Laitn AMerica

Latin American Research Review, Special Issue © 2010 by the Latin American Studies Association. BRAZILIAN DEMOCRACY AS A LATE BLOOMER Reevaluating the Regime in the Cardoso-Lula Era Timothy J. Power University of Oxford Abstract: This article reviews the state of Brazilian democracy at the close of the Cardoso-Lula era. Brazil has now completed a quarter century of competitive poli- tics, the longest democratic period in the country’s history. Although evaluations of the regime’s prospects were often pessimistic in the 1985 –1993 period, the per- formance of democracy improved markedly after the Plano Real stabilization plan in the mid-1990s, which was followed by signifi cant policy achievements under presidents Fernando Henrique Cardoso (Party of Brazilian Social Democracy, or PSDB) and Luiz Inácio Lula da Silva (Workers’ Party, or PT). Since 1995, the axis of national politics has turned on the competition between the PSDB and allies versus the PT and allies. Under this emerging bicoalitional architecture, several key policy domains have been objects of consensus between the two camps, which has led to major policy advances; however, certain policy areas remain outside the zone of consensus and pose enduring challenges. Despite the improving quality of democracy, the mass public continues to display a surprisingly high level of indif- ference to the regime type.

After twenty-one years of military dictatorship, Brazil became a po- litical democracy on March 15, 1985. 1 As the age of democracy passes the quarter century mark, interpretations of the regime are noticeably more positive than they were in the early posttransition years. In the late 1980s I thank Leonardo Avritzer, Philip Oxhorn, João Paulo Peixoto, Nancy Postero, Jeffrey W.

Rubin, Laurence Whitehead, Joel Wolfe, and several anonymous reviewers for comments on an earlier draft of this article.

1. In this article, I understand democracy as political democracy rather than economic, cultural, or social democracy. Political democracy, or polyarchy, as Dahl (1971) termed it, is defi ned by the existence of political procedures that permit robust political contesta- tion and broad participation (i.e., competitive elections with the possibility of alternation in power) and by the simultaneous presence of legal and constitutional guarantees (i.e., civil and political rights) that allow the central democratic institutions to function unimpeded.

Although such a minimalist defi nition of democracy is often dissatisfying in normative terms, it is analytically profi table to divorce the procedural side of democracy from the substantive outputs of the regime, precisely because such a defi nitional separation allows us to assess the interrelationships between the two. Such a causal assessment is an objective of the present study of Brazil. P5338.indb 218 P5338.indb 218 12/20/10 11:44:47 AM 12/20/10 11:44:47 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 219 or early 1990s, the typical analyst might have paraphrased the old joke abo ut Brazil as the perpetual “country of the future” and wondered aloud whether the country would always be an unconsolidated democracy. But in 2010, this is no longer the case. The regime initiated in 1985 is no longer a new democracy in any meaningful sense of the word, either in historical perspective with earlier Brazilian regimes or in comparative perspective with neighboring Latin American countries. Apart from sheer longevity, the regime has also placed some signifi cant policy achievements on the table, whether in macroeconomic performance, social welfare, executive- legislative relations, or global activism. Moreover, with the passage of time, Brazilian democracy has come to be viewed relatively favorably in regional perspective, having managed to avoid some of the more spec- tacular ills that have affl icted several neighboring countries (e.g., fi nancial default, party system collapse, populism, secessionism, and replacement of presidents by dubious constitutional means). The post-1985 regime is defi ned positively not only by what it is but also by what it is not.

Through sheer perseverance, the current Brazilian regime has already passed what might be called the Woody Allen test of democratic consoli- dation: 90 percent of life is simply showing up. It has also passed more conventional tests of democratic consolidation, such as Samuel Hunting- ton’s (1991) two-turnover rule (two successfully handled alternations in partisan control of the government) or Juan Linz’s and Alfred Stepan’s (1996) more subjective litmus test, which is whether relevant actors accept democracy as “the only game in town.” In sharp contrast to the 1980s, when analysts fretted over various potential veto players that might trun- cate or undermine democracy (e.g., the armed forces, right-wing landown- ers, organized business groups, the Globo television network), one of the most outstanding features of Brazilian democracy today is the absence of any major antisystem actor with political clout. Yet despite passing vari- ous tests of endurance and having shown objectively improved perfor- mance since the mid-1990s, the democratic regime continues to confront unresolved questions and daunting challenges. It is only through juxta- posing both achievements and challenges that we can understand Brazil- ian democracy as a lived regime. A comprehensive assessment of Brazilian democracy would require a lengthy monograph, and space considerations preclude an adequate re- counting of political history since 1985. This article therefore limits itself to identifying some key themes and issues that are central to a holistic appraisal of the regime. In doing so, I put forth three arguments and one puzzle. The three arguments are as follows. First, despite ongoing short- comings and challenges, the post-1985 regime represents the high-water mark of democratic experience in Brazilian history. Second, and general- izing rather broadly, the regime has had at least two identifi able phases:

a relatively underperforming period from 1985 to 1993, followed by an P5338.indb 219 P5338.indb 219 12/20/10 11:44:48 AM 12/20/10 11:44:48 AM 220 Latin American Research Review objectively more successful period from 1994 to the present. Third, after twenty-fi ve years of democracy, the tally sheet of regime performance is becoming increasingly clear. By this, I mean that it is reasonably possible to distinguish the policy domains in which palpable successes have been achieved from those in which reforms or “solutions” have been notably lacking. In (briefl y) reviewing policy domains of both types, I reach the rather unsurprising conclusion that the areas of success are the ones in which a broad, cross-party consensus has emerged since the mid-1990s, and the areas of stagnation are the ones in which necessary reforms have failed to acquire stable political sponsorship. These three arguments are accompanied by a puzzle, one that is con- nected to a disjuncture of democracy. Thanks to the pathbreaking work of scholars such as O’Donnell (1993), Holston and Caldeira (1999), and many others, we often think of the main disjuncture in contemporary Brazil as the contradiction between liberal democratic forms and illiberal demo- cratic practice (i.e., competitive politics combined with a woefully inad- equate rule of law). There is little doubt that these scholars are correct, but here I draw attention to a second disjuncture, one that has received much le ss at tent ion i n t he l iterat u re. Si mply put, Bra zi l ia n demo c rac y i s st rongly legitimate at the elite level but weakly legitimate at the mass level. Al- though this sort of contradictory pattern can emerge in the life cycle of al- most any political regime, in Brazil, the data have been remarkably consis- tent. Year in and year out, in good times and in bad, Brazil ranks near the bottom of Latin America in terms of popular preference for democracy as a regime type, and we do not know why. In this article, I develop the three arguments outlined earlier and try to shed some light on the enduring puzzle. I give the greatest amount of attention to identifying the areas of consensus and dissensus in public policy. However, I also attempt to unravel the puzzle of why—despite elite support for democracy, objective political stability, reasonable policy suc- cesses, the absence of important antisystem actors, a string of largely un- controversial landslide elections, and occasionally stratospheric approval ratings for presidential incumbents—ordinary Brazilians seem decidedly unimpressed with political democracy.

BRAZILIAN DEMOCRACY IN HISTORICAL PERSPECTIVE Apart from the current regime, Brazil has had only one other experi- ence with political democracy, lasting from 1946 to 1964. Comparing the 1946 r eg i me to t he 1985 r eg i me i s a u s ef u l exerc i s e b e c au s e it de mon st rate s how the criteria for “democraticness” have shifted over time. When con- trasted with the vigor of the current regime, the democracy of the imme- diate postwar years appears rather limited. P5338.indb 220 P5338.indb 220 12/20/10 11:44:48 AM 12/20/10 11:44:48 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 221 A key difference concerns the constraints of the past: it appears that the preceding authoritarian experience shaped the 1946 regime much more than the 1985 regime. This is true both in politics and in economics. All the elected presidents of the 1946 regime save one (Jânio Quadros, who served for seven months) emerged from the political machine that Getúlio Vargas forged between 1930 and 1945. In the current democracy, although the fi rst two presidents (José Sarney and Fernando Collor) were oligarchic clients of the 1964–1985 military regime, politicians who built their careers in opposition to dictatorship have governed Brazil uninterruptedly since 1992. In the economic sphere, the elites of the 1946 regime followed the basic lines of import substitution and state-led development as laid down in the Vargas years, whereas the current regime has invested far more en- ergy in economic reform and has made a partial but still important course correction in the direction of a market-friendly model. Because democ- racy is a moving target, the early (and mostly negative) assessments of Brazilian democracy published in the late 1980s and early 1990s could not yet perceive these important discontinuities with the past. Moreover, although the 1946 democracy is commonly described as hav- ing lasted for eighteen years, that period was in fact marked by several instances of what the contemporary political science literature would call presidential interruptions (Hochstetler 2006) or breakdowns (Llanos and Marsteintredet 2010), not to mention two hastily implemented modifi ca- tions of the system of government. The suicide of Vargas in 1954, the op- position to Juscelino Kubitschek’s and later João Goulart’s taking of offi ce, and the aborted presidency of Quadros created signifi cant political insta- bility. Nine different men took the presidential oath of offi ce between Jan- uary 1946 and September 1961 (Eurico Gaspar Dutra, Vargas, Café Filho, Carlos Luz, Nereu Ramos, Kubitschek, Quadros, Ranieri Mazzilli, and Goulart). The system of government changed from presidential to parlia- mentary in 1961 and then back again in 1963. It is true that, since 1985, un- elected presidents have ruled Brazil almost a third of the time (both José Sarney and Itamar Franco were originally elected to the vice presidency), but these periods saw no interruption of democratic legality, and the latter instance emerged from a presidential impeachment process conducted in the bounds of the constitution (Pérez-Liñán 2007; Weyland 1993). But the most important advantage of the current regime over its post- war predecessor lies in the robustness of democratic procedures. The two d i me n sio n s of p olya r c hy t h at D a h l (1971) ide nt i fi ed, participation and con- testation, have undergone signifi cant change. Figure 1 is a Dahl-inspired diagram illustrating the two dimensions in all three regimes since the war: the democracy of 1946, the military regime of 1964, and the democ- racy of 1985. The cases in the scatterplot are congressional elections, of which there have been sixteen without interruption since 1945. If we take P5338.indb 221 P5338.indb 221 12/20/10 11:44:48 AM 12/20/10 11:44:48 AM 222 Latin American Research Review 10.08.0 6.0 4.0 2.0 10.0 20.030.0 40.050.0 60.070.0 Participation Competition 1945 1950 1958 1954 1962 1966 1970 1974 1978 19821986 19941998 1990 20022006 Figure 1 Competition and Participation under Three Regimes, 1945 –2006 Notes:

