideologies will prioritize social policy (Huber and Stephens
2001; Korpi 1983; Stephens 1979), and explanations centered on income distribution, specifically
the median voter theory (MVT) (Meltzer and Richards 1981), which predicts that more unequal
democracies will engage in more economic redistribution than egalitarian ones, and the structure
of inequality hypothesis (Lupu and Pontusson 2011), which predicts that middle-income voters
will support redistribution when the distance between them and the rich is greater than the distance
between them and the poor. To accomplish this, it conducts a time-series cross-sectional analysis
of the determinants of total social spending and its three component parts – education, healthcare
and social security spending – in 18 Latin American countries from 1990 to 2010.
43
To my
knowledge, this is the first test of the structure of inequality on a sample of developing countries.
In contrast to the research on Western Europe, this chapter finds only weak support for
PRT. Having a left-of-center president is associated with slightly higher social spending than
having a non-left president. However, distinguishing between center- and far-left presidents
Social Policy and Presidential Ideology in Latin America
64
reveals that the effect of ideology is driven almost entirely by the administrations of far-left
presidents. In contrast, having a center-left president has had no effect on social spending levels
during the last two decades. Disaggregating spending reveals that presidential ideology operates
exclusively through education. In fact, all else equal, countries with far-left presidents were found
to spend as much as 48 percent more on education than countries with other types of presidents.
Income distribution was found to be a more robust determinant of social spending levels
than presidential ideology. However, social spending is not influenced by the level of inequality,
as predicted by the MVT, but rather by the structure of inequality. As the gap between the incomes
of the rich and those near the median income becomes larger, the latter become more likely to
align with the poor in favor of increased social spending. Disaggregating total social spending
reveals that the structure of inequality operates through education and social security spending.
The chapter proceeds as follows. Section 1 presents the hypotheses associated with the
theories tested. The next section focuses on issues of case selection, data, control variables, and
the main challenges associated with time-series cross-sectional data. Section 3 presents the
findings of the statistical analysis. The chapter then concludes by explaining the significance of
the findings for the study of social policy in developing countries, comparing these findings to the
findings in the broader literature, and offering suggestions regarding future avenues for research.

2. Theory and Hypotheses
Initial research attributed cross-national and over-time variation in social policy and
economic redistribution to the “logic of industrialism” and posited that welfare state development
was a product of modernization and its associated social and demographic changes.
Industrialization created the need for a more educated workforce and policies to protect workers
Social Policy and Presidential Ideology in Latin America
65
from workplace-related risks and old age, which in turn required the expansion of social spending
(Wilensky 1975). This largely functionalist approach came under criticism for failing to take into
account the role of politics and economic conflict in the construction of welfare states.
In place of the modernization approach, two distinct approaches to theorizing social policy
have been proposed. The first, power resource theory (PRT), is a political sociological approach
that places emphasis on class conflict and political ideology. The other is a political economy
approach, most commonly associated with the median voter theory (MVT), which focuses on
patterns of income distribution. These two theories dominate current discussions of social policy.
Thus, what follows presents the hypotheses derived from these two approaches to theory.
2.1. Power Resource Theory: Presidential Ideology
PRT initially sought to account for the construction of welfare states in Western Europe in
terms of popular mobilization and politics. In a clear elaboration of Marxist class analysis, this
theory attributes the expansion of social policy to the relative strength of the working class and,
especially, their social democratic party allies (Korpi 1983; Stephens 1979). In a nutshell, social
policy hinges on the power of the left. However, to extend the application of PRT from its home
base in Western Europe to the developing world, an important adjustment is required.
The core hypotheses of PRT can be stated as follows: governments with a left-of-center
ideology are expected to spend more on social policy than those with centrist or rightist ideologies.
But in developing regions, such as Latin America, legislators may not be the key actors that shape
social policy. In particular, as is well known, Latin America’s legislatures are “reactive,” in that
their role in the policymaking process is largely limited to amending and vetoing executive
proposals rather than designing and enacting their own proposals (Cox and Morgenstern 2001).
Social Policy and Presidential Ideology in Latin America
66
Thus, to properly test PRT in a context such as Latin America, it is important to focus not on the
ideology of parties in congress but instead on presidential ideology.
2.2. Political Economy of Income Distribution: Level and Structure of Inequality
The most discussed alternative to PRT is the median voter theory (MVT), which
exemplifies a political economy approach focused on patterns of income distribution (Acemoglu
and Robinson 2006; Boix 2003; Meltzer and Richard 1981; Perotti 1996). The MVT predicts that
demand for redistribution within a democratic society will be a function of inequality, which in
turn influences the interests and political actions of the pivotal median voter. The more unequal a
society, the lower the median voter’s income is relative to the mean income level. And, as a result,
the more the median voter stands to gain from redistribution through taxes and transfers. Thus, the
model predicts the median voter’s decisive role in the political process will lead to increased
redistribution in highly unequal societies, a prediction that is highly relevant to the developing
world and Latin America in particular. Thus, a key alternative hypothesis to PRT is as follows: the
more unequal a country is, the more income will be redistributed through social policy.
44

The political economy of income distribution is not exhausted by the MVT, however. In a
ground-breaking recent paper, Lupu and Pontusson (2011) argue that it is not the level of
inequality, as posited by the MVT, but rather the structure of inequality that drives demand for
redistribution. They argue that voters with incomes near the median will seek redistribution when
the distance between their share of national income and that of the poor is small relative to the
distance between their share of the national income and that of the rich. Following Lupu and
Pontusson, three results are expected: 1) the ratio of the income share of rich to the median voter
will have a positive effect on social spending 2) the ratio of the income share of the median voter
Social Policy and Presidential Ideology in Latin America
67
to poor will have a negative effect on spending, and 3) the ratio between these two measures, the
skew of the income distribution, will have a positive effect on social spending.
2.3. Political Assumptions: Cumulative Democracy and Voter Turnout
Both PRT and income distribution explanations presuppose that a country has a democratic
form of government. However, they differ in their conception of how democracy operates.
Whereas PRT conceptualizes democracy as working over a longer period of time as left-leaning
parties use their time in government to gradually increase spending, coverage and redistribution;
income distribution explanations assume the effect of democracy is more or less instantaneous:
voters are seen as explicitly using elections to set the level of redistribution. Thus, it is important
to formulate hypotheses that capture this difference.
The initial formulations of PRT simply assumed that politics was democratic, a reasonable
assumption in post-World War II Western Europe. But subsequent theorizing has explicitly
acknowledged the importance of democracy. Democracy is seen as affecting social spending
through two channels: electoral competition, which provides politicians with incentives to deliver
popular social programs; and, interest group freedom, which allows groups to lobby and, if
necessary, protest on behalf of their preferred policies (Haggard and Kaufman 2008). Thus, at the
very least, a democratic government is a necessary pre-condition for PRT. But a complementary
argument pushes the point further. The full effect of democracy on social spending might not be
apparent in the short-run. Rather, spending levels are the cumulative result of years of political
competition and contestation (Gerring, Thacker and Alfaro 2010; Huber, Mustillo and Stephens
2008; Huber and Stephens 2012). Thus, a test of PRT in the context of developing societies should
acknowledge its scope condition, by making democracy a case selection criteria, and the fact that
even a dominantly democratic region such as Latin America has both old and very young
Social Policy and Presidential Ideology in Latin America
68
democracies, by addressing a second hypothesis: a country’s accumulated stock of democracy
should have a positive effect on social spending levels.
The political assumption of income distribution explanations draws attention to voter
turnout, which is seen as magnifying or suppressing the effect of income distribution on
redistribution and spending. Around the world, the poor are less likely to vote than the rich. If the
left’s natural constituency (the poor) participates in elections at a lower rate than the rich, its
preferences will be less like to be taken into account and social spending and redistribution will be
lower than predicted by the MVT. Thus, voter turnout acts as an inverse proxy for the income bias
that prevails in most elections (Lupu and Pontusson 2011; Mahler 2008). Based on this, countries
with higher turnout are expected to have higher social spending. This same logic should apply to
the structure of inequality hypothesis.
45

3. Empirical Methods
To assess the two theoretical approaches outlined above, this chapter conducts a time-series
cross-sectional regression analysis of the determinants of social spending in 18 Latin American
democracies during 1990-2010. Before turning to the analysis, however, a few words are in order
concerning case selection, data, control variables, and analytical techniques.
3.1. Case Selection
The choice of Latin America obeys three broad reasons. Latin America is a developing
region and thus serves to counterbalance the heavy focus on Western Europe in the welfare state
literature. Indeed, as Carnes and Mares (2007, 868-69) assert, “the most exciting research
opportunities in the study of welfare states lie in examining the variation in policies of social
protection in developing economies.” Furthermore, Latin America stands out among developing
Social Policy and Presidential Ideology in Latin America
69
regions as the most democratic and thus complies with the scope condition for the theories tested:
that the country under consideration is a democracy. Finally, though data availability does impose
some restrictions, they do allow for a fairly comprehensive analysis, covering 18 countries during
21 years (for an N of 302).
46

The region also exhibits substantial variation in terms of both the key explanatory variables
– presidential ideology, income distribution and democratic trajectory – and the dependent variable
– social spending levels. With regard to ideology, many of the region’s countries have in recent
years elected left-of-center governments that campaigned on promises of higher social spending
and income redistribution (Castañeda 2006; Flores-Macías 2012; Levitsky and Roberts 2011;
Weyland, Madrid and Hunter 2010). However, the variation in policy choices made by this recent
wave of left-of-center governments is such that scholars and commentators speak of two lefts: a
moderate, social democratic variety, which shares the goals and values of the left in industrialized
countries, and a more radical or populist variety, for which there is no clear equivalent among rich
countries.
47
With regard to income distribution, despite notable improvements over the last decade
(López-Calva and Lustig 2010), Latin America remains the world’s most unequal region. Regional
averages, however, mask substantial cross-national variation in both the level and structure of
inequality. The region’s countries also exhibit substantial variation in terms of democratic
trajectory, running the gamut from Costa Rica, which has been consistently democratic since 1949,
to Mexico, which democratized in 2000, to countries that experienced democratic breakdowns
during the period of study (Honduras and Peru).
There is also a great deal of variation in terms of welfare state trajectory and spending
levels. In the 1920s, Southern Cone countries were the first developing countries to establish
comprehensive pension systems (Mesa-Lago 1989) and by 1980 several of the region’s countries
Social Policy and Presidential Ideology in Latin America
70
had full-fledged welfare states (Huber and Stephens 2005, 554). Within-region variation in
spending is jarring. According to the United Nations Economic Commission for Latin America
(ECLAC), average total social spending as a share of GDP during 2005-2009 ranged from almost
25 percent in Brazil to less than 5 percent in Ecuador. Whereas countries in the Southern Cone
spend roughly $1,200 per person per year, the poorer Andean and Central American countries
spend less than $200. In brief, Latin America is a well-suited region in which to test and refine
theories developed in Western Europe.
3.2. Data
Turning to the data used in the statistical analysis, a few comments are in line with regard
to the data used for the dependent and explanatory variables. This study’s main dependent variable
is total social spending measured as a percentage of a country’s GDP. Additional tests will assess
the determinants of spending in per capita 2005 dollars and spending on education, healthcare and
social security as a share of GDP. These dependent variables were all compiled by ECLAC’s
CEPALSTAT database (ECLAC 2012). These are not perfect measures; indeed, the problems with
using social spending as a proxy for welfare state development are well known. Spending levels
tell us nothing about policy effectiveness and little about how funds are prioritized or targeted.
And even the most disaggregated spending data are still highly aggregated, hiding differences
between programs that primarily benefit the poor and those that benefit the well-off. Regardless
of these problems, spending data are widely available, cover a large sample of countries over an
extended period of time, and, ultimately, constitute a reasonable proxy for social policy.
Most of the measures of explanatory variables are well known and the summary in Table
3-1 is largely self-explanatory. Other variables (italicized in what follows) deserve some
discussion. Left-of-Center President is measured using a dummy variable coded 1 for presidents
Social Policy and Presidential Ideology in Latin America
71
labeled as “center-left” or “left.” Similarly, Center-Left President is measured using a dummy
variable coded 1 for presidents labeled as “center-left” and Far-Left President is measured using
a dummy variable coded 1 for presidents labeled as “left.” These data, which are derived from
consultations with country experts, are part of Dataset on Political Ideology of Presidents and
Parties in Latin America (Murillo, Oliveros and Vaishnav 2010).
48

Table 3-1: Key Explanatory Variables
Theory/Variables Measure and Source

Power Resource Theory
Left-President
President’s ideology is “center-left” or “left” (Murillo, Oliveros and
Vaishnav 2010).
Center-Left President
President’s ideology is “center-left” (Murillo, Oliveros and Vaishnav
2010).
Far-Left President President’s ideology is “left” (Murillo, Oliveros and Vaishnav 2010).

Income Distribution Explanations
Gini Coefficient Gini coefficient (CEPALSTAT).
Income Share: Rich to Poor
Ratio of the income share of quintile five to quintile one
(CEPALSTAT).
Income Distribution Skew
Ratio of the income share of the ninth decile to the fifth divided by the
ratio of the income share of the fifth decile to the first (CEPALSTAT).
Income Share:
Rich to Median Voter
Ratio of the income share of the ninth decile to the fifth (CEPALSTAT).
Income Share:
Median Voter to Poor
Ratio of the income of the fifth decile to the first (CEPALSTAT).

Political Assumptions
Cumulative Democracy Years of democracy since 1946 (Cheibub, Gandhi and Vreeland 2010).
Voter Turnout Voter turnout relative to the voting age population (IDEA).

Since social spending does not necessarily imply redistribution, spending is not the ideal
dependent variable for testing the MVT. In fact, Latin America’s social security-centered social
policy is not very redistributive (Breceda, Rigolini and Saavedra 2009; Goñi, López and Servén
2011) and some programs are even regressive or “reverse Robin Hood” (Lindert, Skoufias and
Shapiro 2006). The ideal measure for assessing the determinants of redistribution compares pre-
and post-tax inequality, as is done in recent studies of OECD countries. However, this kind of data
are not available cross-nationally over an extended period of time for Latin American countries.
Social Policy and Presidential Ideology in Latin America
72
Despite these limitations, it is still worthwhile to assess whether the level and structure of
inequality affect social spending.
49

3.3. Control Variables
This study relies on a number of standard control variables, which are related to the initial
theory about the impact of the logic of industrialization on social policy as well as some more
recent theories that can be considered as alternatives to PRT and income distribution explanations.
In brief, the control variables and data sources for these variables are as follows.
Logic of Industrialism: Economic development is operationalized as GDP per capita in
constant 2005 international dollars. Demographic controls capturing the shares of the population
living in urban areas, over 65 and under 15 are also included. These variables are from the World
Bank’s World Development Indicators (WDI).
Taxation: Timmons (2005, 531) argues that it is “fundamentally flawed” to study spending
and taxation in isolation from one another. Government spending reflects the interests of those
who pay for it. Since much of the burden of consumption taxes falls on the poor, countries that are
dependent on these taxes, such as those in Latin America, should be expected to spend more on
policies that benefit the poor, namely social policy. Timmons also argues that, because the vast
majority of the income tax burden falls on wealthier individuals who have little need for social
policy, a country’s share of income taxes should have no effect on social spending. This should
hold true in Latin America, where high minimum exemption levels and a plentiful deductions limit
the income tax base to the wealthiest individuals (Goñi, López and Servén 2011, 1562).
Consumption taxes and income taxes are respectively measured as indirect and direct taxes as a
share of GDP. The models also control for social security payroll taxes relative to GDP. These
variables are compiled by CEPALSTAT.
Social Policy and Presidential Ideology in Latin America
73
Ethnic divisions: Ethnic divisions may reduce support for redistribution and thus social
spending because tax payers may be less willing to finance spending that benefits people different
from themselves (Luttmer 2001). Alesina and Glaeser (2004) attribute much of the size difference
between U.S. and European welfare states to ethno-racial fragmentation. In addition, ethnically
fragmented societies provide fewer public goods (Alesina, Baqir and Easterly 1999; Easterly and
Levine 1997; Lieberman 2009). Thus, there is reason to expect that ethnically diverse Latin
American countries will have lower levels of social spending than more homogenous ones. Ethnic
divisions are measured using Lieberman’s (2009, 47) ethnic boundary index.
International Constraints: Political decisions regarding budgets and spending in
developing countries that are highly integrated to the global economy, such as those in Latin
America, should be highly responsive to international factors. There has historically existed a
positive correlation between economic openness and government spending among developed
countries, which has been interpreted as an attempt by governments to compensate workers
exposed to international competition (Cameron 1978). Garrett (2001) contrasts this “compensation
hypothesis” with the “efficiency hypothesis” – the view that increased international competition
under globalization has unleashed a “race to the bottom” in terms of social policy (Rudra 2002).
Openness to trade is measured as exports plus imports as a share of GDP. The models also control
for current account balance as a share of GDP and debt service as a share of gross national income.
These three variables are from the WDI.
Veto Players: Veto players are the political actors that must agree with a proposed policy
change in order for it to be enacted. As the number of veto players increases, the probability of
policy change decreases. The effect of veto players on social spending is ambiguous: they reduce
the probability of both major social policy expansion and retrenchment. The goal in including this
Social Policy and Presidential Ideology in Latin America
74
variable is to control for the region’s institutional diversity. Veto players are measured using an
index ranging from 0 to 3. Countries are coded as 0 if they have unitary government 1 for unitary
government/bicameral legislature or federalism/unicameral legislature 2 for federalism/bicameral
legislature, and 3 for strong federalism/bicameral legislature.
3.4. Analytic Techniques
Finally, the technique used in this study requires some discussion. As has been discussed,
the type of pooled time-series cross-sectional data used in research on the determinants of social
spending violates several basic assumptions of ordinary least squares regression (Hicks 1994, 172).
The data are often: 1) temporally auto-regressive (a country’s spending level during a given year
is largely determined by its spending level during the previous year); 2) cross-sectionally
heteroskedastic (errors have different variances across units and these variances tend to be
correlated with the size of the unit studied – meaning that error terms for countries with high
spending levels tend to be larger than for countries with low spending levels), and, 3) cross-
sectionally correlated (spending in one country may influence spending in another). Furthermore,
these data may: 4) conceal unit and period effects (spending may be influenced by conditions
unique to a country or a time period); and, 5) reflect causal heterogeneity across time, space or
both (the causal process that explains the dependent variable may not be the same for all
observations in the sample).
Taking these issues into account, this analysis follows Beck and Katz (1995) in estimating
the regression parameters using ordinary least squares with panel-corrected standard errors.
However, the approach followed differs from their recommendations in two ways: I do not use a
lagged dependent variable to address serial auto-correlation or rely on country fixed effects to deal
with omitted variable bias. A lagged dependent variable is not included for fear that doing so will
Social Policy and Presidential Ideology in Latin America
75
suppress the explanatory power of the independent variables (Achen 2001) and possibly bias
estimates (Plümper, Troeger and Manow 2005). Instead, the models correct for first-order-auto-
regression (ar1).
Country fixed effects are not used for fear that doing so may eliminate too much of the
cross-sectional variance, make it impossible to estimate the effect of time-invariant variables (e.g.
the ethnic boundaries index), and severely bias estimates of partly invariant variables (e.g. the veto
player index) (Beck 2001). Furthermore, Plümper, Troeger and Manow (2005) warn that the use
of country dummies may generate inefficient and potentially biased estimates if, as is the case with
my data, most of the variance in the dependent variables already existed during the first period of
observation. Regardless, fixed effects will be used as a robustness check.

4. Tests and Results
This section first presents the results of a series of baseline models that incorporate only
the two main explanatory variables – presidential ideology and income distribution– and control
for per capita income. A wide range of control variables are then incorporated into the models.
Then, as robustness checks, the models are re-tested using an alternative dependent variable and
country fixed effects. Finally, to ascertain the areas of spending the explanatory variables work
through, total spending is disaggregated into education, healthcare and social security spending.
4.1. Baseline Models
The analysis begins by estimating stripped-down baseline models. This decision was
influenced by Achen’s (2002) Rule of Three, which recommends that statistical models not
grounded in a well-developed formal theory include no more than three explanatory variables. At
the risk of omitted variable bias, this serves to show that the findings are not dependent on other
Social Policy and Presidential Ideology in Latin America
76
variables included in the models. Given the high likelihood of omitting a key variable, discussion
of these findings will be limited to their statistical significance. Discussions of effect magnitude
are reserved for more complete models. The findings of these models are presented in Table 3-2.
Table 3-2: Total Social Spending to GDP (Baseline Models)
Fixed Effects? NO NO NO NO YES

Exp. Variable (1-1) (1-2) (1-3) (1-4) (1-5)

Left President 0.653+ 0.666+ 0.684+ 0.711+ 0.939**
(0.357) (0.353) (0.358) (0.367) (0.327)
Gini Coefficient -0.481 ----- ----- ----- -----
(1.327)
Quintile5/Quintile1 ----- 0.086 ----- ----- -----
(0.060)
Decile9/Decile5 ----- ----- 1.779* ----- -----
(0.869)
Decile5/Decile1 ----- ----- -0.315* ----- -----
(0.143)
Income Dist. Skew ----- ----- ----- 2.669* 2.500*
(1.241) (1.019)
GDP per Capita 0.001*** 0.001*** 0.001*** 0.001** 0.001***
(0.000) (0.000) (0.000) (0.000) (0.000)
Constant 11.639*** 10.118*** 7.617** 9.551*** 5.787***
(0.850) (0.992) (2.661) (1.014) (1.031)

R2 0.187 0.184 0.200 0.202 0.731
N 302 302 302 302 302
Notes: +p <=0.10 *p <=0.5; ** p <=0.01; *** <=0.001 in two-tailed test.

With regard to PRT, having a left president has a small positive effect on spending levels
in most models, albeit at the 90 percent confidence level. Only once fixed effects are included
(model 1-5) does the coefficient reach standard significance levels. Disaggregating the measure of
left president into its two component parts (center-left and far-left) reveals that the effect of
ideology is driven entirely by the small subset of far-left presidents, namely Hugo Chávez in
Venezuela (1998-2006), Evo Morales of Bolivia (2006-2010) and Rafael Correa of Ecuador (2007-
2010), which represent just 21 country-years (roughly 7 percent of total cases and 21 percent of
cases of left government).
50
When the effects on spending of the two lefts are analyzed separately,
only far-left president has a significant positive effect.
51

Social Policy and Presidential Ideology in Latin America
77
Another notable finding is that spending levels are influenced not by the level of inequality
but by its structure. In contrast to the predictions of the MVT, the Gini coefficient, which is not
statistically significant, has a negative sign (model 1-1). Spending increases in line with the ratio
between the income shares of decile 9 and decile 5 and decreases in line with the ratio of the
income shares of decile 5 to decile 1 (model 1-3). Similarly, the ratio of these two measures,
income distribution skew, has a positive and significant effect on spending (model 1-4). This
finding is robust to the inclusion of country fixed effects (model 1-5).
4.2. Full Models
The preceding analysis will now be replicated incorporating additional variables of
theoretical interest, such as democratic history and voter turnout, as well as a wide range of control
variables. Table 2-3 summarizes the main results of these models.
The analysis fails to find an effect on spending of having a president of the left (models 2-
1 and 2-3). Disaggregating this measure confirms that, although center-left presidents do not affect
spending, far-left president do. Countries governed by the far-left are expected, all else equal, to
spend an additional 1.8 percentage points. This effect is substantively quite large, amounting to 15
percent of the average level of spending during the period of study (models 2-2 and 2-4).
Incorporating the additional variables only serves to strengthen the relationship between
income distribution skew and spending (models 3-3 and 3-4). The relationship is now significant
at the 99 percent level and the coefficient increases in size substantially from 2.5-2.7 in the baseline
models to 3.4-3.5 in the full models. Holding other variables constant, moving from a sample
minimum skew (0.34) to the sample mean (0.68) increases total social spending by 1.2 percentage
points (10.0 percent of the mean spending level) (see figure 2-1). Moving from the sample mean
to the sample maximum (0.99) increases spending by 1.05 percentage points (8.7 percent of mean

Social Policy and Presidential Ideology in Latin America
78
Table 3-3: Total Social Spending to GDP (Full Model)
Fixed Effects? NO NO NO NO

Exp. Variable (2-1) (2-2) (2-3) (2-4)

Left President 0.485 ----- 0.591 -----
(0.366) (0.373)
Center-Left President ----- 0.103 ----- 0.218
(0.365) (0.368)
Far-Left President ----- 1.811* ----- 1.859*
(0.838) (0.793)
Gini Coefficient -1.336 -1.051 ----- -----
(1.452) (1.447)
Income Dist. Skew ----- ----- 3.483** 3.390**
(1.108) (1.071)
Cum. Democracy 0.150*** 0.147*** 0.162*** 0.159***
(0.017) (0.018) (0.019) (0.019)
Voter Turnout 0.998 0.599 1.237 0.837
(1.285) (1.276) (1.302) (1.288)
Consumption Taxes 0.411*** 0.394*** 0.386*** 0.368***
(0.107) (0.106) (0.101) (0.100)
Income Taxes 0.064 0.076 0.080 0.092
(0.055) (0.055) (0.058) (0.056)
Social Security Taxes 0.720*** 0.724*** 0.717*** 0.722***
(0.105) (0.102) (0.108) (0.105)
Ethnic Diversity -5.987*** -6.187*** -6.407*** -6.580***
(1.179) (1.164) (1.169) (1.147)
Openness to Trade -0.018* -0.019* -0.018* -0.018*
(0.008) (0.007) (0.008) (0.007)
Current Account Bal. 0.006 -0.001 -0.003 -0.009
(0.020) (0.020) (0.021) (0.020)
Debt -0.016 -0.008 -0.015 -0.007
(0.031) (0.028) (0.031) (0.028)
Pop. over 65 0.382** 0.460*** 0.383*** 0.456***
(0.120) (0.129) (0.112) (0.120)
Pop. under 15 -0.119 -0.143+ -0.092 -0.117
(0.079) (0.076) (0.090) (0.085)
Urban Population -0.142*** -0.163*** -0.133*** -0.153***
(0.029) (0.032) (0.029) (0.030)
GDP per Capita 0.000 0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000)
GDP Growth -0.010 -0.009 -0.005 -0.004
(0.016) (0.016) (0.015) (0.015)
Veto Players 2.984*** 3.082*** 2.911*** 3.007***
(0.334) (0.358) (0.292) (0.318)
Constant 13.412** 15.462*** 8.552+ 10.789*
(4.440) (4.435) (5.000) (4.770)

R2 0.625 0.625 0.631 0.630
N 299 299 299 299
Notes: +p <=0.10 *p <=0.5; ** p <=0.01; *** <=0.001 in two-tailed test.

Social Policy and Presidential Ideology in Latin America
79
spending). In contrast, the Gini coefficient remains insignificant and negatively signed (models 2-
1 and 2-2).
Figure 3-1:
Estimated Effects of Income Distribution Skew
on Total Social Spending to GDP 1990-2010 (Model 2-3)

Source: Table 2-3, Model 2-3.
Notes: All other variables kept at their listwise mean.

With regard to the other explanatory variables, whereas democratic trajectory has a positive
and highly significant effect in the four models, voter turnout fails to achieve significance in any
of the models. An additional year of democracy increases spending by 0.15 percentage points.
There is also support for the taxation social contract hypothesis (Timmons 2005).
Consumption taxes, which fall heavily on the poor, are associated with increased social spending.
Moving from a sample minimum level of consumption taxes (2.07 percent of GDP) to the sample
mean (8.06 percent of GDP) increases spending by 2.3-2.5 percentage points (between 19.0 and
20.7 percent of the mean spending level). Moving from the sample mean to the sample maximum
(13.86 percent of GDP) increases spending by 2.2-2.4 percentage points (between 18.2 and 19.8
percent of mean spending). In contrast, income taxes have no effect on spending levels.
9
10
11
12
13
14
15
0.34
0.37
0.40
0.43
0.46
0.49
0.52
0.55
0.58
0.61
0.64
0.67
0.70
0.73
0.76
0.79
0.82
0.85
0.88
0.91
0.94
0.97
95% CI Effect of Income Distribution Skew 95% CI
TotalSocial Spending as % of GDP
IncomeDistribution Skew
SampleMean
Social Policy and Presidential Ideology in Latin America
80
Finally, the evidence backs the hypothesis that more ethnically divided countries spend
less on social policy. Ethnic diversity has a very large negative effect on spending. All things equal,
homogenous countries like Argentina or Costa Rica are expected to spend an additional 4.5-4.7
percentage points more than divided countries such as Ecuador or Guatemala. This effect, which
is significant at the 99.9 level, amounts to between 37.2 to 38.9 percent of the sample mean.
4.3. Robustness Checks
To assess the robustness of the previous findings, the models tested in this section
incorporate country fixed effects and an alternative dependent variable: total social spending per
capita in 2005 dollars (see Table 3-4). Whereas spending as a share of GDP captures the relative
importance of social spending within a national economy, spending per capita measures the
absolute value of resources transferred to beneficiaries (Kaufman and Segura-Ubiergo 2001, 561).
Under fixed effects, having a far-left president ceases to have a significant effect on social
spending as a share of GDP (model 3-1). However, having a far-left president has a positive and
significant effect on per capita social spending, regardless of whether fixed effects are included.
All else equal, countries with far-left presidents are expected to spend an additional $73.86 per
individual (about 15.5 percent of the mean spending level of $476) (model 3-2). This effect is
substantially smaller when fixed effects are incorporated ($39.41).
Income distribution skew has a positive effect on spending as a share of GDP under fixed
effects and on spending in per capita terms. All else equal, moving from the minimum to the
average skew increases spending by $39.35 or 8.3 percent of the mean spending level. Moving
from the mean to the maximum level increases spending by about $36.00 (7.6 percent of the mean
level). However, introducing fixed effects substantially reduces the size of the estimated
coefficients (from 3.39 to 2.15 percentage points of GDP and from $115.73 to $78.59).

Social Policy and Presidential Ideology in Latin America
81
Table 3-4: Total Social Spending (Robustness Checks)

Total Social
Spending/GDP
Total Social
Spending P.C.
Total Social
Spending P.C.
Fixed Effects? YES NO YES

Exp. Variable (3-1) (3-2) (3-3)

Center-Left President 0.256 15.919 12.915
(0.292) (12.706) (12.362)
Far-Left President 0.554 73.861** 39.413**
(0.405) (22.676) (13.160)
Income Dist. Skew 2.150** 115.734** 78.592*
(0.801) (44.802) (35.407)
Cum. Democracy 0.360*** 6.378*** 11.149**
(0.066) (0.066) (3.713)
Voter Turnout 0.596 -19.598 -88.121*
(0.989) (40.340) (39.199)
Consumption Taxes 0.091 11.545*** 0.208
(0.077) (3.437) (2.677)
Income Taxes 0.198*** 5.547* 5.316*
(0.057) (2.811) (2.267)
Social Security Taxes 1.206*** 30.803*** 49.134***
(0.160) (4.446) (7.276)
Ethnic Diversity ----- -307.576*** -----
(0.603)
Openness to Trade -0.009 -1.335*** -1.115***
(0.009) (0.224) (0.269)
Current Account Bal. -0.083*** -0.846 -2.154***
(0.017) (0.767) (0.631)
Debt -0.007 0.485 0.798
(0.025) (1.128) (1.159)
Pop. over 65 0.954* 35.049*** 142.429***
(0.403) (5.773) (33.946)
Pop. under 15 0.051 -4.139 8.564
(0.110) (3.146) (5.999)
Urban Population -0.247*** -11.393*** -19.067***
(0.073) (0.969) (3.362)
GDP per Capita -0.001** 0.085*** 0.060***
(0.000) (0.006) (0.009)
GDP Growth -0.007 -0.437 0.123
(0.017) (0.639) (0.695)
Veto Players -1.920** 134.215*** -107.130***
(0.624) (13.023) (22.804)
Constant 11.844 719.324*** 789.231+
(10.064) (149.795) (429.512)

R2 0.93 0.819 0.910
N 299 299 299
Notes: +p <=0.10 *p <=0.5; ** p <=0.01; *** <=0.001 in two-tailed test.
Country dummies not shown.

Social Policy and Presidential Ideology in Latin America
82
Democratic history retains a positive and significant effect in these models. Contrary to
expectations, voter turnout is negatively signed in regressions on per capita spending. This effect
reaches statistical significance under fixed effects (model 3-3).
The most notable difference between these and previous models involves the taxation
variables. Against Timmons’ (2005) predictions and findings, in both models of the determinants
of per capita spending, income taxes as a share of GDP have a statistically significant positive
effect on spending. The same is true when fixed effects are incorporated to the model of total social
spending as a share of GDP (model 3-1). In addition, consumption taxes cease having a significant
effect on total spending (as a share of GDP or in per capita dollars) under fixed effects. Timmons’
(2005) original paper did not test the robustness of his findings to the use of fixed effects. In light
of this evidence, it may be worthwhile to re-examine his original findings.
4.4. Disaggregated Spending
Finally, to uncover the channels through which the explanatory variables affect total social
spending, the dependent variable is disaggregated into its component parts: education, healthcare
and social security (as a share of GDP) (see Table 3-5). The most striking finding is that the effect
of ideology on total spending is driven almost entirely by education spending (model 4-1). All else
equal, having a far-left president increases education spending by 1.85 percentage points or 47.8
percent of the mean level of education spending (3.87 percent of GDP)! The magnitude of this
effect is nearly identical to the effect on total social spending of having a far-left president, a sign
that the latter effect is driven primarily by education. Furthermore, the effect on education spending
of having a center-left president is also positive and significant, albeit at the 90 percent level.
Ideology has no effect on healthcare (model 4-2) or social security (model 4-3) spending.

Social Policy and Presidential Ideology in Latin America
83
Table 3-5: Disaggregated Spending to GDP

Education
Spending/GDP
Healthcare
Spending/GDP
Soc. Security
Spending/GDP
Fixed Effects? NO NO NO

Exp. Variable (4-1) (4-2) (4-3)

Center-Left President 0.298+ 0.082 -0.071
(0.160) (0.103) (0.193)
Far-Left President 1.847*** 0.250 -0.104
(0.382) (0.217) (0.298)
Income Dist. Skew 1.318* -0.351 2.006**
(0.650) (0.322) (0.651)
Cum. Democracy 0.028*** 0.030*** 0.079***
(0.006) (0.007) (0.009)
Voter Turnout -0.047 -0.046 1.222*
(0.567) (0.426) (0.584)
Consumption Taxes 0.190*** 0.076* 0.200***
(0.046) (0.034) (0.042)
Income Taxes 0.029 0.020 -0.016
(0.033) (0.019) (0.038)
Social Security Taxes 0.150** 0.153** 0.426***
(0.053) (0.054) (0.049)
Ethnic Diversity -1.080** -2.233*** -3.826***
(0.406) (0.338) (0.521)
Openness to Trade 0.004 -0.007* -0.017***
(0.003) (0.003) (0.004)
Current Account Bal. -0.011 -0.007 0.011
(0.008) (0.006) (0.011)
Debt 0.004 0.005 -0.009
(0.009) (0.009) (0.014)
Pop. over 65 -0.192** 0.098* 0.401***
(0.061) (0.047) (0.069)
Pop. under 15 -0.035 -0.060* -0.077*
(0.039) (0.028) (0.036)
Urban Population -0.052*** -0.086*** -0.017
(0.014) (0.011) (0.012)
GDP per Capita 0.000* 0.000** -0.000***
(0.000) (0.000) (0.000)
GDP Growth -0.007 0.001 0.001
(0.007) (0.004) (0.008)
Veto Players 0.621*** 0.633*** 1.904***
(0.114) (0.123) (0.177)
Constant 5.262* 8.537*** -0.607
(2.168) (1.559) (2.251)

R2 0.355 0.458 0.793
N 299 299 286
Notes: +p <=0.10 *p <=0.5; ** p <=0.01; *** <=0.001 in two-tailed test.
Nicaragua does not report social security expenditures.

Social Policy and Presidential Ideology in Latin America
84
The disaggregated models also reveal that the effect of income distribution skew on total
spending operates through education and social security, but not through healthcare. Democratic
history has a positive and significant effect on the three forms of spending. And, in line with
Kenworthy and Pontusson’s (2005) predictions, voter turnout has a significant positive effect on
social security. The effects of the other variables of interest do not vary with the type of spending.
Table 3-6: Findings on the Determinants of Social Spending
Theory/Variable
Expected
Outcome
Finding Notes

Power Resource Theory
Left-Leaning President + +/None
Significant, but generally only at the
90% level. Operates through
education.
Center-Left President + None
Positive effect at the 90% level on
education.
Far-Left President + + Operates through education spending.
Income Distribution Explanations
Gini Coefficient None None Has a negative sign.
Income Share:
Rich to Poor
None None
Income Distribution Skew + +
Operates through education and
social security spending.
Income Share:
Rich to Median Voter
+ +
Significant at the 90% level in some
models.
Income Share:
Median Voter to Poor
- -

Political Assumptions
Cumulative Years of
Democracy
+ +
Voter Turnout + None
Negative effect on total spending PC
under FE. Positive effect on social
security spending.

