The Politics of Social Policies in Latin America

Few things seem as directly meaningful as whether our children live to adulthood, whether we can feed our families, get medical care when seriously ill, and sleep in conditions of basic dignity. Over the last quarter century, these conditions have undoubtedly improved for most of the approximately 650 million people who live in Latin America. While a broad variety of factors influence these social conditions, including the economy, climate, technological change, and even charity, public policies are central mediating factors. Social policies can produce remarkably different outcomes for people who otherwise live in similar conditions. As Alberto Díaz-Cayeros, Federico Estévez, and Beatriz Magaloni note, government policy makes a huge difference in whether “children die from preventable diseases” (2). Until the end of the twentieth century, Latin American countries were characterized, in broad strokes, by “truncated” or “dualistic” welfare states.1 During the twentieth century, states extended social security to formal-sector workers and their dependents, but given the high levels of informality, the majority of

the population remained uncovered. Meager public systems targeted to the poor, many times through clientelistic policies, provided inadequate transfers and services to these "outsiders." The divide was so stark that until recently, "Latin American social policy … has mostly historically worked backward, making preexisting economic and social inequalities wider rather than narrower" Estévez,and Magaloni,28).
Since then, the region has witnessed a fairly massive expansion of social policies to include more of the mostly low-income outsiders than ever before. By 2012, seventeen countries in Latin America had adopted conditional cash transfer programs (CCTs) aimed at reducing child poverty and encouraging health checkups and school attendance, collectively reaching twenty-seven million of the poorest households in the region (De La O,3). Health care coverage for outsiders increased dramatically as well, as did old-age pensions for those who were not part of the contributory system. These expansions, together with economic growth and employment, reduced both infant mortality and poverty. According to World Bank data, the infant mortality rate in Latin America and the Caribbean has declined from 43.8 per 1,000 live births in 1990 to 14.9 per 1,000 live births in 2016. Extreme poverty dropped by half, from 26.3 percent in 1995, to 13.3 percent by 2011. 2 Even inequality went slightly down, in a region known for the highest inequalities in the world. 3 Yet, concurrently with the expansion, dramatic variation remained in coverage, generosity, and effectiveness. The books under review here chart aspects of this expansion and the variation in it, and seek to explain its causes and in some cases implications. I will discuss the findings of the books in relation to two broad themes: first, the causes of policy expansion and the dynamics of policy design; and second, broadening the conceptualization of what we consider welfare policy.
The books chart a similar landscape: policy expansion, with variation. What has been remarkable about this expansion, in comparison to the region's past, is not only the active inclusion of outsiders but the choice of some of the governments, in Ana De La O's words, to "tie their own hands" in policy design; that is, to reduce executive discretion that allows for the clientelistic manipulation of policies for political ends.
The books challenge two prevalent explanations for social policy expansion since the turn of the millennium: the commodity boom and the shift to the left. Undoubtedly, the commodity boom of the first decade of the twenty-first century gave governments more room to maneuver. However, as Candelaria Garay shows in Social Policy Expansion in Latin America, "the occurrence and the timing of expansion … do not coincide neatly with the timing of the boom," as "incumbents adopted new social policies for outsiders at times of both economic strain and abundance" (9). Four of the books-The Political Logic of Poverty Relief, by Díaz-Cayeros, Estévez, and Magaloni; Crafting Policies to End Poverty in Latin America, by Ana De La O; Social Policy Expansion in Latin America, by Candelaria Garay; and Forbearance as Redistribution, by Alisha C. Holland-also explicitly downplay the role of left/right ideology in the policy process. This runs counter to recent works that have highlighted the role of ideology in social policy expansion, for example, Democracy and the Left by Evelyne Huber and John Stephens, which finds that together with years of democracy, left political strength contributed to more social spending and to a reduction in poverty and inequality. 4 Other recent works have also highlighted the centrality of left governments in bringing about more active propoor policies. 5 What these four books do instead is highlight the importance of electoral competition. When governments need the votes of outsiders, and/or when they face electorally successful opposition parties, they are more likely to behave well and deliver good policies, these books argue. They make a compelling case for electoral competition in explaining policy expansion, although the role of ideology versus other factors in policy design remains less settled.