Dots represent elections to the national Chamber of Deputies. Competition is mea- sured as the mean number of candidates per seat in Chamber; participation is the elector- ate as a percentage of the general population. Sources: Laboratório de Estudos Experimen- tais (www.ucam.edu.br/leex) and IPEADATA (www.ipeadata.gov.br).

the average number of candidates per seat in the Chamber of Deputies as a simple measure of competition (e.g., W. G. Santos 2007), and we take the proportion of the general population eligible to vote as an intuitive mea- sure of political participation, we can see a clear U-shaped pattern since 1945. Contestation stabilized in the 1950s before falling sharply under the artifi cial two-party system the military imposed in 1965 and then rising dramatically after 1982. The percentage of Brazilians eligible to vote has risen at every election since 1958, 2 and the expansion of the suffrage has been nothing short of spectacular in the past three decades. Although in 1945 only 16 percent of Brazilians could vote, this rose to 27 percent in 1966, 51 percent in 1986, and an incredible 67 percent in 2006. The main changes under the current regime have been the granting of suffrage to 2. The brief decline after 1954 is more apparent than real: it is explained by major elec- toral re-registration in 1956–1958, which purged deceased voters from the rolls and insti- tuted a new identifi cation system (Nicolau 2002). P5338.indb 222 P5338.indb 222 12/20/10 11:44:48 AM 12/20/10 11:44:48 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 223 illiterates in 1985 and the reduction of the voting age to sixteen in 1988, but these reforms cannot fully explain either the pace or the extent of this remarkable transformation. A good deal of it is owed to socioeconomic modernization as well (Nicolau 2002). What the Dahl-like diagram cannot capture is the meaningfulness of elections, which approached zero between 1966 and 1978 as the military repressed civil liberties and harassed the political opposition. But leaving aside the authoritarian period at the bottom of the U-curve and compar- ing only the two democratic regimes, it is clear that the republic of 1985 is far more polyarchic than the republic of 1946. By the standards of today, the postwar democracy appears rather anemic, and the current regime has given Brazil levels of political participation and contestation never be- fore seen in the country’s history. COMPARING PHASES OF PERFORMANCE , PRE - AND POST -1994 Although democratic procedures have been increasingly robust over time, the performance of the regime has had distinct phases of perfor- mance. At the risk of overgeneralization, it appears in hindsight that 1994 is a clear dividing line between an earlier phase of democracy character- ized by political upheaval and weak economic performance and a second phase with much more encouraging results. The year 1994 is the year of the Plano Real stabilization plan, which ended hyperinfl ation; and from 1995 onward, only two presidents have governed Brazil: Fernando Hen- rique Cardoso and Luiz Inácio Lula da Silva. Thus the essential compari- son is between the Cardoso-Lula era and the more precarious period that preceded it. In terms of economic growth, democracy began with a roaring start, with year-over-year gross domestic product (GDP) growth in the range of 7–8 percent for 1985 and 1986. 3 However, the period also saw mounting infl ation and currency devaluation, and the collapse of the Cruzado Plan in late 1986 inaugurated a long period of stagfl ation. The worst period stretched from 1987 through 1992, when average GDP growth was nega- tive ( –0.14 percent) and hyperinfl ation sailed out of control, averaging 1300 percent annually during this period and spiking to nearly 2500 percent in 1993. However, after the Plano Real ended hyperinfl ation and instituted a fi scal adjustment, Brazil grew at an average annual rate of 3.2 percent between 1994 and 2008, a far cry from the 1930–1980 period when growth averaged 7 percent annually but still quite respectable. Infl ation from 1995 through 2008 averaged only 8 percent annually, a rate roughly equivalent 3. All economic and social data in this section are drawn from the Instituto de Pesquisa Econômica Aplicada (IPEA) Web site (http://www.ipeadata.gov.br), unless otherwise noted. P5338.indb 223 P5338.indb 223 12/20/10 11:44:49 AM 12/20/10 11:44:49 AM 224 Latin American Research Review 0.64 0.62 0.60 0.58 0.56 0.54 0.52 19761978 1980 1982198419861988 19901992 199419961998 20002002 2004 20062008 Figure 2 Gini Coeffi cient of Income Inequality under Dictatorship and Democracy, 1976 –20 08 Source: IPEADATA (www.ipeadata.gov.br).

to three days of infl ation in the fi nal month of the Sarney presidency in 1990. Economic performance since the Plano Real, though clearly uneven, can be characterized as moderate growth with low infl ation.

Social indicators have also improved considerably in the Cardoso-Lula era since 1995. Democratization coincided with rising inequality, with the Gini coeffi cient reaching a dismal 0.635 during the hyperinfl ation of 1989 (fi gure 2). But the Gini has been falling more or less steadily since the mid- 1990s. The end of high infl ation can explain the initial decline (infl ation functioned as a de facto tax on the poor), but the trend was reinforced by cumulative educational reforms and most especially by the conditional cash transfer (CCT) social programs the Lula government pursued ag- gressively since 2003 (Hall 2006; Soares et al. 2006). Although the signa- ture CCT program, Bolsa Família, has garnered signifi cant international at t e nt io n , m o s t a n a ly s t s h ave ove rlo ok e d t h e r ol e o f m i n i mu m wag e p ol i c y since the Plano Real. Both Cardoso and Lula pursued a consistent policy of wage hikes over and above the infl ation rate, such that the real mini- mum wage increased in every year from 1994 through 2009. In fact, by late 2009, the real minimum wage had more than doubled in value since the introduction of the Plano Real (fi gure 3).

By almost all measures, ordinary Brazilians lived much better in the second decade of democracy than they did in the fi rst. This is true not only in social indicators but also in terms of access to good and services that affect daily life. To take one example, the 1995 constitutional amend- P5338.indb 224 P5338.indb 224 12/20/10 11:44:49 AM 12/20/10 11:44:49 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 22\b 500 450 400 350 300 250 200 1501964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 Figure 3 Real Minimum Wage Levels under Dictatorship and Democracy, 1964 –2009 Source: IPEADATA (www.ipeadata.gov.br).

ment that ended the state monopoly on telecommunications caused a mini-revolution. On the eve of privatization, Brazil had 13 million fi xed- line telephones (roughly one phone per twelve inhabitants) and cellular telephony was restricted to a privileged few. By mid-2010, Brazil had 43 million landlines in service and an astonishing 180 million mobile phones (a cellular penetration rate of more than 90 percent), which when added together comes to more than one telephone per inhabitant. 4 The country has also doubled automobile production since 1994, from 1.5 mil- lion to 3 million vehicles annually. 5 The improving economy has put car ownership within the reach of many: there are now 25 million privately owned automobiles in Brazil. In a country of continental proportions, it is also interesting to note the revolution in air travel. In 1985, Brazil had only 11 million air passengers (Instituto Brasileiro de Geografi a e Estatística [IBGE] 1990). By 2003, this had risen to 71.2 million passengers, but then another massive increase of 55 percent in only fi ve years (to 110.5 mil- lion passengers in 2007), caused the air-traffi c system nearly to collapse in 2007. 6 Stable macroeconomic conditions, falling interest rates, and rising personal income at the lower end of the income distribution have also 4. Data from the Agência Nacional de Telecomunicações (http://www.anatel.gov.br).

