The Oxford Handbook of the Welfare State

Francis G. Castles (ed.) et al.

https://doi.org/10.1093/oxfordhb/9780199579396.001.0001
Published: 2010 Online ISBN: 9780191594014 Print ISBN: 9780199579396

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

CHAPTER

44 Latin America 
Evelyne Huber, Juan Bogliaccini

https://doi.org/10.1093/oxfordhb/9780199579396.003.0044 Pages 644–655

Published: 02 September 2010

Abstract
This article starts by describing the origins and development of Latin American welfare states before
moving on to outline more recent reforms of social policy regimes in the area. It also explains the
reasons why some policy regimes are more e ective than others and the evidence concerning the
impact of programmes in the di erent countries. It discusses future directions for research on Latin
American policy regimes. The Costa Rican and Uruguayan reforms go far towards basic universalism,
whereas the universalism of the Chilean reform remains con ned to the speci ed set of illnesses. In
the area of labour market policy, governments have had great di culty making the transition from
traditional forms of employment protection to more exible labour markets and policies that would
still protect workers from falling into poverty with their families. Another major area of developing
research is the comparative analysis of political struggles over speci c policy initiatives.

Keywords: Latin American welfare states, social policy, Costa Rica, Uruguay, Chile, labour market, reforms,
universalism
Subject: Comparative Politics, Regional Political Studies, Politics
Series: Oxford Handbooks

Introduction

THIS chapter is organized into three sections. The rst discusses the origins and development of Latin
American welfare states before moving on to outline more recent reforms of social policy regimes in the
area. A second section examines reasons why some policy regimes are more e ective than others and
discusses evidence concerning the impact of programmes in the di erent countries. The chapter concludes
by discussing future directions for research on Latin American policy regimes.
 Development and Reforms of Latin American Welfare States

Origins of the Welfare State

The concept of the Welfare State (Estado de Bienestar) has come into widespread use relatively recently in the
Latin American literature. Traditionally, studies of social security dominated the literature. This re ected
reality in so far as states—to the extent that they began to take responsibility for the welfare of their
p. 645 citizens—emphasized employment based social insurance, that is, social security, over non‐contributory

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

social assistance. In healthcare, the dominant model was social security coverage for the employed
population and dependents. Public health services in charge of preventive care and provision of care for the
uninsured tended to be severely underfunded.

More recent authors have developed the concept of social policy regimes or models of social policy to denote
1
the totality of social programmes. Typically, these authors are concerned with the distributive e ects of the
structure and generosity of social programmes and therefore include not only transfers and health services
but also education in their analyses. The distribution of access to quality education is an important
determinant of poverty and inequality in Latin American societies. It is worth noting that price controls and
subsidies constituted an important instrument of social protection for the poorer sectors of the population
before the debt crisis of the 1980s and the subsequent opening and deregulation of the Latin American
economies. However, the general welfare state/social policy literature did not include systematic analyses of
price controls and subsidies.

The theoretical perspective o ered by the pioneering works that sought to explain the formation of social
security regimes was a combination of economic development and pressure group politics (Mesa‐Lago 1978;
Malloy 1979; Isuani 1985). Social strati cation came to be re ected in the social security systems of Latin
American countries because the most important pressure groups over time managed to extract from the
state their own schemes of social protection. Typically, the military received the earliest and best social
protection, accompanied or followed by the judiciary and other high‐ranking civil servants, then
professionals and other white‐collar sectors, and nally unionized blue‐collar workers in the urban sector
and on rural plantations (Mesa‐Lago 1978). Only later and only in some countries did the bulk of the rural
population become incorporated into social security. Extensions of social security protection were not
always a response to direct pressure from organized groups; the alternative consisted of state incorporation
attempts of groups that elites perceived to be potential threats or power bases, or concerns with state
building and modernization (Mesa‐Lago 1978; Spalding 1978; Papadópulos 1992).

