Content uploaded by Wendy Hunter
Author content
All content in this area was uploaded by Wendy Hunter on Nov 14, 2017
Content may be subject to copyright.
Rewarding Lula: Executive Power, Social Policy, and the Brazilian Elections of 2006
Author(s): Wendy Hunter and Timothy J. Power
Source:
Latin American Politics and Society,
Vol. 49, No. 1 (Spring, 2007), pp. 1-30
Published by: Distributed by Wiley on behalf of the Center for Latin American Studies at
the University of Miami
Stable URL: http://www.jstor.org/stable/4490505
Accessed: 05-07-2017 00:02 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Center for Latin American Studies at the University of Miami, Wiley
are collaborating with
JSTOR to digitize, preserve and extend access to
Latin American Politics and Society
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
Rewarding Lula:
Executive Power, Social Policy,
and the Brazilian Elections of 2006
Wendy Hunter
Timothy J Power
ABSTRACT
This article analyzes Luiz Inaicio da Silva's resounding reelection vic-
tory in the wake of corruption scandals implicating his party and
government. Voters with lower levels of economic security and
schooling played a critical role in returning Lula to the presidency.
Least prone to punish the president for corruption, poorer Brazil-
ians were also the most readily persuaded by the provision of mate-
rial benefits. Minimum wage increases and the income transfer pro-
gram Bolsa Familia expanded the purchasing power of the poor.
Thus, executive power and central state resources allowed Lula to
consolidate a social base that had responded only weakly to his ear-
lier, party-based strategy of grassroots mobilization for progressive
macrosocietal change. Although Lula won handily, the PT's delega-
tion to Congress shrank for the first time, and the voting bases of
president and party diverged. The PT benefited far less than the
president himself from government investment in social policy.
O ne year before the October 2006 presidential elections, Brazilian
opinion polls showed incumbent Luiz Inaicio Lula da Silva of the
PT (Partido dos Trabalhadores, Workers' Party) losing by a significant
margin to either Jose Serra or Geraldo Alckmin, potential challengers
both from the PSDB (Partido da Social Democracia Brasileira). Yet
Brazilian voters ultimately returned Lula to the presidency in a landslide
vote. On October 29, 2006, in a runoff, some 61 percent of all valid
votes went to Lula, compared to the 39 percent cast for his ultimate
opponent, former Sio Paulo governor Alckmin. In this stunning per-
formance, Lula gained more than 12 percentage points over his first-
round showing on October 1.
Besides Alckmin, the contest featured two former PT politicians:
Heloisa Helena, of the left-wing PSOL (Partido Socialismo e Liberdade),
and Cristovam Buarque, who ran on the PDT ticket (Partido Democradtico
Trabalhista). The 2006 election marked the fourth successive presidential
contest involving the PT and PSDB as the two frontrunners. Although the
2006 second-round vote margin of 61-39 is identical to that in the 2002
presidential runoff, the similarity of these aggregate figures masks crucial
1
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
2 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
Table 1. Presidential Election Results, 2006
First Round Runoff
(October 1) (October 29)
% Valid % Valid
Candidate Party Votes Vote Votes Vote
Lula da Silva PT 46,662,365 48.61 58,295,042 60.83
Geraldo Alckmin PSDB 39,968,396 41.64 37,543,178 39.17
Heloisa Helena PSOL 6,575,393 6.85
Cristovam Buarque PDT 2,535,844 2.64
Others 251,762 0.26 - -
Valid votes 95,996,733 100.00 95,838,220 100.00
Eligible voters 125,900,000
Voter turnout (%) 83.25 81.00
Notes: Blank and spoiled ballots are excluded from the table. Invalid votes totaled
an additional 8.41% of total votes cast in the first round and 6.03% in the runoff.
Source: TSE 2006.
differences in the support bases of the two parties between these two
contests. Indeed, government-versus-opposition dynamics exerted a dra-
matic influence on the fate of both parties.
Focusing on the 2006 presidential election, this article analyzes why
Brazilian voters-especially those in the lowest income and education
brackets, and in the North and Northeast regions of the country-sup-
ported Lula in such a resounding fashion. It pays central attention to how
voters of different socioeconomic and regional backgrounds reacted to
the corruption scandals and the key economic and social policy devel-
opments that took place under the first Lula administration, 2003-6. In
this vein, it explains why PSDB candidate Alckmin faced such an uphill
battle to make inroads into President Lula's core support base. By ana-
lyzing the lower house elections, this article also seeks to shed light on.
how Brazilian voters evaluated the PT as a party. In this regard, it tries
to explain why the popularity of Lula and his government did not trans-
late into a higher vote share for the party's congressional delegation.
Several factors make Lula's emphatic victory especially remarkable.
Most noteworthy is that it came in the wake of the series of corruption
scandals that implicated the party, Lula's advisers, and nearly the presi-
dent himself. Allegations of serious and systematic malfeasance first
emerged in June 2005, centering on the monthly payments the govern-
ment made to deputies from allied parties to secure their support in
Congress (the so-called mensaldo). Shortly thereafter, news broke of the
caixa dois (second cash till), whereby PT mayors lined the party's cam-
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 3
Figure 1. Evaluation of Lula Government, January 2003-October 2006
60
40
- POSITIVE
3 0 - ------------- -- ------------ ---- - - -
S 3- o- NEGATIVE
o
= 20 v--- ----- W. ---- ---- 0-i0 -- - el
10
0
Month
Source: CNT/Sensus <www.sensus.com.br>
paign coffers with illegal revenues from bus companies and trash col-
lection agencies contracted in their cities. Although it had gone on since
about 1994, such a practice was especially unexpected from the PT,
which had built a reputation for "clean government" during its two
decades in political opposition.
Amid investigations over these and related charges, several historic
PT figures who held important positions in the government (including
presidential advisers Jose Dirceu and Luis Gushiken), the president of
the party (Jos6 Genoino), and later Finance Minister Ant6nio Palocci all
resigned under a cloud. When Lula's approval ratings sank during the
second half of 2005, many observers felt that his presidential days were
numbered. Yet beginning in December of that year, his popularity
began to rebound. By August 2006, his ratings had returned to pre-
scandal levels, in time to secure reelection solidly in late October. How
did Lula manage to come back and carry the day?
A HISTORIC SHIFT IN VOTER SUPPORT
Beyond the aggregate numbers of votes Lula commanded, also striking
is the shift in the composition of his electoral base since 2002. The pres-
ident enjoyed a groundswell of support among Brazilians in the lowest
categories of income and educational achievement. For example, vote
intentions in the final two days of the runoff campaign had Lula secur-
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
4 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
ing 69 percent of the vote among those with a household income of less
than 2 minimum wages (saldrios minimos), and 59 percent among those
in the 2-5 minimum wage category. His vote share among those
between 5 and 10 minium wages was 49 percent, and among those with
over 10 minimum wages, 44 percent.
Stated differently, Alckmin led Lula by 12 percentage points (56-44)
among those with a family income of over 10 minumum wages
(Datafolha 2006). The relationship is almost perfectly linear. Given that
47 percent of all voters live in households that earn 2 minimum wages
and under (equivalent to 700 reais or 327 dollars), the numerical advan-
tage undoubtedly rests with securing the support of Brazil's poor. The
breakdown by education yielded similar results. In the same poll, Lula
commanded 67 percent of the vote among those with primary school-
ing or less, but Alckmin led by 6 percentage points (53-47) in the group
with some university education or higher.
The scale of Lula's victory in the impoverished North and Northeast
was sweeping. Between 60 and 85 percent of all valid votes cast went
to Lula in the Northeast states. States in the North and Northeast score
toward the bottom end of the Human Development Index (HDI)
devised by the United Nations Development Program.' Accompanying
the heightened backing among the poor was a corresponding loss of
standing among voters with higher levels of income and educational
achievement. Voters with such a profile are concentrated more in the
South and Southeast regions, the major states of which were won by
Alckmin. Not just Slo Paulo, Alckmin's home state, but also Rio Grande
do Sul, home to PT stronghold Porto Alegre and a state Lula had carried
in 1989, 1994, 1998, and 2002, went to the PSDB candidate. In a coun-
try where cross-cutting cleavages very frequently offset voting based on
socioeconomic indicators, the sharp splits displayed in this election are
remarkable indeed. In terms of region and class, this election repre-
sented the clearest division between the "two Brazils" since Lula lost to
Fernando Collor de Mello in 1989. Yet in that memorable election, Lula
stood on the opposite side of the divide.