5. Data from the Associação Nacional dos Fabricantes de Veículos Automotores (http:// www.anfavea.com.br). 6. Data from Empresa Brasileira de Infra-Estrutura Aeroportuário (http://www.infraero .gov.br). P5338.indb 225 P5338.indb 225 12/20/10 11:44:49 AM 12/20/10 11:44:49 AM 226 Latin American Research Review caused a massive boom in the credit market. The size of the credit market expressed in relation to GDP rose from 24 percent in 2003 to more than 45 percent in 2009, and much of the growth has been in so-called crédito popu- lar (Loyola 2009). First-time access to telephones, cars, airplanes, and loans may not fascinate scholars of regime change, but in each case, there was a clear democratization of the sector involved—something immensely im- portant to average citizens. The changing socioeconomic context since 1994 is also refl ected in the political system, with signs of stabilization in electoral volatility and presi- dential competition (the following section deals with executive-legislative relations). Although electoral volatility for the Chamber of Deputies was notoriously high in the early years of democracy, averaging 43 percent in the fi rst three electoral cycles (Paiva and Bohn 2007), the party system showed improved consolidation after 1994. Not only did hyperinfl ation come to an end in that year, thus reducing the space for populist adven- turism and the politics of outbidding a la O’Donnell’s (1994) delegative democracy, but also a constitutional amendment provided for concur- rent elections to the presidency and the legislature. These two contextual changes had strong impacts on the electoral and party systems. Volatility declined sharply, averaging 31 percent in the next three cycles and falling to 27.6 percent in 2006, and presidential politics began to display char- acteristics of a two-party system—all the more surprising given one of the most fragmented party systems in the world (the effective number of electoral parties in 2006 exceeded ten). The last four presidential contests (1994, 1998, 2002, and 2006) were fought essentially between the same two parties: Cardoso’s PSDB and Lula’s Workers’ Party. Note that, in 2006, the PT and PSDB jointly received 90.3 percent of the vote in the fi rst round of the presidential race, a bigger two-party share of the vote than the United States witnessed in the 1990s and a rather strong indication that the coun- try does not object to the many economic and social reforms that Cardoso and/or Lula have instituted since 1994. In summary, the year 1994 stands as a clear turning point in the evolu- tion of the current democratic regime. The political, economic, and social fundamentals of the regime were precarious in the fi rst decade of democ- racy, but all three dimensions have shown marked improvement over the past fi fteen years, a period in which only two presidents have governed Brazil.

AC H I E V E M E N TS AND CONSENSUS As noted previously, the Plano Real and the election of Cardoso re- booted Brazilian democracy in the mid-1990s, thus leading to a new phase in which the PSDB fi rst instituted major reforms and the PT and allies P5338.indb 226 P5338.indb 226 12/20/10 11:44:50 AM 12/20/10 11:44:50 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 227 later expanded on and consolidated. Because of these important policy convergences, it is possible to speak of an implicit cross-party consensus that has emerged around several key issues of democratic governance in the Cardoso-Lula era. Here I stress the word implicit, because the consen- sus that has emerged is by no means a formal political pact but a shared understanding of the basic objectives of policy and the standard operat- ing procedures used for implementation. Furthermore, by using the term consensus I do not mean to imply that the Cardoso and Lula governments have been identical in their outputs. I reject that position as too simplistic:

the different emphases and styles of these two administrations show the consensus to be somewhat elastic, preserving some space for innovation, experimentation, and credit claiming. That being said, the similarities clearly outweigh the differences. The fi ve central points of the emerging consensus in the Cardoso-Lula era are (1) macroeconomic policy, (2) so- cial policy, (3) a revised federal pact, (4) coalitional presidentialism, and (5) renewed international projection. I discuss only the fi rst four of these dimensions here. 7 Macroeconomic Management The fundamentals of macroeconomic policy date from the imple- mentation of the Plano Real in the mid-1990s, although some initiatives can be traced to the failed efforts at structural adjustment by the Collor administration in 1990–1992. As fi nance minister under Itamar Franco, Cardoso rescued some Collor-era initiatives while rejecting the populist Collor’s “neoliberalism by imposition” (Weyland 2000, 41) in favor of a reform program negotiated in concert with state governors and Congress.

This program, which earned support stretching from the center-left to the right in Congress (but not from the PT-led left bloc), was based on a fi scal adjustment combined with the introduction of a new currency, the real. A new tax on fi nancial transactions and dramatically improved revenue col- lection procedures undergirded the fi scal adjustment. The Central Bank, granted effective if not statutory independence and that quickly became known for implementing some of the world’s highest interest rates, was to aggressively defend the real. After Cardoso was elected president in 1994, the reform program was accelerated by conventional initiatives of 7. Although the substantive content of PSDB and PT foreign policies clearly differ, Bra- zil’s renewed and vigorous global activism is a consensual objective of the Cardoso-Lula era, especially in the present decade. However, because of space considerations and be- cause the theme of this special issue of Latin American Research Review is democracy, I do not discuss foreign policy here. For analysis of recent trends, especially presidentialization of the foreign policy-making process, see Cason and Power (2009). P5338.indb 227 P5338.indb 227 12/20/10 11:44:50 AM 12/20/10 11:44:50 AM 228 Latin American Research Review privatization, state reform, and market liberalization, many of which re- quired amendments to the Constitution of 1988. To underscore the partial withdrawal of the state from macroeconomic management, the Cardoso administration created independent regulatory agencies in nine sectors, including telecommunications, energy, and petroleum (Amaral, King- stone, and Krieckhaus 2008). Arguably, the fi rst Cardoso government (1995–1998) had more of an impact on the Brazilian development model than did any other presidency since the days of Vargas. From 1995 through 2002, the PT led the left-wing faction in Congress (holding 25–30 percent of the seats during the period) that vociferously and nearly unanimously opposed all of the major Cardoso reforms. How- ever, in his fourth run for the presidency, Lula released the Carta ao Povo Brasileiro in June 2002, which soberly laid out a rationale for leaving most reforms intact and maintaining the basic lines of macroeconomic policy (Spanakos and Rennó 2006). Acknowledging that the “turbulence of fi - nancial markets” had tied Brazil’s hands, Lula agreed to continue infl a- tion targeting and lent his support to the fi nal International Monetary Fund (IMF) agreement that Cardoso had negotiated in August 2002, which aimed to bolster the real by requiring Brazil to run a primary fi scal surplus of 3.75 percent of GDP (Barros, Baer, and Pio 2003). Upon tak- ing offi ce in January 2003, Lula voluntarily raised the surplus target to 4.25 percent, e s s ent ia l ly advo cat i ng a st ronger do s e of med ic i ne t h a n wh at the IMF had actually prescribed. Lula awarded the Ministry of Finance to the pragmatic Antônio Palocci, a former PT mayor with excellent con- nections in the private sector. More important, he appointed Henrique Meirelles—a former executive at FleetBoston Financial who had just been elected to Congress on the PSDB ticket—to the Central Bank. Meirelles was given free rein to maintain infl ation targets using the same aggres- sive interest-rate policies of the Cardoso era, essentially reaffi rming the Central Bank’s de facto independence. By the end of Lula’s fi rst year in offi ce, fi nancial markets, domestic capitalists, and Wall Street seemed largely convinced of the new PT-led government’s commitment to re- sponsible economic policies. The honeymoon was an extended one: in Lula’s fi rst fi ve years the real gained some 58 percent in value against the U. S. dol l a r. There are clearly differences between the policies of Lula and those of his predecessor: Lula has quietly overseen the expansion of public em- ployment and spending, for example, and has been less sympathetic to independent regulatory agencies. Nonetheless, a consensus has formed around the ideas of continued infl ation targeting, fi scal responsibility with a primary surplus, monetary policy in defense of the real, an improved tax take, and no reversals of the 1990s privatizations. Equally important is that Lula has decisively ratifi ed the Cardoso-era practice of keeping the economic team almost completely insulated from politics. P5338.indb 228 P5338.indb 228 12/20/10 11:44:51 AM 12/20/10 11:44:51 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 229 Social Policy Social policy is another area benefi ting from broad cross-party consen- sus. However, different from the simple sequential pattern of macroeco- nomic policy, wherein the PSDB-led coalition instituted new policies and the PT-led coalition basically preserved them, social policy has evolved from a genuinely interactive process in which both major parties inno- vated, emulated, and expanded signature programs over the past fi fteen years (M. A. Melo 2008). The result has been improving quality of policy, especially in conditional cash-transfer programs. In the mid-1990s, PSDB and PT governments in Campinas and Brasília, respectively, innovated basic income programs and CCTs requiring school attendance (Bolsa Es- cola). After Bolsa Escola showed promise, Cardoso implemented a feder- a l i z e d ver sion. I n re t u r n, whe n i n 20 03 t he Lu la gover n me nt merge d B ol s a Escola with several smaller CCTs to create the renowned Bolsa Família program (Hall 2006), the political opposition was reciprocally supportive.

The political ingenuity of its administration can partly explain the cross- party appeal of Bolsa Família: it is funded federally but distributed mu- nicipally, giving all of Brazil’s 5,500 mayors, whatever their party, a chance to bask in the glow of this hugely popular program. However, cross-party support for Bolsa Família mirrors that for the Fundo de Manutenção e Desenvolvimento da Educação Básica e de Valorização do Magistério (FUNDEB), an educational fund that has few of the electoral payoffs asso- ciated with the signature CCT, which suggests that expansion of the social safety net is indeed a point of consensus that transcends politics.