The second theme I discuss is broadening the conceptualization of what we consider welfare policy, which Holland has brilliantly done in her book Forbearance as Redistribution. Holland examines an area that has been neglected in welfare policy research in Latin America: in the absence of formal access, the informal (and illegal) securing by the poor of two of the most basic material needs: housing and employment. These legal violations "function as a means of securing basic welfare" (27). Given the lack of alternatives for the poor, governments often choose to engage in "forbearance," which is "intentional and revocable government leniency toward violations of the law" (6) on squatting and street vending. Debunking explanations based on state capacity, Holland shows how governments at the local and national levels respond to electoral incentives in deciding when to enforce: "Politicians turn to forbearance when welfare policies are inadequate and they need the poor's support to take or retain office" (3). She also examines the mechanisms that make it difficult for governments to get out of what she calls the "forbearance trap" and the conditions under which they are able to move to more institutional policies and more consistent enforcement of laws.

Explaining Policy Expansion and Design
The books highlight the importance of electoral democracy in policy reforms. Candelaria Garay focuses on the big picture, seeking to explain social policy expansion toward outsiders in pensions, income support, and health care from the 1980s to 2010 in Argentina, Brazil, Mexico, and Chile. She defines outsiders as workers outside the formal labor market and their dependents, and defines social policy expansion as "the creation of new social benefits or the extension of preexisting ones to a significant share-at least 35 percent-of the outsider population" (29). Garay includes three dimensions in her characterization of social policy expansion: scope of coverage (proportion of relevant outsider population), generosity of benefits (amount of transfer/service provided compared to benefits for insiders, defined as low-income formal sector workers), and participation of beneficiary organizations in policy implementation. To give the reader an idea, she classifies scope of coverage as "high" if it reaches at least 70 percent of the relevant outsider population; moderate if it reaches between 50 and 70 percent; and low if it reaches between 35 and 50 percent. Programs that reach less than 35 percent of the outsider population are not considered "expansions" (Garay, 340).
Garay argues that expansion occurred in democratic regimes when incumbents considered social policy expansion a powerful instrument to elicit outsiders' electoral support when a credible challenger threatened to defeat the incumbent party by courting outsider voters, and/or necessary to mitigate intense social mobilization. In the absence of these conditions, expansion of large-scale nondiscretionary benefits was unlikely (18). When sustained social mobilization was exercised "by a coalition of social movements and labor unions making demands on the state for broad, universal benefits for outsiders" (40-41), policy was likely to become more inclusive. Garay highlights the different ways that social movement coalitions can influence policy design: either by allying or being part of the government, or from the streets. The latter, she finds, can be more effective because incumbents want people off the streets. Thus, in both Chile and Mexico, which lacked such mobilization, "policy design" was negotiated "exclusively with the parties in Congress," resulting in "restrictive policy," whereas in Brazil and Argentina, negotiations "with social movements in addition to, or instead of, parties in Congress" (19) produced "inclusive policy." De La O's study seeks to explain the expansion in antipoverty policies specifically between 1990 and 2011, with a regional-level quantitative analysis; process tracing in Colombia, Guatemala, Peru, Argentina; and an in-depth study of Mexico. She argues that in the case of CCTs, it was economic crises, in contexts of electoral competition, that brought concern about poverty to the fore of the political agenda. Governments have to appear to be doing something, she argues, and electoral candidates have to promise to do something. In this context, CCTs are an easy policy to propose as they are relatively inexpensive, especially "in lieu of more comprehensive policy responses" (47), and they do not affect the interests of vested organized actors (14).
Augusto De Venanzi's book Sistemas políticos y bienestar social explores social policy expansion and outcomes in three major Latin American countries with governments of different political orientation after 2000: "moderate left" Brazil, "radical left" Venezuela, and right-wing Colombia. By choosing his cases on the basis of governing party ideology, De Venanzi, unlike the above-mentioned authors, makes that ideology a central part of his analysis. Through an examination of each political system, he highlights the differences between the two lefts, with many policy and development indicators placing Brazil and Colombia closer to each other than to Venezuela.