The uneven pace of industrialization led to uneven development of social security schemes. In the pioneer
countries (Argentina, Brazil, Chile, Uruguay, and Cuba), the main social security schemes were established
in the 1920s and grew in a gradual and fragmented manner. In the second group (Colombia, Costa Rica,
Mexico, Paraguay, Peru, and Venezuela), the main schemes were installed in the 1940s and tended to grow
in a less fragmented manner. In the remaining countries, the least economically developed ones, social
security schemes appeared even later (Mesa‐Lago 1989: 3–6). The pioneer countries plus Costa Rica had
achieved the highest levels of social security health coverage by 1980, over 60 per cent of the population;
Colombia, Guatemala, Mexico, Panama, Peru, and Venezuela covered between 30 and 60 per cent of their
p. 646 populations; in the rest of the countries coverage was below 30 per cent. The English‐speaking Caribbean
was an important exception, with coverage of 80 per cent or higher, despite the low level of economic
development. The gures for pension coverage of the economically active population showed a similar
pattern (Mesa‐Lago 1994: 22).

More recently authors have used a power constellations perspective, developing their analyses in dialogue
with welfare state theories formulated for advanced industrial countries, and focusing on structural
 changes (industrialization and urbanization) leading to changes in political regimes (democratization)
and/or power distributions (strength of labour and the left) and political alignments or coalitions (Dion
2005; Filgueira and Filgueira 2002; Haggard and Kaufman 2008; Huber et al. 2008; Pribble 2008; Segura‐
Ubiergo 2007). In particular, they have emphasized the political economy of Import Substitution
Industrialization (ISI), which on the one hand fostered growth of urban labour and unionization and
enabled employers to pass on high social security taxes to consumers behind high tari walls, but on the
other hand developed only a limited capacity to absorb the swelling ranks of the urban labour force and thus
left a rapidly growing urban informal sector and the still large rural sector without coverage. These authors
have also identi ed di erent paths to extensive social security coverage, through either high

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

industrialization and strength of labour and left‐of‐centre parties, with or without long records of
democracy, or in the absence of high industrialization through long records of democracy and strength of
left‐of‐centre parties (Huber and Stephens 2009; Haggard and Kaufman 2008; Pribble 2008; Segura‐Ubiergo
2007; Martínez Franzoni 2008).

Reforms of Social Policy Regimes

The literature here is quite voluminous. The studies of the de ciencies of the social security systems as they
had developed by the early 1980s emphasize the large scal de cits faced by the social security schemes in
the more advanced countries. These schemes had matured and the ratio of active to inactive a liates had
deteriorated; the surpluses in the social security funds had been poorly invested or spent on other purposes
altogether; life expectancy rose; expenses for costly curative healthcare escalated; employers evaded or
greatly delayed contributions, particularly during periods of high in ation; and administrative costs were
exorbitant (Mesa‐Lago 1994). The debt crisis and the ensuing austerity and structural adjustment policies
aggravated these problems. Rising unemployment and growing informalization caused declines in social
security contributions; the scal crisis of the state reduced state subsidies; high in ation eroded the value of
bene ts or put great strain on the systems and created a debt to bene ciaries. The de cits of the social
security systems aggravated the general scal crisis of the state and attracted the attention of the
International Financial Institutions (IFIs), particularly the International Monetary Fund (IMF).

The policy prescriptions issued by the IFIs and the United States Agency for International Development (US
AID) applied neoliberal principles to both economic and social policies. Essentially, these agencies pushed
for cuts in government expenditures, liberalization of trade and capital ows, privatization of state
p. 647 enterprises, relaxation of economic regulations, and incentives for foreign direct investment, along with
a reduction of the role of the state in nancing social policy and in providing social services in favour of
greater involvement of the private sector in nancing and providing social insurance and social services
(Williamson 1990; World Bank 1994 b). These prescriptions, however, were di erentially implemented,
dependent on the balance of power between the IFIs and internal advocates of neoliberalism on the one
hand, and stake holders in the existing systems and advocates of more state‐oriented reforms on the other
hand (Huber 2005).