8 This point was underscored by the presidential debates of 2006, in which the PSDB candidate Geraldo Alckmin attempted to loosen Lula’s ownership of Bolsa Família not by criticizing the program but by claiming that the PSDB would expand it beyond the PT’s ambitious plans (Hunter and Power 2007). Virtually no political actor in Brazil opposes the social policies of the Lula government, which built on and expanded initiatives of the Cardoso period, which themselves were heavily infl uenced by innovations by the PT at the subnational level. The result is a social safety net that provides a guaranteed income to more than 12 million families, covering nearly a quarter of the national population. The two substantive points are basic incomes for the poor and aggressive investment in the educational system to improve human capital. Combined with the equally consensual policy to gradually raise the minimum wage above the rate of infl ation, these policies have reduced poverty and inequality in the Cardoso-Lula era. 8. The FUNDEB succeeded the Fundo de Manutenção e Desenvolvimento da Educação Fundamental e de Valorização do Magistério (FUNDEF) in 2006. Both are multisourced funding schemes for subnational school districts, though the FUNDEB was expanded to include preschool and secondary education. P5338.indb 229 P5338.indb 229 12/20/10 11:44:51 AM 12/20/10 11:44:51 AM 230 Latin American Research Review The New Federal EquilibriumSamuels and Abrucio (2000) observe that subnational political actors largely drove the Brazilian transition to democracy. Democratization oc- curred subnationally before the military withdrew from power nation- ally, leaving pro-decentralization governors and municipalista mayors in a strong position to dominate constitutional redesign. Under their infl u- ence, the National Constituent Assembly of 1987–1988 devolved more than half of all revenues to the states and municipalities while leaving the fed- eral government’s policy responsibilities largely unchanged. This created a fi scal hole for the central government while encouraging subnational executives to embark on a long spending spree in the late 1980s and early 1990s, a period in which state payrolls expanded dramatically. Governors became notorious for using subnational state-owned banks to fund their pet projects. Losses by the banks would require periodic bailouts by the federal government, a practice that governors came to expect (Rodden 2003). To consolidate the fi scal adjustment required by the Plano Real in 1994, the federal government began an aggressive fi scal recentralization pro- gram that continued throughout the Cardoso years. First came the Fundo Social de Emergência in 1994, which allowed the federal government to de-earmark revenues and permit the president’s discretionary spending.

In 1996, the Lei Kandir reduced the right of states to tax exports, generat- ing another de facto recentralization, and in the same year, Cardoso in- stituted the Programa de Incentivo á Redução do Setor Público Estadual na Atividade Bancária (PROES), which incentivized the restructuring, federalization, privatization, and even closure of state banks (Beck, Cri- velli, and Summerhill 2005). In 1999 came the Lei Camata, which limited state spending on personnel to 60 percent of total outlays. This was su- perseded in 2000 by the landmark Lei de Responsabilidade Fiscal (LRF), which maintained limitations on subnational public employment while imposing hard budget constraints on the states. States are now prohibited from incurring public debt in excess of 120 percent of current revenue; they cannot lend to or borrow from other governments at any level; and in a given fi scal year, they cannot borrow in excess of the total amount of capital expenditures (the so-called golden-rule clause). A partial fi scal recentralization and a sharp reduction of the blackmail potential of subnational governments thus characterized the new federal equilibrium. Because these developments greatly improved the fi scal po- sition of the central government, they had the effect of consolidating the Real Plan and improving the investment climate overall. In doing so, they later facilitated the PT-led government’s efforts to establish its economic bona fi des and reduced the dependency of the president on governors. By putting the fi scal house in order and increasing the breathing room for P5338.indb 230 P5338.indb 230 12/20/10 11:44:52 AM 12/20/10 11:44:52 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 231 the president, Cardoso’s LRF ironically made it easier for a left-wing gov- ernment to subsequently come to power in Brazil. For all of these reasons and more, the Lula government has chosen to preserve intact the broad outlines of the new federal pact.

Coalitional Presidentialism The oft-noted similarities between Cardoso and Lula also extend to the realm of political management. Both executives have aimed to secure gov- ernability (understood here in the informal Brazilian sense of the term, that is, legislative support for presidential initiatives) via the rubric of presidencialismo de coalizão. The core insight of coalitional presidentialism is straightforward: because the party of the Brazilian president will rarely if ever have a majority in Congress, the executive must share power with other parties. Presidents must behave like European prime ministers, fashioning multiparty cabinets and voting blocs on the fl oor of the legis- lature (Abranches 1988). The idea is not a new one, as coalitional presiden- tialism was also practiced in the 1946 democracy (Figueiredo 2007), but its early record in the post-1985 regime was uneven. Only after the 1992 impeachment of Collor showed the consequences of failing to build coali- tions and only after presidents had gone through a learning process of how to manage the new governing tools made available by the Constitu- tion of 1988 did coalitional presidentialism come into its own as a coherent modus operandi. Cardoso wrote the user’s manual for the new and improved variant of coalitional presidentialism, and Lula read it carefully. In 1994, Cardoso shocked the political world by partnering his PSDB (previously perceived as a social democratic party) with the conservative Partido da Frente Lib- eral (PFL) and the clientelistic Partido Trabalhista Brasileiro (PTB). This stroke established two principles of successful alliance management: leg- islative coalitions are not only oversized (meaning that they amass more nominal members than what would be necessary to pass legislation) but also disconnected (meaning that the participating parties are not ideologi- cally adjacent to one another). Cardoso rejected the alternative conception of alliance formation (i.e., a compact, minimum-winning, ideologically co- herent governing coalition) in favor of a diverse base aliada that controlled nearly 70 percent of congressional seats. The exaggeration of coalitional size no doubt refl ects the bandwagon behavior of politicians who want to support the executive, but it also suggests a presidential expectation that indiscipline and defections are likely to occur. Oversized and dis- connected coalitions are thus insurance policies: the goal is to protect the president’s legislative program via political overcompensation. Before 2002, Lula had never advocated a coalition that reached outside the family of ideological left-wing parties. In breaking with this tradition P5338.indb 231 P5338.indb 231 12/20/10 11:44:52 AM 12/20/10 11:44:52 AM 232 Latin American Research Review in 2002, and in accepting a running mate from the clientelistic PL, the PT won the presidency for the fi rst time. In power, the PT’s coalitional strat- egy has been strikingly similar to that of the PSDB in the 1990s. The Car- doso and Lula coalitions both had or have one large steering party that is virtually the only source of ideas and policy (the PSDB and PT, respec- tively), and one large support party that is mostly silent on programmatic issues but offers a network of local elites and political machines spread throughout the national territory (the PFL and PMDB, respectively). The combination of these two types of parties gave each president only about 40–45 percent support in Congress, however. The coalitions were there- fore rounded out by incorporating the opportunistic smaller parties of the center-right, such as the PTB, PL, and PP, parties that (true to form) appeared in both the Cardoso and Lula alliances. This addition then cre- ated a support coalition of 65–70 percent of the seats in Congress, allow- ing each president to dominate a marginalized opposition (the PT-led left and the PSDB-PFL center-right, respectively) that could lay claim to only a third of the legislature. As formateurs, Cardoso and Lula clearly worked from the same play- book. The remarkable legislative success of these oversized, disconnected coalitions has provided a great deal of fodder for students of Brazilian politics, who in recent years have endeavored to discover the secrets that make this system work. Generalizing broadly, three answers have come forth. One view emphasizes the ample agenda-setting powers of the pres- ident that the Constitution of 1988 afforded (e.g., Figueiredo and Limongi 1999), a second view stresses the allocation of ministerial portfolios as a way to ensure coalitional discipline (e.g., Amorim Neto 2002), and a third view draws attention to the ample discretionary powers of the president in budgetary matters (e.g., Pereira and Mueller 2004). Although the litera- ture still lacks a unifying approach showing how all of these presidential tools work together to generate governability (Raile, Pereira, and Power 2011), virtually all of the research produced so far agrees that presidents deploy their tools with an essentially coalitional logic in mind (Power 2010). The success of Cardoso and Lula in constructing and cultivating their respective coalitions will probably have a strong demonstration ef- fect on future executives: the general lesson is that ways of coping with a fragmented multiparty system have now been identifi ed, and therefore minority governments and ideologically monolithic coalitions are likely to be discounted in the future. In summary, macroeconomic policy, social policy, a revised federal pact, coalitional presidentialism, and renewed international projection are elements of a broad political consensus in Brazil and have become hall- marks of the more successful, post-1994 phase of democracy. What these policy domains have in common is that, regardless of who originated the core initiatives, they were able to secure political sponsorship from both P5338.indb 232 P5338.indb 232 12/20/10 11:44:52 AM 12/20/10 11:44:52 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 233 major coalitions—captained by the PSDB and PT, respectively—that have governed Brazil since the mid-1990s. The PSDB coalition can claim pa- ternity of economic stabilization and the reform of fi scal federalism, and the PT coalition is closely associated with advances in social policy and international leadership, but the central point here is that each coalition has either imitated and/or lent support to the innovations of the other.