The Quest for Universal Social Policy in the South, by Juliana Martínez Franzoni and Diego Sánchez-Ancochea, presents a normatively ambitious project in which the authors seek to uncover the conditions under which universal social policies, defined as "policies that reach the entire population with generous transfers and services" (6), come into being in the Global South, with case studies of Costa Rica, Uruguay, South Korea, and Mauritius. They define universalism as a set of "desirable policy outputs: similar, generous entitlements for all" (4). In doing so, they explicitly distinguish themselves from the welfare state literature that they argue sets the Nordic type of universalism (mechanisms and outputs) as the model to aspire to. In other words, it doesn't matter how states get to universalism, whether through social insurance or social assistance; what matters is the result.
Martínez Franzoni and Sánchez-Ancochea examine universal outputs primarily through health care policy (and secondarily through pensions) with measures of access, generosity, and equity for salaried workers, the self-employed, and the poor. Like the other authors under review, they argue that democracy is important in bringing these issues onto the political agenda. They also highlight the role of "progressive political leadership," that is, government ideology, in promoting universalism. To make their case, they compare trajectories in their case studies, which, they argue, "stand out for having created active social policies at one point or another" (13).
In the context of restrictive policy legacies and truncated welfare states, policy expansion is a necessary but not sufficient condition to reach the poor and the outsiders; equally important in terms of actual outputs is, as many of the authors stress, policy design.
Garay highlights the "preferences of those engaged in the design process, and their institutional power." She recognizes that party ideology may play a role in policy design, as can opponents, in the form of program stakeholders and subnational authorities. While neither of these opponents were able to veto expansion when outsider competition and mobilization were present, they could influence policy design, she argues. This said, she finds that neither party ideology nor stakeholder opposition overall "[altered] the policies significantly" and was not decisive in producing inclusive versus restrictive policy (73).
Martínez Franzoni and Sánchez-Ancochea examine, aside from democracy and "progressive political leadership," the role of what they call policy architecture-which they define as the combination of instruments that define who has access to what benefits, and how-in promoting universal outputs. In order to "deliver universal outputs through a diverse set of instruments including social insurance and social assistance," unification is "a key feature that must accompany any given policy instrument. Unification takes place when all beneficiaries receive the service or cash transfer in a similar fashion and the state plays a major role in defining benefits, acting as direct provider and effectively regulating the market" (19). Fragmentation of the policy architecture, defined as "the existence of diverse mechanisms of access, funding, entitlements, and providers, as well as the presence of a large outside market option" (19), on the other hand, is a threat to universalism. They argue that the "presence of state actors capable of promoting unified architectures through the adaptation of international ideas" (4-5) is important in influencing the policy architectures and, in turn, the potential for universal outputs (57).
Alejandro Bonvecchi, Julia Johannsen, and Carlos Scartascini title their book "Who Decides Social Policy?" They provide an overview of political economic and social networks analysis of social policy, and (with others) conduct meticulous case studies of specific antipoverty policies in Bolivia and Peru. They find that officials in Bolivia are more linked to social actors and more often take their views into account than in Peru.
Both Díaz-Cayeros and colleagues and De La O construct formal models to explain the individual-level incentives that drive better, or worse, policy among policymakers, arguing that electoral competition drives the latter. Díaz-Cayeros and his coauthors examine social policies in Mexico, including both antipoverty cash transfers as well as public goods provision such as water, sanitation, and electricity. They develop a formal model of vote buying to predict when governments target core supporters with clientelistic policies and when they deliver "pork" to a constituency. In either case, political rather than technocratic or "propoor" criteria dominate. Through analysis of subnational variation in Mexico over time, they show how the macropolitical shifts in Mexico in the 1990s (including the Zapatista rebellion and threat of social unrest, voter rebellion against corruption, the PRI losing Congress, and the threat of the PRI losing the presidency) changed the political context and electoral calculations of the governing parties, forcing them to deliver better public policies. Electoral competition was the key component: "the share of public goods over particularistic transfers in a given geographic area will increase as a party's core base proves insufficient to guarantee victory" (12). These shifts had a direct impact on people's lives: "municipalities that experienced partisan turnover in power are better at preserving infants' lives and providing basic services than one party rule municipalities" (19).