Taxation Social Contract Hypothesis
Consumption Taxes + +/None
Positive and significant, except under
FE. Insignificant under FE.
Income Taxes None None/+
Insignificant, except under FE.
Positive and significant under FE.
Ethnic Divisions
Ethnic Boundaries Index - -

Social Policy and Presidential Ideology in Latin America
85
5. Conclusion: Implications of the Findings
This chapter has sought to broaden the discussion of the politics of social policy beyond
advanced industrialized states by testing key hypotheses concerning the determinants of total social
spending and its component parts – education, healthcare and social security spending – through a
time-series cross-sectional analysis of 18 Latin American democracies. The overall findings of this
study are summarize in Table 3-6. In a nutshell, this chapter finds that explanations based on
income distribution constitute a more consistent predictor of variation in social spending in post-
market reform Latin America than explanations based on presidential ideology.
The first major finding is that there is weak support for power resource explanations (Korpi
1983; Stephens 1979). With regard to ideology, there is some evidence that social policy
expenditures are higher under left-of-center presidents than under centrist and rightist presidents.
However, this effect is largely explained by a small subset of cases with far-left presidents, namely
Venezuela under Chávez, Bolivia under Morales and Ecuador under Correa. Indeed, having a far-
left president has a large positive effect on total social spending. More interestingly, this effect
operates almost exclusively through education spending. All else equal, having a far-left president
is estimated to increase education spending by nearly 50 percent!
These findings challenge recent work by Birdsall, Lustig and McLeod (2012, 168), which
claims that the social democratic left tended to outspend its populist counterpart.
52
The finding that
the left’s effect on spending operates primarily through education, the most progressive spending
category, challenges earlier work that found that presidents of the left tended to cut spending on
human capital (Huber, Mustillo and Stephens 2008; Kaufman and Segura-Ubiergo 2001; Segura-
Ubiergo 2007).
53
Thus, it appears that far-left governments have, so far, kept their promise of
increasing social spending. This increase has operated through education spending, which, in
Social Policy and Presidential Ideology in Latin America
86
contrast to the social security spending (Lindert, Skoufias and Shapiro 2006), has been shown to
reduce both poverty and inequality (Birdsall, Lustig and McLeod 2012, 167).
The limited impact of ideology on spending raises doubts regarding how well PRT travels
to the developing world. Given its long history of labor-affiliated parties, Latin America should be
an easy context in which to study the effects of ideology on spending. Additionally, over the last
decade, much of the region has been consistently governed by the left. However, the effect of
ideology on spending is explained not by the social democratic left, which shares the values of its
rich-country counterparts, but rather by the populist left, which does not fit neatly into the
categories laid out by PRT. While espousing redistribution like its social democratic cousin, the
populist left is personalistic, rather than centered on strong programmatic parties. Its social base is
the poor employed in the informal sector, not the traditional working class.
54
Whereas moderates
have long-standing alliances with organized labor, radicals have had looser ties with labor and these
have soured as leaders consolidated and expanded their power. Interpreting spending levels under
Latin America’s far-left as confirmation of PRT involves a fair amount of conceptual stretching.
The second major finding is that there is robust evidence that income distribution
influences spending. However, what matters is the structure of inequality, not its level. Echoing
Lupu and Pontusson’s (2011) findings for industrialized countries, there is strong and robust
evidence that the skew in the distribution of income has a positive effect on social spending levels
among Latin American countries. Citizens near the median income level appear more likely to ally
with the poor in favor of spending when the distance between their share of a country’s income
and that of the poor is small relative to the distance between their share of the income and that of
the rich. Tests on disaggregated measures of spending reveal that skew’s effect operates through
Social Policy and Presidential Ideology in Latin America
87
education and social security spending. Thus, inequality does affect social spending through the
channel of the median voter, but not in the way that Meltzer and Richards (1981) posited.
The failure to find an effect of level of inequality on social spending is consistent with
existing evidence from industrialized countries. A cursory look at OECD countries reveals that the
most egalitarian countries engage in the most redistribution. Cross-national studies controlling for
a variety of confounding variables provide limited evidence of Meltzer and Richard’s predictions
(Kenworthy and McCall 2008; Lübker 2007). Moene and Wallerstein (2003) even find a negative
relationship between earnings inequality and spending. Survey research on Latin America has
shown that the poor are no more likely than the non-poor to vote for higher taxes or to think that
the distribution of income is unfair (Kaufman 2009).
The robust evidence of the applicability to Latin America of the structure of inequality
hypothesis complements the strong evidence of this relationship found among advanced
industrialized countries (Lupu and Pontusson 2011). Future research should test whether this
finding holds in other regions and cross-regionally. Given its less rigid and precise actor categories
– income strata rather than organized and self-aware classes – less conceptual stretching is required
when applying income distribution explanations to developing regions. But this strength is also a
weakness. Further theorizing is needed to identify the key actors and flesh out the causal
mechanisms though which alliances between the citizens near the median income level and the
poor (the rich) act to increase (decrease) social spending and income redistribution.
Additionally, this chapter’s findings support past research on how democracy and ethnic
fragmentation influence social spending. Findings on the cumulative effect of democracy reinforce
earlier research on spending (Huber, Mustillo, and Stephens 2008; Huber and Stephens 2012) and
human development (Gerring, Thacker and Alfaro 2012).
55
The robust negative effect of ethnic
Social Policy and Presidential Ideology in Latin America
88
divisions on spending supports work by Huber and Stephens (2012) that finds that ethnic diversity
negatively affects healthcare spending. But, in contrast to their work, my research also finds a
negative effect on education and social security spending. Ardanaz (2009) finds a negative
correlation between ethnic fractionalization and individual-level support for inequality-reducing
policies in the region. My findings provide evidence of an empirical implication of his work.
Research on the politics of social policy has generated substantial cumulative knowledge.
Amenta (2003) attributes this success to the combination of a consensus on the scope of the
phenomenon to be explained and disagreement over possible explanations. However, work on this
research agenda has been heavily biased in favor of advanced industrialized countries. In recent
years, thanks to increased data availability, research on developing countries started to take off.
By contrasting the applicability to developing countries of two of the leading explanations of social
policy in industrialized countries, this chapter has sought to reduce this geographic bias. Applying
these theories to a different context allows us to examine them under a new light. Doing so will
help us continue to advance one the most productive research agendas in comparative politics.
The remainder of this study will focus on the politics of combating poverty through
conditional cash transfer programs, which constitute “the most important innovation in social
policy of the last fifteen years” in Latin America (Lustig 2011, 19).
Social Policy and Presidential Ideology in Latin America
89
Chapter 4.
Does the Left Actually Redistribute More?:
Adoption and Scope of Conditional Cash Transfer Programs 1996-2010

1. Introduction
The last two decades have witnessed an unprecedented decline in poverty levels among
developing countries (Chen and Ravallion 2012). This decline is in part explained by the rapid
spread of large-scale anti-poverty programs, which, it is estimated, are responsible for preventing
150 million people from falling into poverty (Fiszbein, Kabur and Yemtsov 2014). Latin America
has been at the forefront of this trend. Beginning in the late 1990s, governments in the region
started adopting innovative programs that award poor families stipends conditional on their
children remaining in school and attending regular medical checkups. These conditional cash
transfer (CCT) programs, which now cover nearly a quarter of the region’s population (Stampini
and Tornarolli 2012), constitute “the most important innovation in social policy of the last fifteen
years” (Lustig 2011, 19). In addition to increasing school enrollment (Baird et al., 2013; Saavedra
& García, 2012), CCTs are credited with helping pull 15.5 million Latin Americans – about 13
percent of the region’s population – out of poverty (Stampini & Tornarolli, 2012).
56
They have
also contributed to the recent decline of income inequality in what has historically been the world’s
most unequal region (López-Calva & Lustig, 2010; Lustig, López-Calva & Ortíz-Juárez, 2013;
Soares et al., 2009). And their success in Latin America has led countries in settings as diverse as
Sub-Saharan Africa, Southeast Asia, and even New York City to emulate these programs (Fiszbein
& Schady, 2009; Hanlon, Barrientos & Hulme, 2010).
Social Policy and Presidential Ideology in Latin America

90
The effect of CCTs has been well studied and this research supports the claim that they
help to reduce poverty. But the emergence of these policies in Latin America is a puzzle.
Historically and still today, Latin American social spending has been dominated by social
insurance policies that, despite being financed by all tax payers, primarily benefit the middle class.
Latin America’s social security-heavy welfare regimes are not very redistributive and key
programs are even regressive (Breceda, Rigolini and Saavedra. 2009; Goñi, López and Servén
2008). This chapter moves beyond the standard discussion about the impact of CCT to explore this
important puzzle.
This chapter aims to answer two related questions pertaining to the politics of CCTs in
Latin America. First, what explains their rapid spread in the region? And, second, what explains
the large variation that exists among CCT programs in terms of the share of a country’s population
that they cover? To answer these questions, I conduct two statistical analyses on 18 Latin American
countries from 1996, the year before the launch of the first national-level CCT (Mexico’s
Progresa/Oportunidades), until 2010.
57
The first question is addressed through an event history
analysis, the second through a time-series cross-section regression analysis.
Studying CCTs offers an opportunity to overcome one of the main limitations that has
plagued the recent wave of quantitative research on Latin American social policy: the reliance on
highly aggregated social spending data (Dion 2010; Garay 2010). As Esping-Andersen (1990, 19)
famously remarked, “expenditures are epiphenomenal to the theoretical substance of welfare
states.” Spending levels are useful inasmuch as they serve as proxies for how much a country
prioritizes social policy, but tell us little about how funds are spent or whether governments are in
fact redistributing income.
Social Policy and Presidential Ideology in Latin America

91
The problems associated with spending data are particularly acute when studying Latin
America. The ideal measure for testing the determinants of income distribution involves
comparing pre- and post-tax inequality, as is done in recent studies of OECD countries. However,
this kind of data are not available cross-nationally for Latin America over an extended period of
time. With nearly 75 percent of their resources reaching the poorest 40 percent, CCTs constitute
one of the most redistributive tools available to Latin American policymakers (Goñi, López and
Servén 2008, 21). It therefore stands to reason that the determinants of CCT adoption and coverage
will paint a more accurate picture of the drivers of income redistribution in the region and in
developing countries more generally than social spending levels.
To preview this chapter’s findings, there is a disconnect between the drivers of CCT
adoption and the factors that explain CCT coverage. Presidential ideology does not influence the
probability of CCT adoption, but it does explain CCT coverage levels. Coverage, measured as the
share of the population enrolled in CCTs, tends to be 3.5 percentage points higher (about 19 percent
of the average coverage level in 2010) in countries governed by left-leaning presidents. In other
words, though Latin American presidents of all ideological stripes were equally likely to adopt
these policies, left-leaning leaders were more likely to deepen them by expanding their coverage.
The findings from this statistical analysis are confirmed by qualitative research. Brief case studies
of Brazil, Ecuador and Argentina demonstrate how governments of the left have taken more
aggressive steps toward extending a basic income floor to their populations.
The chapter proceeds as follows. Section 1 summarizes the effects of CCTs. This is
followed by a section explaining their adoption. After a brief review of the relevant literature and
a presentation of the hypotheses tested, an event history analysis of the determinants of CCT
adoption is performed. Section 3 analyzes the determinants of CCT coverage. Like the previous
Social Policy and Presidential Ideology in Latin America

92
section, it begins with a literature review before presenting the hypotheses and conducting a
statistical analysis. This is followed by case studies demonstrating how CCT coverage has
increased under left-leaning governments in three Latin American countries. The chapter then
concludes by explaining the significance of the findings for the study of social policy in developing
countries and offering suggestions regarding future avenues for research.

2. Conditional Cash Transfer Programs in Latin America
CCTs aim to simultaneously reduce poverty and increase human capital. In the immediate
term, transfers relieve poverty by boosting the incomes of the poor and economically vulnerable.
In the long-term, conditionality seeks to increase the human capital of beneficiary children, which
in turn is expected to help them achieve higher future earnings and thus prevent the inter-
generational transmission of poverty.
Proposed and designed by Latin American scholars and policymakers, CCTs were first
adopted on a nationwide scale in Mexico in 1997.
58
By 2010 approximately 130 million Latin
Americans – one in four – benefitted from CCTs (Stampini and Tornarolli 2012). While it is still
too early to measure their effects on inter-generational poverty transmission, CCTs have been
successful at their twin goals of reducing poverty and increasing school attendance and use of
medical facilities. Their effects on poverty and inequality are all the more remarkable considering
that, on average, spending on these programs represents less than 0.5 percent of GDP or roughly
3 percent of total social spending.
59

CCTs and other non-contributive programs – such as recently adopted minimum pension
schemes (Carnes and Mares 2014) – mark a break from Latin American countries’ typical approach
to social policy. Said approach has traditionally been centered on contributory social insurance for
Social Policy and Presidential Ideology in Latin America

93
formal-sector workers, which make up as much as half of the region’s labor force, and, at best,
provided minimal safety nets for the poor (Barrientos and Santibáñez 2009; Ferreira and Robalino
2011; Haggard and Kaufman 2008). These programs also represent a qualitative leap in terms of
their ambition. Never before had the region’s governments attempted to provide an income boost
to such a large share of their populations. In targeting the most vulnerable members of society and
providing benefits completely de-linked from employment status, CCTs constitute a path-breaking
development in Latin American social policy and may even mark a shift in the region’s welfare
regime (Barrientos 2009).
These programs are succeeding at reaching the poorest members of society (Coady, Grosh
and Hoddinott 2004; Goñi, López and Servén 2008; Lindert, Skoufias and Shapiro 2006). In 2010,
CCTs benefited roughly half of the region’s poor, boosting their incomes by one third on average
(Stampini and Tornarolli 2012). In contrast to the social security policies that dominate the region’s
budgets, CCTs are highly redistributive. Whereas roughly 70 percent of social spending in the
region benefits the richest 40 percent, just 8 percent reaches the poorest 20 percent. In contrast,
nearly three quarters of resources spent on CCTs go to the poorest 40 percent (Goñi, López and
Servén 2008, 19-21). Thus, CCT coverage can be analyzed as a proxy for income redistribution in
Latin America.
Compared to the voluminous literature on the effectiveness of CCTs, relatively little has
been written on the politics of CCTs (Martínez Franzoni and Voorend 2011, 280). By studying the
factors that lead to their adoption and influence their coverage levels, it is possible to obtain a
clearer explanation of cross-national variation in income redistribution in the region. The following
section analyzes the determinants of CCT adoption.

Social Policy and Presidential Ideology in Latin America

94
3. Explaining CCT Adoption
While multiple case studies have detailed the factors that prompted the adoption of specific
programs, there exist only two prior cross-national quantitative studies on the determinants of CCT
adoption (Díaz-Cayeros and Magaloni 2009; Sugiyama 2011).
60
Existing explanations can be
categorized as political, economic and related to policy diffusion.
Political explanations encompass both the incentives faced by politicians (supply) and the
actions taken by the broader population (demand). Focusing primarily on the experiences of
Mexico and Brazil, both of which adopted programs during periods of increased political
competition, the case study literature has emphasized the importance of elite competition for votes.
Research on Mexico’s Progresa stresses the importance of growing competition during the final
years of the Partido Revolucionario Institucional (PRI) regime (Díaz-Cayeros, Estévez and
Magaloni 2012; Dion 2010). The Brazilian case was similar. Facing increasing competition from
the left-wing Partido dos Trabalhadores (PT), President Fernando Henrique Cardoso (1995-2003)
of the centrist Partido da Social Democracia Brasileira (PSDB) began providing federal funding
for municipal CCTs in 1997 and launched a CCT at the federal level, Bolsa Escola, in 2001
(Bandeira Coêho 2012; Garay 2012; Melo 2008).
Power resource theory (PRT), the dominant explanation of social policy among
industrialized countries (Huber and Stephens 2001; Korpi 1983), predicts that redistributive
policies will be more likely and more generous under left-leaning governments. However, the
experiences of Mexico and Brazil point to a limited role of ideology in the adoption of CCTs. In
Mexico, President Ernesto Zedillo (1994-2000), a center-right politician from the catch-all PRI,
adopted Progresa to stave off competition from both the right and the left. In Brazil, Cardoso, a
centrist, adopted Bolsa Escola in an attempt to weaken the left. Overall, left-leaning presidents
Social Policy and Presidential Ideology in Latin America

95
enacted only 38 percent (6 of 16) of CCTs in the region. Right leaning presidents enacted 44
percent (7 of 16) (see Table 4-1).
Table 4-1: Presidential Ideology and CCT Adoption
Country Program Name President in Power Ideology
1

.
Argentina Plan Familias/AUH Nestor Kirchner Center-Left

Bolivia Bono Juancito Pinto Evo Morales Left

Brazil Bolsa Escola/Bolsa Familia Fernando Henrique Cardoso Center

Chile Chile Solidario Ricardo Lagos Center-Left

Colombia Familias en Acción Andrés Pastrana Center-Right

Costa Rica Avancemos Oscar Arias Center

Dominican Republic Solidaridad Leonel Fernández Center

Ecuador Bono de Desarrollo Humano Lucio Gutiérrez Center-Right

El Salvador Red Solidaria Antonio Saca Right

Guatemala Mi Familia Progresa Álvaro Colom Center-Left

Honduras Programa Asignación Familiar Carlos Roberto Flores Center-Right

Mexico Progresa/Oportunidades Ernesto Zedillo Center-Right

Panama Red Oportunidades Martín Torrijos Center-Left

Paraguay Tekopora Nicanor Duarte Center-Right

Peru Juntos Alejandro Toledo Center-Right

Uruguay Asignaciones Familiares Tabaré Vázquez Center-Left
.
Notes:
1
. Source: Dataset on Political Ideology of Presidents and Parties in Latin America (Murillo, Oliveros
and Vaishnav 2010).

Turning to the demand-side, PRT also predicts that redistributive policies will be the
product of popular mobilization. However, mobilization has not been central to existing accounts
of CCT adoption. Díaz-Cayeros, Estévez and Magaloni (2012) do cite the 1994 Zapatista Rebellion
as a factor explaining Mexico’s shift from clientelistic to programmatic social policies. In a related
argument, Feitosa de Britto (2008, 187) speculates that increasing urban violence may explain
CCT adoption. In her words, “fear encourages elites to favor public policies that tackle poverty
directly and, in particular, that keep poor children and adolescents in school.”
Economic explanations are centered on the market reforms of the 1980s and 1990s, which
significantly increased economic vulnerability in the region (Barrientos 2013; Dion 2010).
Social Policy and Presidential Ideology in Latin America

96
Vulnerability increased demand for social assistance through two channels. First, increased labor-
market informality, a product of trade liberalization and public-sector downsizing, increased the
share of workers excluded from contributory social insurance. Second, low growth and periodic
financial crises during the 1990s increased the need for emergency safety nets, many of which
became permanent.
Economic stagnation and financial crises that disproportionality hurt the poor increased
public demand for safety nets. The launch of Mexico’s Progresa has been partially attributed to
the country’s 1994-1995 “Tequila Crisis” (Cortés and Rubalcalva 2012; Díaz-Cayeros, Estévez
and Magaloni 2012, 182). Several now permanent CCTs, most notably Colombia’s Familias en
Acción (Brearly 2011) and Honduras’ Programa de Asignación Familiar (Moore 2008), were
originally temporary responses to crises.
In an argument that combines political and economic factors, Barrientos (2013, 12) posits
that the spread of non-contributory social programs in the developing world may be explained by
a shift in the composition of government revenues away from payroll and corporate taxes toward
consumption taxes, which fall heavily on the poor. This echoes earlier work by Timmons (2005),
which argues that, since much of the burden of consumption taxes falls on the poor, countries that
are highly dependent on those taxes will spend more on policies that benefit the poor such as social
policy.
Whereas the explanations presented above look at domestic factors, policy diffusion
arguments conceptualize CCT adoption as a process driven by learning from and/or emulation of
neighboring countries. Diffusion research seeks to explain situations in which, over a short period,
nation-states or other units, adopt similar policies or institutions (Elkins and Simmons 2005).
Social Policy and Presidential Ideology in Latin America

97
Diffusion offers an explanation as to why neighboring countries facing different conditions
sometimes adopt strikingly similar policies and institutions.
Figure 4-1: Number of CCTs
Operating in Latin America (1996-2010)

Source: News reports.
Notes: The tally includes Costa Rica's Superémonos and Nicaragua's Red de
Protección Social CCTs. The former was cancelled in 2003. The latter was
cancelled in 2007.

There are strong reasons to believe that the spread of CCTs constitutes a diffusionary
process. Weyland (2007, 18-19) defines policy diffusion as having three characteristics: an s-
shaped curve pattern, geographic clustering of policies and commonality among diversity. First,
Figure 1, which provides a tally of the number of CCTs operating in the region, resembles an s-
shaped curve. Second, it constitutes an example of geographic clustering that over the span of a
decade the vast majority of the countries in the region came to possess the same type of policy.
Third, given that CCTs were adopted by governments of the left (Bolivia and Guatemala) as well
as the right (El Salvador and Mexico) and by the region’s richest countries (Argentina and Chile)
as well its poorest (Bolivia and Honduras), their proliferation in Latin America is an example of
commonality among diversity.
0
2
4
6
8
10
12
14
16
18
Jan‐96
Jan‐97
Jan‐98
Jan‐99
Jan‐00
Jan‐01
Jan‐02
Jan‐03
Jan‐04
Jan‐05
Jan‐06
Jan‐07
Jan‐08
Jan‐09
Jan‐10
Jan‐11
Social Policy and Presidential Ideology in Latin America

98
Furthermore, there is substantial evidence that, in adopting CCTs, late adopters were
emulating the successful experiences of Mexico and Brazil (Lana and Evans 2004; Martínez
Franzoni and Voorend 2011). These experiences were publicized by the media, academics and
international institutions; furthermore, the World Bank and Inter-American Development Bank
provided technical and financial support for the adoption of CCTs throughout the region
(Ancelovici and Jenson 2013; Lana and Evans 2004; Martínez Franzoni and Voorend 2011;
Sugiyama 2011).
3.1. Hypotheses and Data
Dependent variable: CCT adoption is measured dichotomously and is defined as taking
place during the year a country’s government announces the launch of a nationwide anti-poverty
program that provides monetary transfers conditional on school attendance. This excludes sub-
national CCTs, pilot versions of programs and CCTs without an educational conditionality.
3.1.1. Political Variables
Presidential ideology: The logic of PRT leads to the expectation that countries with left-
leaning governments will be more likely to adopt CCTs. Ideology is operationalized using a dummy
variable coded one if the president is left-leaning and an index of presidential ideology ranging
from one (right) to five (left). Both measures are drawn from the Dataset on Political Ideology of
Presidents and Parties in Latin America (Murillo, Oliveros and Vaishnav 2010).
61

Political competition: Based on the Mexican and Brazilian experiences, presidents facing
elections should be more likely to adopt a CCT. Competition is measured using a dummy variable
coded one during years in which there were presidential, legislative or nationwide referendum
elections. CCTs are politically popular policies. Thus, weak presidents looking to increase their
political position should be more likely than stronger presidents to adopt a CCT. This is measured
Social Policy and Presidential Ideology in Latin America

99
through a dummy coded one if the president’s party or legislative coalition lacks a legislative
majority.
Number of protests: Politicians facing mass mobilizations can utilize social policy to
assuage or even coopt protestors. Thus, the greater the number of protests reported during a given
year, the more likely a government will be to adopt a CCT. Protests are measured as the number of
strikes, riots and demonstrations in a country during a given year lagged by one year as reported
by the Cross-National Time-Series data archive (Banks 2011).
62

Crime: If CCTs are a response to crime, countries with high levels of violence should have
a higher probability of adopting CCTs. Violence is operationalized as the homicide rate per
100,000 inhabitants lagged by one year as reported by the United Nations Crime Trends Surveys
(UNDOC 2014).
3.1.2. Economic Variables
Level of inequality: CCTs are among the most redistributive policies available to Latin
American policymakers. The median voter theory (MVT) (Meltzer and Richard 1981) predicts that
the most unequal democracies will, all else equal, be the most redistributive. This logic can be
extended to argue that the more unequal a country, the higher the probability that it will adopt a
CCT. Inequality is measured as the Gini coefficient lagged by one year as reported by the United
Nations Economic Commission on Latin America and the Caribbean’s CEPALSTAT Database.
Recession: If, indeed, they are a response to crises, CCTs should be more likely to be
adopted during periods of low or negative growth. This is measured using a dummy variable coded
one during years with economic growth below one percent. The variable is lagged by one year and
calculated from the World Bank’s World Development Indicators (WDI).
63

Social Policy and Presidential Ideology in Latin America

100
Consumption taxes: Following Barrientos (2013), the higher a government’s level of
consumption tax revenues, the higher the probability that it will adopt a CCT. Consumption taxes
are measured utilizing indirect taxes as a share of GDP lagged by one year as reported by
CEPALSTAT.
64

3.1.3 Policy Diffusion
If the spread of CCT programs has been an example of diffusion, the probability of adopting
a program should increase as the share of countries in the region with CCTs increases. Past
diffusion research has hypothesized that countries receive most of their policy cues from peer
states, which in this case, would be other Latin American countries (Brooks 2005). The existence
of shared political, economic and cultural traits, as well as constant interaction at the governmental
and societal levels, makes Latin American governments keenly aware of what other countries in
the region are doing and can lead them to emulate policies perceived as successful.
Following past studies on the topic (Brooks 2005; Carnes and Mares 2014), diffusion is
operationalized as the share of countries in the region with CCTs (excluding the country in
question) lagged by one year.
65
A squared version of the variable is also tested to assess whether
this effect is non-linear. A cubed term is also tested because s-shaped curves such as those
associated with diffusion processes resemble cubic functions.
3.1.4. Control Variables
Youth population: measured as the share of the population under 15 lagged by one year as
reported by the WDI.
GDP per Capita: measured as the natural log of GDP per capita at purchasing power parity
in constant 2005 international dollars lagged by one year as reported by the WDI.

Social Policy and Presidential Ideology in Latin America

101
3.2. Estimation Technique
An event history analysis utilizing binary time-series cross-section methods (Beck 1998;
Beck, Katz and Tucker 1998) is conducted to assess the determinants of CCT adoption. Before
proceeding, three methodological issues need to be discussed. First, to deal with heteroskedasticity
common to cross-national data, Huber-White robust standard errors are calculated. Second, in
response to Beck’s (1998) concerns regarding the need to avoid simultaneities among the
covariates and the duration to be explained as well as more general concerns about endogeneity,
most independent variables are lagged by one year. Third, to address the temporal dependence
inherent to this data (i.e. the passage of time influences the probability of adopting a CCT), the
conditional hazard rate based on the underlying time trend is accounted for by fitting the logit
models with a cubic polynomial approximation (Carter and Signorino 2010).
3.3. Results
The event history analysis summarized in Table 4-2 yields two main conclusions. First, it
confirms that diffusion is a major driver of CCT adoption. This echoes earlier findings on CCT
diffusion by Sugiyama (2011). Second, on the domestic front, the analysis supports MVT and
challenges PRT. Income inequality has a positive effect on the likelihood of adoption. In contrast,
ideology has no effect.
66
The finding on ideology provides additional evidence in support of earlier
research on CCT adoption (Díaz-Cayeros and Magaloni 2009; Sugiyama 2011). Indeed, while not
statistically significant, both left-leaning president and the index of presidential ideology have a
consistently negative sign.

Social Policy and Presidential Ideology in Latin America

102
Table 4-2: Determinants of CCT Adoption 1996-2010
Independent Variable (1-1) (1-2) (1-3) (1-4) (1-5)
1

Left President -0.579
-----
-0.418 0.136
-----
(0.839) (0.897) (1.066)

Presidential Ideology
----- -0.840 ----- -----
-0.276
(0.548) (0.552)

Election Year -0.825 -0.829 -0.916 -0.993 -1.150
(0.661) (0.697) (0.794) (0.758) (0.730)

Majority President 0.884 1.148 0.802 0.792 0.841
(0.725) (0.882) (0.731) (0.734) (0.772)

Protests
0.028 0.146
0.033 0.016 0.055
(0.073) (0.154) (0.079) (0.109) (0.126)

Homicides
-----
-0.001
----- ----- -----
(0.029)

Inequality
11.831* 18.070* 10.262+ 13.931* 15.096*
(5.903) (8.405) (6.220) (6.565) (7.000)

Recession -0.701 -0.411 -0.430 -0.226 -0.246
(1.160) (1.564) (1.221) (1.483) (1.520)

Consumption Taxes 0.050 0.156 0.041 0.027 0.072
(0.081) (0.188) (0.084) (0.087) (0.096)

Diffusion 2.930* 4.753* -12.041 -57.441** 0.459
(1.433) (2.272) (11.672) (19.857) (1.102)

Diffusion (Squared)
----- -----
13.943 152.521** 5.504**
(11.016) (50.983) (1.686)

Diffusion (Cubed)
----- ----- -----
-105.097** -2.300**
(37.807) (0.766)

Population under 15 -0.158 -0.216 -0.184 -0.154 -0.150
(0.180) (0.257) (0.184) (0.238) (0.223)

GDP per Capita
-0.000
-0.000
-0.000 -0.000 -0.000
(0.000) (0.000) (0.000) (0.001) (0.001)

Constant -3.668** -8.299* -1.164 1.866 -5.598*
(1.235) (4.085) (2.477) (2.332) (2.684)

Pseudo R2 0.2381 0.2795 0.2631 0.3806 0.3717
N 152 132 152 152 152
Notes: + p = 0.07; *p <= 05; ** p <= 0.01 in two-tailed test.
Cubic polynomials (not shown) are used to address time-dependence.
1
. To reduce multicolinearity, the linear diffusion variable was transformed in accordance with
Kim’s (1999) recommendations. The transformed term was then squared and cubed.

Social Policy and Presidential Ideology in Latin America

103
The other hypothesized explanations of CCT adoption fail to hold up to scrutiny. Neither
the number of protests nor the level of consumption taxes have a significant effect on adoption.
The finding on protests reinforces case study work by Martínez Franzoni and Voorend (2011, 290),
who find that “the formation of CCTs was pretty much the business of poverty experts rather than
grassroots organizations or the poor themselves.” While also failing to reach statistical
significance, the signs of several coefficients consistently go against expectations. With regard to
political competition, majority president and election year have, respectively, positive and negative
signs. Recession has a consistently negative sign.
At the cost of a 20 observations (15 percent of the sample), model 1-2 incorporates the
homicide rate as a proxy for crime level. Said variable is far from statistically significant. Its
incorporation does not alter the previous findings on diffusion and inequality.
Figure 4-2: Estimated Effect of Time, Diffusion and
Income Inequality on CCT Adoption (Model 1-1)

Source: Model 1-1, Table 3-3
Note: all other covariates are held at their means (continuous variables) or
medians (discrete variables).

0.00
0.20
0.40
0.60
0.80
1.00
123 45 678 9 10 11 12 13 14 15
Gini 0.40, 25% with CCTs Gini 0.50, 50% with CCTs
Gini 0.60, 75% with CCTs
Predicted Prob. of CCT Adoption
Years a Country has Been at Risk of Adopting aCCT
Social Policy and Presidential Ideology in Latin America

104
Based on model 1-1, Figure 4-2 simultaneously explores how the share of countries with
CCTs and the level of inequality have, over time, affected the probability of CCT adoption. Under
a baseline scenario, which assumes a Gini of 0.4 (close to the sample minimum) and that a quarter
of countries have CCTs, the probability of adoption is consistently low, peaking at about 6 percent.
Under an intermediate scenario, which assumes a Gini of 0.5 and that half the countries in the
region have CCTs, the probability of adoption peaks at 31 percent. The final scenario, which
assumes a Gini of 0.6 (close the sample maximum) and that three quarters of countries have CCTs,
probability of adoption peaks at about 75 percent.
Given the presence of an s-shaped pattern with two points of inflection, there is reason to
believe that the relationship between the share of countries with CCTs and adoption is non-linear
and possibly even cubic. While there is no evidence of a squared function (model 1-3), there is
evidence of a cubic relationship (models 1-4 and 1-5).
Multicolinearity is the main drawback of incorporating a cubed term. It is problematic in
as much as it leads to coefficient instability and large standard errors. Multicolinearity does not,
however, bias estimates (Arceneaux and Huber 2007). Thus, while the coefficients do not provide
precise estimations of adoption probabilities, they are indicative of a statistically significant
relationship. The diffusion variables are adjusted to reduce multicolinearity (model 1-5) following
the procedure devised by Kim (1999).
67
This provides more reliable coefficients with much smaller
standard errors. However, while the squared and cubed versions of the diffusion variable remain
highly significant, the linear term is no longer significant. Still, these results provide evidence that
CCT adoption is the result of diffusion.
Social Policy and Presidential Ideology in Latin America

105
The findings on the determinants of CCT adoption are presented in Table 4-3. Having
explored the determinants of CCT adoption, the following sections analyze the determinants of
CCT coverage.
Table 4-3: Findings on the Determinants of CCT Adoption
Theory/Variable
Expected
Outcome
Finding Notes

Political Factors
Presidential Ideology Index + None Negative sign
Left-leaning president + None Negative sign
Majority president - None Positive sign
Election Year + None Negative sign
Number of Protests + None
Homicide Rate + None
Economic Factors
Level of Inequality + +
Recession + None Negative sign
Consumption Taxes to GDP + None

Policy Diffusion
% of Countries with CCTs + +
% of Countries with CCTs
2
+ None/+
Only significant when
cubed term is included
% of Countries with CCTs
3
+ +

4. Explaining CCT Coverage
While there is no literature precisely on CCT coverage from which to draw hypotheses, the
factors that influence coverage levels could be expected to be similar to those that determine levels
of social spending and redistribution. Hypotheses on CCT coverage are drawn from the rich
literature on the determinants of social spending in Latin America (e.g. Huber and Stephens 2012;
Segura-Ubiergo 2007) and developing countries more generally (e.g. Haggard and Kaufman 2008;
Rudra 2008).
68

The vast PRT literature (Huber and Stephens 2001; Korpi 1983) attributes the expansion
of and cross-national variation in social policy to the political strength of the left and the working
Social Policy and Presidential Ideology in Latin America

106
class. This perspective leads to the prediction that left-leaning presidents will increase the scope
of CCTs and push them in a more universalistic direction. The other component of this theory,
popular mobilization, is often driven by demands for increased social spending and income
redistribution. This is particularly the case in poorer countries. Governments can neutralize
mobilization by increasing the scope of redistributive programs such as CCTs.
In addition to serving as a defensive strategy, CCTs can be an offensive weapon used by
incumbent presidents to increase their popularity and political leverage vis-à-vis the opposition.
There is substantial evidence from throughout Latin America that incumbent presidents and parties
have reaped electoral benefits from CCTs (de la O 2013; Diaz-Cayeros, Estevez and Magaloni
2009; Zucco 2013). Under this logic, presidents facing elections and/or in a precarious political
situation, such as lacking a legislative majority, could be expected to increase CCT coverage.
CCT coverage may also be influenced by a country’s income distribution. The MVT
(Meltzer and Richard 1981) predicts demand for redistribution within a democracy will be a
function of inequality, which determines the political interests of the pivotal median voter.
Increased redistribution is predicted in highly unequal societies, such as Latin American countries.
Given their laser-like focus on the poor, CCTs constitute one of the most effective and affordable
ways to redistribute income on a large scale. Based on this reasoning, CCT coverage should be
higher in more unequal countries.
A separate line of research has argued that the existence of politically salient ethnic
divisions within a society reduces public support for and, ultimately, the provision of redistributive
policies. Individual support for redistribution, it is argued, tends to be lower when tax payers
perceive that social programs benefit people who are perceived to be racially different from
Social Policy and Presidential Ideology in Latin America

107
themselves (Alesina and Glaeser 2004; Luttmer 2001). This same argument may apply to CCT
coverage.
4.1. Hypotheses and Data
Dependent variables: For each country-year, CCT coverage is measured both as the number
of CCT beneficiaries as a share of the total population and as the ratio of the number of CCT
beneficiaries to the estimated number of people living in poverty (defined as an income of less than
$4 US a day).
69
The number of beneficiaries is compiled using data from Stampini and Tornarolli
(2012) and the websites of CCT programs. Country population was obtained from the WDI. The
poverty data was compiled by SEDLAC (2014).
4.1.1. Political Explanations
Presidential ideology: Based on PRT, presidents of the left will be more likely to ramp up
CCT coverage than presidents of the center or right. As in the CCT adoption models, ideology is
measured using an index of presidential ideology and a dummy variable coded one if the president
is left-leaning (Murillo, Oliveros and Vaishnav 2010).
Protests: Governments are expected to respond to mass protests by increasing CCT
coverage. Protests are measured as the number of strikes, riots and demonstrations in a country
lagged by one year (Banks 2011).
Political competition: Coverage is expected to increase under presidents facing an election
or lacking a legislative majority. As in the CCT adoption models, dummy variables are used to
capture these situations.

Social Policy and Presidential Ideology in Latin America

108
4.1.2. Income Distribution
Level of inequality: In line with the MVT, more unequal Latin American countries are
expected to have higher levels of redistribution and thus CCT coverage. Inequality is measured
using the Gini coefficient lagged by one year.
4.1.3. Ethnic Divisions
The more ethnically divided a country is, the lower the share of its population covered by
CCTs. Ethnic divisions are measured using Lieberman’s (2009, 47) ethnic boundary index, which,
for Latin America, ranges from 0 (extremely weak or non-existent boundaries, i.e. Argentina) to
0.75 (mostly strong and/or multiple ethnic boundaries, i.e. Ecuador).
70

4.1.4. Economic Constraints and Control Variables
Though CCTs are very cost-effective, increasing their coverage will still be constrained by
factors such as tax revenues, social spending levels and the state of the economy. While economic
crises generate demand for safety nets that could make CCT adoption more likely, they impose
resource constraints that might prevent governments from expanding coverage.
Four variables are included to control for resource constraints: size of the tax base measured
as total tax revenues as a share of GDP, social spending measured as total social spending as a
share of GDP, economic growth measured as a percent change in GDP, and, as a measure of
international constraints, debt service measured as a share of gross national income. All four
variables are lagged by one year. The first three are from CEPALSTAT and the last is from the
WDI. In addition, the models control for the share of the population under 15 and GDP per capita
both from the WDI.

