Explaining Social Spending Variation and National Income Inequality in Latin American Nations, 1980-2008 K

Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2012 Universal Human Rights as New Political Power Resources: Explaining Social Spending Variation and National Income Inequality in Latin American Nations, 1980-2008 K. Russell (Kaiser Russell) Shekha

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected]

THE FLORIDA STATE UNIVERSITY COLLEGE OF SOCIAL SCIENCES AND PUBLIC POLICY

UNIVERSAL HUMAN RIGHTS AS NEW POLITICAL POWER RESOURCES: EXPLAINING SOCIAL SPENDING VARIATION AND NATIONAL INCOME INEQUALITY IN 18 LATIN AMERICAN NATIONS, 1980-2008

KAISER RUSSELL SHEKHA

A Dissertation submitted to the Department of Sociology in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded: Summer Semester, 2012

Kaiser Russell Shekha defended this dissertation on June 28th, 2012

The members of the supervisory committee were:

Jill Quadagno Professor Directing Dissertation

William H. Moore University Representative

Daniel Tope Committee Member

Deana Rohlinger Committee Member

The Graduate School has verified and approved the above-named committee members, and certifies that the dissertation has been approved in accordance with university requirements.

ACKNOWLEDGEMENTS

To the Florida State University Department of Sociology I offer my heartfelt gratitude and thanks. Your encouragement and motivation helped me grow into a better person in my academic and personal life. I really want to thank you all for making my family and I feel at home in Tallahassee these last six years. To my dissertation committee, wow what a journey! To Deana Rohlinger, thank you for your ever ready guidance and suggestions and your efforts to help me grow as an academic. To Daniel Tope, your mentorship and guidance have been invaluable. Your savvy understanding of the political world influenced me beyond belief. I would also like to thank Brian Starks, whom sat on my committee and contributed a wealth of guidance. To William H. Moore and John Mayo, thank you so much. To Doug Schrock and Marc Dixon, my first graduate mentors! To Jill Quadagno, I have often wondered what my life, academic and otherwise, would be like without such a phenomenal mentor and guide. I am honored to be your student! I have so many dear friends that watched me as I took my first steps on this ride, joined me along the way, and sang me off to a better tomorrow. To Wes Salinas, Johnny Hardy, Fabian Luna, Elena Guerra, Jimmy Burgan, Chris and R.J. Cobb, Patrick and Dawn, Casey and Shyla , Matt Poston and so many more. To Jason Laguna, Sammy Rastagh, Sarah Conn, Carmen Von Rohr, Jenny Rothenberg, Brandon McKelvey, Orit and Khen Fischer, LJ Roman, and Sean Tabor. To Edward Kring, you taught me so much about teaching!!! To Patrick McGrady and Joellen Pedersen. Thank you for always believing in me and keeping me and my family in your minds and hearts. Your friendship means the world!! To my family, where would I be without my family? From my father Siddiq and Amijan Samina living afar, to my mother Sharon and my dad Bob, I love you so much! To the support of my sisters Sky and Mary Eleanor and their mother Barbara. To my brother, Sean Giles, for always saying what I needed to hear! To Daniel, Brenda, and Nick for being there, always! Most importantly, I need to thank my own little family. Damien Cienega, my son, I did it all to impress you and your love made it possible. Vikky Lea, my best friend, my wife, my partner. I could spend the rest of every day of our lives thanking you and I will! To my little Kaiya Rylea, I’ll never tire of chasing the deer on the hill with you at our new home at Denison University!

iii

TABLE OF CONTENTS

List of Tables vi

Abstract vii

Chapter 1: Universal Human Rights 1 International Human Rights Treaties 2 Determinants of Treaty Ratification 3 National Human Rights Institutions 6 Determinants of NHRI Adoption 8 Guatemala 10 Argentina 12 Research Questions 14

Chapter 2: Political Theories 15 Logic of Industrialism 15 Social Spending Levels 15 National Income Inequality 17 The National Economy 17 Demographics and Society 21 The Economy and Globalization 27 Social Spending Levels 27 National Income Inequality 30 National Power Resources 34 Social Spending and National Income Inequality in Developing Nations 34 Power Resources, Developing Nations, & Social Spending 37 Power Resources, Developing Nations, & Income Inequality 38 Human Rights as a Power Resource 40 Social Spending Levels and National Income Inequality 40 Social Effects of Treaty Ratification 41 Social Effects of NHRI Adoption 43 International Civil Society 45

Chapter 3: Data and Methods 47 Research Design 47 Case Selection 47 Data and Measures 48 Dependent Variables 48 Social Spending Levels 48 The GINI Coefficient 49 Independent Variables 50 Logic of Industrialism 50 Economics/Globalization 51 Power Resources 52

Power Resources and Human Rights 53 Methodological Controls 55 Analytic Plan 56 Analysis of Social Spending Levels 56 Analysis of Income Inequality 58

Chapter 4: Explaining Variation in Latin American Social Spending Levels 65 Latin American Social Spending Systems 65 Social Security and Welfare Spending Determinants 68 Health Spending Determinants 75 Education Spending Determinants 79

Chapter 5: Explaining National Income Inequality in Latin America 83 National Income Inequality in Latin America 84 Determinants of National Income Inequality 85

Chapter 6: Discussion 90 Explaining Social Spending Variation 90 Explaining National Income Inequality 93 Future Research Directions 96

Appendix A: Tables, Boxes, and Graphs 98

Appendix B: Political Coding Report 118

References Cited 121

Biographical Sketch 139

LIST OF TABLES

Table 1: Adoption of NHRIs in Latin America 98

Table 2: Description of Variables 99

Table 3: Summary Statistics after 100 Imputations 101

Table 4: Determinants of Social Security and Welfare Spending in Latin America 102

Table 5: Determinants of Health Spending Levels in Latin America, 1980-2008 104

Table 6: Determinants of Education Spending in Latin America, 1980-2008 106

Table 7: Number of Repressive Regimes by Year 108

Table 8: Determinants of National Income Inequality in 18 Latin American Countries 109

ABSTRACT

Universal human rights include guarantees to adequate income and fair wages, sufficient healthcare, and investments in educations and are extended to all regardless of their social and citizenship status. Universal human rights are centered internationally at the United Nations and institutionalized at the national level through international human rights treaty ratifications and National Human Rights Institution adoptions. In this dissertation I ask if extending power resources theory to universal human rights lead to higher social spending levels and lower levels find that the logic of industrialism, which focuses on the effects of economic growth and associated demographic transitions, globalization theory, which emphasizes the effects of international trade, foreign investment, privatization, and fiscal austerity, and power resources theory, which links national political regimes and political balance of power to spending and inequality, continue to have significant effects on social spending levels and national income inequality in contemporary Latin America. However, the real story is the effects of universal human rights on social spending levels and national income inequality. I find that treaty ratifications improve health and education spending levels while they are associated with declining social security and welfare spending. Treaty ratifications are also associated with lower levels of national income inequality. National Human Rights Institution adoptions are associated with increases in all forms of social spending levels, but have no effect on national income inequality. Overall, universal human rights have mostly positive effects on social spending and reductions in income inequality, and provide positive proof of the continuing extension of social rights to income security, adequate healthcare, and quality educations.

vii

CHAPTER 1: UNIVERSAL HUMAN RIGHTS

After the widespread destruction of WWII, international human rights norms increased in importance with the United Nations creation in 1945. The United Nations’ mission reflected goals of global peace and security for all member nations. The UN also wished to improve the human rights situation for individuals around the world, irrespective of their national background or social status. While state repression, torture, and imprisonment ranked highest among past human rights atrocities, UN members also recognized the importance of individual rights to personal economic, social, cultural self-determination without fear of repression or discrimination by state agencies (Donnelly 1998). The UN Economic and Social Council created the UN Commission on Human Rights (UNCHR) in 1946 which in turn drafted the 1948 United Nations Declaration for Human Rights. The document establishes individual universal rights regardless of nationality, sex, age, or gender (United Nations 1948). Articles 22, 25, and 26 specifically mandate the rights to social security, social assistance, health, well-being, and free and compulsory primary educations (UN 1948). The Declaration embodies the post-WWII push for international human rights, signals commitment on a national level, but lacks legal enforcement. There are two categories of international human rights: Civil and Political Rights (CPRs) and Economic, Social and Cultural Rights (ESCRs). In this study I focus on the effects of international human rights as a power resource on the extension of social rights through spending on old age income, healthcare, and education and reductions in national income inequality. However, I also include human rights indicators that reflect CPRs such as International Covenant on Civil and Political Rights (ICCPR) and The Convention on Torture (CAT), which are international human rights treaties that were created to protect CPRs. In early years Western nations prioritized CPRs while the Soviet Union and the Eastern countries emphasized the primary importance of ESCRs.

Nations serious about embracing human rights institutionalize ESCRs by ratifying human rights treaties and adopting NHRIs. Treaties and NHRIs are vehicles for universal human rights sponsored by the UN but implemented at the national level. Thus universal human rights embodied by treaties and NHRIs represent an extension of what were previously only the rights of citizens in full democracies to all peoples within a nation’s border regardless of their status

and background. In nations with recent turns to democracy or restricted democracies the extension of rights from select groups of citizens to all people is an important factor in ensuring the protection of the most vulnerable including the elderly, women, children, the sick, and the poor. International human rights treaties and NHRIs help to protect, promote and ensure the recognition of the human rights of a country's people.

Countries committing to universal human rights ratify international human rights treaties that protect specific rights or groups of people and adopt NHRIs to monitor rights compliance. They may even go a step further by ratifying optional protocols to the treaties which allow UN bodies to hold national governments directly accountable for their human rights abuses. For example, the 1966 International Covenant for Economic, Social and Cultural Rights (ICESCR) and the International Covenant for Civil and Political Rights (ICCPR) protect the specific rights named in the titles (United Nations 1966). The UN also created Optional Protocols (OPs) that creates a committee at the UN where individuals can directly file their grievances without the aid of an Ombudsman or outside organization. NHRIs are quasi-state institutions promoted by the UN and include the classical Ombudsman, modern Human Rights Commission, and hybrid Human Rights Ombudsman. These institutions may protect the nation’s citizens against public administration, governmental, and private sector violence, abuse, fraud, and discrimination.

International Human Rights Treaties Efforts to cement universal human rights in international law resulted in international human rights treaties meant to lay the legal and social foundations for extending universal human rights around the world. The 1966 International Covenant for Civil and Political Rights (ICCPR) and the International Covenant for Economic, Social, and Cultural Rights (ICESCR) act as the basis for other important human rights treaties. The ICESCR again contains specific language protecting income, health, and education, which are the focus of this study. Other important treaties included in this study are Convention on the Elimination of all forms of Racial Discrimination (CERD), Convention on the Elimination of all forms of Discrimination Against Women (CEDAW), the Covenant on the Rights of the Child (CRC), the International Convention on the Protection of the Rights of All Migrant Workers and Members of Their Families (CMW), and the Agreement establishing the Fund for the Development of the Indigenous Peoples of Latin America and the Caribbean (INDIG). Countries signal commitment to treaty terms through signing, ratification, accession, or succession. Signing signals an interest

to future ratification of the treaty: ratifying the treaty makes it legally binding. Similarly, if a state wishes to comply, but does not wish to sign, it may accede, which is also legally binding (United Nations 2001). Finally, succession is when a state changes regime type or becomes a new state and inherits the responsibilities of treaty ratified or acceded by the previous state (Kamminga 1996). Unfortunately, international human rights treaties lack the mechanisms needed for proper promotion and protection of rights. Adding Optional Protocols (OPs) for some treaties resulted in committees that allow individuals to file their grievances directly with the United Nations. Countries that sign the Optional Protocols willingly submit to the authority of the UN, and thus show a stronger commitment to human rights than do countries that do not sign. The most prominent examples include the ICCPR Optional Protocol (OP) (1966) and the CEDAW OP (1999). The ICESCR OP became available in 2008 with only three nations ratifying as of April, 2011 (treaties.un.org). Both of the committees created for the OPs allow individuals to file directly their grievances against the state. The sample of nations in this study ratified the bulk of human rights treaties and Optional Protocols in the 1980s and 1990s. The high rates of treaty ratification suggest the region’s commitment to international human rights norms, at least on the surface. Every single nation in the sample ratified the ICESCR, ICCPR, and ICCPR OP. These two treaties and Optional Protocol are the foundation of all other treaties and protocols, and further signal the integration of international human rights norms and laws with national political and governmental processes in Latin America. Treaty ratification is a popular tool for Latin American government representatives, citizens, and civil society organizations. Latin America as a region represents fertile testing ground for the extension of social rights through social spending because of its deep integration with international human rights norms through high rates of treaty ratification. Determinants of Treaty Ratification Much of the research on human rights treaties focuses on why states ratify and test multiple theories in one model or study (Cole 2005, 2009; Goodliffe and Hawkins 2006; Hafner-Burton et al. 2008; Hathaway 2007; Powell and Staton 2009; Wotipka and Tsutsui 2008; Vreeland 2008). Realist or rationalist theory stresses the perceived costs of ratification to a state, and enjoys wide empirical support in studies of treaty ratification and effects. Democracies should be more likely to ratify compared to authoritarian regimes, because dictators have more to lose under the

scrutiny of the international eye. While some find that democracies are more likely than autocracies to ratify (Cole 2005), others find that autocracies are also likely to ratify human rights treaties because of weak enforcement mechanisms (Hafner-Burton et al. 2008; Powell and Staton 2009). A variant of the rationalist perspective, liberal theorists ask why any state would willingly give up any legal national sovereignty to an international organization or system (Moravcsik 2000). The “lock-in” thesis supposes that any state would willingly reduce their national sovereignty to reduce the threat of future political instability and human rights abuses. This should be especially true for new democratic regimes, as they wish to set up a legal structure to protect the new political system. Despite some support (Cole 2005; Moravcsik 2000), others studies find little to no evidence to support the thesis (Cole 2009; Goodliffe and Hawkins 2006; Hathaway 2007; Wotipka and Tsutsui 2008). Regardless, results of rationalist studies of treaty ratification suggest the importance of including indicators that measure the strength of democracy, instances of autocratic rule, and regime change. Research in the area provides evidence of their utility for analyzing Latin American social spending (Avelino et al. 2005; Brown and Hunter 1999, 2004; Kaufman and Segura-Ubiergo 2001; Huber et al. 2008a). Another version of rationalist theory focuses on domestic political and judicial institutions within any one regime, and their impacts on treaty ratification. These studies hypothesize that domestic legal and political factors better predict treaty ratification and effects. Open dictatorships with more competition are more likely to ratify (Vreeland 2008), while democratic states with strong legal systems are less likely to ratify (Powell and Stanton 2009). Another study indirectly measures domestic legal enforcement by creating an interaction term between democracy and human rights records, finding that democratic states with poor human rights records are more likely to ratify, but that less democratic states with better records are not (Hathaway 2007). This study also provides support for rationalist arguments of regime change and stability.

Sociological institutionalism is at the other end of the spectrum from rationalist arguments. These theories hypothesize a process of institutional isomorphism through coercion, imitation, or normative pressure (Dimaggio and Powel 1983). World society theory draws on this approach focusing on the mimetic and normative pathways to isomorphism. It uses indicators of civil society connections, measured by memberships in IGOs, INGOs, human rights treaties, transnational advocacy networks (TANs), regional and international conferences, and regional

diffusion rates to explain treaty ratifications and effects (Cole 2005; Hathaway 2007; Powell and Stanton 2009; Wotipka and Tsutsui 2008). Organizational memberships enjoy strong support predicting lower terror rates and lower likelihoods of ratifying treaties because they apply pressure to states to conform to human rights norms, which leads to less repression but also a less likelihood of states ratifying treaties. States that do not wish to comply with human rights norms avoid ratifying treaties to evade the pressure of civil society groups. Regional diffusion rates of organization membership and treaty ratification have a positive association with further ratification across studies because of the pressure for neighboring states to adopt the same practices and policies. The negative effect on treaty ratification is counterintuitive to studies predicting social spending levels. Yet the results above suggest that civil society groups, such as INGOs and transnational advocacy networks (TANs), use human rights law to appeal to national governments to improve conditions for their citizens, whether that is scaling back repression or providing better resources and benefits.

The above studies of treaty ratification point to important themes to consider when analyzing the impact of human rights treaties on social spending. Rationalist theory, which underlines the perceived costs of ratification, finds varying support. While some find that democracies are more likely than autocracies to ratify (Cole 2005), others find that autocracies are also likely to ratify human rights treaties because of weak enforcement mechanisms (Hafner- Burton et al. 2008; Powell and Staton 2009). Regardless, these results suggest the importance of including indicators that measure both the strength of democracy and instances of autocratic rule. Another variant of the rationalist perspective, liberal theory suggests that newly democratized countries seek to ratify treaties to prove legitimacy and “lock-in” human rights policies to prevent future abuses. While the theory has only limited support, it may be worthwhile to include indicators of regime change in studies of social spending. World polity theory uses civil society connections, measured by memberships in IGOs and INGOs and regional diffusion rates. Organizational memberships enjoy strong support predicting lower terror rates and lower likelihoods of ratifying treaties because they apply pressure to states to conform to human rights norms, which leads to less repression but also a less likelihood of states ratifying treaties. States that do not wish to comply with human rights norms avoid ratifying treaties to evade the pressure of civil society groups. Regional diffusion rates have a positive association with ratification across studies because of the pressure for neighboring states to adopt the same practices and

policies. While the negative effect on treaty ratification may be counterintuitive to studies predicting social spending. Regardless, the results above suggest that civil society groups, or transnational advocacy networks (TANs), use human rights law to appeal to national governments to improve conditions for their citizens, whether that is scaling back repression or providing better resources and benefits. Studies of repression and ratification (Goodliffe and Hawkins 2007; Hafner-Burton et al. 2008; Powell and Staton 2009; Vreeland 2008), yield contradictory results for the impact of repression on ratification, but still make a good case for including repression indicators in studies analyzing other outcomes like social spending. Indeed, many of Latin America’s most developed social security systems date back to the 1920s (McGreevey 1990), meaning that autocratic rulers either set up or managed these systems for long periods of time. Related to repression, studies measuring domestic factors like the strength of law, political party representation, judicial independence, and civil liberties (Goodliffe and Hawkins 2007; Hafner-Burton et al. 2008; Hathaway 2007; Powell and Staton 2009; Vreeland 2008) find that local factors do impact ratification rates. Together, this body of literature suggests the use of important factors such as the strength of democracy, other domestic legal and political constraints, the influence of international politics, and international civil society when analyzing social spending patterns in Latin America.

National Human Rights Institutions As human rights norms and law spreads, the need for domestic institutions to handle growing numbers of complaints filed against states and private firms increases. NHRI allow states to control their own domestic human rights agenda by providing its citizens with specific protection of their individual rights, and creates links to the United Nations and international human rights system. NHRIs protect and promote human rights around the world. Set up by national governments, supported by the United Nations, they have a broad mandate of promotion, protection, and education of human rights. They interact and collaborate with domestic civil society, state administrators, and IGOs, INGOs, and TANs, occupying a unique space between global governance institutions, domestic and international civil society, and the state (Smith 2006). The United Nations is the single largest supporter of NHRIs, and is largely responsible for their international development. As early as 1946, the UN National Economic and Social Council recommended that national governments create institutions to work with the United

Nations Commission on Human Rights (UNCHR). In the late 1960s the UN revisited NHRIs, especially within the context of new human rights agreements like the ICCPR and ICESCR. The international community felt the need for national institutions to help countries implement their new commitments to human rights made legally binding by ratifying treaties (Lindsnaes and Lindholt 2000). While efforts to establish NHRIs around the world increased in the 1980s, their roles, functions, and power remained unclear until the early 1990s. The UN and national governments seeking to create strong NHRIs joined in the first International Workshop on National Institutions for the Promotion and Protection of Human Rights in Paris, October 1991. There participants drafted a set of guidelines and recommendations for building effective NHRIs. Known as the Paris Principles, the recommendations focused on the responsibilities, composition, independence and pluralism of the institution, methods of operation, and the quasi- judicial nature of NHRIs (Lindsnaes and Lindholt 2000). Complementing these advances in defining NHRIs, the 1993 Vienna World Conference of Human Rights further elaborated the role of the UN in promoting domestic institutions and providing them with resources (Cardenas 2003; Lindsnaes and Lindholt 2000; Reif 2000). The oldest form is the classical ombudsman, first set up in Sweden in 1809 and spreading through Scandinavia throughout the 20th century (Cheng 1968; Reif 2000). Classical ombudsman did not focus on human rights explicitly. Their purpose was to oversee public administration and ensure the rule of law for the citizens of its country (Cheng 1968; Reif 2000). The second category of NHRIs is the modern Human Rights Commission (HRC) (Hadden 2002; Reif 2000). The HRC has an explicit human rights focus, both in promotion and protection and education of the populace. The commission, depending on the country, may also address discrimination claims while protecting other CPRs and ESCRs. In addition, HRCs may provide policy advice to governments. The Human Rights Ombudsman (HRO) is a hybrid institution, incorporating elements of the classical ombudsman model and the modern HRC. A particular favorite for Latin American countries, these institutions focus both on human rights and the rule of law in public administration and the private sector (Reif 2000; Uggla 2004). Studies point to the weak rule of law throughout Latin America in post-conflict and transitional democracies, and the importance of the HRO in addressing the deficiency (Dodson and Jackson 2004; Hannum

2006; Popkin and Roht-Arriaza 1995; Uggla 2004). The HRO is by far the most popular form in Latin America. NHRIs vary in their scope and mandate from nation to nation. The institutions might monitor and report human rights abuses by the state, provide resolutions to correct abuses, intervene in legal processes, educate citizens of their rights, and propose policy changes that reflect respect for human rights norms and laws. However, limited autonomy from the legislative and executive branches of the government can hamper their ability to achieve the full scope of their mandates in some countries (Reif 2000). Historically, Latin American NHRIs address violations of civil and political rights (CPRs), though in recent years some shifted their focus to economic, social, and cultural rights (ESCRs) (Uggla 2004). Determinants of NHRI Adoption Only one study uses quantitative, comparative analysis to explore the determinants of NHRI formation using a global sample (Koo and Ramirez 2009). The study uses event history analysis to model the likelihood of adopting classical ombudsman offices, the modern human rights commission, and the human rights ombudsman. The analysis combines human rights commissions and ombudsmen, comparing them against the classical ombudsman institution. The results for the analysis of classical ombudsman offices suggests that world polity institutionalism best explains their adoption. None of the neorealist indicators have a statistically significant association with ombudsman adoption, except for national economic development (GDP/capita), which has a negative association. Participation in international organizations, conferences, and agreements has a positive association with the likelihood of adopting a classical ombudsman office. As regional and global adoption increase, so too does the likelihood of any one state adopting an ombudsman office. The results for analysis of adoption of human rights commission and ombudsmen find different results. The strength of democracy and national human rights practices positively predicts the likelihood of NHRI adoption and support the neorealist perspective. Yet world polity institutionalism again better predicts adoption of the modern NHRIs. Participation in international organizations, conferences, and agreements leads to a greater likelihood of modern NHRI adoption, as do global and regional density rates. These analyses suggest that regardless of their form NHRIs both influence and are influenced by the world society.

In Latin America nations began adopting the institution by the late 1980s (see Table 1). Some NHRIs started because of pressure from global governance institutions during democratic transitions, while others began due to pressure from domestic and international civil society. For example, Guatemala began the region’s first NHRI in 1987, the Procurador de los Derechos Humanos (PDH) (Uggla 2004). However, UN sponsored peace agreements in 1994 and 1996 between the Guatemalan government and Guatemalan National Revolutionary Unit (URNG) led to a strengthening of the PDH, especially as a truth commission for post-civil war conflict (Hannum 2006; Popkin and Roht-Arriaza 1995; Reif 2000). Argentina, however, created its HRO in 1993 (the Defensor del Pueblo) after a Presidential decree followed by legislation by Congress, and gave it Constitutional status in 1994 (Reif 2000; Uggla 2004). Other NHRIs in the region started for similar reasons with Colombia’s beginning because of the indirect efforts of the OHCHR (Hannum 2006), Peru’s because of domestic efforts to show democratic processes to international observers during Fujimori’s reign (Pegram 2008), and El Salvador’s initially created to conduct truth commissions in the post civil war transition to peace (Dodson and Jackson 2004). I provide a discussion of NHRIs in Guatemala and Argentina with greater detail below. Despite some differences in the origins of Latin American NHRIs, their structures are similar. Most HROs in the region serve for 5 years, with reelection prospects varying by country. Legislatures appoint or elect the Ombudsman, with some country’s executive leaders intervening in the process. Legislative bodies in most countries keep the right to fire the Ombudsman for negligence or not fulfilling duties. All offices report annually to congress and some also report to regional and international HRCs. The Latin American HRO oversees a staff that assists in their duties. Finances come from annual budgets provided by the legislature, and therefore can fluctuate wildly as a means to control or punish the office. However, international donors contributed as much as 50% of the total budget to Latin American nations in 2004, suggesting that autonomy from the legislative branch of the government also varies between nations depending on funding sources (Uggla 2004). Regardless of the similarities, enough variation exists to assume that each NHRI in Latin America affects different outcomes. The HRO in Latin America has a broad mandate and can initiate several actions and handle various claims. Many states limit NHRIs to annual reports to congress, investigations of claims made by individuals, educational programs, judicial process monitoring, and public

denouncement of human rights abuses. However, some may initiate investigations on their own when they become aware of potential abuses such as in El Salvador. In Argentina the HRO may initiate investigations of private firms, but not against the judicial, legislative, and military branches of the government. However, the Argentinean HRO can launch court actions to address constitutional rights of the individual (Reif 2000). Only three nations in this sample lack an NHRI at the time of this analysis: Chile, the Dominican Republic, and Uruguay. The prevalence of NHRIs in the region and the popularity of the HRO suggest that these institutions are deeply embedded in the national political framework. Whether implemented as a top-down solution to civil war and state violence from the United Nations as in Guatemala, or due to popular support combined with governmental backing as in Argentina, NHRIs seek to extend civil, political, and social rights to all peoples within its mandate, irrespective of their national origin. NHRI impact on the rights of peoples in the region, however, has yet to be fully realized as evidenced by their limited successes discussed below. Guatemala Guatemala was the first Latin American country to start an Ombudsman office in 1987, the Procurador de los Derechos Humanos (PDH) (Reif 2000; Uggla 2004). Congress elects the position once every 5 years, with no possibility of reelection. While Congress controls the budget, 15% of its total budget comes from foreign sources (Uggla 2004). It suffers from a low budget, only $4 million in 2001 (Dodson and Jackson 2004). The PDH may investigate claims of human rights abuses of individuals by public administration officials, recommend actions to address claims, and make annual reports to Congress (Reif 2000; Uggla 2004).

The PDH’s early popularity was because of its most famous director, Ramiro de Leon Carpio. He aggressively intervened in human rights matters, and Congress selected him to be President after the 1993 failed self-coup of President Jorge Serrano. Ironically, after spending 4 years bolstering the institution, he created an executive office to handle human rights monitoring, the Presidential Commission for Human Rights (COPREDEH), which often competed directly with the PDH (Dodson and Jackson 2004).

UN sponsored peace agreements in 1994 and 1996 between the Guatemalan government and Guatemalan National Revolutionary Unit (URNG) led to a strengthening of the PDH, especially in its function as a truth commission for post-civil war conflict (Popkin and Roht-

Arriaza 1995; Reif 2000). The 1994 agreement, or the Comprehensive Agreement on Human Rights, expanded the PDH’s scope to investigation of forced participation in volunteer civil defense committees during the civil war (Reif 2000). Despite this, PDH director Jorge Garcia la Guardia viewed UN presence as another competitor (Dodson and Jackson 2004). Under the 1996 Peace Accords, the UN assumed responsibility for overseeing human rights compliance, though it tried to work closely with the PDH (Dodson and Jackson 2004; Uggla 2004). The office was to resume compliance monitoring in 2000, as the United Nations Verification Mission in Guatemala was to end. However, the UN remains an influential player in Guatemalan human rights monitoring. The UN Office of the High Commissioner of Human Rights (OHCHR) began working in Guatemala in 1994 and negotiated an agreement with Guatemala in 2003 that allowed it to place staff in Guatemalan prosecutor’s offices and initiate judicial authority in drug trafficking and human rights (Hannum 2006).

The PDH enjoys significant autonomy from the judicial and legislative branches of the government (Reif 2000; Uggla 2004). However, wide variation in regional Ombudsman reports of compliance with recommendations makes it difficult to ascertain the institution’s efficacy. Adding to that, reports of physical abuse on Ombudsman staff from police forces and the office’s investigation of high-powered political elites like the former Vice President, Reyes Lopez, suggests the institution works on shaky political ground (Uggla 2004). Indeed the institution was often at odds with the COPREDEH and UN mission, and suffered under the corrupt leadership of Dr. Julio Arango Escobar (Dodson and Jackson 2004).

Despite the variation in state compliance with the office and political hostility, claims remain high, and the office fielded 18,000 claims in 2000. Most claims made in 2000 focus on the violations of civil and political rights (Uggla 2004). This suggests the Ombudsman office gives less priority to violations of ESCRs than it does CPRs, which is not unusual in a transitional democracy still suffering with problems of state repression and judicial corruption.

Overall, the Guatemalan Ombudsman had some success in fulfilling its mandate of protecting human rights. It has worked closely with the UN since the 1990s, benefitted from strong leadership in its early years, targeted high-power public officials, and provided an opportunity for new human rights NGOs to participate (Reif 2000). Despite this, problems such as extrajudicial killings and kidnappings, state torture and illegal imprisonment, police and

prison mistreatment, and discrimination still plagues Guatemala (Reif 2000). Also, falling popularity because of the poor leadership of director, Dr. Escobar, contributed to its difficulties in fulfilling its mandate (Dodson and Jackson 2004). Fortunately, in 2002 the legislature elected the ex-director of the Human Rights Institute at the University of San Carlos, Sergio Fernando Morales Alvarado, as the new director. He brought a measure of hope to the office, though his prognosis of a low budget combined with the planned 2003 withdrawal of the UN team meant he faced great challenges. Regardless, the PDH remains a vibrant force in human rights monitoring and compliance, and is especially important given the continued violence seen in Guatemala.

Argentina Argentina created its human rights ombudsman in 1993 (the Defensor del Pueblo) after a Presidential decree followed by legislation by Congress, and gave it Constitutional status in 1994 (Reif 2000; Uggla 2004). Congress appoints the office once every five years from three candidates chosen by a parliamentary commission (Uggla 2004). The budget comes directly from the parliament’s own funds (Uggla 2004). The Ombudsman’s power comes from the 1994 Constitution, which raised nine human rights treaties, two declarations of human rights, and past human rights guarantees from the old Constitution to constitutional status (Reif 2000). This gives the office a broad mandate to include civil and political rights, economic social and cultural rights, and environmental and consumer rights. It allows investigation of public administration officials and any public or private entity that provides public services. While this may seem like a broader mandate than Guatemala’s Ombudsman, the Defensor cannot investigate the legislative, judicial, or military bodies of the government, severely limiting its capacity to oversee public administration officials (Reif 2000). Similar to other institutions in the region, its resolutions are merely suggestions and are not legally binding (Reif 2000; Uggla 2004). In contrast to other offices around the region, however, the Defensor can launch amparo actions, or court proceedings to protect the rights of an individual, but cannot enforce any actions beyond that (Reif 2000).

Like its Guatemalan counterpart, the Defensor enjoys autonomy from the government because of high public popularity, a well funded budget, strong leadership, and strong Constitutional backing (Reif 2000). However, political manipulation is a problem for the office.

The legislature holds the right to dismiss the Ombudsman with a two-thirds vote, especially if found in negligence of duties (Uggla 2004).

Claims made to the Defensor, while always high, increased rapidly during the 1990s. In 1995 the office fielded 7,256 claims which increased to 30,434 by 1998 (Defensor del Pueblo de la Nacion 1998). Since the institution’s investigatory powers restrict it from investigating the legislature, judiciary, or military, most claims focus on economic administration, which include private companies providing public services like water and national financial institutions (Reif 2000). The other majority of claims target employment discrimination and social security (Reif 2000). For example, in 1996, 31.6% of claims focused on economic administration and 43% of claims concerned employment discrimination and social security (Defensor del Pueblo de la Nacion 1998). In 1998, about 18% of the claims concerned health and social actions, the environment, culture, and education administration (Defensor del Pueblo de la Nacion 1998). The focus of economic, social, and cultural claims throughout the 1990s suggests that protection of ESCRs is a primary concern of the Defensor in Argentina, especially when compared to the focus on CPRs throughout most of the region.

Due to Argentina’s developed democratic institutions, wealthier status than Guatemala and many other Latin American countries, and well regarded and funded Ombudsman office, the institution is strong. Coupled with a relatively weak and corrupt judiciary (Reif 2000), the Defensor has managed to carve out a spot in the national legal structure as a champion for human rights, and continues to field high numbers of claims.

The discussion of human rights treaties and NHRIs above shows the commitment the international human rights communities have in extending rights previously enjoyed only by citizens to all peoples irrespective of their national origin or social status. Early analysts of national rights often focus solely discussion on citizens (Marshall 1964), though some extend the concept of national citizenship because of the influence of universal human rights (Shafir and Brysk 2006). Yet the United Nations and the human rights community seek to transcend the requirement of citizenship and replace it with simply being human. Latin America as a region embraces this mission and shows its commitment to international human rights norms by ratifying nearly all international human rights treaties, forming NHRIs in 14 of 17 nations in this

sample, and giving constitutional and legal status to these instruments and institutions through the domestic political process.