Lula's ability to sweep voters in the lowest education and income
brackets and (relatedly) consolidate his support in the North and North-
east reverses a longstanding pattern. In his four previous bids for the
presidency between 1989 and 2002, Lula's core support base lay with
voters of higher education levels in the more urban industrial states of
the South and Southeast. This stemmed from the PT's initial formation
as a party of organized interests, intellectuals, and progressive middle-
class urbanites. The lopsidedness was especially stark in 1989, when it
was the populist Collor who secured the support of the poorest voters
in the most impoverished regions of the country (Singer 1990). Although
this deficit was less marked by the 2002 presidential election, even that
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 5
Table 2. Voting Intentions in the Final Two Days of the
Runoff Campaign
Voter Category % of Sample % Lula % Alckmin
Male 48.9 64 36
Female 51.1 58 42
Education
Primary school or less 48.0 67 33
Middle to high school 38.3 59 41
Some university or higher 13.7 47 53
Family income
< 2 min. wages 44.2 69 31
2-5 min. wages 36.3 59 41
5-10 min. wages 11.0 49 51
>10 min. wages 6.2 44 56
Racial group, self-identified
White 45.7 52 48
Brown 33.8 67 33
Black 13.5 74 26
Indigenous 2.7 70 30
Yellow 2.4 68 32
Region
Southeast 44.3 57 43
South 15.4 48 52
Northeast 26.6 76 24
North/Center-West 13.7 61 39
City size
<35K voters 37.5 63 37
35-100K voters 16.9 60 40
>100K voters 45.6 60 40
First-round vote
Lula 51.6 97 3
Alckmin 36.1 11 89
Helena 3.8 55 45
Buarque 1.5 72 28
All voters 100 61 39
N = 11,807
Notes: Table includes only those respondents with declared valid votes, excludes
undecided voters and those intending to cast a blank or spoiled ballot (approxi-
mately 6% of the full sample of 12,561 surveyed). Field dates: October 27-28, 2006.
The predicted result for all voters (61-39) corresponds exactly to the national result
on October 29. Value of minimum wage in October 2006 was 350 reais (approxi-
mately $162 at the time). For states in each region, see appendix.
Source: Datafolha 2006.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
6 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
Figure 2. Swing To/Against Lula Comparing 2002 and 2006 Runoffs,
by State HDI
30
MA
PE
20
To AM
BA
10 2PA
* iES
SE
AC SP
SPR R
-20
RR
-30 Rsq = 0.6133
.5 .6 .7 .8 .9
State Human Development Index
Notes: Y axis is the statewide Lula vote share in the 2002 runoff subtracted from his
vote share in the 2006 runoff. For ease of interpretation, not all states are labeled.
Sources: TSE; IPEA.
recently, Lula faced some difficulty gaining ground among groups
whose interests the party claimed to represent.2
Figure 2 compares the 2002 and 2006 presidential runoffs, in each
of which Lula defeated a PSDB candidate by identical national margins
of 61 to 39 percent. The aggregate vote totals are the same, but they
conceal a massive geographic inversion of support across Brazil's 27
states. Note the strong interelection swing toward Lula in states that
score low on the HDI and the strong swing against the president in
Brazil's most socioeconomically modernized states. What accounts for
this striking shift?
As table 3 shows, the relationship depicted in figure 2 is not spuri-
ous. When we control for the lagged dependent variable (Lula's vote
share in the state in 2002), we find that the HDI continues to be a pow-
erful and negative predictor of the Lula vote in 2006. The same is true
of GDP per capita, a more direct measure of state wealth. We also find
that in 2006, Lula performed better in states with more oligarchical tra-
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 7
Table 3. Predicting the Lula Vote in the Brazilian States
Variable Model 1 Model 2 Model 3 Model 4
Lula in 2002 runoff .495*** .428** .408** .449***
HDI 2000 -.751***
GDP per capita 2003 -.522*** -.317*
Democratic tradition -.563*** -.406**
Adjusted R2 .604 .309 .361 .482
N 27 27 27 27
Dependent variable: Lula vote share in 2006 runoff.
Significance levels: *** p < .01 ** p < .05 * p < .10
Note: Entries are standardized regression coefficients.
Sources: IPEA 2006; TSE 2006.
ditions and less competitive politics, even when we control for wealth
and his previous performance in the state.3 How, then, did Lula come
to acquire a geographic profile that somewhat resembles the voting base
of his 1989 opponent, Collor, not to mention the geographic bases of
support for the conservative military regime of 1964-85?
Contrasting Results in the Lower House
While Lula secured a crushing victory in the presidential contest, more-
over, the performance of PT candidates in races to fill all 513 seats in
the Camara dos Deputados was less than stellar. The most valid indica-
tor of electoral support for a given political party in Brazil is the aggre-
gate national vote total for the party's candidates for the Chamber of
Deputies (see appendix for the PT's state-level performance). In terms
of the distribution of seats, the overall legislative results in 2006 differed
only marginally from those of 2002. The PT won enough votes to secure
83 chamber seats, placing it second to the PMDB's 89 representatives.
(In the 2002 election, the PT won 91 seats, making it the single largest
party in the chamber). In the senate, the PT lost 4 seats.
This was the first time in its history that the PT had failed to grow
relative to its previous performance in national legislative elections (see
tables 4 and 5). Given that the PT had held the presidency in the inter-
vening four years and therefore had enjoyed the advantages of incum-
bency, this showing was mediocre in historical perspective. By contrast,
previous governing parties, such as the PMDB and the PSDB, increased
their congressional delegations substantially under the respective presi-
dencies of Jose Sarney (1985-90) and Fernando Henrique Cardoso
(1995-2002).
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
8 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
Table 4. Distribution of Chamber of Deputies Seats
Before and After the 2006 Elections
After After Gain/Loss Gain/Loss
2002 By 2006 over 2002 over
Party Elections mid-2006" Electionsb Elections mid-2006
PT 91 81 83 -8 +2
PFL 84 65 65 -19 0
PMDB 75 83 89 +14 +6
PSDB 70 57 66 -4 +9
PP 49 47 41 -8 -6
PTB 26 43 22 -4 -21
PL 26 36 23 -3 -13
PSB 22 28 27 +5 -1
PDT 21 20 24 +3 +4
PPS 15 16 21 +6 +5
PC do B 12 12 13 +1 +1
PV 5 7 13 +8 +6
PSOLc 7 3 - -4
Others 17 11 23 +6 +12
Totals 513 513 513
'Reflects deaths, resignations, and party switches during the quadrennial legislature
that began in February 2003.
'"Reflects expected distribution of seats immediately after the first round of voting on
October 1; this will change due to party switching.
cThe PSOL did not exist in 2002; it was formed in 2003-4 by dissidents from the rad-
ical wing of the PT.
Source: Camara dos Deputados.
The trend in the party's demographic support base in the Chamber
of Deputies compared to that of Lula is increasingly incongruent. Where
Lula has made impressive inroads into the most backward regions of the
country (os grot&es, the PT's longstanding Achilles' heel), the party's
stronghold remains the more urban, industrialized areas of Brazil. Some
nationalization of the party's congressional delegation since the early
1990s notwithstanding, Lula has managed to secure much broader pen-
etration than his party has (Samuels 2006). This is also true of most pres-
idential candidates in Brazil-including Cardoso-but what is interest-
ing is how the geographic dimension of the Lula-PT relationship has
changed over time.
Table 6 illustrates this point: it reports the correlation between the
PT partisan vote and the Lula presidential vote across Brazil's states in
four successive elections. With respect to the October 2006 election,
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 9
Table 5. Distribution of Senate Seats Before and After the
2006 Elections
After After Gain/Loss Gain/Loss
2002 By 2006 over 2002 over
Party Elections mid-2006a Electionsb Elections mid-2006
PMDB 19 21 18 -1 -3
PFL 19 16 18 -1 +2
PSDB 11 16 13 +2 -3
PT 14 11 10 -4 -1
PDT 5 4 5 0 +1
PTB 3 5 4 +1 -1
PSB 3 2 3 0 +1
PP 1 0 1 0 +1
PPS 1 0 1 0 +1
Others 5 6 8 +3 +2
Totals 81 81 81 - -
3Reflects deaths, resignations, and party switches during the quadrennial legislature
that began in February 2003.
bReflects expected distribution of seats immediately after the second round of voting
on October 29, when some senators were elected state governors and their seats
were slated to be taken by alternate members.
