Trade Liberalization and Regional Dynamics · 2017-08-25 · Trade Liberalization and Regional...

Trade Liberalization and Regional Dynamics∗

Rafael Dix-Carneiro†

Duke Universityand BREAD

Brian K. Kovak‡

Carnegie Mellon Universityand NBER

February 2017

Abstract

We study the evolution of trade liberalization’s effects on local labor markets, followingBrazil’s early 1990s trade liberalization. Regions that initially specialized in industries fac-ing larger tariff cuts experienced prolonged declines in formal sector employment and earningsrelative to other regions. The impact of tariff changes on regional earnings 20 years after lib-eralization was three times the size of the effect 10 years after liberalization. These findingsare robust to a variety of alternative specifications and to controlling for a wide array of post-liberalization shocks. The pattern of increasing effects on regional earnings is not consistent withconventional spatial equilibrium models, which predict that effect magnitudes decline over timedue to spatial arbitrage. We investigate potential mechanisms, finding empirical support fora mechanism involving imperfect interregional labor mobility and dynamics in labor demand,driven by slow capital adjustment and agglomeration economies. This mechanism graduallyamplifies the initial labor demand shock resulting from liberalization. We show that the mech-anism explains the slow adjustment path of regional earnings and quantitatively accounts forthe magnitude of the long-run effects.

∗This project was supported by an Early Career Research Grant from the W.E. Upjohn Institute for EmploymentResearch. The authors would like to thank Peter Arcidiacono, Penny Goldberg, Gustavo Gonzaga, Walker Hanlon,Guilherme Hirata, Joe Hotz, Joan Monras, Enrico Moretti, Nina Pavcnik, Mine Senses, Juan Carlos Suarez Serrato,Lowell Taylor, Gabriel Ulyssea, Eric Verhoogen, and participants at various conferences and seminars for helpfulcomments. Ekaterina Roshchina provided excellent research assistance. Dix-Carneiro thanks Daniel Lederman andthe Office of the Chief Economist for Latin America and the Caribbean at the World Bank for warmly hosting himwhile part of the paper was written. Remaining errors are our own.†[email protected]‡[email protected]

1

Trade Liberalization and Regional Dynamics Dix-Carneiro and Kovak

1 Introduction

Prominent theories of international trade typically focus on long-run equilibria in which the re-

allocation of resources across economic activities is achieved without frictions. These models have

traditionally given little attention to the adjustment process in transitioning from one equilibrium

to another, creating a tension between academic economists advocating trade liberalization and

policy makers concerned with the labor market outcomes of workers employed in contracting sec-

tors or firms (Salem and Benedetto 2013, Hollweg, Lederman, Rojas and Ruppert Bulmer 2014).

While theory tends to focus on long-run outcomes, empirical studies of the labor market effects

of trade liberalization typically emphasize short- or medium-run effects. Frequently changing de-

signs of cross-sectional household surveys forced researchers to focus on relatively short intervals

to guarantee consistency over the periods analyzed (Goldberg and Pavcnik 2007). Thus, although

many countries underwent major trade liberalization episodes throughout the 1980s and 1990s (e.g.

Brazil, Mexico, and India, among others), we still know very little about the evolution of the effects

of these policy reforms on labor markets.

We fill this gap in the literature by using 25 years of administrative employment data from Brazil

to study the dynamics of local labor market adjustment following the country’s trade liberalization

in the early 1990s. We exploit variation in the tariff declines across industries and variation in

the industry mix of local employment across Brazilian regions to measure changes in local labor

demand induced by liberalization. We then compare formal employment and earnings growth

between regions facing larger and smaller tariff declines, while controlling for pre-existing trends in

these outcomes.1 This approach allows us to observe the ensuing regional labor market dynamics

for 20 years following the beginning of liberalization.

The results are striking. We find large and steadily increasing effects on regional earnings and

employment. Regions facing larger tariff declines experience deteriorating formal labor market

outcomes compared to other regions. These effects grow for more than a decade before beginning

to level off in the late 2000s. This pattern is robust to a wide variety of alternative measurement

strategies, weighting schemes, and controls for pre-existing trends across multiple decades. The

growing effects are not driven by post-liberalization shocks such as later tariff changes, exchange

rate movements, privatization, or the commodity price boom of the 2000s. We conclude that

liberalization’s effects on regional earnings and employment grew substantially over time.

This pattern challenges the conventional wisdom that labor mobility gradually arbitrages away

spatial differences in local labor market outcomes (Blanchard and Katz 1992, Bound and Holzer

2000). If that were the case, one would observe declining regional effects of liberalization on

1In this paper, we focus on formal labor market outcomes, covering workers with a signed work card providingaccess to the benefits and labor protections afforded by the legal employment system. See Dix-Carneiro and Kovak(2015b) for an analysis covering the informal labor market, which includes the self-employed and employees withoutsigned work cards.

2


earnings, such that the short- and medium-run estimates of trade exposure in prior work would be

an upper bound on the long-run effects.2 Instead, we document increasing effects of liberalization;

the effect on regional earnings 20 years after the start of liberalization is more than 3 times larger

than the effect after 10 years. Liberalization’s long-run effects on regional labor market outcomes

are therefore much larger than initially supposed.

This surprising finding leads us to evaluate a variety of alternative mechanisms that might

account for the growth in liberalization’s effects on regional earnings. The evidence rules out mech-

anisms based on slow urban decline (as in Glaeser and Gyourko 2005), changing worker composition

(based on observable or unobservable characteristics), and slow responses of trade quantities to tariff

changes. Instead, we find strong evidence for a mechanism involving imperfect interregional labor

mobility and dynamics in labor demand, driven by a combination of slow regional capital adjust-

ment and agglomeration economies. Intuitively, as capital slowly reallocates away from harder-hit

regions, workers’ marginal products steadily fall. Similarly, with agglomeration economies, a nega-

tive local labor demand shock decreases local economic activity, reducing regional productivity, and

further decreasing the marginal product of labor. We find minimal responses of regional working-

age population to regional tariff declines, suggesting imperfect worker mobility across regions. In

this setting, dynamic labor demand, driven by slow capital adjustment or agglomeration economies,

can rationalize the steady relative decline in wages in regions facing larger tariff declines.

We present a wide array of evidence in support of this mechanism. Regions facing larger tariff

reductions experience steady declines in the number of formal establishments and declining average

establishment size, suggesting that capital stocks slowly reallocate away from negatively affected

regions. Capital investment shifts away from these regions on impact, with immediate declines in

establishment entry and job creation. In contrast, establishment exit and job destruction increase

slowly over time, consistent with firm owners waiting for installed capital to depreciate before

contracting or closing down regional establishments. Supporting the presence of agglomeration

economies, we show that employment in a given industry × region pair falls more when other

industries in the region face larger tariff cuts. Regional labor market equilibrium would suggest

the opposite in the absence of agglomeration economies (Helm 2016). Finally, we extend the

specific-factors model of regional economies in Kovak (2013) to incorporate slow factor adjustment

and agglomeration economies. Within this framework, we show that a proxy for regional capital

adjustment quantitatively accounts for a substantial portion of the long-run earnings effects that

we observe. Standard magnitude agglomeration economies and perfect long-run capital mobility

quantitatively account for all of the long-run earnings effects. In contrast to the other alternative

mechanisms that we considered, this dynamic labor demand mechanism is both qualitatively and

2Papers documenting short- and medium-run regional effects of trade exposure include Autor, Dorn and Hanson(2013), Costa, Garred and Pessoa (2016), Edmonds, Pavcnik and Topalova (2010), Hakobyan and McLaren (forth-coming), Hasan, Mitra and Ural (2006), Hasan, Mitra, Ranjan and Ahsan (2012), Kondo (2014), Kovak (2013),McCaig (2011), Topalova (2010), and many others.

3


quantitatively consistent with the observed earnings responses.

Only recently have researchers begun measuring reallocation costs and the dynamics of labor

market adjustment following trade policy reforms. The papers in this literature calibrate or esti-

mate small open economy models in order to study their quantitative implications for welfare and

the implied transitional dynamics when facing hypothetical changes in trade policy.3 We contribute

to this literature by describing empirical transitional dynamics in response to a real-world trade

liberalization. We document the importance of dynamic labor demand in the evolution of liberal-

ization’s effects on labor markets and suggest that incorporating this mechanism into quantitative

models is an important task for future work.

A growing empirical literature finds substantial differences in the effects of trade exposure across

local labor markets with different industry structures.4 Each of these papers measures the effects of

trade shocks over a fixed time window of 7 to 10 years. We contribute to this literature by placing

the single-year estimates from prior work into a dynamic context, documenting the evolution of

trade liberalization’s regional effects over time. This exercise is possible because our data provide

complete yearly coverage of the formal labor market, even at fine geographic levels, and because

Brazilian liberalization represents a discrete shock occurring during a well-defined time period. A

similar analysis would be much more challenging when studying shocks that continually evolve over

time, such as Chinese export growth, because it is difficult to separate the influence of dynamics

from the effects of newly arriving shocks.5

Our paper proceeds as follows. Section 2 describes the history and institutional context of

Brazil’s early 1990s trade liberalization. Section 3 describes the data sources, local labor market

definition, and empirical approach. Section 4 presents i) our main results for liberalization’s effects

on regional earnings and employment, ii) a wide array of robustness tests, and iii) analyses ruling

out the influence of post-liberalization shocks. Section 5 evaluates potential mechanisms that could

account for the growing earnings effects of liberalization. Section 6 concludes.

2 Trade Liberalization in Brazil

Brazil’s trade liberalization in the early 1990s provides an excellent setting in which to study the

labor market effects of changes in trade policy. The unilateral trade liberalization involved large

declines in average trade barriers and featured substantial variation in tariff cuts across industries.

Many papers have examined the labor market effects of trade liberalization in the Brazilian context

3Examples include Artuc, Chaudhuri and McLaren (2010), Caliendo, Dvorkin and Parro (2015), Cosar (2013),Dix-Carneiro (2014), Kambourov (2009), Traiberman (2016), and many others.

4See footnote 2 for citations.5Autor, Dorn, Hanson and Song (2014) discuss this point in their study of the effects of Chinese export growth

across U.S. industries.

4


to take advantage of this variation.6

In the late 1980s and early 1990s, Brazil ended nearly one hundred years of extremely high

trade barriers imposed as part of an import substituting industrialization policy.7 In 1987, nominal

tariffs were high, but the degree of protection actually experienced by a given industry often

deviated substantially from the nominal tariff rate due to i) a variety of non-tariff barriers such

as suspended import licenses for many goods and ii) a system of “special customs regimes” that

lowered or removed tariffs for many transactions (Kume, Piani and de Souza 2003).8 In 1988 and

1989, in an effort to increase transparency in trade policy, the government reduced tariff redundancy

by cutting nominal tariffs and eliminating certain special regimes and trade-related taxes, but there

was no effect on the level of protection faced by Brazilian producers (Kume 1990).

Liberalization effectively began in March 1990, when the newly elected administration of Presi-

dent Collor suddenly and unexpectedly abolished the list of suspended import licenses and removed

nearly all of the remaining special customs regimes (Kume et al. 2003). These policies were replaced

by a set of import tariffs providing the same protective structure, as measured by the gap between

prices internal and external to Brazil, in a process known as tariffication (tarificacao) (de Carvalho,

Jr. 1992). In some industries, this process required modest tariff increases to account for the lost

protection from abolishing import bans.9 Although these changes did not substantially affect the

protective structure, they left tariffs as the main instrument of trade policy, such that tariff levels

in 1990 and later provide an accurate measure of protection.

The main phase of trade liberalization occurred between 1990 and 1995, with a gradual reduction

in import tariffs culminating with the introduction of Mercosur. Tariffs fell from an average of 30.5

percent to 12.8 percent, and remained relatively stable thereafter.10 Along with this large average

decline came substantial heterogeneity in tariff cuts across industries, with some industries such as

agriculture and mining facing small tariff changes, and others such as apparel and rubber facing

declines of more than 30 percentage points. We measure liberalization using long-differences in the

log of one plus the tariff rate from 1990 to 1995, shown in Figure 1. During this time period, tariffs

accurately measure the degree of protection faced by Brazilian producers, and tariff changes from

6Examples include Arbache, Dickerson and Green (2004), Goldberg and Pavcnik (2003), Gonzaga, Filho andTerra (2006), Kovak (2013), Krishna, Poole and Senses (2014), Menezes-Filho and Muendler (2011), Pavcnik, Blom,Goldberg and Schady (2004), Paz (2014), Schor (2004), and Soares and Hirata (2016) among many others.

7Although Brazil was a founding signatory of the General Agreement on Tariffs and Trade (GATT) in 1947, itmaintained high trade barriers through an exemption in Article XVIII Section B, granted to developing countriesfacing balance of payments problems (Abreu 2004). Hence, trade policy changes during the period under study wereunilateral.

8These policies were imposed quite extensively. In January 1987, 38 percent of individual tariff lines were subjectto suspended import licenses, which effectively banned imports of the goods in question (Authors’ calculations fromBulletin International des Douanes no.6 v.11 supplement 2). In 1987, 74 percent of imports were subject to a specialcustoms regime (de Carvalho, Jr. 1992).

9Appendix Figure A1 shows the time series of tariffs. Note the tariff increases in 1990 for the auto and electronicequipment industries.

10Simple averages of tariff rates across Nıvel 50 industries, as reported in Kume et al. (2003). See Appendix A.1for details on tariff data.

5


1990 to 1995 reflect the full extent of liberalization faced by each industry. We do not rely on the

timing of tariff cuts between 1990 and 1995, because this timing was chosen to maintain support

for the liberalization plan, cutting tariffs on intermediate inputs earlier and consumer goods later

(Kume et al. 2003).

As discussed below, along with regional differences in industry mix, the cross-industry variation

in tariff cuts provides the identifying variation in our analysis. Following the argument in Goldberg

and Pavcnik (2005), we note that the tariff cuts were nearly perfectly correlated with the pre-

liberalization tariff levels (correlation coefficient = -0.90). These initial tariff levels reflected a

protective structure initially imposed in 1957 (Kume et al. 2003), decades before liberalization. This

feature left little scope for political economy concerns that might otherwise have driven systematic

endogeneity of tariff cuts to counterfactual industry performance.

To check for any remaining spurious correlation between tariff cuts and other steadily evolv-

ing industry factors, we regress pre-liberalization (1980-1991) changes in industry employment

and average monthly earnings on the 1990-1995 tariff reductions, with detailed results reported

in Appendix B.1. We attempted a variety of alternative specifications and emphasize that the

results should be interpreted with care, as they include only 20 tradable-industry observations.

Most specifications exhibit no statistically significant relationship, but heteroskedasticity-weighted

specifications place heavy weight on agriculture and find a positive relationship. Agriculture was

initially the least protected industry, and it experienced approximately no tariff reduction. It also

had declining wages and employment before liberalization, driving the positive relationship with

tariff reductions. Consistent with earlier work, when omitting agriculture, tariff cuts are unrelated

to pre-liberalization earnings trends (Krishna, Poole and Senses 2011). Given these varying results,

we include controls for pre-liberalization outcome trends in all of the analyses presented below, to

account for any potential spurious correlation. Consistent with the notion that the tariff changes

were exogenous in practice, these pre-trend controls have little influence on the vast majority of

our results.

3 Data and Empirical Approach

3.1 Data

Our main data source for regional labor market outcomes is the Relacao Anual de Informacoes

Sociais (RAIS), spanning the period from 1986 to 2010. This is an administrative dataset assem-

bled yearly by the Brazilian Ministry of Labor, providing a high quality census of the Brazilian

formal labor market (De Negri, de Castro, de Souza and Arbache 2001, Saboia and Tolipan 1985).

Accurate information in RAIS is required for workers to receive payments from several government

benefits programs, and firms face fines for failure to report, so both agents have an incentive to

provide accurate information. RAIS includes nearly all formally employed workers, meaning those

6


with a signed work card providing them access to the benefits and labor protections afforded by

the legal employment system. It omits interns, domestic workers, and other minor employment

categories, along with those without signed work cards, including the self-employed.11 These data

have recently been used by Dix-Carneiro (2014), Helpman, Itskhoki, Muendler and Redding (forth-

coming), Krishna et al. (2014), Lopes de Melo (2013), and Menezes-Filho and Muendler (2011),

though these papers utilize shorter panels. The data consist of job records including worker and

establishment identifiers, allowing us to track workers and establishments over time. We utilize

the establishment’s geographic location (municipality) and industry, and worker-level information

including gender, age, education (9 categories), and December earnings.12

These data have various advantages relative to previous work on the effects of trade on local

labor markets. First, relative to Kovak (2013) and Autor et al. (2013), we can analyze the dynamics

of adjustment to the trade liberalization shock, as RAIS data are available every year. Second, RAIS

is a census rather than a sample, so it is representative at fine geographic levels.13 Third, a rich set

of labor market outcomes can be analyzed with such data, including how liberalization affected job

creation and job destruction rates, the number of active establishments, and the establishment size

distribution. Fourth, the ability to follow workers over time allows us to control for both observable

and unobservable worker characteristics.

As is typically the case in administrative employment datasets, the limitation of RAIS is a lack

of information on workers who are not formally employed, making it impossible to tell whether

a worker is out of the labor force, unemployed, informally employed, or self-employed. This is

important in the Brazilian context, with informality rates often exceeding 50 percent of all employed

workers during our sample period.14 When we need information on individuals who are not formally

employed, or information before 1986, we supplement the analysis using the decennial Brazilian

Demographic Census, covering 1970-2010. While these data provide much smaller samples and do

not permit following individuals over time, they cover the entire population, including the informally

employed, unemployed, and those outside the labor force.15 When possible, we corroborate results

from RAIS using the Demographic Census, finding very similar results across datasets.

Throughout the analysis, we limit our sample to include working-age individuals, aged 18-64.

When studying employed individuals, we omit those working in public administration and those

11See Appendix B.2 for summary statistics on the informal sector, and Dix-Carneiro and Kovak (2015b) for analysescovering the informal labor market.

12RAIS reports earnings for December and average monthly earnings during employed months in the reference year.We use December earnings to ensure that our results are not influenced by seasonal variation or month-to-monthinflation. See Appendix Section A.2 for more detail on the RAIS database.

13The National Household Survey (Pesquisa Nacional por Amostra de Domicılios - PNAD) would be a naturalalternative data source for a yearly analysis, but it only provides geographic information at the state level, does notallow one to follow individual workers over time, and provides a much smaller sample.

14Authors’ calculations using Brazilian Demographic Census.15See Appendix A.3 for more detail on the Demographic Census data and Dix-Carneiro and Kovak (2015b) for

analyses covering the informal labor market.

7


without valid information on industry of employment.16

To analyze outcomes by local labor market, we must define the boundaries of each market. We

use the “microregion” definition of the Brazilian Statistical Agency (IBGE), which groups together

economically integrated contiguous municipalities (counties) with similar geographic and productive

characteristics (IBGE 2002), closely paralleling an intuitive notion of a local labor market. When

necessary, we combine microregions whose boundaries changed during our sample period, to ensure

that we consistently define local labor markets over time. This process leads to a set of 475

consistently identifiable local labor markets for analyses falling within 1986-2010 and 405 markets

for analyses using data from 1980 and earlier.17

3.2 Empirical Approach

Our empirical analysis follows the literature on the regional effects of trade by comparing the evo-

lution of labor market outcomes in regions facing large tariff declines to those in regions facing

smaller tariff declines. Intuitively, regions experience larger declines in labor demand when their

most important industries face larger liberalization-induced price declines (Topalova 2007). Kovak

(2013) presents a specific-factors model of regional economies that captures this intuition (a gen-

eralization of this setup appears below in Section 5.4.1). In this model, the regional labor demand

shock resulting from liberalization is

∑i

βriPi, where βri ≡λri

1ϕi∑

j λrj1ϕj

, (1)

hats represent proportional changes, r indexes regions, i indexes industries, ϕi is the cost share of

non-labor factors, and λri is the share of regional labor initially allocated to tradable industry i. Pi

is the liberalization-induced price change facing industry i, and (1) is a weighted average of these

price changes across tradable industries, with more weight on industries capturing larger shares of

initial regional employment.18 Thus, although all regions face the same vector of liberalization-

16We exclude public administration because the labor market in this field operates quite differently from the restof the market. This choice has no substantive effect on any of our results.

17This geographic classification is a slightly aggregated version of the one in Kovak (2013), accounting for additionalboundary changes during the longer sample period. Related papers define local markets based on commuting patterns(e.g. Autor et al. (2013)). Our local market definition performs well based on this standard as well – only 3.4 and4.6 percent of individuals lived and worked in different markets in 2000 and 2010, respectively. The main regionaldefinition is shown in Figure 2. The analysis omits 11 microregions, shown with a cross-hatched pattern the figure.These include i) Manaus, which was part of a Free Trade Area and hence not subject to tariff cuts during liberalization;ii) the microregions that constitute the state of Tocantins, which was created in 1988 and hence not consistentlyidentifiable throughout our sample period; and iii) a few other municipalities that are omitted from RAIS in the1980s. The inclusion or exclusion of these regions when possible has no substantive effect on the results. We alsoimplemented the main analyses using a more aggregate local labor market definition, “mesoregions” defined by IBGE,and results are nearly identical.

18Following Kovak (2013), we drop the nontradable sector, based on the assumption that nontradable prices movewith tradable prices. We confirm this assumption by calculating a measure of local nontradables prices in Section

8


induced price changes, differences in the regional industry mix generate regional variation in labor

demand shocks.

We operationalize this shock measure by defining the “regional tariff reduction” (RTR), which

utilizes only liberalization-induced variation in prices, replacing Pi with the change in log of one

plus the tariff rate.

RTRr = −∑i

βrid ln(1 + τi) (2)

τi is the tariff rate in industry i, and d represents the long difference from 1990-1995, the period of

Brazilian trade liberalization. We calculate tariff changes using data from Kume et al. (2003), λri

using the 1991 Census, and ϕi using 1990 National Accounts data from IBGE.19 Together, these

allow us to calculate the weights, βri. Note that RTRr is more positive in regions facing larger

tariff reductions, which simplifies the interpretation of our results, since nearly all regions faced

tariff declines during liberalization.

Figure 2 maps the spatial variation in RTRr. Regions facing larger tariff reductions are pre-

sented as lighter and yellower, while regions facing smaller cuts are shown as darker and bluer.

The region at the 10th percentile faced a tariff reduction of 0.2 percentage points, while the region

at the 90th percentile faced a 10.7 percentage point decline. Hence, in interpreting the regression

estimates below, we compare regions whose values of RTRr differ by 10 percentage points, closely

approximating the 90-10 gap of 10.5 percentage points. Note that there is substantial variation

in the tariff shocks even among local labor markets within the same state. As we include state

fixed effects in our analyses, these within-state differences provide the identifying variation in our

study.20

We use the following specification to compare the evolution of labor market outcomes in regions

facing large tariff reductions to those in regions facing smaller tariff declines.

yrt − yr,1991 = θtRTRr + αst + γt(yr,1990 − yr,1986) + εrt (3)

We estimate this equation separately for each year t ∈ [1992, 2010], as reflected by the t subscripts.

yrt is the value of a regional outcome such as earnings or employment, θt is the cumulative effect

of liberalization on outcomes by year t, αst are state fixed effects (allowed to differ across years),

and (yr,1990 − yr,1986) is a pre-liberalization trend in the outcome variable. While the change in

outcome varies with the year t under consideration, the liberalization shock, RTRr, does not.

4.1.19See Appendix A.4 for more detail on the construction of (2). We use the Census to calculate λri because the

Census allows for a more detailed industry definition than what is available in RAIS (see Appendix A.1) and becausethe Census allows us to calculate weights that are representative of overall employment, rather than just formalemployment. That said, shocks using formal employment weights yield very similar results (Appendix Table B6,Panel D).

20A regression of RTRr on state fixed effects yields an R2 of 0.36; i.e. 64% of the variation in RTRr is not explainedby state effects. Our main conclusions are unaffected by the inclusion or exclusion of state fixed effects.

9


Instead, it always reflects the regional measure of tariff reductions during liberalization, from 1990

to 1995. Using this strategy, each year’s θt represents one point on the empirical impulse response

function describing the cumulative local effects of liberalization as of each post-liberalization year.

This methodology captures only relative effects across regions, as does the rest of the literature

examining the regional or sectoral effects of trade.

We use 1991 as the base year for outcome changes, and include state fixed effects to account for

any state-specific policies that might commonly affect outcomes for all regions in the same state,

such as state-specific minimum wages, introduced in 2002 (Neri and Moura 2006).21 We control for

pre-liberalization changes in outcomes (yr,1990 − yr,1986) to address the possibility of confounding

pre-existing trends, and consider longer pre-liberalization trends as a robustness test. For our main

outcomes, we present results with and without state fixed effects and pre-trends, with little effect

on the coefficients of interest. Since many of our dependent variables are themselves estimates, we

weight regressions based on the inverse of their standard error to account for heteroskedasticity.

We also cluster standard errors at the mesoregion level to account for potential spatial correlation

in outcomes across neighboring regions.

To consistently estimate θt, εrt must be uncorrelated with RTRr, conditional on the state fixed

effects and outcome pre-trend. For this identification assumption to be violated, there would need

to be an omitted variable that i) drives wage or employment growth across regions within a state

and ii) is correlated with RTRr but iii) is not captured by pre-liberalization outcome trends. While

such a feature is unlikely to exist, in Section 4.2 we confirm that our results are robust to a wide

variety of potential confounders and alternative specifications.

Our empirical approach is similar to prior studies examining the local effects of trade liberal-

ization, but we make two important contributions to that literature. First, the RAIS data allow us

to calculate changes in regional outcomes in each year following liberalization. We trace out the

dynamic regional response to liberalization as it evolves over time, rather than observing liberal-

ization’s local effect in only one post-shock period, as in the prior literature (e.g. Topalova (2007),

Autor et al. (2013), or Kovak (2013)). The RAIS data also allow us to control for pre-liberalization

trends that might otherwise confound the analysis. Second, we study a discrete, well-defined trade

policy shock that was complete by 1995. This contrasts with Autor et al. (2014), who use U.S.

panel data to study the effects of growing trade with China. They emphasize that the continuously

evolving nature of Chinese trade confounds their ability to study the dynamic response to a trade

shock at any given point in time.

21Using 1991 as the base year allows us to take advantage of more detailed industry information in the 1991 Censuswhen calculating the industry distribution of regional employment (λri), and makes our results comparable withKovak (2013).

10


4 Results

4.1 Main Findings

We begin by examining the effects of liberalization on formal sector earnings and employment

in local labor markets. First, we calculate “regional earnings premia,” which reflect average log

monthly earnings for workers in a given region, controlling for the composition of the regional

workforce.22 For each year t, we regress log December earnings for worker j on flexible controls for

age, sex, and education (Xjt); industry fixed effects (φit); and region fixed effects (µrt).23

ln(earnjrit) = XjtΓt + φit + µrt + ejrit (4)

We estimate this equation separately for each year t ∈ [1991, 2010], allowing the regression coeffi-

cients (Γt) and fixed effects (φit and µrt) to differ across years. The region fixed effect estimates

from these regressions, µrt, represent the regional log earnings premia for the relevant year. By es-

timating these regressions separately in each year, we allow for changes in the regional composition

of workers (X) and changes in the returns to worker characteristics (Γ) over time.24 This approach

ensures that our earnings estimates are not driven by changes in observable worker composition,

changing discrimination, changes in the returns to schooling, or any other changes in the returns to

observable characteristics that operate at the national level. Our dependent variable when studying

earnings is then the change in regional log earnings premium from 1991 to each subsequent year,

1992 to 2010. Table 1 presents summary statistics for this and other main dependent variables

throughout the paper.

Table 2 shows the results of estimating (3) for regional formal sector log earnings premia and

formal log employment. All estimates for the coefficient on RTRr are negative, indicating that re-

gions facing larger tariff reductions experience relative declines in earnings or employment. Consider

Panel A, which presents liberalization’s effect on regional earnings. Columns (1) to (3) examine

changes in earnings from 1991 to 2000, while columns (4) to (6) examine changes from 1991 to

2010, such that the effects cumulate over time. Columns (2) and (4) add state fixed effects, and

columns (3) and (6) add pre-trend controls for the change in the regional outcome from 1986 to

1990. The coefficient estimate of -0.529 in column (3) indicates that a region facing a 10 percentage

point larger tariff reduction (approximately the 90-10 gap in RTRr) experienced a 5.29 percentage

point larger proportional decline (or smaller increase) in formal earnings from 1991 to 2000. This

22Estimating the regional earnings premia for each year separately from the effects of liberalization on regionalearnings reduces the computational demands relative to pooling across years and estimating both steps jointly.

23We use monthly earnings rather than hourly wages because RAIS only provides hours from 1994 onward. Censusresults using hourly wages are similar.

24Appendix B.3 presents the coefficient estimates from (4) for 1991, 2000, and 2010. In Section 5.2, we control forobservable and unobservable worker heterogeneity by pooling across years and including individual fixed effects. Theresults are very similar.

11


magnitude is similar in size to the corresponding estimate in Kovak (2013) (-0.439), which used a

different data source (Census of Population) and covers all workers rather than restricting attention

to the formally employed. The estimate of -1.594 in column (6) indicates that the gap in earnings

growth expanded to 15.94 percentage points by 2010.25

This increase in liberalization’s effect on earnings from 2000 to 2010 is a striking feature of

Table 2. It indicates that the divergent earnings growth in regions facing different tariff reductions

continued well beyond the liberalization period. Figure 3 confirms this pattern by plotting the

coefficients on RTRr (θt) for each year. The points for 2000 and 2010 correspond to the RTRr

coefficients in columns (3) and (6) of Table 2. The vertical lines indicate that liberalization began in

1991 and was complete by 1995. We present coefficient estimates for 1992-94, but these should be

interpreted with care, as liberalization was still ongoing.26 The local earnings effects of liberalization

appear just after liberalization and steadily grow for more than a decade, before leveling off in the

late 2000s, a pattern that is very robust to details of the specification.27 Figure 3 also shows

pre-liberalization coefficients, in which the dependent variable is the change in regional earnings

premium from 1986 to the year listed on the x-axis, and the independent variable is RTRr. If

anything, the relative earnings declines in regions facing larger tariff reductions represent a reversal

of the pre-liberalization trend. Recall that all post-liberalization results control for pre-liberalization

trends, as shown in (3).

It is likely that the prices of local nontradable goods change in response to the regional shocks

to the prices of traded goods (Kovak 2013, Monte 2016). If this is the case, the relative decline in

nominal earnings in regions facing larger tariff reductions may be partly offset by declines in the

local price index. To empirically evaluate this possibility, we construct local price indexes using

housing rents information in the Census, following the approach of Moretti (2013).28 Only the 1991

and 2010 Censuses included rent questions, so we can only calculate the change in rental prices

for 1991-2010. Our local price index uses consumption weights from the Brazilian Consumer Price

Index system (IPC) and accounts for the fact that the prices of non-housing nontradables tend

to move with housing prices. See Appendix A.5 for details on constructing the index. We then

calculate the change in log real earnings as the change in log nominal earnings minus the change in

log local price level. Panel B of Table 2 shows the effect of regional tariff reductions on the change

in real regional earnings for 1991-2010. The effect on real earnings in column (6) is smaller than the

25Appendix B.4 presents an alternative research design at the industry × region level finding similar growth inthe regional earnings effects of liberalization and confirming the importance of cross-industry regional equilibrium indriving the main earnings effects discussed here.

26However, the tariff cuts were almost fully implemented by 1993, so these early coefficients are still informativeregarding liberalization’s short-run effects. When regressing RTRr on an alternate version measuring tariff changesfrom 1990-93, the R2 is 0.93.

27See Section 4.2 for a variety of robustness tests. Appendix B.5 shows the underlying scatterplots, confirming ourchoice of linear estimating equation and showing that the results are not driven by outliers. Appendix B.6 showsthat the same pattern appears when estimating formal earnings or formal hourly wages using Census data.

28As in the U.S., the Brazilian government does not produce local price indexes outside a few large cities.

12


effect on nominal earnings by about 21 percent. This difference confirms that regional nontradable

prices move with tradable prices, falling more in places facing larger tariff reductions. However, the

long-run effects of liberalization on real regional earnings are still large and statistically significant.

Table 2 Panel C and Figure 4 both examine liberalization’s effects on regional log formal em-

ployment. The year 2000 estimate of -3.533 shows that a region facing a 10 percentage point larger

tariff reduction experienced a 35.3 percentage point larger proportional decline (smaller increase)

in formal employment from 1991 to 2000. As with earnings, the employment effect grew substan-

tially from 2000 to 2010, indicating that employment growth continued to diverge for regions facing

different regional tariff changes. Most of this divergence was complete by 2004, after which the

estimates level off.29

Note that since Table 2 and Figure 4 examine formal employment, there are two channels

through which formal employment might decline in regions facing more negative shocks. Formally

employed workers may migrate away from negatively affected places to more favorably affected

places, or existing residents of the region may shift out of formal employment and into non-

employment or informal employment. Table 3 rules out the interregional migration mechanism,

showing that a region’s working-age population did not respond to RTRr.30 We measure working-

age population using Census data, so we can observe individuals outside formal employment, and

control for 1980-1991 and 1970-1980 population pre-trends. None of the population estimates are

significantly different from zero, and the point estimates with extensive pre-trend controls (columns

(3) and (6)) are quantitatively small. These results suggest that workers losing formal employment

in harder hit regions did not leave the region, but transitioned out of formal employment.

Using Census data on informal workers, Panel A of Table 4 confirms that in regions facing larger

tariff reductions informal employment increased relative to the national average. For example, the

estimate 1.196 in column (6) implies that on average a region facing a 10 percentage point larger

tariff reduction experienced an 11.96 percentage point larger increase in informal employment by

2010. Rather than migrating away, many workers who lose formal employment in negatively affected

regions appear to transition into the informal sector in the same region.31 Panel B of Table

4 implements a similar exercise for regional earnings premia in the informal sector. Somewhat

unexpectedly, there is no significant relationship between regional tariff reductions and earnings

in the informal sector. A potential explanation for the lack of effect on informal wages is that

29To assess the scale of our long-run estimates, consider Dix-Carneiro (2014), which studies a very similar settingwith slow adjustment of labor across Brazilian industries rather than regions. After estimating the model’s parametersusing RAIS data, he simulates the economy’s response to a price shock when capital is mobile across industries (seehis Figures 4 and 6). The long-run wage elasticity in the adversely affected sector (High-Tech Manufacturing) is -1.56.This is exceedingly close to our 2010 earnings estimate of -1.594. The long-run employment elasticity in Dix-Carneiro(2014) is -3.2. Although this is somewhat smaller than our 2010 employment estimate of -4.663, the two effects aresimilar in magnitude, suggesting that our findings are reasonable in the context of this type of model.

30Similarly, Autor et al. (2013) find little evidence for population responses to trade shocks in the U.S.31See Dix-Carneiro and Kovak (2015b) for a more extensive discussion of the various margins of labor market

adjustment following Brazilian liberalization.

13


consumers in harder-hit regions experience declining incomes and shift toward lower-priced, lower-

quality goods produced in the informal sector, offsetting wage declines for informally employed

workers.32

Together, these results are quite surprising, particularly compared to the conventional wisdom

from the literature studying local labor demand shocks. The standard framework predicts initially

large wage effects of local labor demand shocks, as labor supply is approximately fixed in the

very short run, after which employment adjustment arbitrages away spatial wage differences, and

observed wage effects fall in magnitude (Blanchard and Katz 1992, Bound and Holzer 2000). This

mechanism is consistent with the steadily growing employment effects in Figure 4, but is at odds

with the growing earnings effects in Figure 3. It predicts large negative coefficients shortly after

liberalization, but then declining magnitude effects as arbitrage partly equalizes earnings growth

across regions. Even in the absence of equalizing migration, as shown in Table 3, one would expect

constant effects over time. Instead, we find continuing divergence in earnings growth for 14 years

following the end of liberalization, with earnings growth in regions facing larger tariff reductions

lagging further and further behind other regions. This pattern means that the local labor market

effects of trade estimated in prior work for a single post-liberalization year actually understate the

longer run effects. The remainder of the paper focuses on examining and explaining this surprising

result.

4.2 Robustness

We first establish that the steadily growing earnings effects are robust to alternative measurement

and specification choices and that they were not driven by confounding effects from other shocks

to Brazilian local labor markets. Detailed analyses appear in Appendix Sections B.7–B.9, and we

summarize the results here.

Appendix B.7 shows that the growing earnings estimates are robust to alternative pre-trend

controls, RTRr shock measures, earnings premium measures, and weighting. We use Census data to

construct longer pre-liberalization earnings trends, from 1970-1980 and from 1980-1991, and control

for these alongside the 1986-1990 RAIS pre-liberalization trends present in our main specification.33

We construct alternative RTRr measures, i) using industry weights, λri, reflecting only formal

employment, ii) using effective rates of protection, which account for the effects of tariffs on inputs

32Burstein, Eichenbaum and Rebelo (2005) show that lower quality goods gain market share in recessions, whileMcKenzie and Schargrodsky (2011) make a similar argument in the context of the 2002 economic crisis in Argentina.While there is little direct evidence on the relative quality of goods produced by formal and informal firms, it is wellknown that informal firms are significantly smaller than formal firms (LaPorta and Schleifer 2014, Meghir, Narita andRobin 2015, Ulyssea 2014), and Kugler and Verhoogen (2011) show that larger firms produce higher quality goodsthan small firms, on average. Moreover, LaPorta and Schleifer (2008) show that informal firms use lower qualityinputs and speculate that they produce lower quality outputs as a result.

33Because 1991 is the base year for our post-liberalization earnings growth outcome, 1980-1991 pre-liberalizationtrends are subject to mechanical endogeneity. We resolve this problem by calculating an alternative earnings growthmeasure with 1992 as the base year. See Panel C of Table B6.

14


and outputs for each industry, and iii) including a zero price change for the nontradable sector. We

also construct alternative earnings premium measures. The first is calculated without controlling

for industry fixed effects, maintaining national industry-level earnings variation in the regional

earnings premia.34 The second measure simply uses mean log earnings, without controlling for

any worker characteristics. Finally, we present results weighting regions equally or weighting by

the region’s 1991 formal employment. In all cases, our main results are confirmed, finding steady

growth in liberalization’s effects on regional earnings. The employment effects are similarly robust

to these alternatives.

Although our findings are robust to these specification and measurement changes, the effects

of liberalization could appear to grow over time because of correlated shocks occurring after trade

liberalization. To explain the smooth growth of the effects in Figures 3 and 4, such confounders

would need to affect industries or regions similarly to liberalization and would need to grow steadily

over time or occur quite regularly. Although these circumstances are unlikely, in Appendix B.8 we

construct controls for a wide variety of salient economic shocks in the post-liberalization period,

demonstrating that they cannot account for the growing earnings effects.

If tariff changes after 1995 were correlated with those occurring during liberalization (1990-

95), they might drive the apparently increasing effects of liberalization, although this is unlikely

since post-1995 tariff changes were very modest. We calculate post-liberalization regional tariff

reductions as in (2), but use tariff reductions between 1995 and each year t > 1995, and include

these post-liberalization tariff reductions as additional controls alongside RTRr. Other potential

confounders are the Brazilian Real devaluations that occurred in 1999 and 2002. If these exchange

rate movements affected industries differently, they might have been correlated with tariff changes

during liberalization. We construct industry-specific real exchange rates as import- or export-

weighted averages of real exchange rates between Brazil and its trading partners. We then take the

change in log real exchange rate from 1990 to year t > 1995, and calculate regional shocks using

weighted averages as in (2). There was also a substantial wave of privatization during our sample

period. We address privatization by controlling flexibly for the 1995 share of regional employment

at state-owned firms or the change in this share from 1995 to t. Controlling for each of these post-

liberalization shocks has little effect on the earnings results, which continue to exhibit substantial

post-liberalization growth in all cases.

The global commodity price boom of the late 2000s is another potential post-liberalization

confounder that might explain our growing earnings results, particularly since agricultural products

faced the most positive tariff change during liberalization. In Appendix B.8.4, we provide extensive

evidence ruling out this possibility. First, the timing of the commodity price boom does not

34By omitting the industry fixed effects, these regional earnings measures include both direct industry effects andlocal general equilibrium effects. As shown in Appendix B.7, the associated estimates are only a bit larger than themain results, indicating that local equilibrium effects account for the majority of the overall effects of liberalizationon regional earnings.

15


correspond to the timing of our effects. Commodity prices were flat or declining between 1991 and

2003, during which our earnings and employment results grew substantially. Commodity prices

then grew sharply after 2004, when our results began to level off.35 We show that the substantial

growth in earnings effects remains when i) dropping regions most exposed to commodity price

growth, by restricting the region sample to include only those with below-median or bottom-

quartile employment shares in agriculture and mining or ii) when restricting our regional earnings

measure only to workers in manufacturing. Finally, we use three approaches to directly control for

the regional effects of the commodity price boom. We control for the share of workers in agriculture

and mining and for changes in regional commodity prices using the measure introduced by Adao

(2015). We also use more detailed commodity price data from the IMF Primary Commodity

Price Series to construct similar regional controls for commodity price changes. Finally, we control

for China’s effects on commodity markets using the import and export quantity measures and

instruments from Costa et al. (2016).36 All of these controls have little influence on the observed

earnings effects of RTRr.

As a final set of robustness tests, Appendix B.9 presents results when splitting the sample

by tradable and nontradable sector and by skill. We find growing earnings effects in all of these

subsamples. This pattern is particularly noteworthy for the nontradable sector, as it confirms that

regional labor market equilibrium transmits the effects of liberalization from the tradable sector

to the nontradable sector, as predicted in the model of Kovak (2013), which is the basis for the

RTRr shock. The earnings effects for more skilled workers are a bit larger than those for less skilled

workers, while the employment effects are larger for less skilled workers. However, these results

should be interpreted with care, as the RTRr shocks are derived from a model with a single type

of labor.37

Together, the results in this section demonstrate the robustness of our main findings to alterna-

tive measures and estimation approaches and rule out a wide variety of salient post-liberalization

shocks as potential confounders. We conclude that the earnings and employment profiles shown in

Figures 3 and 4 reflect growing causal effects of liberalization over time. In the next section, we

consider a variety of potential mechanisms that could drive this growth in liberalization’s effects

on local labor market outcomes.

35A similar argument applies to Bustos, Caprettini and Ponticelli (2016), who study the effects of geneticallymodified crops in Brazil. Genetically modified crops were outlawed before 2003 and only permanently authorized in2005, so this channel cannot explain the substantial growth in earnings effects before 2005.

36Special thanks to Rodrigo Adao for providing commodity price data and code, and to Francisco Costa forproviding the shock and instrument measures from Costa et al. (2016).

37For a more general model with two skill types, see Dix-Carneiro and Kovak (2015a).

16


5 Mechanisms

As mentioned above, the conventional model of local labor markets predicts large effects of liberal-

ization just after the tariff change and smaller effects as labor reallocation arbitrages away spatial

differences in earnings growth. Our findings contradict this prediction, instead exhibiting increas-

ing differences in earnings growth for 15 years after liberalization between regions facing larger

and smaller tariff reductions. In this section, we consider a variety of potential mechanisms that

might explain these growing earnings effects, finding strong empirical support for mechanisms in-

volving imperfect interregional labor mobility and dynamic labor demand, particularly slow capital

adjustment and agglomeration economies.

5.1 Urban Decline

Glaeser and Gyourko (2005) and Notowidigdo (2013) present models of urban decline in which the

slow depreciation of housing stocks drives slow adjustment in local labor markets facing permanent

negative labor demand shocks. In their models, the price of housing falls sharply in depressed

markets, incentivizing individuals to remain in the city in spite of nominal earnings losses follow-

ing the demand decline. As housing slowly depreciates, this incentive dissipates, so population

and therefore employment steadily decline. This mechanism could therefore rationalize the slowly

growing employment effects we document in Figure 4.

However, as in the conventional model of local labor markets, this mechanism predicts the

opposite of what we find for earnings in Figure 3. Although wages fall on impact in regions

facing negative shocks, they recover slowly over time as workers leave the market due to housing

depreciation.38 Moreover, the mechanism depends on declining population in cities facing negative

shocks. In Brazil, overall population growth was large enough during our sample period that out of

475 local labor markets, only 11 experienced population decline between 1991 and 2000, and only 6

did so between 1991 and 2010.39 Table 3 also finds no response of local working-age population to

RTRr. Thus, while the slow housing depreciation mechanism is quite relevant for rust-belt cities

in the U.S., it does not appear to apply in the Brazilian context.

5.2 Changing Composition of Worker Unobservables

Liberalization might cause average earnings to slowly decline in regions facing larger tariff reductions

relative to other regions because of worker selection. Higher-earning workers may be more likely

38Note that Glaeser and Gyourko (2005) do not model a production side and instead directly shock wages oramenities. However, a simple extension of their model to include labor market equilibrium would have the featurescited here, as in Notowidigdo (2013). Glaeser and Gyourko (2005) also argue that local average wages will decline overtime in negatively shocked markets because the most productive workers have the strongest incentive to leave. Asshown in Section 5.2, since we control for worker characteristics when calculating regional earnings premia, selectioneffects of this kind are not driving our results.

39Authors’ calculations using Census data.

17


to leave the formal labor market in harder-hit regions, and this selective worker reallocation may

increase over time. Although we flexibly control for detailed worker characteristics including age,

sex, and education when calculating regional earnings premia in our main specifications, worker

composition may also adjust along unobservable dimensions.

To examine this possibility, we calculate alternative earnings premia, pooling the RAIS data

across years and controlling for worker-level fixed effects, which capture time-invariant worker

characteristics, including unobservables.40 We implement this procedure in two ways. First, we use

a straightforward worker fixed-effects regression.

ln(earnjairt) = αj + ψa + φit + µrt + εjairt, (5)

where αj are worker fixed effects, ψa are age effects (indicators for falling within each age bin shown

in Table B3), and φit are time-varying industry effects. We then calculate the change in log regional

earnings premium using the regional earnings estimates, µrt, and examine their response to RTRr.

As shown in Panel B of Table 5, when controlling for worker unobservables in this fashion, the

earnings effects continue to grow over time.

A limitation of (5) is that it restricts the returns to worker characteristics to be constant over

time. Since the returns to observable characteristics change substantially over time (see Appendix

B.3), we allow for time-varying returns (δt) to worker characteristics (αj) in the following specifi-

cation.

ln(earnjairt) = δtαj + ψa + φit + µrt + εjairt. (6)

δt can vary arbitrarily over time, but does so identically for all workers. This restriction distinguishes

δtαj from worker × time fixed effects, which would absorb all variation in the data. We estimate (6)

using the iterative algorithm described by Arcidiacono, Foster, Goodpaster and Kinsler (2012) and

calculate standard errors using the wild bootstrap method suggested by Davidson and MacKinnon

(2006), with 500 iterations. Panel C of Table 5 presents earnings estimates using the resulting

regional earnings premia. The growth in earnings effects remains, and the results from the more

flexible earnings premium specification in Panel C are quantitatively very close to the baseline

specification in Panel A. These findings rule out worker selection as a mechanism driving the

observed growth in the earnings effects of liberalization.

5.3 Slow Response of Imports or Exports

Although trade liberalization was complete by 1995, it is possible that trade quantities were slow to

respond to the sharp change in trade policy, perhaps because of difficulty in forming new trade links

40For computationally tractability, we draw a 3 percent random sample of all valid individual IDs that appearin RAIS with a positive earnings observation between 1986 and 2010. This procedure yields 450 microregions withformally employed workers earning labor income in December for all years in our sample.

18


with firms abroad. Prices faced by Brazilian producers may evolve slowly in response to tariff cuts

if import quantities respond slowly to liberalization. If so, the slow evolution of imports in response

to the tariff cuts could potentially explain the slow growth in the effects of liberalization over time.

To examine this possibility, we i) study the relationship between regional tariff reductions and trade

quantity measures to determine whether such a slow trade response occurred in practice and ii)

control for changes in trade quantities to see whether they mediate the relationship between tariff

changes and earnings. We follow Autor et al. (2013) by constructing changes in imports and exports

per worker for each industry from 1991 to each subsequent year, using Comtrade data.41 We then

form regional weighted averages of these changes in trade flows, weighting by the industry’s initial

share of regional employment. See Appendix A.6 for details on the construction of these measures.

We first examine the effect of regional tariff reductions on these regional measures of import,

export, and net export growth, looking for evidence of slow growth in trade quantities that might

drive the slow growth in earnings effects. We do so using the trade growth measures as dependent

variables in (3). Figure 5 plots the effects of RTRr on each trade flow measure.42 First, consider

the effects on regional imports (blue circles). As expected, regions facing larger tariff reductions

experienced larger increases in the regional import measure. These import increases occurred

immediately after liberalization, with large positive coefficients already present in 1995. Because

we measure trade flows in $100,000 units, the 1995 coefficient of 0.144 implies that a region facing

a 10 percentage point larger tariff reduction experienced a $1,440 larger increase in imports per

worker. These import effects actually decrease on average until 2003 (coefficient estimate = 0.070),

in sharp contrast to the earnings effects, which grew to more than two-thirds of their long-run level

during the same time period. After 2003, the import effects increase, but this coincides with a

leveling-off in the earnings and employment effects. This timing is inconsistent with slow import

growth driving our results.

The sign of the export effects (red triangles) is positive, indicating that industries experiencing

larger export increases were on average located in the same regions as industries facing larger

tariff reductions.43 This effect works against the hypothesis that slow trade quantity growth drove

relative earnings declines in these regions. After 2003, both the import and export effects grow quite

substantially, following the overall trends in Brazilian imports and exports. Note, however, that

the relationship between RTRr and net exports (green diamonds) falls from 2005 to 2010, again a

time period with substantial growth in the earnings effects. Overall, the evolution of import and

export quantities is not consistent with the hypothesis that slow trade quantity growth explains

our results.

41Appendix B.10 shows results for an ad-hoc alternative functional form using the change in log trade, yielding thesame conclusions.

42In Figure 5 we do not have pre-liberalization trends for trade flows because Comtrade data for Brazil begin in1989.

43The positive sign for the export effect is not driven by any particular industry or industries and is robust todropping agriculture and/or natural-resource industries.

19


To confirm this point, we include controls for regional import and export growth when examining

the effect of RTRr on regional earnings premia. If the growing earnings effects remain when

including these controls, we can be confident that a different mechanism is at play. We examine

the relationship between earnings growth and RTRr, as in (3), including controls for the regional

growth in imports (RegImprt) and exports (RegExprt) from 1990 to year t.

yrt − yr,1991 = θtRTRr + β1RegImprt + β2RegExprt + αst + γt(yr,1990 − yr,1986) + εrt (7)

The import and export coefficients, β1 and β2, are constant over time, allowing us to test whether

the slow evolution of trade flows explains the evolution of earnings growth (since RegImprt and

RegExprt change over time, unlike RTRr). Panel B of Table 6 shows that the effect of RTRr on

regional earnings still grows steadily over time when controlling for changes in regional imports and

exports, implying that slow trade quantity growth is not driving the relationship between the tariff

reductions and earnings.

A remaining concern is that if regional imports and exports are endogenous to regional earn-

ings growth, then the coefficients on RTRr will be biased along with the trade flow coefficients,

invalidating the analysis just described. Panels C and D address this issue following the strategy of

Autor et al. (2013), instrumenting for Brazilian trade flows using trade flows for other countries.44

We consider instruments based on the combination of Argentina, Chile, Colombia, Paraguay, Peru,

and Uruguay (“Latin America”) and on Colombia alone, which liberalized during the same time

period as Brazil and imposed similar tariff cuts across industries (Paz 2014). In each case, we

measure imports and exports between these countries and the rest of the world, excluding Brazil.45

Panels C and D show the results. In both cases, the effects of RTRr continue to grow over time,

with a similar magnitude to the main results, shown in Panel A. These and the preceding results in

this section rule out slow import or export responses as the mechanism driving the slowly growing

earnings effects.

5.4 Dynamic Labor Demand

A remaining potential mechanism driving the growing effects of liberalization on earnings and

employment involves dynamics in labor demand. If labor is imperfectly mobile across regions and

an initial labor demand shock is followed by a dynamic process that amplifies the shock’s effects

over time, one will observe the growing regional earnings and employment effects we document.

We consider two potential sources of these dynamics: agglomeration economies (e.g. Kline and

Moretti 2014) and slow adjustment of capital stocks (e.g. Dix-Carneiro 2014). As we will show,

both appear to play important roles in explaining our findings.

44We also include regional measures of commodity price growth from Adao (2015) in the set of instruments.45Due to Comtrade data availability, the changes in trade flows for Latin America are calculated from 1994 to each

subsequent year and those for Colombia alone are calculated from 1991 to each subsequent year.

20


5.4.1 Evidence on the Importance of Dynamic Labor Demand

To study these mechanisms and formalize our argument, we generalize the specific-factors model

in Kovak (2013) to include agglomeration economies and slow adjustment of labor and capital. We

focus on the formal economy, consisting of many regions, indexed by r, which may produce goods

in many industries, indexed by i. Production in each industry uses Cobb-Douglas technology with

constant returns to scale and three inputs: labor, a fixed factor, and capital. Formal labor, Lr, is

assumed to be perfectly mobile between industries within a region. The fixed factor, Tri, is usable

only in its respective region and industry and is fixed over time. This factor represents inputs such

as natural resources, land, or very slowly depreciating infrastructure and capital that are effectively

fixed over the time horizons we consider. Capital, Kri, is also usable only in its respective region

and industry but may change slowly over time through depreciation and investment decisions.46

Output of industry i in region r is

Yri = AriL1−ϕiri

(T ζiriK

1−ζiri

)ϕi(8)

where ϕi, ζi ∈ (0, 1). Goods and factor markets are perfectly competitive, and producers face

exogenous prices Pi, common across regions and fixed by world prices and tariffs. To allow for

the possibility of agglomeration economies, we allow productivity, Ari, to vary with the amount

of local economic activity. We also allow for factor adjustment by letting Lr and Kri change over

time. Recall that changes in Lr primarily reflect workers entering or leaving the formal workforce

rather than other channels such as interregional migration, as shown in Table 3. We assume that

changes in Kri reflect depreciation and firms’ investment decisions rather than physical mobility

via secondary markets for installed capital.

As shown in Appendix C, factor market clearing, zero profits, and cost minimization imply the

following equilibrium relationship, in which hats represent proportional changes.

wr =∑i

βriPi +∑i

βriAri − δr

(Lr −

∑i

λri(1− ζi)Kri

)(9)

where βri ≡λri

1ϕi∑

j λrj1ϕj

> 0 and δr ≡1∑

j λrj1ϕj

> 0.

wr is the proportional change in the regional wage, and λri is the initial share of regional employment

in industry i. This is an equilibrium relationship because the factor supplies and productivity levels

may respond endogenously to the liberalization shock reflected by Pi.

46We separate fixed factors and variable capital for two reasons. First, our research design is based on regionaldifferences in industry mix, which are driven by fixed factors. Second, including fixed factors in each region ensuresthat all regions maintain some economic activity even when faced with very negative shocks. Hence, this formulationis common in the literature on agglomeration economies (e.g. Helm (2016) and Kline and Moretti (2014)).

21


As a thought exercise, suppose we were to hold productivity and factor supplies constant (Ari =

Lr = Kri = 0). In that case, the wage change equals the simple weighted average price shock in

(1). In this restricted model, there is no scope for dynamic effects of liberalization, and one would

observe a substantial wage effect of liberalization on impact, with no changes thereafter. More

realistically, if productivity or factor supplies evolve over time in response to the liberalization-

induced price shocks, then the effects of liberalization on regional wages can change over time as

well.

First, we consider factor supply responses. Imagine that only regional labor supply responds

to liberalization, while maintaining Ari = Kri = 0. Immediately following liberalization, wages

decline more in regions facing larger tariff reductions, and formal employment falls more in these

regions, as in Figure 4. Equation (9) shows that this change in employment partly offsets the

wage losses experienced on impact, since δr > 0. If employment adjusts slowly, then the observed

wage effects of liberalization get smaller over time. In other words, with labor adjustment only,

the model reflects the conventional prediction that liberalization’s effects on local wages decline

over time. If we allow both regional employment and regional capital stocks to vary in response to

liberalization, complex patterns can emerge, depending on the relative speed of labor and capital

adjustment. For example, if regional labor is held fixed and capital stocks contract more in regions

facing larger tariff declines (as we show below), the marginal product of labor will fall, and relative

wages will decline even further in harder hit regions, as seen in Figure 3. More generally, the model

can qualitatively rationalize growing earnings effects of liberalization if the labor supply elasticity

is finite and capital adjusts more quickly than labor.

Now consider changes in productivity, Ari. We assume that these result from agglomeration

economies in which changes in the amount of local economic activity drive changes in the produc-

tivity of local firms. There is little agreement on the specific source of agglomeration economies,

with various papers arguing that they result from changes in population, overall employment,

or employment in particular industries (Melo, Graham and Noland 2009).47 For agglomeration

economies to be relevant in our context, we must observe effects of regional tariff reductions on

at least one of these agglomeration sources. In Table 3 and Appendix B.11, we show that neither

working-age population nor overall employment (sum of formal and informal) respond substan-

tially to RTRr, while Figure 4 shows that liberalization substantially affected formal employment.

For agglomeration economies to be relevant in our context, agglomeration must apply to regional

formal employment, since other potential sources of agglomeration do not significantly respond to

liberalization. This is plausible, as labor market pooling and knowledge spillovers are more likely to

apply in formal employment than in informal employment, which disproportionately includes agri-

cultural production. In this case, a negative labor demand shock decreases wages on impact, which

47Many papers argue that population or employment density is the relevant quantity, but since we utilize regionswith fixed boundaries, the change in log population or employment density is identical to the change in log populationor employment level.

22


endogenously decreases formal employment and therefore decreases regional productivity through

agglomeration economies. As shown in (9), this productivity decline amplifies the wage decline

from the initial shock, leading to further reductions in local formal employment and productivity,

etc. If this amplification occurs slowly over time, perhaps due to slow labor supply responses or

slow responses of productivity to formal employment (Kline and Moretti 2014), then the observed

effects of liberalization may also grow over time.

Therefore, given imperfect labor mobility across regions, both capital adjustment and agglom-

eration economies could qualitatively explain the earnings and employment patterns in Figures 3

and 4. To provide evidence for the relevance of dynamic labor demand, we rearrange (9) to infer the

labor demand shifts needed to rationalize the changes in earnings with the observed regional tariff

reductions and changes in formal employment. For consistency with the agglomeration literature,

we assume identical factor cost shares across industries (ϕi = ϕ ∀i and ζi = ζ ∀i, which implies

δr = ϕ).48 The economy-wide value of ϕ is 0.544 (see Appendix A.4), and we discuss the value of

ζ in Section 5.4.3. ∑i

βriAri + ϕ(1− ζ)∑i

λriKri = wr −∑i

βriPi + ϕLr︸︷︷︸observed

(10)

The left hand side of (10) captures the overall shifts in labor demand resulting from agglomera-

tion economies and capital adjustment, which we can measure as a residual using the observable

quantities on the right hand side. We measure wr as the change in regional earnings premium,

−∑

i βriPi as RTRr, and Lr as the change in regional formal employment. Figure 6 (solid blue

circles) shows the relationship between this inferred labor demand measure and regional tariff re-

ductions in each year following the start of liberalization. We can infer that labor demand steadily

declined in regions facing larger tariff reductions and that these dynamics were complete by the

late 2000s. Given this evidence for dynamic labor demand in general, we examine evidence for the

two specific sources of dynamics: agglomeration economies and slow capital adjustment.

5.4.2 Evidence for Agglomeration Economies and Capital Adjustment

To examine these mechanisms in more detail, we follow the literature by imposing additional long-

run assumptions that allow us to compare our results to prior work and to quantify the roles of

agglomeration and slow capital adjustment. We assume a constant elasticity long-run agglomeration

function.49

Ari = κLr, κ ≥ 0 (11)

48When assuming identical factor cost shares across industries, our production function is identical to those in Klineand Moretti (2014) and Helm (2016). Hanlon and Miscio (2016) use a slightly different Cobb-Douglas productionfunction, but also assume constant cost shares across industries.

49Kline and Moretti (2014) provide empirical support for a constant agglomeration elasticity.

23


Table 3 shows that working-age population does not substantially respond to liberalization, indicat-

ing that the main margin of labor supply adjustment is workers’ choice of whether to pursue formal

employment within a given region. Table 4 shows informal sector earnings do not substantially

respond to liberalization. Therefore, we assume that changes in formal labor (Lr) depend upon

changes in the regional formal wage (wr), and assume a constant elasticity long-run local formal

labor supply function.

Lr =1

ηwr, η ≥ 0 (12)

Finally, we assume perfectly mobile capital in the long run (Rr = R ∀r, where R is the price of

capital).50 We take 2010 to be the long run (20 years following the start of liberalization), consistent

with the flat earnings and employment responses by the late 2000s. Imposing these assumptions

on the model yields the following expressions for the long-run regional wage change and the change

in employment in a given region × industry combination (derived in Appendix C).

wr =η

η[1− ϕ(1− ζ)]− κ+ ϕζ

∑i

βriPi −ϕ(1− ζ)η

η[1− ϕ(1− ζ)]− κ+ ϕζR (13)

Lri =1

ϕζPi −

1

ϕζ· η[1− ϕ(1− ζ)]− κη[1− ϕ(1− ζ)]− κ+ ϕζ

∑i

βriPi −ϕ(1− ζ)

η[1− ϕ(1− ζ)]− κ+ ϕζR (14)

We test for the presence of agglomeration economies using the change in employment in each

region × industry combination, following an approach similar to that of Helm (2016). As shown in

(14), in the absence of agglomeration (κ = 0), holding fixed an industry’s own price decline, larger

regional tariff reductions increase local industry employment. Intuitively, if other industries in the

same region face larger tariff cuts, more laborers will locally transition into the reference industry

in equilibrium. However, in a setting with agglomeration economies (κ > 0), price reductions

in other industries in the same region reduce the local productivity of the reference industry. If

agglomeration forces are strong enough, larger regional tariff reductions can reduce local industry

employment conditional on the industry’s own price change. We therefore estimate the following

specification.

Lri = γ0 + γ1Pi + γ2RTRr + εri (15)

This expression is the reduced form of (14). γ0 captures the term for R, which does not vary across

industries or regions, and γ2 < 0 implies the presence of agglomeration economies.51 We measure

Lri using changes from 1991 to 2010 to capture long-run adjustment. We control for industry price

changes either directly using tariff reductions (−d ln(1 + τi)), or with industry fixed effects. Since

the nontradable sector does not directly experience a tariff change, we use RTRr to measure its

50Perfect long-run capital mobility is a standard assumption in this literature (Hanlon and Miscio 2016, Helm 2016,Kline and Moretti 2014).

51Recall that RTRr ≡ −∑i βriPi, so γ2 < 0 implies η[1−ϕ(1−ζ)]−κ

η[1−ϕ(1−ζ)]−κ+ϕζ < 0 in (14), which in turn implies κ > 0.

24


price change, following the arguments in (Kovak 2013). The results of estimating (15) appear in

Table 7. In all cases, the coefficient on RTRr is negative and highly significant; an industry’s local

employment actually falls when other industries in the same region face larger tariff reductions,

implying the presence of agglomeration economies. This finding is robust to including state fixed

effects and outcome pre-trends, to using direct industry price change controls or industry fixed

effects, and to restricting attention to tradable industries.52

We also find evidence for slow capital adjustment. Although regional capital stock measures are

unavailable, we can observe changes in the number of formal establishments in a given region, which

are likely to approximate changes in regional capital stocks.53,54 Figure 7 shows that regions facing

larger tariff reductions experienced steady relative declines in the number of formal establishments,

with the effect increasing most quickly in the early 2000s and leveling out later in the sample period.

It is possible that capital simply reallocated from smaller exiting establishments to larger continuing

establishments in harder-hit locations. If this were the case, the change in number of establishments

would not be particularly informative about the change in regional capital stock. However, the

decline in the number of establishments was not offset by increases in the average size of remaining

establishments; if anything these establishments shrank on average. Moreover, Appendix B.13

shows that larger tariff declines drove increases in exit rates throughout the establishment size

distribution. These results strongly support the interpretation that trade shocks induced a gradual

reallocation of capital away from harder hit locations.

To reinforce this conclusion, we present evidence on the margins of capital adjustment. We

expect investment to respond immediately following liberalization, with new investment directed

toward more favorable markets and away from markets facing larger tariff reductions. In contrast,

depreciation takes time to erode the capital stock in a negatively affected region. We confirm these

patterns using measures of regional establishment entry and exit and job creation and destruc-

tion. We measure cumulative entry, exit, job creation, and job destruction by observing changes

from 1991 to each subsequent year, and calculate each measure following Davis and Haltiwanger

(1990).55 We then examine the relationship between the log of each measure and RTRr. Figure

8 reports the results for entry and exit, and Figure 9 shows the results for job creation and de-

52Because we use RTRr to measure the industry-specific price change for nontradable industries, it is not possibleto separately identify the effects of industry-specific and regional tariff reductions for nontradable industries alone.

53It is not possible to construct regional capital stocks in Brazil during our sample period. Capital investmentin manufacturing firms could in principle be constructed from the Annual Manufacturing Survey (PIA) beginningin 1996, but the Brazilian Statistical Agency (IBGE) has a strict policy against constructing PIA variables at theregional level. Moreover, with investment data beginning in 1996, we would not have credible capital stock measuresuntil well after liberalization. Data sources covering non-manufacturing sectors also begin well after liberalization.

54Regional capital could slowly reallocate from firms in the formal sector to firms in the informal sector, but thisis unlikely, as firms in the informal sector are much less capital intensive than those in the formal sector (LaPortaand Schleifer 2014, Fajnzylber, Maloney and Montes-Rojas 2011)

55For establishment entry and exit, the Davis and Haltiwanger (1990) measure reduces to the number of establish-ments that entered or exited between 1991 and year t as a share of active establishments in year t. See AppendixA.7 for details.

25


struction. New investment, as observed in establishment entry and job creation, falls immediately

in negatively affected regions and stays low throughout the sample period. In contrast, the exit

and job destruction effects grow slowly over time as existing establishments in regions facing larger

tariff cuts allow their installed capital stocks to erode through depreciation, directing investment

elsewhere. Together, these results support the conclusion that capital slowly reallocated away from

regions facing larger tariff declines, steadily amplifying the earnings effects of liberalization.

5.4.3 Quantification

The preceding results provide evidence that both slow capital adjustment and agglomeration

economies play qualitatively important roles in driving the evolution of liberalization’s effects on

earnings and employment. We now investigate the extent to which these mechanisms can quanti-

tatively explain the long-run labor market effects we observe.

We begin by examining the share of the long-run change in labor demand that can be explained

by regional capital adjustment. In (10), capital’s contribution to overall adjustment is given by

ϕ(1 − ζ)∑

i λriKri. We proxy for∑

i λriKri using the change in log number of regional formal

establishments (as discussed above) and measure ζ (fixed-factors’ share of non-labor input costs),

using estimates of equipment, structures, and land cost shares from Valentinyi and Herrendorf

(2008).56 We consider three alternative values for ζ, defining fixed factors as i) land only (ζ = 0.152),

ii) land and structures (ζ = 0.545) and iii) land and half of structures (ζ = 0.349).57 Figure 6 shows

the evolution of liberalization’s effect on these capital adjustment measures compared to the overall

labor demand adjustment inferred from (10). Although the shapes of the capital adjustment and

overall adjustment profiles are not identical, they both grow over time and have similar scales.

Depending on the value of ζ, capital adjustment can account for between 47 and 88 percent of the

inferred labor demand adjustment in 2010. While this is a somewhat wide range, it is clear that

capital adjustment accounts for an important share of overall long-run labor demand adjustment,

but that it is unlikely to account for all of the adjustment in the absence of agglomeration.

To quantify the strength of agglomeration economies needed to rationalize the data, we first

need to estimate the inverse labor supply elasticity, η. We do so following (12) by regressing the

1991-2010 change in log formal employment on the change in log regional earnings premium with

RTRr serving as an instrument for wr. The resulting estimate of 0.363 is shown in Panel A of Table

8. Given this value for η, we estimate κ using non-linear least squares based on long-run changes in

regional earnings in (13) or long-run changes in employment in (14). In both cases, the R term is

captured by the intercept, and the regional weighted average price shocks are measured by RTRr.

56Agglomeration estimation exercises regularly require cost share calibrations along these lines, e.g. Kline andMoretti (2014).

57While i) is likely an underestimate because there are fixed inputs other than land (e.g. heavy infrastructure),ii) is likely an overestimate, because some structures depreciate substantially at a 15 year time horizon. Thus, theintermediate value, iii), is our preferred estimate.

26


When estimating equation (14), we include all industries and control for industry price changes

using tariff changes, as in column (3) of Table 7, though the results are nearly identical when using

the alternative approaches in columns (4)-(6) of Table 7. We show estimates for each value of ζ and

bootstrap the entire estimation procedure when calculating standard errors to account for potential

correlation between the η and κ estimates.

The resulting estimates of κ appear in Panel B of Table 8. All of the estimates are positive and

fall within the range of the prior literature (Melo et al. 2009). For example, Kline and Moretti (2014)

find an estimate of 0.2, which is quite close to our wage-based estimate of 0.188 for the intermediate

value of ζ. The value of ζ is important in determining the magnitude of the agglomeration elasticity,

which is unsurprising since Figure 6 showed that capital adjustment explains a smaller share of

overall adjustment for higher values of ζ, leaving a larger role for agglomeration economies.

The estimates in Table 8 and the patterns in Figure 6 show that capital adjustment and standard

agglomeration economies can quantitatively account for the long-run behavior of regional earnings in

response to liberalization. Along with this long-run evidence, Figures 4, 7, and 8 show that regional

labor and capital evolved slowly over time following liberalization and did so in a way consistent with

growing earnings effects of liberalization. In contrast to the other mechanisms that we considered,

dynamic labor demand, driven by slow capital adjustment and agglomeration economies, is both

qualitatively and quantitatively consistent with the earnings responses in Figure 3.

6 Conclusion

This paper documents regional labor market dynamics following the Brazilian trade liberalization

of the early 1990s. Using 25 years of administrative employment data, we find large and growing

effects of trade liberalization on regional formal earnings and employment. Contrary to conventional

wisdom, which assumes wage-equalizing labor adjustment, the regional effects of liberalization grow

for more than a decade before leveling off. This pattern is not driven by post-liberalization economic

shocks and is robust to a wide variety of alternative specifications. After ruling out a number of

potential mechanisms that could generate these growing effects over time, we find strong evidence

in support of a combination of imperfect interregional labor mobility and dynamic labor demand,

driven by slow capital adjustment and agglomeration economies.

Our results have important implications for our thinking about the labor market effects of trade

liberalization. A growing literature has shown in a variety of contexts that trade and trade policy

have heterogeneous effects across regions in the short-run. However, most researchers, ourselves

included, generally assumed that these effects would be upper bounds on the long-run effects, as

labor reallocation would arbitrage away regional differences. This paper finds precisely the opposite.

Short-run effects vastly underestimate the long-run effects, indicating that the costs and benefits

of liberalization remain sharply unevenly distributed across geography, even twenty years after the

27


policy began.

Our empirical results also inform a large and growing literature using structural models of the

labor market to study trade-induced transitional dynamics. We document the importance of re-

gional adjustment to trade liberalization, even in the long run, and highlight margins of adjustment

that have received little attention by this line of work.58 We find evidence for slow capital adjust-

ment in response to trade liberalization, reinforcing the message of Dix-Carneiro (2014) that jointly

quantifying mobility frictions for labor and other factors such as capital is key to understanding

trade adjustment.59 We also find that agglomeration economies are quantitatively important in

accounting for the magnitudes of trade’s effects on regional earnings, suggesting another feature

for inclusion in models examining the effects of trade shocks on labor markets.

58With the exception of Caliendo et al. (2015), this literature has abstracted from geography.59Artuc, Bet, Brambilla and Porto (2014) take an initial step in this direction.

28


References

Abreu, Marcelo de Paiva, “The Political Economy of High Protection in Brazil before 1987,” Inter-AmericanDevelopment Bank Special Initiative on Trade and Integration Working Paper, 2004, (SITI-08A).

Acemoglu, Daron, David Autor, David Dorn, Gordon H. Hanson, and Brendan Price, “Import Compe-tition and the Great US Employment Sag of the 2000s,” Journal of Labor Economics, January 2016, 34 (S1),S141–S198.

Adao, Rodrigo, “Worker Heterogeneity, Wage Inequality, and International Trade: Theory and Evidence fromBrazil,” Unpublished, 2015.

Arbache, Jorge Saba, Andy Dickerson, and Francis Green, “Trade Liberalisation and Wages in DevelopingCountries,” Economic Journal, 2004, 114 (493), F73–F96.

Arcidiacono, Peter, Gigi Foster, Natalie Goodpaster, and Josh Kinsler, “Estimating Spillovers Using PanelData, with an Application to the Classroom,” Quantative Economics, 2012, 3 (3), 421–470.

Artuc, Erhan, German Bet, Irene Brambilla, and Guido Porto, “Trade Shocks and Factor AdjustmentFrictions: Implications for Investment and Labor,” unpublished, 2014.

, Shubham Chaudhuri, and John McLaren, “Trade Shocks and Labor Adjustment: A Structural EmpiricalApproach,” American Economic Review, 2010, 100 (3), 1008–45.

Autor, David, David Dorn, and Gordon Hanson, “The China Syndrome: Local Labor Market Effects of ImportCompetition in the United States,” American Economic Review, 2013, 103 (6).

, , , and Jae Song, “Trade Adjustment: Worker Level Evidence,” Quarterly Journal of Economics,2014, 129 (4), 1799–1860.

Blanchard, Olivier Jean and Lawrence F. Katz, “Regional Evolutions,” Brookings Papers on Economic Activity,1992, (1), 1–75.

Bound, John and Harry J. Holzer, “Demand Shifts, Population Adjustments, and Labor Market Outcomesduring the 1980s,” Journal of Labor Economics, 2000, 18 (1), 20–54.

Burstein, Ariel, Martin Eichenbaum, and Sergio Rebelo, “Large Devaluations and the Real Exchange Rate,”Journal of Political Economy, August 2005, 113 (4), 742–784.

Bustos, Paula, Bruno Caprettini, and Jacopo Ponticelli, “Agricultural Productivity and Structural Transfor-mation. Evidence from Brazil,” American Economic Review, 2016, 106 (6), 1320–1365.

Caliendo, Lorenzo, Maximiliano Dvorkin, and Fernando Parro, “Trade and Labor Market Dynamics,”Unpublished, 2015.

Cosar, Kerem, “Adjusting to Trade Liberalization: Reallocation and Labor Market Policies,” unpublished, 2013.

Costa, Francisco J.M., Jason Garred, and Joao Paulo Pessoa, “Winners and Losers from a Commodities-for-Manufactures Trade Boom,” Journal of International Economics, 2016, 102, 50–69.

Davidson, Russell and James G. MacKinnon, “Bootstrap Methods in Econometrics,” in Terence C. Mills andKerry Patterson, eds., Palgrave Handbook of Econometrics, Vol. 1, Basingstoke: Palgrave MacMillan, 2006,pp. 812–838.

Davis, Steven J. and John Haltiwanger, “Gross Job Creation and Destruction: Microeconomic Evidence andMacroeconomic Implications,” NBER Macroeconomics Annual, 1990, 5.

de Carvalho, Jr., Mario C., “Alguns Aspectos da Reforma Aduaneira Recente,” FUNCEX Texto Para Discussao,1992.

De Negri, Joao Alberto, Paulo Furtado de Castro, Natalia Ribeiro de Souza, and Jorge Saba Arbache,“Mercado Formal de Trabalho: Comparacao entre os Microdados da RAIS e da PNAD,” IPEA Texto ParaDiscussao, 2001, (840).

Dix-Carneiro, Rafael, “Trade Liberalization and Labor Market Dynamics,” Econometrica, 2014, 82 (3).

and Brian K. Kovak, “Trade Liberalization and the Skill Premium: A Local Labor Markets Approach,”American Economic Review - Papers and Proceedings, 2015, 105 (5), 551–557.

and , “Trade Reform and Regional Dynamics: Evidence From 25 Years of Brazilian Matched Employer-Employee Data,” NBER Working Paper, 2015, (w20908).

29


Edmonds, Eric, Nina Pavcnik, and Petia Topalova, “Trade Adjustment and Human Capital Investment:Evidence from Indian Tariff Reform,” American Economic Journal: Applied Economics, 2010, 2 (4), 42–75.

Fajnzylber, Pablo, William F. Maloney, and Gabriel V. Montes-Rojas, “Does formality improve micro-firmperformance? Evidence from the Brazilian SIMPLES program,” Journal of Development Economics, 2011, 94(2), 262 – 276.

Glaeser, Edward L. and Joseph Gyourko, “Urban Decline and Durable Housing,” Journal of Political Economy,2005, 113 (2), 345–375.

Goldberg, Pinelopi and Nina Pavcnik, “Trade, Wages, and the Political Economy of Trade Protection: Evidencefrom the Colombian Trade Reforms,” Journal of International Economics, 2005, 66 (1), 75–105.

and , “Distributional Effects of Globalization in Developing Countries,” Journal of Economic Literature,2007, XLV, 39–82.

Goldberg, Pinelopi Koujianou and Nina Pavcnik, “The response of the informal sector to trade liberalization,”Journal of Development Economics, 2003, 72 (3), 463–496.

Gonzaga, Gustavo, Naercio Menezes Filho, and Cristina Terra, “Trade liberalization and the evolution ofskill earnings differentials in Brazil,” Journal of International Economics, 2006, 68 (2), 345–367.

Hakobyan, Shushanik and John McLaren, “Looking for Local Labor Market Effects of NAFTA,” Review ofEconomics and Statistics, forthcoming.

Hanlon, W. Walker and Antonio Miscio, “Agglomeration: A Long-Run Panel Data Approach,” Unpublished,2016.

Hasan, Rana, Devasish Mitra, and Beyza P. Ural, “Trade Liberalization, Labor-Market Institutions, andPoverty Reduction: Evidence from Indian States,” India Policy Forum, 2006, 3.

, , Priya Ranjan, and Reshad N. Ahsan, “Trade Liberalization and Unemployment: Theory andEvidence from India,” Journal of Development Economics, 2012, 97 (2).

Helm, Ines, “National Industry Trade Shocks, Local Labor Markets, and Agglomeration Spillovers,” Unpublished,2016.

Helpman, Elhanan, Oleg Itskhoki, Marc-Andreas Muendler, and Steven J. Redding, “Trade and Inequal-ity: From Theory to Estimation,” Review of Economic Studies, forthcoming.

Hollweg, Claire H., Daniel Lederman, Diego Rojas, and Elizabeth Ruppert Bulmer, Sticky feet : howlabor market frictions shape the impact of international trade on jobs and wages., Washington D.C.: The WorldBank, 2014.

IBGE, Censo Demografico 2000: Documentacao das Microdados da Amostra, Instituto Brasileiro de Geografia eEstatıstica, 2002.

, Sistema Nacional de Indices de Precos ao Consumidor, Vol. 34 of Serie Relatorios Metodologicos, InstitutoBrasileiro de Geografia e Estatıstica, 2005.

Kambourov, Gueorgui, “Labor Market Regulations and the Sectoral Reallocation of Workers: The Case of TradeReforms,” Review of Economic Studies, 2009, 76 (4), 1321–1358.

Kline, Patrick and Enrico Moretti, “Local Economic Development, Agglomeration Economies and the Big Push:100 Years of Evidence from the Tennessee Valley Authority,” Quarterly Journal of Economics, 2014, 129 (1),275–331.

Kondo, Illenin O., “Trade Reforms, Foreign Competition, and Labor Market Adjustments in the U.S.,” unpublished,2014.

Kovak, Brian, “Regional Efects of Trade Reform: What is the Correct Measure of Liberalization?,” AmericanEconomic Review, 2013, 103 (5), 1960–1976.

Krishna, Pravin, Jennifer P. Poole, and Mine Zeynep Senses, “Wage Effects of Trade Reform with Endoge-nous Worker Mobility,” NBER Working Paper, 2011, (17256).

, , and , “Wage Effects of Trade Reform with Endogenous Worker Mobility,” Journal of InternationalEconomics, 2014, 93 (2), 239–252.

Kugler, Maurice and Eric Verhoogen, “Prices, Plant Size, and Product Quality,” Review of Economic Studies,2011, 79 (1), 307–339.

30


Kume, Honorio, “A Polıtica Tarifaria Brasileira no Perıodo 1980-88: Avaliacao e Reforma,” Serie Epico, March1990, (17).

, Guida Piani, and Carlos Frederico Braz de Souza, “A Polıtica Brasileira de Importacao no Perıodo1987-1998: Descricao e Avaliacao,” in Carlos Henrique Corseuil and Honorio Kume, eds., A Abertura ComercialBrasileira nos Anos 1990: Impactos Sobre Emprego e Salario, Rio de Janiero: MTE/IPEA, 2003, chapter 1,pp. 1–37.

LaPorta, Rafael and Andrei Schleifer, “The Unofficial Economy and Economic Development,” Brookings Paperson Economic Activity, 2008, 47 (1), 123–135.

and , “Informality and Development,” Journal of Economic Perspectives, Summer 2014, 28 (3), 109–126.

Lopes de Melo, Rafael, “Firm Wage Differentials and Labor Market Sorting: Reconciling Theory and Evidence,”unpublished, 2013.

MacKinnon, James G., “Thirty Years of Heteroskedasticity-Robust Inference,” Queen’s Economics DepartmentWorking Paper, 2011, (1268).

McCaig, Brian, “Exporting Out of Poverty: Provincial Poverty in Vietnam and US Market Access,” Journal ofInternational Economics, 2011, 85 (1).

McKenzie, David and Ernesto Schargrodsky, “Buying less, but shopping more: Changes in consumptionpatterns during a crisis,” Economıa, 2011, 11 (2), 1–35.

Meghir, Costas, Renata Narita, and Jean-Marc Robin, “Wages and Informality in Developing Countries,”American Economic Review, 2015, 105 (4), 1509–1546.

Melo, Patricia C., Daniel J. Graham, and Robert B. Noland, “A Meta-Analysis of Estimates of UrbanAgglomeration Economies,” Regional Science and Urban Economics, 2009, 39, 332–342.

Menezes-Filho, Naercio and Marc-Andreas Muendler, “Labor Reallocation in Response to Trade Reform,”NBER Working Paper, 2011, (17372).

Monte, Ferdinando, “The Local Incidence of Trade Shocks,” Unpublished, 2016.

Moretti, Enrico, “Real Wage Inequality,” American Economic Journal: Applied Economics, 2013, 5 (1), 65–103.

Neri, Marcelo and Rodrigo Moura, “Brasil: La institucionalidad del salario mınimo,” in Andres Marinakis andJuan Jacobo Velasco, eds., Para que sirve el salario mınimo?, Organizacion Internacional del Trabajo, 2006,pp. 105–158.

Notowidigdo, Matthew J., “The Incidence of Local Labor Demand Shocks,” Unpublished, 2013.

Pavcnik, Nina, Andreas Blom, Pinelopi Goldberg, and Norbert Schady, “Trade Liberalization and IndustryWage Structure: Evidence from Brazil,” World Bank Economic Review, 2004, 18 (3), 319–334.

Paz, Lourenco, “The impacts of trade liberalization on informal labor markets: an evaluation of the Brazilian case,”Journal of International Economics, 2014, 92 (2), 330–348.

Saboia, Joao L. M. and Ricardo M. L. Tolipan, “A relacao anual de informacoes sociais (RAIS) e o mercadoformal de trabalho no Brasil: uma nota,” Pesquisa e Planejamento Economico, 1985, 15 (2), 447–456.

Salem, Samira and John Benedetto, “The USITC’s Roundtable on the Labor Market Effects of Trade: DiscussionSummary,” Journal of International Commerce and Economics, August 2013.

Schor, Adriana, “Heterogeneous productivity response to tariff reduction. Evidence from Brazilian manufacturingfirms,” Journal of Development Economics, 2004, 75 (2), 373–396.

Soares, Rodrigo R. and Guilherme Hirata, “Competition and the Racial Wage Gap: Testing Becker’s Modelof Employer Discrimination,” IZA Discussion Paper, Februray 2016, (9764).

Stock, James H and Motohiro Yogo, “Testing for Weak Instruments in Linear IV Regression,” in Donald W. K.Andrews and James H. Stock, eds., Identification and Inference for Econometric Models: Essays in Honor ofThomas Rothenberg, Cambridge University Press, 2005, pp. 80–108.

Topalova, Petia, “Trade Liberalization, Poverty, and Inequality: Evidence from Indian Districts,” in Ann Harrison,ed., Globalization and Poverty, University of Chicago Press, 2007, pp. 291–336.

, “Factor Immobility and Regional Impacts of Trade Liberalization: Evidence on Poverty from India,” AmericanEconomic Journal: Applied Economics, 2010, 2 (4).

31


Traiberman, Sharon, “Occupations and Import Competition,” Unpublished, 2016.

Ulyssea, Gabriel, “Firms, Informality and Development: Theory and evidence from Brazil,” unpublished, 2014.

Valentinyi, Akos and Berthold Herrendorf, “Measuring Factor Income Shares at the Sectoral Level,” Reviewof Economic Dynamics, 2008, 11, 820–835.

32


Figure 1: Tariff Changes

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

Chan

ge in

ln(1

+tar

iff),

1990

-95

Agric

ultu

re

Met

als

Appa

rel

Food

Pro

cess

ing

Woo

d, F

urni

ture

, Pea

t

Text

iles

Nonm

etal

lic M

iner

al M

anuf

Pape

r, Pu

blish

ing,

Prin

ting

Min

eral

Min

ing

Foot

wear

, Lea

ther

Chem

icals

Auto

, Tra

nspo

rt, V

ehicl

es

Elec

tric,

Ele

ctro

nic

Equi

p.

Mac

hine

ry, E

quip

men

t

Plas

tics

Oth

er M

anuf

.

Phar

ma.

, Per

fum

es, D

eter

gent

s

Petro

leum

Refi

ning

Rubb

er

Petro

leum

, Gas

, Coa

l

Tariff data from Kume et al. (2003), aggregated to allow consistent industry definitions across data sources. SeeAppendix Table A1 for details of the industry classification. Industries sorted based on 1991 national employment(largest on the left, and smallest on the right)

33


Figure 2: Regional Tariff Reductions

BelémBelém

RecifeRecife

ManausManaus

CuritibaCuritiba

BrasíliaBrasília

SalvadorSalvador

FortalezaFortaleza

São PauloSão Paulo

Porto AlegrePorto Alegre

Belo HorizonteBelo Horizonte

8% to 15%4% to 8%3% to 4%1% to 3%-1% to 1%

mean 10 25 50 75 900.044 0.002 0.012 0.031 0.066 0.107

percentile

Local labor markets reflect microregions defined by IBGE, aggregated slightly to account for border changes between1986 and 2010. Regions are colored based on the regional tariff reduction measure, RTRr, defined in (2). Regionsfacing larger tariff reductions are presented as lighter and yellower, while regions facing smaller cuts are shown asdarker and bluer. Dark lines represent state borders, gray lines represent consistent microregion borders, and cross-hatched migroregions are omitted from the analysis. These microregions were either i) part of a Free Trade Area ii)part of the state of Tocantins and not consistently identifiable over time, or iii) not included in the RAIS samplebefore 1990.

34


Figure 3: Regional log Formal Earnings Premia - 1987-2010

-‐2.0

-‐1.5

-‐1.0

-‐0.5

0.0

0.5

1.0

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Pre-‐liberaliza6on (chg. from 1986)

Liberaliza6on Post-‐liberaliza6on (chg. from 1991)

Each point reflects an individual regression coefficient, θt, following (3), where the dependent variable is the changein regional log formal earnings premium and the independent variable is the regional tariff reduction (RTRr), definedin (2). Note that RTRr always reflects tariff reductions from 1990-1995. For blue circles, the earnings changes arefrom 1991 to the year listed on the x-axis. For purple diamonds, the changes are from 1986 to the year listed. Allregressions include state fixed effects, and post-liberalization regressions control for the 1986-1990 outcome pre-trend.Negative estimates imply larger earnings declines in regions facing larger tariff reductions. Vertical bars indicate thatliberalization began in 1991 and was complete by 1995. Dashed lines show 95 percent confidence intervals. Standarderrors adjusted for 112 mesoregion clusters.

35


Figure 4: Regional log Formal Employment - 1987-2010

-‐6.5

-‐5.5

-‐4.5

-‐3.5

-‐2.5

-‐1.5

-‐0.5

0.5

1.5

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010





Each point reflects an individual regression coefficient, θt, following (3), where the dependent variable is the changein regional log formal employment and the independent variable is the regional tariff reduction (RTRr), defined in(2). Note that RTRr always reflects tariff reductions from 1990-1995. For blue circles, the employment changes arefrom 1991 to the year listed on the x-axis. For purple diamonds, the changes are from 1986 to the year listed. Allregressions include state fixed effects, and post-liberalization regressions control for the 1986-1990 outcome pre-trend.Negative estimates imply larger employment declines in regions facing larger tariff reductions. Vertical bars indicatethat liberalization began in 1991 and was complete by 1995. Dashed lines show 95 percent confidence intervals.Standard errors adjusted for 112 mesoregion clusters.

36


Figure 5: Regional Imports, Exports, and Net Exports Per Worker - 1991-2010

-‐0.4

-‐0.2

0

0.2

0.4

0.6

0.8

1

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Imports

Exports

Net Exports


Each point reflects an individual regression coefficient, θt, following (3), where the dependent variable is the changein regional imports per worker (blue circles), exports per worker (red triangles), or net exports per worker (greendiamonds), measured in $100,000 units. The independent variable is the regional tariff reduction (RTRr), definedin (2). Note that RTRr always reflects tariff reductions from 1990-1995. All regressions include state fixed effects,but do not include pre-liberalization trends due to a lack of Comtrade trade data before 1989. Positive estimatesimply larger increases in trade flow per worker in regions facing larger tariff reductions. Vertical bar indicates thatliberalization was complete by 1995. Dashed lines show 95 percent confidence intervals. Standard errors adjusted for112 mesoregion clusters.

37


Figure 6: Inferred Adjustment and Capital Adjustment Quantification - 1992-2010

-‐3.0

-‐2.5

-‐2.0

-‐1.5

-‐1.0

-‐0.5

0.0

0.5

1.0

1.5

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Pre-‐liberaliza6on (chg. from 1986) Liberaliza6on Post-‐liberaliza6on

(chg. from 1991)

Inferred Adjustment

Capital (establishments) adjustment

ζ = 0.152

ζ = 0.349

ζ = 0.545

Each point reflects an individual regression coefficient, θt, following (3). For the blue profile with solid circles, thedependent variable is the inferred labor demand shifts from agglomeration and capital adjustment, defined in (10).For the gray profiles with hollow markers, the dependent variable is capital’s contribution to overall adjustment, usingthe change in the number of regional formal establishments as a proxy for the change in regional capital,

∑i λriKri.

We present profiles for three values of ζ, specific factors’ share of non-labor inputs, based on Valentinyi and Herrendorf(2008). The independent variable is the regional tariff reduction (RTRr), defined in (2). Note that RTRr alwaysreflects tariff reductions from 1990-1995. All regressions include state fixed effects, and post-liberalization regressionscontrol for the 1986-1990 outcome pre-trend. Negative estimates imply larger declines in residual labor demand orthe number of establishments in regions facing larger tariff reductions. Vertical bar indicates that liberalization wascomplete by 1995. Dashed lines show 95 percent confidence intervals. Confidence intervals for capital adjustmentprofiles shown in Appendix B.12. Standard errors adjusted for 112 mesoregion clusters.

38


Figure 7: Regional log Number of Formal Establishments and log Average Formal EstablishmentSize (Number of Workers) - 1987-2010

-‐5

-‐4

-‐3

-‐2

-‐1

0

1

2

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010



Establishment Size

Establishments


Estab. Size Pretrend

Establishments Pretrend

Each point reflects an individual regression coefficient, θt, following (3), where the dependent variable is the changein regional log number of formal establishments or the change in regional log average formal establishment size.The independent variable is the regional tariff reduction (RTRr), defined in (2). Note that RTRr always reflectstariff reductions from 1990-1995. For blue circles and red triangles, the changes are from 1991 to the year listed onthe x-axis. For purple diamonds and orange squares, the changes are from 1986 to the year listed. All regressionsinclude state fixed effects, and post-liberalization regressions control for the 1986-1990 outcome pre-trend. Negativeestimates imply larger declines in the number of establishments or average establishment size in regions facing largertariff reductions. Vertical bars indicate that liberalization began in 1991 and was complete by 1995. Dashed linesshow 95 percent confidence intervals. Standard errors adjusted for 112 mesoregion clusters.

39


Figure 8: Regional log Cumulative Formal Establishment Entry and Exit - 1987-2010

-‐3

-‐2

-‐1

0

1

2

3

4

5

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Exit

Entry



Exit Pretrend

Entry Pretrend

Exit

Entry


Exit Pretrend

Entry Pretrend

Each point reflects an individual regression coefficient, θt, following (3). The dependent variable is the log cumulativeformal establishment entry or exit from 1991 to the year listed on the x-axis (blue circles and red triangles) or from1986 to the year listed (purple diamonds and orange squares), calculated as in Davis and Haltiwanger (1990). Theindependent variable is the regional tariff reduction (RTRr), defined in (2). Note that RTRr always reflects tariffreductions from 1990-1995. All regressions include state fixed effects, and post-liberalization regressions control forlog cumulative establishment entry or exit during 1986-1990. Positive exit estimates and negative entry estimatesimply larger rates of exit and smaller rates of entry in regions facing larger tariff reductions. Vertical bars indicatethat liberalization began in 1991 and was complete by 1995. Dashed lines show 95 percent confidence intervals.Standard errors adjusted for 112 mesoregion clusters.