Research Questions In this dissertation I ask two major questions: Does extending power resources theory to universal human rights lead to higher social spending levels? Are these new universal human right political power resources associated with lower levels of national income inequality? I also ask if welfare state theories focusing on the effects of industrialization, the national and globalized economy, and traditional national political power resources affect social spending levels and national income inequality. I ask this using a sample of 18 Latin American nations from 1980-2008. I use advanced statistical techniques including non-parametric pooled OLS models that address problems of heteroskedasticity, autocorrelation, and temporally correlated panels and multiple imputation techniques available in STATA 11. Chapter 2 focuses on theories used to analyze social spending and income inequality and my own extension of power resources theory to universal human rights. In Chapter 3 I describe the data and methods used in both of these analyses. Chapter 4 examines the results of my analysis on variations in social security and welfare, health, and education spending levels in Latin America. Chapter 5 discusses the results of my analysis on national income inequality in Latin America. In Chapter 6 I provide my overall conclusions to my research questions, some limitations to my studies, and future directions I will take with this body of research focusing on the continuing expansion of social rights through the extension of universal human rights in Latin America and beyond.

CHAPTER 2: POLITICAL THEORIES

In this chapter I discuss the political, economic, and social theories that explain variation Latin American social spending levels and national income inequality. I draw on multiple perspectives including the logic of industrialism, economic/globalization theories, power resources theory, and my own extensions of power resources to universal human rights. Each section discusses the theory’s impact on social spending levels and national income inequality in separate sub- sections. LOGIC OF INSDUSTRIALISM The logic of industrialism perspective emphasizes the impacts of economic growth and behavior and demographic transitions on social spending levels and national income inequality in developed and developing regions. In the following sub-sections I define the logic of industrialism, its applications to analyses of social spending, its impacts on national income inequality. I discuss this by building on research that uses both developed and developing nation samples. Social Spending Levels The logic of industrialism was the first theory applied discussing welfare state development (Kerr et al. 1960; Pampel and Williamson 1989; Wilensky 1975). Its basic premise assumes that as states industrialize their economies, there will be greater resources and needs for social services, naturally leading to more spending in these areas (Wilensky 1975). Caused by industrialization’s negative effects, namely dissolution of strong family ties and support networks as workers become more and more reliant on wage labor, the need for more social welfare is great. However, industrialization also increases life expectancy, which leads to an increase in the aged and youth populations and means that larger segments of at-risk populations are in need of assistance (Pampel and Williamson 1989). Couple new demographics with a reliance on wage labor versus family support, and growing numbers of aged workers and young people in need of social welfare arises. Critiques of the approach suggest that it cannot adequately address variations in social spending between developed and democratized nations (Cameron 1978; Stephens 1979). However, studies that build on this approach still find support for logic of industrialism indicators.

Though the logic of industrialism is lacking in its explanation of welfare state development, several empirical studies have shown support for some of its key variables. Issues of urbanization and percent workforce in manufacturing are less important in empirical, quantitative studies. However, the percent of population over 65 and GDP growth are strong correlates to welfare development (Aaron 1967; Cutright 1967; Jackman 1975; Wilensky 1975).

The studies above were done in the context of advanced capitalist countries of the global North and West. While these countries do not follow the same trajectories of welfare development due to the timing of industrialization, it can be argued that they all followed similar paths of economic development. In other words, these are a fairly homogenous group of states when thinking about economic development.

The field is lacking in comparative analyses looking at countries, like those in Latin America and the Caribbean, which experienced industrialization at different times and in varying capacities. With the late onset of industrialization in many Latin American countries, one may expect to see strong effects of classic industrialization variables like economic development (GDP/capita) and percent aged (over 65). The few quantitative studies on Latin American and other areas with less developed countries (LDCs) show mixed results depending on the outcome variables used to measure welfare spending effort (Huber et al 2008a; Rudra and Haggard 2005; Segura-Ubiergo 2007). Additionally, these studies focus on the effects of politics on social spending levels, but also include logic of industrialism variables. For example one study shows in an international sample of 57 developing nations that economic growth and youth populations decrease education spending in non-democracies but have no effect in democracies. Conversely, the aged population increases health spending in democracies while economic growth decreases this spending in non-democracies (Rudra and Haggard 2005). Others find no consistent association between social spending levels and economic development or aged and youth populations (Kaufman and Segura-Ubiergo 2001; Segura-Ubiergo 2007). Finally, my model study (Huber et al. 2008a) finds that economic development and youth populations increase health and education spending levels while only aged populations increase social security and welfare.

Regardless of the inconsistent results for logic of industrialism variables in studies of social spending in developing countries, this study includes measures of economic development,

aged, youth, and urban populations. I hypothesize a positive correlation between development (GDP/capita) and health and education spending but not social security and welfare spending because of the resiliency of pension and welfare spending to economic shocks and decline. Aged populations (% 65 or over) will positively influence social security and welfare and health spending levels, while youth populations (% under 14) will increase education spending levels.

National Income Inequality The National Economy Many studies use economic theory and models to predict and explain income inequality in Latin America and other samples. Research on the effects of economic development on national income inequality began as early as the 1950s with Kuznets’ (1955) widely cited study on the inverted u-shape income inequality takes over time as a country grows its economy. Kuznets’ (1955) study of income inequality in the United States and England showed that once economic development began in a particular industry or area of the country, a domino-effect of more growth occurred. This sequential growth initially increased income inequality but eventually led to decreases because the urban labor force expanded and changed the income distribution between higher and lower earners to favor the latter. The Kuznets Curve, popularized by his study, is the inverted U-shaped curve that defines the relationship between GDP and income inequality. A large body of research links Kuznets style development with income inequality (Ahluwalia, 1976; Anand and Kanbur 1993; Bollen and Jackman 1985; Bruno, Ravallion and Squire 1996; Clarke 1995; Crenshaw 1992; Deininger and Squire 1996; Fields 1994; Huber et al. 2006; IADB 1998; Morley 2001; Muller 1985, 1988, 1989; Nielsen 1994; Nielsen and Alderson 1995; Ravallion and Chen 1997; Simpson 1990). Others look beyond simple economic growth and find that rising inflation leads to lower real wages especially for low income earners and thus increases income inequality (Morley 2001; Huber et al. 2006). Some studies push the link between economic development and income inequality further and examine the effects of shifts in the agricultural sector on national income inequality (Alderson and Nielsen 1999; Hubet et al. 2006). Finally, many studies emphasize fiscal policies such as taxation and social spending as determinants of income inequality (Barros et al. 2010; Gasparini and Cruces 2010; Huber et al. 2006; IADB 1999; Morley 2001; Robinson 2010). In this section I discuss studies that analyze the effects of economic development, inflation, and the agricultural economy on income inequality using international and Latin American samples.

Studies of economic development and income inequality do not necessarily agree to the effects a rising GDP per capita will have on income inequality, especially in Latin America. One study notes that if the rate of economic growth in Latin America matched that of industrialized, developed nations then Latin American nations would have much lower income inequality (IADB 1998). In a sample of OECD countries one study (Alderson and Nielsen1999) finds that economic development first decreased and then increased income inequality as the developed nations in the sample reached high levels of development. Morley (2001) shows using the Kuznets Curve that while Latin American nations experienced declining income inequality because of economic growth, this is no longer true in the 1990s. The author suggests that 1990s growth is caused by the increase in skill-intensive production and the increase in educated workers. However, Huber et al. (2006) follow up the above studies and further test the effects of economic development using a sample of Latin American and Caribbean nations, characterized as mostly middle-income countries, and find that increased economic development is associated with moderate and statistically significant reductions in income inequality. Despite these disagreements, I include a measure of economic development (GDP per capita). I hypothesize that as economic growth increases in Latin America, income inequality will go down. However, I believe that the effects will not be large just as Morley (2001) suggests. Economic growth, or rising GDP per capita, is not the only national economic determinant of income inequality in Latin America. Morley (2001) suggests that rising inflation, or the costs of consumer goods and services, will increase income inequality in Latin America. The author contends that rapid, unstable national economic growth coupled with inflation will lead to higher income inequality. He goes on to suggest that inflation increases inequality especially in times of recession, which Latin America experienced during the 1980s. Nations that attempted to control inflation during times of recession by not raising minimum wages also exacerbated national income inequality. Other studies also contend that rising inflation leads to rising income inequality (IADB 1998; de Ferranti et al. 2004). Morley (2001) provides examples of Chile and Argentina to prove that inflation increases income inequality. Chile and Argentina faced recessions in the late-1970s and early 1980s. Morley suggests that the recessions, stagnant minimum wages, and inflation led to the devaluation of real wages and thus increased income inequality in these countries. Huber et al. (2006) use pooled time-series data in Latin America and the Caribbean to statistically test Morley’s (2001) assertions. They find inflation has a

statistically significant but very small effect on income inequality, supporting previous research that suggests high inflation is associated with high income inequality. I include a measure of inflation used by Huber et al (2006) in my study and hypothesize that as inflation rises so will income inequality in Latin America. Extending studies based on the Kuznets Curve, Alderson and Nielsen (1999) postulate that shifts away from an agricultural based economy will decrease national income inequality. The authors suggest this because of the assumed lower income inequality within the rural and agricultural sectors. The authors show that sector dualism, or the absolute difference between the percent of the labor force in agriculture and the percent of GDP from agriculture, can impact national income inequality. They suggest that as developed nations move away from an agricultural economy then labor force income inequality will rise because of the assumption that income inequality within rural areas is lower. Indeed, their results show that agricultural sector dualism leads to higher levels of income inequality using a sample of developed, industrialized nations. Another study tests sector dualism using a sample of Latin American and Caribbean nations but suggests that having a larger agricultural sector increases income inequality because GINI estimates show that income inequality within rural areas is generally higher than within urban areas in Latin America. The study shows that sector dualism in Latin America and the Caribbean from 1970-2000 leads to statistically significant and moderate increases in income inequality. While the studies disagree on how sector dualism increases income inequality, both show that agricultural economic production and labor impacts income inequality. I hypothesize that a dual sector, or large percentage of agricultural workers and production in Latin America, will be associated with higher levels of income inequality. I believe this because of the high inequality within rural areas that dominate agricultural production and labor in Latin American nations. A number of studies examine the impacts of fiscal policy and behavior on income inequality in Latin America (Berry 1998; Barros et al. 2010; de Ferranti et al. 2004; Gasparini and Cruces 2010; Huber et al. 2006; IADB 1998). While many emphasize the impacts of taxation as a means to reduce income inequality, I do not include a measure of tax collection. More importantly, I include measures of social spending. Social spending on social security, welfare, health, and education are generated through taxation and are intended to be progressive

forms of redistributive spending. That is, social spending is meant redistribute incomes more heavily to poor and low-income peoples. In Latin America social spending is generally progressive. One study of Latin American social spending suggests that health and education spending are especially progressive (IADB 1998). Health and education spending in Latin America is progressive because it’s often targeted towards the poor as ways to increase their economic standing by providing adequate healthcare and quality educations that enable low income workers to miss less work due to individual and family health problems and land better jobs because of their educated status (IADB 1998). However, others suggest that social security spending in Latin America is regressive in nature, and tends to redistribute income disproportionately to formal sector workers that already enjoy higher incomes (de Ferranti et al. 2004). Additionally, social security spending in Latin America is regressive because pension schemes are often different based on the occupation. For example, civil servants and the military may enjoy better pension packages compared to other formal sector workers (Huber et al. 2006). Despite growth in social spending across categories, social security and welfare spending most often dominate total social spending for Latin American nations. The dominance of social security spending, considered to be regressive, suggests that regressive redistributive spending in Latin America outweighs progressive redistribution through health and education spending. Studies of social spending and income inequality in Latin America support the assertions above. One points to the stable levels of social spending in Costa Rica during times of economic crisis such as the 1980s as proof that social spending helps to alleviate or stave off income inequality (Berry 1998). Another suggests that economic reforms that improved progressive taxation and social spending increased cash-transfers to the poor through targeted programs to alleviate poverty and income inequality and help account for Argentina’s decline in income inequality in the 2000s (Gasparini and Cruces 2010). Similarly, Brazil experienced declining income inequality in the 2000s in part because of generous government programs that targeted the poor and non-urban, underdeveloped areas (Carvalho et al. 2010). A comparative study using pooled time-series data (Huber et al. 2006) finds that social security and welfare spending as a percentage of GDP increases income inequality in Latin America and the Caribbean except when an interaction term with the historical strength of democracy is included. The findings show that social security and welfare spending alone lead to higher income inequality unless a strong

democracy is present, a point I return to below in my discussion of national power resources and income inequality. Based on the theories of social spending and income inequality and the results discussed above I hypothesize that social security and welfare spending levels, largely regressive, will be associated with increases in Latin American income inequality. Conversely, because of the success of health and education spending programs, largely progressive, I hypothesize health and education spending as a percentage of GDP will be associated with lower levels of income inequality. Demographics and Society Many studies also examine the effects of demographic transitions spurred by industrialization and economic development on income inequality (Alderson and Nielsen 1999; Barros et al. 2010; Behrman et al. 2009; Bitrán et al. 2005; Cogneau and Gignoux 2009; de Ferranti et al. 2004; Esquivel et al. 2010; Fay and Laderchi 2005; Fay and Wellenstein 2005; Frankema 2009; Huber et al. 2006; Jaramillo and Savvedra 2010; Kahat 2010; Laderchi 2005; Leite 2009; Lopez-Calva and Lustig 2010; Morley 2001; Woolcock 2005). Demographic transitions such as migration to urban centers and increased youth populations caused by lower infant and childhood mortality rates are important predictors of educational and occupational opportunity, health conditions, poverty, and income inequality (Alderson and Nielsen 1999; Huber et al. 2006). Increased urbanization is a contentious subject because of the varied effects it can have on poverty and inequality rates (Bitrán et al. 2005; Fay and Laderchi 2005; Fay and Wellenstein 2005; Fay et al. 2005; Laderchi 2005; Woolcock 2005). Most would agree that increased youth populations, with their limited educational attainment and occupational experience, equates to higher rates of poverty and income inequality (Alderson and Nielsen 1999; Huber et al. 2006). Additionally, social demographics such as ethnic diversity and educational attainment also influence poverty and income inequality. While increased ethnic diversity and indigenous populations in Latin America lead to worse income inequality ratios (de Ferranti et al. 2004; Huber et al. 2006; Leite 2009; Morley 2001), making primary and secondary educations available to poor rural and urban youths as well as minority populations may mitigate the associated increases urbanization, youth populations, and ethnic diversity have with income inequality in Latin America (Barros et al. 2010; Behrman et al. 2009; Cogneau and Gignoux 2009; de Ferranti et al. 2004; Esquivel et al. 2010; Frankema 2009; Huber et al. 2006; Jaramillo and Savvedra 2010; Kahat 2010; Lopez-Calva and Lustig 2010). This section reviews

quantitative and case-studies on the effects of urbanization, increased youth populations, ethnic diversity, and schooling on national income inequality in Latin America and international samples. Increasing urbanization in Latin America is associated with higher income, less poverty, and more social services when compared to rural areas. However, roughly 75% of the total population lives in urban areas and absolute numbers of poor people remain very high with 60% of the poor living in cities. The myth of higher rural inequality is dispelled by Gini coefficients that show 6 nations in Latin America and the Caribbean have worse urban income inequality when compared to rural inequality, while three have the same inequality between urban and rural areas (Fay and Laderchi 2005). There are many reasons why increased urbanization in Latin America is associated with higher rates of poverty and income inequality. Urban segregation leads to many low income families living and moving to urban peripheries where the lack of infrastructure, social services, and decent wages increase urban poverty and inequality (Fay and Laderchi 2005). Another reason is the inability of the urban poor to participate in the labor market in the same fashion as non-poor. The urban poor in Latin America are more likely to participate in the informal economy, have less education and job skills, more unemployed household heads, less access to transportation to areas with better paying jobs, and face racial, ethnic, and gender discrimination (Laderchi 2005). Housing and living conditions for the urban poor are often informal and lead to worse health and well-being outcomes that can affect their ability to work and generate income (Fay and Wellenstein 2005). Adding to the burden of inadequate housing, public health challenges in urban areas can also affect a household’s ability to work and generate enough income. The rise of communicable disease rates, less access to and quality of public health services, and to infrastructure services such as clean water sources and electricity are characteristics of the urban poor that lead to greater poverty and inequality (Bitrán et al. 2005). While social networks such as friends and relatives can help alleviate urban poverty by providing much needed economic assistance, increased regional economic integration and the lack of basic infrastructure and social services undermines the efficacy of urban networks. The urban poor find their networks to be too small and poorly resourced to respond favorably to the challenges of living and working in urban markets (Woolcock 2005). Finally, the urban poor often do not benefit from social insurance schemes because of their low participation in the formal market

(Fay et al. 2005). Urban poverty and inequality remain one of the biggest social problems Latin America faces because of the challenges discussed above the urban poor confront. While income inequality is higher in rural areas compared to urban ones, the literature above shows that an inequality of incomes, services, infrastructures, and social networks within urban areas is pervasive and so I hypothesize that increased urbanization is associated with higher income inequality in Latin America. Studies also suggest that growing youth populations lead to higher income inequality in Latin America (Alderson and Nielsen 1999; Huber et al. 2006). Alderson and Nielsen (1999) suggest that growing populations of young people mean more workers with less skills and leads to a greater income gap between older, skilled workers and young, unskilled laborers. The authors prove this in their quantitative analysis of income inequality using a sample of 108 nations from 1947-1996. However, a quantitative analysis of income inequality in 18 Latin America and the Caribbean countries from 1970-2000 shows opposite results from the effects of growing youth populations. In models without political indicators of democracy and left-right legislative power balance, youth populations are associated with lower levels of income inequality and become nonsignificant once politics are introduced (Huber et al. 2006). My data, however, show that youth populations in Latin American nations from 1980-2008 are declining. Of the 18 nations in the sample, 12 of them experienced at least a 10% decline in youth populations from 1980-2008. None of the nations experienced a growth of youth populations. Despite the contradictions in findings and the declining population of young people in Latin America, I include a measure of youth populations as a percentage of the total populations. I hypothesize no association between youth populations as a percentage of the total population and income inequality in Latin America. Universal human rights treaties and institutions focus on the conditions of ethnic minorities and indigenous groups and make strides to increase the political, economic, and social participation of these at-risk groups. The treaty to establish a fund for the indigenous peoples of Latin America and the Caribbean and the treaty on migrant workers include special language protecting racial and ethnic minorities and indigenous peoples. Some studies point to the unequal educational, employment, and living conditions ethnic and indigenous groups in Latin America face as relevant determinants of national income inequality (de Ferranti et al. 2004; Huber et al.

2006; Leite 2009; Morley 2001). To that end, I include a measure of ethnic/indigenous diversity in this study to assess its impacts on national income inequality in Latin America. Ethnic diversity marks the social landscape in Latin America. While some countries such as Argentina have relatively low percentages of indigenous populations, nations such as Mexico and Bolivia are home to many indigenous groups, in contrast to countries such as Brazil and Colombia also include large African and East Indies descended populations. Nearly all of the nations in this sample have sizable Mestizo, or mixed populations. A quantitative case study of Brazil and another using assessing micro-data available for the whole of Latin America point to the severe inequalities in opportunities indigenous and ethnic minority populations face (de Ferranti et al. 2004; Leite 2009). For example, one study (de Ferranti et al. 2004) shows that indigenous and African descended populations have lower levels of education, have less wealth, and higher within group income inequality. This inequality in education, wealth, and income within groups suggests that indigenous or ethnic status severely hampers individuals and households attainment of adequate incomes and they face greater levels of poverty compared to their white and non-indigenous counterparts. Leite (2009) examines racial discrimination in Brazil and its impacts on wage inequality through inequalities in educational opportunities. The study finds that large gaps between private and public school access and quality are major determinants of income inequality in Brazil. The author suggests that public policies targeting the further expansion of public schooling at the primary and secondary levels would help racial groups to close the pre-market educational inequality gap and lead to less wage inequality. Only one quantitative, comparative study also assesses the impact of indigenous and ethnic minority populations on income inequality in Latin America (Huber et al. 2006). It finds that nations with very high or very low ethnic heterogeneity are associated with much higher national income inequality measured by the GINI. Clearly the research above finds that ethnic, indigenous, and racial groups face an inequality of opportunities that begins with education, leads to occupational inequalities, and therefore worsens national income inequality for the region. This is especially true in nations with very large or very small ethnic and indigenous populations. I expect a positive association between ethnic diversity and rising income inequality. The next paragraphs discuss educational attainment as mitigating factor for income inequality caused by demographic transitions and ethnic diversity and discrimination.

Access to and attainment of education is a cornerstone of universal human rights as detailed in the UN Declaration of Human Rights and linked treaties such as the ICESCR and the Convention on Children. It may also help mitigate the effects of income inequality because of urban migration and high youth, ethnic and indigenous populations. Many studies of Latin American countries find that higher rates of secondary schooling attainment lead to reductions in national income inequality (Barros et al. 2010; de Ferranti et al. 2004; Esquivel et al. 2010; Frankema 2009; Jaramillo and Savvedra 2010; Kahat 2010; Lopez-Calva and Lustig 2010). One quantitative analysis of the determinants of income inequality in Latin America finds that higher percentages of primary schooling completion reduce inequality (Morley 2001). However, some question the ability of increased access and attainment of primary and secondary educations to reduce national income inequality in case and comparative studies (Behrman et al. 2009; Cogneau and Gignoux 2009; Huber et al. 2006). Regardless of the disagreements, this study includes a measure of secondary school completion because of education’s centrality in universal human rights efforts and the positive results of some of the studies above. Educational attainment has increased in the majority of Latin American nations in the last few decades. For example, Mexico, Brazil, and Peru expanded basic educations during the 1990s and 2000s, implementing educational policies that led to higher enrollment and attainment rates, and thus higher incomes, for youth in poor and rural areas (Barros et al. 2010; Esquivel et al. 2010; Jaramillo and Savvedra 201; Lopez-Calva and Lustig 2010). Mexico’s Progresa/Oportunidades social program targets rural and urban households living in poverty and has improved enrollments in primary and secondary educations (Esquivel et al. 2010). This is especially true among rural areas (Behrman et al. 2009; Esquivel et al. 2010). However, some ask if the popular anti-poverty program does much to relieve income inequality. One study finds that while rural enrollments increased under Progresa, achievement scores did not (Behrman et al. 2009). These results suggest that youth participating in the program may still be at a disadvantage when trying to compete in the market or for coveted university and college enrollment spots. In Peru, increased access to education resulted in income inequality declines, though this effect was greater in rural areas. The study attributes this to nearly universal primary schooling and a jump from 5% of Peruvians in 1940 to 80% in the late 2000s attending secondary school (Jaramillo and Saavedra 2010). One study of Brazil shows that declines in income inequality during the 2000s can be attributed to increased access to education in the

1990s compared to the 1980s (Barros et al. 2010). Another study of income inequality in Brazil suggests that while pre-World War II generation’s increased access to secondary and higher education increased income inequality, post-war generations experienced greater access to primary educations, especially amongst agricultural families, and reduced income inequality during the mid-1980s and 1990s. However, the study questions whether education programs targeting poor children and leading to more years of schooling will continue to reduce income inequality after considering new efforts to improve the quality of educations around the country (Cogneau and Gignoux 2009). Only two studies use a cross-national, quantitative approach to analyzing income inequality in Latin America. One finds that increased completion of primary schooling reduces income inequality while increased attainment of higher education increases it (Morley 2001). The other, my model study, finds no association between secondary schooling and income inequality in Latin America and the Caribbean (Huber et al. 2006). These two studies, while comparable, also use different income inequality data sources because of updates to GINI estimates, and measure educational access and attainment differently. Despite some differences in findings the literature above suggests that demographic transitions such as increased urbanization and youth populations lead to higher income inequality in the region. The same can be said for increased ethnic or indigenous diversity. However, studies suggest increased primary and secondary schooling reduces overall national income inequality. This downward effect on income inequality from primary and secondary schooling comes from the increased higher educational opportunities and market opportunities and income associated with more schooling. I propose too, that increased educational attainment helps to curb exacerbated inequalities caused by large national urban and youth populations and high ethnic or indigenous diversity. I hypothesize that increased urban populations and ethnic heterogeneity will increase national income inequality while increased youth populations will have no effect. I expect that higher percentages of secondary school completion is associated with lower levels of national income inequality in these 18 Latin America nations from 1980- 2008. The sections above review the literature on the effects of industrialization and its associated economic development and demographic shifts on social spending levels and national income inequality in developed and developing regions of the world. While the effects of

economic growth vary by social spending type, it is associated with lower income inequality. Inflation and a dual sector agricultural based economy are associated with higher levels of income inequality. Demographic transitions such as the growth of youth and aged populations also show inconsistent results in studies of social spending in Latin America, though higher urban and youth populations are associated with higher income inequality throughout the region. Similarly, a very high or very low ethnic minority or indigenous population is associated with worse income inequality. Finally, the growth of educational attainment at the secondary level is mostly associated with reductions in income inequality. The next sections turn to the effects of globalization or a globalized economy on social spending levels and national income inequality. THE ECONOMY AND GLOBALIZATION This section reviews the literature on the effects of globalization on social spending levels and national income inequality and makes predictions about their effects in Latin America. Economic globalization affects social spending levels through trade, foreign investment, budgetary deficits, and pension privatization while analyses of income inequality emphasize trade, investment, and pension privatization. The first sub-section reviews the literature on social spending levels while the next reviews literature discussing income inequality. Each section contains hypotheses about the effects of globalization on social spending levels and national income inequality in my sample of 18 Latin American nations from 1980-2008. Social Spending Levels Growing out of functionalist accounts that emphasized industrialization and changing economies, welfare state research also examines globalization’s effects on social spending levels. This line of research gained popularity in the late 1970’s and continues today in studies of developed and developing nations. While some contend that modern economic globalization has no effect on the welfare state (Cameron 1978; Garret 1998; Rodrik 1997; Stephens 1979), others (Huber and Stephens 2001) show that it did cause a period of retrenchment in spending efforts, though this did not translate to less money spent (Swank 1998). Neoliberal globalization, made popular by the United States and Great Britain and adopted by financial institutions such as the World Bank and IMF, focuses on liberalizing markets. Privatization of state enterprises, deregulation of industries, freeing capital flow internationally and fiscal austerity are pivotal to this economic approach (Prasad 2006). Some contend that competition in liberalized markets

pressure nations to reduce social budgets to remain viable, enter trade agreements, and attract investment (Babb 2005; Stiglitz 2002).

In the Western world, many studies reject claims that modern economic globalization substantially reduces welfare state efforts (Cameron 1978; Garret 1998; Rodrik 1997; Stephens 1979). However, another study (Swank 1998) shows changing social programs that economic openness interacts with political structures making states react differently to the pressure showing that liberal states such as the United States should succumb to economic pressures by reducing spending when compared with the corporatist and social democratic states of continental Europe and Scandinavia. However, in reality, there are no historical reductions in social spending in the United States and democratized Europe. Even countries in Scandinavia experiment with welfare state retrenchment. In the 1980s they responded to the challenges of a deregulated and globalized market by embracing deregulation and fiscal austerity in the form of reduced social spending levels compared to increased revenues. While this retrenchment did not equate to real reductions in welfare spending, it did mean the government was spending at a slower pace than in the 1970s (Huber and Stephens 2001).

Though neoliberal globalization did not initially shake the welfare state in Northern Europe substantially (Kitschelt et al. 1999), for LDCs neoliberal policies have resulted in a decline of the social contract of welfare between states and citizens (Babb 2005; Laurell 2000). Indeed, others have shown empirically that globalization, in the form of trade, debt, and investment, has differing effects on types of welfare spending (Kaufman and Segura-Ubiergo 2001; Rudra and Haggard 2005; Segura-Ubiergo 2007). One study of LDCs shows a negative correlation between trade and debt and education spending, but exert no effect on public pension or health spending, while capital mobility has a weak, but positive correlation to public pension spending (Rudra and Haggard 2005). One analysis of Latin American social spending shows a small association between capital mobility and health and education spending, and a positive correlation between trade openness and health and social security spending levels (Kaufman and Segura-Ubiergo 2001). Another study of fourteen Latin American countries shows a negative correlation between trade and public pensions, but has no relation to health and education expenditures combined (Segura-Ubiergo 2007). I include measures of international trade and foreign direct investment to test whether contemporary liberalization of markets negatively affects social spending levels in Latin America.

Some studies extend this approach by considering the impact of tax revenues, national debt and budgetary deficits on social spending in the Latin American context, suggesting the forces of globalization impact a nation’s ability to generate revenues, manage national debt, and operate with budgetary surpluses (Huber et al. 2008a; Segura-Ubiergo 2007). One study of fourteen Latin American countries from 1973-2003 shows convincingly that issues of fiscal constraint, including the capacity to generate revenue from taxes and fiscal adjustment needs, impact welfare spending. Tax revenues have a positive effect on public pension, health, and education expenditures. On the other hand, fiscal adjustments such as increased national debt and budgetary deficits have a negative effect on public pension, health, and education spending (Segura-Ubiergo 2007). Another finds that running budgetary deficits decreases health and education spending but has no effect on social security and welfare levels. It also finds that the debt crisis of the 1980s also decreased health and education spending (Huber et al. 2008a). Brown and Hunter (1999) show in a comparative study of Latin American social spending that debt service does not influence spending levels. Unfortunately, combining health and education expenditures makes it difficult to compare these results with future studies that disaggregate spending. Regardless, when analyzing welfare expenditure effort in Latin America, these studies provide compelling reasons to include at least some measures of fiscal constraint. For the sake of parsimony, and due to the lack of findings regarding debt service (Brown and Hunter 2004), I include only a measure of budgetary deficits and historical period indicators to control for the 1980s debt crisis in Latin America and the 1990s rise of neoliberalism.

International governance organizations also push neoliberal reforms in Latin America. Structural adjustment programs (SAPs), imposed by the International Monetary Fund (IMF) can push down on social spending (Babb 2005), though one study shows that having IMF obligations is associated with higher health and education spending (Huber et al. 2008a). For example, social security privatization began in Chile in 1991 but was popularized by a larger World Bank initiative which targeted Latin American nations (World Bank 1994). The World Bank’s (1994) plan to privatize public pensions is specific only to social security spending, but also represents another attempt at liberalizing economies to make them more competitive. Privatization of social security and pension systems is meant to grow national economies and reduce social spending (Gill et al. 2005). However, some show that social security spending went up in the short term for some Latin American nations as they reconciled old accounts and invested in the transition to

private retirement accounts (Gill et al. 2005). International governance institutions, while political in nature, often implement economic policies in less developed countries such as those found in this study.

For the most part, I hypothesize the forces of globalization will decrease social spending levels in Latin America. Trade and foreign investment will decrease social security and welfare spending to stay internationally competitive and attractive to investors and trading partners, but increase health and education spending levels because of the necessary investment in training a workforce and maintaining their healthcare in the face of a competitive international economic environment. Average social spending levels do not decline in this sample, and so I expect no effect from budgetary deficits. Pension privatization is only included in the analysis of social security and welfare spending and will have a negative effect on its spending levels. The next sub-section discusses the impacts of globalization on national income inequality.