De la O uncovers a similar dynamic on CCT programs across countries. She argues that the key factor in determining policy design, specifically "why some governments tie their own hands to fight poverty," is whether presidents have majority support in Congress. If presidents have majority support and few institutional checks, they are free to implement more clientelistic policies. Under many circumstances, divided government in presidential democracies tends to produce policy paralysis, as the interests of presidents and opposition legislators do not align. In the specific case of antipoverty policy, as noted earlier, economic crises and associated costs "motivate politicians to surpass the policy gridlock" (14). Policy design, then, is a result of the competing interests of the president and his or her opposition. A president wants full discretion over the policy he or she proposes, while the opposition does not want to give him or her discretion or credit. Given this, "anticipating the reaction of the median opposition legislator, the president is better off proposing a CCT with a stringent design. With this course of action, the president limits his own discretion over the policy, but he also limits other politicians' ability to use the program to build and strengthen patronage bases" (12). Therefore, checks and balances on executive power, and divided government, produce better antipoverty policy.
I learned a great deal from these books. Garay's book provides an excellent overview of broad trends in social policy, with innovative measures for tapping into both the expansion and the explanations for it. For example, she shows how competition for outsider voters in presidential elections exists when the vote margin between the first and second party among outsider voters is less than 10 percent, or if the challenger wins among outsider voters, and convincingly measures it through impressive district-level and individuallevel data.
Ana De La O's sharp institutionalist theory illustrates how institutional checks and divided government can reduce executive discretion in favor of the public good, and Díaz-Cayeros, Estévez, and Magaloni's conceptualization of public and private goods, analysis of voter behavior, and the electoral logic of political parties is brilliantly compelling. Bonvecchi, Johanssen, and Scartascini's book is a good introduction for those who are less knowledgeable about social network analysis and are looking to learn more. De Venanzi's analysis, while rich in the empirical details of the three case studies, is less rigorous in considering alternative explanations and situating itself in the extant literature. Finally, Martínez Franzoni and Sánchez-Ancochea make an excellent case for disaggregating the concept of universalism from the mechanisms by which to get to it. Our understanding of the politics of social policy since the turn of the millennium is much richer now.
This said, I return to my earlier point that the devil is in the details of policy design. Electoral competition and political and economic pressure from below clearly drove the adoption of more pro-poor policy agendas in the last two decades. Yet political ideology may play a more prominent role in policy design than some of the authors give full credit for. For example, some elements of De La O's measurement of her dependent variable, that is, the extent to which operational rules limit executive discretion in an antipoverty program, appear to conflate nondiscretionary with ideological policy decisions. Conditionality itself is an ideological choice, as Martínez Franzoni and Sánchez-Ancochea would argue. Beyond this, how conditionalities are monitored, and how noncompliant households are treated, is also an ideological policy decision. For example, De La O scores programs regarding whether a minimum threshold on school attendance is required, whether there are established protocols for verifying conditionalities, and whether there is evidence that households that do not comply are penalized (33). However, punitive measures do not necessarily reflect less discretionary policy design but may reflect ideological positions on how to treat the poor (41). Might some of the variation on scores be related to strictness on conditionalities, and influenced by political ideology rather than divided government? For example, De La O's scoring of Uruguay's Plan Equidad as below average is surprising. How the individual scores were calculated is not outlined, but one wonders whether Uruguay experts would agree with this assessment.
Similarly, while Garay allows for a larger role for ideology in the details of policy design, there are instances where scholars may disagree with her regarding how "significant change" is classified. In the case of Chilean health care reform, for example, Garay argues that overall stakeholder opposition and influence in policy negotiations "did not alter the policies significantly" (73). Christina Ewig and Stephen Kay, analyzing the influence of business opposition in health and pensions reforms in Chile, would disagree with this, giving more weight to stakeholder influence in policy design and therefore outputs. 6 Of course, these kinds of disagreements are inevitable with a project of such broad scope.