The cross-coalitional consensus was not preordained, as for a good part of this period, the PT opposed the Cardoso economic policies, resisted fi scal recentralization, and shunned the heterogeneous interparty alliances that are now widely considered necessary to govern Brazil. Only in 2002–2003 did it become clear that the PT-led left would become a consistent cospon- sor of these key policies and governing strategies. In doing so, the PT de- fi ned the post-1994 phase of democracy as a single historical experience:

the Cardoso-Lula era. STASIS AND DISSENSUS The emerging points of consensus give Brazilian democracy a solid base on which to build. However, the balance sheet of democracy—even in its rebooted post-1994 incarnation—is not all positive. The regime re- mains challenged by several policy domains on which there has been lit- tle change, or in some cases clear backsliding. I devote less space to these issues because it is precisely the shortcomings, rather than the successes, of Brazilian democracy that have attracted the most attention in the schol- arly literature. Although my discussion here is necessarily incomplete, virtually any list of persistent policy challenges would include the rule of law, judicial reform, corruption, and political reform.

Violence and the Rule of law The interrelated literatures on citizenship, human rights, and the rule of law in Brazil have made a strong case that these have been the most persistent problem areas of the post-1985 regime. A major focus of this work has been on violence and crime. The government’s most reliable time-series database on causes of death (DataSUS) was launched in 1979, and a recent projection suggested that the millionth homicide in this da- tabase was recorded in 2009, as the database completed only thirty years of existence. 9 The murder rate peaked at 28.5 per 100,000 people in 2002, nearing the mark of 50,000 homicides per year, which means that in fi f- teen months Brazil would exceed the total number of deaths incurred by the United States during its fi fteen years in Vietnam. The fl ip side of soc ia l 9. See the interview with Daniel Cerqueira of IPEA, “Estudo aponta 1 milhão de homicí- dios em 30 anos no Brasil,” Estado de São Paulo, January 20, 2008. P5338.indb 233 P5338.indb 233 12/20/10 11:44:53 AM 12/20/10 11:44:53 AM 234 Latin American Research Review violence is police violence, which has worsened under democracy. For ex- ample, the police in São Paulo killed 680 persons per year on average in the 1990–2000 period, a rate more than thirty times higher than that in Los Angeles (Brinks 2006). Brazil has acquired the unenviable reputation of being a world leader in both societal and state violence, but even shocking statistics cannot capture the degree to which the crisis of public security affects daily life. As Holston and Caldeira (1999, 693) note, the “talk of crime” is a pervasive interpersonal discourse.The violence issue is complex, but Pereira (2008, 189) argues that the main causal factors are both “spatial and socioeconomic.” Urbanization, “favelization,” youth unemployment, social inequality, and the illicit drug economy—made profi table by both domestic and U.S. demand as well as Brazil’s largely unpatrolled borders with coca-producing neighbors—all contribute to widespread violence. The strong probability that murderers will successfully avoid either arrest and/or eventual conviction is also a contributing factor (Brinks 2006). Efforts to promote a coherent public se- curity reform are hampered by the extreme fragmentation of the security ap pa r at u s, i n t h at e ac h of Br a z i l’s t we nt y- s e ve n s t at e s h a s t wo p ol ic e f o r c e s (civil and military), and these are under the de facto control of governors and mayors. The central state has only a small federal police that also handles immigration, and only in the current decade has the government developed a small program designed to foster coordination at the subna- tional level (Pereira 2008). The fi scal side of federalism has been reformed repeatedly over the past fi fteen years, but the public-security side of fed- eralism awaits similar intervention. The reluctance of Cardoso and Lula to become directly involved is partly a function of historical deference to governors on police matters but also refl ects the perceived intractability of the security problem.

Judicial Reform The post-1985 democratic regime has also done little to streamline Bra- zil’s notoriously overburdened judiciary. The court system operates at a glacial pace: decisions can take years to be rendered and then spend years more on appeal, such that many actors prefer to avoid the formal judicial system altogether (Taylor 2008). The caseload of certain courts is unsus- tainable. In the current decade, for example, the Supremo Tribunal Fed- eral (STF) has been contending with nearly one hundred thousand new cases per year. Jurisprudence has also been improvisational and incon- sistent: Brazil’s more than eleven thousand judges have hitherto had im- pressive latitude to rule as they wish without regard to precedent. Recent reforms in 2004 and 2007 have aimed to change this situation by allowing the STF to reject cases that are not authentically national in scope, by per- P5338.indb 234 P5338.indb 234 12/20/10 11:44:53 AM 12/20/10 11:44:53 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 23\b mitting the STF to create binding precedent ( súmula vinculante) in certain cases, and by creating an external watchdog agency to monitor the judi- cial branch. Yet the judicial system is still widely criticized for weakening investor confi dence and for aggravating the crisis of public security, and the Brazilian legal profession is notorious for its corporatist resistance to change.

Corruption and Offi cial Impunity Closely related to the inertia of the legal system is the prevalence of cor- ruption (Taylor 2009). In the years since Collor resigned the presidency in the face of corruption charges in 1992, a steady stream of scandals has re- inforced a perception of impunity. To cite only a few of the several dozen prominent scandals in recent years, nineteen legislators were expelled from Congress in the budget scandal of 1994, the Cardoso administration was accused of buying support in 1997 for the amendment allowing presi- dential reelection, three senators were forced from offi ce after violating the secrecy of the Senate’s electronic voting system in 2001, the president of the Chamber of Deputies was forced to resign after extorting the res- taurant franchises in Congress in 2005, aides to Lula were accused in 2005 of operating a scheme of illegal bribes to federal legislators (the so-called mensalão affair) in 2005, no fewer than sixty-nine members of the Chamber of Deputies (some 13 percent of the membership) were accused in 2006 of receivi ng k ickbacks from the sale of over priced ambulances (the so-called sanguessugas, or “bloodsuckers,” affair), and an investigation in mid-2009 showed that the Senate presidency had been issuing secret acts to cover up controversial expenditures and personnel decisions. These Brasília- centered scandals are only the tip of the iceberg: subnational political corruption is pervasive as well. No doubt the perception of widespread corruption is partly owed to the presence of a vigorous news media (Porto forthcoming), but the fact remains that journalists have plenty of raw ma- terial to work with. Failure to control corruption has multiple sources, including the weak- ness of monitoring agencies, the inconsistency of interagency cooperation, and the sluggishness of the court system. The most powerful explanation, however, is the unwillingness of the political class to regulate itself, which in turn is reinforced by the unwillingness of the judicial branch to act against the political class (Taylor and Buranelli 2007). Federal legislators facing corruption charges routinely resign their seats to escape conviction, which allows them to avoid any loss of political rights and to run again in the next election. Senior elected offi cials who are accused of crimes have the right to have their cases heard in a privileged forum ( foro privile- giado ), often the STF itself. The STF is an especially desirable foro because P5338.indb 235 P5338.indb 235 12/20/10 11:44:53 AM 12/20/10 11:44:53 AM 236 Latin American Research Review of its perceived friendliness to the political class: the fi rst post-1988 con- viction of a politician on criminal charges took place only in May 2010. Controlling corruption has not been a success story in the Cardoso- Lula era, and the reasons are partially related to political competition.

Political actors use corruption charges instrumentally: when an incum- bent is accused of wrongdoing, normally the “outs” trumpet and prolong scandals to damage the “ins.” A favorite tool in this regard is the ubiq- uitous congressional inquiry committees (Comissões Parlamentares de Inquérito, or CPIs). A well-rehearsed political drama—one in which all fi ve postauthoritarian presidents have played the starring role—involves the executive branch attempting to mobilize its allies in Congress either to block the installation of a given CPI or, if that strategy fails, to at least select CPI chairs and rapporteurs who are sympathetic to the executive.

Paradoxically, the same bicoalitional architecture that has brought about notable policy consensuses in Brazil since 1994 is also favorable to the politicization of corruption allegations: these accusations offer a non- ideological, policy-neutral tool with which to weaken the other side. The 2006 election campaign in which Alckmin lost badly to Lula was light on substantive issues but heavy on debates about administrative probity.

Meanwhile, Brazil’s reputation continues to suffer. In late 2008, Transpar- ency Inter national (TI) released its Corruption Perceptions Index, which showed that Brazil’s reputation for corruption had worsened for the sixth consecutive year. 10 Ranked forty-fi fth in the world for transparency in 2002, Brazil fell to eightieth place in 2008.

Political Reform Thanks to a large scholarly literature in Brazil and abroad, the question of institutional design is one of the most familiar issue-areas of Brazil- ian democracy. In the English-language literature, Mainwaring (1999) and Ames (2001) drew attention to perceived defi ciencies of the party and elec- toral systems. Brazil’s electoral system of open-list proportional represen- tation (OLPR) combined with a high average district magnitude has been criticized for (1) weakening party discipline by creating an individualis- tic, candidate-centered style of competition; (2) damaging accountability through vote-pooling procedures under which voters may inadvertently contribute to the election of candidates whom they oppose; and (3) fos- tering an oversupply of candidates, thus confusing and alienating vot- ers. The electoral system also lacks a meaningful electoral threshold, thus generating one of the most fragmented national legislatures in the world.