National Income Inequality Many studies examine the effects of globalization on income inequality, particularly the effects of structural adjustment programs promoted by international institutions such as the World Bank, International Monetary Fund, and World Trade Organization (Babb 2005). These studies highlight global trade, foreign investment, and more recently, pension privatization as economic structural adjustments that swept the region in the 1990s. Studies predict greater exposure to international trade and foreign investment will increase income inequality in Latin American nations, though some studies find that increased investments decrease inequality (Alarcon and McKinley 1998; Behrman et al. 2001; Berry and Tenjo G. 1998; Calderon and Chong 2001; Cornia 2004; Gasparini and Cruces 2010; Harrison 2002; Larrea 1998; Morley 2001; Urani 1998; Savvides 1998). One of the goals of pension privatization was to reduce income inequality by expanding coverage of social security to protect more of the population (Arenas de Mesa and Mesa-Lago 2006; Bertranou 2001; Casarico and Devillanova 2008; Dion 2006, 2008; Gill, Packard, and Yerno 2005; James, Edwards, and Wong 2008; Schmaehl 2007; World Bank 1994). Studies find that exposure to international trade leads to higher levels of income inequality in Latin America (Alarcon and McKinley 1998; Alderson and Nielsen1998; Behrman et al. 2001; Berry and Tenjo G. 1998; Calderon and Chong 2001; Cornia 2004; Harrison 2002; Huber et al. 2006; Jensen and Rosas 2007; Larrea 1998; Morley 2001; Savvides 1998; Urani

1998). A study of the impacts of trade on income inequality in Mexico during the mid-1980s and 1990s finds that increased exposure to an export based economy built on international trade leads to an increase in wage inequality amongst most workers. The downward pressure applied to the wages of unionized, border state, and export manufacturing employees is a result of the increase in export-oriented economic production (Alarcon and McKinley 1998). Another suggests that because Latin American exporting is limited to a few primary goods and does not expand export manufacturing beyond textiles that countries in the region will experience downward pressure on wages and increased income inequality because they cannot compete with regions that export a better variety of manufactured goods. The study concludes that Ecuador experienced worse trade terms and low export diversification during the 1980s to 1990s leading to higher rates of income inequality because of the associated redistribution of income from labor to business (Larrea 1998). Still another study notes that Colombia’s increasing income inequality during the late- 1980s and early-1990s coincided with a number of structural reform packages that included trade liberalization (Berry and Tenjo G. 1998). In Brazil, structural adjustments during the 1900s resulted in a large trade surplus but also increased inflation and lower real wages which increased income inequality (Urani 1998). Overall, structural economic reform, of which trade liberalization is a key component, is associated with higher income inequality in almost all Latin American nations with the exception of Costa Rica. Costa Rica, already starting with lower income inequality compared to most Latin American nations, managed to introduce some economic restructuring without widening the income gap. For eight of the eighteen nations in my sample, structural adjustment is associated with a 5-10 point increase in the GINI coefficient, a startling jump. Cross-national, comparative studies of the effects of trade liberalization on income inequality in Latin America and international samples yield contradictory results. One (Calderon and Chong 2001) uses an international sample of developed and developing nations and finds that while trade liberalization overall decreases income inequality, export manufacturing of mostly primary goods is associated with higher income inequality. Two studies (Harrison 2002; Savvides 1998) show that liberalized trade reduces the share of income for labor and is associated with higher income inequality for less developed nations. Morley (2001) and Cornia (2004) employ quantitative techniques to show that increased trade leads to higher income inequality in Latin America while Behrman et al. (2001) show that trade liberalization actually reduces income inequality though the finding is not statistically significant. Considering

the individual experiences of nations in Latin America and the results from cross-national studies on trade liberalization and income inequality, I include a measure of net trade, or total imports minus total exports as a percentage of GDP, and hypothesize that increased trade will be associated with increased income inequality in Latin America. Foreign direct investment (FDI) is mostly associated with worsening income inequality in developed and developing regions such as Latin America (Alderson and Nielsen1998; Gasparini and Cruces 2010; Hanson and Feenstra 1997; Huber et al. 2006; Jensen and Rosas 2007). In developed regions outflows of direct investment depress wages for low-skill workers because of the outsourcing of their work and this is proved in a quantitative study of income inequality amongst developed nations (Alderson and Nielsen1998). In developing regions foreign investment is theorized to increase income inequality because the investments are typically targeted towards capital-intensive production instead of labor-intensive production such as manufacturing. The emphasis on capital-intensive production creates only a few jobs that are much higher paying than labor-intensive production and thus increases income inequality (Huber et al. 2006). Morley (2001) suggests that high rates of foreign investment during the 1990s in Latin America increased overall economic volatility and increased local currency values. He theorizes that greater exposure to foreign investment should decrease income inequality because it should increase the demand for labor. Case and cross-national studies examining the impact of foreign investment on income inequality in developing areas generally agree that FDI is bad for income inequality. One study of Mexican income inequality links FDI to wider wage gaps between low and high skilled workers (Hanson and Feenstra 1997), though another finds higher FDI decreases income inequality from 1990-2000 (Jensen and Rosas 2007). In Argentina FDI as a percentage of GDP increased 44 percent between the 1980s and 1990s and increased the nation’s technological innovation and concentrated incomes at the top of the distribution (Gasparini and Cruces 2010). One quantitative study of developed regions finds that increased outflows of investment increase economic inequality (Alderson and Nielsen1998). Another, using a sample of Latin American and Caribbean nations finds that increased FDI is associated with a moderate and statistically significant increase in income inequality from 1970-2000. The studies above suggest that foreign investment increases economic volatility and increases labor demand for high-skill workers leaving low-skill workers behind and sometimes increasing income inequality. I include a measure of foreign direct investment as a percentage of GDP and

hypothesize that as FDI increases national income inequality in these 18 Latin American nations from 1980-2008 will also increase. Pension reform swept through Latin America in the mid-1990s because of the World Bank’s promotion of a plan to privatize pension and social security schemes (World Bank 1994). Chile was the first adopter in the late 1980s and the World Bank followed their format of multi- pillar pension scheme that emphasizes contributions and investment in private accounts. 15 of the 18 nations in this sample privatized their pension plans in some way (Gill, Packard, and Yerno 2005). One of the main goals was to improve pension coverage to more sectors of the economy such as the agricultural and informal sectors and reduce old-age income inequality and poverty for at-risk populations such as elderly women. However, especially for aging women, this does not seem to be the case. Studies point to lower pension benefits for women in post- reform Chile as late as 2002 (Arenas de Mesa and Mesa-Lago 2006), and lower benefits for women and low-income workers when compared to high-earning men in privatized countries (Arenas de Mesa and Mesa-Lago 2006; Bertranou 2001). However, others find that Chile, Argentina, and Mexico included public benefits and survivors annuities for low-income women and widows, improving their total lifetime benefits (James, Edwards, and Wong 2008). Unfortunately the effects of these inclusions may be subverted by traditional gender roles which demand women leave the workforce for childcare (Dion 2006, 2008) and rising divorce rates which exclude divorced women from their ex-husband’s pensions (Bertranou 2001; James, Edwards, and Wong 2008). In other areas of the world, pension privatization is expected to increase income inequality in Germany (Schmaehl 2007) and other developed nations (Casarico and Devillanova 2008). The evidence above in studies of pension privatization suggests that pension privatization does not benefit women and low-income workers. Unfortunately, no studies have systematically tested pension privatization in models controlling for other economic, social, and political factors in Latin America. Huber et al. (2006) analyze social security and welfare spending in Latin America and the Caribbean from 1970-2000 and find that it increases income inequality. I fill this gap by including a dichotomous measure of pension privatization and hypothesize that it will be associated with increases in income inequality because of the ill effects associated with low income and women workers pension benefits. The sections above examine the impacts of globalization on social spending levels and national income inequality. Studies suggest that trade openness, foreign direct investment, and

pension privatization may increase income inequality, while trade and FDI may lead to lower social spending levels. This is especially true for developing regions such as Latin America. The next sections discuss power resources theory and the impacts of national legislative politics, democratic strength, and repressive regimes on social spending levels and national income inequality in developed areas found in North America and Western Europe and developing countries such as those in Latin America and the Caribbean. NATIONAL POWER RESOURCES Power Resources Theory examines the connections between the state, political parties, organized labor, and their impact on social spending and policy in largely democratized, capitalist states. The theoretical approach developed in the late 1970s to better explain variations in social expenditures across developed and democratized nations in the global West (Stephens 1979; Korpi 1980). The approach originated in critiques of functionalist accounts which emphasized development, national industrialization, and demographic shifts as the main causes in growth of welfare state expenditures (O’Connor 1973; Offe 1972). However, functionalist explanations could not account for differences in expenditure levels and social policies. Power resources moved away from functionalist explanations that did not adequately address the power political actors and mobilized labor have in influencing social policy and spending decisions. Important political actors and groups include leftist and center-left political parties (Kangas 1991; Korpi 1980; Palme 1990), mobilized labor interests in the form of unions (Hicks 1999), and the strength of Catholicism and Christian Democratic parties in power (Huber et al. 1993; Van Kersbergen 1995). Power resources theory is also used to explain variation in income inequality in developed nations found in North America and Western Europe and is more recently expanded to developed nations such as those found in Latin America and the Caribbean. In the following sub-sections I examine power resources literature that analyzes variation in social spending levels and national income inequality in developed nations and then in developing areas. Social Spending and National Income Inequality in Developed Nations Early quantitative, comparative research using power resources shows contradictory results of the effects of strong Left political parties and mobilized labor power on social spending levels (Hicks 1990; Huber and Stephens 1993; Pampel and Williamson 1985, 1989; Pampel et al. 1990). Limits to classifying nations based simply on their social expenditures led to developing

welfare state typologies (Arts and Gelisse 2002; Castles and Mitchell 1992; Esping-Andersen 1990) rooted in T.H. Marshall’s (1964) historical analysis of extending civil, political, and social rights to citizens of Great Britain. Other studies anchored in power resources theory focusing on developing nations such as those found in Latin America expand the theory to include the strength of democracy and repressive regimes as factors because of their more recent transitions to democracy compared to the West (Avelino et al. 2005; Avelino and Hunter 1999, 2004; Kaufman and Segura-Ubiergo 2001; Huber et al. 2008a; Rudra and Haggard 2005; Segura- Ubiergo 2007). This section discusses analyses of welfare states using power resources, its critiques and expansions, and its applications in the developed and developing world.

Early quantitative power resources research does not provide a consistent picture of the effects of the Left and labor on variations in social spending levels in the developed world (Hicks 1990; Huber and Stephens 1993; Pampel and Williamson 1985, 1989; Pampel et al. 1990). While Hicks (1990) and Pampel et al. find positive effects for centrist and Right leaning governments on pension spending, Pampel and Williamson (1985, 1989) show no effects for political parties. Extending the analysis beyond spending to quality of pensions (Myles 1984; Palme 1988) and pension quality versus pension spending (Esping-Andersen 1990; Palme 1990), studies show that the strength of Left and labor positively influence pension quality while there is no effect on spending levels. Huber and Stephens (1993) build on these studies and find that Left parties exert stronger positive effects on pension spending compared to quality, while they find that Christian Democrat parties have the opposite pattern. The authors also find that mobilized labor in the form of voter turnout and strike rates had inconsistent effects on pension spending and quality. While these studies do not always support power resources theory, they suggest the importance of political power and process in determining welfare state spending.

As the Power Resources approach grew in popularity, studies shifted focus to defining types of welfare states based on various criteria. Esping-Andersen (1990) introduced a typology of welfare states that defines states based on eligibility criteria for social programs and benefits. Liberal regimes use means tested or income requirements and target those who cannot succeed in the market on their own. Benefits in these states tend to be minimal and may include employment incentives (Orloff 2002). The US, Great Britain, Canada, Australia, and New Zealand are all liberal welfare states. The Church and traditional ideas about family and gender roles heavily influence conservative or corporatist welfare states, such as Germany, France, and

Italy. These nations protect class and status through social insurance programs. While their benefits may be generous, they reinforce existing social hierarchies and inequalities. The most generous and egalitarian welfare state is the social democratic found in Sweden and the Scandinavian countries. These nations use citizenship as their criteria and have the most generous and the lowest levels of inequality of benefits and thus income inequality. Esping- Andersen’s classification system received criticism for being unable to place nations in the Southern Rim such as Spain in the three welfare worlds (Esping-Andersen 1990). In response, others developed new classifications that seek to encompass more nations within a welfare state framework (Arts and Gelissen 2002; Castles and Mitchell 1992). Esping-Andersen’s typology, however, remains the most popular.

Esping-Andersen’s typology of welfare states developed from T.H. Marshall’s (1964) manuscript that suggested citizens of a nation enjoyed fulfillment of civil, political, and social rights in three waves. Civil rights such as a fair trial developed as the judicial system grew in the 18th century, and political rights including the rights to organize and participate in political office developed in the 19th century. Social rights, which Marshall suggests were only made available in Great Britain at the beginning of the 20th century, include guarantees to old age income security, adequate healthcare, and investments in early education. As stated above, social democratic nations base their eligibility criteria for participation in social programs on citizenship status. Other nations that do not use these criteria often have less generous benefits and fewer social programs compared to social democratic models, and therefore their citizens may not enjoy fulfillment of their social rights.

The power resources literature also examines the effects of national politics on income inequality in developed countries. This body of research suggests that social democratic nations experience less income inequality when compared to conservative and liberal welfare states because of their generous and universal social insurance and benefit programs (Esping-Andersen 1990). Esping-Andersen’s (1990) typology shows that social democratic states, which are the most generous and equal, are best equipped to extend social rights to larger segments of their populations. For example, in Sweden and Scandinavia, which had strong Left political parties and labor, had higher levels of spending, more protections for their citizens, and lower levels of inequality compared to countries in continental Europe, the UK including Canada, Australia, and New Zealand, and the USA (Huber and Stephens 1993). These studies that point to the effects of

national political arrangements, including a strong Left and mobilized labor unions or political parties, on national income inequality in developed nations and so should be further tested in studies determining income inequality in developing regions. The next section turns to the effects of national political power resources on social spending levels in developing regions such as Latin America. Power Resources, Developing Nations, & Social Spending A few studies analyze the effects of national political parties on social spending in Latin America and other developing regions. Huber, Mustillo, and Stephens (2008a) find that legislative balance affects neither social security and welfare nor health and education spending. However, contra to hypotheses, it finds that executive partisan balance has a negative effect on health and education spending. The finding suggests the further to the left a president in Latin America leans, the less they spend on health and education. Two studies find that popularly based presidents, defined as executives closely linked to organized labor or whose constituents comprise the popular sector, increases social security spending but decrease health and education spending levels (Kaufman and Segura-Ubiergo 2001; Segura-Ubiergo 2007). While labor unions remain weak in Latin America and the developing world, one study uses a proxy measure of skilled and unskilled labor versus labor surpluses showing the higher the potential labor power in a country, the more it increases its social spending levels (Rudra 2002), though further analysis shows this to only be true in non-democracies (Rudra and Haggard 2005). Though they differ in their measurements and conclusions, these studies suggest the importance of domestic politics in determining social spending levels in Latin America. Also, studies of advanced capitalist democracies do not discuss the strength of a nation’s democracy. Democracy in advanced countries is usually a constant, with the United States and Western Europe having some of the most developed democratic systems in the world. In Latin America and the Caribbean, which was late to democratize, presence and strength of democracy becomes an important indicator of different types of welfare spending (Kaufman & Segura- Ubiergo 2001; Rudra and Haggard 2005; Segura-Ubiergo 2007). For example, one study finds that democracy has a negative impact on public pension spending, but a positive effect on combined health and education spending (Kaufman & Segura-Ubiergo 2001). Most find that a record of democracy increases social spending in its various forms (Avelino et al. 2005; Brown and Hunter 1999, 2004; Huber et al.2008). However, these studies dichotomize democracy as

being present or absent. In both regions, there is enough variability in the presence and strength of democracy that using a more comprehensive measure would be better for comparative analysis. In fact, states in Latin America have experienced different levels of democratization and authoritarian regimes at different time periods as late as the 1980s and early 1990s (Rueschemeyer et al. 1992). Two studies further extend the national political regime approach to authoritarian and non-democratic regimes, but find contradictory results. One (Rudra and Haggard 2005) finds that labor power only increases social spending in non-democracies, while the other (Huber et al. 2008a) finds that repressive regimes decrease health and education spending levels. Based on the information above, I hypothesize that national political party strength will have no effect on social spending patterns in Latin America. Due to poor data availability on labor unions in Latin America, this study does not address the power of organized labor. Democratic strength, shown to be consistently correlated to social spending levels, will be strongly and positively correlated to all forms of social spending. Repressive authoritarianism will have no effect on social security and welfare, but will reduce health and education spending levels. The next sub-section discusses the literature on national power resources and income inequality in developing regions and Latin America. Power Resources, Developing Nations, and National Income Inequality Studies drawing on the power of politics to affect national income inequality in Latin America find overwhelmingly that a strong democratic presence decreases national income inequality, while repressive regimes increase it (Huber et al. 2006; Rudra 2004). However, some contend that nations such as Argentina, Brazil, and Mexico experienced greater success incorporating the working classes in to the political process when compared to nations such as Bolivia, Ecuador and Peru which are marked by strong ethnic divisions, and nations dominated by elite ruling parties such as Colombia and Venezuela (de Ferranti et al. 2004). Legislative partisan balance, or the balance of power between Left and Right political parties, also affects national income inequality. A strong Left representation in the legislative cabinets is associated with declines in national income inequality in developing regions and Latin America (Ha 2012; Huber et al. 2006; Kemp-Benedict 2011). Conversely, nations with a history of clientelistic, patronage-based systems may not experience the improvements in education, social spending, and public services associated with reductions in income inequality (de Ferranti et al. 2004).

Studies of income inequality in developing regions include democracy as an indicator because of modern regime changes, transitions to and from democracy, and the historical instability of democracy. One study of less developed nations finds that strong democracies have less income inequality (Rudra 2004). My model study (Huber et al. 2006) show that democracy is nonsignificant unless interacted with social security and welfare spending levels suggesting that democracy alone is not redistributive unless coupled with certain forms of social spending. Adding to the explanatory power of regime type, Huber et al. (2006) also include a measure for repressive regimes based on their coding and suggest that repressive regimes redistributed income upwards in countries such as Chile and Argentina. However, they contend that as repressive regimes in Latin America and the Caribbean diminish in numbers their effects on national income inequality will dwindle. Their analysis of 18 Latin American and Caribbean nations from 1970-2000 finds that repressive regimes, while associated with higher income inequality, does not reach significance in their models. The two quantitative studies above find that regime strength, especially a strong democracy, leads to lower income inequality levels in less developed nations. However, this is only true in Latin America and the Caribbean when an interaction term between the historical strength of democracy and social security and welfare spending levels is included. I include a indicator of the history of democracy, measured as the cumulative democracy score from 1960 using the Polity IV data set (Marshall et al. 2009) and hypothesize that a strong history of democracy is associated with lower income inequality in Latin America. I also include a measure of repressive regimes using Freedom House data and hypothesize an associated increase in income inequality, though this should only be true up until 1993 when repressive regimes drop fully from my sample. Research focusing on the effects of legislative power balance finds that Leftist parties, associated with redistribution towards the poor, are statistically associated with reductions in income inequality. One study examines the impacts of Leftist cabinets and executives and finds that they are associated with decreases in income inequality even in the face of globalized, integrated markets that can increase inequality (Ha 2012). Another finds that nations with more redistributive parties has less of a wage gap and lower national income inequality (Kemp- Benedict 2011). My model study (Huber et al. 2006) finds that the history of legislative balance, measured as the cumulative score of Left versus Right parties in power since 1945, is associated with highly significant and moderate decreases in income inequality in Latin America and the

Caribbean from 1970-2000. I include a similar measure that captures the cumulative legislative balance of power from 1960 to the current year and hypothesize that a large score indicative of a historically strong Left cabinet to be associated with lower levels of national income inequality in Latin America from 1980-2008. The above sections discussed the effects of national power resources on social spending and national income inequality in developed and developing samples. Within developing samples, a strong history of democracy, an absence of repressive regimes, and a strong Left cabinet are associated with lower levels of income inequality. I further test these findings in my study. The next section explores universal human rights as a new and viable political power resource that can be used to affect social policies and problems connected to social spending levels and national income inequality. HUMAN RIGHTS AS A POWER RESOURCE Social Spending Levels and National Income Inequality In this section I extend power resources to universal human rights as another viable political tool for citizens to influence social spending and national income inequality. The power resources approach in welfare state literature emphasizes the strong effects national politics, labor parties and unions, Christian politics, and worker movements can have on social spending and inequality. Power resources theory is useful here because it helps us understand how politics, the state, the economy, and citizens interact to influence social spending and national income inequality. Interactions between these forces are especially relevant during the present-day period of neoliberal, economic globalization because as economic openness has increased between states and regions, so too has the need for bodies governing this interdependence. As societies feel the ill effects of neoliberal globalization, civil society responds by trying to ease some of these pressures. Universal human rights recognize Civil and Political Rights and Economic and Social and Cultural Rights, encompass national level rights of the citizen, and transcend the need for national citizenship by basing international rights on personhood status. States, in turn, institutionalize these rights through international human rights treaty ratification and National Human Rights Institutions adoption and commit to promoting and protecting the rights detailed in international law to all people within their borders. Treaties and institutions provide concrete resources and tools for citizens and politicians to extend universal human rights nationally and therefore should have positive effects on social spending levels across time.

Most nations in this study turn universal human rights into institutionalized political tools through treaty ratification and NHRI creation. Power resources theory hypothesizes that as the political tools grow for politicians, citizens, and often mobilized labor, the more likely there is growth in social spending levels and reductions in income inequality. International human rights treaties ratified nationally provide concrete language and rules that the country commits to follow. Politicians and national citizens can use this language when interacting with the government to influence decisions of social spending, policy, and problems such as inequality and poverty. Similarly, NHRIs, which embody aspects of domestic civil society, the state, international civil society, and global governance institutions, provide concrete resources in the form of staff, rights monitoring, educational initiatives, and reporting to national and international bodies which can also influence spending and inequality (Reif 2002). At the international levels intergovernmental organizations (IGOs), international nongovernmental organizations (INGOs), and transnational social movement organizations (TSMOs), a subset of INGOs, are also tools used by national and international players to influence social policy and solve social problems (Smith and Wiest 2005). Two important INGOs include Amnesty International and Human Rights Watch, which both use human rights discourse to frame social problems and injustices. Increasing numbers and national participation of INGOs lead to higher numbers of NHRIs in global samples (Koo and Ramirez 2009). However, studies of conflict showing INGOs may or may not have a positive effect on treaty ratification (Cole 2005, 2009; Wotipka and Tsutsui 2008) and a negative effect on Optional Protocols (Cole 2005, 2009). Regardless many INGOs continue to play important roles in promoting and protecting universal human rights by using their power and networks to interact with domestic civil society, national governments, and international institutions. This dissertation includes a measure of INGOs but emphasizes the effects of treaty ratification and NHRI presence as new tools in the political landscape used to increase social spending levels and indirectly affect national income inequality to help realize universal human rights at the national level. Social Effects of Treaty Ratification Studies of the effects of treaty ratification on CPRs drawn on human rights theory using international and regional samples yield contradictory findings, but suggest that universal human rights treaties influence whether a nation extends or retracts rights to the peoples within its borders. Analyses of the effects of treaty ratification on state led repression and torture find that

ratifying the Convention Against Torture increases torture and genocide (Hathaway 2002), especially if a nation is ruled by a dictator that allows some political dissent and competition (Vreeland 2008). However, treaty ratification decreases repression when ratification is coupled with a strong democracy or civil society (Hafner-Burton and Tsutsui 2005, 2007; Neumayer 2005), though it is unclear whether ratification’s positive effects extend over the years (Hafner- Burton and Tsutsui 2007). Another study examines policies on violence towards women in post-Communist Eastern European countries and finds that compliance with human rights treaties depended on UN involvement, degree of norm acceptance by government officials, and pressures from civil society; concluding that countries’ acculturation into international human rights norms through civil society interactions works better than other approaches (Avdeyeva 2007). An international analysis of judicial strength and effectiveness suggests that nations with stronger judicial systems ratify treaties less but also torture less, indicating that states need to develop their judicial capacity to properly enforce international treaties (Powell and Staton 2009). Clark (2010) concludes that the difference between treaty ratifications and repression scores in an international sample is unaffected by democratization and made worse by civil society engagement measured by international organization counts. These studies, while contradictory in their findings, show the importance of national political context when examining the effects of treaty ratification. With only one significant exception, nations with stronger democracies and civil societies experience fewer episodes of torture and state-led violence. These results suggest that treaty ratification within a democratic context, such as this sample of Latin American nations found in this study, provide new political tools for politicians and the people to use to increase protection and extension of their CPRs. Fewer studies examine extending ESCRs through treaty ratification and continue to provide an inconsistent picture but again suggest the ability of treaties to influence social problems and therefore social spending and income inequality. Some studies point to the improvements in child labor, immunization rates, and school enrollment (Boyle and Kim 2009), access to medicine (Hogerzeil et al. 2006), and public health improvements (Singh, Govender, and Mills 2007). Boyle and Kim (2009) also point to the importance of civil society and democratic strength in predicting less child labor, more immunizations, and higher rates of school enrollment. Others (Palmer et al. 2009) suggest no association between ratification of up to six treaties and HIV prevalence, mortality rates, and life expectancies, though the study lacks

other explanatory variables. A case study of Argentina and the CRC also provides conflicting results. The treaty resulted in improvements of children’s CPRs such as better judicial representation, though it had little effect on their ESCRs like improvements in housing and income. However, the treaty’s constitutional status created opportunities for domestic and international human rights and civil society groups to pressure the government to include more policies to protect children such as an Ombudsman dedicated to their defense (Grugel and Peruzotti 2007). While the studies examining ESCRs show an inconsistent relationship with treaty ratification, the evidence is largely positive that treaties, especially coupled with democratic regimes and strong civil societies, promote rights extension for at-risk populations such as children. Improvements to labor and health conditions and reductions to the inequality of access associated with treaty ratification suggest higher levels of social spending and potential reductions in income inequality. Therefore, I hypothesize treaty ratifications to positively influence health and education spending levels, but to have no effect on social security and welfare levels. Treaty ratifications will have no effect on social security and welfare because of the historical resiliency to short term political and economic shocks and changes. Treaty ratifications will be associated with reductions in national income inequality. Social Effects of NHRI Adoption Scholarship investigating the impact of Latin American NHRIs on extending rights suggests a recent shift from strictly civil and political to economic and social rights. Indeed some even call for an explicit shift in NHRI focus from CPRs to ESCRs (Kumar 2006). In the early days of Latin American NHRIs, offices handled high number of CPRs-related complaints. For example, even during the late-1990s and early 2000s, over 50 % of the 18,000 Guatemalan and 80% of the 8,000 Honduran complaints handled by NHRIs focused on violations of civil and political rights such as police and military violence. Conversely, two-thirds of Peruvian cases during this time period related to economic and social rights. Similarly, the Bolivian and Colombian Ombudsman offices handle growing numbers of concerns about the right to petition public officials, a surprise in Colombia given its atrocious CPRs record (Uggla 2004). In Argentina, the number of economic-based complaints concerning administration of water, electricity, employment, social security, health, environment, culture, and education continues to increase, especially considering the trend to privatize public enterprises promoted by neoliberal reforms (Reif 2000). NHRIs in the region show variation in their ability to initiate actions and

investigate public administration officials. They also show variation in the types of claims they handle. Neither CPRs violations nor ESCRs violations dominate the region. Evidence suggests that as democratic transitions mature and state violence reduces then the focus of NHRIs shifts from CPRs to ESCRs. The gradual process of moving from CPRs to ESCRs within a democratic context shows the strength NHRIs possess to increase the national political tools available and influence social conditions. Given their mandate for protecting the rights of citizens using universal human rights norms and law, NHRIs can play an important role in monitoring and protecting social rights to income security, adequate healthcare, and investments in primary and secondary educations. The increasing number of pension and wages complaints handled by NHRI offices in countries like Peru (Pegram 2008) and Bolivia (Uggla 2004) proves this. In turn, NHRIs provide reports to the legislature about the economic well-being of the nation, and can recommend policies and spending increases to better protect at-risk segments of the population (Reif 2000). Their prominent role in protecting social benefits such as social security indicates that NHRIs have the ability to influence social spending and income inequality. NHRIs also have a specific mandate to educate the populace of their human rights and experience much success in this area (Reif 2000). NHRIs educate the people through printed materials, media advertisements, conventional programs, and annual campaigns on their rights, procedures for grievances, and how to contact the Ombudsman office (Uggla 2004). They also often directly target occupational groups such as doctors and nurses or policy and military with educational programs focused on human rights. Educational initiatives by NHRIs should increase respect for human rights extension and focus attention on policy changes that help. The success of HROs in handling economic and socially based complaints and the outcomes of their educational programs and their impact on the continuing extension of social rights has yet to be properly analyzed. This study seeks fill this gap by analyzing the effects of NHRI presence on social spending levels and national income inequality. The literature above suggests that NHRIs should have a positive effect on social spending levels because of the increasing numbers of complaints about wages, discrimination, public pension fraud, and other economic problems and their mandate to investigate these claims and report to the government with recommendations. NRHIs also impact education spending levels because of their specific mandate of educating the populace of their rights and how to respect and protect them. I

hypothesize that NHRIs in Latin America positively influence all forms of social spending levels and are associated with lower income inequality. International Civil Society International civil society also plays a strong role in human rights development. INGOs such as Amnesty International and Human Rights Watch historically helped develop international and regional human rights regimes, and continue to promote and protect universal human rights today (Tahkur and Heine 2010). However, this role diminishes as nations institutionalize human rights (Risse and Sikkink 1999). In this study all nations have institutionalized human rights at the national level through treaty ratification and NHRI presence human. Therefore, theory suggests that INGOs will not play a significant role in the further extension of rights and social spending levels. However, studies discussed above on the effects of treaty ratification and NHRI formation show empirically that INGOs and civil society can still play a significant, albeit contested, role in extending or retracting universal human rights internationally. Indeed, INGOs also show much success extending education as a right, displaying the ability to spread educational resources across geographical boundaries, especially when working through TANs (Schofer and Meyer 2005). To that end, this study includes a measure of INGOs, and hypothesizes that these organizations will have a positive effect on social spending levels and negative association with increased income inequality in Latin America. The power resources approach consistently shows that the stronger the Left and organized labor are the more likely you will see more protective social policies and higher social spending levels. Therefore, extending power resources to universal human rights is appropriate because the theory was developed to explain all the political tools available to politicians and civil society in order to influence social conditions and problems, especially in terms of welfare state development. While the literature reviewed does not explicitly address the effects of universal human rights on social spending levels or national income inequality, the studies implicitly link human rights to social spending and income inequality through analyses of healthcare and labor conditions. I expect strong and positive correlations between all measures of human rights treaties and NHRIs and health and education spending levels, and no effect for treaty ratifications on social security and welfare. INGOs should have a small positive effect on social spending levels because although all nations in this sample have already institutionalized universal human rights, studies point to their continuing importance in extending universal

human rights (Thakur and Heine 2010). NHRI presence measured by counting years since ratification or creation will have positive effects on social security and welfare spending, though treaties measured in this same way should have no effect. Measuring human rights as years since ratification or NHRI adoption is done because of the historical resiliency of these types of spending to short term political and economic changes. All human rights indicators including treaty ratifications, NHRI adoption, and INGO presence will have negative associations with increased national income inequality. This chapter provided an overview of welfare state theories and my own expansions to them used to explain social spending variation and national income inequality in Latin America and other regions. The logic of industrialism says that industrialization and economic advancement leads to demographic and social changes because of market reliance on wage labor. The economic/globalization perspectives suggest that integration and competition in the global economy affects fiscal capacity and behavior of the state and therefore influences social spending levels and national income inequality. Power resources theory shows the importance of national political processes and power and their impacts on spending and inequality. It suggests a strong political Left and mobilized labor provide politicians and civil society with political tools that lead to higher spending and less inequality and continue to extend social rights to income security, adequate healthcare, and accessible and quality educations. I extend power resources to universal human rights as new political power resources that transcend the need for national citizenship as a requirement for the fulfillment of social rights and show the various positive and negative effects of treaty ratification, NHRI adoption, and INGO presence. In Chapter 3 I discuss the many data sources used in my dissertation, and the indicators used and analysis plans I employ in Chapters 4 and 5.

CHAPTER 3: DATA AND METHODS

In Chapter 3 I detail my research design and discuss the various data sources used, indicators and their measurements, and analysis plans for Chapters 4 and 5. I also include a discussion of my case selection of 18 Latin American nations. Most data were publicly available and I detail the sources below. I also used funds from my National Science Foundation Sociology Dissertation Grant (SES 1003012) to hire an undergraduate research assistant and collect and code new social, political, economic, and human rights data. When appropriate I update these data using original sources as new reporting became available. Each data section is organized by theoretical perspective. Analytically, I use pooled time-series models available in STATA 11 as well as multiple imputation techniques to handle missing data. RESEARCH DESIGN Case Selection The 18 countries chosen for this study are the core of what is generally thought of as Latin America and include Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, the Dominican Republic, Guatemala, Honduras, Nicaragua, Mexico, Panama, Paraguay, Peru, Uruguay, and Venezuela. They are drawn from the geographical regions of the Caribbean, Central and South America, share historical similarities in colonization and language, and have more commonalities culturally with each other when compared to nations traditionally defined as Caribbean. Latin American countries included in this study predominantly speak Spanish and all have histories centered on Spanish or Portuguese colonization. They also tend to have similar cultural patterns, with indigenous and European influences. While Brazil’s colonial ties are to Portugal and not Spain, no study of Latin America would be complete without them. This is due to Brazil’s central role in economic and social forces in Latin America. Undeniably, much recent high profile social movement activity in the region has occurred in Brazil with the rise of the Landless Peasants, the role of President Lula in balancing neoliberal policies with more those more socially conscious, and the significance of the World Social Forum as a major player in transnational social movement activity. Cuba is excluded because it is the outlier in the recent shift of democratization in Latin America and often lacks comparable data caused by lower frequencies of reporting. In 2000, it remains the only state classified by the Freedom House democracy scores as being not free, or authoritarian. Finally, similarities in economic struggles

of dependence and development due to colonial expansion make these 18 Latin American nations comparable.1 Other states in the geographical region are not included for important reasons. Historically the northern area of South America, with French Guiana, Suriname, Trinidad and Tobago, and other nearby island states, has very different economies and cultures. Similarly, states in the Caribbean such as Belize have their own historical differences that distinguish them from Latin America proper, making it difficult to include any one state without including the Caribbean as a whole. Future studies might consider analyzing the whole geographic area as some have done in recent scholarship (Huber et al. 1993; Huber et al. 2006). Data and Measures Most of the data used in this study are publicly available from two sources. The first is the Social Policy in Latin America and the Caribbean Dataset, 1960-2006 (Huber et al. 2008b). The second is the Latin America and the Caribbean Political Dataset, 1945-2001 (Huber et al. 2008c). NSF grants (SES 0241389 & SES 0241389) partly supported both of these data sets. These data compile indicators related to Latin American and Caribbean social policy and politics from many reliable sources like the Economic Commission for Latin American and the Caribbean (ECLAC), the United Nations (UN), World Bank, and International Monetary Fund (IMF). When appropriate, Huber and collaborators added data to cover a larger time period. The indicators used in this study from these data sets include dependent variables that measure types of social spending levels and national income inequality, and independent variables that act as the foundation of welfare state research. Other variables are derived from author coding using funds from my NSF Sociology Dissertation Grant (SES 1003012), made available by outside researchers and other online sources discussed below. See Table 2 below for a description of variables used in this study.