Another variable that gets less attention in this set of books is the type of governing party. How much does it matter whether a governing party is programmatic, clientelistic, or simply a populist vehicle of the president? Ostensibly, programmatic parties are expected to promote their vision of the public good in office, whether left or right, while clientelistic parties reward their supporters with selective benefits, and populist parties are vehicles for the aggrandizement of the president's power. For De La O, this distinction, which is not addressed, appears to be endogenous to the president's ability to command a majority and pass policy. Thus, her theory can help explain the origins of more institutionalized social policies, specifically in less institutionalized systems. My question is this: If parties are programmatic enough, is legislative opposition needed to produce less discretionary policy? That is, can we count on the president to be constrained by his or her party apparatus and still act in favor of the public good even if he or she does not face a legislative veto?
One hypothesis might be that in countries governed by more programmatic parties, divided government is not necessary to achieve more stringent operational rules. This of course does not make De La O's finding of how to move from more clientelistic to more pro-poor policies any less important, and indeed most political parties in Latin America are far from programmatic, but it may constrain the generalizability of the theory.
Relatedly, Garay theorizes the relationship of governing parties to social movement coalitions (and to an extent interest groups), arguing that bottom-up pressure by these movements produces more inclusive policy expansion. However, such dynamics may also be influenced by type of party organization, as Jennifer Pribble shows in her work. 7 Pribble compares pro-poor social policy expansion between two relatively programmatic party systems: Chile and Uruguay. She explains the differences in social policies between the two leftist governments in Chile and Uruguay as partly a function of different party organization: Chile's elite-led governing parties made a point of insulating themselves from social mobilization, while Uruguay's participatory Frente Amplio explicitly mobilized social organizations and welcomed their input in policy reforms.
Martínez Franzoni and Sánchez-Ancochea's finding that progressive political leadership matters is quite convincing, especially given that universalism tends to be a hallmark of progressive political movements. That said, I have questions about their case selection and measurement. They compare trajectories in four countries that are "unique examples of robust social states in the South" (51) and are "potential candidates to deliver universal outputs in health care and pensions" (15). The case selection appears to merge independent and dependent variables, and I am not sure that this comparison allows us to settle the question of which factors are the key drivers of universal social policy.
More broadly, and inevitably, the findings of these excellent works are influenced by both selection of cases and policy type examined. Future research could examine differences in policy logics across policies. For example, the dynamics of service delivery may elicit different types of mobilization (and opposition) than dynamics on cash transfers. Political ideology and party type may play a role in a government choosing, for example, to extend early education and care services to children by employing formal sector workers rather than providing "stipends" to neighborhood mothers to look after children in their home. 8 In my own work with Merita Jokela, we find that governing party ideology and type strongly influence policies on domestic workers, a particularly vulnerable yet sizeable occupational group. Under the institutionalized leftist governments of Uruguay and Brazil, domestic workers' rights were equalized with those of other workers, and by 2012 formalization reached 42 percent in Brazil and almost 50 percent in Uruguay; while their rights languished under the non-leftist governments of Peru and Mexico; by 2012, formalization had plummeted to 1.2 percent in Mexico and remained at 0.9 percent in Peru, despite other pro-poor policy advances in the countries. 9

Reconceptualization of Social Policy and Its Determinants
Holland, in contrast to the other authors, argues that we have an incomplete picture of what constitutes welfare policy and the ways the poor seek to secure their material needs, especially in the Global South. While the books discussed above have recounted the expansion of formal, institutionalized social policies, Holland examines an area that is crucial for material wellbeing-access to housing and employment through informal channels and government forbearance-yet has not witnessed the same kind of formal policy expansion as have cash transfers and access to health care, nor has it received the same kind of academic attention. This inattention belies its importance: "informal welfare policies are at the heart of how governments in much of the developing world address the poor's most basic needs" (10). For example, in Lima alone, where 13 percent of the population live in houses acquired through land invasions and 10 percent work as street vendors, Holland calculates that forbearance results in transfers of about $750 million a year; national-level anti-poverty cash transfers amount to just one-third of that, reaching only 7 percent of households (9).