Alternative prescriptions of a more social democratic nature were o ered, increasingly so by left‐leaning
think tanks in Latin America, but they lacked political and nancial backing. The International Labour O ce
emphasized the principles of equity and solidarity in labour market and social policy (Bertranou et al. 2002),
but it also lacked the nancial clout to be able to compete with the IFIs. In the late 1990s, some forces in the
IFIs engaged in important re‐examinations of the reforms and their shortcomings (Holzmann and Stiglitz
2001; De Ferranti et al. 2004; both volumes published by the World Bank). Finally, after two decades of
disappointment with the e ects of neoliberal social policy reforms, the Inter‐American Development Bank
gave exposure to an alternative vision by supporting a project on Universalismo Básico (Basic Universalism)
(Molina 2006). The coming to power of left‐of‐centre governments in the rst decade of the twenty‐ rst
century reinforced interest in comprehensive and equity‐oriented social policy.
 Pension privatization was at the forefront of neoliberal policy recommendations, but in reality the reforms
ranged all the way from putting existing public systems on rmer nancial footing, as in Brazil, to full
privatization and a closing down of the public system, as in Chile (Müller 2003). Other countries, like
Argentina and Uruguay, kept a basic public tier and privatized supplementary pensions or pensions for
higher income earners. Madrid (2003: 16) developed a useful index of degree of pension privatization based
on the percentage of total contributions going to the public system and the percentage of total members
belonging to the private system, which showed Bolivia, Nicaragua, and Chile scoring highest, followed by El
Salvador, the Dominican Republic, and Mexico. He explains pension privatization with the economic burden
of pension spending, domestic capital shortages, in uence of the World Bank and of economists, and the

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

extent of party discipline in the President's party and his control over the legislature (2003: 59).

Other authors agree on the importance of World Bank in uence and of the networks of domestic
technocrats in close contact with the Bank (Huber 2005; Teichman 2001) but also emphasize domestic
power constellations and constitutional structures, such as the existence of the referendum in Uruguay
(Castiglioni 2005; Hernández 2000, 2003; Huber and Stephens 2009). Finally, Weyland (2006) emphasizes
cognitive processes of decision‐makers, arguing that policy di usion happened because technocrats and
politicians working under time and nancial pressures took cognitive shortcuts to process lessons from the
Chilean example.

Reforms in other social policy areas, particularly health and education, were more varied than in pensions.
p. 648 There was no neoliberal blueprint out there similar to the pension privatization model, though the World
Bank exerted general pressures towards decentralization of service delivery and greater involvement of the
private sector. As Kaufman and Nelson (2004) and the various contributors to their volume argue, reforms
in health and education had to deal with more stakeholders than pension reforms, particularly among the
providers of the services. Typically, these stakeholders were well organized, from doctors and hospital
associations to teachers' unions, and thus had the capacity to exercise e ective resistance.
Administratively, it is more di cult to design functioning systems to deliver quality health and education
than transfers. Accordingly, reforms and their implementation were heavily shaped by the interaction
between coalitional alignments, executive involvement and follow‐up, and policy legacies that manifested
themselves in domestic institutional arrangements and interest groups.

Studies of the outcomes of pension reforms found that they were disappointing in terms of the expected
expansion of coverage, the burden of administrative costs, and competition among pension funds, along
with equity, solidarity, and gendered impacts (Cruz‐Saco and Mesa‐Lago 1998; Ewig 2008; Dion 2006; Huber
and Stephens 2000 b; Kay and SinHa 2008 and contributors to their volume). The report commissioned by
President Bachelet in Chile highlighted these shortcomings and gave rise to a signi cant re‐reform of the
Chilean pension system, expanding coverage for those not covered or having accumulated insu cient funds
in the private system (Consejo de Reforma Previsional 2006). Useful review essays are Dion (2008) and
Mesa‐Lago (2008).