Dependent Variables Social Spending Levels. I use three dependent variables, social security and welfare, health, and education, in this study that measure social spending. All three come from the Huber et al. (2008b) data set on social policy, though I update them as needed using the original sources.

1 Other states in the geographical region are not included for important reasons. Historically the northern area of South America, with French Guiana, Suriname, Trinidad and Tobago, and other nearby island states, has very different economies and cultures. Similarly, states in the Caribbean such as Belize have their own historical differences that distinguish them from Latin America proper, making it difficult to include any one state without including the Caribbean as a whole. Future studies might consider analyzing the whole geographic area as some have done in recent scholarship (Huber, Ragin, Stephens 1993; Huber et al. 2006).

Huber and collaborators operationalize social spending variables by dividing yearly spending on social security and welfare, health, and education by the annual gross domestic product (GDP) in local currency units. The results are three continuous measures from 1980-2008.

To update variables through 2008 I consult the original sources and mimic Huber and team’s same method of data collection while also relying on newer data as they become available. They originally collected and calculated welfare state spending measures using the IMF’s Government Financial Statistics (GFS) and International Financial Statistics (IFS), Rossella Cominetti’s "Social Expenditure in Latin America: An Update" (1996) and Cepal’s “Base de Datos Estadisticas y Indicadores Sociales” (English version) and Social Panorama. For social security and welfare spending, they report that Cepal contains no data. Due to high correlations between the other three sources, but longer time series availability from the IMF, they chose to rely mostly on the IMF data. When possible, they use the other two sources to fill in missing values. For education and health spending they opted to use Cepal data, because IMF data does not cover subnational spending. Compared to Cepal, Social Panorama data begins in 1990 and provides better information on coverage. Therefore they use Social Panorama when it overlapped Cepal data. They chose between Cepal and Cominettii based on which set displayed more consistency with Social Panorama. Social Panorama does not collect subnational spending for Mexico, so they relied on Cepal data, which reports subnational spending. Even when combining these four sources, most countries in their data set do not have social spending data past the year 2000-01. I used the same method of data source selection by consulting the same sources, checking for consistency between data sets, and using the same criteria to decide which source to use to fill in missing values. In the case of Panama’s health spending during the 2000s I added the World Bank’s World Development Indicators as a 5th source because of its consistency with previous years compared to Cepal. The results are three time-series indicators that measure social security and welfare, health, and education spending as percentages of GDP from 1980- 2008 that rely mostly on CEPAL and IMF published data. Social spending level measures are used in Chapter 4 as the dependent variable. In Chapter 5 I combined health and education spending as one indicator and include it and Social Security and Welfare as independent variables.

The GINI Coefficient. The GINI coefficient is a measurement of income inequality within nations and comes from the United Nations University World Income Inequality Database,

WIID2 (UNU-Wider 2008). Measuring income inequality over time and across nations is no simple task, and there is no agreed upon definition of income inequality or its measurement. To that end, the UNU-Wider developed a database of income inequality measurements from different sources. The 2008 update reflects additions from new data sources and updates of existing data sources and represents a substantial improvement from its previous version, the WIID (UNU-Wider 2005), and previously popular income inequality calculations (Deininger and Squire 1996). Just as in the original version, the WIID2 is coded for the quality of the score, whether is covers rural and urban areas or both, the unit of analysis as the individual or household, and the use a household income equivalency adjustment. The data set also includes a new estimation method that predicts GINI values as accurately as reported values using reported decile data (UNU-Wider 2008). I use the values from the new estimation technique, best possible quality estimate available, only use estimates made using income as the concept (versus consumption or other market measures), averaged estimates when more than one per year were available, and collected information on household income equivalency adjustment use which I describe below. While the majority of observations are from the total area covered (urban and rural regions), estimates from Argentina and Uruguay rely mostly on urban areas. Independent Variables Logic of Industrialism. I use many measures from the logic of industrialism theoretical perspective. Economic and demographic indicators reflecting this approach come from the World Bank’s World Development Indicators (WDI) and the IMF’s Global Finance Statistics (GFS) and International Financial Statistics (IFS). The first, Development, is the gross domestic product per capita, or the percentage of GDP per person. Aged population (ln) refers to the percentage of people age 65 and over. Youth population (ln) refers to the percentage of people age 14 and younger. Urbanization (ln) refers to the percentage of people living in urban areas. Each of the latter three measures is continuous, and log transformed to correct for skewed distributions. Secondary School measures the percentage of the population that completed secondary school educations and is measured in five year intervals. These data come from the Barro, Robert and Jong-Wha Lee, April 2010, "A New Data Set of Educational Attainment in the World, 1950-2010.". This results in an incomplete time series of secondary school enrollment for each nation that is handled by STATA 11’s multiple imputation functions, which I describe in more detail below. Following the Huber et al. (2006) model for constructing an indicator for

Ethnic Diversity I collected data on ethnic populations in Latin America from the CIA’s The World Factbook (2000). Huber et al. (2006) use data from a World Bank publication (de Ferranti et al. 2004). I chose to use the CIA information to be able to construct my own version of the indicator using the raw data. Similarly to Huber et al. (2006) I construct a time invariant dichotomous indicator that measures ethnic diversity “…in which the total population of indigenous and African descent less than 20 percent or more than 80 percent (as in the case of some of the English-speaking Caribbean countries) were coded as not diverse, and in which such a population comprising 20 to 80 percent were coded as diverse.” Operationalizing ethnic diversity in this way captures a threshold effect on income inequality between heterogeneous and homogeneous ethnic populations, though this does not vary overtime. Chapter 4 uses Aged population (ln), Youth population (ln), and Urbanization (ln). Chapter 5 uses the original, non- transformed versions of youth and urban populations because visual analysis of the skewed distributions did not present a better alternative. In other words, a non-transformed or log transformed version of either variable would have been suitable and does not change the overall results. Chapter 5 also uses Secondary School and Ethnic Diversity.

Economics/Globalization. I include six measures of neoliberal globalization. Trade and investment information come from Huber et al. (2008b) and are updated using the IMF’s GFS (2011). Trade(ln), a continuous variable, measures total imports and exports as a percentage of GDP and is log transformed to achieve a normal distribution. Foreign Investment, a continuous variable, measures net inflows of foreign direct investment as a percentage of GDP. Deficit, a continuous variable, measures reported net revenue, expenditures, grants, loans, and repayments as a percentage of GDP. I then recode any surplus values to zero, to make the indicator reflect only deficit problems. Inflation is the annual percentage change in consumer costs of goods and services. Sector Dualism measures the absolute difference between agriculture employment as a percentage of total employment and agricultural production as a percentage of GDP (Nielsen 1994; Huber et al. 2006). I use data from Huber et al. (2008b) dataset on Latin American social policy to construct the variable using the absolute value command in STATA 11. All of the above indicators originate and are updated using the World Bank World Development Indicators (2011) Trade and investment information come from Huber et al. (2008b) and are updated using the IMF’s Global Financial Statistics (2011). Pension privatization data comes from the United State’s Social Security Administration website which includes information about the social

insurance and assistance system of each country around in this sample. Pension Reform is a dichotomous indicator showing when or if a country privatized their pension according to the World Bank’s (1994) recommendations. Of the 18 nations in this sample, 7 of them have not privatized their pensions. Chapter 4 uses the measure of deficit. Chapter 5 uses the measures of inflation and sector dualism. Both chapters use the measures of trade, foreign direct investment, and pension privatization. Power Resources. I include multiple measures drawn from the power resources literature. I use two sources to collect data on democracy and repression. Democracy scores come from the Polity IV dataset because these data reach before 1980, and repression from the Freedom House to prevent correlation problems between democratic and repressive indicators. Democracy, refers to the cumulated strength of democracy reported by the Polity IV dataset (Marshall et al. 2009). The Polity IV dataset ranks countries separately on autocratic and democratic practices and tendencies and then combines these rankings into one variable that scores a nation overall on its political and civil freedoms from -10 to +10. I sum the rankings from 1945 to 2008 resulting in a cumulated measure of the strength of democracy. Repressive Regime comes from the Freedom House, an organization that collects political data on most countries of the world. The Freedom House scores countries using a 7 point scale based on their protection of civil liberties and political freedoms. 1 is the best score and 7 is the worst. The organization uses 10 political rights and 15 civil liberties questions. Each question is worth 4 points, making the total points possible 40 and 60 respectively. They then use cutoffs in increments of 6 & 9 respectively to assign a score based on the 7 point scale. Finally, they average the two scores to assign a final strength of democracy score. An average score of 5.5 or above puts a country into the ‘Not Free’ category. I use this cutoff to construct a variable that reflects repressive regimes, or regimes that don’t allow basic political rights and civil liberties. This allows me to capture the effects of democracy with more detail. Following Huber et al. (2008a) I then sum the five years previous up to the year of observation to measure the medium- term effects of repressive regimes for analysis of social security and welfare spending levels, but use the current year for analyses of health and education spending levels. Political party data originally comes from Michael Coppedge’s (1997) dataset on Latin American politics and includes party names, votes and seats won, and a left-right, secular- religious coding scheme. Huber and team’s (2008c) treatment of Michael Coppedge’s data on

Latin American political parties adds cases and years unaddressed by the original researcher, and recodes the variable from vote to seat share. Vote share refers to the percentage of votes each category of political parties receive in an election. Seat share “…refers to the proportion of seats received by each category of party in each lower house legislative session during periods of democratic rule…” (Huber et al. 2008c). The team used a different source than Coppedge to establish the number of seats per party (Dieter Nohlen. 1993. Enciclopedia Electoral Latinoamericana Y Del Caribe. Costa Rica: Instituto Interamericano de Derechos Humanos). They then match the party names to Coppedge’s data to calculate seat share values. Dr. Coppedge provided me with an update to his dataset including seat share by bloc, lists and coding of political parties through the current period. His updated data set only provides information for 11 of the 18 countries in my total sample. Following Coppedge’s coding criteria, my assistant and I use the same strategy as Huber et al. (2008c) to expand the data through 2008 for the 7 missing cases. For the most current data, I draw heavily on Nohlen’s (2005) updated English language volumes and recent volumes of Arthur S. Bank’s (1989-2008) Political Handbook of the World. However, my assistant and I also used other sources listed in Appendix 2. These sources include party names, histories, elections, votes and seats won, and other miscellaneous information needed for coding such as political agenda or platform.

To capture the legislative partisan balance I use data collected and coded by Coppedge (1997), Huber et al. (2008c), and my team split into ideological blocs. The blocs include left, center-left, center, center-right, and right split by secular and religious political parties. It also includes personalist, other, and unknown categories when unable to classify a party along left- right, religious-secular lines. To reduce the number of categories, I first collapse the religious- secular splits. Then, following Huber et al. (2008a) I calculate Legislative Balance by weighting the seat share for each category and year. Right blocs are weighted by -1, center-right by -.1, center by 0, center-left by.5, and left by 1. The closer the resulting number is to 1, the more leftist the legislative balance of power in any given country during one year. I then use Huber’s dataset which includes legislative balance data from 1960 to create a cumulative measure, Legislative Balance(cum), which cumulates the weighted balance from 1960 to the year of the observation for the analysis of social security and welfare spending levels. Both Chapter 4 and Chapter 5 make use of all the cumulated measures of democracy, repressive regimes, and legislative partisan balance.

Power Resources and Human Rights. Human rights measures include INGO counts, international human rights treaty ratifications and NHRI adoptions. International organization data comes from one source, the Yearbook of International Organizations, an annual volume edited by the Union of International Associations (UIA). For a fee ($6750 USD) supported by an NSF Dissertation Improvement Grant (SES# 1003012), the Yearbook staff provided me with counts of INGOs in each Latin American country from 1960-2009. INGOs(ln) is a continuous indicator that is log transformed and measures the number of INGOs in each country. In this case, only currently functioning INGOs are included.

Human rights data comes from two sources. First, data on ratifying human rights treaties come from the online United Nations Treaty Collection. The online collection of treaties details if a country signs, ratifies, accedes, or succeeds to a treaty. This study includes the eight most recognized international human rights treaties, plus a treaty designed specifically for Latin America and the Caribbean2. Data for NHRIs comes from Dr. Jeong-Woo Koo and Francisco O. Ramirez’s (2009) data set that includes establishment dates for classical ombudsman offices, human rights commissions, and modern human rights ombudsman offices. Many countries host both the classical ombudsman and modern commissions and ombudsman offices, and they report this when appropriate within the data set. From 2004, the following cases from my sample do not have entries in Koo and Ramirez’s dataset: Brazil, Chile, Dominican Republic, and Uruguay. I collect additional data by consulting the authors’ original sources which are national websites for each missing country’s NHRI. The links to national NHRIs are found at the International Ombudsman Institute’s website3 and National Human Rights Institutions Forum website4.

I include multiple measures of human rights treaties. The first two, HR treaties and Optional Protocols, measures the count of international human rights treaties and the number of optional protocols to these treaties ratified by any given country. This analysis includes the eight most recognized treaties, plus one specific to Latin America and the Caribbean and the seven

2 International Covenant on Civil and Political Rights; the International Covenant on Economic, Social, and Cultural Rights; the International Convention on the Elimination of All Forms of Racial Discrimination; the Convention on the Elimination of All Forms of Discrimination against Women; the Convention Against Torture; Convention on the Rights of the Child; the Convention on the Elimination of Genocide; Agreement establishing the Fund for the Development of the Indigenous Peoples of Latin America and the Caribbean; Convention on the Rights of Persons with Disabilities; the International Convention on the Protection of the Rights of All Migrant Workers and Members of Their Families 3 (http://www.law.ualberta.ca/centres/ioi/) 4 (http://www.nhri.net/)

established optional protocols. ICESCR, ICCPR, ICCPR OP, CHILD, CEDAW, CEDAW OP, MIGRANT, and INDIGENOUGS are dichotomous indicators showing when a country ratified, acceded to, or succeeded each treaty. I also created a cumulative measure of each that measures the number of years since ratification that includes (cum) after the treaty abbreviation. NHRIs, is a dichotomous indicator showing when a country begins a classical Ombudsman, modern Human Rights Commission, or hybrid Human Rights Ombudsman. I create a similar measure of years since adoption named NHRIs (cum). Together, these variables are used to measure the impact of the human rights project on the universal extension of social rights in the late 20th Century. Chapter 4 uses the dichotomous and cumulative measures of treaty ratifications and NHRI adoption while Chapter 5 uses the dichotomous measures. I do not include treaty ratifications that do not reach statistical significance and all significant ratifications are included regardless of their directional sign.

Methodological Controls. Huber et al. (2008a) include a statistically significant dichotomous methodological control for health and education spending data that comes from Cominetti (1996) when analyzing Latin American social spending patterns. This is due to its consistently higher reporting of these values when compared to the other data sources. If the data comes from Cominetti, they code the variable as 1. They code it 0 for any other data source. I, however, updated these series using new CEPAL reporting and minimized the usage of the Comminetti data as much as possible. Still, the time series of social security and welfare, health, and education spending levels use different sources. While I initially included a data source measure I found that including it in models always resulted in a statistically significant relationship with variation in social spending levels. However, upon inspection of my two main data sources I found that IMF data to be older (1980-1990) while the different iterations of CEPAL data began in 1990. Visual inspection of changes in data sources over time by country displayed in Graphs 1, 2, and 3 confirm this pattern and show that including this data source control is actually an unnecessary period control between IMF reported past years of lower spending and CEPAL reported current years of higher spending levels for social security and welfare and education spending levels, and of no consequence for health spending levels because of the predominance of CEPAL data in this time-series. I drop this variable from my analysis of social spending levels.

Huber et al. (2006) include a statistically significant methodological control for household size in determining income inequality in their analysis of Latin America income inequality. I also use the control called Equivalency Scale to control for differences in household sizes used to calculate the GINI coefficient. The equivalency scale is a dichotomous indicator coded 1 for years where no adjustment was needed and coded 0 if a household per capita adjustment was used to control for differences in household size and income across households used to calculate the GINI. Just as Huber et al. (2006) predict, I expect no use of a household size adjustment to decrease inequality. I use two historical indicators to control for unmeasured period-specific effects on the whole region which I model after Huber et al. (2008a). The first, Debt Crisis, is a control for the debt crisis of the 1980s (1982-1989), when many of the countries in the region began defaulting on foreign debt loans, starting with Mexico. As a result, the IMF and World Bank pressured many Latin American countries beginning in the 1990s to implement structural adjustment programs that centered on privatization of state-run industries, deregulation, the lowering of trade barriers and tariffs, and the cutting of social services (Babb 2005). While Huber et al. (2008a) mark 1990-2000 as a period of debt recovery, I instead mark it as the rise of neoliberal adjustment, called Neoliberal Decade. The years 2001-2008 represents the baseline category. These historical controls are included in both the analyses included in Chapters 4 and 5. ANALYTIC PLAN Analysis of Social Spending Levels I use an unbalanced panel data set with 18 countries and up to 544 observations from 1980-2008. Analytically, I model this study after the only truly comparable analysis (Huber et al. 2008a), but make significant improvements over statistical modeling techniques. I build on their analytical procedures because my dependent variables, and many of my independent variables, come from the same data, and thus also suffer from serially correlated errors (Huber et al. 2008b, 2008c). While the previous study uses ordinary least squares (OLS) with panel-corrected standard errors, I use a non-parametric OLS modeling technique available from a user generated program in STATA that uses Driscoll-Kraay (1998) standard errors (Hoechle 2007).

Standard OLS regression presents several problems for panel data. According to Hicks (1994, 172), using OLS for time-series equations results in errors that suffer from first-order auto-regression, cross-sectional heteroskedasticity, and temporally correlated cross-sections.

Other approaches include the Parks-Kmenta method (1997), later shown by Beck and Katz (1995) to produce overly small standard error estimates and unable to handle contemporaneously correlated panels. To account for these problems, I use the following approaches. I first construct OLS using the regress command in STATA 11. After each model I test for heteroskedasticity, or the correlation of errors. I find correlation of errors in most models for all types of spending and so construct new pooled time-series random and fixed effects models using the xtreg, fe or re commands available in STATA 11. These models build on OLS by relaxing some assumptions about error terms and structures. Each model generated includes a Wald’s test showing heteroskedasticity is still present. I save the estimates from both and perform multiple Hausman tests (Hoechle 2007) to find which model is appropriate. The tests indicate that the model is of a fixed effects nature where the unit effects are correlated with the independent variable effects. These results hold true even after including period indicators measuring the decade effects of the 1980s debt crisis and 1990s rise of neoliberal globalization. Further testing using the xtcsd command in STATA 11 necessitates a better model revealing the presence of group-wise heteroskedasticity and cross-sectional dependence between panels as still present (De Hoyos and Sarafidis 2006). Hoechle (2007) created the STATA command xtscc to analyze pooled time-series data that suffers from first-order autocorrelation, heteroskedasticity, and cross-panel correlation of errors regardless of the temporal spacing. The model builds upon Driscoll and Kraay’s (1998) treatment of balanced panel data by extending it to unbalanced panel data and uses Monte Carlo estimates to prove its superior utility in analyzing small samples in the presence of cross- sectional correlation. The author contends that while the PCSE method handles cross-sectional dependence well, it is at the cost of a parametric approach that assumes equal correlation of cross-sections. The Driscoll and Kraay (1998) standard errors do not use parametric constraints to assume equal spacing of cross-sectional dependence and therefore do not force the assumption that cross-sectional errors are temporally dependent. A number of studies, discussion and conference papers successfully employ this model to analyze social spending in the EU (Klein, Leibrechy, Onaran 2009), employment and wages in Romania (Aparaschiveli et al. 2011; Andreica et al. 2010), the fiscal effects of EU Cohesion Policy (Hagen and Mohl 2009), determinants of voter turnout in the USA (Childers and Binder 2010), the impact on productivity of China’s domestic patents (Zhao and Liu forthcoming), international income inequality

(Mahler 2010), and volatility of capital flows in developing countries (Broto et al. 2011). I use this method as my main statistical approach and include a fixed effects estimator due to the results from the Hausman test above. Huber, Mustillo, and Stephens (2008a) use OLS with PCSE in their main statistical models and check their results robustness with fixed effects models. I do not check for robustness because of the nonparametric nature of the xtscc command created by Hoechle (2007). However, like Huber et al. (2008) I do not include a lagged dependent variable on the right side of the equation. While Beck and Katz (1996) later recommend including a lagged dependent variables, I do not because it suppresses explanatory power for other independent variables (Achen 2000). Achen (2000) cautions that static data may require a lagged dependent variable and dynamic data do not. However, Beck and Katz (2004) show that first-order auto- regression correction implicitly incorporates a lagged dependent variable without constraining the explanatory power of other independent variables. Yet this method provides imprecise estimates, especially when the number of years (T) is near or small compared to the number of cases (N), as is the case in these data. Hoechle’s (2007) technique of using OLS with Driscoll and Kraay (1998) standard errors corrects for first order autocorrelation and thus also implicitly includes a lagged dependent variable. Analysis of National Income Inequality I encounter two major challenges in modeling these data to analyze the determinants of national income inequality in Latin America: correlated errors and missing data. Despite problems with missing data points, I begin by checking traditional OLS and pooled time-series models using fixed and random effect approaches to test for heteroskedasticity, correlated errors within units, and cross-sectional dependence between panels. I decide that a non-parametric model using Driscoll-Kraay (1998) standard errors adapted by Daniel Hoechle (2007) to handle unbalanced panel data with a small N is the best modeling technique for this analysis. However, missing data points still remain a large problem. The GINI indicator is reported in uneven intervals resulting in incomplete cases and necessitating a method to increase the number of observations used in each model. To address missing data points, reduce standard errors and increase the efficiency of my models (Frees 2004, p.10) I implement a set of commands available in STATA 11 called multiple imputations (mi estimate). This package in STATA allows the user to separately impute missing observations using available data in other indicators used in the

analysis and then employ statistical modeling techniques that produce invalid responses with a very small N (StataCorp 2009). While this does not fill in missing data points with the real values, it imputes average simulated values based on available data points in other indicators, increases the number of observations in the analysis, and then allows use of Hoechle’s (2007) xtscc package to analyze pooled time-series data. Standard OLS regression presents several problems for panel data. According to Hicks (1994, 172), using OLS for time-series equations results in errors that suffer from first-order auto-regression, cross-sectional heteroskedasticity, and temporally correlated cross-sections. Other approaches include the Parks-Kmenta method (1986), later shown by Beck and Katz (1995) to produce overly small standard error estimates and unable to handle contemporaneously correlated panels. To account for these problems, I use the following approaches. I first construct OLS models using the regress command in STATA 11. After each model I test for heteroskedasticity, or the correlation of errors. I find correlation of errors in most models and so construct new pooled time-series random and fixed effects models using the xtreg, fe or re commands available in STATA 11. These models build on OLS by relaxing some assumptions about error terms and structures. Each model generated includes a Wald’s test showing heteroskedasticity is still present. Unfortunately, there are not enough observations in the full models to perform the Hausman or xtcsd tests for heteroskedasticity and cross-panel correlation, and so I use a reduced model that excludes the household income equivalency adjustment, secondary school completion, and economic sector dualism indicators to temporarily increase the number of observations. These indicators are reintroduced in the full models after using the multiple imputation functions. I save the estimates from the reduced models and perform multiple Hausman tests (Hoechle 2007) to find whether a random or fixed effects approach is appropriate. The tests indicate that the model is of a fixed effects nature where the unit effects are correlated with the independent variable effects. Further testing using the xtcsd command in STATA 11 necessitates a better model revealing the presence of group-wise heteroskedasticity and cross-sectional dependence between panels as still present (De Hoyos and Sarafidis 2006), leading to the use of Hoechle’s (2007) xtscc command which corrects for heteroskedasticity, autocorrelation, and cross-sectional dependence. In my analysis of national income inequality I do not use a fixed effects model. While Hoechle (2007) suggests that estimating models with pooled OLS presents inconsistent results

when a fixed effects model is more appropriate; however, I do not use a fixed effects model despite the statistical test performed above. I decide against a fixed effects model because the average number of time points (13) caused by missing data in the dependent variable is smaller than the number of units or clusters (18) precluding the usage of a fixed effects model (Beck 2001; Beck and Katz 1995:635–4). Furthermore, some contend when using time invariant variables (Ethnic Diversity), effects of levels of independent variables on levels of the dependent variable prior to the year of observation (the history of democracy, repressive governments, and legislative balance), effects of levels of independent variables on the dependent variable (economic development, social spending, inflation, sector dualism, trade levels, foreign investment, population levels, school completion rates, etc.), and variation in independent variables that is mainly cross-sectional and not temporal (pension reform, human rights treaty ratification, and NHRI adoption) in fixed effect estimations will inappropriately suppress coefficients (Plumper et al. 2005, Huber et al. 2006). In other words, because of the small number of time points relative to clusters, the nature of the variables I use and their relation to the dependent variable I reject the use of fixed effects modeling techniques. However, the problem of omitted variable bias is still present when not using a fixed effect or unit-dummy approach. To control for the possibility of unit effects correlated with independent variable effects I include two historical period effects that theoretically affected all countries in the sample similarly and at the same time. These two historical period controls are the 1980’s debt crisis and the 1990s rise of neoliberal economics which both affected Latin American nations across the sample in a similar fashion. That is, omitted variable bias normally controlled for in fixed effects models or with unit dummies is unnecessary when historical period controls that represent similar effects across the sample are included. My study is methodologically innovative because I integrate Hoechle’s (2007) approach for handling heteroskedasticity, autocorrelation, and cross-sectional dependence with multiple imputation techniques available in STATA 11 to handle missing data points that lead to incomplete cases and a greatly reduced sample size in analytic models. In a study using much of these same data to analyze national income inequality in Latin America from 1970-2000 Huber et al. (2006) chooses not to address missing data points and perform their analysis using 135 observations. They do so by choosing to forgo pooled time-series estimations altogether and instead rely on OLS with robust cluster standard errors and historical period indicators to control

for omitted variables and unmeasured period-specific effects on the whole region. Unfortunately, as discussed above, OLS with robust clusters does not fully account for cross-sectional dependence and so should not be relied on. Others also contend that when using pooled time- series data standard errors reduce and coefficient efficiency increase as the number of observations increase (Frees 2004). My initial full models have only 127 observations because of missing data points from variables GINI, Equivalency Adjustment, Secondary School, Dual Sector, Social Security and Welfare, and Inflation. After implementing multiple imputations in STATA 11the number of observations in each model increases to 234, thereby improving the statistical validity of my results. Multiple imputations gained popularity in studies analyzing surveys with missing response data (Rubin 1976, 1987, 1996; Little 1992; Meng 1994; Schafer 1997; van Buuren, Boshuizen, and Knook 1999; Little and Rubin 2002; Carlin et al. 2003; Royston 2004, 2007; Reiter and Raghunathan 2007; Carlin, Galati, and Royston 2008). However, analyses of political data using various levels of measurements also successfully employ multiple imputation techniques. For example, studies successfully employ multiple imputation techniques to analyze public opinion of pension policy in European democracies (Lynch and Myrskylä 2009), international welfare panel data (Maltitz and Van Der Merwe 2012), voter preference in China (Li 2011), trends in cohort and period protest participation in the USA (Caren et al. 2011), development of better democracy indicators for political analysis (Plumper and Neumayer 2010), the effects of racial minority status on losing on a particular ballot in California (Moore and Ravishankar 2011), and how housing policy reduces voter turnout (Gay 2012). Additionally, some have gone on to further develop multiple imputations to better handle panel and longitudinal data (Honaker and King 2010; Plumper and Neumayer 2010). STATA’s multiple imputation suite checks for missing observations, imputes values based on complete cases from other variables used in the dataset, allows the user to set the number of imputations generated, and perform advanced statistical regression analyses. While it does not generate the true value of a missing observation, STATA’s mi suite calculates multiple possible observations, runs each imputation as a separate analysis, and then pools the completed analyses into one set of reportable results. It is superior to other techniques of handling missing data such as listwise deletion which results in the exclusion of missing cases and may drastically reduce sample size (StataCorp 2009).

More specifically, to impute missing values I use the multiple imputation multivariate normal regression (mi impute mvn) method. This method imputes values of missing observations in continuous variables. The command uses an iterative Markov chain Monte Carlo (MCMC) method to fill in the missing values over a user-defined number of iterations. MCMC uses Bayesian logic to perform Data Augmentation on the incomplete variable so that it becomes amenable to predicting the missing values multiple times. The MCMC process iterates until a stationary distribution of imputed values is achieved. Unlike maximum likelihood estimation, MCMC does not guarantee convergence, and the user must decide the number of iterations appropriate. Two main options for this method are “burnin” and “burnbetween” which set the number of iterations needed for the MCMC to converge to a stationary distribution and the number of iterations between imputations respectively. I keep these set at the default of 100 each, resulting in 10,000 iterations total for my imputation process (StataCorp 2009). Even at 100 imputations my models do not show convergence by the 100th iteration. Thus, I perform a visual analysis detailed in the STATA manual to ensure that the MCMC method of multiply imputing data missing from my dataset is appropriate (StataCorp 2009). I first perform multivariate imputation with the MCMC only option which prevents the process from actually imputing the values. However, it is not feasible to monitor each individual imputation for convergence. I save the Worst Linear Function (WLF) to a new dataset. The WLF is the linear regression performed with the most amount of missing data and is the most likely to converge to stationarity the slowest (Schafer 1997). The new dataset contains three variables: wfl, iter, and m. Iter is the number of iterations used to impute the missing values, m is the imputation numbers which are 0 because this process stopped before actual imputation and corresponds to the number of iterations, and wlf is the estimates for the Worst Linear Function. I use time series commands to set iter to the time variable and plot the WLF across the number of iterations to look for a pattern. If no obvious pattern in the estimates plotted over the burn-in periods emerges then I look to the autocorrelations of the WLF to view the number of iterations between imputations to look for independence (StataCorp 2009). I will explain the imputation of missing values for my dataset. I begin by checking my dataset for missing values and then imputing the missing values. As you can see from Box 1 Secondary School is missing 79% of observations because it is measured in 5 year intervals, Sector Dualism is missing 28%, while INGOs, Inflation, and Social Security & Welfare are

missing 10% or less. Health & Education and Foreign Investment are missing less than 1% and so I do not impute their missing values. Despite large amounts of missing data in the GINI and its household income equivalency adjustor I do not impute these values to preserve the data. After imputing the missing values 100 times over 10000 iterations you can see that the process results in a set of variables with more observations than I started with, enabling me to increase the number of observations in the analytic models. This process does not try to impute every missing observation, is still missing data from the dependent variable and its calculation control indicator, and so does not result in the full 522 observations available in this data set. After running Hoechle’s non-parametric pooled OLS models and a robust cluster random effects robustness check I find that even with 100 imputations (10000 iterations) the models do not converge by the 100th iteration. I impute the values again using the MCMC only option to prevent actual imputation and save the estimates to a new dataset called wlf.dta. I use the same sample by setting the seed (12345) and perform the visual diagnostics discussed above. I first set the data to a time series format with the number of iterations as the time variable. Graph 4 plots the WLF estimates across the number of burn-in iterations and shows no consistent pattern or visible trend in just as I would expect in properly converged models. Graph 5 shows autocorrelation in the WLF to show the number of iterations to use between imputations to guarantee their independence. As you can see the autocorrelations drop off quickly after the first- lag autocorrelation and altogether by twentieth-lag autocorrelation suggesting that less than 100 iterations between imputations is appropriate (approximately 20-30), though I still keep this option set at 100 to ensure statistical validity. Table 3 shows the new summary statistics for the imputed dataset. To calculate the total variation in income inequality explained by each model (r-squared) I follow Hoechle’s (2007) advice. Hoechle suggests that using the adjusted r-squared values from the normal OLS with robust-cluster errors is the best way to report total variance. STATA 11 includes a multiple imputation add-on that automatically calculates and reports the r-squared for each model. I report these values in Table 8. This chapter provided details of my data sources, indicators and their measurement, and analysis plans. For each set of results in Chapter 4, Model 1 includes controls and logic of industrialism approach, Model 2 introduces neoliberal globalization, and Model 3 introduces traditional power resources theory. The following models introduce INGO counts and NHRIs

adoptions in Model 4 and then counts of total treaties ratified and single treaty ratifications. I introduce treaty ratifications separately because of their high inter-correlation. In Chapter 5 I discuss the results of the non-parametric pooled OLS modeling techniques developed by Hoechle (2007) over 8 models. Model 1 begins by introducing the measurement and historical period controls and the economic perspective including national and international economic and welfare state indicators. Model 2 introduces the demographic and society perspective which highlights demographic transitions and social indicators focusing on educations and ethnic diversity. Model 3 introduces the power resource perspective that emphasizes national political regimes and legislative power balances. Models 4-8 introduce single treaty ratifications and a total count of international human rights treaties ratified. Single treaty ratifications not included in the models showed no statistical associations with national income inequality measurements.