Source: Senado Federal.
figure 3 shows that Lula's "overperformance" relative to PT legislative
candidates (that is, the Lula vote share minus the PT chamber vote share
in a given state) covaries negatively with state HDI. What explains the
comparative popularity and greater regional advance of Lula and his
government? What implications does this divergence have for the PT's
future as a party?
Lula's new status as an incumbent rather than a challenger is central
to the massive geographic shift in the Lula vote between 2002 (and
before that) and 2006. Brazilian voters, furthermore, are judging him not
on what he has said-encompassing both criticism of past governments
and promises for the future-but on what he has actually done with
executive power. This "interiorization" or "northeasternization" of sup-
port is typical of governing parties in Brazil. Victor Nunes Leal told us
just that more than 50 years ago in his classic work Coronelismo, enxada
e voto (1949). In some important respects, this is yet another unfolding
of the old story of using the government to build clientelistic support.
It is interesting that the process required gaining control of central
state resources to reinvent the social base that the PT had attempted to
build with the promise of structural reforms, such as land distribution,
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
10 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
Table 6. State-Level Correlation Between PT Partisan Vote and
Lula Presidential Vote in Concurrent Elections
Election Year Correlation
1994 .602***
1998 .527***
2002 .462*
2006 -.193
N = 27 for all years.
Significance levels: *** p < .01 ** p < .05 * p < .10
Notes: PT partisan vote is defined as the party's share of the total statewide vote for
the Chamber of Deputies. To make the years comparable, only the first round of
presidential elections was used.
and the strategy of collective mobilization. A twist in this old story, how-
ever, is that the governista effect benefited Lula the presidential candi-
date but not his party in the election. Whereas in 1994, 1998, and 2002
the Lula vote was strongly and positively associated with the PT vote for
the Chamber of Deputies, in 2006 the correlation was slightly negative.
The PT's partisan vote still varies positively with HDI, but Lula's per-
sonal vote varies negatively with HDI. Thus, while shedding significant
light on the results of the presidential election, the traditional gov-
ernismo-interiorization story does not conform completely to previous
(read: more partisan) patterns of governista vote swings.
UNDERSTANDING LULA'S VICTORY:
THREE FRAMEWORKS
A number of serious corruption scandals bedeviled the Lula government
in the last two years of his first term. The final one broke a mere two
weeks before the first round of the 2006 election. Serious enough to
induce the temporary resignation of party president Ricardo Berzoini,
the charge was that advisers close to Lula had tried to buy an unflatter-
ing dossier on PSDB gubernatorial candidate Jose Serra.4 The media
jumped on the news, showing piles of money alleged to have been the
payoff. Lula's approval ratings took a slight dip thereafter, placing the
president slightly short of a first-round victory with 48.6 percent of the
valid vote. But the overall and more significant trend was that the pres-
ident's approval ratings had been on the rise since December 2005, and
they recovered quickly from this stumble at the eleventh hour. Gaining
12 percentage points after the first round, Lula went on to clinch a stun-
ning runoff victory on October 29.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 11
Figure 3. Lula Overperformance Compared to
PT Legislative Candidates, by State HDI
80 MA
U AM
CE
70 "P
60 AtP ES M
a P
3
501 SE MG
PA
20
401 RO MT
20
10 Rsq 0.4301
.5 .6 .7 .8 .9
State Human Development Index
Notes: Y axis is the statewide Lula vote share in the 2006 runoff minus the aggregate
vote share of PT candidates for the Chamber of Deputies. For ease of interpretation,
not all states are labeled.
Sources: TSE; IPEA.
The Differential Impact of the Corruption Scandals
on Brazilian Voters
Lula responded to the various corruption scandals of his first govern-
ment with a contradictory bundle of messages. On the one hand, he
maintained that he had not known of the manipulations carried out by
some of his closest advisers. On the other, he implicitly acknowledged
these misdeeds by maintaining-in a now notorious interview he gave
to the TV Globo program Fantastico while in Paris in July 2005-that
the PT did not do anything different from what other parties had done
before it and that Brazil's perverse political institutions were ultimately
to blame. Underlying Lula's unsuccessful push for Congress to take up
the issue of political reform in July and August 2006 were a tacit admis-
sion of wrongdoing and a clumsy attempt to shift blame onto the "rules
of the game"-which, the president's defensiveness aside, may indeed
contribute to a permissive environment for corruption in Brazil.5
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
12 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
When he was not invoking o sistema, Lula sought to extricate him-
self by questioning the integrity of other individuals and even of his
own party. The distance he maintained from the party throughout the
campaign-going so far as to reduce the PT symbol of the red star to a
tiny dot on his advertisements and barely even to mention the party in
his free TV time (bordrio eleitoral gratuito)-spoke volumes about his
determination to be reelected.
Why did the corruption charges ultimately not sink Lula's chances
of reelection? Did voters not know much about the scandals? Were they
aware, but did they not care? Did they know about the scandals but
attach greater salience to other issues? The answer to these and related
questions depends on drawing critical distinctions among various seg-
ments of the Brazilian electorate. Voters' socioeconomic standing greatly
influenced not only how much they knew but also how much they
cared about the corruption issue. Low levels of education and political
awareness, together with the priority poor people understandably place
on meeting basic material needs, no doubt helped Lula in less devel-
oped regions of Brazil. Having to face an entire electorate with demo-
graphics similar to those found in the wealthier states of the South and
Southeast would have put Lula's reelection at risk.
Information asymmetries abound in a country of such disparities.
Information levels mediate how voters use corruption as a criterion for
evaluating candidates. There is emerging empirical support for the idea
that voters with greater access to information are more punitive. Recent
work on corruption audits in Brazil's municipalities suggests that a con-
siderable share of Brazilian voters care about graft, and that when they
are empowered with information they are willing to punish corrupt
politicians at the polls. For example, the presence of local radio stations
to disseminate the results of these audits elevates significantly the prob-
ability of an accused mayor going down to defeat (Ferraz and Finan
2006); conversely, without local radio, corrupt mayors fare quite well.
Another study has shown that the likelihood of reelection among cor-
ruption-tainted mayors in the state of Pernambuco is powerfully medi-
ated by the effect of municipal-level HDI, which in part measures liter-
acy (Figueiredo et al. 2006).
The strikingly low circulation level of news-intensive print media
also contributes to informational assymmetries. For example, the daily
with the single largest circulation, the Folha de Sdo Paulo, sells approx-
imately 290,000 copies on weekdays and 360,000 on weekends. The
best-selling weekly magazine Veja (whose coverage of the PT and Lula,
it may be noted in passing, has been relentlessly hostile since 2003) has
a wider circulation, at 1.2 million copies. In a country with a population
of roughly 190 million, this is an exceedingly small number; and need-
less to say, the readership of these news sources is confined largely to
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 13
citizens from Brazil's A and B classes (socioeconomic classifications as
used by polling firms).
Beyond being less informed, it is also true that Brazilians of limited
schooling show greater tolerance of patrimonialism than their more edu-
cated counterparts. The recent Brazilian Social Survey (Pesquisa Social
Brasileira, PESB) bears out this point well. It asked respondents to
agree or disagree with the following statement: "If someone is elected
to public office, he/she should use the office as if it were his/her per-
sonal property, for his/her own benefit." Whereas 40 percent of all illit-
erates and 31 percent of those with a fourth-grade education or less
agreed with the statement, only 5 percent of those with a high school
education and 3 percent of those with a university education did so
(Almeida 2006b).
Mirroring these general patterns, demographic characteristics
appear to be central in citizens' evaluation of the PT's corruption. In the
period between June and September 2005, when news of the PT's men-
saldo and caixa dois activities dominated Brazil's media, polls measur-
ing public confidence in Lula registered a 20-percentage-point decline
(from 56 to 36 percent) among respondents with secondary education
or higher. By contrast, the decline among those with less than a high
school education was only 9 percentage points (from 62 percent to 53
percent) (IBOPE 2006). A related finding, in October 2005, was that
when asked if Lula had "much responsibility" for cases of corruption in
his government, 46 percent of respondents with some college education
said yes, whereas only 31 percent of those with a basic education
agreed (Folba de Sao Paulo 2005).