The permissiveness of party affi liation rules has led to a remarkably high 10. Data are from the Corruptions Perceptions Index Web page at Transparency Inter- national (http://www.transparency.org/news_room/in_focus/2008/cpi2008). P5338.indb 236 P5338.indb 236 12/20/10 11:44:53 AM 12/20/10 11:44:53 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 237 rate of politicians’ party switching, with approximately one-third of fed- eral deputies changing parties in each quadrennial legislative session (C. R. Melo 2004). Campaign fi nance is also highly individualistic and poorly regulated (Samuels 2002). This is not an exhaustive list of the issues involved in reforma política, which the Brazilian political class has been debating for two decades. Re- formers have argued for various rule changes with a view to improving accountability, strengthening party organizations, and improving trans- parency (for arguments that many of the proposed rule changes are un- necessary, see Cheibub 2009; Santos and Vilarouca 2008). Similarly to the pattern observed with regard to corruption control, politicians have been notoriously reluctant to regulate themselves. A political reform package was reported out of a congressional committee in 1995 and debated in var- ious versions through June 2007, when the centerpiece proposal—a move to a closed-list proportional representation system—was fi nally voted down. The long delay in discussing reforms led to the courts becoming involved on several occasions. In 2002, the Supreme Electoral Court (Tri- bunal Superior Eleitoral, or TSE), ruled that parties had to form alliances that were consistent at the national and subnational levels (the so-called verticalization decision), but a Congress-initiated constitutional amend- ment four years later reversed this. In 2006, small parties challenged the implementation of an electoral threshold, and the STF ruled it unconsti- tutional. In 2007, the TSE ruled that electoral mandates belong to parties, not individuals, and that party switchers would lose their elective posi- tions. However, Congress responded by passing an amnesty for party switchers and creating a new legal window for musical chairs in the third year of each legislative term. The pattern in all three cases was that the self- interest of politicians prevailed, even over previously implemented reforms and the determination of activist courts. After the congressional vote of 2007 sank the closed-list proposal, an expansive political reform program seemed off the agenda for the foreseeable future. The four policy domains discussed here—the public security crisis, judicial reform, corruption, and political reform—are issues for which Brazilian democracy has provided few if any answers. Different from the doma i n s of success d i sc u ssed ea rl ier, proposed refor m s i n t hese pol ic y a r- eas have failed to acquire consistent political sponsorship. In three of the problem areas (violence, judicial reform, institutional design), the issues involved simply escape the bicoalitional logic that has dominated Brazil- ian politics since 1994. In the fourth area, corruption, the issue unfortu- nately does comport with the bicoalitional logic because accusations of malfeasance can be an effective political weapon. In the early years of eco- nomic stabilization in the mid-1990s, it might have been argued that these issues were being ignored because of the need to prioritize structural ad- justment issues above all else—a reasonable argument in a crisis-recovery P5338.indb 237 P5338.indb 237 12/20/10 11:44:54 AM 12/20/10 11:44:54 AM 238 Latin American Research Review situation. But even as social and economic conditions have improved over the past fi ve years, none of these four policy domains has seen elites stake their careers on reforms. We are left with an untested yet plausible hypothesis about the post- 1994 political system, which is that major reforms are only sustainable when a policy innovation by one of the two major coalitions (PSDB led and PT led) earns the support of the other. Yet some policy areas are ones that impinge too directly on the political survival of elites (corruption, po- litical reform) or that appear too hopelessly intractable to warrant a major investment of political capital (violence and public security). A question mark that hangs over Brazilian democracy is how these four problem ar- eas might somehow be maneuvered into the zone of bicoalitional con- sensus, despite the observed unwillingness of most elites to confront the issues directly. When elite bargaining fails because one of the two major coalitions does not take the lead, it is unlikely that progress will occur without intense popular pressure, external pressure, or some combination of the two.

A PUZZLE : THE WEAK POPULAR LEGITIMACY OF DEMOCRACY So far, I have put forth three arguments about the post-1985 democ- racy: it has given Brazil its most robust polyarchy ever, it has had distinct phases of performance pre- and post-1994, and the regime’s major suc- cesses and shortcomings were evident by the end of Lula’s fi rst term in offi ce. I conclude the article with a puzzle: Brazilian democracy seems legitimate at the elite level but surprisingly unsupported at the mass level.

Why is this the case? T he idea t h at demo c rac y e njoys st rong bac k i ng by el ite s i s u ncont rover- sial. The Brazilian regime is devoid of any major antisystem actors. Behav- ioral compliance with the rules of political democracy is well entrenched:

there have been several successful alternations in power, not to mention the successful removal of a president by constitutional means. The behav- ioral record is also well backed by the available attitudinal data at the elite level: there has been notable convergence both in the policy stances and in the left-right self-placement of politicians (Power and Zucco 2009). More- over, Brazilian political elites clearly believe that their democratic regime is on solid ground. In a survey I conducted with 122 members of Congress in mid-2005, some 88 percent of the sample agreed with the survey state- ment “In 2005, Brazilian democracy is consolidated.” 11 11. The full range of responses was as follows: 50.4 percent strongly agreed, 37.4 per- cent somewhat agreed, 6.9 percent somewhat disagreed, and only 5.2 percent strongly disagreed. P5338.indb 238 P5338.indb 238 12/20/10 11:44:54 AM 12/20/10 11:44:54 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 239 Yet the Brazilian public remains unimpressed with democracy. Histori- cal data on support for democracy are thin, but every year since 1995, Lati- nobarómetro has asked Latin American citizens in up to eighteen countries whether they believe that “democracy is always preferable,” that “under some circumstances an authoritarian regime might be better,” or that “for people like me it doesn’t make a difference” (Corporación Latinobarómetro 1995–2008). Figure 4 presents the Brazilian data through 2008. Although democracy bests authoritarianism in every year, the indifference category is troublingly high. More alarming is the fact that the percentage of Brazil- ians saying that “democracy is always preferable” is among the lowest in Latin America: Brazil has routinely ranked in the bottom three or four coun- tries in the region every year since 1995. As Weyland (2005, 114–115) puts it, “the cultural foundation of Brazilian democracy is not particularly fi rm.” The puzzle here is driven by comparisons to other Latin American polyarchies: given that Brazilian democracy now appears as objectively rather stable and successful compared to that of its neighbors, why are ordinary Brazilians so skeptical? Although more research is necessary on this puzzle, three factors stand out here: enduring cultural patterns, 60.0 50.0 40.0 30.0 20.0 10.0 1995 1996 1997 1998 1999 2000 2001 2002 20032004 20052006 2007 2008 DEM AUTH INDIFF Figure 4 Popular Preferences for Regime Type in Brazil, 1995 –2008 Notes: Possible responses are “democracy is always preferable,” “under some circum- stances an authoritarian regime might be better,” or “for people like me, it doesn’t make a difference” (DK/NA excluded). N ranges from 800 to 1000 every year. There was no survey in 1999 so the values for that year are interpolated. Source: Latinobarómetro (www.latinobarometro.org). P5338.indb 239 P5338.indb 239 12/20/10 11:44:54 AM 12/20/10 11:44:54 AM 240 Latin American Research Review the relative economic success of the 1964–1985 military regime, and con- temporary performance. With regard to culture, Brazilian skepticism about democracy is consistent with a notable pattern of low trust across the board, in which Brazilians evince mistrust in a wide range of social and political institutions (Moisés 1995). This extends to fellow citizens:

the World Values Survey has twice found Brazil to have the lowest rate of interpersonal trust in the world (Power 2002). A long tradition of both qualitative (e.g., DaMatta 1991) and quantitative (e.g., A. Almeida 2008) so- cial science research has linked the Brazilian syndrome of low trust to so- cial hierarchy, elitism, rule-bending practices such as the jeitinho, income inequality, crime and violence, and “mean world” media effects, among many other variables. Omnidirectional mistrust may explain part of the story, but incon- sistent support for democracy is probably also linked to historical and conjunctural factors. Figure 4 suggests that momentary economic condi- tions at least partly drive regime support: democratic legitimacy reached its lowest point in Cardoso’s second term, which was marked by weak growth and the disruptive energy crisis of 2001, but then rose with the improving economic conditions under Lula. If we are to look at economic factors to solve the puzzle, then the comparison with the 1964–1985 au- thoritarian regime is instructive. The military dictatorship coincided with the fi nal phase of Brazil’s long economic boom from roughly 1930 to around 1980, a period in which growth averaged 7 percent a year. A rela- 16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 -2.00 -4.00 -6.00 123456789 12 1110 13 14 15 16 17 1819 20 21 22 23 24 1964 MIL 1985 DEM Figure 5 Real Economic Growth under Dictatorship and Democracy Notes:

For military regime, graph shows 21 years (1964–1984 inclusive) and for current democracy, graph shows 24 years (1985-2008 inclusive). Source: IPEADATA (www .ipeadata.gov.br). P5338.indb 240 P5338.indb 240 12/20/10 11:44:55 AM 12/20/10 11:44:55 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 241 tively small number of living Brazilians can recall the so-called “miracle” years of 1968–1973, a period now forty years in the past, but there is little doubt that the glory days of the 1950s through the 1970s raised Brazilians’ expectations about what the country could achieve. These became incor- porated into the national mythology. By the 1970s, national expectations about economic performance were not dissimilar to those surrounding the soccer World Cup: a second-place fi nish was a national calamity, and the coach summarily fi red.