CHAPTER 4: EXPLAINING VARIATIONS IN LATIN AMERICAN SOCIAL SPENDING

In this chapter I provide a brief overview of Latin American social spending systems before discuss the determinants of variation in social spending levels for 18 Latin American nations from 1980-2008. The models in Table 4, 5, and 6 show the determinants of social security and welfare, health, and education spending as a percentage of GDP in 18 Latin American nations from 1980-2008. For each table, model 1 introduces the historical period controls and logic of industrialism, models 2 and 3 introduce the neoliberal globalization and traditional power resources perspectives, and the final models introduce human rights as extensions to power resource theory including years since a NHRIs was adopted, counts of treaties and optional protocols, the length of time since ratification for individual treaties, and INGOs. The models provide mixed support for logic of industrialism, neoliberal globalization, power resources, and human rights approaches. These results also provide mixed support for the research questions: do institutionalizing universal human rights through treaty ratification and NHRI presence represent a successful extension of power resources theory by leading to higher social spending levels? Do classical approaches to analyzing welfare state spending through industrialization, globalization, and national political processes also explain variation in Latin American social spending? These results suggest that human rights treaties and institutions have divergent effects on social security and welfare spending; they converge and have positive effects on health and education spending levels, while national economic development, the globalization of national markets, and national political processes still exert many effects.

Latin American Social Spending Systems Latin America is a worthy testing ground for classic welfare state indicators and integration of new theoretical approaches because of its developed social systems when compared to other developing regions such as sub-Saharan Africa and Southeast Asia. However, variation in social spending still characterizes the region. Variation within spending across nations is useful to those looking to explore the determinants of these types of spending in developing regions. This section provides an overview of how social security and welfare, health, and education spending systems developed in Latin America.

For example, Argentina, Uruguay, Guatemala, and El Salvador represent the highest and lowest social spenders respectively. While the highest come from South America and the lowest from Central America, location within Latin America does not decide spending levels. Ecuador, Paraguay, and Peru, located in South America, fall into the lowest half of total spenders, while Costa Rica, Nicaragua, and Panama fall into the upper half. High social security and welfare spending drive high total spending levels in Argentina and Uruguay, while either education or health drive total spending levels in Guatemala and El Salvador. Social security and welfare spending are the most fiercely protected within the region. Pension systems in Latin America date back as far as the 1920’s (McGreevey 1990). Historically politicians from the left and right defended these systems to protect certain occupational groups. Administrations extended the benefits to new groups of employed, like agricultural workers whom often exist on the fringes of the formal economy. The gradual extension to more sectors of the economy means that more and more workers in Latin America have a vested interest in resisting any impulse to lower social security and welfare spending. However, the debt crisis of the 1980s forced many countries in the region to review critically their social spending policies. As they opened their economies to trade and financial liberalization, employers could no longer afford the high pension taxes they paid for their employees, Right-leaning governments began suppressing social security spending, and Chile privatized their pension schemes. In 1994 the World Bank used Chile as a model to develop a plan to help states privatize social security to reduce social spending, increase economic development, and move insolvent and corrupt social security schemes from the government to the private sector. Yet some studies predict the move to privatization will result in short-term increases in social security spending, especially in Chile and countries that followed the model of total privatization (Weyland 2004; Gill, Packard, and Yermo 2005). Healthcare systems in Latin America often developed in step with social security systems. Often the funding for healthcare comes from pension funds, and social security health clinics often get funding from pension expenditures. For example, in 1983 ‘social security institute’ expenditures in 6 of the 18 nations from this study spent more than 50% on maternity and healthcare. All 6 used less than 25% of social security institute spending on pensions. To make matters worse, social security health care coverage rates for the richest provinces was two to ten times higher than in the poorest provinces (McGreevey 1990). While this is good news for

those living in developed, urban regions, it means lower access to social security funded healthcare for much of the poor, rural populations. One reason for historically low health spending could be because social security institute spending often covers healthcare services. Despite healthcare and pension system developing in tandem, their expenditures don’t always vary in time together. Indeed, Argentina and Uruguay, which have two of the highest levels of social security and welfare spending, also have remarkably low levels of health spending. Conversely Guatemala and El Salvador, which have two of the lowest levels of social security and welfare spending, have comparably higher levels of health spending. However health and social security spending levels in El Salvador follow closely together. A marriage of private and public efforts characterizes education systems in Latin America. While most middle and upper class youth attend private primary and secondary schools (Puryear 1997), the poor most often attend public institutions. The poor also drop out before completing secondary educations, with less than 60% of poor children remaining in school by the 5th grade (IADB 1998). Although, states publicly fund university educations, as many as 90% of students in Latin American middle-income countries come from the middle and upper classes (Brown and Hunter 2004). Some contend that regime type is a predictor of education spending levels in Latin America. One study of Latin American education spending shows the strength of democracy positively impacts education spending. Countries that scored highest on democracy rankings on average spent $20 more per student than countries that scored the lowest on democracy rankings. Electoral politics also affects social spending. Electoral politics were used in Fujimori’s authoritarian Peru and Cardoso’s democratic Brazil as a means of securing popular support for the government (Brown and Hunter 2004). The susceptibility of education spending to political manipulations suggests that this type of spending is less able to withstand economic shocks, regime changes, and political shifts as well as social security in welfare spending. Social security and welfare, health, and education systems in Latin America respond differently to similar processes. Social security and welfare spending represent the most resilient form of spending, because of their high popularity, relatively wide coverage of the population, and powerful constituents that seek to protect against political manipulations. Many pension systems in Latin America began in the 1920s, meaning that authoritarian regimes either started or managed them. When democracy swept through the region in the 1980s social security and

welfare systems remained solvent and were favorably viewed under both authoritarian and democratic administrations. This exhibited their resiliency to regime change. Health and education spending levels, however, are more subject to economic shifts, political projects, and regime change. Healthcare systems draw much of their funding and services from social security spending, suggesting that health spending is a low priority for politicians, the rich and the poor in Latin America. Little political protection may leave health spending vulnerable to economic shifts caused by globalization, but also opens an opportunity for human rights workers to bring the focus back to health spending, a core focus of human rights documents and law. Education systems in Latin America are often the pet projects of politicians seeking to curry favor with the general population. When politicians seek the poor’s approval, they allocate spending towards primary and secondary schooling, and when they seek middle and upper class support, they allocate spending towards higher education. During times of economic crises, politicians know that education spending is an easy target compared to social security and welfare. The priority placed on education spending by left and right administrations as a means to increase popularity suggests that human rights institutions and instruments should find it relatively easy to target education systems for more spending. Indeed, education spending effects, when allocated properly, are easy to see due to the direct effect 1 or 2 years of extra schooling has on an individual’s economic productivity (World Bank 1991). The next section first provides an over of the international human rights system, then reviews classical welfare state theory, and introduces universal human rights as new political power resources to influence social policy and conditions and therefore spending levels. The following sections detail data, analytic plans, results, and a discussion of the strengths and limitations of the study and possible future directions I intend to pursue. Social Security and Welfare Spending Determinants Table 4 analyzes the effects of human rights as power resources and classical welfare state theories on social security and welfare spending levels. In Model 1 the debt crisis of the 1980s and the 1990s decade of neoliberal globalization show negative and nonsignificant associations with social security and welfare across models suggesting the continuing resiliency of these types of spending against economic downturn. Social histories of pensions and pensions spending in Latin America suggest that these social entitlements are fiercely protected by the

landed elite and the lower classes because of social security’s relatively wide coverage throughout the region.

The logic of industrialism, which suggests increased industrialization and economic advancement affects welfare state spending, only enjoys partial support in explaining social security and welfare spending levels. All three indicators are logged transformed meaning the coefficient must be divided by 100 to show the change in social security and welfare spending as a percentage of GDP caused by one unit change in GDP per capita, urbanization rates, or aged population percentages. Development or GDP per capita shows a negative and significant association with social security and welfare levels though the effect is negligible with a coefficient of -0.000. The results of the development and historical period indicators support past research suggesting that social security and welfare spending are less subject to the types of economic shocks felt during the 1980s and 1990s throughout the region. Yet the coefficient for development also suggests that despite increases in overall economic growth, social security and welfare spending increased as GDP declined. For example, in Venezuela as GDP per capita declined, social security and welfare as a percentage of GDP rose, while in Argentina during the 1990s a short term decrease in economic development coexisted with a continuing increase in social security and welfare spending levels. While most other countries show more dramatic growth in these spending compared to GDP per capita, Chile and Mexico show the opposite pattern where GDP per capita grows overall and social security and welfare spending levels decline. These results combined imply that social security and welfare spending, while seemingly invulnerable to periods of economic trouble, are also not driven by a country’s economic development. While spending continues up and down at an uneven pace throughout most nations in the sample, economic growth slowly creeps up. The percent living in urban areas shows a small and significantly positive effect on social security and welfare spending with a 0.0418% increase for every 1% increase in the urban population. This result suggests that demographic shifts do impact social spending levels when domestic and international politics are not considered. The aged population shows a positive but nonsignifcant effect on these types of spending. The logic of industrialism theory, emphasizing demographic shifts and economic development, provides mixed support in Model 1 and has the lowest r-squared for all the models. Alone, it does not adequately explain variation in social security and welfare spending levels in these Latin American nations.

The economic/globalization perspective in the contemporary period emphasizes the effects of trade, foreign investment, and budgetary deficits on social spending levels and suggests that Latin American nations experience unfair bargaining conditions because of their lower economic development when compared to donor nations and institutions. It begins in model 2. Indicators from Model 1such as the neoliberal decade shows a moderate and negative association with these types of spending, while the percent of population 65 and older become significant and increase social security and welfare spending levels by .02344% for each 1% increase in aged populations. GDP per capita and urban population effects are consistent with Model 2. The log transformation of net trade inflows and outflows negatively impacts social security and welfare spending. For each one unit increase in net imports and exports as a percentage of GDP, there is a -.0105% decrease in social security and welfare as a percentage of GDP. While this effect is small, it reaffirms that increased global economic integration and competition exerts downward pressure on pension and welfare spending levels to remain attractive to international trade partners. An inflow of foreign direct investment is incorrectly signed according to theory but is nonsignificant. The negative effects of trade suggest support for arguments that neoliberal globalization imposes unfair trade packages and regulations on developing countries which causes them to slash social spending to remain competitive. The deficit indicator which measures the annual budget deficit as a percentage of GDP shows a moderately strong positive effect on social security and welfare spending, suggesting that budgetary constraints do not depress social spending levels, though the finding does not hold across models. Finally, a variable indicating when or if pension privatization occurred shows a small, but nonsignificant, negative effect on social security and welfare spending. Interestingly, when run as a dichotomous indicator of before and after privatization, there is a significant and large positive effect on social security and welfare spending. When pension privatization is enacted there is an associated 1.1-1.2% increase in social security and welfare spending levels across all models, supporting some that suggest privatization incurred large short term costs and not reduce spending (Gill et al. 2005). The results above indicate that neoliberal globalization can have unintended consequences such as increasing social spending despite budgetary constraints and low levels of national economic development, though net trade exerts downward pressure.

Model 3 introduces the power resources perspective. The power resources perspective suggests that a strong democracy and Left political balance should lead to higher social spending levels while repressive regimes may decrease them. In the model, the neoliberal decade, urban and aged populations lose significance. Huber, Mustillo, and Stephens (2008) extend the approach by including measures of democratic strength and repressive regimes. They find the history of democracy from 1945 is the only significant political predictor of these spending levels. The results from my study also show support for the history of democracy from 1960 to the current period. The sum of increases and decreases in Polity IV -10 to 10 score on the strength of democracy increases social security and welfare spending levels by .015%. I measure repressive regimes in this analysis as a sum of the 5 years of ‘not free’ status before the year of observation (Freedom House). This indicator has a stronger and positive effect when compared to democracy. As the sum of the previous 5 years of repression increases, social security and welfare levels go up by .272%. This effect increases in strength once human rights as a power resource are introduced. Studies show that repressive politicians in Latin America come from the left and right and both protect and dismantle social spending, but that repressive, right regimes often increase social security spending when compared to the more social justice oriented left. Repressive regimes may further increase these spending types because of the pressure applied by left political parties and civil society using universal human rights as a new national power resource. The cumulative balance of legislative power from 1960 to the current year between left to right political parties is negatively associated with social security and welfare spending levels, but is nonsignificant in this model and the rest. These results suggest that leftist national politics are less important than regime type and repressive administrations for predicting variation in social security and welfare spending levels in Latin Amerca. The power resources theory indicates that a leftist balance of political power and a strong history of democracy should lead to great social spending levels. However, politically repressive regimes in Latin America are historically associated with the left and right, and are not articulated in power resources theory. These findings imply that a history of democracy helps to protect social rights, but that a historically repressive regime also increases spending that protects old age and indigent rights to income security. This is a curious finding given the social justice, human rights, and economic equality platforms of many leftist political parties in the region. Similarly curious, political repressive regimes seem to extend social rights too. This is because authoritarian regimes created

many pension schemes in Latin America for protected pieces of the population, and modern repressive governments need to keep the support of protected groups that gained in political power over time. Repressive administrations, however, do not necessarily respect human rights law and have reduced in numbers in Latin America and are no longer present in this sample after 1992 (see Table 7).

Models 4-9 provide only partial support for extending power resources to universal human rights. Human rights theory puts states on a loose trajectory towards the consistent promotion and protection of human rights (Risse and Sikkink 1999). I extend power resources to human rights institutionalized at the state level as new political tools available to influence social spending. The logged count of INGOs is signed inconsistently and is nonsignificant in all models, due either to the many categories of INGOs within the measure, or because of the diminishing role international civil society plays after a country institutionalizes universal human rights (Risse and Sikkink 1999). The number of years since forming an NHRI, or typically a Human Rights Ombudsman, is moderately and positively correlated with social security and welfare spending levels across all levels. NHRIs in the region handle growing numbers of complaints related to social security and public benefits (Uggla 2004; Pegram 2008). Human rights language in agreements and laws specifically mandates the protection of social security and welfare income and benefits, so it is no surprise that NHRIs have at least a modest effect on these types of government spending. When an alternate measure is used that simply dichotomizes before and after forming an NHRI the coefficient increases dramatically suggesting a larger initial bump in spending that diminishes over time. A baseline model (4) including NHRIs but no treaties reaches significance and indicate that these institutions exert positive effects on these types of spending independently of treaty ratifications. Model 5 includes a full count of all 7 available Optional Protocols, and has a strong negative effect on social security and welfare spending, while a total count of treaties ratified (not shown in Table 4) does not reach significance. In Model 6 the cornerstone treaty the ICESCR also negatively affects social security and welfare spending. For each year after ICESCR ratification, these types of spending levels reduce by .179%. The same is true for the CEDAW Optional Protocol, the treaty on the Latin American indigenous, and the migrant workers. Model 6 explains the most variation in social security and welfare spending, increasing from 63.3% in Model 1 to 77.9% in Model 6 with the ICESCR. Other treaties such as the ICCPR, the CEDAW, and CRC (children’s rights)

are not included here but have positive associations and fail to reach conventional statistical significance in two-tailed tests. The ICCPR reaches a significance level of .1 in a one-tailed test but has a 95% confidence interval of [-.0110188 to .0828369]. Much like a dichotomous measure for NHRIs, a before and after measure of the ICESCR and ICCPR ratification show a very large and positive association with social security and welfare spending levels, though for the ICESCR this effect becomes negative overtime.

Other indicators in these models from classical welfare state theories suggest their ongoing explanatory power. For example, aged populations remains positively associated with these spending levels. Urban populations, however, switch direction when including human rights indicators suggesting that the more people move to the cities, the lower the social security and welfare spending levels. While historically, governments direct social spending and protections to the urban areas, embracing universal human rights means extending social rights to all peoples within the borders. However, this negative effect loses significance in Models 6 and 8 and is very small throughout the models. GDP per capita remains negatively signed and significant but with a negligible effect. Trade levels remain consistently significant, but with a small negative effect on social security and welfare spending levels, while the budgetary deficit indicator loses significance beginning in Model 4.

The results above are contradictory in answering the question of whether universal human rights as a new power resource extend social rights to income security and social assistance. INGOs, vibrant political resources in international civil society, show no associations with social security and welfare spending levels. NHRIs, with their insider access to national legislatures and executives and quasi-state nature, take seriously their role in protecting social rights and influence national governments to spending more on social security and assistance. Human rights treaties, binding yet difficult to enforce, are meant to be legal vehicles of universal human rights norms that states use to protect the civil, political, and social rights of its citizens and people. Previous research casts doubt on this idea showing that some states actually increase their abuses post-ratification, an occurrence human rights theory allows for (Risse, Ropp, Sikkink 1999). NHRIs, INGOs, and other civil society organizations use human rights language found in treaties on a frequent basis to influence government policies. In this case, years since human rights treaty ratification leads to lower levels of spending, casting doubt on their efficacy in protecting social rights. Optional Protocols to international human rights treaties establish

committees where citizens can directly air their grievances. However, in this analysis they negatively affect social security and welfare spending levels, casting doubt on the power of the protocols to protect social rights. Regardless, NHRIs use human rights language embedded in treaties to protect social rights and continue to increase handling social and economically based complaints about social security and assistance provision. The consistently negative effects of years since treaty ratifications suggests that social security, considered an important social entitlement instead of a right in many nations, needs more attention as a universal human right, a project which NHRIs seem to be taking on with enthusiasm.

However, this does not explain why treaties are negatively associated with these types of spending. In a macro-comparative quantitative analysis such as this one, we can only speculate why this pattern emerges. One hypothesis put forth by studies of treaty ratification suggests that nations may publicly embrace universal human rights to reduce international pressure and continue violating the rights of their citizens (Hathaway 2002; Vreeland 2008). Further studies should analyze comparative historical data to understand fully the relationship between ratifying international human rights treaties and social security and welfare spending levels in Latin America. For example, social spending levels also do not necessarily equate to adequate social security income and welfare benefits, so more analysis of treaty ratification effects on social rights outcomes such as average annual benefits, welfare dollars spent per person, or changes in pension and welfare programs may yield different results. Additionally, ratification may be an inadequate measure. Measures that capture the extent that a nation has internally institutionalized treaty language through national legislation, executive decree, and constitutional status should be developed to better understand the impact universal human rights norms have on government behaviors. In that case, treaty ratification may have a more direct effect on national policy, which would then in theory translate to better protection of social rights.

The results for Table 4 remind scholars to be mindful of relying on any one theoretical perspective too heavily. No one theoretical perspective and its corresponding indicators best explains variation in social security and welfare spending levels in 18 Latin American nations from 1980-2008. The logic of industrialism approach finds consistent and strong support for aged populations, but contradictory results for economic development, and inconsistent results for increasing urban populations. Globalization theory finds that trade reduces social security and welfare spending levels, but that none of the other indicators reaches statistical significance.

Traditional power resources theory explains social security and welfare spending levels in Latin America as outcomes of national political processes. I show that regime strength and repressive administrations increase these spending while legislative cabinet balance has no effect. The human rights approach finds only partial support for the extension of social rights measured by social security and welfare spending in NHRIs. International human rights treaties and optional protocols yield negative associations with social security and welfare spending, implying they lack the strength to enforce norms that protect social rights. Latin American nations are faced with challenges such as large scale demographic shifts, vulnerability to neoliberal globalization, and regime shifts as they try to extend protection of social rights to all peoples as mandated by international human rights agreements. While NHRIs use international human rights agreements to complement this mission, and have positive effects on social rights protection, simply ratifying human rights treaties and Optional Protocols cannot do this alone. Health Spending Determinants Table 5 shows the determinants of health spending level over 11 models. Model 1 begins with the logic of industrialism, economic globalization in model 2, power resources in model 3, and then human rights indicators in models 4-11. The results from Table 5 show that universal human rights as a power resource have strong and positive effects on health spending as a percentage of GDP. This is especially true when compared to social security and welfare spending. However, one must keep in mind that social security and welfare spending respond slower to political and economic changes when compared to health and education spending levels. To better measure the immediate impacts of political and economic changes, I do not use cumulated political and human rights variables with the exception of the historical strength of democracy. Overall, human rights as a power resource leads to higher health spending levels in these 18 Latin American nations from 1980-2008.

Model 1 includes the historical period controls, and logic of industrialism variables. Neither the 1980’s debt crisis nor the 1990s decade of neoliberalism show statistical significance and this holds true across models. Both of these periods are associated with large budget cuts for social spending programs. For the logic of industrialism indicators GDP per capita and the percent of population age 65 and over increase health spending levels. However these effects are small with a one percent increase in GDP per capita resulting in .000% increase in health spending and a one percent increase in aged populations leading to a .017% increase in these

spending. These results hold true across the 10 models. The percent urban population is incorrectly signed and does not reach significance. As Latin America experiences growing numbers of elderly over age 65, it needs to increase health spending to accommodate growing healthcare needs; which is a necessary component in extending social rights to adequate healthcare. While social security and welfare spending seemed impervious to economic decline, they were positively driven by demographics.

Model 2 introduces economic globalization variables. This theoretical perspective suggests that developing countries in Latin America will see negative impacts on social spending because of neoliberal policies pushed by wealthy donor countries and international financial institutions such as the World Bank, IMF, and WTO. The only indicator from this perspective that reaches significance levels is foreign direct investment (FDI). For each one unit increase in FDI there is an associated .096% decline in health spending levels. This effect holds true except for Model 9 with the introduction of the Convention on the Elimination of Discrimination Against Women. While the log transformation of trade flows is signed correctly in Model 2 it does not reach significance. However, net trade leads to lower health spending levels (.003- .006% reduction in health spending for each 1% increase in net trade levels) in Models 4-11 when the human rights indicators are introduced suggesting that they work in competition with extending social rights through universal human rights. The deficit as a percentage of GDP is correctly signed but nonsignificant; suggesting that budgetary constraints do not drive health spending levels, and is consistent across all models. The above results suggest that foreign investment and international trade, with its often strict requirements for economic restructuring and policy exerts downward pressure on health spending levels, while trade and budgetary deficits do not, though the effect is minimal.

Model 3 draws on the power resources perspective including the cumulated strength of democracy from 1960, a measure of repressive authoritarianism for the current year of analysis, and the current legislative partisan balance. Democracy is incorrectly signed, but significant only in model 5, and the effect is very small. While power resources theory predicts that democracy will increase health spending, the effects from this analysis are negligible. A repressive regime is one ranked as not free by the Freedom House. Challenging theory, the variable is significant and positively signed in models 4-10 once I introduce human rights indicators. The effect is large and ranges from 1.03-1.24% increase in health spending levels during repressive administrations.

The large increase in health spending levels during repressive regimes in models with human rights indicators suggests that authoritarian administrations feel pressure to comply with human rights agreements because they are a new power resource that the Left and civil society uses to influence policy and spending decisions. Regardless, the effect is undeniable and suggests a relationship between the extension of social rights to healthcare and repressive regimes interacting with the universal human rights projects. The legislative balance is positive and strongly correlated with health spending across models, though this finding is only true in Models 3, 5 and 8 which include a general count of human rights treaties ratified to date and the single treaty ratification of the ICCPR Optional Protocol. A higher score for this variable indicates a higher proportion of leftist legislatures in power and power resources predict that leftist politicians will spend more on health. For example in the power resources baseline model (3) 1% increase in the proportion of left legislative cabinet members leads to a 0.215% increase in health spending levels. Once the human rights indicators above are introduced, this effect grows to a 0.423% increase in Model 5 and a 0.307% in Model 8. These results show that both leftist national legislatures and repressive regimes extend social rights by increasing health spending and that this effect is stronger when universal human rights are introduced, suggesting authoritarians respond to the pressure of universal human rights in a similar fashion as leftist legislatures.

Models 4-11 introduce human rights indicators as a new power resource which suggests that nations institutionalize international human rights norms by working with INGOs, ratifying treaties and installing human rights institutions and thus provide new tools for the people and political actors to use at the national level. These indicators include dichotomous measures of treaty ratification and NHRI presence. INGO counts show mostly positive but nonsignificant effects on health spending, further showing the inability of this measure to explain social spending. NHRIs show significant and moderately correlated effects on health spending levels across models. The adoption of an NHRI leads to a 0.311%-0.458% increase in health spending. Their mission includes specific language to protect and promote the health of the peoples they service. These results show they live up to this mission within this sample. Model 4 includes a count of all treaty ratifications and is significant and positively correlated with health spending. For each new treaty ratified up to a count of 10 there is a 0.371% increase in health spending levels. Models 5-10 include single treaty ratifications. The ICESCR, ICCPR, ICCPR OP, and

CEDAW show the strongest positive effects on health spending, though all treaties included in Table 5 are significant and positive. The strongest effect, ratifying the ICESCR is associated with a 1.238% increase in health spending as a percentage of GDP. The second largest effect from ratifying the CEDAW leads to a 1.119% increase in these spending. The smallest bump in health spending levels comes from ratifying the treaty establishing a fund for indigenous peoples in Latin America and leads to a 0.247% increase. All of these treaties contain specific language protecting healthcare as a universal human right. Overall, embracing universal human rights through NHRI adoption and treaty and optional protocol ratification provides new power resources to continue extending social rights to adequate healthcare in Latin America.

The results for health spending provide an interesting juxtaposition to the results for social security and welfare spending. While analysis of the determinants of health spending levels in Latin American again show why scholars should not rely solely on one theoretical perspective, this analysis shows clear evidence of the positive effects of institutionalizing human rights norms as a new power resource in Latin American countries on health spending when compared to social security and welfare spending. In terms of demographics and national development, health spending relies on increased economic development while social security and welfare spending increase despite periods of economic decline. Both respond positively to increases in aged populations, though there is no consistency in the effects of increasing urbanization in this sample. While health spending has a very weak and negative association with democratic strength, social security and welfare spending show a weak and positive association. Both forms of spending respond similarly to repressive regimes within the human rights context suggesting that authoritarians that openly embrace human rights increase social spending levels. While their motive is not clear, the effect is strong for social security and welfare, and health spending levels. Legislative cabinet balance between left and right political parties increase health spending in only three models but has no effect on social security and welfare spending levels. Health spending responds differently universal human rights as a new power resource to extend social rights at the national level, and leads to large increases in health spending levels, while NHRIs have a positive effect and treaties have a negative effect on social security and welfare spending levels. Economic, Social, and Cultural rights clearly include social security, welfare, and health as top priorities. Clearly, however, the institutionalization of ESCRs through treaties and institutions as new power resources does far more to extend social rights

through health spending compared to social security and welfare spending. In other words, nations showing their commitment to extending social rights through treaty ratification and human rights institution adoption experience more positive effects on health, versus social security and welfare spending levels, in these 18 Latin American nations from 1980-2008.

Education Spending Determinants Table 6 shows the determinants of Latin American education spending levels from 1980-2008 over 11 models. Model 1 again introduces the controls and logic of industrialism. Model 2 includes the globalization variables. Model 3 adds power resources variables to the analysis. Finally, Models 4-11 introduce human rights as a new power resource to influence social spending. Unlike health or social security and welfare spending, education spending responds differently to international economic turmoil, industrialization, demographics, globalization, and is mostly unaffected by national politics. The combined results suggest that much like health spending, education spending levels respond very positively to human rights as a new power resource used to continue extending social rights at the national level. Model 1 shows that 1990s neoliberalism negatively affect education spending and is true in all but Model 8. This effect diminishes from a 0.8% to a 0.2%-0.4% decline in education spending levels once universal human rights as a new power resource are introduced. This reduction in the negative effect of the neoliberal decade suggests the strength of human rights as tools to be used by national political actors and civil society. The 1980s debt crisis does not reach significance, but shows a positive effect once a human rights treaty count and ICCPR ratification are introduced in Models 5 and 8, suggesting that universal human rights can overpower international economic turmoil to a small degree. However, these results are not consistent across models in significance or direction. GDP per capita shows no effect until human rights as a power resource are introduced, though the effect is negligible. The urban population positively impacts education spending in most models, leading to a .02%-.04% increase in education spending levels with each 1% increase in urbanization. The rapid pace of urbanization forces Latin American nations to continue increasing education spending in cities to fulfill their obligation of extending this important social right to its population. Finally, the percent of population under 14 years reaches significance in Model 2 old, shows on average a .02% increase education spending with each 1% increase in youth, and is consistent except for Model 9 which introduces the CEDAW. However, the modest findings do indicate that

demographic increases in urban and youth populations do impact education spending positively. The results from the controls and logic of industrialism indicators suggest that economic downturns and demographic transitions do have small effects on education spending levels in Latin America, though introducing universal human rights as a power resource can counter some of the ill effects of international economic liberalization. Model 2 includes the globalization indicators. Out of the three indicators, only trade is significant. Critiques of economic globalization point to the negative effects trade and foreign investment can have on developing countries social spending and programs. Others suggest that to remain viable in the international market developing countries must invest in educational programs to produce competitive workers. Net trade flows positively affects education spending levels in Latin America in Models 2, 3, and 9. Models 2 and 3 do not include universal human rights, while Model 9 introduces ratification of the CEDAW suggesting that NHRIs and human rights treaties replace the seemingly positive effects of trade. Foreign direct investment and budgetary deficits, while signed correctly, do not reach statistical significance and show no effect on education spending levels. Neoliberal globalization does not have a large impact on education spending in Latin America during this time period, and the positive effects of trade are replaced by the overwhelming positive effects of universal human rights as a new power resource. Model 3 again introduces the power resources perspective. Here education spending responds consistently only to democracy. The effect is significant, positive, and very small leading to a 0.003%-0.007% increase in education spending levels with each one point increase in the strength of democracy on the Polity IV scale. Much like social security and welfare and health spending levels, repressive regimes mostly show positive associations with education spending levels. While repressive regimes are nonsignificant in the baseline models without human rights, they show a 0.516% increase in education spending levels once the ICESCR is introduced and at the lowest a 0.356% increase once the CEDAW is introduced. These results again suggest that repressive regimes feel pressure from national politicians and domestic civil society to extend national social rights to education after embracing universal human rights through treaty ratification. The indicator for legislative partisan balance is not significant and signed incorrectly across models. Unlike health spending levels, education spending is not as subject to the manipulations of national politicians even within the context of universal human rights embracement, though like social security spending it responds favorably to democratic

strength. Equally interesting, all forms of spending respond favorably to repressive regimes in many models, suggesting that even repressive politicians from the left or right cannot subvert the new political tools represented by universal human rights. Beginning with model 4, I introduce human rights as new political power resources. INGOs again show no association with education spending, suggesting the limits of their capacity for action or the poor quality of this measure. NHRIs and human rights treaties have very specific language mandating the protection of social rights through quality educations. Quality educations are enhanced by high spending levels. These variables show a consistently positive relationship with education spending levels. NHRIs have a strong and positive effect on education spending across models and do not seem affected by treaty ratifications. The consistently show between a 0.56%-0.68% increase in education spending levels once they are adopted. The full count of human rights treaties is significant and positive, as are all of the single treaty ratifications including the ICESCR, ICCPR, ICCPR OP, CEDAW, INDIGENOUS, and CMW. Ratifying the ICESCR shows the greatest effect leading to a 1.604% increase in education spending levels, while the CMW has the smallest positive increase of 0.191%. When compared to the results from health and social security and welfare spending levels, education and health spending are similar in their response to the 4th stage in the search for rights embodied by the institutionalization of international human rights norms. Human rights treaties and institutions primarily drive these types of spending, while national politics and demographics exert greater influence over social security and welfare spending. Table 6 casts doubt on the ability of traditional welfare state theories to explain education spending levels in Latin America from 1980-2008, and suggests that embracing universal human rights as new power resources primarily drives increases in these types of spending . While demographics and globalization do not impact education spending in this sample, democratic strength and repressive regimes lead to higher education spending levels. Human rights as a new power resource have a much larger and consistent impact. Again, human rights treaties and institutions specifically protect old age, sickness, and disability income security, adequate health standards, and compulsory and quality primary and secondary educations. The consistently positive and strong effects of NHRI adoption and treaty ratifications on education spending levels suggests the strength of universal human rights as new power resources to continue extending education as a national social right.

When compared to the analyses of health spending levels, it is clear that education spending levels share a more consistently positive and stronger response to universal human rights when compared to social security and welfare spending levels. Social security and welfare spending levels do not respond favorably to human rights treaty ratification suggesting that variation in these spending types are primarily driven by demographics, globalization, and regimes. All forms of spending respond to demographic changes similarly. While democracy only had small effects on education and social security and welfare spending levels, it has no effect on health spending. In this sample repressive regimes are concentrated at the center of political spectrum, represent only 5% of observations, and increase all types of spending. National political balance only affects health spending levels, and does this inconsistently. Overall, national political arrangements matter much more for social security and welfare spending when compared to health and education spending in Latin America. Universal human rights as a new power resource has across the board positive effects on health and education spending, but only NHRIs positively affect social security and welfare spending, suggesting that social security and welfare spending levels remain impervious to short term political changes represented by ratifying international human rights treaties.