Social Policy and Presidential Ideology in Latin America

109
4.2. Estimation Technique
The determinants of CCT coverage are assessed through a time-series cross-section
regression analysis. Time-series cross-sectional data is prone to violating several basic
assumptions of ordinary least squares regression. Given these issues, regression parameters are
estimated, as Beck and Katz (1995) suggest, using OLS with panel-corrected standard errors
(PCSE). Additionally, to avoid the suppression of the explanatory power of the independent
variables (Achen 2000) and possibly bias estimates (Plümper, Troeger and Manow 2005), a lagged
dependent variable is not used to address serial auto-correlation. Instead, the models correct for
first-order-auto-regression, which deals with serial auto-correlation without sacrificing the
explanatory power of the independent variables (Beck and Katz 2004; Huber, Mustillo and
Stephens 2008). The models are run both with and without country fixed effects.
4.3. Results
The analysis finds robust evidence that CCT coverage, both as a share of the population
and relative to the share of the population living in poverty, tends to be higher in countries with
left-leaning presidents. In addition, coverage increases the further to the left a president is situated.
Thus, although ideology has no effect on the probability of CCT adoption, it plays an important
role in shaping how many people actually receive CCT benefits. The effect of ideology is all the
more noteworthy given that Venezuela, the country most consistently governed by the left during
the period studied, is the only country in the sample to never have adopted a CCT; therefore, it has
a coverage rate of zero in every recorded observation.
71

The most surprising finding is that, contrary to expectations, ethnic divisions have a large
positive effect on both measures of CCT coverage. This runs counter to the implications of past
research on individual support for anti-poverty policies in both the U.S. (Luttmer 2001) and Latin
Social Policy and Presidential Ideology in Latin America

110
Table 4-4: Determinants of CCT Coverage as a % of Total Population (1996-2010)

Fixed Effects?
NO NO NO NO YES YES
Ind. Variable (2-1) (2-2) (2-3) (2-4) (2-5) (2-6)

Presidential Ideology 1.352** ----- 1.213+ ----- 1.885*** -----
(0.445) (0.679) (0.484)

Left President ----- 3.497*** ----- 3.372* ----- 4.251***
(0.992) (1.334) (1.086)

Election Year -0.325 -0.221 -0.235 -0.105 -0.217 -0.126
(0.328) (0.338) (0.452) (0.458) (0.385) (0.405)

Majority President 2.135** 1.992* 2.296* 2.223* 2.505** 2.460**
(0.815) (0.803) (0.970) (0.956) (0.888) (0.883)

Protests -0.016 0.001 -0.015 0.013 -0.009 0.013
(0.132) (0.132) (0.176) (0.172) (0.146) (0.149)

Gini Coefficient ----- ----- -3.831 -3.263 ----- -----
(11.160) (11.353)

Ethnic Divisions 28.131* 27.284* 25.439+ 25.132+ ----- -----
(11.557) (10.979) (13.717) (13.050)

Size of Tax Base 0.830** 0.831** 0.868* 0.855* 1.118** 1.159**
(0.297) (0.297) (0.401) (0.392) (0.369) (0.383)

Social Spending ----- ----- -0.011 0.050 ----- -----
(0.334) (0.331)

GDP Growth -0.059 -0.055 -0.068 -0.061 -0.084 -0.079
(0.068) (0.068) (0.070) (0.070) (0.079) (0.079)

Debt Level -0.097 -0.098 -0.098 -0.096 -0.118 -0.117
(0.089) (0.087) (0.086) (0.084) (0.092) (0.091)

Population under 15 -1.036* -0.950* -0.959+ -0.858 -1.065** -0.966**
(0.438) (0.413) (0.562) (0.540) (0.377) (0.358)

GDP per Capita 0.000 0.000 0.000 0.000 0.002* 0.002**
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001)

Constant -12.511* -9.753* -11.994 -10.293 -28.934*** -24.417***
(5.174) (4.916) (8.418) (7.989) (6.250) (5.817)

R2 0.091 0.106 0.096 0.111 0.401 0.416
N 270 270 253 253 270 270
Notes: + p <= 0.10; *p <= 05; ** p <= 0.01 in two-tailed test. Country dummies not shown in fixed effects
models.

America (Aradanaz 2009). It also runs counter to work on the determinants of social spending
levels (Borges 2014) and of poverty and inequality (Huber and Stephens 2012) in Latin America.
Historically, countries with large indigenous populations have been among the poorest and most
Social Policy and Presidential Ideology in Latin America

111
unequal in Latin America. That these countries have among the highest coverage rates can be
interpreted as evidence that CCTs are expanding where they are most needed.
Also contrary to expectations, having a majority president is associated with an increase in
CCT coverage levels, suggesting that presidents need supportive legislative majorities to ramp up
CCT coverage. In contrast, other political variables fail to yield consistently significant results.
Neither election years nor the number of yearly protests have a significant effect on any measure
of CCT coverage. Similarly, income inequality has no effect. With regard to resource constraints,
there is strong evidence that coverage is a function of available tax revenue. Findings on the other
constraints are less robust.
Figure 4-3: Estimated Effects and 95% Confidence Intervals of Presidential Ideology
on CCT Coverage as a % of Total Population (left) and Poor Population (right)
in 18 Latin American countries (1996-2010)

L = Left, CL = Center-Left, C = Center, CR = Center-Right, R = Right

Source: Table 3-4, Model 2-1 (left) and Table 3-5, Model 3-1 (right).
Notes: All other continuous variables kept at their listwise mean. All other categorical variables are
kept at their listwise mode.

0
2
4
6
8
10
12
14
16
CCT Coverage (% Population)
PresidentialIdeology Index
L CL CCR R
10
15
20
25
30
35
40
45
50
CCT Coverage (% Poor Population)
PresidentialIdeology Index
L CL CCR R
Social Policy and Presidential Ideology in Latin America

112
Table 4-4 summarizes the findings on the determinants of CCT coverage as a share of a
country’s total population. To maximize the number of observations, baseline models that exclude
income distribution and social spending are first tested. Model 2-1 incorporates the presidential
ideology index and finds that, all else equal, moving one ideological position toward the left
increases coverage by 1.35 percent of a country’s population (see Figure 3). This effect is
substantively large: a shift from a right-wing to a left-wing president is associated with an
increased coverage of 5.41 percent of the population, which amounts to almost 30 percent of the
average coverage level during 2010 (18.78 percent). A similar result is obtained by substituting
the ideology index with a dummy variable (model 2-2). Coverage in countries with left-leaning
presidents is expected to be 3.50 percentage points (about 19 percent of the 2010 average) higher
than under presidents of the center or right. The effect of having a president of the left is robust to
every model specification tested. In contrast, when income distribution and spending levels are
included, the significance of the level of the ideology index drops to the 90 percent level (model
2-3).
The effect of presidential ideology on CCT coverage is robust to country fixed effects
(models 2-5 and 2-6) and the presidential ideology index retains significance in models including
income distribution and social spending (not shown). In fact, the coefficients for the ideology
variables are substantially (10 to 20 percent) larger in the fixed effects models. This implies that
the effect of presidential ideology on CCT coverage as a share of the population is driven more by
over-time variation within cases than by cross-national variation.

Social Policy and Presidential Ideology in Latin America

113
Table 4-5: Determinants of CCT Coverage as a % of Poor Population (1996-2010)

Fixed Effects?
NO NO NO NO YES YES
Ind. Variable (3-1) (3-2) (3-3) (3-4) (3-5) (3-6)

Presidential Ideology 3.359* ----- 3.751** ----- 4.997*** -----
(1.346) (1.452) (1.486)

Left President ----- 9.408*** ----- 11.474*** ----- 12.763***
(2.484) (2.797) (2.871)

Election Year 0.427 0.635 0.868 1.271 0.769 1.028
(1.102) (2.484) (1.071) (1.154) (1.381) (1.499)

Majority President 4.265+ 3.701+ 4.414+ 3.975+ 5.304* 4.843*
(2.189) (1.177) (2.377) (2.361) (2.360) (2.318)

Protests -0.133 -0.094 -0.103 -0.023 -0.091 -0.014
(0.391) (0.391) (0.400) (0.387) (0.395) (0.389)

Gini Coefficient ----- ----- -20.277 -20.059 ----- -----
(33.776) (34.065)

Ethnic Divisions 74.437* 71.047* 73.121* 71.562** ----- -----
(36.936) (32.804) (29.600) (27.018)

Size of Tax Base 2.833** 2.997** 2.861** 2.906** 3.470** 3.487**
(1.022) (1.012) (0.936) (0.901) (1.137) (1.131)

Social Spending ----- ----- 1.298+ 1.517* ----- -----
(0.773) (0.741)

GDP Growth -0.399** -0.395** -0.367* -0.351* -0.600** -0.590**
(0.138) (0.140) (0.173) (0.174) (0.208) (0.214)

Debt Level -0.478+ -0.497* -0.453+ -0.463* -0.545+ -0.548+
(0.264) (0.242) (0.257) (0.227) (0.283) (0.287)

Population under 15 -3.377** -2.984** -2.084 -1.616 -1.819 -1.528
(1.089) (1.048) (1.275) (1.215) (1.336) (1.241)

GDP per Capita 0.005*** 0.005*** 0.005*** 0.005*** 0.015*** 0.015***
(0.001) (0.001) (0.001) (0.001) (0.003) (0.003)

Constant -33.309+ -28.803+ -51.688* -48.465* -112.20*** -100.02***
(17.475) (16.819) (20.179) (18.979) (20.445) (19.596)

R2 0.076 0.096 0.105 0.134 0.402 0.411
N 246 246 233 233 246 246
Notes: + p <= 0.10; *p <= 05; ** p <= 0.01 in two-tailed test. Country dummies not shown in fixed effects
models.

Table 4-5 summarizes the findings on the determinants of CCT coverage relative to the
population living on less than $4 U.S. a day. Moving one ideological position toward the left
increases coverage of the country’s poor population by 3.36 percentage points (model 3-1, see
Figure 4-3). Moving from a right-wing president to a left-wing one is be expected to increase
Social Policy and Presidential Ideology in Latin America

114
coverage of the poor population by 13.44 percentage points or by almost 18 percent of the average
coverage level during 2010 (76 percent). Substituting the ideology index with a dummy variable
(model 3-2) yields a similarly strong result. Coverage under a left-leaning president is expected to
be higher by 9.41 percentage points (about 12 percent of the average 2010 level) relative to a
country with a president of the center of the right. The effect of the ideology variables is robust to
every specification tested.
As in the previous models, the effect of ideology on coverage relative to the poor
population is robust to country fixed effects (models 3-5 and 3-6) and the coefficients for the
ideology variables are substantially (22 to 27 percent) larger in said models. This provides
additional evidence that ideology’s effect on CCT coverage is driven more by variation within
cases than across cases.
Further discussion of the surprising and substantively large effect of ethnic divisions on
both measures of CCT coverage is warranted. Holding all other variables constant, CCTs in the
most ethnically divided countries in the region (i.e. Bolivia and Ecuador), are estimated to cover
an additional 20 percent of their population compared to the most homogenous countries (i.e.
Argentina and Costa Rica) (models 2-1 and 2-2). Similarly, CCTs in the most divided countries
are expected to cover an additional 50 percent of their poor population relative to the most
homogenous countries.
One plausible explanation for this finding is that it is a coincidence driven by two outlier
cases: Bolivia and Ecuador, which are among the most ethnically divided countries in the region
and have CCTs with above average coverage levels. While excluding those countries substantially
reduces the magnitude of the ethnic divisions coefficient, it does not alter its statistical significance
(not shown).
Social Policy and Presidential Ideology in Latin America

115
The findings on coverage are summarized in Table 3-6. The following section fleshes out
these findings through illustrative case studies of the experiences of Brazil, Ecuador and Argentina.
Table 4-6: Findings on the Determinants of CCT Coverage
Theory/Variable
Expected
Outcome
Finding Notes

Political Factors
Presidential Ideology Index + +
Coefficient is larger under fixed
effects
Left-leaning president + +
Coefficient is larger under fixed
effects
Majority president - +
Election Year + None
Number of Protests + None
Income Distribution
Level of Inequality + None

Ethnic Divisions
Ethnic Boundaries Index - +
Robust to the exclusion of Bolivia and
Ecuador, which have high coverage
levels and large indigenous
populations

5. Cases Studies: CCT Coverage under Left Presidents
The key finding of the preceding statistical analysis was that CCT coverage tends to be
higher in countries governed by left-leaning presidents. Furthermore, the substantially higher
regression coefficients for presidential ideology in the fixed effects models implies that ideology’s
effect is driven more by within-country, over-time variation than by cross-national variation. To
illustrate these findings, this section traces the evolution of CCT coverage over time in three
countries: Brazil, Ecuador and Argentina.
These cases present substantial variation with regard to the key independent variable of
this study, presidential ideology. During the period studied, Brazil went from being governed by a
centrist president to being governed by successive center-left administrations. The change was
more dramatic in Ecuador, which went from being governed by a president who leaned to the right
Social Policy and Presidential Ideology in Latin America

116
to a far-left president. Finally, Argentina was consistently governed by the left during the period
of study. However, it can be argued that a center-left president there was replaced by one who
leaned farther to the left. The cases are also representative of the substantial variation found in the
region in terms of average income levels, levels of poverty and scope of pre-existing social policy
arrangements (Filgueira and Filgueira 2002; Martínez Franzoni 2011).
Brazil: As noted, Cardoso, a centrist, pioneered federal-level CCTs in Brazil. By the time
he left office at the end of 2002 21.5 million Brazilians were enrolled in CCTs. Though initially
skeptical (Lindert in Brearly 2011, 110), his center-left successor Luiz Inácio Lula da Silva (2003-
2011) ultimately embraced CCTs, merging Bolsa Escola and other smaller programs to create
Bolsa Familia. By the end of Lula’s first year in office, CCTs covered 37.7 million. Coverage
increased steadily during his two terms in office. In June 2006 alone, as a measure banning
enrollment three months prior to the October 2006 presidential election was set to enter into effect
1.8 million families were added to the program’s rolls (Hall 2008, 806).
72
By the time Lula stepped
down at the end of 2010, Bolsa Familia reached 52.2 million Brazilians.
Coverage also increased as a result of changes in the program’s design enacted under Lula
and his successor Dilma Rousseff (2011-present), also of the left-leaning PT. In 2008, the program
was extended to 16 and 17 year-olds. Children in this age group were awarded a larger stipend
designed to offset the opportunity cost of not working. Rouseff further expanded the program by
increasing the maximum number of beneficiaries per family from three to five. Through Rouseff’s
Busca Ativa initiative, officials have sought out families previously excluded as a result of
inadequate identification or lack of knowledge about the program. This increased enrollment by
an additional 800,000 during her first three years in office (SECOM 2013).
Social Policy and Presidential Ideology in Latin America

117
When launched, Bolsa Familia did not initially represent a significant improvement over
Bolsa Escola in terms of benefit generosity. However, adjustments toward the end of the decade
raised the inflation-adjusted value of stipends beyond their original levels. Furthermore, in addition
to the aforementioned higher stipends for teenagers, in 2012 Rouseff substantially increased
stipends for children under six (Simoes 2012).
Ecuador: The election of a leftist president was also followed by substantial increases in
CCT coverage and generosity in Ecuador. Enacted in 2003 by Lucio Gutiérrez (2003-2005), who
governed as center-right despite campaigning as center-left (Murillo, Oliveros and Vaishnav
2010), Bono de Desarrollo Humano CCT initially covered 900,000 mothers. It remained close to
that level until Rafael Correa (2007-present) assumed the presidency and expanded coverage to
1.2 million.
73
However, it should be noted that coverage levels declined significantly in 2013 as a
result of stricter enforcement of the income ceiling for beneficiaries (VAMS 2014, 14).
Correa also tripled the nominal size of the monthly stipend in a country that, as a result of
its use of the U.S. dollar, has had very low inflation levels. During his second week on the job,
Correa doubled the program’s stipend, which had not been modified since its launch (Calvas
Chávez 2010, 52-53), to $30 US. The stipend was further increased in 2009 and 2013 and currently
stands at $50 a month.
Argentina: The models presented above also predict that CCT coverage will increase
following a shift from a moderate left-wing president to a more radical one. There is evidence of
precisely this occurring in Argentina under Cristina Fernández de Kirchner (2007-present), who
succeeded her late husband Néstor Kirchner (2003-2007).
Scholars of Latin American politics struggle to categorize the ideology of the Kirchners
and the Peronist Partido Justicialista to which they belong. The Kirchners do not fit well into the
Social Policy and Presidential Ideology in Latin America

118
dichotomized view of the region’s left popularized by Castañeda (2006) in which leaders are either
moderate social democrats or radical populists.
74
Much of the confusion owes to the fact that,
although the Kirchners have governed from the left, their party is not of the left but, rather, a
populist machine without a set ideology (Etchemendy and Garay 2011; Levitsky 2003).
The statistical analysis presented above follows Murillo, Oliveros and Vaishnav (2010) in
classifying both presidents as center-left. However, there is substantial evidence that Fernández
has governed to the left of Kirchner. While both espoused anti-neoliberal rhetoric, Kirchner did
not roll back the major privatizations of the 1990s (Etchemendy and Garay 2011, 293).
75

Fernández, in contrast, spearheaded the high-profile nationalizations of the pension system in 2008
and the energy firm Yacimientos Petrolíferos Fiscales (YPF) in 2012. On the macroeconomic
front, she has prioritized growth at the cost of high inflation levels and, since 2011, has pursued
strict currency controls. Similarly, content analysis of state of the union addresses reveals
Fernández to be significantly to the left of her late husband (Arnold, Doyle and Wiesehomeier
2014).
While both governments pursued CCTs, these policies played a more important role under
Fernández. Néstor Kirchner came to power in mid-2003 in the wake of the 2000-2001 economic
meltdown. He inherited Programa Jefes y Jefas de Hogar Desempleados, an emergency workfare
program with CCT elements. At its peak, Jefes covered 2 million families (Repetto and Díaz
Langou 2010, 6). As the economy recovered, most of those who lost their jobs during the crisis
graduated from Jefes. In 2004 Kirchner launched the Familias por la Inclusión Social CCT to deal
with the long-term poor who remained in Jefes. Familias never fully replaced Jefes; at its peak,
Familias covered roughly 700,000, while 550,000 remained in Jefes (Repetto and Díaz Langou
2010, 6).
Social Policy and Presidential Ideology in Latin America

119
In 2009, Fernández replaced both programs with an ambitious new CCT, Asignación
Universal por Hijo (AUH). The new program universalized the existing system of family
allowances, which previously covered only formal-sector workers (Bertranou and Maurizio 2012).
At launch, AUH benefited slightly more than 3 million families. Coverage has since averaged
about 3.5 million families.
AUH is substantially more generous than its predecessors. At the outset, it paid 180 pesos
per child per month for up to five children. This compares favorably with Jefes, which paid families
a flat 150 pesos per month, and Familias, which paid families a base of 155 pesos plus 30 pesos
per child per month for up to five children. And, in contrast to those programs, AUH stipends have
been adjusted fairly regularly to compensate for inflation and, following an adjustment in June
2013, surpassed their initial level in real terms.
Fernández has since launched two additional CCTs: Asignación por Embarazo, which
benefits new and expectant mothers conditional on medical checkups, was launched in 2011 and
covers 85,000 mothers (Infobae 2013). Progresar, which benefits 18- to 24-year-olds planning to
finish their schooling, was announced in early 2014.

6. Conclusions: Implications of Findings
Social policy in Latin America and the developing world more broadly has undergone a
“quiet revolution” in recent years (Ferreira and Robalino 2011). By targeting the poorest,
innovative anti-poverty initiatives such as CCTs have succeeded in reducing poverty and
inequality beyond what had been possible through the social policies that preceded them.
Moreover, CCTs have accomplished this despite costing only a fraction of what is spent on the
social security programs that dominate the region’s social policy budgets. The affordability and
Social Policy and Presidential Ideology in Latin America

120
overwhelming popularity of CCTs among beneficiaries and politicians alike increases the
likelihood that they will persist over time and possibly even expand further instead of being rolled
back as economic times get tougher.
The statistical analyses conducted in this chapter reveal that although presidential ideology
influences the scope of CCTs, it has no bearing on decisions regarding whether to adopt this type
of policy. CCT adoption has primarily been a diffusionary process. Latin American governments
were influenced by the perceived success of CCTs implemented by their regional neighbors and
in particular the two largest and most influential countries in the region. Technocrats and public
intellectuals involved in the design of pioneering CCTs such as Mexico’s Santiago Levy
(Ancelovici and Jenson 2013; Brearly 2011) and Brazil’s Eduardo Suplicy (Lana and Evans 2004)
heavily promoted this new approach to combating poverty and increasing human capital.
Multilateral banks wholeheartedly embraced the ideas behind CCTs and provided substantial
technical and monetary support to governments looking to create their own programs (Ancelovici
and Jenson 2013; Lana and Evans 2004; Martínez Franzoni and Voorend 2011; Sugiyama 2011).
In sharp contrast to the adoption of CCTs, the country-specific design of CCTs was heavily
influenced by domestic politics. Specifically, presidential ideology played a key role in
determining program scope. True to their ideological orientation, presidents of the left moved
farther than their centrist and rightist counterparts toward establishing a basic income floor for all
members of society independent of employment status. Furthermore, the case studies provide
evidence of left governments working to raise the real value to CCT stipends. Ironically, CCTs,
which by virtue of being means-tested, represent the antithesis of a decommodifying social
democratic welfare regime (Esping-Andersen 1990), have, when implemented by left-leaning
presidents, represented major steps towards establishing more inclusive welfare regimes in Latin
Social Policy and Presidential Ideology in Latin America

121
America. Under the left, CCTs have served as building blocks for the establishment of basic
universalism in the region (Filgueira et al. 2006).
To my knowledge, this is the first systematic cross-national analysis of the determinants of
social assistance coverage levels among developing countries. This analysis of CCT coverage
provides evidence that even in the era of globalization, ideology continues to shape social policy
in developing countries. These findings lend support to recent work on Latin America by Huber
and Stephens (2012), which, despite failing to find evidence of an effect of ideology on social
spending (see also Borges 2014), finds that left leaning governments have had a negative effect on
both poverty and inequality. Because CCTs are among the most effective tools for income
redistribution available to Latin American policymakers (Goñi, López and Servén 2008), CCT
coverage can be viewed as a proxy for redistribution. Thus, the robust and substantively large
positive effect of ideology on coverage found in this chapter is evidence that the left has engaged
in substantial redistribution.
Taken together, this chapter and the one that precedes it reveal that aggregate spending
numbers mask the substantial autonomy Latin American presidents have in shaping social policy.
Without substantially increasing total social spending, left-leaning presidents have modified social
policy in their countries to achieve substantial reductions in poverty and inequality and carry out
significant income redistribution in favor of the poor. CCTs have played a decisive role in making
these accomplishments possible.
Social Policy and Presidential Ideology in Latin America
122
Chapter 5.
Ideology and Varieties of CCTs:
Adoption and Design of CCTs in Three Latin American Countries

1. Introduction
This chapter conducts in-depth case studies of CCT adoption and design in three Latin
American countries – Argentina, Bolivia and Costa Rica – during the second half of the 2000s.
The case studies are based on elite interviews conducted during field research in those countries.
Whereas the last two chapters used statistical methods for hypotheses testing, this chapter utilizes
case studies to accomplish two tasks: 1) assess the plausibility of the previous chapter’s findings
and 2) construct an improved theory of CCT adoption and design.
The first task seeks to complement the preceding statistical analyses. This is done by using
case studies to flesh out the relationships observed in said analyses. Process tracing (George and
Bennett 2005) will be used to map out the political processes – the actors, their motivations and
the complex sequencing of events – that led to CCT adoption and influenced CCT design. Within-
case process tracing makes it possible to go beyond identifying correlations among variables and
delve into the actual decision-making process.
The second task involves using the case studies to go beyond the statistical analyses and
their inherent constraints in order to engage in theory building. Whereas the previous chapter
limited its analysis of ideology’s effect to the ideology of the president in office, the case studies
also explore the effects of the ideologies of prominent opposition politicians at both the local and
national levels. While incapable of adopting a national-level CCT, opposition politicians can (and
did) place CCTs on the political agenda. In doing so, they may set into motion a process that
Social Policy and Presidential Ideology in Latin America
123
culminates with the decision by a president (left-leaning or otherwise) to adopt nation-level CCT.
This is precisely what happened in Argentina and Bolivia, where, as the case studies will show,
centrist and right-leaning politicians put CCTs on the agenda by enacting popular local-level
programs and proposing national-level CCTs in their campaign platforms. Only later did left-
leaning presidents adopt national CCT programs.
Further, whereas the previous chapter, as a result of data availability, analyzed CCT design
exclusively in terms of scope (measured in terms of coverage), this chapter utilizes more nuanced
indicators, namely the approach to targeting, stipend structure and conditionality enforcement. The
constellation of these three factors produces two distinct varieties of CCTs. The first type, human
capital CCTs, such as those found in Mexico, Colombia and Costa Rica, emphasize the role of
transfers and conditionality as tools to promote human capital accumulation and, over the long
run, prevent poverty. In contrast, basic income CCTs, such as those found in Brazil, Argentina and
Bolivia, emphasizes the role of transfers as a tool to relieve poverty.
The most notable findings of this chapter have to do with presidential ideology. Contrary
to the recent scholarly consensus (Díaz-Cayeros and Magaloni 2009; Sugiyama 2011), the cases
reveal that ideology does matter for CCT adoption. The left’s initial position on CCTs in the three
countries studied ranged from apathetic (Costa Rica and Bolivia) to outright opposition
(Argentina). In the three countries, politicians of the center and right put CCTs on the political
agenda. As noted, in Argentina and Bolivia the left implemented CCTs at the local level and
incorporated them into their national campaign platforms before the left seriously considered them.
There is also evidence that presidential ideology determines CCT design. Left-leaning presidents
in Argentina and Bolivia adopted CCTs that emphasized poverty reduction and broad coverage
over human capital accumulation and targeting of those most in need.
Social Policy and Presidential Ideology in Latin America
124
This chapter proceeds as follows. Section 2 assesses findings and limitations of the
previous chapter’s statistical analyses. The section questions the usefulness of diffusion as an
explanation for policy adoption: that Latin American leaders borrowed policies from abroad does
not negate that adoption was explained by domestic politics. This section goes beyond the previous
chapter’s one-dimensional measure of CCT design by introducing the notion that there are actually
two distinct varieties of CCTs. Section 3 highlights the usefulness of in-depth process tracing as a
complement to statistical analyses and justifies the selection of Argentina, Bolivia and Costa Rica
as cases for in-depth study. The section that follows presents the case studies. The chapter
concludes by taking stock of the findings of the two chapters on CCTs and the existing literature
on the politics of CCTs to sketch out a more complete theory of the relationship between
presidential ideology and CCT adoption and design.

2. State of the Knowledge
The previous chapter provided strong evidence that CCT adoption has been an example of
policy diffusion. As the number of countries in the region with CCTs increased, so did the
probability that countries that had resisted adopting this type of program would follow their
neighbors in adopting one. This finding is neither new – Sugiyama (2011) found earlier cross-
national evidence of this – nor surprising. In fact, it would actually be surprising if Latin American
countries were not, given the high level of cross-national interaction that exists among elites and
civil society, aware of what their neighbors are doing and thus failed to imitate successful policy
innovations that emerged elsewhere in the region. By the mid-2000s, every Latin American
policymaker was familiar with the success of Mexico’s Progresa/Oportunidades and Brazil’s Bolsa
Social Policy and Presidential Ideology in Latin America
125
Escola/Bolsa Familia. The evaluations of these programs (Aguiar and Araujo 2002; Skoufias and
Bonnie McClafferty 2001) had been widely disseminated.
Democratically elected leaders around the region faced the same political incentives: voters
demanded improvement in social policy and would likely reward governments that addressed these
demands. There was also a unanimous consensus in favor of improving human capital. CCTs
potentially offered a way of tackling poverty and human capital simultaneously. Moreover, the
technical and monetary constraints on implementing these policies were relatively low. As has
been widely documented, international financial institutions (IFIs) wholeheartedly embraced
CCTs and were willing to provided technical and monetary support to governments looking to
adopt their own programs (Ancelovici and Jenson 2013; Lana and Evans 2004; Martínez Franzoni
and Voorend 2011; Sugiyama 2011).
Chapter 3 also conducted an analysis of CCT coverage, both as a share of the total
population and relative to the size of the population living in poverty. The key finding involved
ideology: coverage tends to be substantially greater in countries with left-leaning presidents. Thus,
while they share the same basic design and goals, there is substantial variation in the design of the
region’s CCTs.
A major challenge for diffusion research is knowing where to draw the line between an
imported policy and indigenous one that greatly resembles an imported one. On this topic,
Weyland (2005, 17) makes a distinction between model and principle diffusion. The former occurs
when countries “import a neat, concrete, well-defined blueprint, largely replicating the original
model.” Under principle diffusion the emulator adopts the general principle of an innovator’s
policy but does not copy the policy wholesale. Weyland goes on to suggest that the diffusion of
CCTs in the region constitutes an example of model diffusion of Brazil’s Bolsa Familia. The
Social Policy and Presidential Ideology in Latin America
126
following section challenges this notion by demonstrating that there in fact two different types of
CCTs that have diffused throughout the region.
2.1. Varieties of Conditional Cash Transfer Programs
Though all CCTs seek to simultaneously reduce poverty in the short-term and increase
human capital over the long-term, there exists an inherent tension between these two goals (Handa
and Davis 2006; Veras Soares, Perez Ribas and Guerreiro Osório 2010). Prioritizing one goal may
hamper achievement of the other. If the main priority is to reduce poverty as much as possible, it
will be counterproductive to revoke the benefits of non-compliers, which tend to be the most
vulnerable members of society. Given the robust evidence of the effectiveness of conditioning
benefits (Baird, McIntosh and Özler 2011; de Brauw and Hoddinott 2011; Schady and Araujo
2008), when human capital accumulation is the main priority, it makes sense to take a more
draconian approach to enforcement. CCT design can be tailored to prioritize one objective over
the other (de Janvry and Sadoulet 2006).
76

Thus, it is possible to classify CCTs based on whether they prioritize poverty reduction or
human capital development. Based on that criteria, Cecchini and Martínez (2011) distinguish
between CCTs with weak and strong conditionality.
77
However, the differences between the two
types go beyond conditionality and extend to their approaches to targeting, determining stipends
and, ultimately, the very nature of their mission. For that reason, in this chapter I make a distinction
between “human capital” and “basic income” CCTs.
2.1.1. Pioneer Programs: Oportunidades vs. Bolsa Familia
While the two types of CCTs discussed above represent ideal types, Mexico’s
Progresa/Oportunidades, and Brazil’s Bolsa Escola/Bolsa Familia respectively constitute
prototypical examples of these two types.
78
Although the two programs emerged at relatively the
Social Policy and Presidential Ideology in Latin America
127
same time, there was little crosspollination between the two (Sugiyama 2011, 42). This is
explained by the fact that these programs were products of different intellectual traditions. The
former was the product of poverty research on the interaction between nutrition, health and
education. The latter emerged from the basic income movement, which proposed an income floor
for all citizens beginning with families with children.
For Santiago Levy (2006, 10), the architect of Mexico’s Progresa/Oportunidades and
widely credited as the “father of CCTs,” these programs represent the culmination of advances in
scholarly understanding of poverty made during the 1980s and 1990s. Scholars adopted a holistic
understanding of the causes of poverty that emphasized the links between food consumption,
nutrition, health, and education. Providing any of these benefits in isolation, it was argued, would
likely be insufficient to pull families out of poverty.
As Levy was developing what in 1997 became Progresa, a different approach to using cash
transfers to combat poverty was gaining ground in Brazil. In 1991, the same year that Levy (1991)
published a World Bank research paper detailing many of the ideas that came to be part of
Progresa, Eduardo Suplicy, a Senator from the left-wing Partido dos Trabalhadores (PT), proposed
an unconditional basic income guarantee for all Brazilians over 25. The idea, though popular
among legislators, was never implemented. In 1993, economist Jose Marcio Camargo refocused
the proposal toward preventing the inter-generational transmission of poverty. Instead of awarding
transfers to adults, he proposed targeting families with school-age children and conditioning
benefits on school attendance (Bandeira Coelho 2012; Sugiyama 2013). These ideas would form
the basis for a series of sub-national CCTs that quickly spread thought Brazil starting in 1995 and
culminated in the creation of a Bolsa Escola Federal in 2001. Bolsa Escola was expanded and
substantially redesigned under the name Bolsa Familia in 2003.
Social Policy and Presidential Ideology in Latin America
128
The two types of CCTs differ in terms of their approaches to targeting, conditionality and
stipends. Whereas human capital CCTs prioritize minimizing errors of inclusion, basic income
CCTs place more emphasis on avoiding errors of exclusion. Looking once more at the pioneer
programs, whereas Oportunidades relies on a stringent proxy means test based on observable
household characteristics, Bolsa Familia selects beneficiaries based on self-reported incomes
(Lindert et al. 2007; Veras Soares, Perez Ribas and Guerreiro Osório 2010).
79
As a result,
“Oportunidades has more efficient targeting than Bolsa Família, but at the price of the program
covering fewer poor households” (Veras Soares, Perez Ribas and Guerreiro Osório 2010, 177).
The two models also pursue differing approaches to conditionality enforcement. Under
Oportunidades, non-compliance leads to the immediate interruption of benefits until the family is
certified as being compliant. Under Bolsa Familia, payments are suspended for the month but the
stipend accumulates and is reimbursed once the family resumes compliance (Cecchini and
Martínez 2011, 105). As Veras Soares (2011, 58) notes, under Bolsa Familia “noncompliance with
co-responsibilities is a sign that a family faces an additional vulnerability. A social worker should
verify the reasons for the family failing to uphold the co-responsibilities and also help it to
overcome these difficulties.”
Oportunidades and Bolsa Familia also differ in terms of stipend structure. Because the
opportunity costs associated with choosing school over work increase as the student becomes
older, Oportunidades stipends increase with each grade. And, because women have historically
received less schooling than men, they are awarded more generous stipends beginning in seventh
grade. In addition, students receive an additional stipend for school supplies. In line with its roots
in the basic income movement, Bolsa Familia offers an unconditional base stipend for all families
living in extreme poverty, regardless of whether they have school-age children. An additional CCT
Social Policy and Presidential Ideology in Latin America
129
component provides a flat stipend for each child under 15. Only in recent years was the stipend for
16 and 17 year-olds made larger than the one paid to children. Furthermore, the program includes
a supplemental subsidy determined on a case-by-case basis for families that fall below the extreme
poverty line after receiving Bolsa Familia.

3. Research Design
In-depth case studies can serve as a powerful complement to statistical analysis. This
chapter complements the region-wide statistical analyses conducted in the previous chapter with
in-depth case studies tracing the political processes that culminated in the adoption and shaped the
design and evolution of CCTs in Argentina, Bolivia and Costa Rica. The case studies will be used
to assess the plausibility and flesh out the relationships between variables observed in the statistical
analyses (Lieberman 2005). For all their strengths in terms of generalization and precision,
statistical methods are incapable of mapping out the complex sequence of events – the actors, their
motivations and the complex sequencing of events – that culminate in policy adoption. Within-
case process tracing makes it possible to go beyond a correlational analysis and investigate the
decision-making process through which the variables of interest led to the observed outcomes
(George and Bennett 2005). It was, after all, people, not variables, empowered and constrained by
availability of resources, existing institutions and other political actors, who made political
decisions regarding CCTs.
Qualitative methods are necessary to explain the two varieites of CCTs discussed in the
previous section. To statistically assess the determinants of policy adoption in the previous chapter,
it was necessary to utilize a dichotomous dependent variable – either the country had a CCT or it
did not. This one-size-fits-all dependent variable is problematic if one believes that there are in
Social Policy and Presidential Ideology in Latin America
130
fact different types of CCTs. Lieberman (2005, 440) also cites as a strength of small-n analysis the
ability to answer questions left open in large-n analysis when there is insufficient data to assess
statistical relationships. Coverage was the only cross-nationally comparable measure of CCT
design available for statistical analysis.
80
Thus, qualitative analysis is the only option available for
assessing the effect of ideology on the components of CCT type, namely conditionality, targeting
and stipend structure.
Qualitative methods can also reveal key factors absent in statistical analyses. That a given
variable, say for example political competition or economic crises, is not a robust determinant of
CCT adoption at a cross-national level does not rule out the possibility that it may have played a
role (perhaps even a decisive one) in a specific country (Flores-Macías 2012, 59).
3.1. Selecting Cases in the Context of Diffusion
Based on more than 40 interviews with political elites, social policy technocrats and
independent policy analysts, this chapter analyzes the political processes that led to the adoption
and shaped the design of CCTs in Argentina, Bolivia and Costa Rica. Focusing on Latin America
makes it possible to keep constant a wide range of potentially confounding factors, most notably
culture, historical trajectory and political institutions. Furthermore, by including one of the richest
countries in the region (Argentina), one that ranks near the middle (Costa Rica) and one of the
poorest (Bolivia), as well as countries from the Southern Cone (Argentina), Andean region
(Bolivia) and Central America (Costa Rica) my cases approximate a representative sample of Latin
America (see Table 5-1).
Social Policy and Presidential Ideology in Latin America
131
Table 5-1: Case Selection Criteria

Second Generation CCTs Pioneer CCTs
Argentina Bolivia Costa Rica Mexico Brazil
Asignación
Universal por Hijo
Bono Juancito Pinto Avancemos Progresa
(later Oportunidades,
now Prospera)
Bolsa Escola
(now Bolsa Familia)
Context
Sub-Region Southern Cone Andes Central America --- ---
GDP per Capita PPP (2010) $15,091 $4,549 $11,339 $13,932 $11,314
Welfare Regime
1
Conservative/
Corporatist
Informal/
Familialist
Embryonic Social
Democratic
Conservative/
Corporatist
Conservative/
Corporatist
CCT Design Features
Year of Adoption 2009, small
predecessor in 2005
2006 2006 (Pilot), 2007
(Nationwide)
1997 (Rural Only),
2001 (Entire Country)
2001, Significant
Expansion post-2003
Coverage as %
of Population (2013)
9% 21%
2
4% 21% 26%
3

Target Population Children of Informal-
Sector/Unemployed
Children in
Public School
Poor Children
in High School
Poor Children Poor Children,
Childless Extreme
Poor
Targeting Mechanism Public Schools/No
Family Allowances
Public Schools Proxy Means-Test Proxy Means-Test Self-Reported Income
Conditionality Enforcement 80% Unconditional,
20% Conditional
Nominally CCT,
No Enforcement
Lax, Getting Stricter Strict Lax, Getting Stricter,
but Still Forgiving
Stipend Structure Uniform Uniform Variable by Grade Variable by Grade,
Sex
Uniform, except for
16-17
Independent Variables
Initiating Pres. Ideology Center-Left/Left Left Center Center Center
Political Competition
at Launch
Heightened Low, Potentially
Increasing
4

Low High High
% Years of under Left
(CCT Launch-2013)
100% 100% 0% 0% 85%
Ethnic Boundaries Extremely
Weak/Nonexistent
Mostly
Strong/Multiple
Extremely
Weak/Nonexistent
Intermediate
Levels
Mostly
Weak/Permeable

1
. See Huber (1996) and Martínez Franzoni (2008).
2
. At the time, the program only covered up to Grade 8.
3
. Estimate.
4
. While politically dominant, President Evo Morales faced a constant series of elections in the years following CCT adoption.
Social Policy and Presidential Ideology in Latin America
132
The previous chapter found evidence that policy diffusion helps explain CCT adoption.
This finding presents methodological challenges to the selection of cases for in-depth qualitative
study. As Starke (2013, 569) notes, “qualitative diffusion scholars cannot use Mill’s methods of
agreement and difference (also known as the ‘most similar cases design’) as these designs do not
allow for interdependence of cases.” For this type of study, Starke (2013) recommends the non-
random selection of a handful of diverse cases (Seawright and Getting 2008). Starke (2013, 570)
recommends cases that exhibit a “large degree of divergence on central conditioning factors of
diffusion such as geographical location or factors that influence the propensity of policy learning
such as administrative professionalism.”
Chapter 3 demonstrated that, although presidential ideology has no effect on the probability
of CCT adoption, it does have an important effect on the scope of CCTs. The cases selected vary
with regard to the ideology of the presidents in charge of CCT adoption and implementation.
Whereas Argentina and Bolivia were governed by left-leaning presidents during the entire period
of study, Costa Rica was governed by centrist presidents. In line with that chapter’s finding on the
effect of ethnic diversity of CCT coverage, this chapter analyzes one of the most ethnically divided
countries in the region (Bolivia) and two of the most homogenous (Argentina and Costa Rica).
The three countries also vary greatly with regard to their social policy trajectory. Argentina
is a regional pioneer that began its social policy expansion during the early 20
th
century. Costa Rica
is a late bloomer that dramatically expanded social services midcentury. Bolivia remains a laggard
by regional standards. The three countries vary with regard to their welfare regimes (Esping-
Andersen 1990). Argentina’s regime, under which most benefits are tied to employment status,
most closely resembles a conservative corporatist welfare regime, with the caveat that a large share
of the population – workers in the informal sector – has historically been excluded from social
Social Policy and Presidential Ideology in Latin America
133
security (Huber 1996, 159). Costa Rica possesses the region’s only “embryonic social democratic
regime” in which most social services constitute rights unrelated to employment status (Huber
1996, 159). To the extent that it is possible to speak of a Bolivian welfare regime, it is what
Martínez Franzoni (2008) has termed “informal-familialist,” meaning that social programs are
limited in scope and exclude much of the population. Given the absence of state involvement,
families step in to fill the vacuum.
To facilitate comparison, the three cases also span roughly the same time period. Bolivia
adopted its program in the second half 2006. This was also when Costa Rica launched pilot version
of its program. The nation-wide version launched in early 2007. While most of discussion of
Argentina will focus on Asignación Universal por Hijo (AUH) adopted in 2009, said program
marks the continuation and expansion of Familias por la Inclusión Social, which was launched in
2005 but did not cover a substantial portion of the population until 2007.
Furthermore, given my use of elite interviews, focusing on recently launched programs
offers practical advantages. Time is of the essence when attempting to reconstruct policy adoption
processes. Elites involved in policy adoption form part of ad-hoc networks that disappear after a
policy transfer occurs or fails (Evans and Davies 1999, 375; Lana and Evans 2004, 194). Studying
recent programs maximizes the likelihood of identifying and accessing key policymakers involved
in the design and adoption of CCTs. Given people’s cognitive limitations, selecting recent cases
increases the reliability of the information obtained through interviews. And, since I relied heavily
on snowball sampling to select interview sources, recent cases increased the likelihood that
interview sources could guide me to other knowledgeable sources.