Comparison to the growth performance of dictatorship was not favor- able to democracy. It is reasonable to assume that the fi rst fi ve to seven years of a new political regime is a critical window for legitimation, thus shaping the regime’s reputation for performance in the eyes of citizens. As fi gure 5 shows, the adult Brazilians of the early 1990s who could recall two political transitions (1964 and 1985) would conclude that one was associ- ated with an economic boom, whereas the other brought mostly recession and infl ation. Dictatorship outperformed democracy not only in a short posttransition window but for a full seventeen to eighteen years of the life cycle of these two regimes. Figure 6 shows that real GDP per capita, a better indicator of living standards, more than doubled in the fi rst fi fteen years of military rule, reaching a peak of about R$12,300 (expressed in reais of the year 2008) in 1980. 12 However, real GDP per capita then re- 12. Approximately US$6,600 with the average exchange rate for the year 2008.

16000 14000 12000 10000 8000 6000 4000 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Military Regime Democratic Regime Figure 6 Real Per Capita GDP under Dictatorship and Democracy, 1964 –2008 Notes:

Calculated in constant reais of the year 2008. Source: IPEADATA (www.ipeadata .gov.br). P5338.indb 241 P5338.indb 241 12/20/10 11:44:55 AM 12/20/10 11:44:55 AM 242 Latin American Research Review mai ned more or less fl at for the following quarter century. Only under the Lula government did Brazil move appreciably beyond the level recorded at the tail end of the 1930–1980 expansion. 13 These observations are tentative: we cannot be sure if democracy’s weak legitimacy is owed to cultural patterns or to comparisons with the recent past, or whether it would rise with a sustained economic boom.

Even if we have identifi ed the right factors, we cannot be sure how much of the puzzle is explained by each. Nonetheless, it is reasonable to sug- gest not only that weak regime legitimation is consistent with the low levels of trust that Brazilians deposit in all sorts of social, institutional, and political objects (trust in parties and in parliament, it should be noted, is always much less than the regional average), but also that high expec- tations about economic performance were long ago hardwired into the national imagination. These expectations are diffi cult to satisfy. If we also make the reasonable assumption that for legitimation purposes, the early years of democracy are the most important in shaping impressions of the regime, then we may be witnessing something of a lagged or in- ertial effect in democratic support. Had the transition to democracy oc- curred in 1995 rather than 1985, giving citizens only the Cardoso-Lula era on which to base their judgments, the Latinobarómetro data might be different. The problem persists, but there may be a silver lining. Although the Lula government has not matched the stratospheric GDP growth rates of the past, the recent upturn in democratic support since 2003 suggests that poverty reduction and declining inequality may provide a new avenue for legitimation. Brazil has trailed China in growth for twenty-fi ve years, but in the past fi ve years, the surging personal incomes among Brazil’s poor have been nicknamed, appropriately enough, crescimento chinês. Much like the dictators who once jailed him, Lula is bringing more Brazilians into the middle class—but the test of democratic legitimation will be keeping them there (Souza and Lamounier 2010). CONCLUSIONS : REEVALUATING DEMOCRACY IN THE CARDOSO -LULA ERA Before the mid-1990s, scholarly analyses of Brazilian democracy em- phasized various dysfunctions of the regime. In the Cardoso-Lula era, evaluations have become markedly more positive. This change is owed to improved policy performance, which in turn is owed to the emergence of a bicoalitional logic centered on a PSDB-led alliance and a PT-led alliance.

13. To express the same idea somewhat differently, between the month that Lula spent in jail in 1980 and his inauguration as president in 2003, real GDP per capita in Brazil in- creased by a paltry 4.5 percent. In his fi rst six years as president, it rose by 18.7 percent. P5338.indb 242 P5338.indb 242 12/20/10 11:44:55 AM 12/20/10 11:44:55 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 243 Subnationally, this logic is reproduced in only a handful of states (most notably São Paulo), but at the national level—the subject of this article—it has provided the axis of presidential competition since 1994. The experi- ence of the Lula government since 2003 has shown that the differences between the two coalitions are much smaller than previously imagined.

In recent years, Brazil has begun to resemble some Western European de- mocracies (e.g., Germany, Spain) wherein center-left and center-right co- alitions alternate in power but apply broadly similar mixtures of market- friendly and social democratic policies. The historical trajectories and the internal cultures of the key parties remain starkly different (thus helping preserve partisan identities), but their policy outputs are not radically dif- ferent in practice. In this sense, the Cardoso-Lula era since 1994 has been something of a long social democratic cycle that has begun to yield visible results for Brazil. The effects are best understood cumulatively. The two Cardoso administrations stabilized the economy, partially reformed the state and federalism, and took the fi rst steps toward improving Brazil’s social safety net. The two Lula administrations have cemented the broad lines of Car- doso economic policy but have greatly expanded the innovations and coverage of social policy. As Lula basked in high approval ratings right before the onset of the global fi nancial crisis in mid-2008, the results of the nearly fi fteen-year cycle seemed to be the best of both worlds: on the one hand, moderate growth and low infl ation, and on the other hand, declin- ing poverty and inequality. The global economic meltdown of the second half of 2008 seemed to temporarily throw this glowing assessment into question, leaving Lula publicly cursing his luck (and the Bush administra- tion as well). However, Brazil was one of the last major economies to enter the crisis and was one of the fi rst to come out of it: economic growth in 2010 was projected to be the strongest since 1986. A year of crisis manage- ment in 2008–2009 barely put a dent in the government’s popularity, and as Lula moved into the eighth and fi nal year of his presidency, the main lines of policy discussed herein remained unchanged. As with economic and social policy, several other policy domains have entered the zone of bicoalitional consensus, and voters have responded positively: in 2006, more than 90 percent of Brazilians cast their fi rst-round presidential votes for either the PT or PSDB. Yet several other policy areas—including public security, corruption, and political reform—have yet to fi nd political spon- sors in the new coalitional architecture. Challenges and doubts remain. External vulnerabilities persist; many domestic reforms remain to be pursued. Outsider politics could shatter the coalitional logic. The political left may not be able to fi nd another Lula that can unite it or that can convince the left to ally once more with the center and center-right, one of Lula’s more impressive political achieve- P5338.indb 243 P5338.indb 243 12/20/10 11:44:56 AM 12/20/10 11:44:56 AM 244 Latin American Research Review ments. Yet the fact remains that the Cardoso-Lula years have given Brazil a reasonably long cycle of pragmatic social democracy, putting the regime in a far better position from which to face future challenges.

REFERENCES Abranches, Sérgio1988 “Presidencialismo de coalizão: O dilema institucional brasileiro.” Dados 31:

5–38.

Almeida, Alberto Carlos 2008 “Core Values, Education, and Democracy: An Empirical Tour of DaMatta’s Bra- zil.” In Democratic Brazil Revisited, edited by Peter R. Kingstone and Timothy J.

Power, 233–256. Pittsburgh: University of Pittsburgh Press.

Amaral, Aline Diniz, Peter R. Kingstone, and Jonathan Krieckhaus 2008 “The Limits of Economic Reform in Brazil.” In Democratic Brazil Revisited, edited by Peter R. Kingstone and Timothy J. Power, 137–160. Pittsburgh: University of Pittsburgh Press.

Ames, Barry 2001 The Deadlock of Democracy in Brazil. Ann Arbor: University of Michigan Press.

Amorim Neto, Octavio 2002 “Presidential Cabinets, Electoral Cycles, and Coalition Discipline in Brazil.” In Legislative Politics in Latin America, edited by Scott Morgenstern and Benito Nacif, 48–78. Cambridge: Cambridge University Press.

Barros, José Roberto Mendonça de, Monica Baer, and Carlos Pio 2003 “Brazil and the IMF: Virtues and Limits of the 2002 Agreement.” Paper prepared for delivery at the Asamblea General del Club de Madrid, Spain, November 1–2.

Beck, Thorsten, Juan Miguel Crivelli, and William R. Summerhill 2005 “State Bank Transformation in Brazil: Choices and Consequences.” Journal of Banking and Finance 29 (8–9): 2223–2257.

Brinks, Daniel M. 2006 “The Rule of (Non)Law: Prosecuting Police Killings in Brazil and Argentina.” In Informal Institutions and Democracy: Lessons from Latin America, edited by Gretchen Helmke and Steven Levitsky, 203–226. Baltimore: Johns Hopkins University Press.

Cason, Jeffrey, and Timothy J. Power 2009 “Presidentialization, Pluralization, and the Rollback of Itamaraty: Explaining Change in Brazilian Foreign Policy Making in the Cardoso-Lula Era.” Inter- national Political Science Review 30 (2): 117–140.

Cheibub, José Antônio 2009 “Political Reform in Brazil: Recent Proposals, Diagnosis, and a Suggestion.” In Brazil under Lula: Economy, Politics, and Society under the Worker-President, edited by Joseph Love and Werner Baer, 9–25. New York: Palgrave-Macmillan.

Corporación Latinobarómetro 1995 –20 08 Encuesta Latinobarómetro (accessed at http:/www.latinobarometro.org).