CHAPTER 5: EXPLAINING INCOME INEQUALITY IN LATIN AMERICA

In this chapter I provide a short overview of income inequality estimates in Latin America, which is among the worst in the world. I then discuss the results of my analysis on Latin American national income inequality using the GINI coefficient as my dependent variable over nine models using STATA 11 to perform multiple imputations and a non-parametric form of pooled OLS regression that corrects for heteroskedasticity, autocorrelation, and cross-panel dependence. These results come from the pooled coefficients of 100 multiple imputations to ensure their statistical validity. While I report the results of non-parametric pooled OLS techniques, a robustness check using robust cluster random effects models yields the same results showing the strength of my analysis. I find that extending national political power resources to universal human rights in the form of treaty ratifications and NHRI adoptions yields mixed results. While NHRIs show no association with income inequality, single treaty and combined ratifications counts reduce national income inequality significantly in these 18 Latin American nations from 1980-2008. In Table 8, Model 1 introduces the household income equivalency adjustor, the 1980s debt crisis, and 1990s neoliberalism as controls. It also introduces variables that draw on economic and welfare state theory to predict income inequality, as well as economic globalization theory centered on the effects of trade, foreign investment, and pension reform. Model 2 introduces demographic and social indicators associated with the logic of industrialism such as youth and urban populations, school completion rates, and ethnic diversity. Model 3 includes variables drawn from the power resources perspective and includes the historical strength of democracy, repressive regimes, and legislative cabinet balance. Models 4-8 introduce human rights as new power resources and include INGO counts, NHRI adoptions, single treaty ratifications, and a count of total human rights treaties ratified. I find that national development decreases inequality while trade and foreign investment increases income inequality. Social security and welfare spending, largely regressive in Latin America, also increases income inequality. Youth and urban populations increase income inequality, while percentages of secondary school completion and ethnic diversity show no effects. A history of democracy and leftist legislatures also reduces income inequality. Amongst human rights indicators, only treaty ratifications have any effect, and they are associated with reductions in income inequality,

providing further evidence that international human rights treaties continue to provide new political tools in the national arena that politicians and civil society members use to influence social policy and problems. National Income Inequality in Latin America Latin America is a useful testing ground for analyzing income inequality because its rates of inequality are among the worst in the world (de Ferranti et al. 2004; Deininger and Square 1996; IADB 1998; Morley 2001). For example, calculations in the 1990s show the top 5% of earners shared 25% of the total national income while the same earners in Southeast Asia and developed nations share 16% and 13% respectively. These figures suggest that income in Latin America is highly concentrated amongst the top earners. Income concentration amongst the top earners stands in stark contrast to the income concentration amongst the lowest earners. While in other regions of the world the poorest 30% of earners share over 10% of the income, in Latin America during the 1990s they share only 7.5% (Deininger and Square 1996; IADB 1998). World income inequality in the 1990s measured by the GINI coefficient is 0.4 (IADB 1998). As you can see from Graph 6 income inequality in Latin American countries rarely declines to the global average. Only Uruguay and Argentina dip below the global average of 0.4 and most countries hover above or around a GINI of 0.5. In this sample, average income inequality rises from a GINI of 0.503 in the 1980s to 0.526 in the 1990s and the declines again slightly from 2000-2008 to 0.524. Countries such as Bolivia, Brazil, Colombia, Ecuador, Honduras, and Venezuela have a GINI that crests 0.6 in some years, showing the upper limits of income inequality in the region compared to the relatively low global average of 0.4. Stepping back, Latin American income inequality rose after the 1982 debt crisis. While estimates show that income inequality in the region declined 10% from 1970-1982, after 1982 the highest earners increased their share of income by 10% while poorest earners decreased their share of income by 15% (IADB 1998). However, these patterns are not homogenous across the region. Reports suggest that income inequality in Brazil, Chile, and Mexico worsened during the 1980s and then tapered off in the 1990s while Costa Rica and Colombia had a relatively stable GINI throughout the period. I, however, show with updated GINI estimates in Graph 6 that Colombia experienced a decline in income inequality in the 1980s that rapidly rose to 1980 levels by 1990 and did not decline overall through the 2000s. Otherwise, during the 1980s most nations in this sample experienced an overall rise in GINI estimates.

The 2000s mark a period of declining national income inequality as shown in Graph 6. Overall the region saw a small decline from the 1990s average. Countries such as Mexico and the Dominican Republic saw a decline of less than 5 points while others such as Bolivia, Brazil, Chile, Argentina, Ecuador, and El Salvador experienced greater declines in income inequality. Other studies show a similar decline for Mexico in the early 2000s (IADB 1998). Early 2000s estimates of income inequality in Argentina show a marked rise of income inequality from the 1970s until 2002 (de Ferranti et al. 2004) though my own updated estimates show a decline beginning in 2003. Others, such as Colombia, Costa Rica, and Uruguay saw no decline and perhaps a small increase in income inequality. However, the region overall fared better in the 2000s up until 2008 than it did in the 1990s when regional average inequality rose substantially from the 1980s mean. Why is rising and declining income inequality in Latin America important? Studies suggest that higher income inequality leads to higher poverty rates, less economic growth and resiliency to economic shocks, more violence, and exacerbated inequalities in democratic politics, educations, occupations, and health outcomes. Advocates of social welfare and social justice suggest inequality in itself is also bad because it violates contemporary global ethics and does not reflect the views of Latin American people whom overwhelmingly feel that such high inequality is unfair and unjust (de Ferranti et al. 2004; IADB 1998). Indeed, high rates of income inequality experienced in Latin America do not converge with universal human rights norms which declare rights to fair wages and stable incomes for all peoples. Below I discuss the results of the effects of industrialization, a global economy, national political arrangements, and universal human rights on national income inequality in 18 Latin American nations from 1980- 2008. Determinants of National Income Inequality in Latin America Model 1 in Table 8 includes the household income adjustment indicator and two historical period controls for the 1980s and 1990s. The household income equivalency adjustment, used to adjust for household size in GINI calculations, shows a significant and negative effect on income inequality. That is, when no adjustment for household income is made, it reduces inequality scores in Latin America. Neither historical control has any effect on income inequality in any models. Development or GDP per capita, shows a statistically significant but very modest downward effect on national income inequality, suggesting that increased economic output of a

nation simply does not lead to large decreases in inequality. Other logic of industrialism indicators such as rates of inflation and the prevalence of agricultural workers and production, show no association with income inequality. However, inflation is significant in Models 2 and 8, though the associated increase in income inequality is very small. Health and education spending levels are inconsistent in their directional effect and do not reach significance. Increases in social security and welfare spending levels, however, increase the GINI coefficient by 0.201% supporting research that suggests these types of spending are largely regressive in Latin America and do not benefit the growing numbers of workers employed in the informal sector (Huber et al. 2006). Social security and welfare spending levels lose significance temporarily with the introduction of social and demographic indicators but regain significance in Model 3 before losing it again in Model 8. The coefficients for social security and welfare spending increase to the 0.32-0.39 range in Models 3-7. Neoliberal globalization indicators of net trade levels and foreign direct investment have significant associations with national income inequality while pension privatization does not. As net trade levels increase 1% in Model 1 the GINI coefficient decrease by 0.018%. However this effect drops out in Model 2 and then becomes significantly associated with increases in income inequality beginning in Model 3 with the introduction of power resources. Indeed, once I introduce national politics and universal human rights in the models as trade levels increase 1% there is an associated 0.024-0.4% increase in the GINI. A 1 % increase in foreign direct investment is associated with a 0.64% increase in the GINI coefficient, supporting research that suggests these investments benefit only skilled workers and widen the income gap. The increase in income inequality from foreign investment holds across models even with the introduction of national and human rights political power resources though the coefficient is reduced some. The reduction in association between foreign investment and income inequality suggests that democracy, leftist politics, and human rights may mitigate the ill effects of foreign investment on income inequality. Pension privatization, measured dichotomously in this analysis, shows a negative impact on income inequality, but does not reach statistical significance in any of the models. As you can see from the results in Model 1 national economics, neoliberal globalization, and the welfare state influence national income inequality in different ways, though some of these effects change with the introduction of new perspectives. Overall, Model 1 explains 38% of the variation in national income inequality in this sample.

Model 2 introduces demographic and social indicators associated with the logic of industrialism perspective. I introduce these separately from the national economic indicators to show the separate effects social demographics have when compared to national and international economic factors. Studies suggest that increased youth populations lead to higher income inequality because they lack educational resources and formal employment, while urbanization increases national income inequality as unskilled rural workers seek employment in cities and stress the ability of urban areas to provide social programs and assistance to alleviate poverty, poor health, and increase educational access. My analysis partially supports this body of research. Increases in the percent of population aged 0-14 are not associated with the GINI coefficient in Model 2. However, it reaches significance in Models 7-8 though this is only significant at the .1 level in a one-tailed test. A 1% increase in youth populations in Models 7-8 is associated with a 0.4% and 0.37% increase in the GINI. A 1% increase in the percent of the population living in urban areas is associated with a 0.21% increase in the GINI. The effect is consistent across models and the coefficient ranges from 0.19-0.27. Percentages of the total population that completed secondary school, while signed in the correct direction, do not reach statistical significance in any models. Ethnic diversity is not statistically associated with income inequality and is only signed correctly in Models 3-8. In summary, growing populations of youth and urbanized workers lead to higher national income inequality while secondary school completion and ethnic diversity have no association. Model 2 explains 43% of the variation in income inequality in this sample of Latin American nations, a substantial improvement over using only economic factors. Model 3 introduces the national political power resources perspective. As discussed above, many studies show that national politics do influence income inequality in developed and developing nations. A history of democracy and Leftist legislative cabinets are political tools that can be used to counter-effect national and international economics and troublesome demographic conditions. As democracies in Latin America began to mature in the 1980s and 1990s, repressive regimes diminish, and Left political parties with progressive agendas that prioritize social justice and human rights gain power, national income inequality goes down. The effect from the cumulative strength of democracy from 1960 to the year of analysis is highly significant, but only shows a modest association with reductions in national income inequality. Repressive regimes, which become non-existent in this sample after 1990, show a negative

association with income inequality measured by the GINI, but this effect becomes nonsignificant with the inclusion of universal human rights as a new power resource. The cumulative legislative balance, however, decreases national income inequality across models. In Model 3, an increase in Left oriented political parties in the legislature is associated with a 0.48% reduction in the GINI coefficient, suggesting that Left parties and their emphasis on social justice and redistribution provide strong political tools at the national level to influence inequality across the region. The effect of Left cabinets holds across models that include universal human rights. In this sample of Latin American nations the greatest national political power resource for reducing inequality is found in Leftist cabinets, supporting a large body of research of nations in the developed world, and my own research on the positive effects of Left politics on social spending levels in Latin America. Model 3 explains 62% of the variation in Latin American income inequality. This is a substantial increase from Models 1-2 and suggests that national political arrangements and processes play very large roles in affecting national income inequality in Latin America. Models 4-8 introduce universal human rights as new political power resources anchored at the international level with the UN but institutionalized at the national level through international human rights treaty ratifications and NHRI adoption. I also include a measure of INGOs because of their important interactions with subnational, national, and international political structures and actors. Research on the effects of INGOs shows inconsistent relationships with treaty ratifications and human rights outcomes. Indeed, in another study I also find no association between INGO counts and social spending levels in this same sample and time frame (Shekha in progress). This analysis bears the same results. INGOs in Model 4 show a large increase in income inequality but are not statistically significant, though this temporarily changes in Models 6 and 9 with the inclusion of the treaty on the rights of the child and a total count of treaties ratified. I find this indicator to be an inadequate measure of international civil society because of the wide variation in INGO types from affinity groups, to business and trade associations, to Transnational Social Movement Organizations. NHRI adoption, which I hypothesized to have a negative association with income inequality, shows no association in any model and is incorrectly signed. NHRIs, while important political tools available to politicians and civil society, have limited powers and do not legislate or make judicial decisions in regards to social policy and problems focusing on inequality and poverty. In another study, however, I do

show that NHRI adoptions are associated with higher social spending levels in these same Latin American nations, suggesting they are not completely powerless. Many treaty ratifications, however, are significant and show strong associations with reductions in national income inequality. Ratifying the ICESCR, which contains special language extending rights to adequate wages and employment, is associated with a 3.38% reduction in the GINI coefficient, a very large effect. Similarly, ratifying the ICCPR, which focuses on the extension of civil and political rights to choice of employment, movement around the country, and political participation, is associated with a 3.43% reduction in the GINI coefficient. In the same fashion, ratifying the Convention on the Elimination of Discrimination Against Women (CEDAW), provides new political tools for women and mothers to fight employment, education, and health discrimination and is associated with a 2.49% reduction in the GINI. A newer treaty focusing on the rights of migrants is associated with a 1.51% reduction in the GINI coefficient. Finally, a full count of treaties ratified is associated with a 1.11% reduction in national income inequality in Latin America. All of the treaty ratifications are reported as one-tailed tests. In a robustness check using a robust cluster random effects model, all of the above treaties are highly significant in one or two tailed tests. Additionally, the treaty on children, while not significant in analysis using Hoechle’s xtscc command, is highly significant and strongly associated with reductions in income inequality. Other treaties not discussed here fail to reach significance in either analysis. International human rights treaties extend universal human rights to broad populations and represent new political tools available at the national level to influence social conditions and problems such as poverty and inequality. This analysis shows that while INGOs and NHRIs do not contribute to reductions in national income inequality, many treaty ratifications are associated with very large reductions in the GINI coefficient, the standard in measuring national income inequality. Models 4-5 explain 67% of the variation in Latin American income inequality while Models 6-8 explain between 63-64% of the variation. The amount of variation explained by universal human rights is larger than that explained by national power resources alone. The increase in income inequality variation explained from national power resources to universal human rights suggests my inclusion of human rights as determinants of Latin American income inequality is a useful and valid contribution to the literature on inequality.

CHAPTER 6: DISCUSSION

This study asks if extending national power resources to universal human rights through treaty ratification and NHRI adoption increases social spending levels and reduces national income inequality thereby extending national social rights to economic security, adequate healthcare, and quality educations in 18 Latin American nations from 1980-2008. It also asks if traditional welfare state theoretical approaches also adequately explain variation in social spending levels and national income inequality in Latin America. The answer is complex at best though all theoretical perspectives used in these studies find at least partial support, including universal human rights as new political power resources.

Explaining Social Spending Variation Late industrialization and demographic changes in Latin America exert significant effects over social spending patterns. However, these forces do not impact each type of spending the same. Social security and welfare spending seem unaffected by short-term economic and political effects, while health and education spending are more volatile. The resiliency of social security and welfare spending levels over time mean that variables must be operationalized in a cumulative fashion to better measure their effects. However, while social security and welfare spending are impervious to economic downturns, health spending is bolstered by increased GDP per capita and education spending decreased during periods of international economic liberalization. Aged populations have small positive effects in social security and welfare and health spending, while youth populations have similar effects on education spending levels.

Neoliberal globalization is the international liberalization of economic markets. Trade openness has a negative impact on health and social security and welfare spending, while it can increase education spending levels. Budgetary deficits show no associations with any of the spending types in this study. These findings show that globalization inspired by neoliberal doctrine mostly hurt social spending, making it more difficult for nations to protect social rights through spending, though education spending seems largely unaffected.

National politics in the forms of democratic strength, repressive regimes, and national legislative balance also continue to influence social spending levels in Latin America. Democratic strength shows small positive associations with education and social security and

welfare spending levels, but no association with health levels. Repressive regimes increase social spending levels across the board, but only within the context of universal human rights as a power resource for health and education spending levels. Indeed, repressive regimes already lead to increases in social security and welfare spending, but this effect grows with the inclusion of universal human rights. Legislative politics, shown consistently to have effects in studies of developed nations, only have an effect on health spending levels. While this effect is positive it is inconsistent and thus not the primary influence that increases health spending in Latin America.

The real story in Latin America over the last several decades is the extension of citizens’ social rights to income security, adequate healthcare, and quality educations through empowering national political actors and civil society members with a new power resource. Universal human rights are birthed at the international level and institutionalized at the national level through NHRI adoption and international human rights treaties ratification. T. H. Marshall suggests rights are linked to citizenship, and points to processes in British history where rights were extended to more and more groups of citizens in waves over the last few centuries. Universal human rights, centered at the United Nations, transcend the requirement of national citizenship and extend civil, political, and social rights to all peoples regardless of their social status or background. Yet, the nation is not obsolete. Nations still retain their sovereignty and must provide protection of rights to the peoples in its border. Extending national power resources to universal human rights drives this agenda, rectifies neglect, and corrects abuse through international human rights treaty ratification and National Human Rights Institution presence.

INGOs often attempt to provide protection of rights by applying both international and national pressure (Keck and Sikkink 1998). However, human rights theorists suggest their limited importance after a nation institutionalizes human rights norms (Risse and Sikkink 1999). Additionally, the INGO measure in this analysis does not distinguish between organizations representing human rights and those representing professional and business associations. Scholars need to develop a more nuanced indicator for INGOs to better measure their effects on social spending and other outcomes, and some have already begun (Smith and Wiest 2005).

NHRIs have positive effects on all types of spending. Their mandates are broad even if their powers are often limited. As democratic transitions in Latin America mature, NHRIs in the region increasingly concern themselves with violations of ESCRs. This move to protect

economic and social rights, combined with their focus on public administration, human rights abuses, and insider access to legislative and executive bodies makes them powerful advocates for the extension of social rights to income, healthcare, and educations. NHRIs are agents of the modern extension of universal human rights, and this is proven by their consistently positive effects on social security and welfare, health, and education spending levels in Latin American from 1980-2008.

The ICCPR and ICESCR have large positive effects on education and health spending, act as foundation for all following international human rights law, and explicitly mandate the protection of all people within a nation’s borders despite their origin,. However, certain treaties such as the CEDAW, CRC, and CMW seek to extend citizen’s rights to at-risk populations, and call upon nations to protect citizens and residents historically devoid of civil and political power. The call to protect migrant workers, often displaced from their home country, is evidenced by the positive effect of the CMW on education spending levels in Latin America. Other treaties such as CRC, CEDAW, and ICCPR OP also help to extend social rights by increasing health and education spending levels. With the exception of social security and welfare spending, treaty and Optional Protocol ratification leads to increased social spending levels, suggesting them to be stronger legal agreements than some studies indicate (Hathaway 2002; Vreeland 2008). However, the negative effects of treaty ratification on social security and welfare spending levels call to question the ability of the universal human rights regime to enforce the extension of all social rights equally.

Nations committed to the continuing extension of social rights embrace universal human rights as new political power resources that can be used to increase social spending levels. Yet the negative effects of treaties on social security and welfare spending points to important limits of international human rights law. NHRIs seek to compensate for this, but cannot bear the weight of this burden alone, and in this sample INGOs have no effect. Knowing that human rights treaties and optional protocols do not consistently exert positive influences on social spending levels in contemporary Latin America shows that much work needs to be done in fulfilling the 4th stage in T. H. Marshall’s search for rights.

As I alluded to above, this study is not without its weaknesses. The primary improvement can be made in terms of operationalizing human rights indicators. United Nations treaty

ratifications, while great markers of historical embracement of universal human rights, do not fully capture the commitment a nation makes to extending civil, political, and social rights to all peoples within its borders. The International Labor Organization (ILO) also issues international treaties protecting and promoting human rights. Latin American nations participate in these treaties, and so future studies should be careful to include ILO treaties. Another way to better capture this process is to look to national laws that include language from international human rights treaties. For example, Argentina constitutionalized international human rights treaties thereby ensuring their careful consideration in political and legal matters. Dichotomous and cumulative measures of NHRIs may also be improved. Though they converge in many ways in Latin America, these important institutions can vary in terms of leadership, state control, budgets, staff, powers, and mandate. More comprehensive measures that capture all or some of these aspects of NHRI offices should be created to better understand their influence over social spending, policy, and problems. Finally, INGOs and TSMOs, important actors in international civil society, can exert great influence over national policies. My measure of INGOs proved to be inadequate in this study because it is too general. In other words, simple counts of INGOs do not reflect the many different types, of which TSMOs are a small piece. While some have begun strengthening these measures (Tsutsui and Wotipka 2004; Smith and Wiest 2005), I will continue the work begun under my own National Science Foundation dissertation grant to work closely with the Yearbook of International Organization staff to improve the utility of INGO measures.

Explaining National Income Inequality Income inequality in Latin America is among the worst in the world, followed only by sub- Saharan Africa (de Ferranti et al. 2004; IADB 1998). In this dissertation I ask if universal human rights in the form of treaty ratifications and NHRI adoptions act as a new political power resource and are associated with lower levels of national income inequality in 18 Latin American nations from 1980-2008. I also ask if more traditional welfare state theories explain national income inequality in Latin America. Methodologically I innovate by combining new approaches to pooled time-series OLS modeling that correct for heteroskedasticity, autocorrelation, and cross-sectional dependence with multiple imputation techniques to handle large amounts of missing data using command packages found in STATA 11. The economy, globalization, and the welfare state do affect national income inequality estimates. National economic development, or GDP per capita, showed only modest negative

associations with income inequality. Inflation, thought to increase income inequality as it prices commodities out of the reach of lower earners, is positively associated with rising income inequality though the effect is negligible and only reaches significance in two models. Welfare state expenditures and policy, which can be used as a tool to reinforce or reduce income inequality, show inconsistent associations in Latin America. While health and education spending as a percentage of GDP show no association, social security and welfare spending levels have a positive association with increased income inequality in Latin America in most models. Given their largely regressive nature in developed and developing nations, the positive effect of social security and welfare spending on income inequality in Latin America is expected. One study, however, shows that an interaction between historical levels of democracy and these types of spending results in a statistically significant negative association with income inequality in a similar sample of Latin American and Caribbean nations from 1970-2000 suggesting that welfare state expenditures and national regime type to be linked (Huber et al. 2006). Having a dual sector economy where agriculture represents a large portion of economic productivity and employment seems to have no effect on income inequality. Amongst globalization indicators that show the degree of openness and integration a national economy has into the international economic arena predict higher levels of national income inequality and support past research on income inequality in Latin America and the developing world. Both net trade levels and foreign direct investment as a percentage of GDP are associated with higher inequality, though foreign investment has a much stronger association. Pension privatization, meant to reform pension systems in Latin America by increasing access and coverage show no association with income inequality in Latin America. Pension reform, however, is new for the region and the effects, good or ill, will only be fully realized in the future. Demographic transitions and societal conditions do influence national income inequality in Latin America. Demographic transitions such as the increased percentage of populations under 14 years of age and the increasing urbanization of Latin America are associated with higher levels of income inequality. Urbanization, however, is a stronger and more consistent predictor of income inequality in Latin American and supports previous research suggesting that this process exacerbates poverty rates and inequality. Ethnic diversity or heterogeneity, measured as a time invariant indicator of high versus low levels of indigenous, African descended, and mixed ethnicity population proportions, predict higher levels of income inequality but fail to reach

statistical significance in this sample. Secondary school completion percentages of the total population consistently predict lower levels of national income inequality but fail to reach statistical significance and thus do not support research in this area. Overall the demographic transitions in Latin America discussed above either have no association with, or lead to, higher levels of income inequality. National political power resources play an important role in shaping national income inequality in Latin America. A strong and stable history of democracy predicts lower levels of inequality in this sample. Compared to a previous study of a similar sample (Huber et al. 2006) this is a new development in the effects of democratic history on income inequality in Latin America. However, the association is not strong and decreases as I introduce universal human rights indicators. Repressive regimes, thought to increase income inequality, show no association with national income inequality after I introduce universal human rights to the models suggesting that repressive regimes cannot subvert the new political processes represented by the introduction of universal human rights into the national political landscape. Additionally, repressive regimes decline in this sample as human rights treaties are ratified. The real success of the traditional power resources perspective found in this study is the cumulative balance of legislative cabinets on a left-center-right spectrum. A longer history of Left based parties in the legislature predicts lower levels of national income inequality in these 18 Latin American nations from 1980-2008. Indeed, many Leftist political parties in Latin America profess a commitment to social justice and human rights values as found in their party platforms and documentation. Left and Center-Left legislative cabinets represent very strong political tools for politicians and civil society to work through to reduce national income inequality in the region. I find that extending power resources to universal human rights for predicting national income inequality estimates to be inconsistent. NHRI adoption and INGO presence have no statistically significant associations in this sample. Only treaty ratifications and a count of total human rights treaties ratified are associated with lower levels of national income inequality. However, not all treaties and none of the Optional Protocol measures are associated with lower levels of in inequality. The negative associations with increased income inequality, however, are very strong. Ratifying the ICESCR and ICCPR, longtime foundations of international human rights law, is associated with an approximately 3.5% reduction in national income inequality. While these indicators are measured dichotomously, a count of up to 10 international human

rights treaties ratified also predicts lower levels of national income inequality. While NHRIs and INGOs do not represent the powerful political tools hypothesized, treaty ratifications act as important political tools to be used in a continual process of social policy negotiations meant to reduce social problems such as poverty and income inequality. To that end, extending power resource to include treaty ratifications makes sense in terms of predicting lower levels of national income inequality. This study is a strong analysis of the determinants of income inequality in Latin America. I show convincingly, and with new innovations for handing missing data and pooled time-series error structures, that the economy, globalization, the welfare state, demographics and society, national political power resources, and its extension to universal human rights all have varied influences on national income inequality in 18 Latin American nations from 1980-2008. In this study, I show that extending the power resources approach to universal human rights, with their dual emphasis on the rights of the citizen and the person, provides more positive proof that extending human rights through treaties helps to fulfill obligations to social rights such as adequate incomes made by nations in Latin America that embrace democratic ideals and processes. The economy, globalization, welfare policy, and demographic and societal conditions are not rendered obsolete by the inclusion of national political power resources and universal human rights in my analysis. However, political processes, whether anchored nationally or internationally, represent the best power resources available for politicians and civil society to continue extending important social and human rights to fair employment, wages, and income security in Latin America during the contemporary period. Future Research Directions My dissertation is the beginning of a larger research agenda focusing on universal human rights, citizenship, the state, globalization, and civil society in Latin America and the developing world. Future studies will expand my data to include individual and household level indicators and focus on social spending allocation, quality and availability of healthcare services, coverage and benefit levels of social security and welfare, as well as health, education, labor, and related outcomes. I expect to find that the project of extending social rights by embracing universal human rights as new political power resources continues to show positive influences in the arenas of economic security, adequate healthcare, and quality educations. Theoretically, I seek to integrate my extension to power resources with the world society perspective because they both

seek to explain political and social processes in terms of political power and process. While I did not draw explicitly on world society theory to inform my analysis, much of the work done in the sociology of human rights is grounded in world society perspectives and thus calls for reconciliation between welfare state theories and theories linking international politics and the state such as power resources, world society, and world systems.

APPENDIX A: TABLES, BOXES AND GRAPHS

Table 1: Adoption of NHRIs in Latin America

Country Year Argentina 1993 Bolivia 1997 Brazil 1998 Chile None Colombia 1992 Costa Rica 1992 Dominican Republic None Ecuador 1998 El Salvador 1998 Honduras 1985 Guatemala 1992 Mexico 1992 Nicaragua 1999 Panama 1995 Paraguay 1992 Peru 1994 Uruguay None Venezuela 1999

Table 2: Description of Variables Variable Description Dependent Variables Social Security & Welfare spending Social security and welfare spending as a % of GDP. Health spending Health spending as a % of GDP. Education spending Education spending as a % of GDP. GINI .00-.1 coefficient of national income inequality Independent Variables Logic of Industrialism Development Logged GDP per capita. Inflation Annual percentage rise in cost of goods/services. Sector Dualism Absolute difference in agricultural employment production. Aged Population Logged % population aged 65 and over. Youth Population % population aged 14 and under. Urbanization % population living in urban areas. Ethnic Diversity Coded 1 if <20% or >80% ethnic/indigenous Secondary School % population completed secondary school Economy/ Globalization Trade Logged total imports and exports as a percentage of GDP. Foreign Investment Logged net inflows of FDI as a percentage of GDP. Deficit Net revenue, expenditures, grants, loans, and repayments % of GDP. Pension Reform Scored 0 for no privatization and 1 for privatized. Power Resources Democracy (cum) Cumulative Polity IV score -10 to +10 from 1945. Repressive Regime (cum) Previous 5 year cumulative score of Freedom House 'Not Free' (5.5+). Legislative Balance (cum) Weighted proportion of seats held by Left-Right blocs, cumulative from 1960. Human Rights INGOs (ln) Logged count of international nongovernmental organizations. HR Treaties Count of 9 treaties ratified. Optional Protocols Count of 7 OPs ratified. ICESCR 0/1 ratified, or years since ratification. ICCPR 0/1 ratified, or years since ratification. 99

Table 2 Continued: Description of Variables Variable Description Independent Variables Human Rights ICCPROP 0/1 ratified, or years since ratification. CEDAW 0/1 ratified, or years since ratification. CHILD 0/1 ratified, or years since ratification. INDIGENOUS 0/1 ratified, or years since ratification. MIGRANT 0/1 ratified, or years since ratification. NHRIs 0/1 present, or years since adoption.