Social Policy and Presidential Ideology in Latin America
134
3.2. Varieties of CCTs Revisited
The cases also vary with regard to type of CCT. Whereas Argentina and Bolivia’s programs
prioritize poverty reduction, Costa Rica’s prioritizes human capital accumulation. These differing
approaches are evident in the first sentences of the presidential decrees – all three programs were
launched via decree – that gave them life.
81
The first paragraph of the 2006 the Bolivian decree
starts by referencing the “objective of eradicating extreme poverty and the exclusion of people,
families and communities” (Presidencia Bolivia 2006). Only in the second paragraph is there a
mention of “programs and projects destined to strengthen, protect and develop human capabilities,
primarily favoring the country's children.” Similarly, the 2009 Argentine decree begins by
referencing the government’s responsibility to provide “policies that will improve the situation of
children and adolescents facing economic vulnerability” (Presidencia Argentina 2009). It is not
until the seventh paragraph that education is mentioned and even then, it is in the same sentence as
the right to social security coverage. In contrast, the Costa Rican decree begins by stating the goals
of “universalizing secondary education” and increasing education spending (Presidencia Costa
Rica 2006). Poverty is first mentioned in the third article in reference to how increased education
reduces its likelihood.
These differing objectives determine the population targeted by the programs. Whereas
basic income CCTs aim to cover most if not all children, human capital CCTs cover a more
restrictive sample. Among the three countries, only Costa Rica selects beneficiaries using a
rigorous proxy means-test. Relative to its levels of economic and social development, Costa Rica
possesses a strikingly high rate of secondary school desertion (Román 2010). The country’s CCT
was designed specifically to address that problem. As the country has universal enrollment at the
elementary school level, it was not deemed necessary for the program to cover elementary school
Social Policy and Presidential Ideology in Latin America
135
students. In contrast, Argentina, which also has universal elementary school enrollment, covers all
children under 18. Bolivia, which does have significant desertion at all levels, initially only
covered students up to grade five but has since gradually expanded coverage to include all grades.
The programs also differ with regard to the structure of stipends and, to a lesser extent,
conditionality enforcement. With regard to stipends, Costa Rica follows Mexico in offering an
increasing amount with each grade so as to compensate students for income lost while attending
school (Sauma 2013).
82
In contrast, the other countries pay a flat stipend per student, regardless of
grade and sex. No public justification has ever been provided for the stipend amounts. Responding
to critiques for lax enforcement, Costa Rica adopted an electronic system that has improved
enforcement. Argentina adopted a semi-conditional approach (Bertranou and Maurizio 2012) in
that 80 percent of the stipend is unconditional while the remaining 20 percent can only be cashed
at the end of the school year by proving school attendance. While attendance is formally a
requirement in Bolivia’s program, conditions are not clear and conditionality is not enforced.

4. Case Studies
4.1. Argentina’s Asignación Universal por Hijo
Described as “the most important social right created since the return of democracy in
1983” (Etchemendy and Garay 2011, 296), Argentina’s Asignación Universal por Hijo (AUH) is
distinct in that it is both a CCT and an attempt to extend to informal-sector workers the existing
system of family allowances, which had previously been available only to formal-sector workers.
Thus, the program, launched at the end of 2009, represents a concerted attempt to break with Latin
America’s “bifurcated” approach to social policy characterized by permanent and, in some cases,
Social Policy and Presidential Ideology in Latin America
136
generous benefits for formal-sector employees and modest and often transitory social assistance
for labor market outsiders (Barrientos and Claudio Santibáñez 2009, 2).
Figure 5-1: Cash Transfer Programs
in Argentina, 2002-2012

Source: Repetto and Díaz Langou (2010), ANSES

AUH is not explicitly means tested, rather it is open to children whose parents are employed
in the informal sector. Because workers in the informal sector are overwhelmingly poor, the
program is, in practice, well targeted (Salvia, Musante and Jaramillo 2013, 14). By the end of
2013, the program reached 3.3 million beneficiaries or roughly seven percent of the country’s
population (see Figure 5-1). AUH is also one of the most ambitious programs in region in terms
of the generosity of its stipends. The monthly stipend per student is currently 460 pesos (dollars),
which over the course of a year amounts to XX of GDP per capita.
83

AUH and its sister programs Asignación por Embarazo, which pays new and expectant
mothers stipends conditional on attending medical checkups, and Progresar, a new CCT that will
pay unemployed 18-24 years-olds to finish high school or enroll in technical education (the so-
called ninis)
84
, were implemented under president Cristina Fernández de Kirchner (2007-present),
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Jefes y Jefas Plan Familias
AUH Seguro de Capacitación
Number of Beneficiaries (Millions)
Social Policy and Presidential Ideology in Latin America
137
a left-wing politician form the Peronist Partido Justicialista Party (PJ). As in Brazil and, as will be
discussed, Bolivia, the left was not responsible for putting CCTs on the policy agenda and, in fact,
Férnandez and her late husband and predecessor Néstor Kichner (2003-2007) initially opposed the
idea. CCTs were thrust into the national agenda by a coalition of labor unions and academics in
the context of the country’s 2001-2002 economic meltdown and later implemented at the local
level by the center-right government of the city of Buenos Aires in 2005. Ultimately, in response
to a worsening economy and increasing political competition, Férnandez adopted AUH.
4.1.1. The Context
AUH cannot be understood without delving into the economic and, subsequently, political
crisis the country faced during 2001-2002.
85
Argentine GDP fell by over 20 percent during 1999-
2002 (World Bank 2005, 1). In 2002 alone, three-quarters of households experienced a decline in
real income with nearly two thirds suffering declines or 20 percent or more (McKenzie 2004, 720).
More than half of workers suffered wage cuts and a quarter of the population was unemployed
(Agis, Cañete and Panigo 2011, 33).
The economic crisis triggered a political crisis. Massive protests against the government
(cacerolazos) became a regular occurrence. In mid-December 2001, 2.7 million Argentines voted
in an informal referendum on a proposal similar to what would become AUH (Rossi 2013, 145).
However, before the Frente Nacional contra la Pobreza (FRENAPO/National Front against
Poverty), the coalition of labor unions and unemployed workers organizing the referendum, could
present the results to the government, President Fernando de la Rúa (1999-2001) of the centrist
Alianza coalition resigned amid protests. During the span of two weeks, the presidency changed
hands three times before landing on Eduardo Duhalde (2002-2003), a Peronist.
Social Policy and Presidential Ideology in Latin America
138
Duhalde is best remembered for defaulting on Argentina’s foreign debt, devaluing the peso
and launching Programa Jefes y Jefas de Hogar Desempleados (Unemployed Heads of Households
Program), a workfare-based program with CCT elements, which, up until that point, was the most
ambitious anti-poverty initiative in the country’s history. Jefes paid unemployed heads of
households a flat 150-peso ($50 at the time of launch) stipend in exchange for 20 hours of
community labor and conditional of their children remaining in school.
J efes succeeded in restoring social peace (Golbert 2004). The program’s main strengths
were its scope – at its peak in mid-2002 it reached 2 million people, roughly 20 percent of
households (Golbert and Giacometti 2008, 29) – and its targeting – about 90 percent of
beneficiaries had incomes below the poverty line (Galasso and Ravallion 2003). The stipend was
also generous at launch, amounting to half the cost of a basic basket of goods and 20 percent of
the poverty line (Golbert and Giacometti 2008, 33). However, the program was not very successful
at reducing long-term poverty (Galasso and Ravallion 2003; Lo Vuolo 2008). Restrictions on
program size – enrollment was open only for a month between April and May of 2002 and never
reopened – left 80 percent of eligible adults excluded (Galasso and Ravallion 2003, 372). The flat
stipend limited the program’s effect on large families. There was also substantial leakage of
benefits: one-third of participants, primarily married women, were not in fact “unemployed heads
of households” (Galasso and Ravallion 2003, 395-96; Monza and Giacometti 2003). There were
also widespread allegations of politization (Galasso and Ravallion 2003; Giovagnoli 2005;
Giraudy 2007; Golbert 2004; Lo Vuolo 2008).
Néstor Kirchner was elected president in mid-2003 in the aftermath of the crisis. Scholars
of Latin American politics struggle to categorize the ideology of the Kirchners and the Peronist
Party to which they belong. The Kirchners do not fit well into the dichotomized view of the
Social Policy and Presidential Ideology in Latin America
139
region’s left popularized by Castañeda (2006) in which leaders are either moderate social
democrats or radical populists. Much of the confusion owes to the fact that, although the Kirchners
have governed from the left, their party is not of the left but, rather, a populist machine without a
set ideology (Etchemendy and Garay 2011; Levitsky 2003).
As the economy began recovering in 2003 and boomed after 2005, the majority of
beneficiaries “graduated” from Jefes by moving to formal-sector jobs. This raised questions about
with the long-term poor that remained in the program. Kirchner and his team were reportedly not
fond of Jefes or the targeted anti-poverty programs that were spreading across Latin America
(Cortés interview; Golbert interview; Vinocur interview). As Peronists, the Kirchners firmly
believed that poverty should be tackled through the creation of formal employment and universal
social policy. Opposition to CCTs has strongest among Alicia Kirchner – Néstor’s sister and Social
Development Minister during most of the Fernández administration. She repeatedly derided CCTs
as “neoliberal,” “impositions from IFIs,” or “pre-packaged programs” not suited for Argentina’s
reality (Amadeo interview; Gasparini interview; Vinocur interview).
At first, Jefes was allowed the program to fade way: enrollment was never reopened and
the stipend was not adjusted despite inflation averaging 10 percent a year. Bu ultimately two
successor programs, restricted to original Jefes beneficiaries, were created. Beneficiaries deemed
to have favorable employment prospects would be transferred to Seguro de Capacitación y Empleo
(Training and Employment Insurance), which provided training in addition to a stipend. Those
with less favorable prospects (primarily single mothers) were to be transferred to a new CCT, Plan
Familias.
86
The launch of these programs was mired by setbacks and, as a result, they never fully
replaced Jefes (Cruces and Gasparini 2008). In 2009, shortly before being replaced by AUH, Plan
Familias had roughly 700,000 beneficiaries while 550,000 remained in Jefes.
Social Policy and Presidential Ideology in Latin America
140
The Kirchners presided over a dramatic expansion in Argentina’s social policy. Starting in
2004, Néstor Kirchner gradually doubled the share of the population covered by pensions to near
universal levels (Etchemendy and Garay 2011; Golbert and Roca 2010). Kirchner and later
Fernández were also responsible substantial increases in minimum wages and pensions (Golbert
and Roca 2010). However, the reticence to target the poor limited how much the poor benefited
from Argentina’s recovery. Levitsky and Murillo (2008, 28) go so far as to assert that:
“In the area of social policy, Néstor Kirchner missed an opportunity to redistribute
wealth and reduce income inequality. Though widely considered left-of-center, the
Kirchner government neglected social policies aimed at combating poverty.
Indeed, despite unprecedented fiscal health, the government did not invest heavily
in either conditional cash transfers to the poor, or health and education programs
for them. (…) Consequently, although unemployment and poverty rates declined
sharply under Kirchner, these declines were rooted almost entirely in economic
growth. In fact, levels of poverty and inequality remained higher in 2007 than they
were during the mid-1990s.”
4.1.2. The Adoption Process
Fernández eventually succumbed to the appeal of targeted programs. The widespread
regional adoption of CCTs and, more specifically, the universalization of family allowances in
neighboring Uruguay two years earlier (Rofman interview; Straschnoy 2011), may have provided
templates for AUH. However, domestic political factors are what explains why Fernández adopted
the types of policies her husband had rejected. Granted, the 2008 re-nationalization of the pension
system provided the resourced needed to finance AUH (Arza 2012; Etchemendy and Garay 2011;
Social Policy and Presidential Ideology in Latin America
141
Ewig and Kay 2011). However, the decision to adopt was explained by factors: a deteriorating
economy and increased political competition.
87

AUH represented an about-face for the Kirchner governments. Proposals for large-scale
cash-transfer programs presented to the government in the mid-2000s were roundly rejected
(Vinocur interview). However, as the economy slowed in 2009, Fernández came to accept that,
despite years of fast growth, informal labor was not disappearing and that some families would
remain poor even if the head of household was employed (Repetto and Potanza 2011; Roca
interview; Rofman interview).
The political context was equally, if not more, important. AUH was launched during a
moment of great weakness for Fernández. By the end of 2009, two-thirds of Argentines
disapproved of her government and less than 20 percent thought the country was headed in the
right direction (Catterberg and Palanza 2012, 22). Declining poll numbers help explain FPV’s
disastrous showing in the June 2009 mid-term elections, where it temporarily lost its legislative
majority. Thus, as was the case in Brazil (Bandeira Coêho 2012; Garay 2010; Melo 2008) and
Mexico (Díaz-Cayeros, Estévez and Magaloni Forthcoming; Dion 2010; Garay 2010), the
adoption of Argentina’s flagship anti-poverty CCT was a response to increasing political
competition.
Although the opposition was highly fragmented, it was united in its desire to establish a
permanent large-scale anti-poverty program. As noted, prior to the economic meltdown, this idea
had gained traction among civil society groups organized under FRENAPO as well as in academic
circles (Lo Vuolo et al. 1999, Ch. 10; Roca 2011; Rofman, Grushka and Chebez 2001; Vinocur
interview).
88
And, since 2005 the government of the city of Buenos Aires had operated Ciudadanía
Porteña, a CCT that on the eve of AUH’s launch benefited 70,000 homes (MPT 2011, 29).
89

Social Policy and Presidential Ideology in Latin America
142
The creation of an unconditional income floor for children was the central plank of the
presidential campaign of Elisa Carrió, the runner-up in the 2007 elections. Six similar bills were
being discussed in Congress prior to the mid-term elections (Repetto, Díaz-Langou and Marazzi
2009). Thus, a political consensus existed behind this policy. But rather than seeking a cross-party
agreement based on common interests, the government opted to “surprise” the opposition (Repetto
and Potenza 2011) and “beat it to the punch” (Kantor 2012) in order to get the full political credit
for the idea (Bertanou interview; Golbert interview; Repetto interview). AUH was launched by
decree one month before the new legislators took office.
While the political will for large-scale cash transfer program existed, it was made feasible
by the abundance resources provided by nationalized pension system. Nationalization, which
transferred the private system’s assets to the agency that would be in charge of AUH, the National
Social Security Administration (ANSES), allowed the government to hit the ground running.
Access to these funds marked a key difference between the AUH decree and bills in Congress,
which proposed financing the transfers through tax reforms (Barbeito and Lo Vuolo 2009; Repetto,
Díaz-Langou and Marazzi 2009).
The political process that culminated in the adoption of AUH was a closed-door, top-down
affair. As the World Bank’s Rofman explained: “It was not a transparent process. The president
decided. It was implemented. This is how things are decided in Argentina.” Lo Vuolo (interview)
adds: “It was a government decision. There were no actors at the table. The decision group was
very closed. It has even been said that the Labor Minister found out afterward.” Even within
ANSES, only “two or three people at the political level” knew about the program. The technical
team learned about it 45 days before the first payment (Lo Valvo interview)
90
.
Social Policy and Presidential Ideology in Latin America
143
Neither Congress nor civil society were asked for input (Kantor 2012). This was not an
exception, but rather the rule in Argentine policymaking. As Golbert (interview) explains:
“Congress has little say over social policy. It seldom discusses the topic. This government is also
very presidentialist.” Amadeo (interview) concurs: “Congress does not have much oversight power
over social policy.” De la O (Forthcoming) argue that increased legislative oversight explains the
noticeable reduction in discretion present in AUH relative to Jefes. The in-depth analysis presented
in this section contradicts her claim that legislators influenced the design of CCTs in Argentina.
4.1.3. Program Design
AUH differs from other CCTs in the region in that it is designed to act as an extension of
the existing social security system rather than a standalone program and in that is “semi-
conditional” (Bertranou and Maurizio 2012). With regard to the first point, the program awards
beneficiaries family allowances comparable to those offered to formal-sector workers. Established
in 1957, family allowances (asignaciones familiares) are monthly cash payments aimed at
compensating for the costs of raising children (Rofman, Grushka and Chebez 2001). In
incorporating labor market outsiders into existing contributory insurance, AUH marks a departure
from Latin America’s traditionally bifurcated approach to social policy. AUH is semi-conditional
in that 80 percent of the stipend is unconditional while the remaining 20 percent can only be
accessed at the end of the school year after the beneficiary proves he or she regularly attended
school and routine medical checkups.
91

Overall, despite not explicitly targeting on income, beneficiaries are overwhelmingly poor
(see Figures 5-2 and 5-3). In 2012, 43.2 percent of beneficiaries between six and 12 were from the
poorest quartile and 72.9 percent were from the poorest 50 percent. Among teenagers 13 to 17,
37.8 percent of beneficiaries were from the poorest quartile and 65.3 percent were from the poorest
Social Policy and Presidential Ideology in Latin America
144
50 percent (Salvia, Musante and Jaramillo 2013, 14).
92
Additionally, in recent years the share of
beneficiaries from the lowest income strata has increased.
Figure 5-2: Primary School Children
(Ages 6 to 12) Enrolled
in AUH by Income Level
Figure 5-3: Secondary School Children
(Ages 13 to 17) Enrolled
in AUH by Income Level
Source: Salvia, Musante and Mendoza Jaramillo (2013, 14).

The program currently covers 3.3 million people, down from a peak of 3.6 in May 2011.
93

Roughly 3 million joined the program on day one. While incorporating such a large number at
once was an impressive technical feat, it should be noted that many of those beneficiaries were
transferred over from existing social programs that were later phased out (Salvia, Musante and
Jaramillo 2013, 2). However, four years after launch, it is estimated that 20 percent of the country’s
minors (2 million people) remain excluded from the social security system (Salvia, Musante and
Jaramillo 2013, 4). This problem is particularly serious among the children in the poorest quintile
(see Figure 5-4).
There is no consensus on the reasons behind the high exclusion rates. Gasparini
(interview)
94
goes so far as to call this a “mystery.” It is likely that those excluded are the poorest
32.1%
39.1%
43.2%
35.5%
28.8%
29.7%
24.1% 22.7%
20.6%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2010 2011 2012
Very Low Low‐Medium Medium Medium‐High
30.2%
35.7%
37.8%
37.8%
35.9%
27.5%
21.5%
19.2%
26.4%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2010 2011 2012
Very Low Low‐Medium Medium Medium‐High
Social Policy and Presidential Ideology in Latin America
145
people in Argentina (Rofman interview
95
; Tuñón interview
96
), including children lacking proper
documentation or not enrolled in schools (Bustos, Giglio and Villafane 2012). Rofman (interview)
attributes the exclusion to the institutional culture of ANSES. He and every expert interviewed
praised the institution for its technical capability with most going so far as to assert that it was the
only government institution capable of setting up such a large and complex program on such short
notice. Rofman argues that because of its inexperience operating anti-poverty programs, ANSES
lacks the expertise and on-the-ground human resources needed to seek out families the poorest
families and reach the most remote parts of the country.
Figure 5-4: Percent of Argentine Minors (0 to 17)
Covered by Social Security Regimes by Income Quintile (2010)

Source: Bustos, Giglio, and Villafañe (2012, 31).

One key difference between AUH and the other two programs analyzed in this chapter is
that stipends have been adjusted to keep up with inflation. In fact, stipends have matched the
country’s actual inflation as measured by independent analysts and have greatly surpassed the
highly questionable official inflation numbers (The Economist 2012) (see Figure 5-5). It should,
20%
11%
9% 2%
1%
36%
16%
6%
4% 2%
31%
60%
63%
62%
50%
11%
8%
19%
33%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Q1 Q2 Q3 Q4 Q5
Excluded Asignación Universal por Hijo
Contributory Family Allowances Other Subsidies
Tax Deduction for High‐Income Earners Self‐Employed
Social Policy and Presidential Ideology in Latin America
146
however, be noted that two of the adjustments occurred shortly before elections. The adjustment
before the 2013 legislative elections was particularly large (35 percent).
Figure 5-5: Consumer Price Inflation and the Purchasing
Power of AUH (2009-2014)

Source: ANSES, INDEC, Trading Economics.

4.2. Bolivia’s Juancito Pinto
Named after a Bolivian boy believed to have sacrificed his life fending off Chilean troops
during the War of the Pacific (1879-1883), Bono Juancito Pinto (BJP) is Bolivia’s flagship CCT.
Like Argentina’s AUH, BJP is not explicitly mean-tested. But, in contrast AUH, which, by
targeting the families of workers in the informal sector, was implicitly designed to reach the poor,
the only criteria for benefiting from BJP is to be enrolled in a public school. While the program
nominally requires beneficiaries to attend school 80 percent of the time, there is no evidence that
conditionality is actually enforced, challenging its classification as a CCT program.
0.60
0.80
1.00
1.20
1.40
1.60
1.80
2.00
Nov‐09
Feb‐10
May‐10
Aug‐10
Nov‐10
Feb‐11
May‐11
Aug‐11
Nov‐11
Feb‐12
May‐12
Aug‐12
Nov‐12
Feb‐13
May‐13
Aug‐13
Nov‐13
Purchasing Power Relative to Unofficial CPI
Purchasing Power Relative to Official CPI
Presidential
Elections
Legislative
Elections
Social Policy and Presidential Ideology in Latin America
147
Given this de facto lack of conditionality, BJP comes closer than any other education-
linked cash transfer program in the region to being a large-scale basic income program. The
number beneficiaries more than doubled from about 1 million (10 percent of the population) at
launch in 2006 to an estimated 2.2 million (about 20 percent of the population) in 2014 (see Figure
5-6). While the program initially covered only up to grade 5, it has gradually expanded and will,
for the first time in 2014, cover all grades. BJP has, however, been widely criticized for failing to
reduce poverty and increase school enrollment.
Figure 5-6: Bolivia, Bono Juancito Pinto,
Number of Beneficiaries and Grades Covered (2005-2014)

Source: Ministerio de Educación de Bolivia

BJP and Bono Juana Azurduy, its sister CCT aimed at increasing healthcare usage among
infants and their mothers, were adopted and expanded under the left-wing administration of Evo
Morales (2006-present).
97
During the early 2000s, parties of the right, not the left, pioneered a
CCT at the local level and proposed a nationwide program. Like his left-leaning counterparts in
Brazil and Argentina, Morales was initially ambivalent about CCTs.

0
2
4
6
8
10
12
0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
2.50
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014*
Millions of Benefiaries
Grades Covered Number of Beneficiaries
Numberof Grades Covered
Social Policy and Presidential Ideology in Latin America
148
4.2.1. The Context
In addition to being one of the poorest countries in the hemisphere, Bolivia has historically
been one of its most unequal and among the worst performers in terms of poverty reduction and
economic growth. This has generated negative policy feedback loops: high inequality has hindered
both poverty reduction during economic expansions and economic growth during recessions (Gray
Molina and Yañez 2009, 2). Education, as a result of both variation in levels of achievement and
because of low returns in the labor market particularly for indigenous peoples, has been identified
as the single most important cause of inequality in the country (Gray Molina and Yañez 2009, 20).
It is under this context that, beginning in the early 2000s, center-right politicians began proposing
education-linked cash transfer programs. Such programs, however, were absent from Morales’
presidential campaign platforms in 2002 and 2005 (McGuire 2013, 9).
The initial adoption of safety net programs in Bolivia was linked to the severe structural
adjustments undergone during the 1980s and 1990s, which sought to reactivate a long-stagnant
economy and end to rampant hyper-inflation. The social cost of these reforms would eventually
play a decisive role in the election of Evo Morales, a Leftist former coca grower of indigenous
descent (Madrid 2008; 2011). In the midst of structural adjustment, the government and the World
Bank recognized the need to provide basic safety nets covering those excluded from traditional
social policy target primarily at formal-selector workers (Newman, Jorgensen and Pradhan 1991).
Thus, in 1986, Bolivia became the first country in region to adopt an Emergency Social Fund
(ESF). The program funded small-scale, labor-intensive infrastructure projects proposed by
governments and non-governmental organizations. Its impact, however, was constrained by the
inability of the poorest communities to articulate funding proposals (Graham 1992; McGuire
2013).
Social Policy and Presidential Ideology in Latin America
149
The first major cash transfer program, a universal non-contributory pension, was made
possible by the partial privatization of state-owned enterprises under first administration of center-
right President Gonzalo Sánchez de Lozada (1993-1997). A controlling stake in the companies
was sold to private firms. The government’s share was managed by newly established private
pension funds. Starting in 1997, the dividends from these assets were distributed to all Bolivian
over 65 through Bono Solidario (Bonosol), an annual non-contributory pension (Müller 2009,
164). Forgoing means-testing sparked opposition from the World Bank and Inter-American
Development Bank (McGuire 2013, 27).
98
This program, which Morales renamed Bono Dignidad
in late 2007, remains the only universal non-contributory pension program in Latin America
(Müller 2009, 163).
99
Bolivia’s initial experience with a universal cash transfer program set a
precedent that was later followed by Morales when designing BJP and Bono Juana Azurduy.
Upon being voted back into office, Sánchez de Lozada (2002-2003) promised to
complement Bonosol with universal cash transfer programs aimed at increasing the use of
education among girls (Beca Educación) and healthcare among children under five (Bono Salud)
(Pardo 2003, 46). However, in October 2003, before the programs were implemented, Sánchez de
Lozada resigned following the death of 60 people protesting his government’s plans to export
Bolivian gas through Chilean ports.
Although education-linked cash transfers failed to materialize at the national level, they
were successfully implemented at the local level. As in Argentina, a major city, in this case El
Alto, pioneered CCTs. Located in the highlands above La Paz, El Alto is populated primarily by
indigenous migrants from the Altiplano. In September 2003, the city’s center-right mayor Luis
Paredes launched Bono Esperanza, which paid every first-grader in the city four installments of
50 Bolivianos (about $6.50 at the time). The program, which has gone by several names, proved
Social Policy and Presidential Ideology in Latin America
150
popular and was expanded to other grades. Bono Esperanza’s popularity prompted Jorge Quiroga,
Morales’ centrist opponent in 2005, to promise a national version of Bono Esperanza (McGuire
2013, 10). Now that BJP covers every grade, the program, which continues to pay a 200-Boliviano
stipend, focuses on children enrolled in kindergarten and disabled students.
The right had put cash transfers on agenda. As in Brazil, where an initially ambivalent
center-left president Lula da Silva (2003-2011) came to embrace and expand CCTs (Brearly 2011,
110; Lindert et al. 2007, 13; Lustig 2014, 11), the Bolivian left under Morales took time to warm
up to the idea of cash transfers linked to education. The only president to be elected by an outright
majority since Bolivia democratized in 1982, Morales came into office in January 2006 with a
strong mandate to enact a new constitution to refound Bolivia as a pluri-national state, nationalize
the country’s gas reserves and redistribute income towards the poor (Madrid 2011, 239). Thus, on
Labor Day 2006, Morales signed a decree declaring state ownership of the country’s reserves and
authorizing the military and workers from the state-owned energy company Yacimientos
Petrolíferos Fiscales Bolivianos (YPFB) to seize control of gas wells operated by foreign firms.
Under the new rules of the game, private energy firms would continue to extract gas but
would have to sell it YPFB, which would handle all distribution and commercialization activities.
As a result, the Bolivian state came to accrue more than 80 percent of the industry’s profits
(Fundación Jubileo 2012, 25). This increased the government’s intake from taxes and royalties
from 28 to 45 percent of GDP during 2004-2010 (De León Naveiro 2011, 58; Fuentes 2014, 109).
In addition, this period coincided with the culmination of various debt forgiveness schemes that
reduced the government’s debt obligations by more than half, from about 84 to 31 percent of GDP
during 2003-2008 (Overseas Development Institute 2010, 21).
Social Policy and Presidential Ideology in Latin America
151
A boom in hydrocarbon exports, a consequence not just of the government’s appropriation
of the majority of gas revenues, but also of record-high international energy prices, combined with
reduced debt payments and increased remittances from Bolivians living abroad, have fueled the
longest period of continuous economic expansion in Bolivia’s recent history. It was under these
highly propitious circumstances that that Morales launched BJP in October 2006.
As in Argentina and Costa Rica, the key enabling factor behind the launch of BJP was the
availability of fiscal resources, a direct consequence of the gas nationalization. Also, as in
Argentina, prior to the launch of a national program, the opposition government of a major city
had demonstrated the feasibility (and political benefits) of CCTs and the national-level opposition
had put the issue on the national agenda. However, in contrast to Argentina, where President
Fernández launched AUH during a moment of political weakness, Morales launched BJP from a
position of strength, just 10 months removed from his historic electoral victory. During that same
election, Morales’ Movimiento al Socialsmo (MAS) party won an outright majority in the
Congress and a near-majority in the Senate
That Morales launched BJP from a position of strength does not negate that political
competition played in its adoption. Although Morales’ grip on power was never seriously
threatened, he did face constant elections and challenges from opposition-party prefects seeking
greater autonomy for the country’s resource-rich eastern departments. During his first term in
office, Morales faced constant elections that challenged his ability to refound the Bolivia and
consolidate his power. Shortly before the launch, in July of 2006, MAS won a majority in elections
to appoint a constituent assembly. In mid-2008, eastern departments controversially voted to
demand greater autonomy and later Morales prevailed in a recall referendum.
100
In late 2009,
Social Policy and Presidential Ideology in Latin America
152
voters approved the country’s new constitution via referendum. Finally, at the end of that year
Morales was returned to office by a wide margin.

Ironically, even though the resource windfall gave Morales the resources needed to pursue
an ambitious social policy agenda, it also intensified existing conflict over the distribution of rents
(Crabtree 2008; Gray Molina 2010). This constant competition for rents from natural resources,
Laserna (2007; 2008) argues, is a major source of Bolivia’s political and economic backwardness.
Rent-seeking diverts energies from productive endeavors and toward politics, generating a cycle
of perpetual political instability.
101

4.2.2. The Adoption Process
Similarly to AUH in Argentina and, as will be demonstrated in the next section,
Avancemos in Costa Rica, BJP was a high-level presidential initiative launched by decree. Neither
legislators nor representatives of civil society were consulted let alone involved in designing the
program.
102
Legislators remain largely shut out from this process. The budget is decided on an
annual basis by the president and the Finance Minister (Meave interview).
103
The legislature’s
oversight of BJP and the government’s other cash transfer programs is hampered by the fact that
they are not paid for from the central government budget but rather from the budgets of state-
owned enterprises and, in the case of Bono Juana Azurduy, the Central Bank (Aguilar Pacajes
2014, 8). Eight years later, the BJP continues to be run via presidential decree and there appear to
be no plans to enshrine the program in law.
Upon coming to office, Morales was presented with a proposal for a nationwide targeted
CCT, but rejected it in favor of a universal program (Meave interview, Yañez interview).
104
As
Murillo noted, “the decision (to adopt BJP) came from the president and the cabinet. The reasoning
Social Policy and Presidential Ideology in Latin America
153
was that it had to be done and we have to do it. A small team was built. It was a matter of executing
it.”
105
The decree was signed in October 2006 and the first payment was made one month later.
Sources interviewed noted that, although social policymaking in Bolivia has always been
controlled by the president, policymaking under Morales is more centralized than under his
predecessors (Meave interview; Yañez interview). Whereas in the past agencies such as the
Planning Ministry played an active role in designing policies, under Morales the social policy
bureaucracy’s role has limited to implementation. Interviewed shortly after Morales announced
the extension of BJP to ninth graders, Meave noted that “the decision to expand comes from the
president. The president does not consult, he simple does. The decree got to us (the Planning
Ministry) a day before the cabinet discussed it.”
4.2.3. Program Design
BJP’s heterodox design constitutes an attempt to adapt CCTs to a context with low state
capabilities, even by Latin American standards. The Bolivian state’s authority manifests itself
unevenly across the country’s territory, conjuring up images of a “Swiss cheese state” (Crabtree
2008; UNDP 2007). Given the government’s limitations and the limited reach of the country’s
financial system, BJP is paid once a year in cash and the armed forces are charged with personally
delivering payments to schools. Experts consulted praised the military’s role, noting that it is the
only government entity capable of reaching every corner of the country at a low cost (Meave
interview; Yañez interview).
Three factors explain the decision to launch a universal program rather than a targeted one:
the aforementioned limits on government capabilities, the country’s high poverty levels and the
prevalence of a discourse centered on universalism. With regard to the first point, there are
legitimate doubts regarding whether the Bolivian government would be able to pull off a “state-
Social Policy and Presidential Ideology in Latin America
154
of-the-art” CCT with precise means-testing. On the second point, with a yearly GDP per capita of
$2,868, Bolivia is one of the poorest countries in the hemisphere and, as such, given a reasonable
exclusion threshold, a very large share of the population would qualify to receive CCT benefits.
Under these conditions, the administrative costs associated with means testing cease being cost-
efficient (McGuire 2013; Paz Arauco interview).
106
Third, Morales’ campaign platform, centered
on nationalizing the country’s hydrocarbons and redistributing that wealth among all Bolivians,
and the precedent created by Bonosol created the expectation that the program would be universal.
Figure 5-7: Bolivian Cash Transfer Programs
Share of Beneficiaries by Income (2009)

Source: Paz Arauco et al. (2013, 11)

Although the program is not means-tested, it still progressive inasmuch poor families tend
to have more children and better-off parents are more likely to send their children private schools
(McGuire 2013, 14). However, just under half of recipients (46 percent) in 2009 were poor (living
on less than $4.00 per day) (see Figure 5-7). That figure, however, has likely declined as a result
of the program’s gradual expansion to higher grades in which poor children are less represented
(Hermani-Limarino interview).
107
Regardless, the vast majority of the poor population does benefit
20.4%
29.2% 28.9%
23.3%
10.0%
16.8%
20.6%
14.6%
69.6%
54.0%
50.5%
62.1%
0%
20%
40%
60%
80%
100%
Renta Dignidad Bono Juancito Pinto Bono Juana Azurduy All Direct Transfers
Extreme Poor (Less Than $2.5/day) Poor ($2.5 to $4/Day) Non‐Poor
Social Policy and Presidential Ideology in Latin America
155
from BJP. In 2009, the program respectively reached 70.5 percent and 61.6 percent of the Bolivian
population (including those without children) living in extreme (less than $2.50 per day) and
moderate poverty (between $2.50 and $4.00 per day) (see Figure 5-8).
Figure 5-8: Bolivian Cash Transfer Programs
Coverage by Beneficiary Income (2009)

Source: Paz Arauco et al. (2013, 11)

The absence of means testing and the use of the armed forces has allowed the Bolivian
government to keep administrative costs below four percent of the program’s budget (Aguilar
Pacajes 2014, 9; Meave interview), a level comparable to that of Mexico’s Progresa-
Oportunidades, one of the region’s most cost-efficient CCTs (Caldés, Coady and Maluccio 2006).
However, foregoing precise targeting also likely also hampered BJP’s goal of increasing
enrollment. The program pays the same 200-boliviano (about $30) yearly stipend to all students,
even though the opportunity costs of remaining in school differ significantly by student age, sex
and place of residence. Dropout rates, particularly in rural areas, increase sharply at the end of
elementary school and tend to be higher among women and students of indigenous descent.
19.2%
70.5%
5.5%
87.7%
14.3%
61.6%
6.0%
83.2%
19.0%
37.9%
2.8%
67.8%
0%
20%
40%
60%
80%
100%
Renta Dignidad Bono Juancito Pinto Bono Juana Azurduy All Direct Transfers
Extreme Poor (Less Than $2.5/day) Poor ($2.5‐$4/day) Non‐Poor
Social Policy and Presidential Ideology in Latin America
156
More importantly, BJP’s ability to make up for the opportunity costs of remaining in school
is hampered by the stipend’s small size, which, at roughly four percent of average household
consumption, is quite low by regional standards (McGuire 2013, 28). Furthermore, the stipend has
not been adjusted since the program was launched. Average consumer prices increased by 62
percent between launch and December 2013 (Figure 5-9).
Figure 5-9: Consumer Price Inflation and the Purchasing
Power of Bolivian Cash Transfer Programs (2006-2014)

Source: Own calculations based in inflation data from the
Central Bank of Bolivia.