Some studies have looked speci cally at social policy reforms and evaluated their gendered impact. In
general, they have found that greater reliance on the market and the private sector intensi ed the
disadvantage of women. Women in general are in a weaker position in the market because they have more
breaks in their work lives, they are more likely to work in the informal sector, and they are earning less than
men in both the formal and informal sectors. Accordingly, contribution‐based social insurance leaves them
with less protection. In addition, their reproductive role and their social role as caregivers means that they
require more healthcare both for themselves and their children and more support services to be able to
combine family responsibilities and paid work (Castiglioni 2005; Dion 2006; Ewig and Bello 2009; Martínez
Franzoni, 2008).
 Origins and Current Impact of Social Policy Regimes: The Evidence

Origins of E ective Social Policy Regimes

E ective social policy regimes should result in low levels of poverty and inequality and high levels of human
capital, or at least in signi cant and sustained movement towards lowering levels of poverty and inequality
p. 649 and improving the human capital base. Of course, the policy regimes should demonstrate their
contribution to such an outcome in the form of social spending as a percentage of GDP and the allocation of

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

spending in a progressive manner. By these standards, there are no fully successful social policy regimes in
Latin America. To begin with, social expenditures are comparatively low. Total social expenditures (for
social security and welfare, health, and education) in 2002–3 surpassed 15 per cent of GDP only in
Argentina, Brazil, Costa Rica, Panama, and Uruguay; Chile, Bolivia, Colombia, Venezuela, and Mexico spent
between 11 and 15 per cent; the remaining countries spent less than 10 per cent of GDP.

Low spending is intimately related to the weakness of the tax base. As a consequence of trade liberalization,
revenue from trade taxes fell; the share of total tax revenue coming from taxes on foreign trade fell from 18
per cent in 1980 to 14 per cent in the mid‐1990s. Tax reforms reduced marginal tax rates on personal income
and taxes on corporate pro ts, and they increased reliance on value added taxes. However, collection rates
have remained lower than the statutory rates (Lora 2001), and average tax revenue has remained low
compared not only to OECD countries but also to East Asian countries at similar levels of development.
Average tax revenue in Latin America, including social security contributions, in 2002 was 16 per cent of
GDP, compared to 36 per cent for OECD countries in general and 26 per cent for the US and Japan, the two
advanced industrial countries with the lowest tax burdens (Centrángolo and Gómez Sabaini 2007: 53). There
are major di erences between Latin American countries, Brazil being at the top end with a tax burden that
had surpassed 30 per cent of GDP by 2005, but Chile, for instance, remaining at only 18 per cent of GDP.

Correspondingly, outcome indicators of social policy leave much to be desired. All countries have rather
large groups that live in poverty, with Uruguay being the only country in the 1990s where less than 10 per
cent of households were below the poverty level de ned either by a basket of basic necessities (Economic
Commission for Latin America and the Caribbean (ECLAC)) or by the $2 per day in purchasing power parity
(World Bank). The economic crisis of the early 2000s in Argentina and Uruguay saw poverty rates shoot up
to 26 per cent in Argentina and 19 per cent in Uruguay in 2005. In that year, Uruguay and Chile had the
lowest poverty rates in Latin America with 19 per cent, thus leaving close to a fth of their populations in
poverty. Moreover, almost all countries have rather large groups with access to poor quality education and
healthcare only.