100

Table 3: Summary Statistics after 100 Imputations Variable Obs. Mean Std. Dev. Min Max GINI 28684 52.1 5.34 38.7 65.8 Household Equivalency 23735 0.06 0.23 0 1 Debt Crisis 52722 0.28 0.45 0 1 Neoliberal 52722 0.38 0.49 0 1 Development 52722 3042.31 1958.4 632.52 9893.81 Inflation 49972 83.72 720.11 -3154.42 11749.64 Sector Dualism 40976 12.55 15.72 -535.94 143.31 Trade(ln) 52621 3.94 0.54 2.45 5.29 Foreign Invest 52419 2.31 2.54 -12.21 14.92 Health & Educ 52318 5.8 2.4 1.3 13.4 Social Sec & Welfare 49181 3.96 3.71 -5.21 16.06 Pension Reform 52722 0.26 0.44 0 1 Youth Population 52722 36.25 5.96 23.02 47.2 Urbanization 52722 64.73 15.67 34.9 93.32 Secondary School 29208 14.83 19.05 -446.06 107.36 Ethnic Diversity 52722 0.28 0.45 0 1 Democracy (cum) 52722 92.32 355.27 -346 1446 Legislative Balance 52520 -0.97 4.61 -10.14 15.6 (cum) Repressive (cum) 52722 0.39 1.07 0 5 ICESCR 52722 0.95 0.27 0 1 ICCPR 52722 0.88 0.32 0 1 CEDAW 52722 0.88 0.32 0 1 MIGRANT 52722 0.16 0.36 0 1 HR Ratify 52722 5.17 1.95 0 10

101

Table 4: Determinants of Social Security and Welfare Spending in Latin America, 1980-2008 1 2 3 4 Debt Crisis -0.836 -0.975 -0.131 0.092 (0.502) (0.604) (0.438) (0.484) Neoliberal Decade -0.569 -0.607+ 0.030 0.325 (0.340) (0.346) (0.224) (0.348) Development -0.000+ -0.001* -0.001** -0.001+ (0.000) (0.000) (0.000) (0.000) Urban Pop(ln) 4.180** 4.249** 0.606 -4.012* (1.554) (1.573) (1.994) (2.023) Aged Pop(ln) 1.014 2.344* 1.668 2.202* (1.084) (1.178) (1.285) (1.189) Trade(ln) -1.053* -0.885* -1.274** (0.546) (0.431) (0.469) Foreign Investment 0.015 -0.019 0.020 (0.022) (0.030) (0.023) Deficit 0.878*** 0.498+ 0.526 (0.223) (0.263) (0.467) Pension Private(sum) -0.022 -0.012 -0.047+ (0.035) (0.032) (0.035) Democracy(sum) 0.015*** 0.012* (0.003) (0.006) Repressive(sum) 0.272** 0.351*** (0.087) (0.035) Legislative Balance(sum) -0.080 0.011 (0.048) (0.071) INGOs(ln) 0.624 (0.712) NHRIs(sum) 0.122*** (0.019) Constant -12.971 -10.466 3.081 17.578 (6.720) (6.716) (8.661) (9.885) R-squared 0.633 0.636 0.694 0.699 N 481 458 458 409 +p<0.1,* p<0.05,** p<0.01,*** p<0.001|One-tailed: GDP Urban Aged Trade Deficit Pension Democracy

102

Table 4 Continued: Determinants of Social Security and Welfare Spending in Latin America, 1980-2008 5 6 7 8 9 Debt Crisis 0.197 -0.213 0.268 0.145 0.332 (0.329) (0.604) (0.431) (0.408) (0.489) Neoliberal Decade 0.291 0.272 0.415 0.219 0.550 (0.205) (0.361) (0.314) (0.291) (0.376) Development -0.001+ -0.001 -0.001+ -0.001+ -0.001* (0.000) (0.000) (0.000) (0.000) (0.000) Urban Pop(ln) -2.499+ -0.935 -3.162+ -2.289 -3.844* (1.685) (2.152) (2.053) (1.995) (1.958) Aged Pop(ln) 3.353** 5.512** 2.814* 2.738* 2.736* (1.148) (1.580) (1.190) (1.278) (1.349) Trade(ln) -1.159** -1.263** -1.210** -1.272** -1.331** (0.419) (0.464) (0.475) (0.487) (0.426) Foreign Investment 0.001 0.032+ 0.002 0.030+ 0.004 (0.020) (0.024) (0.022) (0.022) (0.022) Deficit 0.472 0.424 0.493 0.447 0.392 (0.427) (0.481) (0.468) (0.491) (0.499) Pension Private(cum) -0.015 0.004 -0.035 0.013 0.030 (0.031) (0.045) (0.032) (0.044) (0.042) Democracy(cum) 0.016*** 0.017*** 0.014** 0.013** 0.015** (0.004) (0.004) (0.005) (0.005) (0.006) Repressive(cum) 0.381*** 0.341*** 0.370*** 0.397*** 0.454*** (0.042) (0.036) (0.035) (0.047) (0.044) Legislative Bal(cum) -0.022 0.036 0.016 -0.028 -0.031 (0.062) (0.060) (0.073) (0.060) (0.072) INGOs(ln) -0.102 0.806 0.269 0.378 0.152 (0.498) (0.907) (0.582) (0.600) (0.647) NHRIs(cum) 0.165*** 0.159*** 0.138*** 0.165*** 0.152*** (0.022) (0.020) (0.020) (0.023) (0.017) Optional Protocols -0.353*** (0.053) ICESCR(cum) -0.179** (0.048) CEDAW OP(cum) -0.115*** (0.024) INDIGENOUS(cum) -0.128** (0.039) MIGRANT(cum) -0.199*** (0.036) Constant 13.673 0.296 15.000 11.239 19.223 (8.251) (10.561) (9.717) (9.876) (9.358) R-squared(OLS) 0.704 0.779 0.701 0.699 0.706 N 409 409 409 409 409 +p<0.1,* p<0.05, ** p<0.01, *** p<0.001 | One-tailed: GDP Urban Aged Trade Deficit Pension Democracy

103

Table 5: Determinants of Health Spending Levels in Latin America, 1980-2008 1 2 3 4 5 Debt Crisis -0.088 -0.022 -0.061 0.187 0.593 (0.331) (0.365) (0.382) (0.314) (0.352) Neoliberal Decade -0.166 -0.067 -0.051 -0.004 0.050 (0.120) (0.137) (0.156) (0.110) (0.171) Development 0.000+ 0.000* 0.000* 0.000*** 0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) Urban Pop(ln) -0.347 0.478 0.357 -0.275 -1.761+ (0.565) (1.012) (0.883) (0.954) (1.180) Aged Pop(ln) 1.779* 2.462** 2.315* 1.321* 0.386 (0.799) (0.865) (0.868) (0.497) (0.801) Trade(ln) -0.046 -0.009 -0.251 -0.676** (0.214) (0.212) (0.232) (0.218) Foreign Investment -0.096* -0.094+ -0.095* -0.077* (0.056) (0.056) (0.049) (0.040) Deficit -0.672 -0.726 -0.477 -0.371 (0.583) (0.596) (0.466) (0.354) Democracy 0.000 -0.001 -0.003* (0.001) (0.002) (0.001) Repressive Regime 0.261 1.038** 1.223*** (0.303) (0.379) (0.316) Legislative Balance 0.215* 0.229 0.423+ (0.118) (0.227) (0.252) INGOs(ln) 0.510 0.475 (0.717) (0.655) NHRIs 0.384* 0.387** (0.154) (0.117) HR Treaties 0.371*** (0.085) Constant 0.559 -3.803 -3.282 -2.423 5.140 (2.833) (4.805) (4.370) (2.661) (3.078) R-squared 0.176 0.167 0.202 0.255 0.268 N 519 489 489 437 437 + p< .1 * p<0.05, ** p<0.01, *** p<0.001 | par(std err) One-tailed: GDP Urban Aged Trade Deficit Pension Democracy

104

Table 5 Continued: Determinants of Health Spending in Latin America, 1980-2008 6 7 8 9 10 11 Debt Crisis 0.112 0.125 0.276 0.178 0.202 0.219 (0.294) (0.320) (0.332) (0.290) (0.320) (0.343) Neoliberal Decade -0.102 -0.073 0.121 -0.000 0.033 0.073 (0.103) (0.100) (0.154) (0.116) (0.120) (0.150) Development 0.001*** 0.001*** 0.000*** 0.000*** 0.000*** 0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Urban Pop(ln) -0.835 -1.215 -1.383+ -1.090+ -0.310 0.084 (0.995) (1.086) (1.022) (0.773) (0.915) (0.952) Aged Pop(ln) 1.311** 1.209** 1.093** 2.694** 1.383** 0.835+ (0.485) (0.475) (0.510) (0.781) (0.447) (0.573) Trade(ln) -0.435* -0.512* -0.369+ -0.244 -0.287 -0.339+ (0.229) (0.230) (0.224) (0.232) (0.236) (0.232) Foreign Investment -0.078* -0.085* -0.090* -0.0928 -0.102* -0.094* (0.045) (0.043) (0.045) (0.050) (0.050) (0.049) Deficit -0.480 -0.481 -0.443 -0.371 -0.459 -0.460 (0.477) (0.468) (0.425) (0.369) (0.463) (0.436) Democracy -0.001 -0.002 0.002 -0.001 -0.001 -0.001 (0.002) (0.002) (0.002) (0.001) (0.002) (0.002) Repressive Regime 1.247** 1.157** 1.162** 1.240** 1.038* 1.005* (0.365) (0.343) (0.340) (0.362) (0.376) (0.386) Legislative Balance 0.220 0.285 0.307+ 0.291 0.245 0.243 (0.214) (0.222) (0.231) (0.226) (0.234) (0.236) INGOs(ln) 0.367 0.465 0.326 -0.008 0.403 0.548 (0.680) (0.721) (0.702) (0.668) (0.704) (0.724) NHRIs 0.345* 0.395* 0.458** 0.443** 0.311+ 0.385* (0.154) (0.144) (0.146) (0.143) (0.167) (0.149) ICESCR 1.238*** (0.198) ICCPR 0.946** (0.269) ICCPR OP 0.764** (0.252) CEDAW 1.119*** (0.141) INDIGENOUS 0.247 (0.121) MIGRANT 0.397*** (0.101) Constant 0.328 1.953 3.749 1.495 -1.503 -3.057 (2.716) (3.144) (2.685) (3.098) (2.540) (2.977) R-squared 0.286 0.284 0.268 0.259 0.265 0.255 N 437 437 437 437 437 437

105

Table 6: Determinants of Education Spending in Latin America, 1980-2008 1 2 3 4 5 Debt Crisis -0.390 -0.186 0.141 0.387 0.823* (0.421) (0.375) (0.385) (0.315) (0.358) Neoliberal Decade -0.795* -0.666* -0.490* -0.288+ -0.175 (0.324) (0.263) (0.239) (0.153) (0.174) Development 0.000 0.000 0.000 0.000+ 0.000+ (0.000) (0.000) (0.000) (0.000) (0.000) Urban Pop(ln) 4.426*** 4.781*** 4.134*** 4.057*** 2.764** (1.030) (1.003) (0.825) (0.976) (0.967) Youth Pop(ln) 1.857 2.142+ 2.736** 2.569** 2.666** (1.681) (1.398) (1.183) (0.869) (0.999) Trade(ln) 0.474+ 0.531+ 0.299 -0.074 (0.303) (0.272) (0.228) (0.217) Foreign Investment 0.009 -0.000 -0.032 -0.017 (0.035) (0.033) (0.032) (0.028) Deficit -0.586 -0.531 -0.530 -0.399 (0.632) (0.599) (0.624) (0.512) Democracy 0.005*** 0.005** 0.003+ (0.001) (0.002) (0.002) Repressive Regime 0.050 0.253 0.397 (0.138) (0.277) (0.231) Legislative Balance -0.260 -0.312 -0.191 (0.213) (0.256) (0.251) INGOs(ln) -0.097 -0.276 (0.272) (0.241) NHRIs 0.622*** 0.626*** (0.100) (0.103) HR Treaties 0.324*** (0.069) Constant -21.292** -25.734*** -25.886*** -24.143*** -18.180** (6.343) (5.009) (3.970) (3.545) (5.471) R-squared 0.103 0.167 0.174 0.244 0.230 N 520 491 491 439 439 +p<0.1, * p<0.05, ** p<0.01, *** p<0.001 | One-tailed: GDP Urban Youth Trade Foreign Invest

106

Table 6 Continued: Determinants of Education Spending in Latin America, 1980-2008 6 7 8 9 10 11 Debt Crisis 0.306 0.372 0.525* 0.344 0.394 0.422 (0.368) (0.296) (0.251) (0.344) (0.285) (0.308) Neoliberal Decade -0.397* -0.310* -0.117 -0.293+ -0.262* -0.237+ (0.172) (0.138) (0.128) (0.167) (0.131) (0.162) Development 0.000* 0.000** -0.000 0.000 0.000+ 0.000+ (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Urban Pop(ln) 3.351*** 2.776** 2.973** 3.376*** 4.024*** (0.939) (1.135) (1.137) (0.803) (0.947) (0.960) Youth Pop(ln) 2.312* 1.381* 2.067* 0.579 2.591** 2.690** (0.959) (0.748) (0.834) (0.914) (0.861) (0.862) Trade(ln) 0.061 -0.077 0.181 0.307+ 0.268 0.257 (0.239) (0.210) (0.206) (0.229) (0.231) (0.224) Foreign Investment -0.011 -0.019 -0.028 -0.030 -0.038 -0.032 (0.028) (0.026) (0.028) (0.033) (0.032) (0.032) Deficit -0.562 -0.567 -0.495 -0.428 -0.516 -0.509 (0.661) (0.658) (0.603) (0.537) (0.622) (0.602) Democracy 0.004** 0.004* 0.007*** 0.004** 0.005** 0.005** (0.002) (0.002) (0.001) (0.002) (0.002) (0.002) Repressive Regime 0.516* 0.383+ 0.356+ 0.394+ 0.255 0.235 (0.249) (0.224) (0.225) (0.259) (0.274) (0.282) Legislative Balance -0.324 -0.242 -0.248 -0.190 -0.295 -0.319 (0.216) (0.235) (0.262) (0.259) (0.266) (0.256) INGOs(ln) -0.312 -0.317 -0.377 -0.504 -0.177 -0.112 (0.233) (0.244) (0.225) (0.295) (0.254) (0.271) NHRIs 0.566*** 0.607*** 0.683*** 0.627*** 0.559*** (0.096) (0.074) (0.098) (0.103) (0.077) (0.099) ICESCR 1.604*** (0.281) ICCPR 1.367*** (0.256) ICCPR OP 0.768** (0.215) CEDAW 0.958*** (0.190) INDIGENOUS 0.216+ (0.126) MIGRANT 0.191* (0.090) Constant -19.542*** -13.011*** -15.942** -12.001** -23.413*** - (2.867) (3.486) (5.096) (4.003) (3.870) (3.577) R-squared 0.260 0.236 0.229 0.228 0.237 0.228 N 439 439 439 439 439 439 +p<0.1, * p<0.05, ** p<0.01, *** p<0.001 | One-tailed: GDP Urban Youth Trade Foreign Invest 107

Table # 7: Number of Repressive Regimes by Year Year Absent Present Total 1980 14 4 18 1981 13 5 18 1982 13 5 18 1983 15 3 18 1984 16 2 18 1985 17 1 18 1986 15 3 18 1987 16 2 18 1988 16 2 18 1989 17 1 18 1990 18 0 18 1991 18 0 18 1992 17 1 18 1993 18 0 18 1994 18 0 18 1995 18 0 18 1996 18 0 18 1997 18 0 18 1998 18 0 18 1999 18 0 18 2000 18 0 18 2001 18 0 18 2002 18 0 18 2003 18 0 18 2004 18 0 18 2005 18 0 18 2006 18 0 18 2007 18 0 18 2008 18 0 18

108

Table 8: Determinants of National Income Inequality in 18 Latin American Countries, 1980-2008 Using Multiple Imputations and Driscoll-Kraay Standard Errors 1 2 3 4 5 6 7 8 No Adjustment -6.867*** -5.870** -3.323 -1.768 -1.782 -2.417 -2.126 -2.073 (1.608) (1.810) (1.984) (1.715) (1.704) (1.849) (2.006) (1.786) Debt Crisis -0.168 -0.726 -0.491 -0.009 -0.018 0.105 -0.078 -1.187 (0.577) (1.035) (0.949) (1.786) (1.731) (1.868) (1.850) (2.050) Neoliberal Decade -0.402 -0.567 -0.762 -0.271 -0.243 -0.490 -1.014 -0.661 (0.662) (0.883) (0.765) (0.887) (0.893) (0.926) (0.998) (0.811) Development -0.001*** -0.002*** -0.001*** -0.002*** -0.002*** -0.001*** -0.002*** -0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Inflation 0.001 0.001+ 0.000 0.000 0.000 0.001* 0.000 0.001+ (0.001) (0.001) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Health/Education -0.047 0.167 0.114 0.146 0.148 -0.055 0.011 0.147 (0.100) (0.153) (0.137) (0.206) (0.201) (0.147) (0.155) (0.180) Soc. Security/Welfare 0.201* 0.140 0.394* 0.334* 0.328* 0.398* 0.362+ 0.246 (0.075) (0.162) (0.158) (0.159) (0.159) (0.164) (0.185) (0.197) Dual Sector 0.100 0.162 0.059 0.044 0.050 0.056 0.071 0.062 (0.125) (0.148) (0.099) (0.093) (0.090) (0.096) (0.105) (0.096) Trade(ln) -1.785** -0.214 2.479* 3.515*** 3.513*** 2.922** 2.806** 3.960*** (0.592) (0.762) (0.899) (0.866) (0.881) (0.976) (0.921) (0.813) Foreign Investment 0.631*** 0.522*** 0.543*** 0.455** 0.451** 0.530*** 0.524*** 0.478*** (0.115) (0.138) (0.143) (0.127) (0.128) (0.132) (0.134) (0.120) Pension Reform -0.146 -0.092 -0.443 -0.464 -0.453 -0.467 -0.113 -0.183 (0.928) (1.153) (0.648) (0.642) (0.632) (0.663) (0.603) (0.566) Youth Population 0.343 0.313 0.328 0.327 0.358 0.398+ 0.371+ (0.203) (0.247) (0.225) (0.224) (0.231) (0.231) (0.197) Urbanization 0.205* 0.254** 0.244* 0.245* 0.191* 0.210* 0.267** (0.082) (0.086) (0.097) (0.095) (0.089) (0.093) (0.094) Secondary School -0.009 -0.081 -0.074 -0.071 -0.083 -0.079 -0.073 (0.081) (0.062) (0.059) (0.059) (0.065) (0.063) (0.059) Ethnic Diversity -0.451 0.995 0.692 0.658 0.034 0.235 1.036 (0.879) (0.866) (1.112) (1.089) (0.964) (1.003) (1.121) Democracy(cum) -0.004*** -0.004** -0.004** -0.003** -0.003* -0.002+ (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Legislative Balance(cum) -0.477*** -0.422** -0.419** -0.403** -0.435*** -0.488*** (0.121) (0.118) (0.117) (0.118) (0.114) (0.112) Repressive(cum) -0.605* -0.447 -0.450 -0.547 -0.351 -0.549 (0.222) (0.409) (0.411) (0.366) (0.398) (0.356) INGOs(ln) 2.841 2.817 3.591 3.134 3.712+ (2.399) (2.377) (2.254) (2.437) (1.959) NHRIs 0.052 0.036 0.243 0.167 0.393 (0.575) (0.575) (0.540) (0.559) (0.516) Ratify ICESCR -3.380* (1.820) Ratify ICCPR -3.426* (1.823)

109

Table 8 Continued: Determinants of National Income Inequality in 18 Latin American Countries, 1980-2008 Using Multiple Imputations and Driscoll-Kraay Standard Errors 1 2 3 4 5 6 7 8 CEDAW -2.487+ (1.643) MIGRANT -1.505+ (0.956) HR Treaties -1.108** (0.453) Constant 61.284*** 29.137* 16.568 -3.194 -3.155 -4.148 -5.600 -11.383 (4.060) (12.696) (15.745) (21.679) (21.586) (21.990) (22.050) (19.079) R-squared 0.38 0.43 0.62 0.67 0.67 0.63 0.63 0.64 N 234 234 234 234 234 234 234 234 + p<0.1 * p<0.05, ** p<0.01, *** p<0.001 | One-tailed test: Inflation, Youth, ICESCR, ICCPR, CEDAW, MIGRANT, HR Treaties | 100 imputations

110

Box 1: Missing Variables and Imputations Analysis of Latin American Income Inequality

Variable Percent Missing Complete Observations Missing Observations Imputed Total GINI 45.6 284 238 0 522 Secondary School 79.3 108 414 183 522 Sector Dualism 28 376 146 30 522 INGOs(ln) 10.5 467 55 14 522 Inflation 9.6 472 50 23 522 Social Sec & Welfare 7.9 481 41 6 522 Health and Education 0.8 NA 4 NA 522 Foreign Investment 0.6 NA 3 NA 522

111

112

113

114

115

116

117

APPENDIX B: POLITICAL CODING REPORT FROM RESEARCH ASSISTANT

The following is a report drafted by my research assistant, Edward Martí Kring. I received funds from my National Science Foundation Sociology Dissertation Improvement Grant (SES 1003012) to hire an undergraduate research assistant to help collect and code Latin American political party data using English and Spanish sources. Edward is a native Spanish speaker and was an invaluable asset in helping collect and code data. Below is the report he drafted that I am submitting to the National Science Foundation. It details his time on the project and types of work he performed.

“My name is Edward Martí Kring and I worked under the direction of Kaiser Shekha as his Research Assistant (RA). During the time of my service as Kaiser’s RA I was completing my bachelor’s degree at the Florida State University in Political Science and International affairs with a concentration in Sociology. My participation as an RA lasted from November 2010, through May 2011 and I was paid through funds from an NSA grant of $800 dollars awarded to Kaiser Shekha. Furthermore, I worked a minimum of 8 hours per week and when my scheduled permitted, I worked up to 13 hours per week.

Our research consisted of collecting data on Political Parties from seven Latin American countries: Dominican Republic, El Salvador, Guatemala, Honduras, Nicaragua, Panama and Paraguay. The goal was to collect data not available from other sources that would help Kaiser Shekha code each of the political parties on spectrum of Left to Right and Religious to Secular. Multiple sources were investigated in order to triangulate data and the variables investigated involved polity party: founding, History, Ideology, Issues/Platform, Religious/Secular principals and any other extraneous variables else that would seem important (i.e., regime change, coups d’état, etc).

Moreover, in order to strengthen the reliability and accuracy of the data collected, as well as to ensure a high level of confidence in our research results we used multiple sources to triangulate our data. The web sources used include: the US State Department, Georgetown University’s Political Database of the Americas, party specific websites, electionmeter.com,

118

britanica.com, truman.edu, nationalencyclopedia.com, archivochile.com, Country-data.com, unchr.org, democracynow.org, asamble.gov, cig.gov. We used many books containing political data in Latin America including the Political Handbook of the World (2008), A political and economic dictionary of Latin America by Peter Calvert, Europa World Year Book 2, by Taylor & Francis Group, South America, Central America and the Caribbean 2003 By Europa Publications, Alexander (1988), Ameringer (1992), Latin American political parties by Robert J. Alexander, Party politics and Elections in Latin America by Ronald H. McDonald, J. Mark Ruhl, Latin American Party Systems By Herbert Kitschelt et al., Building democratic institutions: party systems in Latin America edited by Scott Mainwaring and Timothy R. Scully, and Christian Democracy in Latin America : electoral competition and regime conflicts edited by Scott Mainwaring and Timothy R. Scully.

My methods for reporting consists of a master document, single country documents and an excel spreadsheet codebook. Since most of the documents researched were in Spanish translation was required. I used direct translation as the main source of my translation and Google translate as an auxiliary tool. When I had questions or inconsistencies I addressed them by consulting with Kaiser Shekha for direction, there were also times when we collaboratively decided the best solution. For example, I had an problem where the information for the Partido Acción Nacional (PAN) in Guatemala kept leading me to a party in Mexico (a country that was not part of the study). This was resolved by Kaiser stating that he too “think[s] of PAN as being Mexican” and this was verified by reading in the sources above about Mexican politics and the PAN. We then decided not to include the party in the research for the country in question.

In conclusion my role as a research assistant has been an invaluable learning experience that has enhanced my education at Florida State by providing me with real world research application. I got the opportunity to work with Kaiser Shekha, a talented and passionate individual. I was and continued to be inspired by his intelligence and dedication to academia and the benefits it presents to humanity in our quest for knowledge. I was able to use my Spanish language skills to use and I learned to use triangulation as a method to strengthen research. Overall this opportunity has helped me with my own professional development by empowering me to find the information needed to answer a research question. Finally, working with Kaiser Shekha has been a life changing experience that has taught me the importance of people looking

119

to the past to learn how from it in order to make a better future for the generations that follow. I’m eternally thankful for the opportunity to contribute in someone way to this important research.”

120

REFERENCES CITED

Aaron, Henry. 1967. "Social Security: International Comparisons." Pp. 13-48 in Studies in the Economics of Income Maintenance, edited by Otto Eckstein. Washington, D.C.: Brookings Institute.

Achen, Christopher. 2000. "Why lagged dependent variables can suppress the explanatory power of other independent variables.".

Ahluwalia, Montek S. 1976. Inequality, Poverty, and Development. Washington, D.C.: World Bank.

Alarcon, Diana and Terry McKinley. 1998. "Increasing Wage Inequality and Trade Liberalization in Mexico." Pp. 137 in Poverty, Economic Reform, and Income Distribution in Latin America edited by Albert Berry. Boulder, Colorado: Lynne Rienner Publishers.

Alderson, Arthur S. and François Nielsen. 1999. "Income Inequality, Development, and Dependence: A Reconsideration." American Sociological Review 64:pp. 606-631.

Anand, Sudhir and S. M. R. Kanbur. 1993. "Inequality and Development A Critique." Journal of Development Economics 41:19-43.

ANDREICA, ME, ANTONIE MD, CRISTESCU A,CATANICIU N,2010. "A panel data analysis of the romanian labour market." 1:406.

APARASCHIVEI , Larisa, Maria D. VASILESCU and Nicolae CĂTĂNICIU. 2011. "The Impact of Investments and Gross Value Added upon Earnings." Theoretical and Applied Economics XVIII (2):207-18.

Armada, F., C. Muntaner and V. Navarro. 2001. "Health and Social Security Reforms in Latin America: The Convergence of the World Health Organization, the World Bank, and Transnational Corporations." International Journal of Health Services 31:729-68.

Arts, Wil and John Gelissen. 2002. "Three Worlds of Welfare Capitalism Or More? A State-of- the-Art Report." Journal of European Social Policy 12:137-58.

Avdeyeva, Olga. 2007. "When do States Comply with International Treaties? Policies on Violence Against Women in Post-Communist Countries." International Studies Quarterly 51:877-900.

Avelino, G., D. S. Brown and W. Hunter. 2005. "The Effects of Capital Mobility, Trade Openness, and Democracy on Social Spending in Latin America, 1980-1999." American Journal of Political Science 49:625-41.

121

Babb, Sarah. 2005. "The Social Consequences of Structural Adjustment: Recent Evidence and Current Debates." Annual Review of Sociology 31:199-222.

Banks, Arthur, Thomas Muller and William Overstreet. 2009. Political Handbook of the World 2009. Washington: CQ Press.

------.2009. Political Handbook of the World 2009. Washington: CQ Press.

Barro, Robert,Lee J,2010. "A NEW DATA SET OF EDUCATIONAL ATTAINMENT IN THE WORLD, 1950–2010." NATIONAL BUREAU OF ECONOMIC RESEARCH, Working Paper 15902 http://www.nber.org/papers/w15902.

Barros, Richard, Mirela d. Carvalho, Samuel Franco and Rosane Mendonca. 2010. "Markets, the State, and the Dynamics of Inequality in Brazil." Pp. 134 in Declining Inequality in Latin America: A Decade of Progress? edited by Luis F. Lopez-Calva and Nora Lustig. Washington, D.C.: Brookings Institution Press.

Beck, Nathaniel. and J. N. Katz. 1995. "What to do (and Not to do) with Time-Series Cross- Section Data." American Political Science Review 89:634-47.

Beck, Nathaniel. 2001. "Time-series–cross-Section Data." Statistica Neerlandica 55(2):111-33.

Beck, Nathaniel and Jonathan N. Katz. 1996. "Nuisance Vs. Substance: Specifying and Estimating Time-Series-Cross-Section Models." Political Analysis 6:1-36.

------.2004. "Time-series-cross-section issues: Dynamics, 2004.":1.

Beckfield, Jason. 2003. "Inequality in the World Polity: The Structure of International Organization." American Sociological Review 68:401-24.

Beckfield, Jason. 2008. "The Dual World Polity: Fragmentation and Integration in the Network of Intergovernmental Organizations." Social Problems 55:419-42.

Behrman, Jere R., Susan W. Parker and Petra E. Todd. 2009. "Medium-Term Impacts of the Oportunidades Conditional Cash-Transfer Program on Rural Youth in Mexico." Pp. 219 in Poverty, Inequality, and Policy in Latin America, edited by Stephan Klasen and Felicitas Nowak-Lehman. Cambridge, Mass: The MIT Press.

Berry, Albert. 1998. "Confronting the Income Distribution Threat in Latin America." Pp. 9 in Poverty, Economic Reform, and Income Distribution in Latin America, edited by Albert Berry. Boulder, London: Lynne Rienner Publishers.

Berry, Albert and Jaime Tenjo G. 1998. "Trade Liberalization, Labor Reform, and Income Distribution in Colombia." Pp. 155 in Poverty, Economic Reform, and Income Distribution in Latin America, edited by Albert Berry. Boulder, Colorado: Lynne Rienner Publishers.

122

Bertranou, Fabio M. 2001. "Pension Reform and Gender Gaps in Latin America: What are the Policy Options?" World Development 29:911-23.

Bitrán, Ricardo, Ursula Giedion, Rubi Valenzuela and Paavo Monkkonen. 2005. "Keeping Healthy in an Urban Environment: Public Health Challenges for the Urban Poor." The Urban Poor in Latin America, edited by Marianne Fay. Washington, D.C.: The World Bank.

Boli, John and G. M. Thomas. 1997. "World Culture in the World Polity: A Century of International Non-Governmental Organization." American Sociological Review 62:171-90.

Bollen, Kenneth A. and Robert W. Jackman. 1985. "Political Democracy and the Size Distribution of Income." American Sociological Review 50:pp. 438-457.

Boyle, Elizabeth H. and Minzee Kim. 2009. "International Human Rights Law, Global Economic Reforms, and Child Survival and Development Rights Outcomes." Law & Society Review 43:455-90.

Broto, Carmen, Javier Diaz-Cassou and Aitor Erce. 2011. "Measuring and Explaining the Volatility of Capital Flows to Emerging Countries." Journal of Banking & Finance 35:1941-53.

Brown, David. S. and Wendy Hunter. 1999. "Democracy and Social Spending in Latin America, 1980-92." American Political Science Review 93:779-90.

------.2004. "Democracy and Human Capital Formation - Education Spending in Latin America, 1980 to 1997." Comparative Political Studies 37:842-64.

Bruno, M, Ravallion M,Squire L,1996. "Equity and growth in developing countries : Old and new perspectives on the policy issues." Policy Research Working Paper Series 1563, The World Bank.

Calderón, César and Alberto Chong. 2001. "External Sector and Income Inequality in Interdependent Economies using a Dynamic Panel Data Approach." Economics Letters 71:225-31.

Cameron, D. R. 1978. "Expansion of the Public Economy - Comparative Analysis." American Political Science Review 72:1243-61.

Cardenas, S. 2000. "Combining Legal and Political Approaches: Recent Scholarship on Human Rights in Latin America." Latin American Research Review 35:252-67.

------.2003. "Emerging Global Actors: The United Nations and National Human Rights Institutions." Global Governance 9:23-42.

123

Caren, Neal, Raj A. Ghoshal and Vanesa Ribas. 2011. "A Social Movement Generation: Cohort and Period Trends in Protest Attendance and Petition Signing." American Sociological Review 76:125-51.

Carlin, J. B., Ning Li, Philip Greenwood and Carolyn Coffey. 2003. "Tools for Analyzing Multiple Imputed Datasets." The Stata Journal 3:226.

Carlin, John B., John C. Galati and Patrick Royston. 2008. "A New Framework for Managing and Analyzing Multiply Imputed Data in Stata." The Stata Journal 8(1).

Casarico, Alessandra and Carlo Devillanova. 2008. "Capital–skill Complementarity and the Redistributive Effects of Social Security Reform." Journal of Public Economics 92:672-83.

CASTLES, FRANCIS F. G. and Deborah Mitchell. 1992. "Identifying Welfare State Regimes: The Links between Politics, Instruments and Outcomes." Governance (Oxford) 5:1-26.

Cheng, H. Y. 1968. "Emergence and Spread of Ombudsman Institution." Annals of the American Academy of Political and Social Science 377:20-30.

Childers, M,Binder M,2010. "Engaged by initiatives? how the introduction and use of citizen initiatives increase voter turnout." 1465236.

Clark, Rob. 2010. "TECHNICAL AND INSTITUTIONAL STATES: Loose Coupling in the Human Rights Sector of the World Polity." Sociological Quarterly 51:65-95.

Clarke, George R. G. 1995. "More Evidence on Income Distribution and Growth." Journal of Development Economics 47:403-27.

Cogneau, Denis and Jeremie Gignoux. 2009. "Earnings Inequality and Educational Mobility in Brazil Over Two Decades." Pp. 47 in Poverty, Inequality, and Policy in Latin America edited by Stephan Klasen and Felicitas Nowak-Lehman. Cambridge, Mass: The MIT Press.

Cole, W. M. 2005. "Sovereignty Relinquished? Explaining Commitment to the International Human Rights Covenants, 1966-1999." American Sociological Review 70:472-95.

Cole, Wade M. 2009. "Hard and Soft Commitments to Human Rights Treaties, 1966-2000(1)." Sociological Forum 24:563-88.

Cominetti, Rosella. 1996. "Social expenditure in latin america: An update.".

Coppedge, M. 1997. "A CLASSIFICATION OF LATIN AMERICAN POLITICAL PARTIES." Working Paper #244.

Cornia, Giovanni A. 2004. "Inequality, Growth, Amd Poverty: An Overview of Changes Over the Last Two Decades." Inequality, Growth, and Poverty in an Era of Liberalization and

124

Globalization, UNU-WIDER Studies in Development Economics, UNU-Wider and UNDP, edited by Giovanni A. Cornia. New York: Oxford University Press.

Crenshaw, Edward. 1992. "Cross-National Determinants of Income Inequality: A Replication and Extension using Ecological-Evolutionary Theory." Social Forces 71:339-63.

Cutright, P. 1967. "Income Redistribution - Cross-National Analysis." Social Forces 46:180-90.

De Ferranti, David M., Guillermo E. Perry, Francisco H. G. Ferreira and Michael Walton. 2004. Inequality in Latin America: Breaking with History? Washington, D.C: World Bank.

De Hoyos, R. E. and V. Sarafidis. 2006. " Testing for Cross-Sectional Dependence in Panel-Data Models " STATA JOURNAL 6(4):482-96. de Mesa, Alberto A. and Carmelo Mesa-Lago. Spring. "The Structural Pension Reform in Chile: Effects, Comparisons with Other Latin American Reforms, and Lessons." Oxford Review of Economic Policy 22:149-67.

Defensor del Pueblo de la Nacion. 1998. Argentina.

Deininger, Klaus and Lyn Squire. 1996. "A New Data Set Measuring Income Inequality." The World Bank Economic Review 10:565-91.

DiMaggio, Paul J. and Walter W. Powell. 1983. "The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields." American Sociological Review 48:pp. 147-160.

Dion, Michelle. 2008. "Pension Reform and Gender Inequality." Pp. 134 in Lessons from Pension Reform in the Americas, edited by Stephen J. Kay and Tapen Sinha. New York: Oxford University Press.

Dion, Michelle. 2006. "Women’s Welfare and Social Security Privatization in Mexico." Social Politics: International Studies in Gender, State & Society 13:400-26.

Dodson, M. and D. Jackson. 2004. "Horizontal Accountability in Transitional Democracies: The Human Rights Ombudsman in El Salvador and Guatemala." Latin American Politics and Society 46:1-27.

Donnelly, Jack. 2002. "Human Rights as Politics and Idolatry." International Affairs 78:899-901.

Donnelly, Jack. 1998. International Human Rights. Boulder, Colo.: Westview Press.

Driscoll, John C. and Aart C. Kraay. 1998. " Consistent Covariance Matrix Estimation with Spatially Dependent Panel Data " Review of Economics and Statistics 80(4):549-60.

125

Eckstein, Otto e. 1967. Studies in the Economics of Income Maintenance. Washington: Brookings Institution.

Economic Commission for Latin American and the Caribbean. 2004, 2006. Social Panorama of Latin America. Santiago, Chile: United Nations Economic Commission for Latin American and the Caribbean.

Esping Andersen, Gøsta. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity.

Esquivel, Gerardo, Nora Lustig and John Scott. 2010. "Mexico: A Decade of Falling Inequality: Market Forces Or State Action?" Pp. 175 in Declining Inequality in Latin America: A Decade of Progress? edited by Luis F. Lopez-Calva and Nora Lustig. Washington, D.C.: Brookings Institution Press.

Fay, Marianne and Caterina Laderchi. 2005. "Urban Poverty in Latin America and the Caribbean: Setting the Stage." Pp. 19 in The Urban Poor in Latin America edited by Marianne Fay. Washington, D.C.: The World Bank.

Fay, Marianne and Anna Wellenstein. 2005. "Keeping a Roof Over One’s Head: Improving Access to Safe and Decent Shelter." Pp. 91 in The Urban Poor in Latin America edited by Marianne Fay. Washington, D.C: The World Bank.