Another polling series tracked the government's approval/disap-
proval ratings by three income groups: poor (those earning less than 5
minimum wages), middle class (between 5 and 10 minimum wages),
and affluent (more than 10 minimum wages). Net approval of Lula by
the three groups tracked quite similarly for the first two years, but the
three began to diverge significantly after the mensaldo scandal broke in
June 2005. This suggests a strong causal effect of corruption allegations
on evaluations of the president. Among the poor, although net approval
of Lula slipped in the second half of the year, it always remained posi-
tive and rebounded strongly after January 2006. The score for middle-
class respondents slipped into disapproval from July to December,
falling to just below zero before slowly rising afterJanuary. Net approval
among more affluent Brazilians fell precipitously, reaching a low point
of minus 30 in December 2005. Support for Lula in this group never
returned to prescandal levels (Desposato 2006, 32).
Subsequent public opinion research carried out in March 2006 by a
PT party organ, the Fundag.o Perseu Abramo, assessed how informed
respondents were about the corruption charges against the PT and the
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
14 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
Lula government. Analyzed by region, 42 percent of respondents from
the Northeast were evaluated as "uninformed" (desinformado) com-
pared to only 25 percent from the Southeast (and 24 percent in the
South). Similarly, when asked if Lula knew about the illegal monthly
bribes that the PT gave to deputies from allied parties (the mensaldo),
42 percent of respondents from the Northeast thought he did not know,
as opposed to only 19 percent from the South and 23 from the South-
east (Funda~go Perseu Abramo 2006). In sum, greater knowledge and
repudiation of corruption in the more developed regions probably con-
tributed to Lula's poorer showing in the South and Southeast. Whether
due to lower levels of information and political awareness or greater
levels of tolerance for corrupt behavior, the evidence clearly suggests
that less educated voters were more likely to forgive the president and
his government.6
Moreover, interpretations that focus on the economic and social
policy bases of Lula's reelection suggest that even if Lula lost some
points with the public over corruption, he compensated for that loss in
other realms. From the vantage point of less economically secure voters,
the PT and the Lula government may have been corrupt, but they also
delivered valuable material benefits. This notion is encapsulated in the
popular yet pejorative Brazilian saying rouba masfaz (he steals but gets
things done). It was first applied to legendary Governor Adhemar de
Barros of Sio Paulo in the 1950s and thereafter to many patronage-
wielding politicians from the political right (not to politicians associated
with a previously programmatic leftist party like the PT). Although some
conservative anti-Lula media outlets in Brazil have occasionally used this
notion to scapegoat poorer voters for their vote choices, the concept
also finds some theoretical (and more neutral) reflection in social sci-
ence. Theorists of modernization and cultural change would expect less
economically secure individuals to place basic needs above issues like
ethics and transparency in politics and their more privileged counter-
parts to emphasize such "postmaterial" values comparatively more
(Inglehart and Welzel 2005).
Economic Factors
Most presidential elections in which an incumbent seeks reelection are
plebiscitarian affairs: does the president deserve another four years in
office? To the extent that voters answer this question with a view to eco-
nomic factors, the context in 2006 was very favorable to the reelection
of Lula, especially among the poor majority.
Overall GDP growth in the first Lula term represented an improve-
ment over the second Cardoso government. The international context-
characterized by rising commodity prices and an absence of the global
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 15
financial crises that had plagued Cardoso-was far more favorable to
Brazil after Lula's election in 2002. Real GDP growth in Lula's first year
in office (2003) was only 0.6 percent, but a strong expansion followed
in 2004 (4.9 percent) and a moderate expansion in 2005 (2.3 percent).
Foreign observers have been quick to note that Brazil's growth
rates, while respectable, have lagged behind the equivalent global and
regional rates. For example, in the four years between 2002 and 2005,
Brazilian GDP growth was lower than the Latin American average every
year. Voters, however, are unlikely to engage in such global compar-
isons; they are much more likely to evaluate performance intranation-
ally, across time. In this sense, Lula benefited from the weak perform-
ance of the second Cardoso administration and from his own poor start
in 2003. In the 2001-3 period (Cardoso's final two years and Lula's first),
the average annual growth rate was under 1.3 percent. Using current
estimates for 2006, the average annual growth rate in real GDP for the
final three years of Lula's government would be approximately 3.4 per-
cent, well more than double the rate of the previous three-year period.
Thus, in the Brazilian electoral context, the expansion of 2004-6 is more
impressive than it seems cross-nationally.
Lula's decision in 2003 to maintain the aggressive tight money poli-
cies of the Cardoso government and to run a large fiscal surplus (Hunter
and Power 2005; Bianchi and Braga 2005) contributed to unprecedented
macroeconomic stability in his first term. Inflation, which had been
brought under control in the Cardoso period, declined further under
Lula, dropping below 6 percent in 2005 and falling to 4 percent in 2006.
In inflation-adjusted terms, the Brazilian currency, the real, has appreci-
ated more than 25 percent against the U.S. dollar since 2004, benefiting
consumer spending by making imports cheaper. The cost of these poli-
cies has been exorbitant interest rates-routinely among the highest in
the world-but even these have been falling steadily, from an inflation-
adjusted annualized rate of 19.1 percent in 2005 to an expected rate of
15.3 percent in 2006 (Banco Central 2006a).
The gradual lowering of interest rates, combined with an improving
macroeconomic context, has generated a massive expansion of con-
sumer credit in Brazil. Consumer borrowing as a share of GDP doubled
between 2001 and 2006, with an explosion in credit cards and payroll-
guaranteed loans and a significant expansion of over-the-counter credit
for purchases such as electronics and household appliances. The stock
of outstanding credit increased from R$137 billion to R$177 billion in the
year before the election, and banks have been gradually and cautiously
extending the length of individual loans (Economist Intelligence Unit
2006, 34). The consumer credit market has been accessible to low-
income and first-time borrowers for the first time in recent Brazilian his-
tory, thus boosting consumer confidence.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
16 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
Voter confidence in the economy has also been positively affected
by Lula's generous minimum wage policies. Most Brazilian salaries are
in some way linked to the minimum wage as a unit of reference, so
increases in the minimum wage have a ripple effect in consumer spend-
ing and throughout the economy. From a value of R$200 in early 2003,
Lula authorized a series of increases until the minimum wage reached
R$350 (approximately $162 per month) in April 2006, six months before
the election. The real increase in the purchasing power of the minimum
wage was approximately 23 percent in Lula's first term. This, combined
with cash injections from the poverty-reducing income transfer program,
the Bolsa Familia, has had a palpable effect on the local economies of
smaller and less developed municipalities, which depend heavily on
small-scale personal spending for their livelihood. Thus it is not sur-
prising that retail sales over the past three years have climbed most
dramatically in the North and Northeast of Brazil (Banco Central 2006b).
It is also not surprising that these are the two regions where both voter
turnout and support for Lula rose in 2006 compared to 2002.
Although this article argues throughout that the poor engaged in
"pocketbook voting" in 2006 due to beneficial economic and social poli-
cies, this does not mean to suggest that only the poor exhibited such
behavior.7 The economic recovery under Lula has coincided with rising
personal incomes for the poor but stagnating incomes for the privileged
classes. Newly released data from IPEA show that between 2001 and
2005, the aggregate income of the poorest 10 percent of Brazilians grew
by 35.9 percent, or about 7.9 percent per year. (In Brazil, this rate of
growth for the poor has been nicknamed crescimento chines, since it is
approximately equivalent to the average 8.2 percent annual rise in per
capita income in China between 1990 and 2003.) The total increase in
income between 2001 and 2005 for the poorest 50 percent of Brazilians
was 16 percent. In contrast, the upper 20 percent of the income distri-
bution saw their aggregate income decline (-0.5 percent) over these
same four years, and for the top 10 percent the decrease was sharper
(-1.3 percent) (Estado de Sao Paulo 2006).
Thus, in the 2006 elections, not only did the poor have excellent
reasons to vote for Lula, but the professional middle classes had pock-
etbook reasons to vote against him. Overall, however, the vastly differ-
ent relative sizes of these social classes must be taken into account (see
table 2). If all Brazilians engage equally in economic voting, it is the
behavior of the poor that will have the greatest macropolitical effect.
Therefore the analytical emphasis falls on social policy.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 17
Targeted Social Policy
A further and highly compelling line of explanation for Lula's electoral
success, especially among those at the bottom of the income ladder,
concerns developments in the area of social policy. Partly because the
government kept spending in check during the first two years, it could
spend more amply in the final stretch, which no doubt helped Lula's bid
for reelection. As president, Lula has targeted spending increases at the
poorest families. Social policy developments have reinforced the eco-
nomic trends noted above in effectively expanding the purchasing
power of many Brazilians. They have even contributed to some social
mobility; namely, an increase in the number of families moving out of
socioeconomic classifications D and E into class C.