Compared to the rest of Latin America though, Argentina, Chile, Costa Rica, and Uruguay have clearly
developed the most e ective social policy regimes. Brazil is high on social spending but low on the outcome
indicators (Huber and Stephens 2009). The four comparatively successful countries exemplify three paths
to e ective social policy regimes: signi cant industrialization under populist‐authoritarian auspices
(Argentina), or democratization with (Chile, Uruguay) or without (Costa Rica) signi cant industrialization,
but with the presence of a powerful left‐of‐centre party or coalition (Pribble 2008; Segura‐Ubiergo 2007).
Quantitative studies have con rmed that length of the democratic record and left strength in the legislature
p. 650 are signi cantly associated with lower poverty and inequality in Latin America (Huber et al. 2006).
 Reforms and Distributive Outcomes of Social Policy Regimes
A large number of studies have looked at the distributive impact of speci c social policies (e.g. De Ferranti et
al. 2004; CEPAL: Social Panorama, various years) and a few studies have tried to construct a synthetic view
of the distributive impact of the totality of transfer policies in several countries (Lindert et al. 2006). They
have uniformly found that social security is highly regressive and social assistance highly progressive, but
that social security by far outweighs social assistance and thus gives a highly regressive pro le to the
totality of social transfers, in stark contrast to the impact of transfers in advanced industrial countries. In
education and health, programmes focusing on basic education and primary and preventive healthcare are

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

progressive, whereas expenditures for tertiary education and expensive curative medicine are regressive.
Overall and on average, health expenditures are distributively neutral and education expenditures slightly
progressive. However, there is considerable variation between countries.

As noted quantitative studies have found that length of the democratic record and long‐term strength of
left‐of‐centre parties depress both poverty and inequality, but they have also found that these variables do
not signi cantly in uence social security expenditures. Rather, social security schemes are hard to modify
once put into place, be it by authoritarian or democratic regimes, and they push up expenditures with an
ageing population. There is no conclusive evidence about the resilience of di erent types of social spending
in the face of budget de cits. While Huber et al. (2008) nd that social security expenditures are
comparatively more resilient than health and education expenditures, Kaufman and Segura‐Ubiergo (2001)
contrarily suggest that the area of social security transfers (mainly pensions) is the most vulnerable, while
health and education expenditures are far less so. Another important nding in terms of social expenditures
in the region (and the developing world) is its countercyclical nature (Wibbels 2006). Wibbels found that
given the patterns of integration into global markets, where shocks associated with international markets
are profound and access to capital markets in di cult times is limited (in comparison with developed
nations), Latin American governments have more incentives to balance budgets by cutting social spending,
and so are unable to smooth consumption across the business cycle.

Government spending in a democratic context does reduce income inequality; this is true in a worldwide
sample (Lee 2005) as well as for social transfer expenditures in Latin America (Huber et al. 2006). In Latin
America, length of the democratic record is highly correlated with strength of left‐wing parties. This is due
to the fact that the main alternative to democracy was authoritarianism of the right, not the left. Right‐wing
authoritarian regimes repressed organized labour and left‐wing parties, whereas unions and parties of all
stripes were allowed to form and consolidate under democratic regimes. Left‐of‐centre parties produced
redistribution not by overall higher expenditures (no statistical e ect on social expenditures) but by the
allocation of these expenditures. Of particular importance is spending on non‐contributory programmes for
p. 651 those outside the formal labour market. This includes non‐contributory pensions as well as income
support for working‐age families, such as family allowances and conditional cash transfer programmes.

Conditional cash transfer programmes (CCTs) have become popular in Latin America. As of 2005, at least
nine Latin American countries had introduced a programme that provides cash to poor families (preferably
the mother) in exchange for keeping the children in school and under regular medical supervision for
vaccinations and other preventive care (Lindert et al. 2006). Not all of these programmes were introduced
by left governments; one of the best known is ‘Oportunidades’ in Mexico introduced by President Fox in
2
2002. However, the largest of these programmes was developed under President Lula da Silva of Brazil; by
2006, Bolsa Familia had grown to reach 11 million families, or over 20 per cent of the Brazilian population,
and it was credited with playing an important role in Lula's re‐election that year (Hunter and Power 2007).
Early evaluations of these programmes show that they are not only e ective in reducing poverty but also in
improving school attendance and health outcomes (Rawlings and Rubio 2005).
 Emergency employment programmes introduced in the wake of economic crises constitute another
important kind of anti‐poverty programme. Argentine employment programmes, in their latest incarnation
known as Plan Jefes y Jefas de Hogar Desocupados, grew dramatically from 1993 to 2002 to cover 1.5 million
heads of household, and they survived into the recovery phase. On the one hand, these programmes did play
an important role in poverty alleviation, but, on the other, they also became a political tool in the hands of
the central government to build alliances with provincial politicians and placate areas with strong protest
movements (Giraudy 2007).