Feenstra, Robert C. and Gordon H. Hanson. 1997. "Foreign Direct Investment and Relative Wages: Evidence from Mexico's Maquiladoras." Journal of International Economics 42:371-93.

Fields, Gary S. 1994. "Data for Measuring Poverty and Inequality Changes in the Developing Countries." Journal of Development Economics 44:87-102.

Frankema, Ewout. 2009. "The Colonial Origins of Inequality: Exploring the Causes and Consequences of Land Distribution." Poverty, Inequality, and Policy in Latin America edited by Stephan Klasen and Felicitas Nowak-Lehman. Cambridge, Mass.: The MIT Press.

Freedom House. 2011. "Freedom in the world 2011.".

Frees, Edward W. 2004. Longitudinal and Panel Data: Analysis and Applications in the Social Sciences. New York: Cambridge University Press.

------.2004. Longitudinal and Panel Data: Analysis and Applications in the Social Sciences. New York: Cambridge University Press.

Garrett, Geoffrey. 1998. Partisan Politics in the Global Economy. Cambridge, U.K. ; New York: Cambridge University Press.

126

Gasparini, Leonardo and Guillermo Cruces. 2010. "A Distribution in Motion: The Case of Argentina." Pp. 100 in Declining Inequality in Latin America: A Decade of Progress? edited by Luis F. Lopez-Calva and Nora Lustig.

Gay, Claudine. 2012. "Moving to Opportunity." Urban Affairs Review 48:147-79.

Gill, Indermit S., Truman Packard and Juan Yermo. 2005. Keeping the Promise of Social Security in Latin America. Palo Alto, Calif.: A copublication of Stanford Economics and Finance, an imprint of Stanford University Press, and the World Bank, Washington, DC.

------.2005. Keeping the Promise of Social Security in Latin America. Palo Alto, Calif: A copublication of Stanford Economics and Finance, an imprint of Stanford University Press, and the World Bank, Washington, DC.

Goodliffe, J. and D. G. Hawkins. 2006. "Explaining Commitment: States and the Convention Against Torture." Journal of Politics 68:358-71.

------.2006. "Explaining Commitment: States and the Convention Against Torture." Journal of Politics 68:358-71.

Grugel, Jean and Enrique Peruzzotti. 2007. "Claiming Rights Under Global Governance: Children's Rights in Argentina." Global Governance 13:199-216.

Ha, Eunyoung. 2012. "Globalization, Government Ideology, and Income Inequality in Developing Countries." Journal of Politics 74:541-57.

Hadden, Tom. 2002. "The Role of a National Commission in the Protection of Human Rights." Pp. 291-9 in Human Rights Commissions and Ombudsman Offices: National Experiences and Throughout the World, edited by Kamal Hossain, L. F. M. Besselink, H. S. G. Selassie and Edmond and Volker. The Hague, The Netherlands: Kluwer Law International.

Hafner-Burton, E. M. and K. Tsutsui. 2005. "Human Rights in a Globalizing World: The Paradox of Empty Promises." American Journal of Sociology 110:1373-411.

------.2005. "Human Rights in a Globalizing World: The Paradox of Empty Promises." American Journal of Sociology 110:1373-411.

Hafner-Burton, Emilie M. and Kiyoteru Tsutsui. 2007. "Justice Lost! the Failure of International Human Rights Law to Matter Where Needed most." Journal of Peace Research 44:407-25.

Hafner-Burton, Emilie M., Kiyoteru Tsutsui and John W. Meyer. 2008. "International Human Rights Law and the Politics of Legitimation - Repressive States and Human Rights Treaties." International Sociology 23:115-41.

127

Hannum, H. 2006. "Human Rights in Conflict Resolution: The Role of the Office of the High Commissioner for Human Rights in UN Peacemaking and Peacebuilding." Human Rights Quarterly 28:1-85.

Harrison, AE. 2002. "Has globalization eroded labor's share? Some cross-country evidence.".

HATHAWAY, O. A. 2002. "DO HUMAN RIGHTS TREATIES MAKE A DIFFERENCE?" The Yale Law Journal 111.

Hathaway, Oona A. 2007. "Why do Countries Commit to Human Rights Treaties?" Journal of Conflict Resolution 51:588-621.

Hicks, Alexander. 1994. "Introduction to Pooling." Pp. 169-88 in The Comparative Political Economy of the Welfare State, edited by Alexander Hicks and Thomas Janoski. New York: Cambridge University Press.

------.1999. Social Democracy and Welfare Capitalism. Ithaca, N.Y.: Cornell University Press.

Hoechle, Daniel. 2007. "Robust Standard Errors for Panel Regressions with Cross-Sectional Dependence." The Stata Journal 7(3):1-31.

Hogerzeil, Hans V., Melanie Samson, Jaurne V. Casanovas and Ladan Rahmani-Ocora. 2006. "Is Access to Essential Medicines as Part of the Fulfilment of the Right to Health Enforceable through the Courts?" Lancet 368:305-11.

Honaker, James and Gary King. 2010. "What to do about Missing Values in Time-Series Cross- Section Data." American Journal of Political Science 54:561-81.

Hossain, Kamal. 2000. Human Rights Commissions and Ombudsman Offices : National Experiences Throughout the World. The Hague ; Boston: Kluwer Law International.

Huber, E., C. Ragin and J. D. Stephens. 1993. "Social-Democracy, Christian Democracy, Constitutional Structure, and the Welfare-State." American Journal of Sociology 99:711-49.

Huber, Evelyne, Thomas Mustillo and John D. Stephens. 2008. "Politics and Social Spending in Latin America." Journal of Politics 70:420-36.

Huber, Evelyne, Francois Nielsen, Jenny Pribble and John D. Stephens. 2006. "Politics and Inequality in Latin America and the Caribbean." American Sociological Review 71:943-63.

Huber, Evelyne, John Stephens, Thomas Mustillo and Jennifer Pribble. 2008b. Social Policy in Latin America and the Caribbean Dataset, 1960-2006. University of North Carolina: .

Immergut, Ellen M. 1992. Health Politics : Interests and Institutions in Western Europe. Cambridge England ; New York, NY, USA: Cambridge University Press.

128

Inter American Development Bank. 1998. Facing Up to Inequality in Latin America : Economic and Social Progress in Latin America, 1998-1999 Report. Washington, D.C: Distributed by the Johns Hopkins University Press for the Inter-American Development Bank.

------.1998. Facing Up to Inequality in Latin America : Economic and Social Progress in Latin America, 1998-1999 Report. Washington, D.C.: Distributed by the Johns Hopkins University Press for the Inter-American Development Bank.

International Bank for Reconstruction and Development and World Bank. 1994. Averting the Old Age Crisis : Policies to Protect the Old and Promote Growth. Oxford, England: Published for the World Bank by Oxford University Press.

International Monetary Fund. Various Years. Government Finances Statistics Yearbook. Washington: IMF.

------.Various Years. International Finances Statistics Yearbook. Washington: IMF.

Jackman, Robert W. 1975. Politics and Social Equality : A Comparative Analysis. New York: Wiley.

James, Estelle, Alejandra C. Edwards and Rebecca Wong. 2008. "Reflections on Pension Reform in the Americas: From ‘Averting the Old-Age Crisis’ to ‘Keeping the Promise of Old-Age Security’ and Beyond." Pp. 164 in Lessons from Pension Reform in the Americas, edited by Stephen J. Kay and Tapen Sinha. New York: Oxford University Press.

Jensen, Nathan M. and Guillermo Rosas. 2007. "Foreign Direct Investment and Income Inequality in Mexico, 1990-2000." International Organization 61:pp. 467-487.

Kahat, Jamie. 2010. "Labor Earnings Inequality: The Demand for and Supply of Skills " Pp. 25 in Declining Inequality in Latin America: A Decade of Progress? edited by Luis F. Lopez- Calva and Nora Lustig. Washington, D.C.: Brookings Institution Press.

Kamminga, Menno T. 1996. "State Succession in Respect of Human Rights Treaties." Eur J International Law 7:469-84.

Kangas, Olli. 1991. "the Politics of Social Rights: Studies on the Dimensions of Sickness Insurance." OECD Countries, Anonymous Stockholm: Swedish Institute for Social Research.

Kaufman, R. R. and A. Segura-Ubiergo. 2001. "Globalization, Domestic Politics, and Social Spending in Latin America - A Time-Series Cross-Section Analysis, 1973-97." World Politics 53:553,+.

Keck, M. E. and K. Sikkink. 1999. "Transnational Advocacy Networks in International and Regional Politics." International Social Science Journal 51:89,+.

129

Kemp-Benedict, Eric. 2011. "Political Regimes and Income Inequality." Economics Letters 113:266-8.

Kerr, Clark. 1960. Industrialism and Industrial Man; the Problems of Labor and Management in Economic Growth. Cambridge: Harvard University Press.

Klien, M, Leibrecht M,Onaran Ö,2010. "Globalization, welfare regimes and social protection expenditures in western and eastern european countries." Discussion Papers SFB International Tax Coordination, 30.

Kmenta, Jan. 1997. Elements of Econometrics. Ann Arbor: University of Michigan Press.

Koo, Jeong-Woo and Francisco O. Ramirez. 2009. "National Incorporation of Global Human Rights: Worldwide Expansion of National Human Rights Institutions, 1966-2004." Social Forces 87:1321-53.

Korpi, Walter. 1980. "Social Policy and Distributional Conflict in the Capitalist Democracies: A Preliminary Framework." West European Politics 3:296-316.

Kumar, C. R. 2006. "National Human Rights Institutions and Economic, Social, and Cultural Rights: Toward the Institutionalization and Developmentalization of Human Rights." Human Rights Quarterly 28:755-79.

Laderchi, Caterina. 2005. "Working One’s Way Up: The Urban Poor and the Labor Market." Pp. 47 in The Urban Poor in Latin America edited by Marianne Fay. Washington, D.C.: The World Bank.

Larrea, Carlos. 1998. "Structural Adjustment, Income Distribution, and Employment in Ecuador." Pp. 179 in Poverty, Economic Reform, and Income Distribution in Latin America, edited by Albert Berry. Boulder, Colorado: Lynne Rienner Publishers.

Laurell, A. C. 2000. "Structural Adjustment and the Globalization of Social Policy in Latin America." International Sociology 15:306-25.

Leite, Phillippe G. 2009. "Race Discrimination Or Inequality of Opportunities: The Brazilian Case." Pp. 85 in Poverty, Inequality, and Policy in Latin America edited by Stephan Klasen and Felicitas Nowak-Lehman. Cambridge, Mass.: The MIT Press.

Li, Lianjiang. 2011. "Distrust in Government Leaders, Demand for Leadership Change, and Preference for Popular Elections in Rural China." Political Behavior 33:291-311.

Lindsnaes, Birgit and Lone Lindholt. 2000. " National Human Rights Institutions: Standard- Setting and Achievements." National Human Rights Institutions Articles and Working Papers, edited by Birgit Lindsnaes, L. Lindholt and K. Yigen. Copenhagen: The Danish Centre for Human Rights.

130

Little, R. J. A. and D. B. Rubin. 2002. Statistical Analysis with Missing Data, 2Nd Edition. New York: John Wiley.

Little, Roderick J. A. 1992. "Regression with Missing X's: A Review." Journal of the American Statistical Association 87:pp. 1227-1237.

Lopez-Calva, Luis F. and Nora Lustig. 2010. "Explaining the Decline in Inequality in Latin America: Technological Change, Educational Upgrading, and Democracy." Pp. 1 in Declining Inequality in Latin America: A Decade of Progress? edited by Luis F. Lopez- Calva and Nora Lustig. Washington, D.C.: Brookings Institution Press.

Lynch, Julia and Mikko Myrskylä. 2009. "Always the Third Rail?" Comparative Political Studies 42:1068-97.

Mahler, Vincent A. 2010. "Government Inequality Reduction in Comparative Perspective: A Cross-National Study of the Developed World." Polity 42:511-41.

Maioni, Antonia. 1998. Parting at the Crossroads : The Emergence of Health Insurance in the United States and Canada. Princeton, N.J.: Princeton University Press.

Marshall, Monty G. and Keith Jaggers. 2002. Polity IV Project: Political Regime Characteristics and Transitions, 1800-2002. Version p4v2002e. College Park, MD: Center for International Development and Conflict Management, University of Maryland.

Marshall, T. H. 1964. Class, Citizenship, and Social Class: Essays by TH Marshall. Garden City, NY: Doubleday and Co.

McGreevey, William P. 1990. Social Security in Latin America: Issues and Options for the World Bank. Washington, D.C.: The World Bank.

Meng, Xiao-Li. 1994. "Multiple-Imputation Inferences with Uncongenial Sources of Input." Statistical Science 9:pp. 538-558.

Meyer, J. W., J. Boli, G. M. Thomas and F. O. Ramirez. 1997. "World Society and the Nation- State." American Journal of Sociology 103:144-81.

Mohl, P,Hagen T,2009. "How does EU cohesion policy work? evaluating its effects on fiscal outcome variables." Discussion Paper No. 09-051.

Moore, Ryan T. and Nirmala Ravishankar. 2012. "Who Loses in Direct Democracy?" Social Science Research 41:646-56.

Moravcsik, Andrew. 2000. "The Origins of Human Rights Regimes: Democratic Delegation in Postwar Europe." International Organization 54:217.

131

Morley, Samuel A. 2003. "Distribution and Growth in Latin America in an Era of Structural Reform: The Impact of Globalisation." Pp. 63 in Globalisation, Poverty and Inequality, edited by Richard Kohl. OECD Publishing.

Muller, Edward N. 1985. "Income Inequality, Regime Repressiveness, and Political Violence." American Sociological Review 50:pp. 47-61.

------.1989. "Democracy and Inequality." American Sociological Review 54:pp. 868-871.

Myles, John. 1984. Old Age in the Welfare State: The Political Economy of Public Pensions. Boston: Little, Brown.

------.1984. Old Age in the Welfare State: The Political Economy of Public Pensions. Boston: Little, Brown.

Neumayer, E. 2005. "Do International Human Rights Treaties Improve Respect for Human Rights?" Journal of Conflict Resolution 49:925-53.

Nielsen, François. 1994. "Income Inequality and Industrial Development: Dualism Revisited." American Sociological Review 59:pp. 654-677.

Nielsen, François and Arthur S. Alderson. 1995. "Income Inequality, Development, and Dualism: Results from an Unbalanced Cross-National Panel." American Sociological Review 60:pp. 674-701.

Nohlen, Dieter. 1993. Enciclopedia Electoral Latinoamericana Y Del Caribe. Costa Rica: Instituto Interamericano de Derechos Humanos.

------.2005. Elections in the Americas: A Data Handbook. New York: Oxford University Press.

O'Connor, James. 1973. The Fiscal Crisis of the State. New York: St. Martin's Press.

Offe, Claus. 1972. "Advanced Capitalism and the Welfare State." Politics & Society 2 (4):479- 88.

Orloff, AS. 2002. "Womenís employment and welfare regimes globalization, export orientation and social policy in europe and north america " Social Policy and Development Programme Paper Number 12.

Orloff, Ann S., Project on the Federal Social Role (U.S.), Theda Skocpol and Margaret Weir. 1988. The Politics of Social Policy in the United States. Princeton, N.J.: Princeton University Press.

132

Palme, J. 1988. "The determinants of old age pensions in 18 OECD countries 1930 to 1980.".

Palmer, Alexis, Jocelyn Tomkinson, Charlene Phung, Nathan Ford, Michel Joffres, Kimberly A. Fernandes, Leilei Zeng, Viviane Lima, Julio S. G. Montaner, Gordon H. Guyatt and Edward J. Mills. 2009. "Does Ratification of Human-Rights Treaties have Effects on Population Health?" Lancet 373:1987-92.

Pampel, F. C. and J. B. Williamson. 1988. "Welfare Spending in Advanced Industrial Democracies, 1950-1980." American Journal of Sociology 93:1424-56.

------.1988. "Welfare Spending in Advanced Industrial Democracies, 1950-1980." American Journal of Sociology 93:1424-56.

Pampel, Fred C. and John B. Williamson. 1985. "Age Structure, Politics, and Cross-National Patterns of Public Pension Expenditures." American Sociological Review 50:pp. 782-799.

------.1989. Age, Class, Politics, and the Welfare State. Cambridge [England]: Cambridge University Press.

Pampel, Fred C., John B. Williamson and Robin Stryker. 1990. "Class Context and Pension Response to Demographic Structure in Advanced Industrial Democracies." Social Problems 37:pp. 535-550.

Pegram, Thomas. 2008. "Accountability in Hostile Times: The Case of the Peruvian Human Rights Ombudsman 1996-2001." Journal of Latin American Studies 40:51-82.

PLÜMPER, THOMAS and Eric Neumayer. 2009. "Model Specification in the Analysis of Spatial Dependence." European Journal of Political Research 49(3):418-42.

PLÜMPER, THOMAS, VERA E. TROEGER and PHILIP MANOW. 2005. "Panel Data Analysis in Comparative Politics: Linking Method to Theory." European Journal of Political Research 44(2):327-54.

Popkin, M. and N. Rohtarriaza. 1995. "Truth as Justice - Investigatory Commissions in Latin- America." Law and Social Inquiry-Journal of the American Bar Foundation 20:79-116.

Powell, Emilia J. and Jeffrey K. Staton. 2009. "Domestic Judicial Institutions and Human Rights Treaty Violation." International Studies Quarterly 53:149-74.

Prasad, Monica. 2006. The Politics of Free Markets the Rise of Neoliberal Economic Policies in Britain, France, Germany, and the United States. Chicago: University of Chicago Press.

Puryear, J. 1997. "La edcucacion en america latin: Problemas y desafios." Document 7.

Ravallion, Martin and Shaohua Chen. 1997. "What can New Survey Data Tell Us about Recent Changes in Distribution and Poverty?" The World Bank Economic Review 11:357-82. 133

Reif, Linda C. 2000. "Building Democratic Institutions: The Role of National Human Rights Institutions in Good Governance and Human Rights Protection." Harvard Human Rights Journal 1:1-69.

Reiter, Jerome P. and Trivellore E. Raghunathan. 2007. "The Multiple Adaptations of Multiple Imputation." Journal of the American Statistical Association 102:480:1462-71.

Risse Kappen, Thomas, Steve C. Ropp and Kathryn Sikkink. 1999. The Power of Human Rights: International Norms and Domestic Change. Cambridge ; New York: Cambridge University Press.

Risse, Thomas and K. Sikkink. 1999. " the Socialization of International Human Rights Norms into Domestic Practices: Introduction." the Power of Human Rights: International Norms and Domestic Change, edited by Thomas Risse, Steve C. Ropp and Kathryn Sikkink. New York: Cambridge University Press.

Robinson, James A. 2010. "The Political Economy of Redistributive Policies." Pp. 39 in Declining Inequality in Latin America: A Decade of Progress? edited by Luis F. Lopez- Calva and Nora Lustig. Washington, D.C.: Brookings Institution Press.

Rodrik, Dani. 1997. Has Globalization Gone Too Far? Washington, DC: Institute for International Economics.

Royston, P. 2004. "Multiple Imputation of Missing Values." The Stata Journal 4:227.

------.2007. "Multiple Imputation of Missing Values: Further Update of Ice, with an Emphasis on Interval Censoring." The Stata Journal.

Rubin, D. B. 1976. "Inference and Missing Data (with Discussion)." Biometrika 63:581-92.

------.1987. Multiple Imputation for Non-Response in Surveys. New York: John Wiley & Sons.

Rubin, Donald B. 1996. "Multiple Imputation After 18+ Years." Journal of the American Statistical Association 91:pp. 473-489.

Rudra, N. 2002. "Globalization and the Decline of the Welfare State in Less-Developed Countries." International Organization 56:411,+.

------.2004. "Openness, Welfare Spending, and Inequality in the Developing World." International Studies Quarterly 48:683-709.

Rudra, N. and S. Haggard. 2005. "Globalization, Democracy, and Effective Welfare Spending in the Developing World." Comparative Political Studies 38:1015-49.

RUGMAN, A. 2002. "Globalization and its Discontentsby Joseph E. Stiglitz, Norton, New York (2002), 282 Pp., $24.95." Long Range Planning 35:654-6. 134

Saavedra, Jaime and Mariano Tommasi. 2007. "Informality, the State and the Social Contract in Latin America: A Preliminary Exploration." International Labour Review 146:279,+.

Sanchez, Omar. 2006. "Tax System Reform in Latin America: Domestic and International Causes." Review of International Political Economy 13:772-801.

Savvides, Andreas. 1998. "Trade Policy and Income Inequality: New Evidence." Economics Letters 61:365-72.

Schafer, J. L. 1997. Analysis of Incomplete Multivariate Data. Chapman & Hall/CRC.

Schmaehl, Winfried. 2007. "Dismantling an Earnings-Related Social Pension Scheme: Germany's New Pension Policy." Journal of Social Policy 36:319-40.

Schofer, E. and J. W. Meyer. 2005. "The Worldwide Expansion of Higher Education in the Twentieth Century." American Sociological Review 70:898-920.

Segura-Ubiergo, Alex. 2007. The Political Economy of the Welfare State in Latin America : Globalization, Democracy, and Development. Cambridge ; New York: Cambridge University Press.

SHAFIR, GERSHON and ALISON BRYSK. 2006. "The Globalization of Rights: From Citizenship to Human Rights." Citizenship Studies 10:275-87.

Shin, D. M. 2000. "Economic Policy and Social Policy: Policy-Linkages in an Era of Globalisation." International Journal of Social Welfare 9:17-30.

Simon Kuznets. 1955. "Economic Growth and Income Inequality." The American Economic Review 45:pp. 1-28.

Simpson, Miles. 1990. "Political Rights and Income Inequality: A Cross-National Test." American Sociological Review 55:pp. 682-693.

Singh, Jerome A., Michelle Govender and Edward J. Mills. 2007. "Health and Human Rights 2 - do Human Rights Matter to Health?" Lancet 370:521-7.

Smith, Anne. 2006. "The Unique Position of National Human Rights Institutions: A Mixed Blessing?" Human Rights Quarterly 28:904-46.

Smith, J. and D. Wiest. 2005. "The Uneven Geography of Global Civil Society: National and Global Influences on Transnational Association." Social Forces 84:621-52.

Smith, T. and M. Kimmel. 2005. "The Hidden Discourse of Masculinity in Gender Discrimination Law." Signs 30:1827-49.

StataCorp. 2009. "SC stata - statistical software." Release 11. 135

Stephens, John. 1979. the Transition from Capitalism to Socialism. Urbana: University of Illinois Press.

Stiglitz, Joseph E. 2002. Globalization and its Discontents. New York: W.W. Norton & Co.

Swank, D. 1998. "Funding the Welfare State: Globalization and the Taxation of Business in Advanced Market Economies." Political Studies 46:671-92.

Thakur, Ramesh and Jorge Heine. 2010. "Human Rights and the State in Latin America." Human Rights Regimes in the Americas, edited by Mónica Serrano and Vesselin Popovski. New York: United Nations University Press.

The World Bank. 1991. World Development Report 1991: The Challenge of Development. New York: Oxford University Press.

------.2011. " World development indicators .".

Tsutsui, K. and C. M. Wotipka. 2004. "Global Civil Society and the International Human Rights Movement: Citizen Participation in Human Rights International Nongovernmental Organizations." Social Forces 83:587-620.

Uggla, F. 2004. "The Ombudsman in Latin America." Journal of Latin American Studies 36:423- 50.

Union of International Associations. Annual. "Yearbook of international organization.".

United Nations Centre for Human Rights and United Nations General Assembly Universal Declaration of Human Rights. 1995. National Human Rights Institutions : A Handbook on the Establishment and Strengthening of National Institutions for the Promotion and Protection of Human Rights. New York: United Nations.

United Nations Dept. of Public Information. 1998. The Universal Declaration of Human Rights, 1948-1998. New York: United Nations Dept. of Public Information.

United Nations General Assembly. 1948. "Universal declaration of human rights." G.A. Res. 217A (III), U.N. Doc. A/810 at 71.

United Nations Treaty Section. 1948. "Convention on the prevention and punishment of the crime of genocide." 78 U.N.T.S. 277, 102 Stat. 3045.

------.1965. "International convention on the elimination of all forms of racial discrimination." S. Exec. Doc. C, 95-2 (1978); S. Treaty Doc. 95-18; 660 U.N.T.S. 195, 212.

------.1966. "International covenant on civil and political rights." 999 U.N.T.S. 171; S. Exec. Doc. E, 95-2 (1978); S. Treaty Doc. 95-20, 6 I.L.M. 368 (1967).

136

------.1966. "International covenant on economic, social and cultural rights." S. Exec. Doc. D, 95-2 (1978); S. Treaty Doc. No. 95-19, 6 I.L.M. 360.

------.1979. "Convention on the elimination of all forms of discrimination against women." 1249 U.N.T.S. 13; 19 I.L.M. 33(1980).

------.1984. "Convention against torture and other cruel, inhuman or degrading treatment or punishment." 465 U.N.T.S. 85, 113; S. Treaty Doc. No. 100-20(1988); 23 I.L.M. 1027(1984).

------.1990. "International convention on the protection of the rights of all migrant workers and members of their families." 2220 U.N.T.S. 93, 30 I.L.M. 1517 (1991).

United Nations Treaty Section and United Nations Treaty Section. 2001. Treaty Handbook. New York: United Nations.

United States Social Security Administration. 2011. "Social Security Programs Throughout the World." United States Social Security Administration, 2011 (http://www.ssa.gov/policy/docs/progdesc/ssptw/).

United States. Central Intelligence Agency. 2003. The 2000 CIA World Factbook.

UNU-WIDER. 2005, 2008. "World income inequality database, version 2.0c.".

Urani, Andre. 1998. "The Effects of Macroeconomic Adjustment on the Labor Market and on Income Distribution in Brazil." Pp. 205 in Poverty, Economic Reform, and Income Distribution in Latin America, edited by Albert Berry. Boulder, Colorado: Lynne Rienner Publishers.

Van Buuren, S., H. C. Boshuizen and D. L. Knook. 1999. "Multiple Imputation of Missing Blood Pressure Covariates in Survival Analysis." Statistics in Medicine 18:681-94.

Van Kersenberg, Kees. 1995. Social Capitalism. London: Routledge.

Von Maltitz, Michael J. and Abraham J. Van der Merwe. 2012. "An Application of Sequential Regression Multiple Imputation on Panel Data." South African Journal of Economics 80:77- 90.

Vreeland, James R. 2008. "Political Institutions and Human Rights: Why Dictatorships Enter into the United Nations Convention Against Torture." International Organization 62:65- 101.

Weyland, Kurt G. 2004. Learning from Foreign Models in Latin American Policy Reform. Washington, D.C.; Baltimore: Woodrow Wilson Center Press; Johns Hopkins University Press.

137

Wilensky, Harold L. 1975. The Welfare State and Equality : Structural and Ideological Roots of Public Expenditures. Berkeley: University of California Press.

Woolcock, Michael. 2005. "Calling on Friends and Relatives: Social Capital." Pp. 219 in The Urban Poor in Latin America edited by Marianne Fay. Washington, D.C.: The World Bank.

World Bank. "World Debt Tables.".

Wotipka, Christine M. and Kiyoteru Tsutsui. 2008. "Global Human Rights and State Sovereignty: State Ratification of International Human Rights Treaties, 1965-2001." Sociological Forum 23:724-54.

138

BIOGRAPHICAL SKETCH

Kaiser Russell Shekha

Department of Sociology Pepper Institute of Aging and Public Policy Florida State University

June 2012

EMPLOYMENT

2012 Assistant Professor of Sociology/Anthropology, Denison University (beginning August 2012)

2006- 2012 Graduate Assistant, Department of Sociology, Florida State University

EDUCATION

2012 Ph.D. Sociology, Florida State University

Ph.D. Dissertation: Universal Human Rights as New Political Power Resources: Explaining Social Spending Variation and National Income Inequality in 18 Latin American Nations, 1980-2008

Committee: Jill Quadagno (Chair), Dan Tope, Deana Rohlinger and William H. Moore

2010 Comprehensive Exam: Politics and Social Movements

2008 M.S. Sociology, Florida State University

2005 B.A. Magna Cum Laude in Anthropology, Florida Atlantic University

AREAS OF SPECIALIZATION

Human Rights; Globalization; Latin America; Public Policy; Political Economy; Welfare State; Transnational Social Movements; Quantitative and Qualitative Methods

PUBLICATIONS

Shekha, K. Russell. 2011. “Determinants of Latin American Activism: Domestic and International Political Opportunities and Threats” Sociology Compass 5/8

Quadagno, Jill, Ben Lenox Kail and K. Russell Shekha. 2010. “Welfare States: Protecting or Risking Old Age.” Handbook of the Sociology of Aging, Rick Settersten and Jacqueline Angel, Eds. Springer.

139

UNDER REVIEW

Joellen Pederson and K. Russell Shekha. “Attitudes about Pensions and the Elderly: Testing the Self-interest and political ideology theories in Latin American Welfare States”.

MANUSCRIPTS IN PROGRESS

Shekha, K. Russell. “Determinants of Latin American Social Spending: National and International Human Rights Institutions and Instruments”

Shekha, K. Russell. “Human Rights and National Income Inequality in Latin America, 1980- 2008”

Shekha , K. Russell. “Latin American Social Spending and Religious Politics, 1980-2008”

Shekha, K. Russell. “Guaranteeing ESCRs through Universal Human Rights in Latin America, 1980-2010”

Shekha, K. Russell. “Furthering Citizenship: How Universal Human Rights Transcends National Citizenship by Extending Social Rights”

K. Russell Shekha and Doug Schrock. "Farmworker Solidarity Activists' Human Rights Discourse"

GRANTS, HONORS, AND AWARDS

2011 SSSP Sociology and Social Welfare Division Student Paper Competition: “Human Rights vs. Neoliberalism: Welfare State Spending as Institutionalization of Human Rights Norms in 18 Latin American Nations, 1980-2008”. $250 2011 Mildred and Claude Pepper Dissertation Fellowship – Sponsored by the Pepper Institute of Aging and Public Policy, Florida State University. $19,000 2010 National Science Foundation Sociology Dissertation Improvement Grant: “Determinants of Latin American Welfare State Spending: Global and Religious Politics” (with Jill Quadagno). $9650 2005 John Q. Adams Award for International Travel, Florida Atlantic University, $500

CONFERENCE PRESENTATIONS

Summer 2011 - SSSP Sociology and Social Welfare Division Student Paper Competition – K. Russell Shekha. "Human Rights vs. Neoliberalism: Welfare State Spending as Institutionalization of Human Rights Norms in 18 Latin American Nations, 1980-2008” – Student Paper Competition Presentation at the annual meeting of the Society for the Study of Social Problems, Las Vegas, NV

Summer 2011- “Differentiating Effects of International Nongovernmental Organizations on Social Spending in Latin America: Social Justice INGOs VS. Business and Professional INGOs” 140

Roundtable discussion at annual meeting of the American Sociological Association, Las Vegas, NV

Spring 2011- “Latin American Social Spending and Political Parties” Paper presented at annual meeting of the Pacific Sociological Association, Seattle, WA

Spring 2011- “Human Rights and Income Inequality in Latin America” Paper presented at annual meeting of the Southwest Sociological Association, Las Vegas, NV

Summer 2010 - “Determinants of Latin American Welfare State Spending: Global and Religious Politics and International Human Rights” Paper presented at the annual meeting of the American Sociological Association, Atlanta, GA

Summer 2009- “Global Politics and the Latin American Welfare State” Roundtable discussion at annual meeting of the American Sociological Association, San Francisco, CA

Summer 2008- “Latin American Transnational Social Movement Organizations, 2000” Paper presented at the Society for the Study of Social Problems, Boston, MA

Summer 2007- “Pathways to Activism – The Student Farmworker Alliance” with Daphne Holden, Doug Schrock, and Marc Dixon. Paper presented at the Society for the Study of Social Problems, New York.

Spring 2007- “Transnational Social Movements and Neoliberalism – A Comparative Study” Paper presented at the Southern Sociological Society annual meeting, Atlanta.

TEACHING

Instructor The Global Justice Movement (Face-to-face and online) Social Statistics (Online)

Teaching Assistant - Undergraduate Aging & Life Course (Online) Social Psychology of Groups (Face-to-face and online) Deviance and Social Control (Online) Population and Society Social Problems

Teaching Assistant - Graduate Field Methods Multi-Variate Analysis Focus Groups (Invited Guest Lecturer)

RESEARCH EXPERIENCE

2009 Research Assistant for Dr. Jill Quadagno. Researched Latin American welfare states. 141

2008 Research Assistant for Dr. Dan Tope. Organized data on State Children’s Health Insurance Plans.

2007 Research Assistant for Dr. Doug Schrock. Performed in-depth in-person and phone interviews.

PROFESSIONAL SERVICE

National 2006- present American Sociological Association 2006-present Society for the Study of Social Problems 2006-present Southern Sociological Society 2010-present Sociologists Without Borders 2011 Pacific Sociological Association 2011 Southwest Social Science Society 2011 United State Human Rights Network

University 2009 Sociology Graduate Student Union (SGSU) – Treasurer 2008 SGSU – Professional Development Committee (Co-Chair) 2008 Organized SGSU Professional Development Colloquium. 2008 Organized Research-Talk Qualitative Training Session. 2007 Qualitative Research Group (Co-Chair)

142