It is questionable whether the label of CCT actually applies to BJP. There is no data on
conditionality enforcement and experts consulted agreed that the requirement that students attend
school at least 80 percent of the time is not enforced (Hermani-Limarino interview; Murillo
Zambrana interview; Paz Arauco interview). The closest measure of conditionality enforcement
comes from a survey of 3,666 students, which showed that 99.2 percent had received the stipend
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
1.05
1.10
1.15
1.20
1.25
Nov‐06
Mar‐07
Jul‐07
Nov‐07
Mar‐08
Jul‐08
Nov‐08
Mar‐09
Jul‐09
Nov‐09
Mar‐10
Jul‐10
Nov‐10
Mar‐11
Jul‐11
Nov‐11
Mar‐12
Jul‐12
Nov‐12
Mar‐13
Jul‐13
Nov‐13
Purchasing Power of Juancito Pinto Stipend
Purchasing Power of Renta Dignidad
Note: BJPMonthly Stipend = 200 Bolivianos/Year
Bonosolbecomes Renta
Dignidad: 33% increase
25% increase
Social Policy and Presidential Ideology in Latin America
157
Tapia, Murillo, and Flores 2011, 24). Lack of proper documentation was the main reason cited in
the handful of cases of non-payment. Only six students said they had been excluded for failing to
meet the enrollment requirements.
108

4.3. Costa Rica’s Avancemos
In contrast to the Argentine and Bolivian CCTs, which aim to cover a large segment of the
school-aged population, Costa Rica’s Avancemos targets only low-income high school students
and is thus much smaller than the region’s other programs. Avancemos was originally intended to
be the strictest human capital CCT in the region (Víquez interview
109
), but over time, has moved
slightly closer to being a basic income CCT.
Figure 5-10: Costa Rica, Avancemos,
Number of Beneficiaries and Beneficiaries Relative
to the Total Public High School Population Covered (2005-2013)

Source: FODESAF

Despite the broadening of its target population and, until recently, weak conditionality
enforcement, the program is still very much a human capital CCT. Stipend amounts are
0%
10%
20%
30%
40%
50%
60%
0
30
60
90
120
150
180
210
2005 2006 2007 2008 2009 2010 2011 2012 2013
Coverage
Beneficiaries (Thousands)
Coverage as a % of High School Students Number of Beneficiaries
Social Policy and Presidential Ideology in Latin America
158
differentiated by grade level and were explicitly calculated to cover the opportunity cost of
remaining in school and outside workforce. However, as in Bolivia, stipends have not been
adjusted since the program launched in 2006. Against the original idea of designing a strict human
capital CCT, governments have consciously chosen to expand coverage rather than maintain the
purchasing power of the initial target population (Arias interview;
110
Li interview;
111
Marín
interview
112
). Coverage nearly doubled between 2007, its first full year of operation, and 2010
from just under 100,000 (27 percent of public high school students) to 185,000 (47 percent).
Coverage has since stagnated (see Figure 5-10). Overall, the program is estimated to benefit
140,000 poor and economically vulnerable families.
In sharp contrast to the left-leaning leaders of Argentina and Bolivia who were skeptical,
if not outright hostile, towards CCTs before ultimately adopting them, Costa Rica’s centrist
president Oscar Arias (2006-2010) was a strong advocate and included what eventually became
Avancemos in his campaign platform. Despite the rapid increase in coverage mentioned above,
the program remains quite small by regional standards. Thus, the Costa Rican case sheds light on
the role of ideology on CCT adoption and coverage. A non-left president was enthusiastic about
adopting a CCT, yet neither he nor his centrist successor, have dramatically expanded coverage as
a share of the population. Chapter 3 confirmed that the left is no more likely than the right to adopt
these programs (see also Díaz-Cayeros and Magaloni 2009; Sugiyama 2011). Though, as
Fernández in Argentina, Morales in Bolivia and da Silva in Brazil (Brearly 2011) show, the left
tends to adopt CCTs somewhat begrudgingly after politicians of the center or right, place them on
the agenda. However, also as shown in Chapter 3, once the left is in charge of a CCT, it tends to
rapidly increase coverage.

Social Policy and Presidential Ideology in Latin America
159
4.3.1. The Context
Costa Rica stands out among Latin American countries for coming the closest to having a
social democratic welfare regime (Esping-Andersen 1990; Huber 1996). At the center of this
regime is the National Health Service (Caja Costarricense del Seguro Social), which provides
universal public healthcare with standardized benefits de-linked from employment status and is
funded in an entirely solidaristic manner. The country was also regional pioneer with regard to
targeted anti-poverty programs (Franzoni and Sánchez-Ancochea 2012; Román Vega 2012). The
Social Development and Family Allowance Fund (FODESAF), a permanent and well-funded
social assistance fund has, since its creation in 1974, provided funding for a wide range of anti-
poverty initiatives.
113

Avancemos was proposed in 2006 in the context of a country with a universal welfare
regime yet dismally low levels of high school completion. In 2010, less than half of 19-21 year-
olds had graduate. More worrying, the rate of completion drops to just under a quarter among the
poorest quintile (Trejos 2012, 17). These figures are surprisingly low given the country’s level of
human development, which far surpasses what would be expected given its economic development
(Sandbrook et al. 2006, Ch. 4; World Bank 2001, Ch. 8). These indicators present a serious obstacle
to the country’s current development strategy centered on the attraction of high-tech foreign direct
investment (Paus 2005; Rodríguez-Clare 2001).
Costa Rica’s underperformance with regard to high school education begins with the 1980s
debt crisis. This was truly a lost decade: the economy shrank by nearly 50 percent in 1980 and
GDP per capita did not recover to pre-crisis levels until 1987. Social spending per capita declined
sharply and has yet to match its inflation-adjusted pre-crisis levels (Trejos 2012, 5). As result of
the crisis, the country halted an ambitious expansion of high school education started in the 1970s
Social Policy and Presidential Ideology in Latin America
160
(Román 2010). There was a sharp drop in coverage, as tens of thousands of teenagers left school
to help their families only to never return. When all was said and done, the debt crisis set Costa
Rican high school education back by more than two decades. Among the poorest quintile, the share
of the 19-21 year olds that had completed high school did not surpass its 1980 level (16.3 percent)
until after the turn of the millennium (Trejos 2012, 17).
The country had more success at fighting poverty, which, after peaking at over 50 percent
returned to its pre-crisis levels of around 30 percent in 1987. Starting with the first administration
of Oscar Arias (1986-1990), each president launched his own flagship anti-poverty program
funded through FODESAF. These programs proved to be short-term and were, with few
exceptions, dismantled by the following administration. Improved growth and targeted anti-
poverty programs brought poverty down to 20 percent during the mid-1990s. With the exception
of a few years of very fast growth towards the end of the 2000s, poverty has since remained stuck
at that level, plus or minus 1.5 percentage points.
Superémonos, the country’s first CCT linked to education, was one of the many short-term
targeted programs. Small and not very publicized, the program was launched in 2000 to little
fanfare by the administration of Miguel Angel Rodríguez (1998-2002). During the school year,
beneficiary families received a monthly coupon worth 10,000 colones ($30 at the time) in
groceries. The education conditionality was strictly enforced with families being required to prove
attendance twice during the school year. Superémonos, in sharp contrast to other short-term
initiatives, was subject to an external evaluation that found that it reduced the dropout rate among
beneficiaries by 5.0-8.7 percentage points during 2000-2001 (Duryea and Morrison 2004). There
was also some evidence that it increased the likelihood of beneficiaries passing the school year,
but had no effect of child labor. The program peaked at 12,000 beneficiaries before being cancelled
Social Policy and Presidential Ideology in Latin America
161
by President Abel Pacheco (2002-2006), who, despite being from the same party as Rodríguez,
shut down all of his predecessor’s social policy innovations.
4.3.2. Adoption Process
In contrast to the two left-leaning presidents studied above, Arias, a centrist, proposed a
CCT in his campaign platform. High school desertion was a personal concern for Arias (interview),
who, during his first term (1986-1990), witnessed the emergence of the “lost generation” that
dropped out because of economic hardship. Thus, also in contrast to CCTs in Argentina and
Bolivia, human capital formation – not poverty reduction or the creation of a basic income floor –
was the central goal of what eventually became Avancemos. However, as noted, the program’s
focus shifted toward poverty reduction as it gained popularity and as economic circumstances
worsened towards as a result of the global financial crisis.
Of the three explanatory variables analyzed in this study – presidential ideology, economic
conditions and political competition – only the first seems to have influenced CCT adoption in
Costa Rica. Economic conditions affected the program’s scope but not its adoption. Political
competition did not influence adoption or scope. Aware of the experiences of Brazil and Mexico,
Arias included CCTs in his campaign platform as a way of tackling Costa Rica’s stubborn high
school desertion problem. It should also be noted that, whereas the centrist Arias embraced CCTs,
his center-left opponent Ottón Solís made no reference to them in his platform. This constitutes
additional evidence that Latin America’s left has been hesitant to adopt CCTs. With regard to
economic conditions, Costa Rica adopted Avancemos during a period of fast growth. However,
once the economy slowed in 2008, the program was greatly expanded and used as a countercyclical
stimulus policy. Political competition does not appear to have affected either adoption of scope.
While Arias lacked a legislative majority, the only election he faced during his term – the 2007
Social Policy and Presidential Ideology in Latin America
162
referendum of CAFTA-DR – took place while Avancemos was a new and relatively unknown
program.
As in the other two cases, availability of resources, in this case as a result of the
unprecedented fiscal surplus Arias inherited from his predecessor, Abel Pacheco, was a key
enabling factor that made possible a substantial expansion in social policy and the adoption of
Avancemos. Total social spending increased at a rate of 5 percent a year during his administration
(Trejos 2012, 4). In addition to Avencemos, Arias increased non-contributory pensions by 150
percent.
As was the case in Argentina and Bolivia, Avancemos was strictly a presidential initiative.
It was enacted by presidential decree with neither legislators nor civil society being consulted.
Arias (interview) was blunt about the decision to sidestep the legislature: “if we would have done
it through a law we would still be discussing it.” Arias delegated the task of designing the program
to a tight-night group of technocrats and academics specializing in social policy overseen by
Fernando Zumbado, the Social Development and Housing Minister.
The Avancemos technical team began working immediately after Arias was elected in
February 2006 (Víquez interview). They envisioned the program as a highly targeted intervention
designed to reduce the high school dropout rate and convince dropouts to return to school. Poverty
reduction was not an explicit goal, but rather a positive externality from the stipend (Barahona
interview;
114
Víquez interview), which was in turn carefully calibrated to cover the opportunity
cost of not working (Barahona interview; Sauma 2013; Sauma interview
115
). The technical team
was keenly aware of and influenced by the experiences of other countries with CCTs, particularly
Mexico and Chile, and, at the IDB’s expense, met with Miguel Szekély, an economist who played
a prominent role in the development of Mexico’s Progresa/Oportunidades (Barahona interview;
Social Policy and Presidential Ideology in Latin America
163
Sauma interview; Víquez 2010). However, the team was adamant about tailoring the program to
Costa Rica’s needs and not very open to foreign suggestions (Barahona interview; Martínez
Franzoni and Voorend 2011, 288; Víquez 2010).
Almost immediately, there were clashes between the technical team and politicians. A pilot
program was launched covering 3,500 13-17 year olds from poor urban schools selected by
UNICEF was launched in July 2006. The pilot was to generate experimental data that would serve
as inputs with which to assess the program’s impact and improve its design (Barahona interview;
Román 2010; Sauma interview) as in the much-publicized Mexican case (Skoufias and
McClafferty 2001). Arias believed the pilot project was too small and had wanted at least 6,000
beneficiaries (Li interview; Trejos interview
116
). Zumbado, who was carefully crafting a public
image as the “Minister of the Poor” (La Nación 2008) and was seen as a likely successor to Arias
also pushed for a bigger program (Sauma interview; Valverde interview;
117
Víquez interview).
118

Tensions regarding the program’s scope and mission continued as the pilot moved forward.
By then end of the pilot, Avancemos had nearly 10,000 beneficiaries. At the end of its first full
year in operation (2007), coverage reached nearly 100,000. Plans for an experimental evaluation
were dropped as the staff’s energy focused on rapidly expanding the program’s scope (Valverde
interview). Conditionality enforcement was de-emphasized (Víquez interview) and the
government spent heavily on advertising the program (Valverde interview).
The program’s transformation from a small selectively targeted human capital intervention
to the government’s flagship anti-poverty program remains a sore issue for most members of the
program’s original technical team (Barahona interview; Sauma interview: Valderde interview;
Víquez interview). As Víquez (interview) explained, “Arias needed two or three big
Social Policy and Presidential Ideology in Latin America
164
accomplishments. We needed a full administration to measure the program’s impact. Arias was
not going to wait. The program lost its vision. It became a machine for giving away money.”
This transition towards an anti-poverty program sped up as the economy slowed in 2008
and entered into recession in 2009. The expansion of Avancemos to 200,000 beneficiaries and an
additional increase in non-contributory pensions were the central planks of the social policy
portion of the Arias administration’s Plan Escudo (Shield Plan) stimulus package (Villareal F. and
Gómez A. 2010). Arias was overly ambitious and ultimately had to scale down the expansion plans
(La Nación 2009). Still, by 2010 the program was reaching 185,000 beneficiaries.
4.3.3. Program Design
In contrast to AUH and BJP, Costa Rica’s Avancemos was originally envisioned strictly
as a human capital CCT. The original decree launching the program stressed the goal of using cash
transfers to reduce high school desertion and made no explicit reference to the goal of reducing
poverty.
119
Avancemos is also is unique in the region in that it targets only high school students.
Given Costa Rica near universal rate of elementary school completion, Arias saw no need to
include it in a CCT. However, as noted above, the program gradually moved closer to the basic
income side of the spectrum. Whereas during the pilot project the program was restricted to
teenagers 13-17 year-olds living in poverty, it has since been extended to all “economically
vulnerable” 12-25 year-olds.
In contrast to the other CCTs studied in this chapter, Avancemos, like
Progresa/Oportunidades selects beneficiaries utilizing a proxy means test – a comprehensive 63-
variable questionnaire. As a result, the program has succeeded in targeting the poorest high school
students. High poverty areas are over-represented: 70 percent of rural students (Román 2010) and
practically all students in primarily indigenous communities (IMAS/MEP 2009) are covered. But,
Social Policy and Presidential Ideology in Latin America
165
as noted, the program’s eligibility criteria have expanded over time. At launch 80 percent of
beneficiaries scored 1 or a 2 (out of 4) on the means-test. This figure has dropped to 60 percent in
recent years (see Figure 4-11).
120

Also as in Progresa/Oportunidades, monthly stipends increase with each passing grade,
ranging from 15,000 colones ($28) for seventh grade to 50,000 colones ($94) for twelfth grade.
121

These amounts were explicitly calculated to cover the opportunity cost of remaining in school at
the time of launch. This approach, however, has been criticized on the grounds that desertion
actually peaks at the start of high school and gradually tapers off from there (Dinarte Romero 2009;
Garnier interview;
122
Li interview; Sauma interview).
123
During 2006-2012, 15-20 percent of
seventh graders dropped out each year compared to less than 5 percent of eleventh and twelfth
graders (Mata Hidalgo and Hernández Romero 2013, 60).
Figure 5-11: Costa Rica, Avancemos Beneficiaries by
Proxy-Means Test Score (2006-2011)

Source: Villalobos (2012)

A bigger problem is that, as in Bolivia, stipends have remained constant. Average
consumer prices increased by 63 percent between launch and December 2013 (see Figure 5-12).
22%
18% 19%
13%
16%
25%
62%
53%
54%
40%
44%
46%
13%
25%
24%
26%
20%
21%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2006 2007 2008 2009 2010 2011
Group 1 Group 2 Group 3 Others
Social Policy and Presidential Ideology in Latin America
166
The program’s pilot project called for stipends to be updated once a year, but this provision was
dropped once the program launched at the national level (CGR 2012, 12). Arias and his successor
Laura Chinchilla (2010-2014) opted for increased coverage (poverty reduction) over compensating
for the opportunity cost of remaining in school (human capital) (Arias interview, Cordero
interview;
124
Li interview). The program’s expansion was a central to Arias administration’s
countercyclical response to the slowdown that followed the financial crisis. Arias (interview) noted
that the recession made necessary to increase coverage, even at the cost of decreased purchasing
power for existing beneficiaries.
Figure 5-12: Costa Rica, Consumer Price Inflation and
Purchasing Power of Avancemos (2006-2014)

Source: Central Bank of Costa Rica

Avancemos also underwent major administrative changes that shifted its focus more
toward poverty reduction. At the onset the program administrated by a small group of technocrats
0.60
0.70
0.80
0.90
1.00
1.10
Jul‐06
Nov‐06
Mar‐07
Jul‐07
Nov‐07
Mar‐08
Jul‐08
Nov‐08
Mar‐09
Jul‐09
Nov‐09
Mar‐10
Jul‐10
Nov‐10
Mar‐11
Jul‐11
Nov‐11
Mar‐12
Jul‐12
Nov‐12
Mar‐13
Jul‐13
Nov‐13
Purchasing Power of Avancemos Stipends (July 2006 = Baseline)
Note:Stipend Vaue: 15,000‐50,000
Colones Depending on Grade
Social Policy and Presidential Ideology in Latin America
167
and jointly implemented by the Social Aid Institute (IMAS), agency in charge of anti-poverty
initiatives, and the Education Ministry’s Scholarship Office (FONABE), which had experience
managing merit-based scholarships. While poorer and less technically proficient than IMAS,
FONABE had the on-the-ground expertise needed to locate students in need of stipends and
effectively enforce conditionality. However, coordination between the two proved extremely
difficult (CGR 2008; 2009; Defensoría 2010; Román 2010) and, in 2009, as part of the decision to
rapidly scale-up the program, Avancemos was folded into IMAS and thus lost its foothold in the
schools (Morales interview;
125
Trejos 2012). Teachers, which had opposed to the additional work
created by the conditionality resented having to “work for IMAS,” further complicating
conditionality enforcement (Morales interview; Vargas interview
126
).
Avancemos has been widely criticized for lax conditionality enforcement (CGR 2008;
2012; Dinarte Romero 2009; Román 2010; Víquez 2010). However, this should be kept in
perspective, particularly when compared with enforcement in Argetina, let alone Bolivia. For
example, a 2011 audit by the Comptroller General’s Office (CGR 2012) revealed that stipends
were paid to 11,708 students (6.3 percent of beneficiaries that year) not enrolled in a school. In
addition, as many as 10 percent of beneficiaries had failed more than one grade, placing them in
violation of the program’s rules, which have since been modified to allow students to fail up to
two grades.
IMAS has invested heavily in recent years in an information technology platform for
keeping track of school attendance. In the past principals had to turn in school attendance lists to
IMAS regional offices which would manually input them into the nationally database (Herrera
interview
127
). Principals now update the rosters in real time, making it easier to identify dropouts,
Social Policy and Presidential Ideology in Latin America
168
particularly those who quit mid-year. This is expected to substantially improve conditionality
enforcement.
Over time Avancemos has become less ambitious, eliminating a health conditionality that
was never enforced. Though Costa Rica has universal healthcare, teenagers underutilize
healthcare. It was hoped that stipend could be leveraged to increase their use of preventative care
(Víquez interview). It was also hoped that clinics would inform dropouts about the program and
possibly entice them into returning (Barahona interview). Again, bureaucratic infighting got in the
way. The financially autonomous National Healthcare Service repeatedly questioned the plan’s
feasibility (Herrera interview; Marín interview, Vargas interview). Thus, there is a clear pattern of
addressing obstacles by weakening the program.

5. Towards a Theory of CCT Adoption and Design
The previous chapter and past research (Ancelovici and Jenson 2013; Sugiyama 2011)
concur that the rapid spread of CCTs in the region during the 2000s was an example of policy
diffusion. However, explanations based on diffusion can only tell us so much. That the ideas
behind CCTs were available to and ultimately caught the attention of politicians and technocrats
in Argentina, Bolivia and Costa Rica does not explain the timing of CCT adoption or the
substantial variation in the design of the programs they adopted. Granted, the much-publicized
success of the Mexican and Brazilian programs (Aguiar and Araujo 2002; Skoufias and Bonnie
McClafferty 2001) and the willingness of multilateral banks to provide funding and technical
support (Ancelovici and Jenson 2013; Lana and Evans 2004; Martínez Franzoni and Voorend
2011; Sugiyama 2011) made CCTs increasingly attractive to Latin American policymakers. Yet,
Social Policy and Presidential Ideology in Latin America
169
adoption timing and program design were, like any other policy process, the product of a domestic
political calculus.
Taken together, the three case analyzed above and the existing literature on CCT adoption
in Mexico and Brazil discussed elsewhere in this study paint a picture of the adoption process that
goes beyond simply stating that policy diffusion took place.
5.1. International Factors: Putting CCTs on the Table
The Success of Pioneering CCTs: In the late 1990s technocrats and public intellectuals
from different intellectual traditions analyzing poverty in Mexico and Brazil came to a similar
conclusion: give the poor money conditional on school attendance. However, these two traditions
placed different weights on the two goals of what came to be known as CCTs. The Mexican model
(human capital CCTs) emphasizes the role of transfers and conditionality as tools to promote
human capital accumulation and, over the long run, prevent poverty. In contrast, the Brazilian
model (basic income CCTs) emphasizes the role of transfers as a tool to relieve poverty. Despite
these differences, the programs implemented in both countries succeeded in reducing poverty
(Aguiar and Araujo 2002; Skoufias and Bonnie McClafferty 2001) and proved politically
popular.
128

Selling the Success of Pioneering CCTs: The proven success of the pioneering CCTs
caught the attention of IFIs. These programs were compatible with the dominant thinking in the
social policy community at the time: they were obligations-based, market and incentive-based, and
targeted the poor (Adato and Hoddinnott 2010, 13-14). Additionally, by addressing concerns such
as human capital and female empowerment (by giving the money to mothers), they fit well with
the evolving discourse on poverty reduction (Feitosa de Britto 2008, 185). IFIs embraced CCTs
and offered support to countries looking to adopt their own. At the same time, the norm
Social Policy and Presidential Ideology in Latin America
170
entrepreneurs behind the original CCTs, most notably Mexico’s Santiago Levy (Ancelovici and
Jenson, 2013; Brearley, 2011) and Brazil’s Eduardo Suplicy (Lana and Evans, 2004) promoted
these programs throughout the region.
5.2. Domestic Factors: The Real Determinants of Adoption
Ideology Actually Does Matter: The scholarly consensus, further supported by the previous
chapter, holds that presidential ideology has no effect on the probability of CCT adoption (Díaz-
Cayeros and Magaloni 2009; Sugiyama 2011). This chapter has found that, whereas the right and
center were predisposed in favor of CCTs, the left, likely a result of its deeply-held universalistic
aspirations and, more tentatively, enduring distrust of IFIs, was more guarded, if not outright
opposed to them. It has been documented that Brazil’s center-left President Luiz Inácio Lula da
Silva (2003-2011), the man most closely associated with the expansion of CCTs in that country,
was not initially enamored with these programs and instead favored food stamps (Brearly 2011,
110; Lindert et al. 2007, 13; Lustig 2014, 11). Similarly, Argentina’s Kirchners were initially
openly hostile to “pre-packaged” programs from IFIs. CCTs were absent from the campaign
platforms of Bolvia’s Evo Morales (McGuire 2013, 9) and Costa Rica’s Ottón Solís.
As the cases reveal, in both Argentina and Bolivia the center and the right put CCTs on the
political agenda by implementing them at the local level and including them in their national
platforms. In sharp contrast, Costa Rica’s Oscar Arias, a centrist, proposed a CCT in his campaign
platform and began implementing it immediately after taking office. CCTs are extremely popular
policies. Once they entered the political conversation it was only a matter of time before they were
adopted at the national level – even by governments that originally opposed them.
Political Competition Matters (Sometimes): The case study literature on the pioneering
countries cites electoral competition as a decisive factor in the adoption of CCTs in Mexico (Díaz-
Social Policy and Presidential Ideology in Latin America
171
Cayeros, Estévez and Magaloni Forthcoming; Dion 2010; Garay 2012) and Brazil (Bandeira
Coêho 2012; Garay 2012; Melo 2008). Yet the cross-national analysis conducted in the previous
chapter found no evidence of this. The case studies reveal that political competition played a
central role in the adoption of AUH. Bolivia’s Morales has found himself in a position of
remarkable political strength since coming to office. However, BJP was adopted near the start of
a period of seemingly perpetual electoral contests. Competition does not appear to have been a
factor in the adoption of Avancemos in Costa Rica.
5.3. Ideology as a Determinant of CCT Design/Type
The previous chapter revealed that presidential ideology has a substantively large effect on
the scope of CCTs. Specifically, CCT coverage, both as a share of the population and relative to
the share of the population living in poverty, is higher in countries governed by the left. More
tentatively, this chapter provides evidence that whereas left presidents favor basic income CCTs,
on-left presidents prefer human capital CCTs. Left governments in Argentina and Bolivia adopted
programs without explicit means testing, a flat stipend structure that does not consider the varying
opportunity costs of remaining in school, and limited (and in the case of Bolivia essentially no)
conditionality enforcement. In contrast, Costa Rica’s Avancemos has explicit means testing and
narrow targeting, a variable stipend structure and, after some initial problems, a sophisticated
platform for enforcing conditionality.
Does this finding travel outside these three cases? A rigorous cross-national test of this
proposition is beyond the scope of this chapter. However, a cursory analysis of the three largest
CCTs in the region supports this explanation. The prototypical human capital CCT, Mexico’s
Progresa/Oportunidades was adopted by a center-right president. Despite some changes that have
moved it closer to the basic income side of the spectrum (Yanes 2013), it remains the template for
Social Policy and Presidential Ideology in Latin America
172
human capital CCTs. Colombia’s Familias en Acción was adopted in 2001 by a center-right
president, Andrés Pastrana. It follows the Mexican mold by assessing eligibility through a proxy
means test, paying differentiated stipends calculated based on opportunity cost and strict
conditionality enforcement. Colombia and Mexico have the only CCTs in the region were
conditionality violations result in an immediate cessation of benefits (Bastagli 2009, 133).
Granted, the prototypical basic income CCT, Bolsa Escola/Bolsa Familia was adopted by
the centrist Fernando Henrique Cardoso. However, unlike the transition from Progresa to
Oportunidades which changed the program’s scope but did not alter its design,
129
the transition
from Bolsa Escola to Bolsa Familia involved significant changes that moved the program further
in the direction of the basic income ideal. The most notable changes were the inclusion of children
under six and the addition of an unconditional basic benefit for people living in extreme poverty,
regardless of whether they have children. Thus, while it would an overstatement to claim that Lula
adopted an entirely new program, he transformed the one he inherited into one that came closer to
providing a basic income floor for all Brazilians.

6. Summing Up: The Left’s Love/Hate Relationship with CCTs?
Building off the findings of the statistical analyses conducted in the previous chapter and
based on in-depth qualitative case studies on the adoption of CCTs in Argentina, Bolivia and Costa
Rica, this chapter has sought to sketch out a theory of CCT adoption and design that goes beyond
simply stating that countries emulate and/or learn from their neighbors.
Taken together, the last two chapters reveal a complex relationship between CCTs and the
central explanatory variable of this study – presidential ideology. The previous chapter
demonstrated that left-leaning presidents are just as likely to adopt national-level CCTs as their
Social Policy and Presidential Ideology in Latin America
173
counterparts from the center or right. This chapter has shown that presidents of the left have often
been skeptical if not outright opposed to these programs. Politicians of the center and right were
the ones that put CCTs on the political agenda by adopting them at the local level and incorporating
them into their campaign platforms. Leaders of the left, even the ones that eventually adopted
large-scale CCTs, were slower to embrace these policies. However, as demonstrated by Chapter
3, once in charge of a CCT, left presidents tended to increase coverage to levels beyond those
achieved by presidents of the center or right.
The concluding chapter takes stock of the findings from this and the previous chapters and
draws some general conclusions on the role of presidential ideology in post-market reform Latin
America.

Social Policy and Presidential Ideology in Latin America
174
Chapter 5 Appendix.
Effects of the Selected CCTs
on Education Poverty and Inequality

This section briefly reviews the literature on the effects of the CCT programs studied in
Chapter 5.
1. Effects of AUH
There is evidence that, despite coverage problems, AUH has significantly reduced poverty
in Argentina. However, limited government data on the program, let alone systematic evaluations
of its effectiveness, make it difficult to precisely estimate its effects. As the program aims to be
universal and was rolled out at a national level all at once, there is no comparison group with which
to conduct an experimental evaluation. Because the program replaced several other programs at
the national and local levels it is difficult to calculate a baseline level against which to compare its
effects (Tuñón Interview). Furthermore, four years after AUH was launched, the National Institute
of Statistics and Censuses (INDEC) has yet to incorporate a question about the program in annual
household surveys.
130

The first generation of studies on AUH attempted to estimate its potential effects on
poverty, indigence and income inequality by simulating how it would boost beneficiary incomes
(Agis, Canete and Panigo 2009; Arceo, González and Mendizábal 2010; Bertranou and Maurizio
2012; Calabria et al. 2010; Gasparini and Cruces 2010; Roca 2010). These estimates varied greatly
but agreed that the program would substantially boost the incomes of the poorest Argentines.
However, these estimates were overly optimistic as they assumed that the program would reach
all eligible Argentines (Calvi, Cimilo and Chitarroni 2011).
Social Policy and Presidential Ideology in Latin America
175
The most rigorous evaluations have been conducted by the Observatorio de la Deuda Social
Argentina (ODSA-UCA), which incorporated an AUH question into its annual surveys. These are
not, however, experimental or even quasi experimental studies, but rather simulations as in
previous studies, albeit based on more precise data. ODSA estimates that during 2010-2012, the
program, boosted the incomes of poor families by 10.4 percent and indigent families by 21.2
percent (ODSA-UCA 2013, 9). Furthermore, it is estimated that it cut the shares of population
under 17 living in poverty and indigence by 6.2 and 49 percent, respectively (ODSA-UCA 2013,
11).
ODSA-UCA also conducted the most rigorous assessment thus far of the program’s effects
on education and nutrition (Salvia, Musante and Mendoza Jaramillo 2013). As is to be expected in
a country with universal elementary school enrollment, AUH has not affected enrollment among
children under 12. The program has, however, had a sizable effect on enrollment among teenagers.
Among the poorest half of the population, enrollment increased 5.1 percentage points (17.7
percent) among AUH beneficiaries compared to a 6.9-percentage-point (11.9 percent) decrease
among non-beneficiaries. The share of teenagers not attending school decreased by 4.0 percentage
points (75.4 percent) among beneficiaries and increased by 5.8 percentage points (71.6 percent)
among non-beneficiaries. AUH is succeeding in keeping beneficiaries in the classroom. However,
by failing to cover the entire eligible population, the program is failing to meet its full potential.
These effects are likely driven by the income boost AUH provides and not the enforcement
of conditionality. There are no data on the number of children who have stopped receiving stipends
due to non-compliance. The closest thing to hard data on the matter is a survey of 571 school
principals in which only two percent of elementary and 10 percent of high school principals
reported being aware of at least one student whose stipend was revoked (OEBA 2012). The view
Social Policy and Presidential Ideology in Latin America
176
that conditionality is not being enforced was shared by all the experts interviewed. For Maurizio
(interview)
131
, improving human capital has been an afterthought compared to the goal of
extending the right to social policy to previously excluded groups. In her opinion and that of
Bertranou (interview), the conditionality exits in name only to grant the program legitimacy among
the middle class taxpayers. Gasparini (interview) suspects that the government is unwilling to cope
with the possible backlash associated with taking benefits away from the poor. It does, however,
appear that the government is taking steps towards strengthening conditionality. At the end of
2012, ANSES began expelling beneficiaries who had never collected the 20 percent tied to school
attendance (Parrado interview).
132

Beyond schooling, AUH has had a significant positive effect on beneficiaries’ food security
(Salvia, Musante and Mendoza Jaramillo 2013). With regard to the labor market, there is little
evidence that the program disincentivizes employment (Bustos, Giglio and Villafañe 2012; Bustos
and Villafañe 2011). There is, however, evidence that it makes informal-sector workers less likely
to formalize, though it does not make formal-sector workers to become informal (Garganta and
Gasparini 2012).

2. Effects of BJP
It is difficult to quantify BJP’s effectiveness. No studies were conducted prior to its launch,
let alone a pilot project (Murillo interview; Yañez interview) and no official evaluations have been
conducted. The government’s antipathy toward program evaluations has been a source of
frustration among the social policy bureaucracy (Meave interview) and was the main reason why
the World Bank, which was originally scheduled to cover half the cost of Juana Azurduy (Jornada
2009), backed away from the project (Yañez interview).
Social Policy and Presidential Ideology in Latin America
177
The consensus among researchers is that the program has not been very effective at either
reducing poverty or increasing enrollment. Paz Arauco et al. (2013) attribute the limited effect on
poverty of BJP and other transfers programs to the lack explicit targeting of the poor and
insufficiency if the amounts transferred. Given the program’s universal design, half of
beneficiaries are non-poor (see Figure Y). On the one hand, though not means-tested, by targeting
school-aged children, BJP is one of Bolivia’s most progressive programs (Paz Arauco et al. 2013).
On the other, by providing the equivalent of just $0.18 per day, BJP’s stipend does not come
anywhere near providing the income boost needed to seriously reduce poverty.
There is disagreement regarding whether better targeting would increase the program’s
impact. Providing larger stipends the poorest could lead to greater reductions in poverty and
inequality levels (Paz Arauco et al. 2013; Yañez interview). Against this view, McGuire (2013,
31) argues that targeting and stipend size are unrelated issues. Assuming the program’s budget
remains fixed, means-testing to ensure that the entirety of the budget goes to the poor would only
double the stipend to $60 a year. Even assuming the Bolivian state can operate such a program,
targeting would significantly increase administrative costs. Moreover, budgets are not fixed and
policies that benefit the middle class are more likely to see their budgets increase over time. Thus,
universal transfers may actually be more generous than targeted ones and thus more beneficial for
the poor (Gelbach and Pritchett 2002).
It is unclear to what extent BJP is responsible for the significant decline poverty and
inequality experienced over the last decade. Ex-ante simulations of the effects of BJP and Bonosol
based on household surveys attribute to the programs a reduction in the Gini coefficient of 1.5
percentage points – from 0.550 to 0.535 – and reductions of about one percentage point in the
levels of moderate and extreme poverty – from 52.7 percent to 51.9 percent and from 34.7 percent
Social Policy and Presidential Ideology in Latin America
178
to 33.6 percent, respectively (Medinaceli and Mokrani 2010). However, this effect is driven
primarily by Bonosol, which, despite having less than half of BJP’s beneficiaries, has five times
the budget. A separate study looking exclusively at BJP yields a more optimistic finding: the
program was responsible for a 1 percent drop in inequality. That result is larger in the countryside
(2.4 percent) than in cities (0.7 percent) (Yañez Aguilar 2012, 100). The program has had no effect
on moderate poverty and reduced extreme poverty by just 1 percent from 39 to 38 percent (Yañez
Aguilar 2012, 99). This limited impact is attributed to stipend size.
BJP’s impact on to education are mixed as well. Morales asserts that the program cut the
dropout rate among grades the first five grades from 6 percent to 1.8 percent (Press Release). The
elementary school dropout rate began declining five years prior to BJP and this trend has
continued. Worryingly, as the dropout rate fell so did school enrollment, going from 94.1 percent
in 2005 to 82.3 percent 2011 (Políticas Públicas Populi 2012, 3). Thus, the declining dropout rate
is at least partly explained by the lower number of school-age children enrolling in school.
Presumably, the children that are no longer enrolling are ones with a higher risk of dropping out
(Políticas Públicas Populi 2012, 4).
Most research finds a positive, albeit small, effect on enrollment in early grades, an effect
that all but disappears in higher grades. This is evidence that the stipend offered is too small to
keep older children in school, an effect that has likely worsened as inflation has eaten away the
stipend’s purchasing power. A survey of more than 200 teachers reveals that they perceive that
BJP has led to increased enrollment, assistance, completion of eighth grade, and student motivation
(Tapia, Murillo and Flores 2011). Quantitative research reveals that the program’s positive effect
on enrollment is driven almost entirely by the first three to five years of elementary school
(Hermani-Limarino 2009; Vera Cossío 2011; Yañez Aguilar 2012) and, more tentatively, among
Social Policy and Presidential Ideology in Latin America
179
girls living in rural areas (Vera Cossío 2011). These findings offer evidence that BJP has
encouraged parents to enroll their children at the recommended age. Late enrollment, which is
associated with worse learning outcomes, has traditionally been a problem in Bolivia, particularly
in rural areas (Hermani-Limarino interview; Meave interview).
While Bolivia has made strides in recent years to increase primary school coverage, school
attendance lags well behind enrollment rates. Grigoli and Sbrana (2013) find that, although
students who received BJP in 2006 were more likely to be enrolled in 2007, they were no more
likely to actually attend.
133
Thus, the stipend enticed children to enroll, but, in the absence of
conditionality enforcement failed to influence attendance.
134
This problem could be addressed,
they note, by increasing the stipend and paying it in multiple installments each conditioned on
attendance.
While BJP’s effect on education has been disappointing, its effect on child labor has been
nonexistent (Grigoli and Sbrana 2013; Vera Cossío 2011). There is, however, evidence that it may
reduce the number of hours worked by children in the first three years of elementary school (Vera
Cossío 2011, 21). This is also likely a consequence of the stipends, which fail to cover the
opportunity costs of working. Children living in extreme poverty, seem to both going to school
and working, a finding that challenges the conventional wisdom (Grigoli and Sbrana 2012). There
is also some preliminary evidence that the supply of labor and the hours worked decrease during
the months when the stipend is paid (Hermani-Limarino interview).