The left‐wing Frente Amplio (FA) government in Uruguay launched a particularly comprehensive anti‐

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

poverty programme in 2007, to replace the two‐year emergency social programme it had launched upon
taking o ce in 2005. The new Plan Equidad is available to all low‐income households and includes family
allowances contingent on children's school attendance, non‐contributory social assistance pensions for
individuals from 65 years of age, incentives for attendance in secondary schools, expansion of preschool
education, support for employment, nutrition cards, and subsidies for electricity and water costs (Cuenca
2007). Estimates put potential bene ciaries at up to one million citizens, or about one third of the
population (Campodónico 2007). The Chilean government under Socialist President Lagos launched a
similarly comprehensive programme, Chile Solidario, but on a much smaller scale, directed only at
households in extreme poverty, or less than 6 per cent of the population. President Bachelet then attempted
to extend the programme by incorporating new groups into the programme, primarily the homeless.

Universal pre‐school education has been on the agenda of progressive governments, both for reasons of
preparing children from poorer backgrounds to succeed in school and of enabling mothers to enter gainful
p. 652 employment. Thus, free high quality public pre‐school education is a means to combat poverty and
inequality both in this generation and the next. Latin America has signi cantly improved pre‐school
coverage during the last two decades. In 2005, more than 84 per cent of children in the region attended the
last year of pre‐school education (ECLAC 2007). In Uruguay, pre‐school education coverage of 5‐year‐old
children from the lowest quintile increased from 64 per cent in 1991 to 94 per cent in 2007, and coverage of
4‐year‐old children increased from 27 per cent to 72 per cent in the same period (Cardozo 2008). The gap
between the lowest and highest income quintile narrowed from 62 to 23 points for 4‐year‐old children and
from 33 to 5 points for 5‐year‐olds. This improvement is mainly a consequence of the expansion of public
coverage by the 1995 reform. In Chile, while pre‐school education experienced no signi cant improvements
in coverage during the 1990s (ECLAC 2002), the Bachelet government announced a major expansion, setting
the goal of incorporating 180,000 children at the 4‐year‐old level into pre‐school education for 2010, the
equivalent of 37 per cent of non‐matriculated children in 2008 (OAS 2008).

Healthcare reforms have varied widely, heavily contingent on policy legacies. All of the more advanced Latin
American countries have been facing growing elderly populations requiring more healthcare services. The
cutbacks in public expenditures in the wake of the debt crisis had left public systems with severely
underpaid personnel, outdated equipment, and facilities in dire need of repair. Signi cant growth in public
health expenditures from the early 1990s on remedied some of these de ciencies but was unable to close the
quality gap between the public and private systems. Those who could a ord it continued to resort to private
alternatives.

The most recent round of Costa Rican and Uruguayan reforms arguably went furthest towards providing
high quality healthcare for the entire population. The Chilean reform aimed in the same direction but
encountered stronger private sector opposition. Under the previous reform implemented by the Pinochet
regime, formal sector employees could choose to have their mandatory health insurance contributions go to
the public system or a for‐pro t private provider that could charge additional premiums and discriminate on
the basis of risk. Both the Costa Rican and the Uruguayan systems were insurance‐based and covered
virtually the entire economically active population; contributions in Costa Rica were subsidized for low‐
income earners in the informal sector. The non‐insured poor in all three countries had a right to free care in
 public facilities. The Costa Rican system had been a uni ed public one since the 1970s, which means that the
poor were served by the same facilities as the insured population, whereas the insured in Uruguay relied
mostly on not‐for‐pro t private insurers and providers.