Absolute poverty clearly declined under Lula's presidency, by some
estimates by as much as roughly 15 percent in the first three years (Neri
2006). The share of Brazilians living under the "line of misery" (roughly
speaking, a family income equivalent to one-third of the minimum
wage) fell from 28 to 23 percent between 2003 and 2005. In certain
states, misery diminished dramatically (e.g., 15 points in Amazonas, 10
points in Bahia). Poverty reduction on this order is a significant achieve-
ment in both Brazilian historical perspective and Latin American
regional perspective. At the state level, Lula's second-round vote share
correlates at -.635 with HDI and at .642 with the state-level decline in
the misery rate between 2003 and 2005 (IPEA 2006. All correlations sig-
nificant at .01). It is not surprising that Lula did well in these places.
It is notable that the major developments in social policy under the
PT-led government are a far cry from the kinds of reforms the PT advo-
cated when it was an opposition party. They are not structural reforms
based on the premise of a significant redistribution of property or
income away from the rich to the poor. Indeed, one of the main reforms
advocated by the party in the past, agrarian reform, has faded from
center stage. The World Bank, the Inter-American Development Bank,
and even the International Monetary Fund have praised the social safety
net emphasis of social policy under Lula, especially the Bolsa Familia.
Still, it is not clear that the major social programs of the PT-led gov-
ernment are based on a broader vision of how to modernize Brazil.
While Cardoso's was a government of reforms, Lula's has been a gov-
ernment of programs, a qualitative difference. The reform path pursued
by the Cardoso government required entering into lengthy negotiations
and achieving high-threshold votes in Congress in order to implement
unpopular policies, such as privatizing state firms, breaking up monop-
olies, and amending labor laws. The Lula government's social policy
accomplishments are programs controlled by the executive, which do
not require legislative wrangling to be enacted. Targeted social policies,
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
18 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
moreover, produce more immediate electoral returns than the structural
reforms of the Cardoso years, and their very nature-being so concrete,
visible, and immediately beneficial-invites far wider credit claiming by
politicians. Lula has proven most adept in this regard. By sharing the
day-to-day operation of Bolsa Familia with Brazil's 5,500 municipal
mayors, the president has allowed local elites to reap some of the ben-
efits of this hugely popular program.
The social centerpiece of the first Lula administration, the Bolsa
Familia (Family Grant) is a conditional cash transfer (CCT) program
modeled along the lines of the well-known Oportunidades program in
Mexico. Its intention is to boost the income of poor families, on the
conditions that parents enroll their children (ages 6 to 15) in school and
that the minimum attendance rate is 85 percent. Parents of children
under 7 years old are expected to have them receive regular medical
check-ups and the basic childhood vaccinations. Pregnant women are
required to receive prenatal care and to attend classes on maternal and
child health. (For a basic description of the program, see O Globo
2006.) In most cases, the money is transferred to the mother of the
family via an electronic account. Eligibility for the program depends on
a family's earning less than R$120 per month. Those whose household
income is between R$60 and R$120 are classified as poor and can earn
up to R$45, depending on the number of people in the household.
Those whose household income is less than R$60 are assessed as
"extremely poor" and (again depending on the number of family mem-
bers) can earn up to R$95 per month (Ministerio do Desenvolvimento
Social 2006, 29). With more than 80 percent of the funds going to fam-
ilies under the poverty line (judged as between one-half and one min-
imum salary), the Bolsa is a highly targeted program. Increasingly, it is
viewed as an important contributor to reducing poverty and even
inequality (Soares et al. 2006).
The application process is straightforward. Brochures are written in
simple, plain language, complete with illustrations and a toll-free number
to call in the event of difficulties. Notably, receiving the Bolsa Familia
requires no political brokers or intermediaries. In this important respect,
it differs from traditional clientelism. Also unlike traditional patron-client
relationships, in which the client's receipt of patronage is conditional on
demonstrating political loyalty, there is no penalty for voting against the
government. Lula rapidly expanded the program's coverage, even in
states dominated by some of his staunchest political opponents.
The Ministry of Social Development (MDS) administers the Bolsa
Famrilia in conjunction with Brazil's municipal governments. The latter
are in charge of monitoring whether recipients meet the specified con-
ditions. The basic idea originated with a similar program, the innovative
Bolsa Escola, instigated by Cristovam Buarque when he was PT gover-
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 19
nor of Brasilia in 1995. It spread to a number of municipalities before
Cardoso federalized it in his second presidency. The contribution of
Lula's first administration was to unify into the Bolsa Familia what had
been four separate programs and to expand its coverage greatly over
time.8 At the end of 2004, the program was reaching 59 percent of eli-
gible families, and at the end of 2005, some 77 percent. By the end of
2006, some 11.1 million families were receiving Bolsa Familia benefits;
this number approximates 100 percent of those currently eligible.
The MDS estimates that the average recipient family has 4.1 mem-
bers, which, if correct, would imply that approximately one-quarter of
Brazil's citizenry is covered by the Bolsa Familia. But the geographic dis-
tribution is heavily skewed, reflecting Brazil's notorious regional inequal-
ities. In the largest municipalities and cities, only 2.7 million families are
enrolled, representing only 24 percent of the total recipients. Minas
Gerais and the states of the Northeast have been heavy beneficiaries.
Although the program still formally holds recipients to certain stan-
dards of conduct, given the extraordinarily rapid expansion of coverage,
it is not surprising that some degradation of oversight has occurred. As
one (anonymous) ministry official explained, "most municipalities
simply do not have the capacity to monitor new recipients at such a
rate" (MDS 2006b). This has led critics to contend that the BF is evolv-
ing into pure assistencialismo (a term used to describe the politics of
paternalistic "handouts"). While the program's original aim was to alle-
viate poverty in the short term while building human capital over the
long term, recent developments raise questions as to how much of the
latter is actually occurring.
What is clear, however, is how much political capital the Bolsa
Familia generated for Lula in his bid for reelection. As late as early 2005,
analysts observed how few social accomplishments the PT-led govern-
ment could actually claim (see, e.g., Hunter and Power 2005). After the
much-ballyhooed Fome Zero (Zero Hunger) program failed to get off
the ground in 2003-4, Lula was receiving low marks in this area. He set
his sights on reversing this situation in the second half of his adminis-
tration. Staking his prospects on the Bolsa Familia was a cost-effective
strategy. Although the Bolsa takes up a sizable share (38.0 percent in
2006) of Brazil's growing social assistance budget, the latter is a small
percentage (roughly 7 percent) of the overall social budget. Put in dif-
ferent terms, the Bolsa accounts for just 2.3 percent of direct monetary
transfers, compared with the 82 percent consumed by (far more regres-
sive) pensions (Hall 2006, 692-94).
The election results testify to the wisdom of Lula's acceleration of
social policy in the second half of his first term. An impressive state-
level correlation exists between Bolsa penetration in 2006 and vote
swing toward Lula compared to 2002. At the state level, Lula's vote share
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
20 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
correlates at .621 with the share of families covered by the Bolsa. It is
astonishing that in states like Bahia and Maranhdo, long ruled by oli-
garchic families, Lula was able to command 78 and 85 percent of all
votes, respectively. In Bahia, PT candidate Jacques Wagner managed to
win the governorship even against the candidate of the powerful Mag-
alhdes clan, in the election's first round, no less.9 It is also notable that
where it expanded later in the game, the Bolsa packed more of a punch
at the polls. In other words, the higher the percentage of state families
added to the Bolsa Familia rolls in the 18 months prior to the election,
the higher the vote swing toward Lula and against the PSDB.
One interesting and important effect of the Bolsa was to increase
voter turnout. Comparing the 2006 runoff to that of 2002 across Brazil's
27 states, interelection increase in turnout correlates at .544 with Bolsa
penetration. Looking only at the second round in 2006, a possible alter-
native explanation is that turnout rose mostly in states having guberna-
torial runoff elections, a factor known to lure voters back to the polls.
For example, in 2002, second-round presidential turnout in states with
gubernatorial runoffs was 6.14 percent higher than those without. In
2006, this difference diminished to 2.97 percent. In other words, states
without gubernatorial runoffs saw a much smaller dropoff in voter
turnout this year, even though in both 2002 and 2006 Lula was obvi-
ously sailing to a 20-point victory against his PSDB opponent. When
both Bolsa Familia and gubernatorial runoffs are used together to pre-
dict turnout in the 2006 second presidential round, Bolsa penetration
remains the superior explanation for voters' returning to the polls on
October 29. The bottom line is that Lula's social policies appear to have
significantly boosted enthusiasm and turnout among Brazil's poor
majority in the 2006 elections.