3. Effects of Avancemos
As in Argentina and Bolivia and in contrast to the pioneering programs in Mexico and
Brazil, the impact of Avancemos on education and poverty has not been formally evaluated. The
Social Policy and Presidential Ideology in Latin America
180
powerful Comptroller General’s Office has repeatedly demanded an evaluation of the program
(CGR 2008; 2012). The cancellation of the original plan to conduct an experimental analysis based
on the pilot has been attributed to the Arias administration’s rush to expand the program (Barahona
interview; Morales interview; Víquez interview). In April 2010, the IDB agreed to provide $500
million to evaluate Avancemos yet said evaluation never took place (La Nación 2010). Program
officials consulted could not recall what happened. More recently, evaluation has been hampered
by bureaucratic disagreements between IMAS and the Education Ministry over which institution
should be responsible.
With regard to education, independent assessments agree that the program has made
beneficiaries less likely to drop out. Beginning with descriptive data, between 2006 and 2010 the
share of 13-17 year-olds not attending school who cited costs as their main reason for dropping
out decreased from 24 to 20 percent among the population as a whole and from 28 to 15 percent
among the poor (Mata Hidalgo and Hernández Romero 2013, 61-62). In addition, a survey of rural
teachers’ perceptions on the program’s effects reveals that educators overwhelmingly believe that
the program has reduced desertion (Vargas Sandoval et al. 2010)
Statistical analyses also point to a significant drop in desertion. Using matching, Mata
Hidalgo and Hernández Romero (2013) find that Avancemos in 2006 was responsible for 14-16
percent of students remaining in school in 2007. The size of this effect appears to have decreased
over time: Avancemos in 2008 was responsible for 11-13 percent of students remaining in school
in 2009. This could be attributed to the decrease in the stipend’s purchasing power. A less
sophisticated analysis not accounting for selection bias concludes that a student receiving
Avancemos had, all else equal, a 53 percent higher chance of remaining in school (Vargas, Slon
and Fundación Acción Jóven 2010).
Social Policy and Presidential Ideology in Latin America
181
Avancemos also appears to have contributed to improvements in other schooling
indicators. Also utilizing matching, Meza-Cordero (2014) finds that students that benefited from
the program during four consecutive years completed between third of and half a year of additional
schooling relative to comparable teens without Avancemos. This effect is driven primarily by male
students, which are more likely to dropout than girls. Males who benefited for four years were
found to have completed between 0.51 and 0.69 additional years of schooling. Encouragingly, but
also more speculatively, Mata Hidalgo and Hernández Romero (2013) find that Avancemos was
solely responsible for 77-87 percent of students who returned to high school in 2007 and 2009
after not having attended the previous year.
There is consistent evidence that the program has had a very small negative effect on both
poverty and extreme poverty. Microsimulations attribute to the program reductions in both types
of poverty of 0.2-0.5 percentage points during the years 2007-2010 (Sauma 2011). This effect is
driven primarily by families with multiple children (Sauma interview). There is also preliminary
evidence that the program has had a negative effect on child labor among the poor (Morales Dada
2007).
Social Policy and Presidential Ideology in Latin America
182
Chapter 6.
Conclusion and Pending Research Agenda

This study has sought to untangle the relationship between two of the most significant
developments in Latin American politics over the last 15 years: the widespread election of left-
leaning presidents (Castañeda 2006; Flores-Macías 2012; Levitsky and Roberts 2011; Weyland,
Madrid and Hunter 2010) and the implementation of large-scale anti-poverty programs (Barrientos
and Santibañez 2009; Sugiyama 2011). In so doing, this project contributes to answering a broader
question within comparative political economy regarding the effects of ideology on social policy
in the era of globalization.
As a whole, this study demonstrates that in Latin America, as in industrialized countries
(Huber and Stephens 2001; Pierson 1996), the effect of ideology on social spending has weakened
significantly under globalization. However, spending levels mask the substantial autonomy Latin
America’s presidents have in shaping social policy. Left-leaning leaders are associated with an
expansion in the size of conditional cash transfer (CCT) anti-poverty programs. And it is precisely
these programs that explain why inequality in Latin America, historically the world’s most unequal
region, decreased during the 2000s just as it was increasing in both industrialized countries and
the largest emerging markets (Lustig, López-Calva and Ortíz-Juárez 2013; Soares et al. 2009;
World Bank 2014, 27). Without substantially increasing total social spending, left-leaning Latin
American presidents have modified social policy in their countries to achieve substantial
reductions in poverty and inequality and carry out significant income redistribution in favor of the
poor.
The next section provides a brief summary of the study’s main findings.
Social Policy and Presidential Ideology in Latin America
183
2. Key Findings
Chapter 2 revisited the existing literature on the determinants of social spending and social
policy more broadly in industrialized countries and tested its applicability to post-market reform
Latin America, and developing countries more generally. Specially, the chapter contrasted the
applicability of power resource theory (Esping-Andersen 1985; Huber and Stephens 2001; Korpi
1983; Stephens 1979), which predicts that governments with left-leaning ideologies will prioritize
social policy, and explanations centered on income distribution and the median voter (Meltzer and
Richard 1981). A time-series cross-sectional analysis revealed that income distribution was a more
consistent determinant of spending levels than was presidential ideology. However, the factor that
determines spending levels is the structure of inequality, not its level. In line with Lupu and
Pontusson’s (2011) pioneering work on industrialized countries, as the gap between the incomes
of the rich and the median voter grows, the latter becomes more likely to align with the poor in
favor of increased spending.
While less robust, there is also evidence that countries governed by left-leaning presidents
spend more on social policy. This effect, which operates primarily through education spending, is
small and explained almost entirely by the recent wave of far-left presidents, namely Hugo Chávez
in Venezuela, Evo Morales in Bolivia and Rafael Correa in Ecuador. That this effect operates
through education, the most progressive spending category, could be taken as evidence that the
region’s far-left governments have kept their promises to increase spending and redistribute
income. Education, in contrast to the social security programs that dominate the region´s social
policy budgets (Lindert, Skoufias and Shapiro, 2006), has been shown to reduce both poverty and
inequality (Birdsall, Lustig and McLeod 2012, 167; Rudra 2004).
Social Policy and Presidential Ideology in Latin America
184
Given the limitations inherent to spending data, Chapters 3 and 4 analyzed the effect of
presidential ideology on the adoption and design of CCTs. These programs, which award families
stipends conditional on their children remaining in school, constitute “the most important
innovation in social policy of the last fifteen years” (Lustig 2011, 19). In addition to increasing
school enrollment (Baird et al. 2013; Saavedra and García 2012), CCTs are credited with helping
pull 15.5 million Latin Americans – about 13 percent of the region’s population – out of poverty
(Stampini and Tornarolli 2012). Given their precise targeting – 75 percent of CCT spending
reaches the poorest 40 percent of the population – CCTs can be analyzed as a proxy for income
redistribution in Latin America (Goñi, López, and Servén 2008, 19-21). A homegrown Latin
American innovation, CCTs have since been adopted in settings as diverse as Sub-Saharan Africa,
Southeast Asia, and even New York City (Fiszbein and Schady 2009; Hanlon, Barrientos and
Hulme 2010).
Compared to the voluminous literature on the effectiveness of CCTs, relatively little has
been written on the politics behind these programs (Martínez Franzoni and Voorend 2011, 280).
Chapter 3 sought to fill that void in existing knowledge and provide a clearer explanation of cross-
national variation in income redistribution in Latin America. The statistical analyses conducted
revealed that, although presidential ideology influences the scope of CCTs, it has no bearing on
decisions regarding whether to adopt this type of policy.
CCT adoption has primarily been a diffusionary process. Latin American governments
were influenced by the perceived success of CCTs implemented by their regional neighbors and
in particular the two largest and most influential countries in the region – Mexico and Brazil.
Technocrats and public intellectuals involved in the design of pioneering CCTs, such as Mexico’s
Santiago Levy (Ancelovici and Jenson 2013; Brearly 2011) and Brazil’s Eduardo Suplicy (Lana
Social Policy and Presidential Ideology in Latin America
185
and Evans 2004), heavily promoted this new approach to combating poverty and increasing human
capital. Multilateral banks wholeheartedly embraced the ideas behind CCTs and provided
technical and monetary support to governments looking to create their own programs (Ancelovici
and Jenson 2013; Lana and Evans 2004; Martínez Franzoni and Voorend 2011; Sugiyama 2011).
Presidential ideology played a key role in determining program scope. True to their
ideological orientation, presidents of the left moved farther than their centrist and rightist
counterparts toward establishing a basic income floor for all members of society independent of
employment status. Specifically, CCT coverage, measured as a percent of a country’s population
benefiting from CCTs, was higher in countries governed by the left by 3.50 percentage points –19
percent of the average coverage level of a CCT in 2010 – relative to countries with presidents of
the center or right. Brief case studies on the evolution of CCT coverage under left presidents in
Brazil, Ecuador and Argentina further supported the findings of the statistical analysis.
Based on field research, Chapter 4 conducted in-depth case studies of the CCT adoption
and design process in Argentina, Bolivia and Costa Rica. The case studies sought to both
complement the preceding chapter by tracing the political processes that led to CCT adoption and
influenced CCT design and go beyond the statistical analyses in an effort to construct an improved
theory on the politics of CCTs. The chapter began by arguing that policy diffusion fails to provide
a satisfying explanation of adoption: that the ideas behind CCTs were available to and ultimately
caught the attention of politicians and technocrats throughout the region does not explain the
timing of CCT adoption or the variation in program design. The policymaking process surrounding
CCTs, like any other policy process, was the product of a domestic political calculus.
The cases revealed that ideology has in fact been an important determinant of CCT
adoption, but not by the expected way. By putting CCTs on the political agenda at the local and
Social Policy and Presidential Ideology in Latin America
186
national levels, centrist and right-wing politicians in Argentina and Bolivia set into motion a
political process that culminated with the adoption, by left-leaning presidents, of CCTs in those
countries. Process tracing demonstrated that, whereas politicians of the right and center were
predisposed in favor of CCTs, the left, likely a result of its deeply-held universalistic aspirations
and, more tentatively, enduring distrust of IFIs, was more guarded, if not outright opposed to them.
Argentina’s Kirchners were initially hostile to “pre-packaged” programs promoted by IFIs. CCTs
were altogether absent from the campaign platforms of Bolivia’s Evo Morales (McGuire 2013, 9)
and Costa Rica’s Ottón Solís. This finding challenges the conventional wisdom the prevailing
view, further supported by the statistical analysis in Chapter 3, that ideology has no effect on the
probability of CCT adoption (Díaz-Cayeros and Magaloni 2009; Sugiyama 2011).
With regard to CCT design, the cases provided evidence that that left presidents tend to
adopt CCTs that emphasize the role of transfers as a tool to relieve poverty. In contrast, centrist
and right-leaning leaders prefer CCTs that emphasize the role of transfers and conditionality as
tools to promote human capital accumulation and, over the long run, prevent poverty. Specifically,
left governments in Argentina and Bolivia adopted programs without explicit means testing, a flat
stipend structure that does not take into account how the opportunity costs of remaining in school
vary with age, and limited (and in the case of Bolivia essentially no) conditionality enforcement.
In contrast, Costa Rica’s Avancemos, adopted by a centrist government, has explicit means testing
and narrow targeting, a variable stipend structure and, after some initial problems, a sophisticated
platform for enforcing conditionality.
Taken together, Chapters 3 and 4 reveal a complex relationship between CCTs and
presidential ideology. It has been demonstrated statistically that left-leaning presidents are just as
likely to adopt national-level CCTs as their counterparts from the center or right. However, the
Social Policy and Presidential Ideology in Latin America
187
case studies reveal that presidents in Argentina and Bolivia, just as in Brazil (Brearly 2011, 110;
Lindert et al. 2007, 13; Lustig 2014, 11), were initially skeptical, if not outright opposed, to these
programs. Politicians of the center and right in the three countries put CCTs on the national
political agenda. Leaders of the left eventually embraced these programs too. However, as
demonstrated by Chapter 3, once in charge of a CCT, left presidents tended to increase coverage
to levels beyond those achieved by presidents of the center or right. The increase in coverage is in
large part a consequence of their emphasis on reducing poverty over increasing human capital,
which leads them to adopt CCTs with broader targeting and laxer conditionality enforcement.

3. Avenues for Further Research
Three future projects will stem directly from this study. First, having explored the
determinants of social policy at the national level, future research will analyze social policy from
the perspective of the individual. Chapter 2 found a substantively large and robust effect of the
structure of inequality on aggregate levels of social spending. However, the theory that this finding
supports operates at the individual level: members of the middle class, displeased with the growing
gap between their wealth and that of the rich, align with poor individuals to steer democratic
governments toward increasing spending and redistribution. A more convincing test of this theory
must, therefore, take place at the individual level. Such a project is facilitated by the increased
availability of comparable surveys spanning multiple Latin American countries, specifically
Latinobarometro and the Latin American Public Opinion Project, and statistical packages capable
of estimating multilevel models.
Second, having demonstrated through case studies that ideology did in fact play a part in
bringing about the widespread adoption of CCTs in the region, the next step will be to demonstrate
Social Policy and Presidential Ideology in Latin America
188
this relationship statistically on a larger sample of countries. Doing so will require research aimed
at identifying the actors that first introduced CCTs into the political agenda of each country.
Third, future research should delve deeper than this study into explaining the effect of
ethnic diversity on CCT coverage. Chapter 3 found a substantively large and statistically robust
positive relationship between ethnic diversity and CCT coverage. Said finding goes against the
predictions of existing theories (Alesina and Glaeser 2004; Luttmer 2001) and past empirical
findings on the determinants of social spending (Chapter 2), poverty (Pribble, Huber and Stephens
2009) and inequality (Huber et al. 2006). Disappointingly, the case study of Bolivia, one of the
region’s most ethnically divided countries, yielded no additional information on this topic.

4. Concluding Remarks
The issues addressed in this study and outlined in the proposed research agenda discussed
above have a substantive impact on the daily lives of millions of Latin Americans, particularly the
more than one-quarter of the population living in poverty. CCTs can push a family above the
poverty line. Access to quality education directly affects an individual’s economic prospects.
Social policy can also be analyzed through the lens of democratic quality. Surveys show that most
Latin Americans equate democracy with the existence of a minimum-income floor (Huber and
Stephens 2010, 155). Unless their most basic needs are met, individuals cannot fully exert their
rights as citizens.
It is thus no wonder that research on the politics of social policy has developed into one of
the most vibrant research agendas within comparative politics (Amenta 2003; Pierson 2000).
However, most research on this topic has focused on the experiences of a handful of industrialized
countries (Carnes and Mares 2007). By better specifying the complex relationship between
Social Policy and Presidential Ideology in Latin America
189
presidential ideology and social policy in post-market reform Latin America, this study has
contributed toward reducing the geographic bias in this substantively important and theoretically
rich body of research.
Social Policy and Presidential Ideology in Latin America
190
References