The Costa Rican reform expanded primary care centres throughout the country in order to improve access to
healthcare for all. The Uruguayan reform introduced centralized nancing, with public and private
providers alike receiving a per capita payment for patients under their care, and those making the
mandatory contributions having a choice of providers. The reform also strengthened the network of
primary care clinics (Pribble 2008). The Chilean reform under President Lagos guaranteed universal

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

p. 653 coverage of fty‐six common illnesses, with co‐payments being capped in a progressive manner
according to family income, and time limits set within which treatment has to be provided. If the public
sector is unable to provide care, treatment in the private sector will be covered. In Uruguay and Chile, the
groups that are left out of these reformed systems are the non‐poor informal sector workers; in Costa Rica,
strong e orts have been made to include them through subsidies for their contributions. In sum, the Costa
Rican and Uruguayan reforms go far towards basic universalism, whereas the universalism of the Chilean
reform remains con ned to the speci ed set of illnesses.

Brazil's health system illustrates the gap between constitutional rights to universal free healthcare and the
practice of accessibility of services. The Brazilian constitution of 1988 granted this right to all citizens and
established a uni ed public health system like the Costa Rican one, but implementation was left up to the
legislature and the Ministry of Health. Little progress was made before the administration of President
Cardoso who gave backing to a strong Minister of Health to push forward with improvements in universal
basic healthcare. As a result, indicators like infant mortality, vaccination, and maternal health improved
signi cantly (PNUD 2005). However, despite the fact that they and their employers are required to pay into
the public system (in contrast to Chile), higher income earners heavily use private insurers and providers
and rely on the public system for the most expensive kinds of care. Moreover, gaps in access to complicated
care remain huge across classes and gaps in access even to primary care remain great across regions
(Arretche 2006).

The e ects that democracy and the strength of left parties have on reducing poverty and income inequality
work not only through allocation of social expenditures but also through other legislation, such as the level
of minimum wages and support for unions and coordinated wage setting policies. In Chile and Costa Rica
the minimum wage is an important instrument that can keep a small household above the poverty line. The
ratio of median wage for all wage and salary earners to the minimum wage in 2002 was roughly 2.8:1 in
Chile (Marinakis 2006: 6) and 2.2:1 in Costa Rica (Estado de la Nación 2004: 409–10). Since 1990, the real
minimum wage has stayed roughly constant, with a slight upward trend in Costa Rica, whereas it has
increased steeply in Chile; in Argentina it deteriorated dramatically after 2001. In Uruguay, the national
minimum wage is not a relevant oor; it deteriorated consistently and strongly from 1985 to 2005, when the
FA government doubled it. More important have historically been tripartite wage setting councils, from
which the state withdrew under National Party President Lacalle in 1992, but which were re‐established
quickly by the FA government (Rodriguez et al. 2007). Under the FA government, rural and public sector
workers were included for the rst time in collective negotiations, and salaries improved in real terms.

Globalization and Social Policy

The debate about globalization and social policy has overcome initial extreme positions, both with regard to
p. 654 advanced industrial and developing countries. As a result of much research, an intermediate position has
gained ground which holds that globalization has no direct e ects, either negative or positive, on social
spending, but rather that domestic forces and institutions serve as mediating factors (see e.g. the
contributions to Glatzer and Rueschemeyer 2005).
 In statistical analyses, Garrett (2001) in a sample of developed and developing countries, and Kaufman and
Segura‐Ubiergo (2001) in a set of Latin American countries nd negative e ects of trade integration (i.e.
changes in trade openness) on changes in general government consumption and in social spending,
respectively. Kaufman and Segura‐Ubiergo also nd negative e ects of the level of trade openness and of the
interaction of trade and capital market openness on changes in social spending. Avelino et al. (2005) nd a
negative e ect of trade openness but a positive e ect of nancial openness on levels of social spending in 19
Latin American countries between 1980 and 1999. Kaufman and Segura‐Ubiergo (2001) nd unstable e ects
of levels of capital controls on changes in social expenditures and weakly positive e ects of changes in
capital controls. Rudra (2004) nds a positive e ect of capital ows on the level of social expenditures in a

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

worldwide sample, and Garrett (2001) nds no signi cant e ect of levels and changes in capital ows on
levels of government consumption.