Dahl, Robert 1971 Polyarchy: Participation and Opposition. New Haven, CT: Yale University Press.

DaMatta, Roberto 1991 Carnival, Rogues, and Heroes: An Interpretation of the Brazilian Dilemma. South Bend, IN: University of Notre Dame Press.

Figueiredo, Argelina 2007 “Government Coalitions in Brazilian Democracy.” Brazilian Political Science Re- view 1 (2): 182–206.

Figueiredo, Argelina, and Fernando Limongi 1999 Executivo e legislativo na nova ordem constitucional. São Paulo: Editora Fundação Getúlio Vargas. P5338.indb 244 P5338.indb 244 12/20/10 11:44:56 AM 12/20/10 11:44:56 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 24\b Hall, Anthony2006 “From Fume Zero to Bolsa Família: Social Policies and Poverty Alleviation under Lula.” Journal of Latin American Studies 38 (3): 689–709.

Hochstetler, Kathryn 2006 “Rethinking Presidentialism: Challenges and Presidential Falls in South Amer- ica.” Comparative Politics 38 (4): 401–418.

Holston, James, and Teresa P. R. Caldeira 1999 “Democracy and Violence in Brazil.” Comparative Studies in Society and History 41 (4): 691–7 29.

Hunter, Wendy, and Timothy J. Power 2007 “Rewarding Lula: Executive Power, Social Policy, and the Brazilian Elections of 2006.” Latin American Politics and Society 49 (1): 1–30.

Hu nt i ng ton, Samuel 1991 The Third Wave: Democratization in the Late Twentieth Century. Norman: University of Oklahoma Press.

Instituto Brasileiro de Geografi a e Estatística 1990 Estatísticas históricas do Brasil: Séries econômicas, demográfi cas e sociais de 1550 a 1988, 2nd. rev. ed. of vol. 3. Rio de Janeiro: Instituto Brasileiro de Geografi a e Estatística.

Linz, Juan J., and Alfred Stepan 1996 Problems of Democratic Transition and Consolidation: Southern Europe, South America, and Post- Communist Europe. Baltimore: Johns Hopkins University Press.

Llanos, Mariana, and Leiv Marsteintredet 2010 Presidential Breakdowns in Latin America: Causes and Outcomes of Executive Instabil- ity in Developing Democracies. London: Palgrave.

Loyola, Gustavo 2009 “Desafi os do crédito em tempos de transição.” Presentation to SERASA-Experian, São Paulo, March 24.

Mainwaring, Scott 1999 Rethinking Party Systems in the Third Wave of Democratization: The Case of Brazil. Stanford, CA: Stanford University Press.

Melo, Carlos Ranulfo 2004 Retirando as cadeiras do lugar: Migração partidária na Câmara dos Deputados (1985– 2002). Belo Horizonte: Editora Universidade Federal de Minas Gerais.

Melo, Marcus André 2008 “Unexpected Successes, Unanticipated Failures: Social Policy from Cardoso to Lula.” In Democratic Brazil Revisited, edited by Peter R. Kingstone and Timothy J.

Power, 161–184. Pittsburgh: University of Pittsburgh Press.

Moisés, José Álvaro 1995 Os brasileiros e a democracia: Bases sócio-políticas da legitimidade brasileira. São Paulo:

Editora Ática.

Nicolau, Jairo 2002 “A participação eleitoral no Brasil.” In A democracia e os três poderes no Brasil, edited by Luiz Werneck Vianna, 255–296. Belo Horizonte: Editora Universidade Federal de Minas Gerais.

O’Donnell, Guillermo 1993 “On the State, Democratization, and Some Conceptual Problems.” World Develop- ment 21 (8): 1355 –1369.

1994 “Delegative Democracy.” Journal of Democracy 5 (1): 55–69.

Paiva, Denise, and Simone R. Bohn 2007 “Sistema partidário e volatilidade eleitoral no Brasil 1982-2006: Um estudo so- bre a dinâmica inter-regional.” Paper presented at the Associação Nacional de Pesquisa e Pós-Graduação em Ciências Sociais (ANPOCS), Caxambu, Minas Gerais, Brazil, October 22–26.

Pereira, Anthony W. 2008 “Public Security, Private Interests, and Police Reform in Brazil.” In Democratic Brazil Revisited, e d ite d by Pe ter R. K i ng stone a nd Ti mot hy J. Power, 185 –208. Pit t s - burgh: University of Pittsburgh Press. P5338.indb 245 P5338.indb 245 12/20/10 11:44:56 AM 12/20/10 11:44:56 AM 246 Latin American Research Review Pereira, Carlos, and Bernardo Mueller2004 “The Cost of Governing: Strategic Behavior of the President and Legislators in Brazil’s Budgetary Process.” Comparative Political Studies 37 (7): 781– 815.

Pérez-Li ñá n, A n í ba l 2007 Presidential Impeachment and the New Political Instability in Latin America. New York: Cambridge University Press.

Por to, Mau ro Forthcoming “The Media and Political Accountability.” In Corruption and Democracy in Brazil: The Struggle for Accountability, edited by Timothy J. Power and Matthew M.

Taylor. South Bend, IN: University of Notre Dame Press.

Power, Timothy J. 2002 “La confi anza interpersonal brasileña en una perspectiva comparativa.” América Latina Hoy 32 (December): 91–115.

2010 “Optimism, Pessimism, and Coalitional Presidentialism: Debating the Institu- tional Design of Brazilian Democracy.” Bulletin of Latin American Research 29 (1):

18 – 33.

Power, Timothy J., and César Zucco Jr. 2009 “Estimating Ideology of Brazilian Legislative Parties, 1990–2005.” Latin American Research Review 44 (1): 218 –246.

Raile, Eric D., Carlos Pereira, and Timothy J. Power 2011 “The Executive Toolbox: Building Legislative Support in a Multiparty Presiden- tial Regime. Political Research Quarterly 64 (2).

Rodden, Jonathan A. 2003 “Federalism and Bailouts in Brazil.” In Fiscal Decentralization and the Challenge of Hard Budget Constraints, edited by Jonathan A. Rodden, Gunnar S. Eskeland, and Jennie Litvack, 213–248. Cambridge: Massachusetts Institute of Technology Press.

Samuels, David 2002 “Pork Barreling Is Not Credit Claiming or Advertising: Campaign Finance and the Sources of the Personal Vote in Brazil.” Journal of Politics 64: 845–863.

Samuels, David, and Fernando Abrucio 2000 “Federalism and Democratic Transitions: The ‘New’ Politics of Governors in Bra- zil.” Publius: The Journal of Federalism 30 (2): 43–61.

Santos, Fabiano, and Márcio Grijó Vilarouca 2008 “Political Institutions and Governability from FHC to Lula.” In Democratic Brazil Revisited, edited by Peter R. Kingstone and Timothy J. Power, 57–80. Pittsburgh:

University of Pittsburgh Press.

Santos, Wanderley Guilherme dos 2007 Governabilidade e democracia natural. Rio de Janeiro: Editora Fundação Getúlio Va r g a s .

Soares, Fábio Veras, Sergei Soares, Marcelo Medeiros, and Rafael Guerreiro Osório 2006 “Programas de transferência de renda no Brasil: Impactos sobre a desigualdade.” IPEA Texto para Discussão 1228 (October).

Souza, Amaury de, and Bolívar Lamounier 2010 A classe média brasileira: Ambições, valores e projetos de sociedade. São Paulo: Editora Campus.

Spanakos, Anthony, and Lucio Rennó 2006 “Elections and Economic Turbulence in Brazil: Candidates, Voters, and Inves- tors.” Latin American Politics and Society 48 (4): 1–26.

Taylor, Matthew M. 2008 Judging Policy: Courts and Policy Reform in Democratic Brazil. Stanford, CA: Stan- ford University Press.

2009 “Corruption, Accountability Reforms, and Democracy in Brazil.” In Corruption and Democracy in Latin America, edited by Charles Blake and Steven Morris, 150– 169. Pittsburgh: University of Pittsburgh Press.

Taylor, Matthew M., and Vinicius Buranelli 2007 “Ending Up in Pizza: Accountability as a Problem of Institutional Arrangement in Brazil.” Latin American Politics and Society 49 (1): 59–87. P5338.indb 246 P5338.indb 246 12/20/10 11:44:57 AM 12/20/10 11:44:57 AM BRAZILIAN DEMOCRACY AS A LATE BLOOMER 247 Weyl a nd, Ku r t1993 “The Rise and Fall of President Collor and Its Impact on Brazilian Democracy.” Journal of Interamerican Studies and World Affairs 35 (1): 1–36.

2000 “The Brazilian State in the New Democracy.” In Democratic Brazil: Actors, Institu- tions, and Processes, edited by Peter R. Kingstone and Timothy J. Power, 36–57.

Pittsburgh: University of Pittsburgh Press.

2005 “The Growing Sustainability of Brazil’s Low-Quality Democracy.” In The Third Wave of Democratization in Latin America: Advances and Setbacks, edited by Fran- ces Hagopian and Scott Mainwaring, 90–120. New York: Cambridge University Press. P5338.indb 247 P5338.indb 247 12/20/10 11:44:57 AM 12/20/10 11:44:57 AM