Because governments have done so little in delivering access to affordable formal housing and employment, what we see is widespread squatting and informal employment in less developed and highly unequal countries. Governments, especially those with a basic level of state capacity, have a choice on how to deal with them: whether to enforce the law, with dire consequences for the poor, or to "forbear" by being "intentionally lenient" in addressing these violations. Forbearance also includes whether and how to provide post hoc accommodations, such as land titling and infrastructure provision in the case of squatting. What Holland examines, then, is less an expansion of formal policies per se, but continued informal and ad hoc policy by both local and national governments to respond to the informal ways the poor seek to secure their material needs. Holland argues that forbearance is more likely when a state has complementary rather than substitutive social policy. The former is inadequate and/or targeted at other class groups; for example, more than 90 percent of housing subsidies in Peru went to the middle class in 2011 (61). When only complementary social policy exists, the political costs of enforcement are higher: as the director of a city planning agency in Lima states: "We understand that in a poor country, market mechanisms are insufficient to result in an orderly occupation of land, but we can't repress invasions until there are other housing options" (129). These land invasions help meet the unmet needs.
Government forbearance, Holland shows, is recognized by the poor as a pro-poor strategy. Like most of the other authors, Holland challenges "the assumption that the central axis of politics forms between the left, in favor of tax-based redistribution, and the right, which resists it." Because of truncated welfare states, "there is little reason to expect the Latin American poor to support more state spending if they are not its beneficiaries" (5). Rather, she shows through experimental and survey data how class divisions are sharper around legal violations regarding squatting and street vending. "Voters perceive politicians who enforce as taking ' anti-poor' positions" (6).
Forbearance, Holland shows, "occurs through state inaction toward legal violations, which makes its costs less visible, its benefits less secure, and its politics less apparent" (10). It can also empower exploitative intermediaries, "but precisely the fact that these informal transfers do not require a welfare bureaucracy or legislative debate also makes their implementation more credible, especially in contexts where politicians' promises of state welfare solutions often fail to come to fruition" (10).
Substitutive social policy actually addresses the poor's welfare needs. An example of this is Chile's postdemocratic transition housing policy, where governments have, since the early 1990s, provided institutional channels through which the poor could realistically access housing. This made local government opt for enforcement of squatting laws, without political costs.
Chilean housing policy is the exception that proves the rule. In the absence of a broad, effective employment policy in Chile or in the other two countries, there is no across-the-board enforcement on street vendors. Rather, as in Bogotá and Lima, district-level mayors in Santiago respond to their voting base; when a mayor's support base depends on the urban poor, he or she is less likely to enforce strict policies against street vending.
Holland uses the variation in city-level electoral systems across the three capital cities to explain why some mayors enforce and others forbear on street vending specifically. Enforcement variation in Bogotá, which is governed by just one mayor, is influenced by whether the poor are part of any given mayor's winning coalition. Thus, enforcement in Bogotá varies more over time than across space, as in the smaller districts of Lima and Santiago, where mayors respond to their local and highly geographically segregated support bases.
Local governments, of course, have little control over national-level policy trends. Holland brings her argument to the national level by showing how what she calls demand displacement over time results in the "forbearance trap." After violating property laws such as land invasions, the poor can push for post hoc policies such as official recognition of a land occupation or delivery of infrastructure and basic services to occupied land. These efforts to formalize informal property and business have policy feedback effects, reorganizing the demands of the "outsiders" (240) and thus deepening the forbearance trap. Broadly, "complementary social policy moves the target of the poor's claims away from welfare bureaucracies and toward local forbearance, post-hoc policies, and, in some cases, constitutional rights claims" (27).