Acemoglu, Daron, and James A. Robinson. 2006. Economic Origins of Dictatorship and
Democracy. New York: Cambridge University Press.
Achen, Christopher H. 2000. ‘‘Why Lagged Dependent Variables Can Suppress the Explanatory
Power of Other Independent Variables.’’ Presented at the Annual Meeting of Political
Methodology Section of the American Political Science Association, Los Angeles,
California.
__________. 2001. ‘‘Why Lagged Dependent Variables Can Suppress the Explanatory Power of
Other Independent Variables.’’ Presented at the Annual Meeting of Political Methodology
Section of the American Political Science Association, Los Angeles, California.
__________. 2002. “Toward a New Political Methodology: Microfundations and ART.” Annual
Review of Political Science 5: 423-450.
Adato, Michelle and John Hoddinot, ed. 2010. Conditional Cash Transfers in Latin America.
Baltimore: Johns Hopkins University Press/International Good Policy Research Institute.
Adserà, Alícia and Carles Boix. 2002. “Trade, Democracy, and the Size of the Public Sector: The
Political Underpinnings of Openness.” International Organization 56, 2:229-262.
Agis, Emanuel, Carlos Cañete, and Demian Panigo. 2009. “El Impacto de la Asignación
Universal por Hijo en Argentina.” Unpublished Document.
Aguiar, Marcelo, and Carlos Henrique Araujo. 2002. “Bolsa-Escola: Educación para Enfrentar la
Pobreza.” UNESCO.
Aguilar Pacajes, Hernán. 2014. “Evolución de Impacto del Bono Juancito Pinto en Bolivia.”
Paper presented at the 7th Encuento de Economistas de Bolivia, August 21.
Social Policy and Presidential Ideology in Latin America
191
Allan, James P. and Lyle Scruggs. 2004. “Political Partisanship and Welfare State Reform in
Advanced Industrial Societies.” American Journal of Political Science 48, 31:496-512.
Alesina, Alberto, Reza Baqir, and William Easterly. 1999. “Public Goods and Ethnic Divisions.”
The Quarterly Journal of Economics 114, 4:1243-1284.
Alesina, Alberto, and Edward L. Glaeser. 2004. Fighting Poverty in the US and Europe: A World
of Difference. New York: Oxford University Press.
Altman, David, and Rossana Castiglioni. 2009. “Democratic Quality and Human Development
in Latin America: 1972–2001.” Canadian Journal of Political Science 42, 2:297-319.
Alzúa, María Laura, Guillermo Cruces y Laura Ripani. 2013. “Welfare Programs and Labor
Supply in Developing Countries: Experimental Evidence from Latin America.” Journal of
Population Economics 26, 4:1255-1284.
Amenta, Edwin. 2003. “What We Know About the Development of Social Policy: Comparative
and Historical Research in Comparative and Historical Perspective.” In Comparative
Historical Analysis in the Social Sciences, ed. James Mahoney and Dietrich Rueschemeyer.
New York: Cambridge University Press. 91-130.
Ancelovici, Marcos, and Jane Jenson. 2013. Standardization for Transnational Diffusion: The
Case of Truth Commissions and Conditional Cash Transfers. International Political
Sociology 7, 3: 294-312.
Arceneaux, Kevin, and Gregory A. Huber. 2007. “What to Do (and Not Do) with
Multicollinearity in State Politics Research.” State Politics & Policy Quarterly 7, 1:81-101.
Arceo, Nicolás, Mariana González, and Nuria Mendizábal. 2010. “La Asignación Universal por
Hijo a un Año de Su Implementación.” Documento de Trabajo No 7, Centro de Investigación
y Formación de la República Argentina.
Social Policy and Presidential Ideology in Latin America
192
Ardanaz, Martin. 2009. “Preferences for Redistribution in the Land of Inequalities.” Unpublished
manuscript, Department of Political Science, Columbia University.
Arnold, Christian, David Doyle, and Nina Wiesehomeier. 2014. “Did Cristina Move Further to
the Left this Year?” The Presidential Speeches Blog, March 14. Available at:
http://textospoliticospresblog.net/did-cristina-move-further-to-the-left-this-year/.
Arza, Camila. 2012. “The Politics of Counter-Reform in the Argentine Pension System: Actors,
Political Discourse, and Policy Performance.” International Journal of Social Welfare 21,
Issue Supplement S1:S46–S60.
Austen-Smith, David. 2000. “Redistributing Income under Proportional Representation.” Journal
of Political Economy 108, 6:1235-126.
Avelino, George, David S. Brown, and Wendy Hunter. 2005. “The Effects of Capital Mobility,
Trade Openness and Democracy on Social Spending in Latin America, 1980-1999.”
American Journal of Political Science 49, 3:625-641.
Baez, Javier E., Adriana Camacho, Emily Conover, and Román A. Zárate. 2012. “Conditional
Cash Transfers, Political Participation, and Voting Behavior.” Impact Evaluation Series No.
66, Poverty Reduction and Economic Management Network, World Bank.
Baird, Sarah, McIntosh, Craig. & Özler, Berk. 2011. “Cash or Condition?: Evidence from a Cash
Transfer Experiment.” Quarterly Journal of Economics 126, 4:1709-1753.
Baird, Sarah, Francisco H. G. Ferreira, Berk Ozler, and Michael Woolcock. 2013. “Relative
Effectiveness of Conditional and Unconditional Cash Transfers for Schooling Outcomes in
Developing Countries: A Systematic Review.” Campbell Systematic Reviews 2013, 8.
Social Policy and Presidential Ideology in Latin America
193
Bandeira Coêho, Denilson. 2012. “Brazil: Basic Income - A New Model of Innovation
Diffusion.” In Basic Income Worldwide: Horizons of Reform, ed. Matthew C. Murray and
Carole Pateman. Baskingstoke, UK: Palgrave Macmillan. 59-80.
Banks, Arthur. S. 2014. Cross-National Time-Series Data Archive. Jerusalem, Israel.
Barbeito, Alberto C., and Rubén M. Lo Vuolo. 2009. “Ingreso Ciudadano para la Niñez:
Reelaborando Ideas para Construir una Sociedad Más Igualitaria.” Documento No. 19,
Centro Interdisciplinario para el Estudio de las Políticas Publicas.
Barrientos, Armando. 2004. “Latin America: A Liberal-Informal Welfare Regime?” In Insecurity
and Welfare Regimes in Asia, Africa and Latin America, ed. Ian Gough, Geof Wood,
Armando Barrientos, and Philippa Bevan. New York: Cambridge University Press. 121-168.
__________. 2009. “Labour Markets and the (Hyphenated) Welfare Regime in Latin America.”
Economy and Society 38, 1:87-108.
__________. 2008. “Latin America: Towards a Liberal-Informal Welfare Regime.” In Insecurity
and Welfare Regimes in Asia, Africa and Latin America: Social Policy in Development
Contexts, ed. Ian Gough and Geof Wood. New York: Cambridge University Press. 121-168.
__________. 2013. Social Assistance in Developing Countries. New York: Cambridge University
Press.
Barrientos, Armando, and and Lloyd-Sherlock, Peter. 2002. “Non-Contributory Pensions and
Social Protection.” Issues in Social Protection Discussion Paper No. 12, International Labour
Office.
Barrientos, Armando, and Claudio Santibáñez. 2009, “New Forms of Social Assistance and the
Evolution of Social Protection in Latin America.” Journal of Latin American Studies 41, 1:1-
26.
Social Policy and Presidential Ideology in Latin America
194
Bartels, Larry. M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age.
Princeton: Princeton University Press.
Bastagli, Francesca. 2009. “Conditionality in Public Policy Targeted to the Poor: Promoting
Resilience?” Social Policy and Society 8, 1:127-140.
_________. 2010. “Poverty, Inequality and Public Cash Transfers: Lessons from Latin America.”
Paper presented at the conference on Experiences and Lessons from Social Protection
Programmes Across the Developing World: What Role for the EU?, Paris, France, June 17-18.
Beck, Nathaniel. 1998. “Modelling Space and Time: The Event History Approach.” In Research
Strategies in the Social Sciences, ed. Elinor Scarbrough and Eric Tanenbaum. New York:
Oxford University Press. 191–213.
_________. 2001. “Time-Series-Cross-Section Data: What Have We Learned in the Past Few
Years?” Annual Review of Political Science Political Analysis 4:271-293.
Beck, Nathaniel, and Jonathan N. Katz. 1995. “What to Do (and Not to Do) with Time-Series
Cross-Section Data.” American Political Science Review 89, 3:634-647.
Beck, Nathaniel, Jonathan N. Katz, and Richard Tucker. 1998. “Taking Time Seriously: Time–
Series–Cross–Section Analysis with a Binary Dependent Variable.” American Journal of
Political Science 42, 4:1260-1288.
Benderly, Beryl Lieff. 2011. “A Bargain or a Burden? How Conditional Cash Transfer (CCT)
Program Design Affects the Women Who Participate in Them.” World Bank.
Bertranou, Fabio, and Roxana Maurizio. 2012. “Semi ‐conditional Cash Transfers in the form of
Family Allowances for Children and Adolescents in the Informal Economy in
Argentina.” International Social Security Review 65, 1:53-72.
Social Policy and Presidential Ideology in Latin America
195
Birdsall, Nacy, and Augusto de la Torre. 2001. Washington Contentious: Economic Policies for
Social Equity in Latin America. Washington, DC: Carnegie Endowment for Peace and Inter-
American Dialogue.
Birdsall, Nancy, Nora Lustig, and Darryl McLeod. 2012. “Declining Inequality in Latin America:
Some Economics, Some Politics”. In Routledge Handbook of Latin American Politics, ed.
Peter Kingstone and Deborah J. Yashar. London: Routledge. 158-180.
Birdsall, Nancy, and Miguel Székely. 2003. “Bootstraps not Band-Aids: Poverty, Equity and
Social Policy Latin America.” Working Paper No. 24, Center for Global Development.
Blaydes, Lisa, and Mark Andreas Kayser. 2011. “Counting Calories: Democracy and Distribution
in the Developing World.” International Studies Quarterly 55, 4:887-908.
Blustein, Paul. 2005. And the Money Kept Rolling In (and Out): Wall Street, the IMF, and the
Bankrupting of Argentina. New York: Public Affairs.
Boix, Carles. 2003. Democracy and Redistribution. New York Cambridge University Press.
Bohn, Simone R. 2011. “Social Policy and Vote in Brazil: Bolsa Família and the Shifts in Lula’s
Electoral Base.” Latin Research Review 46, 1: 54-79.
Bradshaw, Sarah, and Ana Quirós Víquez. 2008. “Women Beneficiaries or Women Bearing the
Cost? A Gendered Analysis of the Red de Protección Social in Nicaragua.” Development and
Change 39, 5:823-844.
Brearly, Emily. 2011. “The Politics of Poverty: the Political Economy of Social Protection in Latin
America and the Rise of Conditional Cash Transfers.” Doctoral Dissertation, John Hopkins-
School for Advanced International Studies.
Social Policy and Presidential Ideology in Latin America
196
Breceda, Karla, Jamele Rigolini, and Jaime Saavedra. 2009. “Latin America and the Social
Contract: Patterns of Social Spending and Taxation.” Population and Development Review 35,
4:721-748.
Brooks, Sarah M. 2002. “Social Protection and Economic Integration: The Politics of Pension
Reform in an Era of Capital Mobility.” Comparative Political Studies 35, 5:491-523.
_________. 2005. “Interdependent and Domestic Foundations of Policy Change: The Diffusion of
Pension Privatization around the World.” International Studies Quarterly 49, 2: 273-294.
_________. 2007. “Globalization and Pension Reform in Latin America.” Latin American Politics
& Society 49, 4:31-62.
_________. 2008. Social Protection and the Market in Latin America: The Transformation of
Social Security Institutions. Cambridge: Cambridge University Press.
Brown, David S., and Wendy Hunter. 1999. “Democracy and Social Spending in Latin America,
1980-92.” American Political Science Review 93, 4:779-790.
Bustos, Juan Martín, Georgina Giglio, and Soledad Villafañe. 2012. “Asignación Universal por
Hijo: Alcance e Impacto por Regiones del País.” In Trabajo, Ocupación y Empleo:
Investigaciones sobre Protección Social y Relaciones Laborales, ed. Ministerio de Trabajo,
Empleo y Seguridad Social. Buenos Aires: Ministerio de Trabajo, Empleo y Seguridad Social.
17-41.
Calabria, Alejandro, Analía Calero, Vanesa D’Elia, Julio Gaiada, and Sergio Rottenschweiler.
2010. “Transferencias Condicionadas de Ingreso en Argentina: La Asignacion Universal por
Hijo para Porteccion Social.” ANSES.
Social Policy and Presidential Ideology in Latin America
197
Caldés, Natália, David Coady, and & John A. Maluccio. 2006. “The Cost of Poverty Alleviation
Transfer Programs: A Comparative Analysis of Three Programs in Latin America.” World
Development 34, 5:818-837.
Calvas Chávez, Glenda Viviana. 2010. “Evaluación del Impacto del Bono de Desarrollo Humano
en la Educación.” Master´s Thesis, FLACSO Ecuador.
Calvi, Gabriel, Elsa Cimillo, and Horacio Chitaroni. 2010. “Alcances y Límites de la AUH en los
Primeros meses de su Implementación.” Paper presented at the Congreso Nacional de Estudios
del Trabajo, Buenos Aires.
Camacho, Adriana, Emily Conover, and Alejandro Hoyos. Forthcoming. “Effects of Colombia’s
Social Protection System on Workers’ Choice between Formal and Informal Employment.”
World Bank Economic Review.
Camargo, Gabriel Nicolás, and Juan Carlos Rodríguez-Raga 2011. “Familias en Acción y
preferencias políticas, Colombia 2010.” Apuntes de Cultural Política No. 1, Observatorio de
la Democracia, Universidad de los Andes.
Cameron, David R. 1978. “The Expansion of the Public Economy: A Comparative Analysis.”
American Political Science Review 72, 4:1243-1261.
Campos, Luis, Eleonor Faur, and Laura C. Pautassi. 2007. “Programa Familias por la Inclusión
Social: Entre el Discurso de Derechos y la Práctica Asistencial.” Colección Investigación y
Análisis, Centro de Estudios Legales y Sociales.
Carnes, Matthew E., and Isabela Mares. 2007. “The Welfare State in Global Perspective.” In The
Oxford Handbook of Comparative Politics, ed. Carles Boix and Susan C. Stokes. New York:
Oxford University Press. 868-885.
Social Policy and Presidential Ideology in Latin America
198
__________. 2014. “Coalitional Realignment and the Adoption of Non-contributory Social
Insurance Programmes in Latin America.” Socio-Economic Review 12, 4:695-722.
Carter, David B., and Curtis S. Signorino. 2010. “Back to the Future: Modeling Time Dependence
in Binary Data.” Political Analysis 18, 3:271-292.
Castañeda, Jorge G. 2006. “Latin America’s Left Turn.” Foreign Affairs 85, 3: 28-43.
Catterberg, Gabriela, and Valeria Palanza. 2012. “Argentina: Dispersión de la Oposición y el Auge
de Cristina Fernández de Kirchner.” Revista de Ciencia Política 32, 1:3-30.
Cecchini, Simone, and Aldo Madariaga. 2011. Programas de Transferencias Condicionadas:
Balance de Experiencia Reciente en América Latina y el Caribe. Santiago, Chile:
CEPAL/ASDI.
Cecchini, Simone, and Rodrigo Martínez. 2011. Protección Social Inclusiva en América Latina:
Una Mirada Integral, un Enfoque de Derechos. Santiago: CEPAL/ASDI.
Cheibub, José Antonio, Jennifer Gandhi, and James Raymond Vreeland. 2010. “Democracy and
Dictatorship Revisited.” Public Choice 143, 1-2:67-101.
Chen, Shaohua, and Martin Ravallion. 2012. “An Update to the World Bank's Estimates of
Consumption Poverty in the Developing World.” Briefing Note 03-01-12. Development
Research Group, World Bank.
Coady, David, Margaret Grosh, and John Hoddinott. 2004. “Targeting Outcomes Redux.” World
Bank Research Observer 19, 1:61-85.
Collier, David and Richard E. Messick. 1975. “Prerequisites versus Diffusion: Testing Alternative
Explanations of Social Security Adoption. American Political Science Review 69, 4:1299-
1315.
Social Policy and Presidential Ideology in Latin America
199
Contraloría General de la Republica (CGR). 2008. “Informe Sobre el Diseño y Ejecución del
Programa de Transferencia Monetaria Condicionada Denominado ‘Avancemos.’” División de
Fiscalización Operativa y Evaluativa, Área de Servicios Sociales.
__________. 2012. “Informe Sobre la Gestión del Programa de Transferencia Monetaria
Condicionada Denominado ‘Avancemos.’” División de Fiscalización Operativa y Evaluativa,
Área de Servicios Sociales.
Cornia, Giovanni Andrea. 2010. “Income Distribution under Latin America's New Left
Regimes.” Journal of Human Development and Capabilities 11, 1:85-114.
Cornelius, Wayne A. 2004. “Mobilized Voting in the 2000 Elections: The Changing Efficacy of
Vote Buying and Coercion in Mexican Electoral Politics.” In Mexico's Pivotal Democratic
Election: Candidates, Voters, and the Presidential Campaign of 2000, ed. Jorge I. Domínguez
and Chappell H. Lawson. Stanford: Stanford University Press. 47-65.
Corrales, Javier, and Michel Penfold. 2011. Dragon in the Tropics: Hugo Chavez and the Political
Economy of Revolution in Venezuela. Washington, DC: Brookings Institution Press.
Cox, Gary, and Scott Morgenstern. 2001. “Latin America’s Reactive Assemblies and Proactive
Presidents.” Comparative Politics 33, 2:171-189.
Collier, David, ed. 1979. The New Authoritarianism in Latin America. Princeton: Princeton
University Press.
Cortés, Fernando, and Rosa María Rubacalva. 2012. “El Progresa Como Respuesta a la Crisis de
1994.” In Pobreza, transferencias condicionadas y sociedad, ed. Mercedes González de la
Rocha and Agustín Escobar Latapí. Mexico City: Centro de Estudios Superiores en
Antropología Social. 27-50.
Social Policy and Presidential Ideology in Latin America
200
Crabtree, John. 2008. “Introduction: A Story of Unresolved Tensions.” In Unresolved Tensions:
Bolivia Past and Present, ed. John Crabtree and Laurence Whitehead. Pittsburgh: Pittsburgh
University Press. 1-7.
Cruces, Guillermo, and Leonardo Gasparini. 2008. “Programas Sociales en Argentina:
Alternativas para la Ampliación de la Cobertura.” Documento de Trabajo No. 77, Centro de
Estudios Distributivos, Laborales y Sociales, Universidad de la Plata.
__________. 2012. “Políticas sociales para la reducción de la desigualdad y la pobreza en América
Latina y el Caribe. Diagnóstico, propuesta y proyecciones en base a la experiencia reciente.”
Documento de Trabajo #142, Centro de Estudios Distributivos, Laborales y Sociales, Facultad
de Ciencias Económicas, Universidad Nacional de La Plata.
de Brauw, Alan, and John Hoddinott. 2011. “Must Conditional Cash Transfer Programs be
Conditioned to be Effective? The impact of Conditioning Transfers on School Enrollment in
Mexico." Journal of Development Economics 96, 2: 359-370.
de Hoop, Jacobus, and Furio C. Rosati. 2014. Cash Transfers and Child Labor. World Bank
Researcher Observer 29, 2: 202-234.
de Janvry, Alain, and Elisabeth Sadoulet. 2006. “When to use a CCT versus a CT approach?”
University of California, Berkeley and DECRG World Bank
de la O, Ana L. 2013. “Do Conditional Cash Transfers Affect Electoral Behavior? Evidence from
a Randomized Experiment in Mexico.” American Journal of Political Science 57, 1:1-14.
de León Naveiro, Omar. 2011. “Bolivia: La Construcción Participativa del Desarrollo.” in Nuevas
estrategias economicas en America Latina, ed. José Déniz Espinós, Omar de León y Antonio
Palazuelos Manso. Madrid: Catarata.
Social Policy and Presidential Ideology in Latin America
201
Defensoría de los Habitantes de la Repúbica de Costa Rica (Defensoría). 2010. Informe de Labores
2009-2010. San José: Defensoría de los Habitantes.
Díaz-Cayeros, Alberto and Beatriz Magaloni. 2009. “Poverty, Inequality, and Democracy: Aiding
Latin America’s Poor.” Journal of Democracy 20, 4:36-49.
Diaz-Cayeros, Alberto, Federico Estevez, and Beatriz Magaloni. 2009. “Welfare Benefits,
Canvassing, and Campaign Handouts.” In Consolidating Mexico's Democracy: The 2006
Presidential Campaign in Comparative Perspective, ed. Jorge I. Domínguez, Chappell H.
Lawson, and Alejandro Moreno. Baltimore: Johns Hopkins University Press. 229-245.
Diaz-Cayeros, Alberto, Beatriz Magaloni, and Federico Estevez. Forthcoming. Strategies of Vote
Buying: Democracy, Clientelism and Poverty Relief in Mexico. Unpublished Manuscript,
Stanford University.
Dinarte Romero, Geannina. 2009. “La Efectividad de la Gerencia Pública en Proyectos Sociales
Selectivos para la Solución de los Problemas de las Poblaciones Beneficiarias: Evaluación del
Proyecto ‘Avancemos’ desde un Enfoque basado en Objetivos.” Master’s Thesis, Instituto
Centroamericano de Administración Pública.
Dion, Michelle L. 2010. Workers and Welfare: Comparative Institutional Change in Twentieth-
Century Mexico. Pittsburgh, University of Pittsburgh Press.
Draibe, Sonia Miriam. 2006. “Brasil: Bolsa-Escola y Bolsa Familia.” In Transferencias con
Corresponsabilidad: Una Mirada Latinoamericana, ed. Ernesto Cohen and Rolando Franco.
Mexico: FLACSO. 139-78.
Duryea, Suzanne, and Andrew Morrison. 2004. “The Effect of Conditional Transfers on School
Performance and Child Labor: Evidence from an Ex-Post Impact Evaluation in Costa Rica.”
Working Paper No 606, Inter-American Development Bank Research Department.
Social Policy and Presidential Ideology in Latin America
202
Easterly, William, and Ross Levine. 1997. “Africa's Growth Tragedy: Policies and Ethnic
Divisions.” The Quarterly Journal of Economics 112, 4:1203-1250.
Economic Commission for Latin America and the Caribbean. 2012. CEPALSTAT Database.
Santiago, Chile.
Elkins, Zachary, and Beth Simmons. 2005. “On Waves, Clusters, and Diffusion: A Conceptual
Framework.” Annals of the American Academy of Political and Social Science 598, 1:33-51.
Escobar Latapí, Agustín, and Mercedes González de la Rocha. 2008. “Girls, Mothers, and Poverty
Reduction in Mexico: Evaluating Progresa-Oportunidades.” In The Gendered Impacts of
Liberalization: Towards Embedded Liberalism?, ed. Shahra Razavi. New York:
Routledge/UNRISD. 435-68.
Esping-Andersen, Gøsta. 1985. Politics against Markets: The Social Democratic Road to Power.
Princeton: Princeton University Press.
__________. 1990. The Three Worlds of Welfare Capitalism. Princeton: Princeton University
Press.
Estevan, Fernanda. 2013. “The impact of conditional cash transfers on public education
expenditures: A political economy approach.” European Journal of Political Economy 32:268-
284.
Estevez-Abe, Margarita, Torben Iversen and David Soskice. 2001. “Social Protection and the
Formation of Skills: a Reinterpretation of the Welfare State.” In Varieties of Capitalism: The
Institutional Foundations of Comparative Advantage, Peter A. Hall and David Soskice, ed.
New York: Oxford University Press. 145-183.
Social Policy and Presidential Ideology in Latin America
203
Etchemendy, Sebastián, and Candelaria Garay. 2011. “Argentina: Left Populism in Comparative
Perspective, 2003-2009.” In The Resurgence of the Latin American Left, ed. Steven Levitsky
and Kenneth Roberts. Baltimore: John Hopkins University Press. 283-305.
Eterovic, Dalibor S., and Cassandra M. Sweet. 2014. “Democracy and Education in Twentieth-
Century Latin America.” Economics & Politics 26, 2:237-262.
Evans, Mark, and Jonathan Davies. 1999. “Understanding Policy Transfer: A Multi ‐Level, Multi ‐
Disciplinary Perspective.” Public Administration 77, 2:361-85.
Ewig, Christina and Stephen J. Kay. 2011. “Postretrenchment Politics: Policy Feedback in Chile’s
Health and Pension Reforms.” Latin American Politics and Society 53, 4: 67-99.
Feitosa de Britto, Tatiana. 2008. “The Emergence and Popularity of Conditional Cash Transfers
in Latin America” In Social Protection for the Poor and Poorest: Concepts, Policies and
Politics, ed. Armando Barrientos and David Humle. Baskingstoke: Palgrave MacMillan. 321-
362.
Fenwick, Tracy Beck. 2013. “Stuck between the past and the future: Conditional cash transfer
programme development and policy feedbacks in Brazil and Argentina.” Global Social Policy
13, 2:144-67.
Ferreira, Francisco H. G., and David Robalino. 2011. “Social Protection in Latin America:
Achievements and Limitations.” In Oxford Handbook of Latin American Economics, ed. José
Antonio Ocampo and Jaime Ros. New York: Oxford University Press. 836-862.
Federación Internacional de Administradoras de Fondos de Pensiones (FIAP). 2011. “Non-
Contributory Pension Programs in FIAP Countries —Part I: Latin America.” Santiago, Chile:
FIAP.
Social Policy and Presidential Ideology in Latin America
204
Filgueira, Fernando F. 2005. “Welfare and Democracy in Latin America: The Development, Crises
and Aftermath of Universal, Dual and Exclusionary Social States.” UNRISD Project on Social
Policy and Democratization.
Filgueira, Carlos H., and Fernando Filgueira. 2002. “Models of Welfare and Models of Capitalism:
The Limits of Transferability.” In Models of Capitalism: Lessons for Latin America, ed.
Evelyne Huber. State College, Penn State Press. 127-58.
Filgueira, Fernando, Carlos Gerardo Molina, Jorge Papadópulos, and Federico Tobar. 2006.
“Universalismo Básico: Una Alternativa Posible y Necesaria para Mejorar las Condiciones de
Vida.” In Universalismo Básico: Una Nueva Política Social para América Latina, ed. Carlos
Gerardo Molina. Washington, DC: Inter-American Development Bank. 19-55.
Fiszbein, Ariel, Ravi Kabur, and Ruslan Yemtsov. 2014. Social Protection and Poverty Reduction:
Global Patterns and Some Targets. World Development 61:167–177.
Fiszbein, Ariel, and Norbert Schady with Francisco H.G. Ferreira, Margaret Grosh, Nial Kelleher,
Pedro Olinto, and Emmanuel Skoufias. 2009. Conditional Cash Transfers: Reducing Present
and Future Poverty. Washington, DC: World Bank.
Flores-Macías, Gustavo. 2012. After Neoliberalism?: The Left and Economic Reforms in Latin
America. New York: Oxford University Press.
Fuentes, Federico. 2014. “‘Bad Left Government’ versus ‘Good Left Social Movements’?:
Creative Tensions within Bolivia’s Process of Change.” In Latin America's Radical Left:
Challenges and Complexities of Political Power in the Twenty-first Century, ed. Steve Ellner
and William I. Robinson. New York: Rowman & Littlefield Publishers. 103-125.
Fundación Jubileo. 2012. "Renta Hidrocarburifera: Mas Allá de las Regalias y del IDH." Serie
Debate Publico 24.
Social Policy and Presidential Ideology in Latin America
205
Galasso, Emanuela, and Martin Ravallion. 2004. “Social Protection in a Crisis: Argentina's Plan
Jefes y Jefas.” World Bank Economic Review 18, 3:367-399.
Garay, María Candelaria. 2010. “Including Outsiders: Social Policy Expansion in Latin America.”
Doctoral Dissertation, University of California, Berkeley.
Garrett, Geoffrey. 1998. Partisan Politics in the Global Economy. New York: Cambridge
University Press.
__________. 2001. “Globalization and Government Spending around the World.” Studies in
Comparative International Development 35, 4:3-29.
Garganta, Santiago, and Leonardo Gasparini. 2012. “El Impacto de un Programa Social sobre la
Informalidad Laboral: El Caso de la AUH en Argentina.” Documento de Trabajo No. 133,
Centro de Estudios Distributivos, Laborales y Sociales, Universidad Nacional de la Plata.
Gasparini, Leonardo, and Guillermo Cruces. 2010. “Las Asignaciones Universales por Hijo en
Argentina Impacto, Discusión y Alternativas.” Económica 57:104-146.
Gelbach, Jonah B. and Lant Pritchett. 2002. “Is More for the Poor is Less for the Poor?: The
Politics of Means-Tested Targeting.” Topics in Economic Analysis & Policy 2, 1:1-26.
George, Alexander L. and Andrew Bennett. 2005. Case Studies and Theory Development in the
Social Sciences. Cambridge: MIT Press.
Gerring, John, Strom C. Thacker, and Rodrigo Alfaro. 2012. “Democracy and Human
Development.” Journal of Politics 74, 1:1–17.
Giovagnoli, Paula. 2005. “Poverty Alleviation or Political Networking? A combined Qual-Quant
Analysis of the Implementation of Safety Nets in post-crisis Argentina.” Working Paper No.
05-66, Development Studies Institute, London School of Economics and Political Science.
Social Policy and Presidential Ideology in Latin America
206
Giraudy, Agustina. 2007. “The Distributive Politics of Emergency Employment Programs in
Argentina (1993-2002).” Latin American Research Review 42, 2:33-55.
Golbert, Laura. 2004. “Derecho a la Inclusión o Paz Social? Plan Jefas y Jefes de Hogar
Desocupados.” Serie Políticas Social No. 84, División de Desarollo Social, CEPAL.
Golbert, Laura, and Claudia Giacometti. 2008. “Los Programas de Transferencia de Ingresos
Condicionados: El Caso Argentino.” Instituto Fernando Henrique Cardoso and Corporación
de Estudios para Latinoamérica.
Golbert, Laura, and Emilia Roca. 2010. De la Sociedad de Beneficencia a los Derechos Sociales.
Buenos Aires: Ministerio de Trabajo, Empleo y Seguridad Social.
Goñi, Edwin, J. Humberto López, and Luis Servén. 2011. “Fiscal Redistribution and Income
Inequality in Latin America.” World Development 39, 9:1558–1569.
Goodman, Margaret S., Samuel Morely, Gabriel Siri, and Elaine Zuckerman. 1997. Social
Investment Funds in Latin America: Past Performance and Future Role. Washington, DC:
Inter-American Development Bank.
Gray Molina, George. 2010. “The Challenge of Progressive Change under Evo Morales.” In Leftist
Governments in Latin America: Successes and Shortcomings, ed. Kurt Weyland, Raúl L.
Madrid and Wendy Hunter. New York: Cambridge University Press. 57-76.
Gray Molina, George, and Ernesto Yañéz. 2009. “The Dynamics of Inequality in the Best and
Worst of Times, Bolivia 1997-2007.” UNDP Bureau for Development Policy/Poverty Group.
Green, Tina. 2006. “Do Social Transfer Programs Affect Voter Behavior? Evidence from Progresa
in Mexico, 1997-2000.” Working Paper, University of California, Berkeley.
Social Policy and Presidential Ideology in Latin America
207
Green-Pedersen, Christoffer. 2008. The Politics of Justification: Party Competition and Welfare-
State Retrenchment in Denmark and the Netherlands from 1982 to 1998. Amsterdam:
Amsterdam University Press.
Grigoli, Francesco and Giacomo Sbrana. 2013. Determinants and Dynamics of Schooling and
Child Labor in Bolivia. Bulletin of Economic Research 65, s1:s17–s37.
Grimes, Marcia and Lena Wängnerud. 2010. “Curbing Corruption through Social Welfare
Reform? The Effects of Mexico’s Conditional Cash Transfer Program on Good Government.”
American Review of Public Administration 40, 6:671-690.
Hacker, Jacob S. and Paul Pierson. 2002. “Business Power and Social Policy: Employers and the
Formation of the American Welfare State.” Politics and Society 30, 2:277-325.
Haggard, Stephan, and Robert R. Kaufman. 2008. Development, Democracy, and Welfare States:
Latin America, East Asia, and Eastern Europe. Princeton: Princeton University Press.
Haggard, Stephan, Robert R. Kaufman, and James D. Long. 2013. “Income, Occupation, and
Preferences for Redistribution in the Developing World.” Studies in Comparative
International Development 48, 2:113-140.
Hall, Anthony. 2008. “Brazil’s Bolsa Família: A Double-Edged Sword?” Development and
Change 39, 5:799-822.
Handa, Sudhanshu and Benjamin Davis. 2006. “The Experience of Conditional Cash Transfers in
Latin America and the Caribbean.” Development Policy Review 24, 5:513-536.
Hanlon, Joseph, Armando Barrientos, and David Hulme. 2010. Just Give Money to the Poor: The
Development Revolution from the Global South. Sterling, Virginia: Kumarian Press.
Social Policy and Presidential Ideology in Latin America
208
Hernani-Limarino, Werner L. 2009. “Assessing the Impact of School Subsidies in Bolivia: A
Reduced Form Non-Parametric Approach.” Paper presented at the annual meeting of the Latin
American and Caribbean Economics Association, Buenos Aires, Oct. 1-3.
Hicks, Alexander. 1994. ‘‘Introduction to Pooling.’’ In The Comparative Political Economy of the
Welfare State, ed. Alexander Hicks and Thomas Janoski. New York: Cambridge University
Press. 169–188.
Huber, Evelyne. 1996. “Options for Social Policy in Latin America: Neoliberal versus Social
Democratic Models.” In Welfare States in Transition, ed. Gøsta Esping-Andersen. London:
Sage Publications. 141-191.
__________. 2005. “Globalization and Social Policy Developments in Latin America.” In
Globalization and the Future of the Welfare State, ed. Miguel Glatzer and Dietrich
Rueschemeyer. Pittsburgh: University of Pittsburg Press. 75-105.
__________. 2009. “Including the Middle Classes? Latin American Social Policies after the
Washington Consensus.” In Doing Good or Doing Better: Development Policies in a
Globalizing World, ed. Monique Kremer, Peter van Lieshout, and Robert Went. Amsterdam:
Amsterdam University Press. 137-55.
Huber, Evelyne, Thomas Mustillo and John D. Stephens. 2008. “Politics and Social Spending in
Latin America.” Journal of Politics 70, 2:420-436.
Huber, Evelyne, Francois Nielsen, Jenny Pribble and John D. Stephens. 2006. “Political
Determinants of Inequality in Latin America and the Caribbean.” American Sociological
Review 71, 6:943-963.
Huber, Evelyne, and John D. Stephens. 2001. Development and Crisis of the Welfare State.
Chicago: University of Chicago Press.
Social Policy and Presidential Ideology in Latin America
209
__________. 2005. “Welfare States and the Economy.” In The Handbook of Economic Sociology
(second edition), ed. Neil J. Smelser and Richard Swedberg. Princeton: Princeton University
Press. 552-574.
__________. 2010. “Successful Social Policy Regimes? Political Economy, Politics and Social
Policy in Argentina, Chile, Uruguay, and Costa Rica.” In Democratic Governance in Latin
America, ed. Scott Mainwaring and Timothy Scully. Stanford: Stanford University Press. 155-
209.
__________. 2012. Democracy and the Left: Social Policy and Inequality in Latin America.
Chicago: University of Chicago Press.
Hunter, Wendy, and Timothy J. Power. 2007. “Rewarding Lula: Executive Power, Social Policy,
and the Brazilian Elections of 2006.” Latin American Politics and Society 49, 1: 1-30.
Infobae. 2013. “Más de 12 mil mujeres empezarán a cobrar la asignación por embarazo.”
Infobae.com, December 2.
International Institute for Democracy and Electoral Assistance (IDEA). Voter Turnout Database.
Stockholm, Sweden.
Iversen, Torben. 2005. Capitalism, Democracy, and Welfare. New York: Cambridge University
Press.
Iversen, Torben, and Thomas R. Cusack. “The Causes of Welfare State Expansion:
Deindustrialziation or Globalization? World Politics 52, 3:313-349.
Iversen, Torben, and David Soskice. 2006. “Electoral Institutions and the Politics of Coalitions:
Why Some Democracies Redistribute More Than Others.” American Political Science Review
100, 2:165-181.
Social Policy and Presidential Ideology in Latin America
210
Kantor, Mora. 2012. “El Proceso de Formulación de la Asignación Universal por Hijo y el Plan
Jefas y Jefes de Hogar Desocupados en la Argentina: Una Perspectiva Comparada.” Paper
presented at the Latin American Studies Association Conference, San Francisco, CA, May 23-
26.
Kaufman, Robert R. 2009. The Political Effects of Inequality in Latin America: Some
Inconvenient Facts. Comparative Politics 41, 3:359-379.
Kaufman, Robert R., and Alex Segura-Ubiergo. 2001. “Globalization, Domestic Politics, and
Social Spending in Latin America: A Time-Series Cross-Section Analysis, 1973-97.” World
Politics 53, 4:553-587.
Kenworthy, Lane, and Leslie McCall. 2008. “Inequality, Public Opinion and Redistribution.”
Socio-Economic Review 6, 1:35–68.
Kenworthy, Lane, and Jonas Pontusson. 2005. “Rising Inequality and the Politics of Redistribution
in Affluent Countries.” Perspectives on Politics 3, 3:449-471.
Kim, Doo-Sub. 2009. “A Standardization Technique to Reduce the Problem of Multicollinearity
in Polynomial Regression Analysis.” Department of Sociology, Hanyang University.
Kitschelt, Herbert and Steven I. Wilkinson. 2007. “Citizen–Politician Linkages: An Introduction.”
In Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political
Competition, ed. Herbert Kitschelt and Steven I. Wilkinson. New York: Cambridge University
Press. 1-49.
Korpi, Walter. 1983. The Democratic Class Struggle. London: Routledge, Kegan & Paul.
__________. 2006. Power Resources and Employer Centered Approaches in Explanations of
Welfare States and Varieties of Capitalism. World Politics 58, 2:167-206.
Social Policy and Presidential Ideology in Latin America
211
Korpi, Walter and Joakim Palme. 2003. “New Politics and Class Politics in the Context of
Austerity and Globalization: Welfare State Regress in 18 Countries, 1975–95.” American
Political Science Review 97, 3:425-446.
La Jornada. 2009. “El Banco Mundial financiará el 50% del bono ‘Juana Azurduy’ este 2009.” 15
de abril.
La Nación. 2008. “Zumbado se Forja Imagen de ‘Ministro de los Pobres.’” April 16.
__________. 2009. “Gobierno Desiste de Dar Beca a 40.000 Colegiales Pobres.” September 21.
__________. 2010. “Avancemos Será Sometido a Examen.” April 22.
Lake, David A. and Matthew A. Baum. 2001. “The Invisible Hand of Democracy: Political Control
and the Provision of Public Services.” Comparative Political Studies 34, 6:587-621.
Lana, Xenia, and Mark Evans. 2004. “Policy Transfer between Developing Countries: The
Transfer of the Bolsa-Escola Programme to Ecuador.” In Policy Transfer in Global
Perspective, ed. Mark Evans. Aldershot: Ashgate. 190-210.
Lance, Justin Earl. 2014. “Conditional Cash Transfers and the Effect on Recent Murder Rates in
Brazil and Mexico.” Latin American Politics and Society 56, 1: 55-72.
Laserna, Roberto. 2007. Entre el Ch’enko y el Rentismo: ‘Riqueza Nacional para la Ciudadanía’,
una Opción para la Democracia y el Desarrollo en Bolivia. La Paz: Fundación Milenio.
__________. 2008. “Entre el Ch’enko y el Rentismo: Democracia y Desarrollo en Bolivia.” In
Bolivia 25 Años Construyendo la Democracia, ed. Vicepresidencia de la República. La Paz:
Vicepresidencia de la República.87-110.
Lavinas, Lena. 2013. “21
st
Century Welfare.” New Left Review 84, 6: 5-40.
Social Policy and Presidential Ideology in Latin America
212
Layton, Matthew L., and Amy Erica Smith 2011. “Social Assistance Policies, Economic Crisis,
and the Presidential Vote in Latin America.” Paper presented at the Annual Meeting of the
American Political Science Association, Seattle, Washington, September 2.
Levitsky, Steven. 2003. Transforming Labor-Based Parties in Latin America: Argentine Peronism
in Comparative Perspective. New York: Cambridge University Press.
Levitsky, Steven, and Kenneth Roberts, ed. 2011. The Resurgence of the Latin American Left.
Baltimore: John Hopkins University Press.
Levy, Santiago. 1991. “Poverty Alleviation in Mexico.” Policy, Research and External Affairs
Working Paper No. 679, The World Bank.
__________. 2006. Progress against Poverty: Sustaining Mexico’s Progresa-Oportunidades
Program. Washington, DC: Brookings Institution Press.
__________. 2008. Good Intentions, Bad Outcomes: Social Policy, Informality, and Economic
Growth in Mexico. Washington, DC: Brookings Institution Press.
Lindert, Kathy, Emmanuel Skoufias, and Joseph Shapiro. 2006. “Redistribution of Income to the
Poor and the Rich: Public Transfers in Latin America and the Caribbean.” Social Policy
Discussion Paper #0605. Washington, DC: World Bank.
Lieberman, Evan S. 2005. “Nested Analysis as a Mixed-Method Strategy for Comparative
Research.” American Political Science Review 99, 3:435-52.
__________. 2009. Boundaries of Contagion: How Ethnic Politics Have Shaped Government
Responses to AIDS. Princeton: Princeton University Press.
Lindert, Kathy, Emmanuel Skoufias, and Joseph Shapiro. 2006. “Redistribution of Income to the
Poor and the Rich: Public Transfers in Latin America and the Caribbean.” Washington, DC:
World Bank.
Social Policy and Presidential Ideology in Latin America
213
Lindert, Kathy, and Vanina Vincensini. 2010. “Social Policy, Perceptions and the Press: An
Analysis of the Media’s Treatment of Conditional Cash Transfers in Brazil.” SP Discussion
Paper No. 1008, Social Protection and Labor Division, World Bank.
Lindert, Peter H. 2004. Growing Public Social Spending and Economic Growth since the
Eighteenth Century. New York: Cambridge University Press.
Linos, Elizabeth. 2013. “Do Conditional Cash Transfer Programs Shift Votes? Evidence from the
Honduran PRAF.” Electoral Studies 32, 4:864-74.
López-Calva, Luis Felipe, and Nora Lustig, ed. 2010. Declining Inequality in Latin America: A
Decade of Progress? Washington, D.C: Brookings Institution.
Lo Vuolo, Rubén. 2008. “Why Basic Income is Better than Renewed Policy Promises for Latin
American Informal Security Regimes.” Paper represented at the 12th Congress of the Basic
Income Earth Network, Dublin, Ireland June 20-21.
__________. 2013a. “Epilogue.” In Citizen´s Income and Welfare Regimes in Latin America:
From Cash Transfers to Rights, ed. Rubén Lo Vuolo. Baskingtoke: Palgrave MacMillan. 259-
266.
__________. 2013b. “Introduction.” In Citizen´s Income and Welfare Regimes in Latin America:
From Cash Transfers to Rights, ed, Rubén Lo Vuolo. Baskingtoke: Palgrave MacMillan. 1-28.
__________. 2013c. “The Argentine ‘Universal Child Allowance’: Not the Poor but the
Unemployed and Informal Workers.” In Citizen´s Income and Welfare Regimes in Latin
America: From Cash Transfers to Rights, ed, Rubén Lo Vuolo. Baskingtoke: Palgrave
MacMillan. 51-66.
Social Policy and Presidential Ideology in Latin America
214
Lo Vuolo, Rubén, Alberto Barbeito, Laura Pautassi, and Corina Rodríguez. 1999. La Pobreza...
de la Política contra la Pobreza. Buenos Aires. Centro Interdisciplinario para el Estudio de
Políticas Públicas.
Lübker, Malte. 2007. “Inequality and the Demand for Redistribution: are the Assumptions of the
New Growth Theory Valid? Socio-Economic Review 5, 1:117–148.
Lupu, Noam, and Jonas Pontusson. 2011. “The Structure of Inequality and the Politics of
Redistribution.” American Political Science Review 105, 2:316-336.
Lustig, Nora, coordinator. 2011. “Fiscal Policy and Income Redistribution in Latin America:
Challenging the Conventional Wisdom.” Working Paper #1124, Economics Department,
Tulane University.
__________. 2014. “Scholars Who Became Practitioners: The Influence of Research on the
Design, Evaluation, and Political Survival of Mexico’s Antipoverty Program.”` In Scholars,
Policymakers, and International Affairs: Finding Common Cause, ed. Abraham F. Lowenthal
and Mariano E. Bertucci. Baltimore: Johns Hopkins University Press. 105-118.
Lustig, Nora, Luis F. López-Calva, and Eduardo Ortíz-Juarez. 2013. “Declining Inequality in Latin
America in the 2000s: The Cases of Argentina, Brazil, and Mexico.” World Development
44:129-141.
Luttmer, Enzo F. P. 2001. “Group Loyalty and the Taste for Redistribution.” Journal of Political
Economy 109, 3:500-528.
Madrid, Raúl. 2003. Retiring the State: The Politics of Pension Privatization in Latin America and
Beyond. Stanford: Stanford University Press.
Social Policy and Presidential Ideology in Latin America
215
__________. 2011. “Origins and Policies of the Movimiento al Socialismo.” In The Resurgence
of the Latin American Left, ed. Steven Levitsky and Kenneth M. Roberts. Baltimore: Johns
Hopkins University Press. 239-259.
Mahler, Vincent A. 2008. “Electoral Turnout and Income Redistribution by the State: A Cross-
National Analysis of the Developed Democracies.” European Journal of Political Research
47, 2:161-183.
Mahon, James E. 2012. “Tax Incidence and Tax Reforms in Latin America.” Update on the
Americas, Woodrow Wilson Center.
Malloy, James. 1979. The Politics of Social Security in Brazil. Pittsburgh: University of Pittsburgh
Press.
Manacorda, Marco, Edward Miguel and Andrea Vigorito. 2011. Government Transfers and
Political Support. American Economic Journal: Applied Economics 3, 3:1-28.
Mares, Isabela. 2003. The Politics of Social Risk: Business and Welfare State Development. New
York: Cambridge University Press.
__________. 2009. “The Comparative Political Economy of the Welfare State.” In Comparative
Politics: Rationality, Culture, and Structure, 2nd Edition, ed. Mark Irving Lichbach and Alan
S. Zuckerman. New York: Cambridge University Press. 358-375.
Martínez Franzoni, Juliana. 2008. “Welfare Regimes in Latin America: Capturing Constellations
and Markets, Families, and Policies.” Latin American Politics and Society 50, 2:67-100.
Martínez Franzoni, Juliana and Diego Sánchez-Ancochea. 2012. “The Road to Universal Social
Protection: How Costa Rica Informs Theory.” Working Paper No. 383, Kellog Institute, Notre
Dame University.
Social Policy and Presidential Ideology in Latin America
216
Martínez Franzoni, Juliana, and Koen Voorend. 2011. “Actors and Ideas behind CCTs in Chile,
Costa Rica and El Salvador.” Global Social Policy 11, 2-3: 279–298.
Mata Hidalgo, Catherine, and Karla Hernández Romero. 2013. “Evaluación de Impacto de
Transferencias Monetarias Condicionadas en Costa Rica: Programa Avancemos.” Licenciature
Thesis, Universidad de Costa Rica.
McGuire, James W. 2010. Wealth, Health and Democracy in East Asia and Latin America. New
York: Cambridge University Press.
__________. 2013. “Conditional Cash Transfers in Bolivia: Origins, Impact and Universality.”
Paper presented at the annual meeting of the International Studies Association, San Francisco,
California, April 3-6.
McKenzie, David. 2004. “Aggregate Shocks and Labor Market Responses: Evidence from
Argentina's Financial Crisis.” Economic Development and Cultural Change 52, 4:719-58.
Medinaceli, Mauricio, and Leila Mokrani. 2010. “Impacto de los Bonos Financiados con la Renta
Petrolera.” Revista Umbrales 20:223-263.
Melo, Marcus André. 2008. “Unexpected Successes, Unanticipated Failures: Social Policy from
Cardoso to Lula.” In Democratic Brazil Revisited, ed. Peter R. Kingstone and Timothy J.
Power. Pittsburgh: University of Pittsburgh Press. 161-184.
Meltzer, Allan H., and Scott F. Richard. 1981. “A Rational Theory of the Size of Government.”
Journal of Political Economy 89, 5:914-927.
Mesa-Lago, Carmelo. 1978. Social Security in Latin America: Pressure Groups, Stratification and
Inequality. Pittsburgh: University of Pittsburgh Press.
__________. 1989. Ascent to Bankruptcy: Financing Social Security in Latin America. Pittsburgh:
University of Pittsburgh Press.
Social Policy and Presidential Ideology in Latin America
217
__________. 1994. Changing Social Security in Latin America: Toward Alleviating the Social
Costs of Economic Reform. Boulder, CO: Lynne Rienner Publishers.
__________. 2007. “Pension and Healthcare Reform.” Latin American Research Review 42,
2:181-201.
Meza ‐Cordero, Jaime A. 2011. “The Effects of Subsidizing Secondary Schooling: Evidence from
a Conditional Cash Transfer Program in Costa Rica.” Unpublished Document, University of
Southern California.
Milanovic, Branko. 2000. “The Median-Voter Hypothesis, Income Inequality, and Income
Redistribution: an Empirical Test with the Required Data.” European Journal of Political
Economy 16, 3:367-410.
Ministerio Público Tutelar de la Ciudad Autonóma de Buenos Aires (MPT). 2011. Programa
Ciudadanía Porteñas ¿Con Todo Derecho?: Fortalezas y Debilidades de las Transferencias
Monetarias Condicionadas en Niños, Niñas y Adolecentes. Eudeba: Buenos Aires.
Moene, Karl Ove, and Michael Wallerstein. 2003. “Earnings Inequality and Welfare Spending: A
Disaggregated Analysis.” World Politics 55, 4:485-516.
Molyneux, Maxine. 2006. “Mothers at the Service of the New Poverty Agenda:
Progresa/Oportunidades, Mexico's Conditional Transfer Programme.” Social Policy &
Administration 40, 4:425-49.
Montero, Alfred P. 2010. “No Country for Leftists? Clientelist Continuity and the 2006 Vote in
the Brazilian Northeast.” Journal of Politics in Latin America 2, 2:113-153.
Monza, Alfredo, and Claudia Giacometti. 2003. “Los Beneficiarios del Plan Jefes y Jefas de
Hogar.” Documentos de Trabajo No 1, Proyecto Cooperación OIT/Gobierno Argentino
(MTSS).
Social Policy and Presidential Ideology in Latin America
218
Moore, Charity. 2008. “Assessing Honduras’ CCT Programme PRAF, Programa de Asignación
Familiar: Expected and Unexpected Realities.” Country Study 15, International Poverty
Centre.
Morales Dada, María José. 2007. “Trabajo Infantil y Programas de Transferencias Monetarias
Condicionadas: el Caso del Programa AVANCEMOS de Costa Rica.” Programa
Avancemos/Organización Internacional del Trabajo.
Müller, Katharina. 2009. “Contested Universalism: from Bonosol to Renta Dignidad in Bolivia.”
International Journal of Social Welfare 18:163-172.
Murillo, María Victoria, Virginia Oliveros, and Milan Vaishnav. 2010. Dataset on Political
Ideology of Presidents and Parties in Latin America. Columbia University.
Mussa, Michael. 2002. Argentina and the Fund: From Triumph to Tragedy. Washington, DC:
Peterson Institute for International Economics.
Niedzwiecki, Sara. 2010. “Social Policy in South America: The Effect of Organized Labor and
Civil Society Groups on Pension and Health Reforms in Argentina and Brazil (1988-2008).
M.A. Thesis, University of North Carolina, Chapel Hill.
Newman, Newman, Steen Jorgensen, and Menno Pradhan. 1991. “How Did Workers Benefit from
Bolivia's Emergency Social Fund?” World Bank Economic Review 5, 2:367-393.
Nupia, Oskar. 2011. “Anti-Poverty programs and Presidential Election Outcomes: Familias en
Acción in Colombia.” Center of Economic Development Studies, Universidad de los Andes.
Obinger, Herbert, Carina Schmitt, and Peter Starke. 2013. “Policy Diffusion and Policy Transfer
in Comparative Welfare State Research.” Social Policy & Administration 47, 1:111-29.
Observatorio de la Deuda Social de la Universidad Católica Argentina (ODSA-UCA). 2013.
“Estimación del Impacto de la AUH y las Pensiones para Madres con Siete Hijos Sobre los
Social Policy and Presidential Ideology in Latin America
219
Ingresos Familiares, la Indigencia y la Pobreza Urbana en la Argentina del Bicentenario.”
Buenos Aires.
Observatorio de la Deuda Social de la Universidad Católica Argentina (ODSA-UCA). 2014.
“Estimaciones de Tasas de Indigencia y Pobreza (2010-2013): Totales Urbanos.” Buenos
Aires.
Observatorio de la Educación Básica Argentina (OEBA). 2012. “Informe de resultados: Módulo
del Programa Asignación Universal por Hijo (AUH).” Centro de Estudios en Políticas
Públicas, Universidad de Buenos Aires.
Overseas Development Institute. 2010. “Bolivia Case Study for the MDG Gap Task Force Report.”
Case Study for the MDG Gap Task Force Report.
Paus, Eva. 2005. Foreign Investment, Development, and Globalization: Can Costa Rica Become
Ireland? Baskingstoke: Palgrave Macmillan.
Pardo, Marcia. 2003. “Reseña de Programas Sociales para la Superación de la Pobreza en América
Latina.” División de Estadística y Proyecciones Económicas, CEPAL.
Pastor, Robert. 2001. Exiting The Whirlpool: U.S. Foreign Policy Toward Latin America And The
Caribbean, 2nd Edition. Boulder: Westview Press.
Paz Arauco, Verónica, George Gray Molina, Wilson Jiménez Pozo, and Ernesto Yañez Alguilar.
2013. “Explaining Low Redistributive Impact in Bolivia.” Working Paper No. 6, Commitment
to Equity, Inter-American Dialogue/Tulane University Center for Inter-American Policy and
Research.
Perotti, Roberto. 2006. “Growth, Income Distribution, and Democracy: What the Data Say.”
Journal of Economic Growth 1, 2:149-187.
Social Policy and Presidential Ideology in Latin America
220
Perry, Guillermo, William F. Maloney, Omar S. Arias, Pablo Fajnzylber, Andrew D. Mason, and
Jaime Saavedra-Chanduv. 2007. Informality: Exit and Exclusion. Washington, DC: The World
Bank.
Persson, Torsten, and Guido Tabellini. 2003. The Economic Effects of Constitutions. What Do The
Data Say? Cambridge: MIT Press.
Pierson, Paul. 1994. Dismantling the Welfare State? Reagan, Thatcher and the Politics of
Retrenchment. New York: Cambridge University Press.
__________. 1996. “The New Politics of the Welfare State.” World Politics 48, 2: 143-179.
__________. 2000. “Three Worlds of Welfare State Research.” Comparative Political Studies 33,
6/7:791-821.
Plümper, Thomas, Vera Troeger, and Philip Manow. 2005. “Panel Data Analysis in Comparative
Politics: Linking Method to Theory.” European Journal of Political Research 44, 2:327-354.
Políticas Públicas Populi. 2012. “El Bono Juancito Pinto: Algunas Reflexiones Socio-
Económicas.” Asuntos Públicos 13, 6:1-4.
Ponce, Juan, and Arjun S. Bedi. 2010. “The Impact of Conditional Cash Transfer Program on
Students’ Cognitive Achievements: The Case of the 'Bono de Desarrollo Humano' of
Ecuador.” Economics of Education Review 29, 1:116-125.
Pribble, Jennifer. 2021. “Worlds Apart: Social Policy Regimes in Latin America.” Studies in
Comparative International Development 46, 2:191-216.
__________. 2013. Welfare and Party Politics in Latin America. New York: Cambridge
University Press.
Pribble, Jennifer, Evelyne Huber & John D. Stephens. 2009. “Politics, Policies, and Poverty in
Latin America.” Comparative Politics 41, 4:387-407.
Social Policy and Presidential Ideology in Latin America
221
Przeworski, Adam, Michael E. Alvarez, Jose Antonio Cheibub, and Fernando Limongi. 2000. New
York: Cambridge University Press.
Rawlings, Laura B. and Gloria M. Rubio. 2005. “Evaluating the Impact of Conditional Cash
Transfer Programs.” World Bank Research Observer 20, 1:29-55.
Repetto, Fabián and Gala Díaz Langou. 2010. “Desafíos y Enseñanzas de la Asignación Universal
por Hijo para Protección Social a un Año de su Creación. Documento de Políticas Públicas
Nº88. Centro de Implementación de Políticas Públicas para la Equidad y el Crecimiento.
Repetto, Fabián, Gala Díaz Langou, and Vanesa Marazzi. 2009. “¿Hacia un Sistema de Protección
Social Integral? El Ingreso para la Niñez es Sólo la Punta del Ovillo.” Documento de Políticas
Públicas/Análisis No. 67, Centro de Implementación de Políticas Públicas para la Equidad y
el Crecimiento.
Repetto, Fabián, and Fernanda Potenza Dal Masetto. 2011. “Protección Social en la Argentina.”
División de Desarrollo Social, CEPAL.
Rodríguez-Clare, Andrés. 2001. "Costa Rica’s Development Strategy based on Human Capital and
Technology: How it Got there, the Impact of Intel, and Lessons for Other Countries." Paper
prepared for the Human Development Report of 2001, UNDP.
Rofman, Rafael, Carlos Grushka, and Victor Chebez. 2001. “El Sistema de Asignaciones
Familiares como Herramienta Central en la Política Social Argentina.” Paper presented at the
6th Congress of Centro Latinoaméricano para el Desarrollo sobre la Reforma del Estado y la
Administración Pública, Buenos Aires, November 5-9.
Rodrik, Dani. 1997. Has Globalization Gone Too Far? Washington: Institute for International
Economics.
Social Policy and Presidential Ideology in Latin America
222
__________. 1998. “Why Do More Open Economies Have Bigger Governments?” Journal of
Political Economy 106, 5:997-1032.
Román, Isabel. 2010. “Sustentabilidad de los Programas de Transferencias Condicionadas: la
Experiencia del Instituto Mixto de Ayuda Social y ‘Avancemos’ en Costa Rica.” Serie de
Políticas Sociales No. 160, CEPAL.
Román Vega, Isabel. 2012. Social Protection Systems in Latin America and the Caribbean: Costa
Rica. CEPAL.
Rossi, Federico M. 2013. “Juggling Multiple Agendas: The Struggle of Trade Unions against
National, Continental, and International Neoliberalism in Argentina.” In Transnational
Activism and National Movements in Latin America: Bridging the Divide, ed. Eduardo Silva.
London: Routledge. 141-60.
Rudra, Nita. 2002. “Globalization and the Decline of the Welfare State in Less-Developed
Countries.” International Organization 56, 2:411-445.
__________. 2004. “Openness, Welfare Spending, and Inequality in the Developing World.”
International Studies Quarterly 48, 3: 683-709.
__________. 2007. “Welfare States in Developing Countries: Unique or Universal?” Journal of
Politics 62, 2:378-396.
__________. 2008. Globalization and the Race to the Bottom in Developing Countries: Who
Really Gets Hurt? New York: Cambridge University Press.
Rudra, Nita, and Stephan Haggard. 2005. “Globalization, Democracy and Effective Welfare
Spending in the Developing World.” Comparative Political Studies 38, 9:1015-1049.
Social Policy and Presidential Ideology in Latin America
223
Saavedra, Juan Esteban, and Sandra García. 2012. “Impact of Conditional Cash Transfer Programs
on Educational Outcomes in Developing Countries: A Meta Analysis.” Working Paper WR-
921-1. Rand Corporation.
Salvia, Agustín, Bianca Musante, and Alejandro Mendoza Jaramillo. 2013. “Análisis de Impacto
de la AUH en Materia de Inseguridad Alimentaria y Déficit Educativo.” Observatorio de la
Deuda Social, Universidad Católica de Argentina.
Sandbrook, Richard, Marc Edelman, Patrick Heller, and, Judith Teichman. 2006. Social
Democracy in the Global Periphery: Origins, Challenges, Prospects. New York: Cambridge
University Press.
Sauma, Pablo. 2011. “Pobreza, Desigualdad en la Distribución del Ingreso y Empleo en un
Contexto de Lenta Recuperación Económica y Crisis Fiscal.” Ponencia para el XVII Informe
Estado de la Nación en Desarrollo Humano Sostenible.
__________. “Análisis de la Suficiencia de los Montos de la Transferencia de Avancemos.”
Unpublished Document, Avancemos.
Schady, Norbert, and María Caridad Araujo. 2008. “Cash Transfers, Conditions, and School
Enrollment in Ecuador.” Economía 8, 2:43-70.
Schneider, Ben Ross, and David Soskice. 2009. “Inequality in Developed Countries and Latin
America: Coordinated, Liberal and Hierarchical Systems.” Economy and Society 38, 1:17-52.
Seawright, Jason, and John Gerring. 2008. “Case Selection Techniques in Case Study Research:
A Menu of Qualitative and Quantitative Options.” Political Research Quarterly 61, 2:294-308.
Segura-Ubiergo, Alex. 2007. The Political Economy of the Welfare State in Latin America:
Globalization, Democracy and Development. New York: Cambridge University Press.
Social Policy and Presidential Ideology in Latin America
224
Simoes, Carla. 2012. “Rousseff Expands Brazil Welfare Benefits to Target Child Poverty.”
Bloomberg.com, May 14.
Siri, Gabriel. 2000. “Employment and Social Investment Funds in Latin America.” International
Labour Office.
Skocpol, Theda and Edwin Amenta. 1986. “States and Social Policies.” Annual Review of
Sociology 12:131–57.
Skoufias Emmanuel, and Bonnie McClafferty. 2001. “Is PROGRESA Working? Summary of the
Reuslts of an Evaluation by IFPRI.” Food Consumption and Nutrition Division, International
Food Policy Research Institute.
Soares, Sergei, Rafael Guerreiro Osório, Fábio Veras Soares, Marcelo Medeiros, and Eduardo
Zepeda. 2009. “Conditional Cash Transfers in Brazil, Chile and Mexico: Impacts upon
Inequality.” Estudios Económicos. Special Issue:207-224.
Smith, Peter H. 2012. Talons of the Eagle: Latin America, the United States, and the World, 4th
Edition. New York: Oxford University Press.
Social Communication Office of the Presidency of Brazil (SECOM). 2013. “Brazil without
Extreme Poverty (Brasil Sem Miséria): With the Support of Federal Programs, More than 22
million Brazilians have Overcome Extreme poverty.” Highlights 6, 1:12-15.
Socio-Economic Database for Latin America and the Caribbean (SEDLAC). 2014. “Tasas de
pobreza oficiales (nacionales).” La Plata: Argentina.
Stampini, Marco, and Leopoldo Tornarolli. 2012. “The Growth of Conditional Cash Transfers in
Latin America and the Caribbean: Did They Go Too Far? Policy Brief No. IDB-PB-185, Social
Sector, Social Protection and Health Division, Inter-American Development Bank.
Social Policy and Presidential Ideology in Latin America
225
Standing, Guy. 2008. “How Cash Transfers Promote the Case for Basic Income.” Basic Income
Studies 3, 1: 1-30.
Starke, Peter. 2013. “Qualitative Methods for the Study of Policy Diffusion: Challenges and
Available Solutions.” Policy Studies Journal 41, 4:561-582.
Stasavage, David. 2005. Democracy and Education Spending in Africa. American Journal of
Political Science 49, 2:343-58.
Stephens, John D. 1979. The Transition from Capitalism to Socialism. London: Macmillan.
Stokes, Susan C., Thad Dunning, Marcelo Nazareno, and Valeria Brusco. 2013. Brokers, Voters,
and Clietelism: The Puzzle of Distributive Politics. New York: Cambridge University Press.
Straschnoy, Mora. 2011. “Asignación Universal por Hijo para Protección Social y Nuevo Régimen
de Asignaciones Familiares: un Análisis Comparativo de los Casos de Argentina y
Uruguay.” Revista Catédra Paralela 8:111-126.
Sugiyama, Natasha Borges. 2011. “The Diffusion of Conditional Cash Transfer Programs in the
Americas.” Global Social Policy 11, 2-3: 250–278.
__________. 2013. Diffusion of Good Government: Social Sector Reforms in Brazil. Notre Dame,
Indiana: University of Notre Dame Press.
Sugiyama, Natasha Borges, and Wendy Hunter 2013. Whither Clientelism?: Good Governance
and Brazil’s Bolsa Família Program. Comparative Politics 46, 1: 43-62.
Swank, Duane. 2002. Global Capital, Political Institutions, and Policy Change in Developed
Welfare States. New York: Cambridge University Press.
__________. 2013. Diffusion of Good Government: Social Sector Reforms in Brazil. Notre Dame:
University of Notre Dame Press.
Social Policy and Presidential Ideology in Latin America
226
Swenson, Peter. 1991. “Bringing Capital Back in, or Social Democracy Reconsidered: Employer
Power, Cross-Class Alliances, and Centralization of Industrial Relations in Denmark and
Sweden.” World Politics 43, 4:513-544.
__________. 2002. Capitalists against Markets: The Making of Labor Markets and Welfare States
in the United States and Sweden. New York: Oxford University Press.
Tanzi, Vito. 2000. “Taxation in Latin American in the Last Decade.” Working Paper No. 76,
Center for Research on Economic Development and Policy Reform, Stanford University.
Tapia Huanaco, Blanca, Orlando Murillo Zambrana, and Susana Flores Sotomayor. 2011. Bono
“Juancito Pinto”: Evaluación de Resultados. Observatorio Social de Políticas Educativas de
Bolivia (OSPE-B). La Paz: Bolivia.
Timmons, Jeffrey F. 2005. “The Fiscal Contract: States, Taxes, and Public Services.” World
Politics 57, 4:530-567.
Trejos, Juan Diego. 2012. “La Inversión Social en el 2011: Evolución en un Contexto de Lenta
Recuperación Económica y Crisis Fiscal.” Ponencia para el XIIX Informe Estado de la Nación
en Desarrollo Humano Sostenible.
United Nations Development Program (UNDP). 2007. El Estado del Estado en Bolivia: Informe
Nacional sabre Desarrollo Humano en Bolivia. La Paz: UNDP.
United Nations Economic Commission for Latin America and the Caribbean (CEPAL). 2014.
“CEPALSTAT Database.” Santiago, Chile.
United Nations Office on Drugs and Crime (UNDOC). 2014. “United Nations Surveys on Crime
Trends and the Operations of Criminal Justice Systems.” Geneva, Switzerland: Untied
Nations.
Social Policy and Presidential Ideology in Latin America
227
Valencia Lomelí, Enrique. 2008. “Conditional Cash Transfers and Social Policy in Latin America:
An Assessment of their Contributions and Limitations.” Annual Review of Sociology 34:475-
99.
Vargas, Juan Rafael, Pablo Slon, and Fundación Acción Jóven. 2012. “¿Protegen los programas
sociales de la deserción educativa en Costa Rica?” Ponencia para el IV Informe Estado de la
Educación.
Vargas Sandoval, Yensi, Marianela González Zúñiga, Marilyn Sánchez Fallas, and Marvin
Rodríguez Vargas. 2012. “Las Transferencias Económicas del Programa AVANCEMOS del
Instituto Mixto de Ayuda Social y su Aporte en el Componente Educativo y Psicosocial
(Período 2006-2010), los Casos de: Liceo de Aserrí Liceo Experimental Bilingüe de Pococí y
Colegio Técnico Profesional de Guápiles.” Facultad de Educación/Instituto de Investigación
en Educación, Universidad de Costa Rica.
Vera Cossío, Diego A. 2011. “Matriculación y Trabajo Infantil en Bolivia: Un Análisis Quasi
Experimental.” Serie de Documentos de Trabajo sobre Desarrollo, Instituto de Estudios
Avanzados en Desarrollo.
Veras Soares, Fábio, Rafael Perez Ribas, and Rafael Guerreiro Osório. 2010. “Evaluating the
Impact of Brazil’s Bolsa Família: Cash Transfer Programs in Comparative Perspective.” Latin
American Research Review 24, 2:173-90.
Viceministerio de Aseguramiento No Contributivo y Movilidad Social, República de Ecuador
(VAMS). 2014. “Reporte Anual VAMS.” Quito, Ecuador.
Villalobos, María Lourdes. 2012. “El Planeamiento, Ejecución y Evaluación de las Transferencias
Monetarias Condicionadas: Orígenes y Transformaciones del Programa Avancemos, Costa
Social Policy and Presidential Ideology in Latin America
228
Rica.” Unpublished Document, Instituto de Investigaciones en Ciencias Económicas,
Universidad de Costa Rica.
Villareal F. Evelyn, and Gómez A. Steffan. 2010. “Costa Rica 2009: Enfrentando la Crisis
Internacional en Plena Campaña Electoral.” Revista de Ciencia Política 30, 2:275-296
Villatoro, Guillermo. 2007. Las Transferencias Condicionadas en América Latina: Luces y
Sombras. Santiago, Chile: CEPAL/IPEA.
Víquez, Diego. 2010. “Estratificación Social y Movilidad Social Ascendente: Implicaciones para
la Convivencia Democrática y la Cohesión Social.” In Costa Rica en Tiempo Preelectoral:
Cohesión Social, Reforma del Estado y Descentralización: Retos y Perspectivas desde
Diversos Actores, ed. Yahaira Ceciliano. San José: FLACSO. 29-42.
Weyland, Kurt. 2005. “Theories of Policy Diffusion: Lessons from Latin American Pension
Reform.” World Politics 57, 2:262-295.
__________. 2007. Bounded Rationality and Policy Diffusion: Social Sector Reform in Latin
America. Princeton: Princeton University Press.
__________. 2011. “The Performance of Leftist Governments in Latin America: Conceptual and
Theoretical Issues.” In Leftist Governments in Latin America: Successes and Shortcomings,
ed. Kurt Weyland, Raúl L. Madrid, and Wendy Hunter. New York: Cambridge University
Press. 1-27.
Weyland, Kurt, Raúl L. Madrid, and Wendy Hunter, ed. Leftist Governments in Latin America:
Successes and Shortcomings. New York: Cambridge University Press.
Wibbels, Erik. 2006. “Dependency Revisited: International Markets, Business Cycles, and Social
Spending in the Developing World.” International Organization 60, 2:433-468.
Social Policy and Presidential Ideology in Latin America
229
Wibbels, Erik and John S. Ahlquist. 2011. “Development, Trade, and Social Insurance.”
International Studies Quarterly 55, 1:125-149.
Wilensky, Harold L. 1975. The Welfare State and Equality: Structural and Ideological Roots of
Public Expenditure. Berkeley: University of California Press.
Williamson, John. 1989 “What Washington Means by Policy Reform.” In Latin American
Readjustment: How Much has Happened?, ed. John Williamson,. Washington, DC: Institute
for International Economics.
World Bank. 1982. World Development Report, 1982. Washington, D.C.: World Bank.
__________. 2001. World Development Report 2000/2001: Attacking Poverty. New York: Oxford
University Press.
__________. 2005. “Argentina: Seeking Sustained Growth and Social Equity: Observations on
Growth, Inequality and Poverty.” Report No. 32553-AR, Argentina, Chile, Paraguay, and
Uruguay Country Management Unit and Poverty Reduction and Economic Management Unit.
__________. “Social Gains in the Balance: A Fiscal Policy Challenge for Latin America and the
Caribbean.” Latin America and Caribbean Region.
__________. 2014. World Development Indicators. Washington, D.C.: World Bank.
Yanes, Pablo. 2013. “Targeting and Conditionalities in Mexico: The End of the Cash Transfer
Model.” In Citizen’s Income and Welfare Regimes in Latin America: From Cash Transfers to
Rights, ed. Rubén Lo Vuolo. Baskingstoke: Palgrave MacMillan. 67-85.
Yáñez Aguilar, Ernesto. 2012. “El Impacto del Bono Juancito Pinto: Un Análisis a Partir de Micro
Simulaciones.” Revista Latinoamericana de Desarrollo Económico 17:75-112.
Yaschine Arroyo, Iliana. 2012. “¿Oportunidades?: Movilidad Social Intergeneracional e Impacto
en México.” Doctoral Dissertation, Colegio de México.
Social Policy and Presidential Ideology in Latin America
230
Yi, Dae Jin. 2013. “Politics and Income Inequality: Does Politics Still Matter in New
Democracies?” New Political Economy 18, 5: 653-79.
Zarate Tenorio, Bárbara. Forthcoming. “Social Spending Responses to Organized Labor and Mass
Protests in Latin America, 1970-2007.” Comparative Political Studies.
Zedillo, Ernesto. 2009. "Keynote Address: A Look at Progresa’s Genesis.” Presentation at
conference on The Origins, Implementation, and Spread of Conditional Cash Transfer
Programs in Latin America, University of Texas at Austin, April 17.
Zucco, Cesar. 2008. “The President’s ‘New’ Constituency: Lula and the Pragmatic Vote in Brazil’s
2006 Presidential Elections.” Journal of Latin American Studies 40, 1: 29–49.
__________. 2013. “When Payouts Pay Off: Conditional Cash Transfers and Voting Behavior in
Brazil 2002–10.” American Journal of Political Science 57, 4: 810–822.
Zucco, Cesar Jr. and Timothy Power. 2013. “Bolsa Família and the Shift in Lula’s Electoral Base,
2002-2006: A Reply to Bohn.” Latin American Research Review 48, 2:3-24.
Social Policy and Presidential Ideology in Latin America
231
End Notes