In sum, the social policy story is arguably the single most plausible
explanation for Lula's reelection. Put simply, the poor are significantly
better off now than they were when Lula assumed the government in
January 2003, and this group exerted the most influence in reelecting
him. It bears repeating that Lula did best among voters with a family
income of fewer than two minimum wages (approximately $327 per
month), a segment that comprises roughly 47 percent of the electorate.
Analysts need not reach for explanations based on concepts like iden-
tity politics or populism to understand this result. Lula's own humble
nordestino origins may well enhance his credibility as a caretaker of the
poor. Among the millions who struggle daily against grinding poverty,
many no doubt take comfort in his presence and find his example uplift-
ing. Yet it is worth recalling that before Lula held executive power, his
personal story did not compel poor people to vote for him at such high
rates. Therefore a fairly straightforward "pocketbook" explanation goes
a long way toward accounting for Lula's victory.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 21
Along similar lines, a measure of Lula's recent success is frequently
attributed to the use of populist tactics. If populism is a way to achieve
and exercise centralized political power based on creating direct links
between leader and mass, on mobilizing the latter through anti-elite dis-
course, and sometimes on using economic and social policy as an
instrument (Weyland 2001), the definition applies only partially to Lula's
situation. His increasing autonomy from his party in recent years
notwithstanding, a higher degree of organizational intermediation exists
between him and the masses than with figures like Peru's Alberto Fuji-
mori (1990-2000) or Venezuela's Hugo Chaivez (1999-present). By the
same token, Lula has not centralized governmental decisionmaking to
nearly the same degree, or drawn attention to himself through spectac-
ular acts like inaugurating new schools with helicopter rooftop landings
(Fujimori) or using inflammatory "friend versus foe," class-based rheto-
ric (Chivez). While the very notion of lulismo-that is, support for Lula
as distinct from support for his party (petismo)-suggests a personalis-
tic element to his appeal (Samuels 2006), the main story of Lula's reelec-
tion can be told without reference to the concept of populism.
Alckmin's Uphill Battle
In light of the three factors discussed above--corruption, economic
trends, and social policy-and their differential impact on Brazil's poor
majority, it is perhaps not surprising that Geraldo Alckmin could not gain
much traction in the campaign. Although Alckmin did appeal to voters
at the upper end of the income and education ladders, in numerical
terms they were too small a group to prevail. Notably, even in some
states where the PSDB's main coalition partner, the PFL, had strong roots,
Alckmin could not advance strongly. Given the indisputable economic
and social policy benefits that the Lula government had brought to less
privileged voters, one of the only legs for Alckmin to stand on was to
criticize Lula for the corruption that had taken place under his watch.
Alckmin did this in no uncertain terms in the October debates. (Lula had
skipped all the multicandidate debates in the first round.) Yet as we have
seen, even if an ethical discourse mattered to certain segments of the
electorate, it did not approach the salience of other pocketbook issues.
It did not help Alckmin that few voters were at all familiar with him
before the publicly sponsored electoral propaganda began to air on tel-
evision in August, or that as a well-educated, upper-middle-class doctor
from Sio Paulo, he had little resonance with poorer Brazilians from the
country's interior. This stood in marked contrast to Lula's ability to reach
out to such people by emphasizing his own humble Northeastern ori-
gins. Internal conflicts in Alckmin's own party probably also contributed
to his defeat. These conflicts stemmed in large measure from the future
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
22 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
presidential aspirations of Governors Jos6 Serra (Sio Paulo, elected
2006) and Aecio Neves (Minas Gerais, elected 2002, reelected 2006),
neither of whom bent over backward to reinforce the campaign of their
PSDB colleague. A frequent allegation is that both had their sights set
on running in 2010, and neither would want his path to the presidency
blocked by the prospect of a second Alckmin term.10
Although Alckmin performed respectably well in the first round, in
the runoff he allowed Lula to put him on the defensive by linking him
to the policies and performance of the Cardoso years. (Ironically, one
of Alckmin's initial advantages as a candidate was thought to be his own
weak connections to the former president, who was generally unpopu-
lar in the 1999-2002 period.) While such factors impinged on the out-
come, Alckmin's more fundamental impediment undoubtedly was that
he was a challenger going up against an incumbent who, by virtue of
his record in office, was popular with a majority of Brazilians.
COMPARATIVE POPULARITY:
LULA VS. His PARTY
If the October 29 election was a resounding endorsement of Lula, PT
candidates for the Chamber of Deputies did not, on the whole, experi-
ence similar fortunes. What explains Lula's popularity compared to that
of the political party he founded in 1980?
The first point to understand in this regard is that since his surprisingly
strong showing against Collor in 1989, Lula has generally enjoyed broader
popular acceptance than has the party organization. Although such a bifur-
cation is typical of presidential systems everywhere, Lula's first-round vote
share in all previous presidential contests far outstripped that of the party
in simultaneous races, and Lula grew faster than the PT over time. From a
slightly different perspective, although partisan identification with the PT
is higher than that of other parties, it has never exceeded 24 percent of the
entire electorate. Thus, especially in majoritarian races (governors and big-
city mayors) and plurality contests (senators and mayors in smaller munic-
ipalities), the PT needs to bring in voters who are unconnected to the party
or even to its ideological traditions. Lula has been more successful in this
regard than most of his fellow PT politicians.
Some overlap between the two groups notwithstanding, David
Samuels has argued that the voting base for Lula (lulistas) is actually
quite different from that which supports the party for more partisan rea-
sons (petistas). Whereas "the average petista participates actively in pol-
itics, is highly knowledgeable about politics, identifies as left of center,
supports clean government, and believes that his or her vote can make
a political difference" (Samuels 2006, 19), these characteristics do not
define the average Lula voter, especially in 2006. It appears that while
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 23
the PT's unique historical identity as a left party (albeit one that moder-
ated over time) appeals to a more select crowd (that is, those with the
above qualities and who tend to vote for the party for the Chamber of
Deputies), Lula has moved much more toward a "catch-all" profile.
The narrowcasting of the PT versus the broadcasting of Lula was
already visible in the 2002 election, as Lula deliberately sought the
autonomy necessary to moderate his program and profile to win the
presidency (Almeida 2006a). After entering the government in 2003, he
fell under increasing pressure to become "normalized" under the logic
of the situagdo. The party did not follow suit nearly to the same degree.
More Brazilians, especially poorer people, were impressed by what Lula
did with executive power than by the ideas and images that the PT
sought to project. Part of this can be explained by the political culture
of a presidentially centered system. Yet Lula's presidential style, and his
compelling personal story of rising from the depths of misery, also con-
tributed to his following (Hunter and Power 2005). In short, most Brazil-
ians are far more interested in Lula's personal story and the concrete
benefits he has provided as president than they are in the PT's partisan
story and the ideological cues it offers to voters.
Another factor in the incongruence between the popularity of Lula
and his party is that the 2005-6 corruption charges stuck to the PT more
than they did to the president himself. The main individuals "nailed" for
corruption were top PT figures, including the last two presidents of the
party and the treasurer. The media quickly constructed this as a partisan
affair. The media also went after the PT because it was, and still is, a
much easier target than the president. Lula was invariably accused of
sins of omission, not commission. While this image is not favorable, it
is qualitatively different from charges of direct involvement. Lula's
hands-off, Reaganlike management style in his first term also reinforced
this distinction, insofar as his personal distance from the various opera-
tional nuclei of his government afforded him a cushion of "plausible
deniability" when the corruption allegations first surfaced. Lula is ill
suited to charges of being a mastermind of anything, although the label
seemed particularly well suited to the political style of Jose Dirceu, his
once all-powerful former chief of staff.
The ever greater incongruence between Lula and the party still con-
cerns PT partisans and militants. While they recognize that Lula's
charisma and personal popularity brought the party a degree of visibil-
ity and standing that it probably could not have attained otherwise, they
view the president as transcending the previous "marriage of lulismo
with petismo" (Folha Online 2006). Not only have the policy positions
Lula has endorsed since 2003 diverged significantly from those of the
party's core, but the party organization recognizes that the electoral for-
tunes of a single politician are insufficient to guide it into the future.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
24 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
This argument has particular resonance given that the PT has no obvi-
ous presidential candidate to succeed Lula in 2010. Especially in a pres-
idential system, where the "overhang" of a winning presidential candi-
date can bestow advantage on a party, the lack of a future candidate,
combined with the party's major reputational setbacks, is no doubt a
cause for serious concern.
LULA DE Novo, COM A FOR~A DO POVO
This article has sought to explain Lula's resilience in the wake of the
corruption scandals plaguing his government. His smashing reelection
victory constitutes the clearest evidence yet of his comeback. The voters
who carried Lula to victory appear to have been strongly influenced by
the government's social policies, especially the Bolsa Familia, in con-
junction with minimum wage increases and their many ripple effects
throughout the economy, including raises in public pensions. Impres-
sive levels of macroeconomic stability made these policies possible and
gave voters confidence in the government, despite the many corruption
charges it faced. With the economic fundamentals in place, Lula was
able to use incumbency to his advantage.
In assessing incumbent advantages in 2006, it can be noted in pass-
ing that the Bolsa Familia, a program based on monthly cash transfers
to the poor, simply could not have worked in the hyperinflationary envi-
ronment that prevailed in Brazil before the presidency of Fernando
Henrique Cardoso. Years of painful reforms and newfound stability have
afforded Brazil some space for policy experimentation, and this has
redounded largely to Lula's advantage. The irony was not lost on Ger-
aldo Alckmin, who carried the standard of Cardoso's PSDB in 2006. In
the end, successful poverty-reducing policies, together with Brazil's
demographics, put challenger Alckmin in a most difficult position.
If Cardoso won the 1994 election because of the Plano Real, and if
Lula won in 2002 because he symbolized a change from eight years of
government by the PSDB-PFL, Lula's victory in 2006 may well go down
in history for its association with the Bolsa Familia. And if one of the
main legacies of the Cardoso era is price stabilization, which Brazilians
have come to expect as a matter of course, it is entirely possible that
income transfer programs along the lines of the Bolsa will assume the
form of a new entitlement for poorer Brazilians. Like Social Security in
the United States or the National Health Service in Britain, the Bolsa
Familia may indeed become the "third rail" of Brazilian politics, a polit-
ical totem that self-styled reformers can touch only at their own peril.
To the extent that the paternity of the Bolsa Familia becomes entwined
with the growing legend of Lula-a candidate in all five of Brazil's
modern democratic elections, and the victor in the last two-this pro-
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 25
gram may well become a political heirloom over which Lula's would-be
successors will do battle. These speculations, of course, are premature,
but there is no doubt that Bolsa Familia has had immediate and palpa-
ble effects on the president's political fortunes.
Notably, Lula was able finally to conquer states in which he had
been trounced previously as an opposition candidate. Some of them, for
example Maranhio and Bahia, are among the least competitive, most
oligarchic states in Brazil, if not Latin America as a whole. What does
this tell us? On the one hand, it suggests that voters in those contexts
do not have deep affective ties to their patrons and their parties, or at
least that these ties do not run deeply enough to determine their vote
choices. It also suggests that ideological proclivities are not, first and
foremost, what shapes these choices. The poorest, least educated citi-
zens in Brazil have seen material improvements in their lives since 2003.
In 2006, they made clear that they are available to be mobilized by
politicians who provide them with poverty-reducing, equity-enhancing
benefits. These voters also showed that they will turn out in unprece-
dented numbers to return those politicians to office, even though the
social benefits in question are provided through universal (means-
tested) criteria and require none of the political intermediation that has
shaped Brazilian politics for centuries.
It remains to be seen whether Lula's advance in the least developed,
most oligarchic states will disrupt well-established local networks of
clientelism and prompt the demise of longstanding oligarchies. Yet the
twin democratic values of accountability and responsiveness appear
alive and well, at least in this election. Kitschelt (2000) argues that clien-
telism of the kind employed by Lula is a fully representative "citizen-
politician linkage," not to be diminished or disparaged as being some-
how inferior to a programmatically based linkage. Historical petistas
who spent two decades in the opposition defending a left project of
structural transformation instead of the arguably palliative and shorter-
term (though popular) programs Lula has pursued might disagree.
In sorting out this debate, relevant issues to consider include how
fundamental and long-lasting the impact of Lula-style benefits will be on
human development, and whether the money allocated to ends like min-
imum wage increases and conditional cash transfer programs would be
better spent on reforms that would enhance the life conditions of poor
Brazilians over the long term. At the same time, it is not clear precisely
what form a more structural path of change could or should take. What
economic models and social policy alternatives would even be reason-
able to pursue? Regardless of which side of this debate one embraces,
one thing is not in dispute: in Latin America's largest democracy, where
both continuity and change are ever apparent, it was a burgeoning citi-
zen-politician linkage that returned Lula so resoundingly to office.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
O
z
c,,
z
>
0
O
z
0
t'
??
Appendix. State Abbreviations and Background Data
Lula PT Chamber Party of
Voters HDI Rank Runoff Vote Winning
Code State Region (millions)a 2000 in HDI Vote Share Share Governor
AC Acre North 0.41 0.643 18 52.37 3.5 PT
AL Alagoas Northeast 1.86 0.579 26 61.45 11.8 PSDB
AM Amazonas North 1.78 0.681 15 86.80 9.3 PMDB
AP Amapa North 0.36 0.705 13 70.40 6.5 PDT
BA Bahia Northeast 9.11 0.631 20 78.08 16.7 PT
CE Ceara Northeast 5.36 0.624 22 82.38 11.5 PSB
DF Distrito Federal Center-West 1.66 0.815 1 56.96 19.0 PFL
ES Espirito Santo Southeast 2.34 0.730 10 65.54 13.3 PMDB
GO Goias Center-West 3.73 0.736 9 54.78 17.8 PP
MA Maranh~o Northeast 3.92 0.569 27 84.63 2.7 PDT
MG Minas Gerais Southeast 13.68 0.736 8 65.19 9.9 PSDB
MS Mato Grosso do Sul Center-West 1.56 0.746 7 44.98 16.1 PMDB
MT Mato Grosso Center-West 1.94 0.722 11 49.69 7.8 PPS
PA Pard North 4.16 0.678 16 60.12 18.1 PT
PB Paraiba Northeast 2.57 0.587 24 75.01 9.5 PSDB
PE Pernambuco Northeast 5.83 0.646 17 78.48 12.5 PSB
PI Piaui Northeast 2.07 0.582 25 77.31 21.8 PT
PR Parand~ South 7.12 0.752 6 49.25 13.8 PMDB
RJ Rio de Janeiro Southeast 10.89 0.780 5 69.69 18.9 PMDB
RN Rio Grande do Norte Northeast 2.10 0.631 21 69.73 25.2 PSB
RO Rond6nia North 0.99 0.690 14 55.83 7.9 PPS
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
z
H
01
zI
01
0
01
01
z
0
Lula PT Chamber Party of
Voters HDI Rank Runoff Vote Winning
Code State Region (millions)a 2000 in HDI Vote Share Share Governor
RR Roraima North 0.23 0.720 12 38.51 7.9 PSDB
RS Rio Grande do Sul South 7.75 0.786 4 44.65 19.7 PSDB
SC Santa Catarina South 4.17 0.788 3 45.47 9.8 PMDB
SE Sergipe Northeast 1.30 0.622 23 60.16 18.8 PT
SP Sdo Paulo Southeast 28.04 0.792 2 47.74 16.5 PSDB
TO Tocantins North 0.88 0.639 19 70.27 11.3 PMDB
aTotal electorate is 125.9 million.
Sources: Electoral data from TSE 2006. Human Development Index (HDI), calculated by IPEA on the basis of the 2000 national census. See IPEA
2006.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
28 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
NOTES
The authors thank several persons with whom they engaged the topic of
this article intensely during the second half of 2006: Leslie Bethell, David Fleis-
cher, Vinicius Torres Freire, Kathryn Hochstetler, Marcus Melo, Rachel
Meneguello, Anthony W. Pereira, Carlos Pereira, David Samuels, Andrew J.
Stein, Kurt Weyland, and Cesar Zucco. They are also grateful to five anonymous
reviewers of LAPS for helpful comments on an earlier draft. The authors assume
all responsibility for errors of fact or interpretation.
1. The HDI is a composite index based on three indicators of human well-
being: life expectancy, educational attainment, and per capita income. The scale
ranges from 0 to 1, with lower values indicating poorer human development.
See IPEA 2006.
2. For a detailed study of Lula's campaign and the PT's marketing strategy
in the 2002 election, see Almeida 2006a.
3. In table 3, the indicator of political democracy is adapted from Vanhanen
(1990), who operationalized Dahl's 1971 concept of polyarchy as (participation
x contestation). The dimension of contestation incorporates both the closeness
between the top two candidates for governor and the overall vote dispersion of
the field of candidates, giving equal weight to both factors. Participation is meas-
ured simply as voter turnout. Following Vanhanen, the measures of participa-
tion and contestation are then multiplied to produce a democracy score for each
election, and the state average is determined for the two decades preceding
Lula's election ("democratic tradition"). Turnout is the average of all six legisla-
tive elections, 1982-2002; competitiveness is measured for the same years for
governor, except in DF, TO, RR, and AP, which had only four gubernatorial elec-
tions beginning in 1990.
4. Defeated by Lula in the 2002 presidential election, Serra was elected
mayor of Sdo Paulo in 2004. He served only 15 months, resigning in March 2006
to run for the state governorship, which he won by defeating PT senator Aloizio
Mercadante in the first round.
5. For a good overview of how institutional design leads to "exchange pol-
itics" in Brazil, see Geddes and Ribeiro 1999; a broader, anthropological study
of clientelism in Congress is provided by Bezerra 1999. For specific effects of
the party and electoral systems on corruption, see Mainwaring 1999, 175-218;
Ames 2001, 56-61.
6. Interesting, however, is the finding of PT pollster Gustavo Venturi that a
tendency among more educated citizens to be more critical was offset some-
what by a higher propensity to attribute the PT's conduct to structural con-
straints, such as the stiff requirements of campaign financing and the difficulties
all Brazilian presidents face in forging a legislative majority (Venturi 2006,
20-26).
7. We are grateful to Anthony W. Pereira for this insight on pocketbook
voting by the middle class.
8. The four programs were the Bolsa Escola, Bolsa Alimenta~ao, Auxilio
Gas, and Cartao Alimenta~ao.
9. PT candidates also won gubernatorial office in Acre, Piaui, Pard, and
Sergipe, all Northern or Northeastern states (see appendix).
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
HUNTER AND POWER: REWARDING LULA 29
10. To overcome this concern by Serra and Neves, Alckmin was reportedly
ready (if elected) to reimpose Brazil's historical ban on immediate reelection of
presidents, removed by Cardoso in 1997.
REFERENCES
Almeida, Alberto Carlos. 2006a. Por que Lula? 0 contexto e as estrategiaspoliti-
cas que explicam a eleigdo e a crise. Rio de Janeiro: Record.
. 2006b. Brasil: a luta entre o arcaiico e o moderno. Revista Insight Inteligdn-
cia: September. <www.insightnet.com.br/inteligencia>
Ames, Barry. 2001. The Deadlock of Democracy in Brazil. Ann Arbor: University
of Michigan Press.
Banco Central do Brasil. 2006a. Effective Interest Rates. <www.bcb.gov.br/
?indicators>
. 2006b. Sales Volume Index in the Retail Sector. <www.bcb.gov.br/?indicators>
Bezerra, Marcos Otivio. 1999. Em nome das bases: politica, favor e dependincia
pessoal. Rio de Janeiro: Relume Dumari.
Bianchi, Alvaro, and Ruy Braga. 2005. The Lula Government and Financial Glob-
alization. Social Forces 83, 4: 1745-62.
Dahl, Robert. 1971. Polyarchy: Participation and Opposition. New Haven: Yale
University Press.
Datafolha. 2006. Opinion Poll Project No. PO3369. <www.datafolha.com.br>
Desposato, Scott W. 2006. From Revolution to Rouba Mas Faz? Re Vista: Harvard
Review of Latin America 5, 1 (Spring-Summer): 29-32.
Economist Intelligence Unit. 2006. Country Report: Brazil. London: EIU. 34.
September.
Estado de Sdo Paulo. 2006. Para os mais pobres, um crescimento chines. Inter-
view with IPEA economist Ricardo Paes de Barros. November 12.
<www.estadao.com.br>
Ferraz, Claidio, and Frederico Finan. 2006. Exposing Corrupt Politicians: The
Effects of Brazil's Publicly Released Audits on Electoral Outcomes. Unpub-
lished mss. University of California, Los Angeles. September.
Figueiredo, Carlos Mauricio, Marcus Melo, and Carlos Pereira. 2006. Does Cor-
ruption Enhance the Role of Reelection Incentives? Evidence from Brazil's
Municipal Audit Reports. Paper prepared for the 10th Annual Conference of
the International Society for New Institutional Economics (ISNIE), Boulder,
September 21-24.
Folha de Sdo Paulo. 2005. Cresce percepCgo de responsabilidade de Lula na cor-
rupgdo. October 23.
Folba Online. 2006. "Petismo" deve prevalecer sobre "lulismo," diz dirigente do
PT. November 13. <www.folha.com.br>
Fundagao Perseu Abramo. 2006. Imagem partidiria e cultura politica. Public
opinion poll. March. So Paulo: Funda~go Perseu Abramo.
Geddes, Barbara, and Artur Ribeiro Neto. 1999. Institutional Sources of Corrup-
tion in Brazil. In Corruption and Political Reform in Brazil: The Impact of
Collor's Impeachment, ed. Keith S. Rosenn and Richard Downes. Coral
Gables: North-South Center Press. 21-48.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
30 LATIN AMERICAN POLITICS AND SOCIETY 49: 1
O Globo (Rio de Janeiro). 2006. A bolsa e os bols6es. Caderno especial, eleigoes
2006. August 12.
Hall, Anthony. 2006. From Fome Zero to Bolsa Famfilia: Social Policies and
Poverty Alleviation Under Lula. Journal of Latin American Studies 38, 3
(November): 689-709.
Hunter, Wendy, and Timothy J. Power. 2005. Lula's Brazil at Midterm. Journal
of Democracy 16, 3 (July): 127-39.
Inglehart, Ronald, and Christian Welzel. 2005. Modernization, Cultural Change,
and Democracy: The Human Development Sequence. New York: Cambridge
University Press.
IBOPE. 2006. Time series data. <www.ibope.com.br>
Instituto de Pesquisa Econ6mica Aplicada (IPEA). 2006. Indicadores sociais:
IDH. <www.ipeadata.gov.br>
Kitschelt, Herbert. 2000. Linkages Between Citizens and Politicians in Democra-
tic Polities. Comparative Political Studies 33, 6-7: 845-79.
Leal, Victor Nunes. 1949. Coronelismo, enxada e voto: o municipio e o regime
representativo no Brasil. Sdo Paulo: Alfa e Omega.
Mainwaring, Scott. 1999. Rethinking Party Systems in the Third Wave of Democ-
ratization: The Case of Brazil. Stanford: Stanford University Press.
Ministerio do Desenvolvimento Social (MDS). 2006a. Programa Bolsa Famflia,
guia do gestor. Brasilia: MDS.
- . 2006b. Ministry official (anonymous). Author interview, Brasilia, August 16.
Neri, Marcelo Cortes, coord. 2006. Poverty, Inequality, and Labor Dynamics: The
Second Real. Report. Rio de Janeiro: FundaCdo Geftilio Vargas.
Samuels, David. 2006. Sources of Mass Partisanship in Brazil. Latin American
Politics and Society 48: 2 (Summer): 1-27.
Singer, Andre. 1990. Collor na periferia: a volta por cima do populismo? In De
Geisel a Collor: o balango da transiado, ed. Bolivar Lamounier. Sio Paulo:
Sumare. 135-52.
Soares, Faibio Veras, Sergei Soares, Marcelo Medeiros, and Rafael Guerreiro
Os6rio. 2006. Programas de transferencia de renda no Brasil: impactos
sobre a desigualdade. IPEA Textopara Discussdo 1228 (October).
Tribunal Superior Eleitoral (TSE). 2006. Election data. <www.tse.gov.br>
Vanhanen, Tatu. 1990. The Process of Democratization: A Comparative Study of
147 States, 1980-1988. New York: Crane Russak.
Venturi, Gustavo. 2006. A opiniao puiblica diante da crise. Teoria e Debate 19,
66 (April-June): 20-26.
Weyland, Kurt. 2001. Clarifying a Contested Concept: Populism in the Study of
Latin American Politics. Comparative Politics 34 (October): 1-22.
This content downloaded from 128.83.63.20 on Wed, 05 Jul 2017 00:02:11 UTC
All use subject to http://about.jstor.org/terms