Holland also shows how Colombia's Constitutional Court plays a role in both improving the well-being of the poor and in fostering the forbearance trap. In 2008 the court ruled that local governments could not deny individuals access to water or electricity based on the legal status of the land where they live, as such denials violate the right to a minimum living standard. As a result, the time that settlements waited to receive basic services dropped from around five years in the 1980s to months in 2010, making land invasions more attractive (134)(135). Holland qualifies such clear policy directions as "quasi-programmatic." This framework sheds light on what from a left-right perspective might be considered surprising electoral support, or lack thereof, among the poor. She shows how forbearance-as an explicit political strategycontributed to the popularity that Peruvian president Alberto Fujimori, certainly not identified as left-wing, had among the poor in the 1990s. Fujimori's government granted more than a million titles to squatters and officially encouraged families to occupy empty lots, with city governments (much more critical of this policy) bearing the costs of providing infrastructure after households built their homes (242)(243). In contrast, a decidedly leftist mayor of Bogotá, Antonius Mockus, managed to alienate the urban poor despite his claims to represent them, through his strict "rule of law" enforcement policies on housing and employment (208). His attempts to provide substitutive social policies in these two realms was not enough to win them over: "The fact that the poor turned against Mockus in his reelection confirms that enforcement had the predicted electoral repercussions" (174). As Holland states, "including forbearance toward the poor in the picture also makes sense of why governments have made progress in reforming some social areas, such as cash transfers and health care, while other 'big-ticket' policies with effective informal welfare substitutes, such as housing and employment policy, remain focused on formal-sector workers" (9).
Holland stresses that she is not making a normative argument in favor of forbearance; rather, she acknowledges that it can reinforce inequalities and give power to exploitative intermediaries. Given this, a key question is under what conditions countries exit the forbearance trap. Holland explains Chile's successful move toward substitutive housing policy by the legacy of Chile's socialist years and the military regime (and the desire by key actors not to "rock the boat"), and the ability of Chile's programmatic political parties to establish more institutional and long-term policies. Holland also uses the case study of Istanbul, Turkey, to show how a megacity exited the forbearance trap, through economic growth in a less unequal society, where by the 2000s, the urban poor no longer formed a necessary element of a winning political coalition. Instead, and in a less positive vein, a change in the ethnic composition of new urban migrants (more ethnic Kurds) has not elicited the same sympathy among Sunni Muslim Turks, and city mayors have not courted their votes through forbearance in the same way.
For many reasons, Holland's book is sure to become an instant classic and required reading in graduate courses across universities. Looking over the sheer quantity, quality, and originality of the data Holland has collected will be worth the reader's time. She conducted a public opinion survey, collected original data on legal violations, enforcement actions, and political dynamics through a structured survey of 149 district bureaucrats in the three cities, as well as interviews with over 100 bureaucrats and politicians. But pick it up for the originality of the conceptualization, the theoretical framework, and empirical findings. In the limited space here, I cannot do justice to the sophistication and nuance of the arguments that she makes.

Conclusion
These seven books represent a major advancement in our understanding of social policy in the region. Given that most of the authors find that political ideology was not central to policy expansion in the areas that they investigated, a natural follow-up question regards the recent shifts to the right in Argentina, Brazil, and Chile, in light of these analyses. Will their social policies remain the same? Of course, once policies are in place, it has been well established that it is much more difficult to touch them, particularly when they are broadly popular. Sara Niedzwiecki and Jennifer Pribble have shown that popular social policies have been maintained under the right-wing governments of Chile (2010-2014) and Argentina (2015-). 10 If the authors who find ideology to be unimportant are right, the rightward shift also should not inhibit further expansion, and even the economic slowdown of the past years should not be determinant. Rather, we should expect the continued political importance of outsiders and electoral competition to drive further policy reforms.
This brings me to a second, broader point. Holland has compellingly broadened our understanding of what constitutes welfare policy, especially in developing countries. Yet one area that none of the books directly examine is the politics of the bulk of social expenditures, such as (often subsidized) contributory pensions, which eat up the budgets of many Latin American countries. One could argue that there are phases of policy expansion, and that for many countries, the "easy," that is, politically less costly and financially manageable phase is over, and given the massive, often regressive bulk of social expenditures that go to contributory systems, further expansion will hit a wall unless states grapple with their extant social security systems. This remains a looming challenge, particularly in countries with more advanced demographic transitions and