The inconsistency of the ndings has various explanations, such as the particular cases and time periods
examined, di erent sets of control variables, di erent operationalizations of the dependent variables, and
di erent analytic techniques. Arguably the most serious reason is the use of di erent control variables, in
particular the omission of domestic political variables. Results of statistical analyses that include
urbanization, democracy, and long‐term strength of di erent political tendencies in the legislature or the
executive, indicate that trade openness and foreign direct investment have no e ects on social
expenditures, neither social security and welfare nor health and education expenditures (Huber et al. 2008).

Future Directions in Research on Latin American Social Policy Regimes

The distributive impact of social policy.

So far, comparative studies of the distributive impact of social policies have been hampered by the lack of
comparability of the data. There are excellent country case studies based on analyses of national household
surveys, and some attempts to compare the results of these analyses across countries, but these national
surveys use di erent concepts and/or equivalency scales for household size, which makes strict comparison
essentially impossible. Now the Luxembourg Income Study is adding several Latin American countries to
p. 655 their data base, which will be a huge step forward. It will allow systematic comparative examination of
the distributive impact of social policies on di erent social groups (by income level, gender, level of
education, geographic location, etc.).

Nexus to labour market policy.

It is clear that overcoming poverty requires employment, both at the individual and the societal level. At the
societal level, the transition from ISI to the neoliberal model caused deindustrialization, a loss of jobs in the
formal sector, and rapid growth of informal employment. Even in the post‐adjustment period, job growth
was concentrated in the informal sector, with low productivity and therefore low wages (Tokman 2002). The
commodities boom starting in 2004 stimulated growth in the Latin American economies, but it did little to
improve the structural weaknesses of Latin American economies. Most governments remained reluctant to
engage in industrial policy, and private sector initiatives for productivity increases are few and far between.
Research into options for job creation linked to active labour market policies to make quali ed workers
available will be an essential component of the entire welfare state research programme in Latin America.

In the area of labour market policy, governments have had great di culty making the transition from
traditional forms of employment protection to more exible labour markets and policies that would still
protect workers from falling into poverty with their families. Unemployment insurance is a new feature and
has been introduced in only some Latin American social policy regimes, and even where it exists, coverage
and bene ts are very small. Active labour market policies in the form of job training and support for job
placement are few and far between. Some pilot programmes have linked conditional cash transfer
programmes with labour market training and jobs in the public sector. This is clearly an important direction
both for practical policy experiments and for research.

The politics of social policy.

Another major area of developing research is the comparative analysis of political struggles over speci c
policy initiatives. Understanding these struggles is important both theoretically, to help scholars build

Downloaded from https://academic.oup.com/edited-volume/34386/chapter/291616258 by Columbia University user on 28 February 2023

better models of welfare state development in Latin America, and practically, to help governments
committed to welfare state construction build the most e ective political alliances. As more materials, such
as legislative debates, party positions, and newspaper archives become available on line, the details of the
politics of policy will become more amenable to systematic comparative analysis. These kinds of data
sources will never substitute for interviews with those directly involved in the struggles, but they will
constitute the basis for triangulation of information and for systematic testing of generalizations.

Notes
1 The literature on advanced industrial societies distinguishes between three or more welfare state regime types (e.g.
Esping‐Andersen 1990); the literature on Latin America uses the concept of regimes simply to denote the total
configuration of social policies.
2 The programme was based on a previous one, PROGRESA, initiated in 1997.