1
. It should be noted that in Chile the government of the Michelle Bachelet of the center-left
Socialist Party was followed by that of Sebastián Piñera (2010-2014) of the center-right
Renovación Nacional party. Bachelet was subsequently reelected and returned to the presidency
in March 2014. In Paraguay, Fernando Lugo a left-leaning former priest was elected president in
2008 but impeached in 2012.
2
. Weyland, Madrid and Hunter (2010) call these, respectively, the moderate and contestatory left.
For a critique of this dichotomy, see Levitsky and Roberts (2011) and Flores-Macías (2012).
3
. The left also gained power in Guatemala following the 2007 election of Alvaro Colom of the
center-left National Unity of Hope (GANA) party. However, voters returned the right to power
during the 2011 elections.
4
. The most notable example is democracy, which is a constant, not a variable, in the advanced
industrialized countries.
5
. In contrast to other work on developing countries, Brown and Hunter (1999) find openness to
trade leads to higher social spending.
6
. Pensions are technically deferred earnings.
7
. Schneider and Soskice (2009) posit that Latin America’s mix of majoritarian executives and
proportional legislatures contributed to the persistence of high inequality in the region.
8
. A growing literature is dedicated to categorizing Latin American welfare regimes. These
typologies are generated either inductively based on case knowledge (Filgueira 2005; Huber and
Stephens 2009) or statistically through the use of cluster analysis (Martínez Franzoni 2008; Pribble
2011; Rudra 2007). The main strength of this typological approach is that it goes beyond aggregate

Social Policy and Presidential Ideology in Latin America
232

spending measures in mapping out the general contours of countries’ social policies. Rather than
being based on spending levels, typologies are derived from measures of social policy success and
priorities.
9
. See also Malloy (1979) for a detailed analysis of the emergence of social insurance in Brazil.
10
. The emblematic examples of this were Argentina and Chile under bureaucratic authoritarian
regimes during the 1970s and 1980s, see Collier (1979).
11
. Segura-Ubiergo (2007) defines the existence of welfare states in the region dichotomously. The
countries with welfare states are Argentina, Chile, Costa Rica, Uruguay, and, more controversially,
Brazil.
12
. In fact, there is evidence that union membership is associated with reduced support for
redistribution in developing countries (Haggard, Kaufman and Long 2013) and that labor strength
negatively affects income equality and even delays improvements in infant mortality (McGuire
2009). For a dissenting view on the role of unions in Latin America, see Huber and Stephens (2012)
and Niedzwiecki (2010).
13
. Following Levitsky and Roberts (2011), I classify the Peronists in Argentina and the Alianza
Popular Revolucionaria Americana (APRA) in Peru as populist machines with flexible ideologies.
While they leaned left, it would be hard to argue that they were social democratic.
14
. This presumed link is based on the fact that Latin America’s ISI economies adopted social
security-intensive social policies while export-oriented East Asian countries adopted human
capital-intensive social policies.
15
. In turn, countries with large domestic markets, scarce labor, unequal land distribution, and less
resource wealth are more likely to adopt ISI.
Social Policy and Presidential Ideology in Latin America
233

16
. Human capital investment is defined in the literature as the sum of spending on education and
healthcare. In chapter 3, I will disaggregate this type of spending into its component parts.
17
. Haggard and Kaufman (2008) do find that democracy has a negative effect on social security
spending among Latin American countries during 1980-2000.
18
. In contrast. Yi (2013) finds no effect on democracy on inequality among a sample of 26 new
democracies.
19
. The impact on social spending of another component of PRT, labor unions, remains contested.
Given the large disparities that exist among developing countries in terms of union autonomy and
capacity, Rudra (2002) compiles a measure of “potential labor power” (PLP), which is positively
related with social spending. However, PLP’s positive effect decreases as economic integration
increases. Wibbels (2006) confirms the positive effect on spending of PLP, but finds a negative
effect of union density on total social spending. Using her own index of labor strength,
Niedzwiecki (2010) finds an association between strong unions and increased social spending.
20
. See also Huber, Mustillo and Stephens (2008).
21
. While openness to trade has been at the center of research on globalization’s effect on social
spending, scholars have also analyzed other aspects. Looking at economic volatility, Wibbels
(2006) finds that economic shocks lead Latin American governments to cut human capital
investment while leaving social security transfers untouched. This may be explained the
differences in the political power the differing constituencies behind these forms of spending
(Zarate Tenorio Forthcoming). Evidence of a negative effect of foreign direct investment (Huber,
Mustillo and Stephens 2008; Niedzwicki 2010) and capital mobility are largely inconclusive
(Avelino, Brown and Hunter 2005; Kaufman and Segura-Ubiergo 2001; Rudra 2002; Rudra and
Haggard 2005; Wibbels 2006).
Social Policy and Presidential Ideology in Latin America
234

22
. See end note #2.
23
. There is disagreement on how the two lefts have fared with regard to the reduction of inequality.
Whereas Cornia (2010) finds that populist governments were more successful than social
democratic governments, Birdsall, Lustig and McLeod (2012) find the opposite.
24
. Other commonly used measures include spending per capita and as a share of total government
spending.
25
. Examples of research that has overcome these difficulties include: Eterovic & Sweet (2014);
Huber and Stephens (2012); Wibbels and Ahlquist (2011).
26
. Venezuela is a notable exception to this trend. Nicaragua adopted a CCT program in 2002 but
cancelled it in 2006.
27
. These calculations include only the 13 Latin American countries for which Stampini and
Tornarolli (2012) had data. The countries excluded are: Bolivia, Dominican Republic, El Salvador,
Honduras, and Nicaragua.
28
. It should be noted that much of what we know about the impact of CCTs on education, health
and nutrition draws on the experiences of Mexico and, to a lesser extent, Brazil.
29
. This has occurred despite the fact that few CCT programs explicitly cite reducing child labor as
one of their goals. Notable exceptions include Costa Rica’s Avancemos and Brazil’s Programa de
Erradicação do Trabalho Infantil (PETI), which was rolled into Bolsa Familia in 2003.
30
. There is evidence that social policies explicitly targeting informal-sector workers disincentivize
formalization (Garganta and Gasparini 2012) and may even promote informality (Camacho,
Conover and Hoyos Forthcoming; Levy 2008).
31
. Much maligned for ignoring poverty and redistribution and most closely associated with trade
and investment liberalization, privatization and conservative fiscal and monetary policy, the
Social Policy and Presidential Ideology in Latin America
235

Washington Consensus actually called for the “redirection of public spending from subsidies
(“especially indiscriminate subsidies”) toward broad-based provision of key pro-growth, pro-poor
services like primary education, primary health care and infrastructure investment” (Williamson
1989).
32
. Technically, the first CCT implemented in the region was Chile’s Subsidio Único Familiar in
1981. It provided the equivalent to $6 per month to indigent mothers with school-age children
conditional on school attendance, pregnant women and women caring for people. At its peak, it
covered less than 1,000 women (Lavinas 2013, 11).
33
. The actual figure is 51.37 percent. This assumes a poverty line of $2.50 (2005 dollars PPP).
Using a poverty line of $4.00 (2005 dollars PPP), the average share of the region’s poor covered
drops to 39.04 percent. On average, CCTs boosted the incomes of the poor by 34.5 percent
(Stampini and Tornarolli 2012).
34
. For a dissenting views of the Brazilian and Mexican cases, respectively see Bohn (2011) and
Green (2005). Looking at Honduras, Linos (2013) finds no effect of CCTs on the presidential
elections.
35
. It is worth restating that the PRI, not the PAN, launched Progresa/Oportinidades.
36
. Research on Honduras by Linos (2013) constitutes a notable exception. She finds that a CCT
increased the likelihood of mayoral reelection by 39%.
37
. These are Chile, Colombia, Mexico, and Uruguay.
38
. These are Brazil, Chile and Uruguay.
39
. It should be noted that part of the reduction in Latin America’s inequality that occurred between
the 2000s and 1990s may have been due to a reduction in the skill premium, meaning that in that
Social Policy and Presidential Ideology in Latin America
236

growth in the supply of skills outpaced the demand for them. Having more education did not in fact
translate into higher earnings during that period (Cornia 2010; Lopez-Calva and Lustig 2010, 13).
40
. Hanlon, Barrientos and Hulme (2010, 130) concede that conditionality may be useful for
keeping secondary students in school. Teenagers may not properly value the benefits of remaining
in school.
41
. Molyneux’s (2006, 432) critique is directed specifically at Mexico’s Progresa/Oportunidades,
which in addition to the standard education and health conditionality, requires that participating
mothers attend workshops on health and contribute “a set amount of hours of work to the program,
typically cleaning buildings or clearing rubbish.”
42
. The views of Latin Americans are remarkably similar to those of Europeans and sharply contrast
with those of Americans (Ardanaz 2009, 23).
43
. The 18 countries are: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican
Republic, El Salvador, Ecuador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay,
Peru, Uruguay, and Venezuela.
44
. Huber & Stephens (2012, 35) posit that PRT and MVT make opposite predictions based on level
of inequality. A highly unequal distribution of income, they argue, leads to a highly unequal
distribution of political power, which in turn reduces the political system's responsiveness to
demands for redistribution. In contrast, left governments reduce inequality in the distribution of
political power and thus improve the prospects for redistribution.
45
. Kenworthy & Pontusson (2005) provocatively argue that voter turnout may be the missing link
that connects PRT with the MVT. Nonetheless, there is a significant difference in the political
arguments of PRT and MVT in that the former envisions a lengthy democratic process while the
latter only posits an event, voting.
Social Policy and Presidential Ideology in Latin America
237

46
. The remaining countries in the region are excluded for not being democratic or for lack of data.
While presidential ideology data cover the 1980s, reliable and disaggregated data on income
distribution is available only as of 1990.
47
. Weyland, Madrid & Hunter (2010) call these, respectively, the moderate and contestatory left.
For a critique of this dichotomy, see Levitsky & Roberts (2011) and Flores-Macías (2012).
48
. The dataset was been updated to 2010.
49
. Income distribution data is seldom published on a yearly basis. As a result and because income
distribution changes slowly, missing observations were estimated using linear interpolation in
order to make the data compatible with the time-series, cross-sectional research design. Linear
interpolation was used on all the measures of income distribution: Gini coefficient, quintile 5 to
quintile 1 ratio, decile 9 to decile 5 ratio, decile 5 to decile 1 ratio, and income distribution skew.
50
. Venezuela has not reported its level of social spending since 2006. There is disagreement over
how to categorize the left-leaning Argentine presidencies of Néstor Kirchner (2003-2007) and
Cristina Fernández de Kirchner (2007-present). The models presented follow Murillo, Oliveros &
Vaishnav (2010) in coding these governments as center-left. However, coding them as far-left did
not substantially alter the results of the models tested.
51
. Given space limitations, baseline models distinguishing between center-left and far-left
presidents are not presented.
52
. They find that populists outspent non-left governments, but were outspent by social democrats.
The paper’s main finding is that social democrats were more successful at reducing inequality than
populist. Cornia (2010) finds the opposite.
53
. The negative effect on human capital spending in Huber, Mustillo & Stephens (2008), disappears
in their later work (Huber and Stephens 2012, 138-39). Importantly, the latter work finds a negative
Social Policy and Presidential Ideology in Latin America
238

effect of left governments on both poverty and inequality. Based on these findings, they conclude
that the region’s left-leaning governments have made social policy more progressive.
54
. Chávez and Alvaro Garcia Linera, Vice-President under Morales, have argued that “the
organized working class has failed to live up to the revolutionary expectations inherent in
traditional Marxism” (Ellner, 2014, 9).
55
. This finding lends additional support for past findings that show that having a democratic
government has a positive effect on spending levels (Avelino, Brown & Hunter, 2005; Brown &
Hunter, 1999; Kaufman & Segura-Ubiergo, 2001; Rudra, 2002; Segura-Ubiergo, 2007; Stasavage,
2005; Zarate Tenorio, forthcoming).
56
. Defined as less than $2.5 a day.
57
. The 18 countries are: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican
Republic, El Salvador, Ecuador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay,
Peru, Uruguay, and Venezuela.
58
. Brazil pioneered CCTs at the sub-national level starting in 1995 (Sugiyama 2013).
59
. Own calculations utilizing data from Cruces & Gasparini (2012, 55) and CEPALSTAT.
60
. Carnes & Mares (2014) analyze the adoption on non-contributory pensions in Latin America.
61
. Data has been updated up to 2010.
62
. Strikes are defined as involving at least 1,000 workers and affecting multiple employers.
Demonstrations are peaceful gatherings of 100 or more with the purpose of voicing opposition to
government policies or authority. Riots are violent demonstrations or clashes involving 100 or more
(Banks 2011).
Social Policy and Presidential Ideology in Latin America
239

63
. A dummy variable is used rather than the GDP growth rate because the hypothesis assesses
whether crises make adoption more likely, not whether adoption becomes more likely as the
economy worsens. Using GDP growth does not alter the findings.
64
. Indirect taxes primarily consist of taxes on goods and services (CEPALSTAT 2014).
65
. No distinction is made between whether a peer is also a contiguous neighbor. Given that the
drivers behind diffusion are the presence of the shared political, economic and cultural traits, and
constant interaction, Latin American countries should be as likely to emulate non-contiguous peers
as contiguous neighbors.
66
. Sub-national level research on Brazil does find that left-leaning government are more likely to
adopt CCTs (Sugiyama 2013).
67
. Kim (1999) subtracts the mean from each observation of the linear variable and then divides the
result by the original variable’s standard deviation. The result is then squared and cubed.
68
. Also see: Avelino, Brown & Hunter (2005); Brown & Hunter (1999); Kaufman & Segura-
Ubiergo (2001); Wibbels (2006), Zarate Tenorio (Forthcoming).
69
. The latter is calculated by dividing the number of CCT beneficiaries by an estimate of the
poverty headcount. The latter is estimated by multiplying the share of the population living in
poverty time the total population. This estimate does not measure whether CCT beneficiaries are
actually people living in poverty.
70
. No Latin American country receives the highest possible score (1.0 or extremely strong/multiple
boundaries).
71
. Chávez’s presidency (1999-2013) spans 80 percent of the time period studied.
72
. Regardless of this provision, there is evidence that Bolsa Familia played a decisive role in Lula’s
reelection (Hunter & Power 2007; Zuco 2008).
Social Policy and Presidential Ideology in Latin America
240

73
. The program covers the elderly and disabled. As these are unconditional, they are excluded from
the tally of program coverage.
74
. Weyland, Madrid & Hunter (2010) call these, respectively, the moderate and contestatory left.
For a critique of this dichotomy, see Levitsky & Roberts (2011) and Flores-Macías (2012).
75
. Notable exceptions include the nationalizations of the national water utility and postal service
(Etchemendy & Garay 2011, 293).
76
. Bastagli (2010) is more optimistic about the possibility of striking a balance between both goals.
77
. Cecchini and Martínez (2011) distinguish CCTs from a third type of program that includes a
CCT component but goes further by aiming to connect beneficiaries to a wide range of social
services and assistance programs. Chile Solidario constitutes the best of example of this kind of
program. For a detailed analysis, see Larrañaga, Contreras and Ruiz-Tagle (2012).
78
. However, these programs are not static and there is evidence of convergence in recent years
(Cecchini and Martínez 2011; 90 Yanes 2013).
79
. Self-reported incomes are cross-checked with a federal database. However, the database does
not include informal-sector workers, who make up the bulk of Brazil’s poor. This targeting method
has been criticized for giving beneficiaries an incentive to under-report incomes.
80
. De la O (Forthcoming) creates a cross-national measure of CCT design covering the degree to
which a program’s operational rules that suppress administrative discretion.
81
. CCTs are enshrined in law in only three countries in the region – Brazil, Chile and Uruguay
(Cecchini and Madaraiga 2011, 161-65).
82
. However, as will be discussed later, stipends have not been adjusted since the pilot program was
launched.
Social Policy and Presidential Ideology in Latin America
241

83
. There is a limit of five beneficiaries per family. There is a separate allowance for families with
seven or more children.
84
. This describes those who neither study nor work (ni estudia, ni trabaja).
85
. For analyses of the causes of the crisis, see Blustein (2005); Mussa (2002).
86
. Plan Familias paid a monthly stipend of 155 pesos a month plus 30 pesos per child up to 305
pesos a month.
87
. Garay (2010, 88-90) argues that popular protests prompted the launch AUH. None of the sources
interviewed or the secondary sources consulted emphasized the role of popular mobilization. In
contrast to Jefes (Golbert 2004), civil society groups had no say in AUH’s design or implementation
(Kantor 2012).
88
. Two types of proposals stemmed from academia: those that, like FRENAPO, proposed
universalizing family allowances (Rofman, Grushka and Chebez 2001) and those advocating for a
basic citizens income (Lo Vuolo et al. 1999).
89
. AUH beneficiaries cannot receive other cash transfers. The number of beneficiaries declined
sharply once the more generous AUH launched.
90
. Ezquiel Lo Valvo is Manager of Planning at ANSES.
91
. To obtain the conditional 20 percent, children six and under must certify that they are vaccinated
and be registered with SUMAR, which provides preventative care for mothers and children.
Children 5-18 must demonstrated school attendance. A major problem with withholding the 20
percent is that it loses much of its value as a result of Argentina’s high inflation (Lo Vuolo 2013c).
92
. Given the questionable nature of Argentina’s poverty data (ODSA-UCA 2014), when possible
I rely on data generated by independent researchers.
Social Policy and Presidential Ideology in Latin America
242

93
. The size of the beneficiary population changes from month to month depending on employment
trends. In theory, people who obtain formal-sector jobs are automatically transferred to contributory
family allowances and vice-versa.
94
. Leonardo Gasparini is Director of Center for Distributive, Labor and Social Studies (CEDLAS)
at Universidad Nacional de La Plata in La Plata.
95
. Rafael Rofman is Lead Social Protection Specialist at the World Bank’s Buenos Aires office.
96
. Ianina Tuñón is Coordinator of the Infancy Barometer at the Universidad Católica Argentina’s
Observatorio de la Deuda Social.
97
. Juana Azurduy de Padilla was a female pro-Independence fighter in Bolivia’s War of
Independence.
98
. Declaring it unsustainable, right-wing president Hugo Banzer (1997-2001) initially cancelled
the program before reinstating it six month later as the less generous Bolivida. Sánchez de Lozada
restored it to its previous name and value upon starting his second term.
99
. Morales reduced the age requirement to 60 years.
100
. Voters in those departments overwhelmingly voted for autonomy, but the results were declared
unconstitutional by the National Electoral Court.
101
. Laserna’s solution is to transfer natural resource rents directly to Bolivians to spend as they see
fit. This, he predicts, would reduce rent-seeking.
102
. The absence of civil society from BJP’s design is surprising given the historically close links
between Morales and social movements (see Fuentes 2014).
103
. Erick Meave is economist at the Bolivian Planning Ministry specializing in social policy.
104
. Ernesto Yáñez is a former vice-president of Bolivia’s Central Bank who has written extensively
on social policy in Bolivia.
Social Policy and Presidential Ideology in Latin America
243

105
. Orlando Murillo is an education economist who has worked for the Bolivian Education Ministry
and was the coordinator of the Observatorio Social de Políticas Educativas de Bolivia (OSPE-B),
an education think tank.
106
. Verónica Paz Araujo is a consultant for the United Nations Research Institute for Social
Development and a Bolivian representative for Commitment to Equity, a joint initiative between
Tulane University and the Inter-American Dialogue.
107
. Werner Hernani-Limarino is the lead researcher at Fundación ARU, a Bolivian think tank.
108
. The main problem with this survey is that children who altogether left the school system were
not interviewed.
109
. Diego Víquez served as executive president of IMAS when Avancemos launched and was part
of the team that designed the program.
110
. Oscar Arias was president of Costa Rica during 1986-1990 and 2006-2010.
111
. José Antonio Li replaced Víquez as IMAS president.
112
. Fernando Marín replaced Li as IMAS president.
113
. FODESAF was the region’s first social policy fund (Román 2012). Starting with Bolivia in the
1980s (Graham 1992; Newman, Jorgensen and Pradhan 1991), most of the region’s countries
adopted these programs to ameliorate the social costs associated with structural adjustment
(Goodman et al. 1997; Siri 2000).
114
. Manuel Barahona is an economics professor at Universidad Nacional and part of the team that
designed Avancemos.
115
. Pablo Sauma is an economics professor at Universidad de Costa Rica and part of the team that
designed Avancemos.
Social Policy and Presidential Ideology in Latin America
244

116
. Juan Diego Trejos is an economics professor at Universidad de Costa Rica and part of the team
that designed Avancemos.
117
. Carla Valverde was part of the Avancemos technical team during 2008-2009.
118
. Zumbado’s political aspirations were dashed when he was forced to resign in mid-2008 for
allegedly misusing a donation from the government of Taiwan.
119
. Poverty is referenced in two contexts: as a requirement for benefiting from the program and a
consequence of failing to complete high school.
120
. As part of this broadening of coverage, even students not deemed “vulnerable” may benefit
from the program if deemed appropriate by officials following an interview or home visit.
121
. In contrast to Progresa/Oportunidades, stipends are not differentiated based on sex. Women
have lower rates of desertion than men in Costa Rica.
122
. Leonardo Garnier was Minister of Education during 2006-2014.
123
. Nor is it necessarily true that seventh graders drop out to enter the workforce (Garnier
interview). Additionally, lack of interest and difficulty together constitute a more important
explanation for desertion than cost (Mata Hidalgo and Hernández Romero 2013, 61-62).
124
. Juan Manuel Cordero was vice-minister of Labor and Social Security during the Chinchilla
administration.
125
. María José Morales was part of the Avancemos technical team during 2006-2008.
126
. Olga Sonia Vargas was coordinator of Avancemos at IMAS until 2012.
127
. Rosibel Herrera is the current coordinator of Avancemos at IMAS.
128
. See Chapter 1, section 4.2.1 for a detailed analysis.
129
. The key difference between the two is that Oportunidades covers urban households, whereas
Progresa was limited exclusively to rural households.
Social Policy and Presidential Ideology in Latin America
245

130
. The usefulness of an AUH survey question is unclear given the well-documented problems
with Argentina’s data.
131
. Roxana Maurizio is a Researcher in Economics at the Science Institute of Universidad Nacional
General Sarmiento.
132
. Mónica Parrado is Director of Family Allowances and Unemployment Benefits for Buenos
Aires at ANSES.
133
. Grigoli and Sbrana (2012) find that, although still less likely to enroll, children living extreme
poverty enrolled in BJP were more likely to attend.
134
. During its first year, the 200-Boliviano stipend was paid in two 100-Boliviano stipends.

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Reconstituting the fiscal bargain: the politics of taxation in Latin America's emerging markets

PDF

Economic policy-making in a changing environment: the politics of success in Peru vis-à-vis China

PDF

Politics behind appointments: policy, patronage, and party systems in the Andes

PDF

Power to the masses? The rise of direct democracy in Latin America

PDF

Habits and policy: the socio-cognitive foundations of foreign policy stability

PDF

Budget institutions and the positive theory of fiscal policy

PDF

Flinging the boomerang: locating instability and the threat potential of identity-bias in US national security policy

PDF

The Catholic church in Latin America: an evaluation of the institutional and political impacts of progressive church reforms

Asset Metadata

Creator Borges-Herrero, Fabián A. (author)

Core Title Social policy and presidential ideology in Latin America: the political economy of social spending and anti-poverty programs

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Politics and International Relations

Publication Date 11/19/2014

Defense Date 10/22/2014

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

Tag CCT,CCTs,conditional cash transfers,Latin America,Latin America left turn,Latin American politics,left turn,mixed methods,OAI-PMH Harvest,poverty policy,presidential ideology,social policy,social spending,welfare states

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Munck, Gerardo L. (committee chair), Pastor, Manuel, Jr. (committee member), Sellers, Jefferey M. (committee member), Wise, Carol (committee member)

Creator Email borges.fabian@gmail.com,borgeshe@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-519000

Unique identifier UC11298518

Identifier etd-BorgesHerr-3096.pdf (filename),usctheses-c3-519000 (legacy record id)

Legacy Identifier etd-BorgesHerr-3096.pdf

Dmrecord 519000

Document Type Dissertation

Format application/pdf (imt)

Rights Borges-Herrero, Fabián A.; Borges-Herrero, Fabian A.

Type texts

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

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA