Cash Transfers and Formal Labor Markets: Evidence from Brazil — Policy Research Working Paper 9778
Summary
This World Bank Policy Research Working Paper 9778, dated September 2021, is titled Cash Transfers and Formal Labor Markets: Evidence from Brazil, by François Gerard, Joana Naritomi, and Joana Silva. The abstract says the paper studies the impact of the large-scale program Bolsa Familia on local labor markets in Brazil. It says the analysis finds that an expansion of Bolsa Familia increased local formal employment, using variation in the size of the reform across municipalities. It says the evidence is consistent with multiplier effects of cash transfers in the local economy, which dominate potential negative effects on formal labor supply among beneficiaries.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
- Issuer
- Source documents
- Document type
- Cash Transfers And Formal Labor Markets Evidence From Brazil
- Case
- Cash Transfers And Formal Labor Markets Evidence From Brazil
Cite this document
“Cash Transfers and Formal Labor Markets: Evidence from Brazil — Policy Research Working Paper 9778.”
Original at doi.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/cash-transfers-and-formal-labor-markets-evidence-from-brazil/
, SHA-256 808b46f071f88118eec34cd758cbc064de18e9de6e98684577a31028cf4ef191.Full text
Public Disclosure Authorized
Policy Research Working Paper 9778
Public Disclosure Authorized
Cash Transfers and Formal Labor Markets
Evidence from Brazil
François Gerard
Joana Naritomi
Public Disclosure Authorized
Joana Silva
Public Disclosure Authorized
Latin America and the Caribbean Region
Office of the Chief Economist
September 2021
Policy Research Working Paper 9778
Abstract
Cash transfer programs have expanded widely in developing labor market, where earnings are more easily verifiable.
countries and have been credited for sizable reductions in Yet, the analysis finds that an expansion of Bolsa Familia
poverty. However, their potential disincentive effects on increased local formal employment, using variation in the
beneficiaries’ labor supply have spurred a heated policy size of the reform across municipalities. The evidence is
debate. This paper studies the impact of a large-scale pro- consistent with multiplier effects of cash transfers in the
gram Bolsa Familia in Brazil on local labor markets in a local economy, which dominate potential negative effects
context where such concerns could be particularly strong: on formal labor supply among beneficiaries.
eligibility is means-tested and the paper focuses on the formal
This paper is a product of the Office of the Chief Economist, Latin America and the Caribbean Region. It is part of a larger
effort by the World Bank to provide open access to its research and make a contribution to development policy discussions
around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The
authors may be contacted at jsilva@worldbank.org.
The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development
issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the
names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.
Produced by the Research Support Team
Cash Transfers and Formal Labor Markets:
Evidence from Brazil *
François Gerard
Queen Mary University of London & CEPR
Joana Naritomi
London School of Economics & CEPR
Joana Silva
World Bank & Catolica Lisbon & CEPR
Originally published in the Policy Research Working Paper Series on September 2021. This
version is updated on October 2021.
To obtain the originally published version, please email prwp@worldbank.org.
* f.gerard@qmul.ac.uk, J.Naritomi@lse.ac.uk, jsilva@worldbank.org. We are grateful to Fernanda
Brollo, Adriana Camacho, Claudio Ferraz, Giacomo de Giorgi, Gustavo Gonzaga, Paul Niehaus, Luis Hen-
rique Paiva, Rodrigo Soares, Serguei Soares, Pablo Surico, Carlos Vegh, Guillermo Vuletin, and participants
in several seminars for helpful comments. Daniel M. Angel, Sebastian Melo, Alejandra Martinez, Samira
Noronha, Rafael Prado Proenca, Lorenzo Pessina, and Divya Singh provided outstanding research assis-
tance. We thank the Brazilian Ministries of Social Development and Labor for data access. We remain
responsible for any errors. We thank the LSE Inequalities Institute (III) for financial support. We declare
that we have no relevant or material financial interests that relate to the research described in this paper.
Cash transfer programs are the main type of safety net in middle-income countries,
covering about 30 percent of the poor (Honorati et al. 2015). It is well established that
these programs reduce poverty and improve the lives of beneficiaries in several important
dimensions.1 However, their potential impacts on the labor market have spurred a heated
policy debate. Fears over “lazy welfare recipients" that often fuel controversy over the
future of these programs are generally unwarranted (Banerjee et al. 2017). Yet, concerns
remain that the targeting of some of these programs, such as the use of means testing,
could distort beneficiaries’ incentives to work, especially in the formal economy where
earnings are more easily verifiable (Levy 2008; for recent evidence, see Bergolo and Cruces
2021). By contrast, some studies argue that cash transfers can help beneficiaries find better
jobs by unlocking liquidity constraints (Baird et al. 2018), or that they can stimulate labor
markets through increases in labor demand due to multiplier effects in the local economy
(e.g., Neri et al. 2013, Egger et al. 2019). In practice, these effects can coexist, and there is
limited evidence on their relative importance to inform the policy debate.2
This paper contributes to the literature by studying the effects of a large-scale means-
tested cash transfer program – Programa Bolsa Familia (PBF) in Brazil – on aggregate for-
mal employment in local labor markets. This is the margin of employment most likely
to be affected by disincentive effects from means testing. Formal job creation has also
been a primary focus of economic policies in Latin America over the past decades (Levy
2008), as formal jobs are more likely to provide workers social security coverage and bet-
ter working conditions (e.g., Perry et al. 2007) and are associated with higher output and
total factor productivity (e.g., Ulyssea 2020). We use longitudinal administrative data
covering the universe of formal workers and PBF beneficiaries, along with the benefits
disbursed through the program. With these data, we estimate the impacts of a reform
that increased the total number of PBF beneficiaries by more than 17 percent (or almost
two million additional families) in 2009. We find that higher PBF payments increased local
formal employment, by exploiting variation in the size of the expansion across Brazilian
municipalities. The evidence is consistent with multiplier effects of cash transfers in the
local economy, which dominate potential negative effects on formal labor supply among
PBF beneficiaries.
Our main empirical strategy relies on the fact that, because of budgetary constraints,
PBF is not an entitlement program. The total number of beneficiaries is set nationally and
1 See Bastagli et al. (2016), Fiszbein and Schady (2009), and Garcia and Saavedra (2017) for systematic
literature reviews and Viana et al. (2018) for studies on the Bolsa Familia program in Brazil.
2 This debate is particularly heated in Brazil, with right-leaning politicians emphasizing negative effects
on the labor supply of Bolsa Familia beneficiaries, and left-leaning politicians highlighting positive effects on
beneficiaries and local economies (we provide some illustrative quotes in Online Appendix A).
2
specific rules guide the allocation of slots across municipalities. The 2009 reform increased
the national cap and changed the methodology used to allocate slots. In our analysis, we
use a Difference-in-Differences design that exploits the variation in the intensity of the
reform across municipalities caused by this change in methodology. The number of bene-
ficiaries increased sharply in 2009 for municipalities in the top 50 percent of our measure
of treatment intensity, and it was more than 10 percent higher in 2010-2011 compared
with 2007-2008. By contrast, it was similar between these two periods for municipalities
in the bottom 50 percent of our measure of treatment intensity. We estimate a differential
increase in total PBF payments of 13.6 percent between the two groups.
Our main finding is that this differential increase in total PBF payments led to an in-
crease in the number of private-sector formal jobs that reached 2 percent by 2011, two
years after the reform. The effect on total payroll is slightly smaller (1.7 percent) because
the increase in formal employment is concentrated in lower-skill occupations. We find no
effect on public employment. The results are robust across various specifications and def-
initions of treatment intensity, and the formal employment effect is larger when we com-
pare groups of municipalities with a larger differential increase in PBF payments (those in
the top 75 percent versus the bottom 25 percent of our measure of treatment intensity).
The results are consistent with larger PBF outlays stimulating local labor demand
through a multiplier effect. Indeed, we find a similar increase in formal employment
among workers who were never part of the program.3 We provide three additional pieces
of evidence of an overall increase in activity in the local economy. First, we show that our
results are not simply driven by a reallocation of formal employment across neighboring
municipalities within a greater local economy. The results are similar if we replicate the
same analysis at a higher level of aggregation than the municipality (i.e., micro-regions).
Second, we estimate effects on municipal gross domestic product (GDP) and taxes paid
in the municipality. The results are in line with our formal employment results, with in-
creases in local GDP and local taxes reaching 1.7 and 2.7 percent by 2011, respectively.
Third, we document a clear increase in the municipal GDP attributed to non-tradable in-
dustries after the reform. Point estimates for the relative increase in formal jobs are also
larger for non-tradables versus tradables, although our estimates become noisy when we
explore heterogeneity across industries.4
In the last section of the paper, we explore whether the overall increase in formal em-
3 The type of “market externalities” studied in, e.g., Lalive et al. (2015), through which decreases in job
search among beneficiaries can increase job-finding among non-beneficiaries, could rationalize this specific
result but not the overall increase in formal employment (considering all workers together).
4 Non-tradable industries are more informal (Dix-Carneiro and Kovak 2019), such that a similar relative
change in formal employment could be consistent with a larger overall impact on non-tradable industries.
3
ployment occurs despite negative effects of PBF on beneficiaries’ formal labor supply or
are partly driven by positive responses among beneficiaries.5 We compare formal em-
ployment outcomes for families that are eligible for different benefit amounts within a
local labor market through a Regression Discontinuity (RD) design. We exploit income
per capita cutoffs that determine eligibility for benefits per child conditional on school at-
tendance and health checks (“poverty line”) and for an additional unconditional benefit
(“extreme poverty line”). We find no evidence that benefits increase formal labor sup-
ply. In fact, our results are consistent with recent evidence in the Latin American context
that substitution effects created by means testing can induce beneficiaries to reduce their
formal labor supply to retain eligibility (Bergolo and Cruces 2021).
Taken together, our results indicate that cash transfer programs can have positive ef-
fects on formal labor markets and the local economy at large, even if they create potential
disincentive effects on beneficiaries’ labor supply. To put the magnitudes in perspective,
we estimate the implied cost per formal job, following the literature on the impact of local
government spending on local labor markets.6 We obtain a point estimate of US$5,600
per year, or about 2.1 times the minimum wage at the time. This is much lower than
estimates for the United States,7 but closer to the estimates in Corbi et al. (2019) for the
impact of inter-governmental transfers on formal employment in Brazilian municipalities
(US$8,000). Direct transfers to poor families with likely high marginal propensity to con-
sume may imply larger multipliers than transfers to local governments. Interestingly, we
also find that the cost per formal job is much lower in municipalities with lower informal-
ity rates at baseline. This is consistent with the possibility that more jobs are at the margin
of formalization in those labor markets. By contrast, the cost per formal job is closer to
U.S. estimates in more informal labor markets. This is consistent with the possibility that
our analysis misses economic activity generated in the informal sector in those labor mar-
kets, and that it takes larger demand shocks to create formal jobs where informality is
pervasive.
This paper contributes to the literature on the aggregate effects of social transfers on
5 A recent literature has argued that there is no micro-evidence that cash transfers discourage recipients
from working (e.g., Banerjee et al. 2017). However, many of the programs discussed in this literature only
generate income effects because targeting is performed through proxy-means tests that are reassessed infre-
quently and do not include labor market outcomes. By contrast, means-tested programs can generate both
income and substitution effects as workers may change their labor supply, particularly in the formal sector
where earnings are more easily verifiable, to obtain or retain eligibility.
6 This literature often focuses on the “cost per job” rather than the “GDP multiplier,” in part because
local GDP is measured more noisily (Zidar, 2019; Chodorow-Reich, 2019).
7 For instance, Chodorow-Reich et al. (2012), Suárez Serrato and Wingender (2016), and Zidar (2019)
obtain estimates of US$30,000. Our 95 percent confidence interval is relatively wide, however, going up to
US$26,833.
4
local labor markets. While most of the literature focuses on the effects of social protection
policies in developed countries (e.g., Hagedorn et al. 2013, Lalive et al. 2015, Chodorow-
Reich and Karabarbounis 2016, Marinescu 2017), a smaller, but growing, literature studies
developing countries. Angelucci and De Giorgi (2009) highlight the importance of con-
sidering the effect of cash transfers on the overall local economy. They find evidence of
indirect effects of the Progresa program in Mexico (which is not means-tested) on the con-
sumption of ineligible households living in the same villages. The effect operates entirely
through the insurance and credit markets in their setting, however, while we provide evi-
dence of multiplier effects impacting the local labor market. A possible reason for this dif-
ference is that their setting consists of poor rural villages, while most PBF families live in
urban areas. Bosch and Campos-Vazquez (2014) estimate the aggregate effects of extend-
ing health insurance coverage to workers outside the formal sector in Mexico. They find
reductions in local formal employment, driven by negative (substitution) effects on for-
mal labor supply among beneficiaries. In-kind transfers, especially in the form of health
insurance coverage, may not lead to similar multiplier effects in the local economy as cash
transfers.8 Imbert and Papp (2015) and Muralidharan et al. (2017) estimate the aggregate
effects of a public employment program on local labor markets in India. Such programs
deliver benefits in the form of (manual) employment, and therefore operate through a
different channel, driving up private-sector wages and earnings by improving workers’
outside option (Ravallion 1987, Basu et al. 2009). These papers also focus on rural agricul-
tural labor markets, whereas we study urban and rural formal labor markets, spanning
different types of sectors and skills. More recently, Egger et al. (2019) study how a one-
off nongovernmental transfer affects local recipients and rural economies in Kenya. They
find large multiplier effects on consumption and assets but no effects on employment.
Our analysis studies a persistent expansion of a large-scale government welfare program
that has been running for more than 15 years.
Our results also contribute to the extensive literature on the effect of social transfers on
beneficiaries’ labor supply in developing countries. Recent studies argue that cash trans-
fers do not discourage work among beneficiaries (see Baird et al. 2018 for a review of the
literature). However, much of that discussion focuses on programs that use proxy-means
testing to define eligibility based on geographic location and household assets (e.g., Baner-
jee et al. 2017), such that these programs only generate income effects. We study a context
in which disincentive effects are likely stronger: under means testing, labor supply could
be affected by substitution effects as well, particularly in the formal sector where earnings
8 Health insurance cannot be exchanged for cash, and it may improve families’ health care rather than
imply large changes in their budget constraint, depending on their health care spending at baseline.
5
are third-party reported to the government. Several other papers in the Latin American
context have found formal labor supply responses to cash transfer programs that changed
the relative returns of formal employment (e.g., Garganta and Gasparini 2015; Bergolo
and Cruces 2021; Gerard and Gonzaga 2021). Our findings highlight the importance of
also accounting for aggregate effects to capture the impact of these policies on formal la-
bor markets in full. This point is particularly policy-relevant: as countries develop and
income becomes more verifiable across the income distribution (Jensen 2019), the use of
means testing for targeting social transfers is bound to expand around the world.
The paper is organized as follows. Section 1 describes the institutional background
and data. It also documents the extent to which PBF beneficiaries participate in the formal
labor market. Section 2 lays out our main empirical strategy. Section 3 presents the results,
and Section 4 provides additional analyses to explore mechanisms, including formal labor
supply responses among PBF beneficiaries. Section 5 concludes.
1 Institutional details and data
This section provides important background information for the empirical analysis. We
begin by briefly introducing the Bolsa Familia program (PBF) and some of its relevant in-
stitutional details for our period of analysis, which is from the beginning of 2007 to the
end of 2011. We then describe the various data sets that we use. Finally, we use these data
to provide some descriptive statistics on PBF beneficiaries and their participation in the
formal employment sector.
1.1 Bolsa Familia Program (PBF)
PBF is the largest conditional cash transfer program in the world and the main social
program for the poor in Brazil. It was created in 2004, bringing together and expanding
existing social transfers (Bolsa Escola, Bolsa Alimentação, and Auxílio Gás). As of 2012, it
reached 13.9 million families or around a quarter of the Brazilian population, for a cost of
about 0.6 percent of GDP. Pioneered by Brazil and Mexico, conditional cash transfer pro-
grams have expanded widely over the past 20 years and became the main type of safety
net in middle-income countries, covering 30 percent of the poor (Honorati et al. 2015).
Nowadays, they exist in almost every country in Latin America, as well as in Bangladesh,
Cambodia, Indonesia, Malawi, Morocco, Pakistan, and South Africa, among other coun-
tries.
There are two important features of PBF for our study. First, it is means tested. Second,
6
due to budgetary restrictions, it is not an entitlement program.
A. Means testing. Eligible families are classified into “poor” and “extreme poor” cat-
egories based on two income per capita thresholds.9 Extreme poor families are eligible
for an unconditional monthly benefit (the “basic benefit”) and monthly benefits per child
conditional on the child’s school attendance and health checks (the “variable benefits”).10
Poor families are only eligible for the conditional variable benefits.11 Thus, in terms of
the eligibility criteria, PBF is more similar to welfare programs in richer countries; most
developing countries rely instead on proxy-means testing. For example, in Mexico and
Colombia, families are assigned a poverty index score computed from information col-
lected through a national survey conducted every three to four years (see, e.g., Cama-
cho and Conover 2011). One reason for these differences in targeting between richer and
poorer countries comes from differences in how easily governments can observe families’
income. Indeed, income tends to become more verifiable as countries develop (Jensen
2019). Brazil is thus an interesting policy case as it lies in the middle of the development
path: it has a large informal economy, but it also has very developed administrative data
systems, including detailed records of formal employment income.
At the core of the targeting of PBF is Cadastro Unico, the “census of the poor,” which
is a federal registry that is continuously updated. Municipalities must register families
through interviews, verify the information provided by them, and ensure that it is up-
dated following changes, for example, in family income or family size. This process is
based on a questionnaire that asks questions about income, housing, and assets, among
other characteristics. Importantly, Cadastro Unico is not used exclusively for PBF. Many
social programs with different eligibility criteria use this registry. For this reason, Cadastro
Unico aims to include all families with income per capita below one-half of the minimum
wage (R$255 in 2010), which is much higher than the poverty threshold (R$140 in 2010).
As a result, it includes families that are not eligible for PBF. To be eligible for any govern-
ment benefit that uses Cadastro Unico, families must have a valid registration (complete
and up-to-date); at a minimum, a valid registration must be updated every two years.
The income per capita definition in Cadastro Unico is computed based on information
reported by families. During the interview, the applicant reports the value for each of
9 These income thresholds are defined at the national level. During the period of analysis, they were
revised once in 2009: the extreme poverty threshold was increased from R$60 to R$70 (∼US$35), and the
poverty threshold from R$120 to $140 (∼US$70).
10 The variable benefits include specific amounts for pregnant women, lactating women, children
younger than age 15 years, and adolescents ages 15 to 18.. The number of variable benefits that a family can
receive is capped by the total amount and a maximum number of children.
11 During the period of analysis, the basic benefit increased from R$58 to R$62 in 2008. Variable benefits
per child increased from R$18 to R$20 in 2008. Variable benefits per adolescent increased from R$30 to R$33
in 2009.
7
seven income categories – earnings in the last month, earnings in the last 12 months, dona-
tions, pensions, unemployment benefits, alimony, and other income – for each household
member. Income per capita is computed in three steps. The first step takes the minimum
between the average monthly income in the last 12 months and the last month’s income for
each individual. The second step adds all other income categories to this value. The third
step sums the individual monthly income across all household members and divides it by
the number of household members, including all adults and children registered under the
same household (names and ID numbers are recorded for each household member).12
There are two main strategies to prevent income misreporting. First, the income ques-
tions come at the end of the questionnaire so that answers to the previous questions about
assets and social demographics help the interviewer gauge the veracity of the total re-
ported income. The law suggests as best practice that registration should be done in the
family’s home to facilitate verification of answers to the questionnaire.13 Second, audits
can be conducted following citizens’ complaints or red flags arising from cross-checking
the information in Cadastro Unico with data from formal employment records and the
Brazilian Social Security System. Government employees may visit families to update
their information or require them to update their information at the local administration
office. Nonetheless, the large informal sector in the Brazilian economy, and limited ca-
pacity to cross-check information systematically and follow up on each case, leave scope
for discrepancies between the information in Cadastro Unico and families’ true income per
capita. We return to this discussion in Section 4.2. For more details on eligibility and
enforcement, see Lindert et al. (2007).
B. Benefits are not an entitlement. Eligible families are not entitled to become PBF
beneficiaries. The federal budget finances the program and sets the total number of slots
for PBF beneficiaries at the national level accordingly. Figure 1a displays the total number
of slots and the total number of beneficiaries over time. The total number of slots was set
in 2003 (before the start of the program) at 11.2 million. It was then revised in 2006 (11.1
million), 2009 (13 million), and 2012 (13.7 million). In the first few years, as beneficiaries of
other welfare programs were transferred to PBF, and municipalities were registering new
families in Cadastro Unico, the number of beneficiaries gradually increased. It only reached
the national cap in mid-2006, after the first reform. The national cap was then binding
until the 2009 reform when the total number of slots increased by about 15 percent. The
expansion was rolled out progressively between the second and fourth quarters of 2009, as
shown in Figure 1b. The total number of beneficiaries decreased in the first quarter of the
12 This process refers to the procedure used during our period of analysis.
13 See link to Portaria 177, article 5, which states that registration should be performed in the family’s
home as a default option.
8
F IGURE 1: T OTAL NUMBER OF PBF SLOTS AND BENEFICIARIES OVER TIME
(a) From 2004 to 2014 (levels) (b) Around the 2009 reform (relative changes)
Note: This figure presents the evolution of the total number of PBF slots and beneficiaries at the national level, in levels since the
creation of PBF in 2004 (panel a) and in relative changes around the 2009 reform (panel b). Vertical lines in panel (a) indicate the timing
of reforms of the program, which updated the total number of slots and municipal quotas (in 2006, 2009, and 2012). Vertical lines in
panel (b) indicate the start and the end of the rollout of the 2009 reform studied in the paper.
year, due to a “cleanup” of Cadastro Unico through cross-checks with other administrative
records. It then increased by more than 10 percent as the expansion was rolled out, and
continued to increase until it reached the national cap by the end of 2011.14 In this paper,
we focus on the impacts of the 2009 PBF expansion highlighted in Figure 1b.
The total number of slots is divided across municipalities based on the “municipal
quotas” that were set in 2003, and revised in 2006, 2009, and 2012. These are not strictly
speaking quotas, but they determine the allocation of slots across municipalities. In par-
ticular, slots that are not yet assigned to a family, or that reopen whenever a family leaves
the program, are assigned to municipalities based on a priority ratio — the number of PBF
beneficiaries divided by the municipal quota (lowest values first) — and on local demand
defined by the number of eligible families in the municipality that are registered in Cadas-
tro Unico and not yet receiving PBF benefits. Therefore, eligible families can only become
PBF beneficiaries if there are available slots assigned to their municipality.
Section 2 provides more details on the rules used to allocate the municipal quotas as
they generate the variation at the core of the empirical analysis. Specifically, the 2009
PBF expansion was also associated with a change of methodology, which led to large
differences in the additional numbers of PBF beneficiaries across municipalities.
14 After that, the federal government let the number of beneficiaries exceed the cap. The increase in the
total number of slots in 2012 only caught up with the actual number of beneficiaries at the time, which in
2014 was temporarily allowed to exceed the cap by 0.3 million, but has since converged to the 2012 cap.
9
1.2 Data
The analysis in this paper draws on three main administrative data sets.
A. Cadastro Unico. First, we use snapshots of Cadastro Unico, the census of the poor
described in the previous subsection, in December 2008 and August 2010. They include
family-level information on per capita income, family composition, and geographical lo-
cation. Each individual in the family is identified with an ID number that can be matched
to the other two administrative records below.
B. Bolsa Familia payment sheets. Second, we use administrative data from the former
Brazilian Ministry of Social Development (MDS) on the universe of PBF benefits paid. The
data identify the amount received by each family in each month by type of benefit (basic
benefit and variable benefits). We use monthly data from January 2007 to December 2011,
which allows us to calculate, for example, the number of PBF beneficiaries and the total
PBF payments per municipality for each month in our period of analysis.
C. Relação Anual de Informações Sociais (RAIS). Third, we use the Brazilian matched
employee-employer data set from the former Brazilian Ministry of Labor (MTE), which
covers the universe of formal employment spells in each year. For each worker, the data
include information on occupation, wage, race, gender, industry, municipality, as well as
hiring and termination dates. RAIS allows us to calculate municipal and individual formal
labor market outcomes for our period of analysis.15 For instance, we can compute the
number of formal private-sector jobs, overall and by worker category, in each municipality
in a given month. We can also match individuals in RAIS and Cadastro Unico to compute
how many adults in each family are formally employed in a given month.
We use other sources of data to complement these administrative data. The total num-
ber of PBF slots and the municipal quotas were made publicly available by the former
Ministry of Social Development. We use microdata from the 2000 Brazilian census to
compute poverty rates and informality rates for each municipality. Finally, we use munic-
ipal data on estimated population growth, GDP, and taxes paid from the Brazilian Census
Bureau (IBGE).
1.3 Descriptive statistics on PBF beneficiaries
Table 1 displays descriptive statistics for three groups of families using the snapshot of
Cadastro Unico from August 2010: all families, families below the extreme poverty line,
15 Formal workers are those with a signed working card (carteira assinada). Accurate information in RAIS
is required for workers to access the benefits and labor protections afforded by the legal employment system.
Employing firms face fines for failure to report. RAIS omits interns, domestic workers, and other minor
employment categories, along with those without signed work cards, including the self-employed.
10
and families between the poverty line and the extreme poverty line. In 2010, the poverty
line was R$140 per capita per month (∼US$70), and the extreme poverty line was R$70
(∼US$35).
Table 1 first highlights that Cadastro Unico includes many families that are not PBF
beneficiaries. Around 61 percent of all families in Cadastro Unico in August 2010 were PBF
beneficiaries. PBF is not an entitlement, such that families can be eligible without becom-
ing beneficiaries. For instance, 21 percent of families below the extreme poverty line in
Cadastro Unico in August 2010 were not PBF beneficiaries. Comparing the demographic
characteristics of PBF families in Table 1 (discussed below) with the same characteristics
for all families in Cadastro Unico in these income per capita brackets (see Online Appendix
Table A.1) shows that PBF families have lower socioeconomic status on average than fam-
ilies that were eligible but were not beneficiaries.
The top panel in Table 1 also shows that PBF families mostly live in urban areas, have
low high-school completion rates, and have low income levels. In particular, their average
monthly income per capita in Cadastro Unico in August 2010 was R$53.9, and it was only
R$36.2 among the extreme poor. The middle panel in Table 1 shows that the average
monthly benefits received by PBF families (R$101.93; calculated over the following 12-
month period) were substantive compared with their average monthly income (R$206.3).
Extreme poor families have lower income levels and are eligible for both the unconditional
basic benefit and the conditional variable benefits, such that they received higher benefits
than poor families in absolute terms (R$112.45 versus R$68.66) and relative to their income
levels (83 percent versus 17 percent).
The bottom panel in Table 1 sheds some light on PBF families’ participation in the for-
mal labor market, by matching all adult household members in Cadastro Unico in August
2010 to their formal employment spells in RAIS over the following 12-month period. This
exercise reveals that although lower-income families are in general more likely to work
informally in middle-income countries (Perry et al. 2007), a relevant share of PBF families
interact with the formal labor market. Yet, their attachment to the formal labor force re-
mains limited. For instance, 24 percent of extreme poor families and 31 percent of poor
families had at least one adult formally employed for at least one month in the 12-month
window. On average, extreme poor families had a total of 2.45 adult-months in formal
employment over the period; this figure is twice as large among poor families, highlight-
ing the positive correlation between formal employment and income. Finally, conditional
on working formally, adults in PBF families earned more than the minimum wage on av-
erage (R$510 per month in 2010). However, because the income definition for eligibility
is the minimum between the current monthly income and the average income over a 12-
11
month period, working in a formal job does not necessarily disqualify a family from PBF
benefits as long as their per capita income remains below the eligibility threshold.
TABLE 1: PBF BENEFICIARIES AND FORMAL EMPLOYMENT
Note: The table presents summary statistics for families registered in Cadastro Unico in August 2010. We use the information on monthly
per capita income (excluding PBF benefits) to classify families as extreme poor (below R$70) or poor (between R$70 and R$140). The
top panel only uses information from the snapshot of Cadastro Unico, the middle panel matches these families to PBF payment sheets
in the following 12-month period, and the bottom panel matches these families to the formal employment data (RAIS) over the same
period. Here we report statistics for PBF beneficiaries (based on the PBF payment sheets in August 2010); Online Appendix Table A.1
reports the same statistics for all families in each income group. Monetary values are reported in nominal terms.
2 Empirical strategy
To estimate the impact of PBF on local formal labor markets, we exploit the timing of the
2009 PBF expansion and variation in the number of new PBF beneficiaries across munic-
ipalities due to changes in the methodology used to allocate municipal quotas. We lay
out our empirical strategy in this section. First, we provide details on the allocation of
municipal quotas. Second, we present our difference-in-differences (DD) research design.
2.1 Allocation of PBF quotas across municipalities
Each municipality is assigned a quota of the national slots based on poverty estimates
performed by IBGE. These municipal quotas are not strictly speaking quotas, but they
determine the allocation of any PBF slot available at the national level (when new slots
are created or an existing slot is made available after a family leaves the program) across
12
municipalities.16
A. Initial methodology. When the government set the initial number of national slots
in 2003, and when it first revised this number in 2006, it followed a similar strategy to as-
sign the municipal quotas. First, IBGE calculated the number of poor families in each state
based on the poverty threshold17 using microdata from the National Household Sample
Survey (PNAD), a yearly household survey that is representative at the level of the 27
Brazilian states, but not at more disaggregated geographical levels. Specifically, IBGE
used PNAD 2001 and PNAD 2004 for the 2003 and 2006 quotas, respectively. Second, it
assigned quotas across municipalities within a state using microdata from the 2000 Brazil-
ian census, the only source of data representative at the municipal level at the time. IBGE
calculated the number of households in each municipality with total income below twice
the minimum wage. Compared with 2003, it simply updated these figures in 2006 using
an estimate of the population growth in each municipality between 2000 and 2003 (MDS
2012).
Formally, the quotas in 2003 and 2006 were calculated as follows:
2000
P oorms 2001
Quota2003
ms = 2000 · P oors (1)
∑k∈s P oor ks
2000 · n [2000,2003]
P oorms ms
Quota2006
ms = · P oors2004 (2)
[2000,2003]
∑k∈s P oorks · nks
2000
where P oorms2000 is the number of poor families in each municipality m in state s based on
the 2000 census; P oors2001 and P oors2004 are the number of poor families in each state based
[2000,2003]
on PNAD 2001 and PNAD 2004, respectively; and nms is an estimate of the popula-
tion growth in each municipality between 2000 and 2003. By construction, the number of
poor families in each state equals the sum of the quotas across all the municipalities in the
state.
B. Change of methodology in 2009. At the time of the 2009 PBF expansion, the IBGE
changed its methodology to assign municipal quotas.
The first step remained essentially unchanged. IBGE used PNAD 2006 to calculate the
16 The calculations underlying the allocation of municipal quotas are conducted by IBGE using data from
the Census and other household surveys, so there is no clear room for political manipulation in this process.
Brollo et al. (2020) show that there is evidence of political manipulation in other aspects of the program,
i.e., in the enforcement of the school attendance conditionality for existing PBF beneficiaries (to keep their
variable benefits). The quota allocation does not consider information related to the implementation of the
program, so this type of manipulation is not a concern for our analysis.
17 The poverty threshold that was used to compute national slots and state quotas corresponded to half
of the minimum wage per capita.
13
number of poor families in each state. However, as shown in Figure 1, the national number
of slots was increased compared with previous years, and it actually became greater than
the total number of poor families across all states. Concretely, this higher number of slots
was obtained by scaling up the number of poor families by a factor of 1.18 to account for
“transitory poverty” (see Soares 2009 and MDS 2009).18
The second step was fundamentally different. Instead of relying on the number of poor
families in each municipality based on the 2000 census, IBGE adopted a “poverty map”
methodology developed by World Bank researchers (Elbers et al. 2003). The methodology
essentially worked as follows (MDS 2012). IBGE developed a statistical model to predict
the number of poor families in each municipality in the 2000 census based on prior data
on municipal-level variables, as well as household-level variables from PNAD, although
PNAD is not representative at the municipal level. The idea was to capture an overall wel-
fare measure, including information on education, health, and other indicators, to com-
pute poverty shares. It then used the model and data from 2006 (municipal-level variables
2006
and PNAD) to predict the number of poor families in each municipality in 2006, Pd oorms
(IBGE 2019). The municipal quotas were then calculated as before:
2006
Pd
oorms
Quota2009
ms = 2006 · 1.18 · P oors2006 (3)
∑k∈s Pd
oorms
IBGE returned to using census data to construct the municipal quotas in 2012 (based on
the 2010 census). Thus, this poverty map methodology was only used in 2009.
2.2 Research design
In our empirical analysis, we exploit the timing of the 2009 PBF expansion and differ-
ences in the relative size of the expansion across municipalities induced by the change of
methodology described above through a difference-in-differences strategy.
A. Cross-sectional variation in municipal quotas. To capture the variation across mu-
nicipalities induced by this change of methodology, for our research design, we calculate
the difference between the actual 2009 quotas and counterfactual quotas computed as
if the previous methodology had been maintained. Specifically, following the approach
used to update the municipal quotas in 2006, we define the counterfactual quotas as fol-
18 The concept of “transitory poverty” is that more families may regularly fall below the poverty line than
the number of families observed below the poverty line at any single time (Soares 2009).
14
F IGURE 2: VARIATION IN THE SIZE OF THE 2009 EXPANSION ACROSS MUNICIPALITIES
(a) Distribution of the relative difference between (b) Relative change in PBF beneficiaries by reform
ms ) size (above/below the median of ∆Quotams )
actual and counterfactual 2009 quotas (∆Quota2009 2009
Note: The figure presents the variation in the size of the 2009 PBF reform across municipalities induced by the change in the method-
ology used to update municipal quotas. Panel (a) displays the distribution of ∆Quota2009 ms across municipalities, that is, the difference
between their actual 2009 quota and their counterfactual 2009 quota (computed as if the previous methodology had been maintained)
relative to their population in 2006. Panel (b) displays the average of the relative change in the number of PBF beneficiaries for munic-
ipalities in the top 50 percent and bottom 50 percent of the distribution of ∆Quota2009
ms .
lows:
2000 · n [2000,2006]
P oorms ms
CountQuota2009
ms =
[ 2000,2006 ]
· 1.18 · P oors2006 , (4)
∑k∈s P oorms
2000 · n
ms
[2000,2006]
where nms is the estimated population growth between 2000 and 2006 from IBGE.
The state-level constant, 1.18 · P oors2006 , is simply the sum of the 2009 quotas across the
municipalities in each state.19 We then rank the municipalities based on the difference
between the actual and counterfactual quotas relative to their population in 2006:
Quota2009 2009
ms − CountQuotams
∆Quota2009
ms = . (5)
P op2006
ms
This variable captures the relative change in quota in 2009 coming from the change in
methodology, that is, abstracting from any change coming from population growth.
Figure 2a displays the distribution of ∆Quota2009
ms across municipalities, which is more
or less centered around zero. We note that no beneficiary was forced to leave the program
as a result of the update of municipal quotas in 2009, even in municipalities for which
19 Although the methodology used to assign municipal quotas in 2009 is described in several documents
(see references in the text), we did not obtain the statistical model used by IBGE. Therefore, we cannot
2006 2006
recover Pdoorms or exploit idiosyncrasies of that model to isolate exogenous variation in Pd
oorms .
15
quotas were reduced in 2009 (these municipalities became less likely to be allocated new
PBF slots). Figure 2b shows that this variation led to systematic differences in the evo-
lution of the number of PBF beneficiaries across municipalities. It displays the average
changes in the number of PBF beneficiaries in each municipality between 2007 and 2011
relative to the average in the pre-reform period for two groups of municipalities: those in
the top and bottom 50 percent of the distribution of ∆Quota2009ms . The two groups shared
the same trend before the reform, including the 5 percent drop in the number of beneficia-
ries in the first quarter of 2009 highlighted in Figure 1b. The number of PBF beneficiaries
increased rapidly in the following three quarters of 2009 for municipalities in the top 50
percent group, by more than 15 percent. In contrast, the number of PBF beneficiaries only
returned to pre-2009 levels in the bottom 50 percent group. The 2009 reform thus led to
large increases in the number of PBF beneficiaries, and thus in the total amount of PBF
payments, in some municipalities, but not in others.
We present two maps of Brazil to compare the geographical variation in the size of
the program across municipalities prior to the reform and in the relative change in quota
induced by the change in methodology in 2009. Figure 3a displays the number of PBF
beneficiaries as a share of the population in each municipality in 2008. There is a clear
geographical pattern in line with PBF targeting goals: poorer areas in the North and the
Northeast have relatively more PBF beneficiaries. By contrast, Figure 3b shows that there
is no such clear geographical pattern for the distribution of ∆Quota2009ms : in many areas,
contiguous municipalities fall in nonadjacent quartiles of the distribution of ∆Quota2009
ms .
B. Difference-in-differences strategy. In our analysis, we restrict attention to the five-
year period from January 2007 (after the 2006 quota reform; the most recent data used to
compute the 2009 quotas were also from 2006) to December 2011 (before the 2012 quota
reform). We estimate variants of the following difference-in-differences specification:
ym,s,t =αm + θp + φp,s + ∑ βp · 1{t ∈ p} · T reatm,s
p6=2009q1
+∑ ∑ k
γp,k · 1{t ∈ p} · Xm,s + εm,s,t , (6)
k p6=2009q1
where ym,s,t is an outcome of interest in month t for municipality m in state s. Our main
outcomes of interest are the logarithm of the number of PBF beneficiaries, of the total PBF
payments, and of the number of private-sector formal employees. Municipality fixed ef-
fects αm control for time-invariant characteristics of municipalities. We bin time in six pe-
riods, p = {2007, 2008, 2009q1 , 2009q2−q4 , 2010, 2011}, to trace the evolution of the outcome
before the 2009 reform, during its roll-out in 2009, and in the following years. Period
16
F IGURE 3: G EOGRAPHICAL VARIATION IN THE SIZE OF PBF ACROSS MUNICIPALITIES
PRE - REFORM AND IN THE SIZE OF THE 2009 EXPANSION
(a) Number of PBF beneficiaries as a share of the population in 2008
!
(
(
!
(
!
!
( (
(
!
( !
!
!
(
!
(
(
!
( !
!
( !
!
( !
!
(
( !
((
!
(
(
! ®
!
( ( !
(
! (!
(!
! (
!
( ! (
!
( !
(
(
!
!
( (!
! (
( !
! !
( (
(
! !
( !
(
( !
! (
! !
( (! (! (!
((!
(!
! (!
((
!
(
!(
!(
!
!
((
!!(!
(! (
!
(!
(
(!((
!
((!
! (!
( ( !
! (! (
! (!(
(!
(! (
!
!
( ( !
! (! (! (!((!
((
! !
(
(
!
!
(
(
!!
!
(! (
! ((!
! !
(! (!
( (
!(!
! !
(!
!
( ( !
! ( ! ( !
( ((!
! (! (!(! ((!
! (( !
(!
(
!
(! (! (
!
(!(!(
!
( !
( ! ( (!(! (! ! ((!
! (!
! ( !
!
(!
((!
! ( (!
((
!
(
(
! ( !
! ( (!
! ( ! (
!
( !
( (!
! ( ! ( ( !
! (!
! (( ! (! (
(!( !
(
(!
! (
!
( !
(
! !
((!
! (!
(!(
! (! (! (!
(!
(
! ( ! !
(!
(! (! !
(!
(!
(!(!
! ( !
( (!
(! (! (!(! (!
(! (
!
( (
!!(!
(! (! (!
(!(!
( (!
!(
!
( !
( !
!
( ( !
! ( !
( (!
!
!
( ! ( ! ( ! (!( !
(
( !
(( !
! ( ! ( !
! (!
(( (!
! (!
!
((!
(
((
! !
(
(
!
!
( (
!!
(!
(!
! ((!
(!(!
(!(!(!(
(
! !
(!
(!
(
!(!
((
!!
!
(
((!
!
(
( (
!
(
! (
!
!(
(!!
(((
! !
(!(!(!
!(!
(!
( ( !
( (
! !
( ! !
( (
! (
! !
(! (((! (!
(!
((
!(!
( (!
! ((!
! ( (!( (!( (
!!
(
(
!(
!!
(!
( (
(!
!
(!(! (!( !
(!
( !
(!
((!
(! ((
!!
(!
( !
( (!!
(
!
( (
! (
! !
(
(
! (!
! ( ( !
! ( ! (! (!
! (
!! !
!
((
! (
!
!
( !
((
! (!
! (! !
(!
!
(
! !
(
!
(!
(!
(
(
! ! (
(!
! (
!
!
((!
!
(
!
((
!(
!!
(
!
(! !
(! (!
((!
!
(!
(
((
!!
( (
(
!
!
(
!
(
(
!
(
!(
!!
( (
!
( !
! !
( !
( ( !
! ( ! ( ! ( ! ( !
(! ( (
! !
(!
(!
((
!(
!!
(!(!
! (! ( (
(!( !
(!!
(
!
(((!
! (!
! (!
( !
(
(!(!!
(!
( (!
(!
!
(! ! !
((
!
(
!(
!(
!(!
! (!
(!
((
( !
( ( !
( (
! (!
(!
! (!(
!
((
!!
((
! (
(! (! (
! (!
(
! !
( !
( !
(!
(!
( ((! (!
(! ((! (
(! !
( (
!! (
!
!
(!
(!(!
(
(
! !
( !
( !
( (!
! (! (! (
!
(!(!!
( (!
! (!(!
( (! (! ( (
! !
(!( (!
!((
! (!
! (!
! ( !
(!
(
!(! (!
! (
(! (!
! (!
! (((!
! (
!
(! ((
(!
! !
(
(!(!
(!
((
!
(!
! (!
!(!
(
!
((
!
(!
!(
!
( !
( !
((
! (!
(!
(
(
!
(
!(
!
(
!!
( (
! (
!
!
( ! !
( (
! !
(!
(!
( (
!
(
! !
(
!
( !
((
! !
(!
((
!
(
! !
( ( !
(
(
! ! !
( (
! (
! (!
! (!
!(! !
(!((
!! (!
!(! (!
! ((
!(
! !
(
!
(
(
! (
!
(!
! !
(
!
(!
(
(
!
!
(
!
(
!
((!
!
(
!(
(
!
!
( (
!
(!
(
!(!(!
!
( (
(
! !
( !
( !
( (
!(!
!
(( !
!
(! (! (!(!
(
!
(
(
!!
(!
(
!
((
!!
((
!
(!
! ( (!
! (! (! (! (! (( !
(!
(((
!(
!
(
!
!
((
!
(
!
(
!(
!(
!!
((!
! ((!
(
!
(
!
(!
( (
! (
(
!
!
((
!
!
(!
(
(
!
(
!
!
(
!
(
!
(
!
(
(
!(
!!
(
!
(
!
(
(
!
(
!
!
(
!
(
!
( !
(
!
(! ( (!
! (! (
!(
!
!
(
(
!
(
!(
! (
!
!
(
(
!
!
( !
(!
!
( (! ( !
(!( (
! !
(! !
( (! ((
! !
(
(
! !
(!
(
(
!!
(
!
(
(
! (
! (
! !
( !
(!(
(
! !
(
!
(!
(!
(!
(!( (
!
(
!
(
!
(
!
!
(
!
(
(
!
!
(
(
!(
!
!
((
!(
!
(
!
(
!
!
(
!
((
!
(
!
(
!
(
!
!
((
!(
!
(
!(
!!
((
!
(
!
(
!
(
!
!
(
(
!
!
((
!
(
!
!
(
(
!
(
!
(
!
!
(
(
!
(
!
(
!
(
!
(
!
(
!
!
(
(
!
(
!!
(
!
(
!
(
(
!(
!
!
(
!
(!
(
(
!
!
(
!
((
!
!
( !
( ( !
! ( ! (
! !
(
(!
! ((!
! ((!
! (! ( !
(!
(! (
! !
(! (!(
(!
! (
!
( !
(!
(!((!
!( (!
! !
(! (!
( (!
!((
!(
! (
!!
( (
!
(
!!
(
!
(!
((
!
(
!(
!
!
((
!(
!(
!(
!
!
( !
((
!!
(
!
( (
!!
(!
(
(
!(
!
(
!!
(
!
(!
(
!
(!
(
!
((
!
!
(
!
(
(
!!
(
!
((
!
!
(
(
! ( !
! (
(
! !
( !
(
(!
(! (! (
(!
(
! (! (!
! (
(!
! (!(! (! ( (!
! (
!
(! ((
! ((!
!
(
! (! (! (!( (!
!
!
(
(
(
!(
!!
(
!(
(!
(
! !
(
!
((
!
(
!
(
!
(
!
(
!(
!
!
(
(
!
!
(
(
!(
! (!
!
( ((!
! (!
!
(
((!
!
(
!
(
(
!(
!
!
((
!
(
!
(
!
!
( !
(!
(
(
!
(
!
(
!(
!
!
( !
(
(!
!
(
!
(
!(!
((
!
!
(
!
(
(
!!
(
(
!
(
!
!
(!
(
(
!
!
(!
(
(
!
(
!!
(
!
(
!
(
!
((
!
!
(
!
(
!
(
!
((
!
!
(
(
!
(
!(
!!
(
(
!
(
!
!
((
!
(
!
!
(
!
(
!
(
(
(
!
!
(
!
(
!
!
(
!
(
(
!
(
!
!
(
!
(
(
!
(
!
!
(
(
!
(
!
!
(
!
(
(
!
(
!
!
(
!
(
!
(
!
(
!
(
!
(
(
!
!
(
(
!
!
(
(
!
!
( (
! !
( ( !
! (! (( !
! !
( !
(! ((!
! (! (
! (
! !
( (!
! !
(! !
( (
! !
(!
(!
((
!!
(
!
( (
!(
! (
!!
( (!
! ((
!(
!!
(
(!
( !
(
!
( !
((
!
!
( !
(
!
(!( !
!
( !
( (
! (! (! !
(!
(
(!
! (! (!(! (
( !
! (! (!(!(! ((! (!
! ((
(!(!
! ((!
!
(!
!
(!
((
(!
! (
! (!(!
!
( (!(((!
! !(!
(! ((
!
(
!
(
!!
(
!
((!
!
(!
! ( ((!
!
(!
!
(
!
((
!
!
(
(
(
!
(
!!
(!
(!
(
!
((
!(
(!
! (
!(
!
(
!(
!
!
( (!
!
!
((!
(
!
!
((
!
(
!
(
!
(
!
!
(
!
(
!
(
(
!
(
!
!
(
(
!
!
(
(
!
!
(
(
!
(
!!
(
(
!
(
!
!
(
(
!
!
(
(
!
!
(
!
(
!
(
(
!
!
(
!
(
(
!
(
!
( (
! ( !
! ( !
( (!
( (
! !
( ! ( !
( !
( !
( (
! !
( !
(!
( (
! (
! ( (
! !
((
! !
( !
( ( !
( !
( !
( !
( !
( !
( !
(
!
( (
!!
(!
((
!
(
!
(
!(
!!
(
!
(!
(
(
!
(
!(
!(
!
!
(
(
!
!
((
!
(
!!
(
(
!
!
( !
( ( !
! ( ! (! (
!(! ( !
( !
( ! !
( ! (! (! (! (! (! ((! (! (! (!!
( !
(( ! !
(!
(!(
!(
!
(
!(
!!
(
!
((
!(
!!
((
!
(
!
!
(! !
( (
! !
(! (
! (!
! (!(!(
!
(!(! (!
!
(
!
(!
!
(
( ( !
! ((! (!
(! ( (
!(
! !
(! (!
! !
(! ( ! (
! (! ((!
! (! (!
! ( !
(
!
(
(!
(
! (!
!
(!(
!
(
(!
(
!
((!
!
((
!
(
!
(
(
!
!
(
(
!
(
!
(
!
!
(
!
(
(
!
!
(
!
(
!
(
!
(
!
(
!
(
!
(
!
(
(
!
(
!
(
!
(
!
(
!
!
(
!
(
!
(
(
!
!
(
(
!
!
(
!
(
!
(
(
!
(
!
!
(
(
!
(
!
( !
( ( !
( !
( (
! !
( (
! ( !
( !
(( ( (
! !
( (
! !
( (
! !
( (
!!
(!
( !
((
!(
!(
! (
! !
( (
!
!
(
!
( (
!
!
(!
(
!
( (
!
!
(
(!
! ( (!
! ( ! (!
! ( ! ( (!
! !
( ! ( !
!
( ! (!
((!
(!
! (!
((! (!
! ( ! ( ! (!(! (! ( ! (! (!(!(!( (!
! ( (!
! !
(!
!
(((!
!(
!
(
(
!!
(!
(
!
(
!
(
(
!(
!
(
!
(
!
!
(
!
((
!
(
!!
(
!
(
(
!!
(!
(
(
!
(
!
!
(
!
(
!
(
(
!
(
!!
(
(
!
(
!
(
!(
!!
(
!
( ! ( ! ( ( !
( (
! !
(
( !( (
! (
!
(!( !
( !
( !
( !
!
(( (!
! (! ( !
( !
( (
! !
( !
( !
( !
( (
! !
( !
( !
(! ( !
( (!( (
!
(
!(
!!
( (
!
!
(
!
(!
((
!
(
!
(
!
(
!
(
!
!
(!
(
(
!
!
((
!(
!
!
((
!
(
!
!
((
!
(!( ( !
! (!
(!((! !
(!
( ! ( !
( !
( !
(( ( !
! ( ! ( ! ( !
!
( (! (!( ! (!
! (!
! (! (!
! (!
(! (! (! !
((
!(
! (!
! (
!(
!!
((!
!
!
( (
(
!(
!
!
(
!
(!(
(!!
(
!
(!
(
(
!
( !
! (! ( (!(! (!
! (((! (! ( ( ! ((! ( !
( !
( (!!
( !
(!
(!
(((
!(
!
(
!(
!
!
(!
( (
!
(
! (
! !
(
( !
! (! ( ! (
!(!
(! ( !
! ( !
!
( !( ! ( ( !
! ( ( !
( ! ( ! ((! (
! ( ! (! (! ( ! (! (! (! (! (
! !
(! !
(!
((
(
(!
! (! (
(!
! (!(!((
!(
!(
!(!(!
(!
(
(
!!
(!
(
!
(!
(
!
((
!
!
((
!
(
!
(
! (!
! !
( (!(!
(! ( ( ( !
! (! (!
! (!
! !
(! (!
(!
((!
(( ! ( (!
! !
(
(
!(
!!
( (!
(
(!
! !
(!
(
! (
!
(
!
(
!
(
! (
!
(
!
!
(! (
(!
! (
! (
(
!(!
! (!(
!
(
!
!
((
!
(
!
!
(
!
(
(!
!
(((
!
(!
! ( ! (!
(!(
!
( (!
(!
!
( ( (!
! (! ( ! ( !
(!
! (! ( !
(( !
! !
(!(!( ! ( !
!
(( ( !
! (!
! ( (
(!
(
!
(
!
((!
! (!
! (!
((
(
!(
!!
(
(!!
(
(
(
!(
!(
!
!(
(!
((!
!
!
(
!
(
(!
!
(
(
! (
!
( !
((!
! (!
!
((!
(
(!
!
(
(
!
!
(
(!
! ((
(( ! ( ! ( !
! ( (( ! ( !
! (!
! (! ( (( ! ! (( ! (!(! (
! ( (((!
! !
(!(! (
!!
(!
(!
((
!!
(
!
((
!!
(! (!
!
( !
! ( ! (! ! ( !
! (
! ( !
! ( ! (! (!( (!
! (!
! ! ( (
!
(!(!
! (!
! (
(! (! !
(! !
( (! (!
((
(
!
!
(!
(
(
!(
! (
!(
!
!
((
!(
!((
!
!
(!
(
( ! ( !
! ( !
(( ! ( !
! (!
! (! (! (!
( !
( (! ((
( !
!
(
! ( ! (!
( !
( ((!
! (! !
(!
( (! (!
! (! (
(
! (!(! !
( ! (
! !
(!
(
!
(
(
!
(
!
(
!
!
(
!
(
(
!!
((
!
!
(!
(
!
(
(
!
(
!
!
(!
(
!
((
!(
!
(!
! (!
(! ( ( !
( !
( ! ( ! (
(
! (! (! (! (!
! (!
! ( (!
! (!
(! (! (!(!
( (
! (!
(! ! (!
! (!
((
(!
!
(
!(
!
(
(
!(
!
!
(
!
(
!
((
!
!
(!
((
!
(
!
(!
!
!
(
( (
!
! ( !
! ( ! ( ! ( ! (
! !
( (
! (
! (
!((
!
(
! !
(! ( (
! (
! !
( ( !
! (
!
(! (
! !
( (!
! (!
!
( ( (! (!
(
(
!
!
(! (!
!
(
!( (
!
!
(
(
! !
(
(
!
(
!
!
(
!
((
!(
!
!
(!
(
( !
(! !
(! ((! (! (! !
( !
(
(
! (!
! (
! !
(! ( ! (! (! (! (!
(!
!
(!
(! (! (! (! ( !
(!!
( !
((!
( ! (!
(!
((! (! (!
(!
! (!
(!
(!(! (!(!(! (!
! (!
(!
(
(
!
!
((
!!
((
!
!
((!
! (
(
!
!
( ! ( (!
( ! ( ! ( !
! ( ! ( ! (! (!!
( ((
!( !
! ( ! ( ! !
(! (! ((! (!
! (!
! (( (! (!
(! (
! !
( (
! !
(
(
!
!
( ! (!
(!
(( !
! ( ! (( ! ( (!
! ((! (! ( (!( ( ! ( (!(! !
(
(! ( (!
( !
( ! (!
! ( !
( ((
! (
!!
((
! (
!
( ! (( (!
! (( !
! ( (! ( ( ! (! ((! (
! (!
!( (!
! (!
(!
((
( !
! ( !
! (! ((! (!
((
(
! (!
( !
(!
(!(!(! (!(! ( (!
(!
(!
((
!
!
((
!
(!
! ((
!
!
((
!
(
!
(
!
!
(! (!(!
(!
! ((!
(!(! (!(!
(( ! ( ! ((! (! (! (
! (
!
(
!(
!(
!(
!(!
! (!
(! (! (! (! (! ( ! ! (
! !
(
(!(!(! !
(! !
( (!
! (! (!
! (
! (!
(!
(
(!
!
(
(
!
(!
( !
( (
(!
! (!
! !
(
(! ( ! !
(!
! ! ((
! (
!(!
! (
!
(!
(
!
(!
! !
(! ! (!
! ( (! (( ! (! ! (
!
!
((!
! ( (!
! ((!
! (!
(!
(! ((!
! !
(!
(
! ( !
(! (
(!(! ( (
!(!
! (!(!
(! !
(
! !
((! ( ! (!
(!
((! ! (!
! (!
(( !
(! !
( (
! !
(
((
!
( ( (!
(!
! ((!
!( ! ((! (! (! (! !
(! (
! (!
! (!(! ! (
!(!( ( !
( (!
!(! ( (!
! (! (
!
!
( ! (!
! (!
! ( ! (!
((!
((!
! (!
(! ((
(!
!
(!
!
!
(
(!
( ( !
(!
(
!
(!
((
! !
((
(!
! ((
!(
!
!
(!
(
(
!!
(
(
!
(
!(
!(
!
(
!
!
(
!
(
!
((
!
(
!(
!
(
!!
(
!
(!(!
( (!
! ( !
! ((!
! (!
! ((
(!
!
(!
(!(
(!
(
(
!
!
(! (! ((! (
(!
! (
!(!
! ((!
! (
(
! !
(! ( (!
!
!
(
(!
((!
! (! (!
(!
! (
(
( ! ( ( !
! (!
(!
!
( !
! (! !
(!
! ( (
!!
(!!
(
(
!
(
!(
(!(
!(
!
(
!(
!
(
!
(
!
!
(!
(!
(!(!
!
(
(!( ! (
(
! (( !
! (!(!
(
(
!
(!
! ((!
!
(
! (! (!
!
(
(!
!
!
( ((
!
(
(
!(
! !
(! (
!
(
(!
! (!
!
(
(!
(
(
!
(
! !
(
(
!(
! !
(
(
! !
(!
!
((! (
!
( ! ( ( (
! !
( !
( (
! (
!(
! (
! !
(!
( (
! !
( !
((! (! ( (
! (
! !
( !
( !
( (
! !
( !
(
(
!( (! (! (! ( !
( (( !
(
!
(! !
(!(! (!!
((
!(!
! (! ((!
! (!(! (!(! (!
! ( ! !
(! (!
(!
!
(!
!
(!
(! (
(!
!
((!
( !
(
!
(! !
(
(!
! (! (!(! ( !
(
(!
! (! (!
! (
! ! ( ((! !
( !
( (! !
(! ( !
(! (! ( ((
Quartiles
(!
! ( (!
! ( (
! !
( (! (( ((
!!
( !
( (
! !
( !
(! (! !
( (!
( !
!
(! (! (! ( !
(!
(! !
(! ((
!
(!
! ((
!
!
(!
(
!
( (
!(! (! (
(
!(!
! (
!
(
!
(
!
((!
! (!
(( (!
!
(! (! (!
((
! (!
!
( (!
! (! (
(!!(
(! !
!
(!
(
(!
(!
!
( ((
!!
(
(
!
!
((
!(!
!
(
!(
(
!!
((
(
!!
(!
(!
!
(
(! (
(
!!
(((
!!
(
!
(!
(!
(
! (! (
(!
!(!(!(
!
(
!
!
( ( !
!
!
( !
(! (
( !
! (
!
(
! (
(!
! (! !
( (!
! ( (!
! ( (
!!
(
(
!
!
(
(
!
!
(
! (!
! (
!
(
(
!
(!
!
(
(!
((
!
!
((
!!
(
!
(!
(!
(
!
(
! (
!
(!
!(!
(
!(
!
(!
(
!(!
!
(
(
!
( !
(
!
(
(
!!
(
(
!
!
(
!
( (
!
(
!
!!
(
(!
(
(
!
(
!
!
(
(
!
!
(
(
!
!
(
(
!
!
((
!(
!(
!
(
!
!
(!
((
!!
(!
(
(
!
!
(
(
!(
!
!
(
(
!
(
!
!
(
!
( (
!(!
! !
(
!
(
(
!
(!
(!
((
!
(
!
(
!
( !
! ( !
(! (! ( ! (!
! (!
! ((! (!
! (!
! (! !
(! (
!
(
! (
!(
! !
(
(
! (
! !
((!
(! ( !
((
! !
(
(!
!(!
(!
(
(
!!
(!
(!
(
! (!
(! (!!
( (!
! ( (
!
( !
! (! ( !
! ( ! (! (! ( (
! !
(! ((
! (!
! (! (!
(!
(!
( (
! (
!
((
! (!
! (((!
(!
(
(
!(
!!
(!
(
(
!!
(!
((
! !
((
!(
!(
!
(
!(
!!!
(!
((
!!
((
!!
(!((
!
!
(!
(!
(
!
(!
(!
( !
( (
! (!
( (!
! ( (
! !
(
(
! ((
(! (! (
!!
(!
((
!(
!
(
!(
!
(
!!
((
!(
!
!
( (
! (! (
! (!(! !
(
!
(! ((!
(!(
! (!
(
! (! !
(! (!
!(!( !
( (!
!
!
(((
!(
!!
((
! (
!(
!
!
((
!(
!
(
!(
!!
(!
( (
!
(
!
(
! (
!
(
! (
!!
(!
(!
(!
( !
((!
!(!
((
!(
!(
! (!
(! (
!!
(!((!
! ((
!(
!
( !
! (!(!( ! ( !
! (! !
( (!
! (
!
((
!
(
!
(
!!
(
(
!(
!
(
!
!
(
!
(
!
(
!
(!
(!
(
!
(!
(
!
(
!
(
(
!
!
((
!!
(!
(
!
(
!
(
!
(!
(((
!
!
(
!
(
(
!
!
!
((
!((
! (
! (!(!
! (
!
(
(
!
(
(
!!
((
!!
(
(
!
!
( !
((
!!
( ((
! !
((!
! ((!
!(!
(
!
(!
(! (
!
(! (
!(
!
!
( !
(
(
!
!
(
(
!
!
(
(
!!
(!
(
(
!
(
!
!
(
(
! !
(
(
!(
!
(
!!
((
!(
!(
! !
(
(
!(
!!
(
(
!(
!(
!(
!
!
(
(
!
(
!
(
!!
(
!
((
!(
!
!
(
(
!(
!
(
!
!
(
!
((
!!
(
!
(
!
(
!
(
!
(
!
(!
(! ((
!(
!(
!
(
! !
(
!
((
!(!
!
(
(!
(
!
((
!
!
(
((
! (!
!
( ( (
!(
! !
((
! !
( !
( (
! (
!!
( (
! (
!(
! (
! (
! (
!(
! !
( (
! !
( !
(!
( (
! !
((
!(
! (
! !
( (
! !
( !
First
!
( (
!!
( !
(!
( !
( !
( !
( !
( (
!(
! (!((!( !
(! ( !
( !
((
! (
! !
( (
! (
! (
! !
( (
!
( ! (! ( ! ( ! (
(! (! ((
!(!
!(!
(!
(!
(!
(
!
!
( !
(
!
(
(
(!
!
(!
((
!
(
! !
(
!
(!
((
!
!
(
(
!!
(
(
!(
!
(
!!
(
(
! !
(
!
(!
(
(
!
(
!
!
((
!(
!
!
( !
(
(
(! !!
(
(!!(
(!
!
( (
!!
(!
( (
!(
!!
(
(
! !
(
(
!!
(
(
!!
!
( !
(! (!
!
!
((( (
!(
!
(
!
(!!
(
!
(!
(
(
!(
! (
!(
!
(
!(
!(
!(
!!
(
!
(!
(
(
! (
!
(
!
!
(!
((
!!
(
(
!!
(
(
!
!
(!
((
!
(
!
(
!
!
(
!
(
(
!!
(!
(
(
!
!
(!
(!
(
(
!!
((
!
(
!!
((
!
!
(!
(
(!(
!
((!
! (
(!
! (!
(
(
!
!
(!
(!
(
!
(!( ! !
(
(
!(
!
(
!
(
!!
(!
((
! !
(
!
((
!(
!!
( !!
(!
( (
!
(
! !
((
!!
(
!
((
!!
( (
!(
!!
( (
!
(
!!
!
( (
!
(
!(
!(
!
((!
!(!
((
!(
!
!
(!
((
!
(
! (
!(
!
(
!!
(!
(
!
(!
(
(
!!
(!
((
!(
!
(
! !
((
!! (
! (
(
!(
!(
!!
((
!(
! (
!
!
((
! !
( !
(
(!
! (
!
(
!
(!
(!
(
(
!(
!!
(!
(!
(!
(!
(!
(!
(
!
(
(
! (!
!
(
! !
((!
(! (
!
(
! (!
!(! ( ! (!(!
(
!
((
!(
! (
!
(
!(
!
!
(!
(
!
(!!
((
!
!
((
!!
(
(
!(
!(
!(
!(!
! ((
!!
(!
(!!
(
(
! !
( !
(
(
! !
((
!!
(
(
! (
(!
!
(!
( (
(! (!
!( !
(
!
((
!
(
!!
(!
( (!
!( (
!(
!!
(!
(
(
! !
(!
(!
(
!!
(!
(!
(!!
(!
( !
( (
! !
(
(
!(
! (! !
(
!
(!
((
!
!
( !
(!
(!
((
!(
! (
!
(
! ! (!
(! (
!
(! (!(!( ! (
!(!
(! (! (! (! (!(! (!
!
( !
(
(
!!
(!
((
!(
! (
!
!
(
(!
(! ( (
!(
!
!
( !
(
!
( !
(
(
!
(
!(
! !
((
!
(!
(
(
!!
( !
(
(
(! (
!
(
!
(
! !
(!
(!!
( (!
!(!!
(!
(!
(
(
!!
(
!
((
!(!
(
!
(
!
(
(
!
(
!
(
!(
(!
! !
(
!
( !
((
!
!
(
(
!(
!(
! !
(
(
!!
( !
(
!
((
!
((
!
(
!
!
(!
(
!
((
!(
!(!
(
!
(!!
(
(
!
((
!(!
!((
! !
(!(
!
((
!
(
!!
(
!
( (
!
!
((
!!
(!
(
!
((
!!
(!
(
(
!
!
(!
((!
!
!
( (
((!
! (
(
!
(!
! ( !
( !
(! (!
! !(
(! !
(!(!
(
(
!
!
(!
((
!! (!
(! (
!
((
!!
( (
!
(!
! (
!(
!
!
((
!
!
(!
(
!
((
!(
!(
!! (
!(
!(
!!
(
!
((
!(!
! !
(!
( (
!
((!
!(
!
( (
!!
(!
(
! !
!
(
(
! !
(!
!
((
(!(
!
(
!(
!!
(!
(!
(
(
!!
( (
!
!
(
!
(!
(
(
!!
(!
(
(
!!
(
(
!(
!
!
((
!
(
!!
(
(
!
(
! !
(!
(
(
! !
(
(
!(
!(
!!
(
(
!
(
! (
!
(
!(
!!
(
(
! !
((
! (
! (
! (
!!
( (
!(
!(!
!
Second
(!(! (!
(! ( ( (
!(
! (
!!
( (
! (
! !
( ( (
! (
! (
! (
! ((
!(
! !
((
! !
( !
((
! (
! (
! !
( (
! !
( !
( !
( !
((
!
(!
! ( !
!
(! (!(
!
(!
(
(
!
!
(
!!
(
(!
( (
!
(
(
!(
! (
!
(
!
!
(
!
(
(
!
(
!(
!
!
(
!
(
(
!
!
(
(
!
(
!
!
(
(
!
(
!
(
!
(
!
!
((
!!
(
!
(
!
((
!
!
(!!
(
((
!
!
(!
((
!
(
!!
(!
(!
(
(
! !
(
!
(!
(
!
(!
((
!!
(
!
(
(
!
(
!!
(!
(!
(
!
(!
(
!
(!
((
!(
!(
!
!
(!(
!(
!!
(
(
!(
!
!
( !
(
!
(
(
!
!
(
!
((
!
!
((
!
(
!!
((
!
!
(!
(
(
!!
(
(
!(
!(
!!
((
!
!
(
(
!(
!
!
(
(
!
(
!
!
((
!
!
(
(
!
!
((
!
(
! (
!(
!
!
((
!
(
!
!
(
!
(!
(
!(
!
(! !
(
(
!
!(
!
(!
!
(
((!
!
!
( (
! !
(
(!!
(
( !
((
! (
!
(
!!
( (
!!
((
!!
((
!!
((
!(
!(
!!
(!
((
!
(!
! (!
! (!(! (!(!
(!
(
!
( !
(!
(
!
(!
(
(
! (
!
(
!!
((
! !
(!
(
(
! (
!(
! !
((
!!
( !
( (
!(
!
!
( (!
! ( !
((
! !
((
!
!
( !
( ((! !
((
!!
(
!
( !
((
! (
(!
!
(
!
( (
!
(
!(
!!
((
!(
!(
! (
! (
(!
(
! ((!
(
! (! (! !
((
!!
((
!(
!(
!!
((
! !
(
!
(
!
((
!
!
( !
((
!
(
!
(
!(
!
(
!(!
! !
((
!
(!(!
(!
(!
(
(
!!
( (
!!
( (!
!(!
((
!!
(
(
!!
(
(!
((!
! (! (!
(
(!
(!
((!
(
!
(!
((!
!
(
! (
(! !
(!
(
!
((
!
!
((
!
(
!(
!(
!
(
!
!
((
!
!
((!
! (!
(
! (
! (
! (
!!
(
(!
! ( (!
((!
! (
!
(!
((
!
!
(
(
!
(
!
(
!(!
!((
!!
(!
( !
(!(
(!!
(
(
!(
!!
(
!
(
(
!
!
(
!
(
!
(
(
!!
(
(
!
!
(
!
(!
(
!
( !
((! !
(
(! ( (
(!
!
!
( !
(
(!(!
(
(!
!( (!
! (
! !
(!
(!
(!
(!(! (!
! !
((!
! !
( (
!
!
(!
(
(!
((
!
!(
(!!
((!
! ( !
( !
(
(
! (
! !
(
(
(! !!
(
(!
((
!!
((
!(
!(
!!
(!
(
(
!!
(!
((
!!
(
(
! !
(!
((
!
!
((
!!
( !
((!
(!( ! (!(! ( (
! (!
! (!(
(!
!
( ( (!
!
(!
(
((!
!
(!
( !
(
Third
!
( !!
( !
((
!!
( (!((!
( ! ((! !
( (! ( !
(
!!
(
(
! (
!
(!
! (
! (
!
(!
(
!
(!
( !
((!
!( (
!! (! (!
! ! (!
! ( (
!
(! (!
(!
(
! (!(!((
!(!
!
(
!
(
!
(!
(!(
(! (
(! ! !
(!
(!
( ( !
((!
!
!
(
(
(!
(! !
(!(!
!((!
! (!(
(!
!((!
!
(! ( (!(!
(
! (
!
!
(
!
( !
(! (!
(
! (
(!
!
(
! !
((
! !
(!
!(
!
(!
(
! (
(
!
!
(
!
(
!
(
! (
!
(!
!
(!
! ((
! (
!
!
((!
! !
(! (! !
(
(!(!
(! (!
!((!
(! !
(
!
(
!
(
(
!!
(!!
(
((
!(!
! (!
(!
!(!
(
!
(!
( (
!
!
(!
(!
( (!
((
!
(
!(
! (!
!
!
(
(( !
( !
( (!
!(!
((!
! (
!
(!(!
(
!
(
!
(
!
(
(
! !
( !
(
(
!
(
! !
((!
(
!
!(
!
( (
!!
( !
( !
( !
( !
(! !
( !
( !
(!
(!( ! (!
! !
( (
! !
(
(
! !
(
!
(
Fourth
!
((
!
!
( (
!
!
((
! !
((
!!
(
!
(!
(
!
(!
(
(
! !
(
!
( !
( ( !
(!
( (
! (
! !
( !
( ( (
! (
!
(
!(
! !
((
!(
!(
!
!
(
(
!(
!(
!
(
!(
!!
(
!
(
!
(
!
(!
(!
(
(
!!
((
!(
!
(
!
(
!!
((
!
!
(
(
!!
((
!
(
! (
!
!
( !
(!!
( (!
!
!
( !
(!
( !
(! !
(!( !
(!
((
!(
!
!
((
!(
!(
!(
!
!
(!
(
(
!(
!
(
! !
(
!
(!
(!
((
!!
(
!
(
(
!(
!
!
((
!
(
!(
!!
((
!
!
( (
!(
!
(
!!
(!
(!(
!
(!!
(
!
((
!
(!
(!
(!
(
(
!(
!!
( ( (
!
(!
( (
!!
(
!
( !
(!
(!
(!
( !
(
(
!(
!!
((
!(
!
(
!!
(!
((
!(!
!
(
!
!
(
(
!
(
!
(
(
!
(
!
(
!
(
!(
!
(
! !
(
(
!!
(
(
!
(
!(
!
(
!!
(!
( !
(
!
(!
(
(
!
(
!
(
!(
!
(
!!
(!
(
!
((
!
!
(!
((
(
!!
(
!
(
!
(
(
!
!
((
!
!
( (
!
!
(
(
!(!
!
(
! (!((!
! (
(!
! (
!
(
!
((
!
(
!(
!
!
(!
(
(
!
!
(
!
(
(
!
!
((
!(!
!
(
! (
(!
! (
!(
!!
(
!
(
!
(!
(!
(
!
((
!(
!(
!
!
((
!
(
!!
(
(
! !
(!
( (
! !
((
!
!
(!
(!
(
(
! (
!(
! (
!(
!(
! (
!(
! (!
! ( (
! (!
! ( !
( !
( (
!
(!
(
(
!
!
((
!
!
(
!
((
!
(
!
!
(
(
!
(
!
(
!
!
((
!
(
!
!
((
!!
((
!
(
!(!
!((
!
!
(!
((
! (
!
!
(
!
(!
(!
(!
(!
(
!
(
(
!!
(!
((
!
!
(!
(
!
((
!
!
((
!
!
(
!
(
(
!
!
(
!
(
!
( !
(
(!
! (!
! (!
( (!
! (! (
(!
(!
! (!(
!
( !
(!
(
(
! (
!!
((
!(
!
(
!
(
!
(!(!
(!
(
(
!!
((
!!(!
!
(!
( ( !
(
(!
!(
(!
((
! (
!
(
! (
!(
!(!
!((
!(
!(
(!
! !
((
!
(
!!
(!
(!
(
!
(!
( (! (! (
! (
! !
((
!
(
!(
!
!
( (
!
(
!(!
!
(
!
(
! (
!
(! (!
( (
!(
!!(
(! !(
! !
(!!
(
!
(
(!
((
!
!
(!
( (
!
(
(!
! (
!
(!
! (
!
(
!
!
(!
(
(
!(
!
(
!
!
(
(
!(
!
!
(!
(
!
(!
((!
! ( ((
! !
((
!(
!
(
!(
!(
!
(
!!
((
!
!
(
!
(!
( (
!
(!
! !
(!(!
(
!
(!(!
!
( ( !
((
!
(
!!
(
!
(!
(!
(!
( (
!
(
!(
(
!(
!
(
!(
!
(
!(
!
(
!
(
!!
(!
(
(
!
!
(
!
((
!
!
(
(
!!
(
!
(
!
(
(
!
(
!
!
(
!
(
(!
!((!
! (! (!(!
( !
(
!
(
(
!(
!
(
!(
! (
!
( !
(!
!
( !
(!(
(
!(!
!(!
(
!
((!
!
(
!
(!
!
((
!
( !
(
!
(!
(!
(
!
(
!
(!
( !
(
(!
! (!
(
(
!!
((!
(!
!
(
!
(
(
!
!
(!
(!
(
!
((
!
!
(
(
!
!
(
(
!
!
((
!
(
!
!
(!
(!
(
!
(!
(
(
!!
(
!
(
!
(!
(
(
!(
!
!
( (
(!
! !
((
!!
(
!
((
!
(
!
!
(
!
(
(
!
(!
! ( (
!(!
! ( !
(
(!
!
((
!
(
!
(
(
!
(
!
!
((
!
!
(
!
((
!
(
!!
(
!
(!
(
!
((
! (
!!
( (
!
(
!(
!
( !
! ( ! (!
(! (( ! (! !
(!
((!
!
(!
!( ( ! (
(
!
(!
! ( ! (!
! (! (!
(!
(!
!
( ((
!
(!
!( !
(
!
( !
(!(! !
(!( !
( ! (
(!
! (!
(!
(! !
((!
((!
! (!
( !
(
(!
(( !
! (
(
!
(
!
0 125 250 500 750 1,000
Miles
(b) Relative difference between actual and counterfactual 2009 quotas
(∆Quota2009
ms )
!
(
!
(
!
(
!
(
( !
!
(
!
(
!
(
!
(
(
!
(
!
®
(
!
( !
! ( ! (
( !
! ( ! ((
!
( !
! ( !
(
!
(!(!(
( !
! (
(
! (
!
!
(
(
! (
! (!
! (
( !
! (
(
! (
! (!
!
!
( ! (! (! (!
((!
(!
! ( !
((!
((
! (!
!
!
( (!(
!
(
!
( !
(
( !
( (!
! ( ! (! !
(
(
! !
(
(!((!
(!(
(!
(
!!
((!(!
!
( (
! (!
! (! (! (((
!(
! !
(!
(
!
(
(
!!
(
!! (
!((!
! (!
! (!(!
((!
! (!
!
(
! !
( (
! !
( (
! ((!
! (! (!(! ((!
! (( (
!!
(
(!
! (! ((!
! (!(
!
( ( !
! ( (!(! (! ! ((!
! (!
! ( !
!
(! ((!
! ( (!
((
(
!
!
( ( !
! ( (!
! ( ! (
( !
! ( ! (! ( ! ( ( !
! (!
! (( ! (! (
(!
( !
((!
! (!
( !
!
( (
!(!
! (!(!
(
! (! (!(!
(!
(
! !
( ! ( !
! ( ! ( ! (! (!
!(! (! !
((!
!(!(!
! ( !
(!
( (!
(! (!(! (!
(!(!(!
( !
( !
(!
(!(!(! (!
(!(!
( !
(!
(
(
! ( !
!
(
! !
( (
! ( ( ( !
! (!( !
( !
(! ( !
( ! !
(( !
(! (!
!
( ((
!
((
! !
(
(
!
(
! (
!!
(!
! ((!
(! (!
(!
(!
!
(! (
(
! !
(!
(!
(!(
!
( (
!((!
!
(
!
( !
(
!
( (
!(
(!
! (
!((
! !
(!(!(!
! ((!
!
!
(
( (!
! (! ( ! (
!
( ( !
! (
(
!
!
( ! (! (! ( ! !
( ! !
(! (! (((!
! (
!
(!(!
(
!
!(!
(
(
! (!
!
(
!
((!
! ( (!
! ( (
!(
!(
!(
!
(
(
!
(
!!
((
!!
(
(
!(!
!
(!
!
( (! (!
! (
(
!!
((
!
!
((
!
!
(
(
!
(
!
(!
! (! !
(!((
!
( (
!(
!
!(
(!
!
( (!!
(
!
(
(
(
!!
( !
(
(
! !
( !
( (
! ( (
! !
( !
(
!
( (
! (!
( (
! (
! (
!!
!
(!
(! (! (!((!
! ( (
! (
! (!(!
(!
(! !
( (!
!
(
!
!
(
(
(
!(
!(
!
(
! (!
! (
!
(!
! ((
!
!
((
!
!
(
(
!!
(!
(!
(
(!
(
! !
(
(!(
!
( ! ( (
! (
! ( !
! ( ! ( ! (! (!( ! (!((!
!
(!
(!(
!
(!
!
((!
(!
(!
!
( ( (! (
(! (!(!
(
! (!
(((
!
(!
!
(
! (
! (
(!
! (!
(!
(!
(!
((!
(
(! (
(!
(!(!
(!
(
!(!
! (
(
!!
(!(
(
! !
( (
! !
( (!
! (! (! !
(
(
! !
( !
( !
( (
! (
! (!
!
!
( (
!
(!( (!
!(!
( (
! !
(!( !
(!
(!! !
(! (
(! (! ((! (!(! (! ((!(!
(! (!
(
! !
(
(!
( (
!(!
!((
! (
!
!
(!
!
( !
(!( !
(!((
!(
(! !
(! (
!
(
! (
! (( !
! !
((!
! !
(!
( (!
(! (!
(
!
! (
(!
! (!((!
! (!
! ! ! (! (!
!( (!
!
((!
((!
!
(
!
!
((
!
( !
(
!
(
!
(((
(
!
!!
(!(
(!
!
( !
( ( !
!
(!
!
( (
!
(! (!
!
(
!
(
!!
((
!
(
!
!
( !
(
!
(
( !
(
! ( ! ((!
! (!
(!
(
(!
(!(! !
( !
(!
(!
(
(
!!
(
(
!
!
((
!
!
( !
((
(
!(!
!
!
( !
(! (! (
(!
! (!
(
(! !
(!
( (!(!
(!
!
!
(!
(!
( (
!
(
(
!
!
(
(
!(
!
!
(
!
(
(
!(
!(!
!
!
(
(
(
!!
(!
(
!
(!
!(
!
(
!
!
(
(
!
(
(!(
!!
(!
(
(
!
!
(
(
!
!
(
!
(
(
!
(
!
(
!
(
!
(
!
(
!
(
!
(
!
!
(
!
(
!
(!
(
(
!
!
(!
(
!
(
(
!
(
!
!
( !
( !
( ( ( (
! (
! !
( !
( !
( !
( (
! !
( !
( (
! (
! !
( (! (! !
( (
! !
( (
!!
(! (
!!
(!
(
!
((
!
(
!(
!(
!
(
!(
!
!
((
!
!
((
!!
( (
!!
(!
(
!
((
!!
( !
(
(
!
(
!(
!
(
!(
!
!
(
(
!(
!!
(
(
!!
(
!
(
(
!
(
! !
( !
( ! (!
! (
(!
(
!(!(! (!
! ( ! ( ! (! (! (!
! (!(!(! ( (!
! ((!
!(!( !
( ! !
(!!
(! !
(! (!!
(!
(
(
!!
( (
!!
((
!(
! !
(
!
((
!!
(
!
(!
( !
(
(
! (
!
!
((
!!
(
(
!
(
!!
(
(
!!
((
!
!
(
(
!
!
( !
( ( !
! ( ! (((
! !
(
(!
!
!
( ( (
! !
(! (! ( (
! (!
(!
(
! ( (!
! ((!
! (
(
! (
!
!
( (!
!
!
( (!
(!(
(
! !
(
!
( !
(! (!
!
( (
(
!( (
!(!(((
! !
(!
(
!
(
(
!
(
!
(!
((
!!
(
!
(
!
(
(
!
!
(
!
((
!
(
!!
(!
(
(
!(
!
(
!
(
!
(
!
(
! (
!
(
!
(
!!
(
(
!
(
!!
(
(
!
(
!
(
!(
!
(
!
!
(!
(!
(
!
(!
(
(
!
!
(
(
!
(
!
!
(
!
(!
(
!
((
!
!
(
(
!
!
(
!
((
!
!
(
(
!
(
!
!
(
!
(
(
!
!
(
(!( (
! !
( ( !
! !
( (
! !
(
( ! (! (
! (
! (
! !
(
(! ( ! (
! (
! (! !
(
(! (
! !
( (
! (!
! !
(
(!(!(!(
!
(!(
!
((!
!
((
!
(!(
!(
!!
(!
(! (
((!
(!
(!
!
(
(!(!
( !
(!(
!
(!
!(
!
(
(
! !
( (
!
!
(!
((
!(
!
(
(!(
!(!
!
(
!(
(
!(
!(
!!
(
(
!
!
(
!
(!
(!
(
(
!!
(
(
!
!
(!
(
!
((
!
(
!(
!
!
((
!
!
(!
(!
(
(
!
!
((
!
!
(
!
((
!(
!
(
!!
(
(
!
!
(
!
((
!
!
(!
(
!
(
!
!
(! !
( (
!
(
! (
!
!
(
!
( ( ! (! (!
(( !
! (!
(
! !
(! (!
! ( !
(( !
!
( !
(!(! (! ((!
! ( (
! !
(
!
( (
!
(
!
!
( !
(!
(! (
!((
!(
!!
(
(!
!
!
(
(!
! (
! (
(!
! !
(!
(!((
!!
(
(
!
!
(
(
!
!
((!
! (!
!
(
((
!!
!
(!
(
!
((
!
(
!
(
!
(
!
!
(
(
!(
!
!
(!
(
(
!
(
!
!
(
!
((
!
(
!
!
(
!
(
(
!!
(
(
!
!
(!
(
!
(
!
(
(
!
!
(
!
!
(
!
(
(!
(!
((
!(
!
(
(!
!(
!
(
!(
!
!
(
!
(
!
(
(
!
(
!
(
!
!
(
!
(
!
(!
(
(
!
(
!
!
(
!
(
!
(!
(
!
(
!
(
!
(
!
(
(
!
(
!
(
!
!
(
!
(
(
!
(
!
(
!
(
!
( ! !
( !
( ! (
! !
( (
! ( (
! ( (
! (
! !
( !
((
! !
( !
( !
( (
!( (
! ( (
! (
! !
( (
!(
!
!
(!
( (
!!
( !
(!
( !
(
(
!(
!
(
!!
(
!
(
(
! !
(
!
((
!
!
(!
(
!
((
!
( (
(( (!
! ( !
! (! (! (!
!
( (! (!(! (!
! (! (!
! (! ((! ((!(!( (! (!(!(
!
(!
((!
! !
(!
(!
(!
(!
(!
((
!
(
!!
(
(
!(
!
! (
!
!
(
!
( !
( ! ( ! (!
!
(
(!!
(!
(
!
(!
(!(!
(
!(! (
( ! ( !
(
! (! (
!
(!(! (!(! (
!(! (! (! ((
(! !
( (!
! !
(
! (!
! (
(!
! (! (
!(!
! !
(
(!
(!
(
(
!(
!(
(
!!
(
(
!
!
(
(
!
!
(
!
(!
(
(
!
(
!
(
!
(
!
!
(
!
(
(
!
(
!
(
!
!
(
(
!
(
!
(
!
!
(
!
(
!
(
(
!
!
((
!
!
(
!(
(!
!
((
!
(
!
!
(
(
!
!
( (
! !
( (
! ( !
( ! !
( !
( (
! (
! (
! !
( !
( !
(! (
!(
!(
!!
((
!
(
!(
!(
!
(
!!
((
!!
(
!
(
(!
! ( ( !
! (!(!( (!
(
!
(! (!(! ( ( (!
! (! (!
(!((
!(!( ( !
! ( ! (!(!(!(! !
( (
(! (!
((
! !
(
(
! !
(
!
( !
(!
((
!
(
!!
(
(!
! (
!
( ! !
(!( (
! (
! !
(
!
( (!
! ( ( !
! ( (
!
( !
! (!
!
( (!
!
(
(!
(
! !
(! ( ( !
! ( ! (! (! (! ( (!
! (! !
(!( (
! ( (!
! (! (!
!(
(
!(
(!
(
!(
!!
!
((
!
(
!
(
!
(
!
(
!
!
((
!
!
(
(
!
(
!
(
!
(
!
!
(
(
!
!
(
(
!
(
!
(
!
!
(
(
!
(
!
(
!
(
!
(
!
!
(
!
(
(
!
(
!
!
(
(
!
!
(
(
!
!
(
(
!
!
(
(
!
!
(
!
(
!
(
(
!
(
!
(
!
(
!
!
(
(
!
!
(
(
!
!
( (
!
!
( !
(! !
( ! ( ! ( (
! (
! (!
! (! (( ! ( (! (! (
! !
(! !
((
(!(
!!
(
(!
! (
(!(
!
( !
(!
(
( !
!
(!
( !((!(!
! (!
(!
( ! ( ! ( ! ( (!(!(! ( !
! ( ! ( !
! ( ! ( ! (!
(!
!
(( ! (!(!
!
(!(( ! ( ! (! (!(! !
(! ((!
(
! ( !
! (! !
(! (!
! !
(! !
(!
(!
( (
!
(!!
(
!
((
!
!
(!
((
!
!
(
(
(
!
!
(
!
(
!
(
!
!
(
(
!
(
!
!
(
(
!
!(
(!(
!
!
(
!
((
!
(
!
( !
! (! ( !
! ( ( !
( !
( !
( ( !
( !
( (
!!
( !
( (
! (
! (
! !
( ( ( (
! !
( !
( !
(
!
( !
( !
( !
( !
( (
!
(!
((
!(
!(
!
!
((
!(
!
((
!(
!
!
(
!
(!
(
!
(
(
!
!
((
!(
!
(
!!
(
!
(
(
!(
! !
(
( (
!(!
(! !
( ! ( ! ( !( ! ( ( ( !
!
( ! ( ! ((! (( ( !
! (! ( ! (! !
(! !
( (!
( (!
! (!
! (! !
(! !
( (
!(
!
(
!(
!
!
( (!
! !
( (!(!
(! ( ( ( !
! (! (!
! (!
! (!
! (!
(! (( !
((! (
( !
(!
! (
(
!(
! !
( ((
!
!
(! !
(! (! (
!
!
(
(
!
!
( !
(
!
(
(! (
(!
! (
! (
!
( ((
!
(!
!
!
(
(
(
!(
!
!
(
(
!
(
!(
!
(!
(
!(
!( !
(! (!
(!(
!
( !
( ( (!
! (! ( ! ( !
!
(! (! ( !
(( !
! (!
! (!( ! ( !
!
(( ( !
! (!
! ( (
!!
( (!
! (!
! (!
((
(
!!
(!
( (
!
(
!!
(!
(
(! (((
!(!
!
(!
!(
(
! ( !
((!
! (!
!
(((
! (
!!
(
!
!(!
(!(!((
(( ! ( ! ( !
! ( (( ! ( !
! (!
! (! ( (( ! ! (( ! (!
(
(! (! (
!
( ( (((!
! !
(!((!
! ( ! (
! !
(
!
((
!!
(!
(
(
!(
!(
!
!
((!
! (
!(!
! (
( !
! ( ! (! ! ( !
! (
! !
( ! ( ! (! (!( (!
! !
(! ! ( !
(
(!(!
! (!
! (
(! (! !
(! !
( (! (!
((
(
!
(
!!
(
!
(!
( !
(!
(!
(
!
(!
( (!
!
(
!
((
( ! ( !
! ( !
(( ! !
( ! (!
! (! (!(!
( !
( (! (
!
((
( !
! ( ! (!
( !
(! ((! (! (!
!
( ( (!
! (! (!
( (! (! (
! ! !
( (
!!
(!
(
(
!
(
!
!
(
!
(
(
!
!
(
(
!!
(!
(
(
!!
(
!
(
!
(
(
!
!
((
!(
!
!
((
!
!
(!(!
(! ( ( !
( !
( ! ( ! (
(
! (! (! (! (!
! (!
! ( (!
! (!
(! (! (!(!( (
! !
((!
! ! !
(! (! ( !
(!
!
( !
(
(
(
!!
(
(
!
(
!
(
!
!
(!
((
! !
(
!
(
!
!
(
!
(
(!
!
(
! !
( ! ( ! ( ! ( ! (! (! ( (
! (!
! (! ( !
! ( ( !
(! !
( (!
(!
! (( (! (!(!
(! (!(
(
!(
! !
(
(
!!
(!
( !
(!
(
(!((
(
! ( ! ( (!
! (!
! (
! (!((!
! (!
!
(
(! (! (!
(!(!(!!
( ! ( ! (!
! (!
(!
! (!
(!(! (!! (!
! (!(
!
(! (! ( !
(
!
((
!
(
!(
!
!
((
!
!
(!
(
!
(! (
!
!
(
!
( ( !
! (
!
(( !
( !
( !
( ( !
( (
! (
!(
! !
( ! ( !(
(!(
! (
! (
! ( !
( !
(
!
( !
( ( (
! (
! !
( (
! (
! (
! !
( (
!(
!!
( !
(!
( (
!
!
(
!
( ! (!
! (!( ! (
! (!
! ( ! (! (
(! ( ! ( ! ( ((! (! (! !
(! !
(!
(! (! (! (! (! (!
! (! !
( (
!
(!
(!
( ! ( !
( (! (!
! ((! (! ( ( ( ! ( (!(! !
(! ( ( ( ! (
(!( (
!!
(!
((
!!
( (
!
( !
! (( (!
(! !
(! (!
(! (( !
! ( ( ! (! ((! (
!!
( !
(!
!
((!
(!
!
(!
(!
(!
(!
((
!
( ! ( !
! (!
( !
( (! (!
! (!(
!
( (!
(
(! (!
!(!(! (! (!(! (
(!
(!
(
(
!(!
!
(
! ((
!
(
!!
(
(
!
(
!
(
!(
!
(!
! (!
!
( (
(!
!(!
!
(
(! (! (( !
(! !
( ( !
! ((!
! (! (! !
((
!!
(!
(
(!((
!(
! (!
! (! (! (! ( (! (! (
!(!( !
(! (
!(!(! !
(!((!
! (! (! ( ! (
! (
! !(
(! (
(!(
!
(!
! ( ! !
(((!
! ( ! ( ! (! ( ((!
! !
(!
(! ((
! !
(
!
!
(!
(
!(
(!
!
(
!(! (! ( !
(!
!
(!
(! (!
!
( (!
( !
( ! (!
! (! !
(! (
!!
(!
(!
! (! ((!
!
(!
!(! (!
(! (!
! (
!
(!
(
! ((
!!
(!(!(! (
(!
! (!(!
!
(
! (!((!
! !
(
(! ( ! (! !
((! (! (! (!
(! !
(!
(! ((
! !
( (!
!
(
!(
(!( (
!(
!(
! (
!
!
(!
(!
(!
((!
!
(
( ( ((!
! (!
((! (!
! (
(
(!
(!( (
! (!(! !
((
! (
!( (!
! (
! (!
! (( (
(!
(!
! (
(!(! (
!
( ! ( ! (
! (!
! !
( ! (! ((! !
( ! (!(! ((!
!
(
! (!(!
(! (
!
( ! !
((
!(
!
!
(
!
(
!
(!
((
!
!
(
!
(
!
(
!
(
(
!!
(
!
((
!
(
!! !
(
(!(!
!
(! ( ! !
(!
(( !
!
(! (!
( (!
! (!
(! (!
! ((!
! ( ( (!( ( !
!
((
!!
(! (!
! (
(! (! ((
!!
(( (
!(
!(
!(
!!
( ((
! !
(( ( !
!
(! (
((!(!
! (
(
!(!
! (! ((!
( (
(!(
!!
((!
! (! !
(
(
!
(! !
(!
(
(!
!
(
( !
((
! ! !
(! (
( !
! (! (! (! ( ! (! ( (!
! (
! !
(!(!
((
!(!
! (! (!
!
(! (!(! (! ( !
!
((! (! (!
! (! ((!
! (! (!
! !
( (!
! !
( (!
!
(!
((! (
! (!
!
(
!
(!
((! (
!
(!
(! (!
! ( !
( ! (!
! !
(!
( (! (!
! (
!
(
!
(! (! !
((
!! (
! (!( (!
! (
! (
(
! (
! (
! (
! (
!(!
!
(
(
!
(
(
!
!
( !
(
(!
( !
(
!
(!
! (
(
! (
! (!
! (!
( ( !
! ( !
! ( ! (! (!
! ((!
! (!( !
( (!
! (
! (!
! (! (! (! (! ( !
((
!(
!!
( (
!!
((
!(!
! ( (
!!
(! (!
! (!
! (! (!(
!(!
(
(!
!
( !
( (!
! ((! ( !
( !
((!
! (!
( (!(!
!
( (
!
(!
((
!
(!
((!
!
( (! (
!
((!
! (!
(
(
!(
!
(!
(!
(
! (!
!
((!
( !
( (
!((
! !
(
!
!
( !
(!
(
(!
!
!
(
(!(!
(
(
!!(
(!!
(
!
(!
(
!!
(!
( (
(
!
(
!(
!
(
!
!
(
(
!(
!
!
(!
(
!
((
!
(
!!
(
!
( (!
!
!
(
( (
(
!
(
!
(
!
((
!
!
(!
(
!
((
!
(
!
!
(
!
(!
( (
!(
!(! (
!
!
(
!
( (
! (
! (
! (
! (
! (
!(
! !
( !
( (
! (
!
(
!
(! (! (
(!
! ( ! ( ! (! (! (! ( !
!
(! (! (!(!
!
( (!
((!
(
(! (!(!
( (
!(
! !
(! (! !
((
!!
(!!
(!
(!(!
! ( !
(
(
! (!
!
(
! (
!
((
! !
(
(
!!
((
!
!
((
!(
!
!
(!
((
!(!
! ( (
!
!
((!
!(!
((!
! (
!(
(!(!
!(
Quartiles
! (! (! !
(! !
(! ( ! (! (! !
(! (!
! !
(!
(
!(
! !
(! !
(! !
((
!
( (
!!
(!
(!
((
!
( ( (!
(
!(!(!
(!!
(
(
!!
( !
(
!
((
!!
((
!(
! !(!
(! (! !
((
! !(
(! !
(
( !
! (! ( !
! ( ! ( ( !
(! ((
! !
(! (! ( !
( ((!
! (!(( ((
!(
!(
!(
!!
(!
((
! !
(!
((
!(
!(
!(
!!!
(!
(!
(!
((
! (!
(
!
( (
!(
!
(
!!
(!
( (
! (
! (
! ( (!
! ( !
( (
!
!
( ((
(! (! (
!(
!(
!(
!
(
!
!
(!
(
(
!!
((
!!
(
!
( (
! (! !
( !
(! (
!(!
! (( !
(! !
((!
(
! (! (!
! (!
! (!( !
( (
! !
(!
(!
(!
(!
((
! (
!
(
!!
(!
((
!!
(!
((!
(
!
!( (
!
!
(
(
! (
!
(
! (
!(
!
(
! (
! (
!!
(!
((
!!
( (
!!
( (!
(! (
!
!
( !
((!
! ((
!(
!(
!
( !
! (!(!( ! ( !
! (! (
! !
(!(
(
!(
!
!
(
(
!(
!
(
!!
(
!
(
(
!
!
(
!
(
(
!
!
((
!
(
!(
!
(
!
!
(
!
(
!
(!
(!
(!
(
!
!
(
!
(
(!
((
(
!
(
!
(
!!
(
(
!
(
!(
!((
! (
! (! (!
!(
(
(
!
!
(!
(
!
((
! (
!
!
(
(
!(
!(
!!
(((
! (!
!
(
! ((!
! (
!
(
!
(!
(! (
(!
! ( !
(
(
! (
!
(
!
!
(
(
!!
((
!
!
(
(
!
!
(!
(
!
(
!
( (
!
(
!!
(
!
(!
(!
((
!!
( (
!
!
((
!!
(
!
((
!(
!(
!
(
!
(
!
!
(
(
!
!
(
!
( !
(
!
((
!
!
((
!
!
(
(
!
!
(
(
! !
(
!
(!
(
(
!!
((
!! ((
! !
(
!
(!
( !
(
(
!(
!(!
(
!
(!
(
(
!!
(
( ! ( (
!(!
((!
!
(!
( (!
(
!
!(!
((
!(
!
!
((
!(
!(
!(
!
!
(!
(
!
((
!
(
!!
(
!
(!
(
!
((
!
!
( (
!
!
(!
(
!
(!
((
!!
(!
(!
(!
( (
!!
( (
!
(!(
(
!(
!!
(!
( (
!(
!
(
!(
!
!
((
!
(
!
!
((
! (!
!( (!
! (
!
(! !
(!
(!
((
!!
((
!!
(
(
! (
! !
(
!
((
!!
( !
(
(
! (
!
(!
!
(
(
!
(!
((
!
(
!(
!
(
!!
((
!
!
((
!
!
(!
(
!
((
!!
(
(
!(
!
(
!!
(
!
(!
(!
(
(
! !
(
(
!
!
( (
!
!
(!
(
(
!!
((
!(
!
(!( (
(!
!!
(
!
(
!
(
(
!!
((
!!
( !
! (! ( ! (!
! (! (
!
((
!
(
!
!
(! ((
!(
!!
(
(
!!
((
! !(
!
(!
(
(
!!
(!
((
!
!
( (
!
(
!!
(!
(
!
(
!
((
!
(
!(
!(
!!
(
(
!
(
!!
( !
(
!
(
!
(!
(
!
((
!!
(
(
!
!
((!
!((!
!
(
! (!
(!(
!
(!
(!
(
!
( !
(
!
(!
(
!
((
!
!
((
!
(
!(
!!
(
!
(
(
!(
!
!
(
(
!!
(
!
((
!(
!
(
! !
(!
(((
! (
!
!
((!
! (!
((
!!
(
!
(
!
(
!
((
!(
!(
! !
(
(
!(
!(
!! !
(
!
(
(!
!
!
(
!
(
!
((
!!
(
(
!!
((
!!
(
!
(
(
!
!
(
!
((
!
(
!
!
(!
((
!
!
((
!
!
(
(
!!
(
!
(!
!
(
(
!
(!
!
( (! !
( !
(
(
!
(
!
(!( !
(
(
! (
!!
(!
( !
(!
( (
! !
( !
( !
( !
( ! !
( !
( !
( !
( (
!!
( (
! !
( !
(!(
(! !
( !
( (
! (
! !
( (!(!
First
(!
!(! ( ! (!
! (!
! (
!
((
!
!
(!
((
! (
!
!
(!
(
(
!
(
!!
(
!
(!(
!
!
(!
(
(
!!
(
!
((
!(
!(
!(!
! ((
!
!
( !
(!(
!
(
! (
! (
!(
! (
!!
(
(
!!
( (
(!
!
( !
(
(!((
!(!
!
!
(( !
(
!
((
!
(
!
!
((
!!
((!
! ((
!!
(
(
!(
!!
(
(
! !
(
(
!
(
! !
(
(
! !
((
!
(
!!
(!(
!!
( !
( !
( !
(
!
((
! (!
( (
!
(
!
(
! !
(
(
!!
((
!!
(!
(!
((
!!
(!
( ! ((!
! (
!
(
!
(! (!(!( ! (
(!
!(!
!
(
!
(
(!
!
(
(!
!
(
(!
!
( (! (
(!
(
(
!
!
(!
(
(
!!
(
(
!
!
(!
(
!
((!
!(
(!
(!
!
(
(!
(
!
(!
(
!
(
(
!
(
!(
!
!
(
!
(
!
( (
!
!
(
!
(
(
!
(
!
!
(
!
(
(
!
(
!
!
(
(
!
(
!
(
!
(
!
(
(
!
!
(
(
!
(
(!
! !
(
!
(
(
!(
!
(
!(
!
!
(
!
(
!
(
!
!
(
(
!
!
(
(!
!
(!
!
(!
(
!
!
((
!
!
(!
(!
(!
(
(
! (!
!
(
!
(
!
!
(
(
!
(
!
(
(
!
(
!
!
(!
!
(
(
!
(
(!
!
(!
(
(
!
!
(
!
((
!
!
(
(
!
(
!
!
((
!
(
!
!
(
(
!
(
!
(
!
(
!
(
!
!
(!
(
(
!
!
(
!
(
!
( !
(
(
!
(
!
(
!
(
!(
!
(
!
(
!
(
!
!
((
!(
(
!(!
!
(
!
!
(
((
!
!
(
!
(
(
!
!
!(
(!
( !
(
!
(
!
(
!
(!
!
(
!
(
!
(
(
(!
!
(!
(
(
!
(
!!
(
(
!
!
(
(
!!
(
(
! (
!(
!
!
(
(
!
(
!
!
(
!
((
!(!
!
(
!(
!
(
(
!
!
(!(! (
!
( (!
(! ( (
!
!
(! !
( (
!
!
( (
!(
!
!
( !
( !
(!
(! (
! !
( (!
! (
! (
!(
!!
((
! !
( !
(
!
(!
( !
( ((
! !
(!
( !
(
!
(
!
((
! (
!
(
!(
!(
!!
(
(
! (
! (
!
(
(!
!! (
!!
(!
((
! !
( !
((
!
!
(!
(!
(!( (
!!
((
!(
! (
!
(
!(
!(
!(!
! (!
( !
( (
! !
(
(!
(!
! (! !
(! (!(
!
(
!
(
!
(
!
(
(
!
!
((
!!
(
(!
!
(( !
(
!
(
(
!
!
(
!
(
!
(!
(
(
!
(
!
!
(
(
!
(
!
!
(
!
(
(
!
(
(
!
!
!
(
(
!!
((
!
!
(
(
!!
(
(!
! (
!
((
(
!(
!(
!
(
!!
(
(
!!
(
!
( (
!
!
((
!(
(
!(
!
!
(!
(
(
!
!
(!
!
(!
((
!
(!
(!
(
(
! !
(
(
!
!
(!((
(!
!
!
(
(
(!(
!!
(
!
(!
(!
((
!!
(
(
!(
!
!
(!
(
(
!!
(!
(
!
((
!
(
!
!
(
(
! (
!
!
((
!
(
!
!
(!
(!
(
(
!
!
(
!
(
!
(!
(
!
(
(
!!
(
!
(
!
(
!
(
[Text truncated at 2 MB: about 20,174 of 31,935 words are shown here. The complete text is in the original document held by the archive.]File and source
- File
- cash-transfers-and-formal-labor-markets-evidence-from-brazil.pdf
- Size
- 9,103,582 bytes
- SHA-256
- 808b46f071f88118eec34cd758cbc064de18e9de6e98684577a31028cf4ef191
- Original
- doi.org