40


Figure 9: Regional log Cumulative Job Creation and Destruction - 1987-2010

-‐4

-‐3

-‐2

-‐1

0

1

2

3

4

5

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Job Crea)on

Job Destruc)on



Job Crea)on Pretrend

Job Destruc)on Pretrend

Each point reflects an individual regression coefficient, θt, following (3). The dependent variable is the log cumulativejob creation or destruction rate from 1991 to the year listed on the x-axis (blue circles and red triangles) or from1986 to the year listed (purple diamonds and orange squares), calculated as in Davis and Haltiwanger (1990). Theindependent variable is the regional tariff reduction (RTRr), defined in (2). Note that RTRr always reflects tariffreductions from 1990-1995. All regressions include state fixed effects, and post-liberalization regressions control forlog cumulative job creation or destruction during 1986-1990. Positive job destruction estimates and negative jobcreation estimates imply larger rates of job destruction and smaller rates of job creation in regions facing larger tariffreductions. Vertical bars indicate that liberalization began in 1991 and was complete by 1995. Dashed lines show 95percent confidence intervals. Standard errors adjusted for 112 mesoregion clusters.

41


Table 1: Descriptive Statistics

Panel A: Liberalization Shock 1991-1995

Regional tariff reductions (RTRr) 0.044(0.039)

Panel B: Main Outcome Variables 1991-1995 1991-2000 1991-2005 1991-2010

Change in log formal earnings premium 0.258 0.305 0.401 0.712(0.161) (0.174) (0.189) (0.201)

Change in log informal earnings premiuma -0.050 0.161(0.135) (0.197)

Change in log formal employment 0.268 0.599 0.976 1.308(0.377) (0.549) (0.576) (0.614)

Change in log informal employmenta 0.269 0.291(0.162) (0.228)

Change in log num. formal establishments 0.358 0.728 1.055 1.271(0.230) (0.318) (0.389) (0.444)

Change in log avg. formal establishment size -0.180 -0.260 -0.220 -0.128(0.279) (0.387) (0.375) (0.384)

Change in log formal job destruction -1.014 -0.824 -0.966 -1.135(0.398) (0.421) (0.466) (0.516)

Change in log formal job creation -0.608 -0.099 0.143 0.299(0.387) (0.270) (0.221) (0.176)

Change in log formal establishment exit -1.206 -1.092 -1.183 -1.305(0.226) (0.285) (0.343) (0.405)

Change in log formal establishment entry -0.397 0.052 0.251 0.366(0.287) (0.175) (0.139) (0.114)

Change in log working-age populationa 0.198 0.388(0.103) (0.178)

Panel C: Region Characteristics 1991 1995 2000 2005 2010

Average Formal Earnings (2010 R$) 755.98 1,105.83 944.18 939.93 1,152.40(273.08) (394.71) (323.97) (480.00) (469.95)

Formal Employment 30,466 34,929 40,100 51,631 70,170(152267) (161657) (163917) (197206) (269602)

Share Agriculture/Mininga 0.397(0.194)

Share Manufacturinga 0.113(0.077)

475 microregion observations. See the text for descriptions of all measures. a Calculated using the Census, soestimates are available only for 1991, 2000, and 2010.

42


Table 2: Regional log Formal Earnings Premia and Employment - 2000, 2010

Change in outcome: (1) (2) (3) (4) (5) (6)

Panel A: log Formal Earnings PremiaRegional tariff reduction (RTR) -0.451*** -0.638*** -0.529*** -1.885*** -1.736*** -1.594***

(0.152) (0.154) (0.141) (0.316) (0.184) (0.169)Formal earnings pre-trend (86-90) -0.312** -0.418***

(0.149) (0.144)State fixed effects (26) ✓ ✓ ✓ ✓

R-squared 0.040 0.225 0.268 0.320 0.501 0.537

Panel B: log Formal Real Earnings Premia (regional deflators following Moretti (2013))Regional tariff reduction (RTR) -1.594*** -1.382*** -1.260***

(0.306) (0.180) (0.168)Formal earnings pre-trend (86-90) -0.359***

(0.133)State fixed effects (26) ✓ ✓

R-squared 0.238 0.449 0.477

Panel C: log Formal EmploymentRegional tariff reduction (RTR) -3.748*** -3.545*** -3.533*** -6.059*** -4.675*** -4.663***

(0.516) (0.563) (0.582) (0.560) (0.660) (0.679)Formal employment pre-trend (86-90) -0.0331 -0.0319


R-squared 0.072 0.291 0.291 0.149 0.409 0.410

1991-2000 1991-2010

Negative coefficient estimates for the regional tariff reduction imply larger declines in formal earnings or employmentin regions facing larger tariff reductions. Microregion observations: Panels A and C, 475; Panel B, 456 (omits a fewsparsely populated locations with insufficient data to calculate regional price deflators). Regional earnings premiacalculated controlling for age, sex, education, and industry of employment. Panels A and B: efficiency weightedby the inverse of the squared standard error of the estimated change in log formal earnings premium. Pre-trendscomputed for 1986-1990. Standard errors (in parentheses) adjusted for 112 mesoregion clusters. *** Significant atthe 1 percent, ** 5 percent, * 10 percent level.

43


Table 3: Regional log Working-Age Population - 2000, 2010

Change in log Working-Age Population: (1) (2) (3) (4) (5) (6)

Regional tariff reduction (RTR) 0.333 -0.061 0.018 0.392 -0.175 -0.059(0.243) (0.330) (0.204) (0.319) (0.473) (0.294)

Population pre-trend (80-91) 0.406** 0.328* 0.632*** 0.531**(0.164) (0.171) (0.225) (0.235)

Population pre-trend (70-80) 0.297*** 0.137*** 0.445*** 0.190**(0.072) (0.047) (0.087) (0.073)

State fixed effects (26) ✓ ✓ ✓ ✓ ✓ ✓

R-squared 0.654 0.557 0.678 0.666 0.554 0.685

1991-2000 1991-2010

Positive (negative) coefficient estimates for the regional tariff reduction imply larger increases (decreases) in populationin regions facing larger tariff reductions. Outcomes calculated using Census data. 405 microregion observations.Efficiency weighted by the inverse of the squared standard error of the dependent variable estimate. Pre-trendscomputed for 1980-1991 and 1970-1980. Standard errors (in parentheses) adjusted for 90 mesoregion clusters. ***Significant at the 1 percent, ** 5 percent, * 10 percent level.

44


Table 4: Regional log Informal Employment and Earnings Premia - 2000, 2010


Panel A: log Informal EmploymentRegional tariff reduction (RTR) 2.017*** 1.706*** 1.593*** 2.122*** 1.448*** 1.196*

(0.431) (0.344) (0.532) (0.468) (0.491) (0.705)Informal employment pre-trend (80-91) 0.069 0.050 0.149 0.109

(0.115) (0.114) (0.132) (0.126)All employment pre-trend (70-80) 0.121** 0.110** 0.263*** 0.239***

(0.056) (0.044) (0.080) (0.063)State fixed effects (26) ✓ ✓ ✓ ✓ ✓ ✓

R-squared 0.579 0.589 0.592 0.524 0.552 0.562

Panel B: log Informal Earnings PremiaRegional tariff reduction (RTR) -0.027 -0.217 -0.034 0.352 0.054 0.338

(0.161) (0.160) (0.163) (0.256) (0.298) (0.251)Informal earnings pre-trend (80-91) -0.191*** -0.193*** -0.288*** -0.291***

(0.049) (0.048) (0.086) (0.084)All workers' earnings pre-trend (70-80) 0.008 -0.016 0.001 -0.035

(0.064) (0.060) (0.109) (0.102)State fixed effects (26) ✓ ✓ ✓ ✓ ✓ ✓

R-squared 0.676 0.654 0.676 0.690 0.667 0.690

1991-2000 1991-2010

Positive (negative) coefficient estimates for the regional tariff reduction imply larger increases (declines) in informalearnings or employment in regions facing larger tariff reductions. Outcomes calculated using Census data. 405microregion observations. Regional earnings premia calculated controlling for age, sex, education, and industry ofemployment. Efficiency weighted by the inverse of the squared standard error of the dependent variable estimate.Pre-trends computed for 1980-1991 and 1970-1980. Standard errors (in parentheses) adjusted for 112 mesoregionclusters. *** Significant at the 1 percent, ** 5 percent, * 10 percent level.

45


Table 5: Mechanisms: Changing Worker Composition - 1995, 2000, 2005, 2010

Change in log Formal Earnings Premia: 1991-1995 1991-2000 1991-2005 1991-2010(1) (2) (3) (4)

Panel A: Main specificationRegional tariff reduction (RTR) -0.096 -0.529*** -1.294*** -1.594***

(0.120) (0.141) (0.139) (0.169)Panel B: Earnings premia controlling for individual fixed effects (fixed returns)

Regional tariff reduction (RTR) -0.193* -0.514*** -1.119*** -1.271***(0.115) (0.144) (0.147) (0.172)

Panel C: Earnigns premia controlling for individual fixed effects (time-varying returns)Regional tariff reduction (RTR) -0.230** -0.551*** -1.322*** -1.454***

(0.093) (0.098) (0.094) (0.119)Formal earnings pre-trend (86-90) ✓ ✓ ✓ ✓State fixed effects (26) ✓ ✓ ✓ ✓

Negative coefficient estimates for the regional tariff reduction (RTRr) imply larger declines in formal earnings inregions facing larger tariff reductions. Microregion observations: Panel A, 475; Panels B and C, 450 (omits regionswith insufficient observations to identify region-year fixed effects in any particular year). Regional earnings premia:Panel A: calculated controlling for age, sex, education, and industry of employment; Panels B and C: controlling forindividual fixed effects. Standard errors (in parentheses) adjusted for 112 mesoregion clusters. Efficiency weightedby the inverse of the squared standard error of the estimated change in log formal earnings premium. See text fordetailed description of each panel. *** Significant at the 1 percent, ** 5 percent, * 10 percent level.

46


Table 6: Mechanisms: Slow Response of Imports or Exports - 1995, 2000, 2005, 2010



(0.120) (0.141) (0.139) (0.169)

Panel B: Controls for trade quantities (OLS)Regional tariff reduction (RTR) -0.089 -0.521*** -1.287*** -1.562***

(0.112) (0.138) (0.181) (0.221)Import quantity control

Export quantity control

Panel C: Latin America IVRegional tariff reduction (RTR) -0.129 -0.569*** -1.342*** -1.757***



First-stage F (Kleibergen-Paap)

Panel D: Colombia IV Regional tariff reduction (RTR) -0.049 -0.488*** -1.372*** -1.502***



First-stage F (Kleibergen-Paap)

Formal earnings pre-trend (86-90) ✓ ✓ ✓ ✓State fixed effects (26) ✓ ✓ ✓ ✓

(3.268)876.2

1.668

-0.149(3.861)93.04

-3.489(2.427)5.379*

-0.382(2.242)0.142

(3.355)

(2.631)

Negative coefficient estimates for the regional tariff reduction (RTRr) imply larger declines in formal earnings inregions facing larger tariff reductions. Panel A replicates the earnings results in columns (3) and (6) of Table2. Panels B-D include regional import and export quantity controls as in (7). We instrument for the potentiallyendogenous import and export controls using regional measures of commodity price growth from Adao (2015) andwith regional trade flows for other countries. “Latin America” consists of Argentina, Chile, Colombia, Paraguay,Peru, and Uruguay. We measure imports and exports between Latin America or Colombia and the rest of the worldexcluding Brazil. Due to Comtrade data availability, changes in Colombian trade flows are measured from 1991 toeach subsequent year and Latin American trade flows from 1994. We allow for time-varying first-stage coefficients,so we have 2 endogenous variables (RegImprt and RegExprt) and 57 instruments for Colombia (3 instruments × 19years) and 48 instruments for Latin America (3 instruments × 16 years). First-stage Kleinbergen-Paap F statistics arecompared to the Stock and Yogo (2005) critical value of 21 to reject 5 percent bias relative to OLS. Standard errors(in parentheses) adjusted for 112 mesoregion clusters. Efficiency weighted by the inverse of the squared standarderror of the estimated change in log formal earnings premium. *** Significant at the 1 percent, ** 5 percent, * 10percent level.

47


Table 7: Test for Agglomeration Economies

Change in log Region × IndustryEmployment: (1) (2) (3) (4) (5) (6)

Regional tariff reduction (RTR) -7.751*** -6.084*** -6.183*** -6.333*** -6.708*** -6.704***(0.625) (0.623) (0.631) (0.646) (0.675) (0.694)

Industry tariff reduction -1.790*** -1.666*** -1.669*** -2.017***(0.294) (0.290) (0.291) (0.332)

Formal employment pre-trend (86-90) -0.106*** -0.147*** -0.110*** -0.150***(0.036) (0.032) (0.037) (0.032)

Industry fixed effects (20) ✓ ✓

State fixed effects (26) ✓ ✓ ✓ ✓ ✓

All Industries Tradable Industries

Negative coefficient estimates for the regional tariff reduction imply the presence of agglomeration economies, following(15). Observations represent region × industry pairs. The dependent variable is the change in log formal employmentin a given region × industry pair from 1991 to 2010. Columns (1) - (4) cover all industries, including the nontradablesector, while columns (5) and (6) restrict attention to tradable industries. For tradable industries, industry tariffreductions are given by the decline in the log of one plus the tariff rate. For the nontradable sector, the industrytariff reduction is measured using RTRr. Standard errors (in parentheses) adjusted for 112 mesoregion clusters. ***Significant at the 1 percent, ** 5 percent, * 10 percent level.

48


Table 8: Agglomeration Elasticity Estimates

Panel A: Inverse labor supply elasticity (η) 0.363***(0.060)

Panel B: Agglomeration elasticity (κ)(1) (2) (3)

Specific factors' share of non-labor inputs (ζ): low (0.152) mid (0.349) high (0.545)

Wage-based agglomeration elasticity (κ) 0.042* 0.188*** 0.333***(0.023) (0.023) (0.025)

Employment-based agglomeration elasticity (κ) 0.215*** 0.330*** 0.461***(0.032) (0.038) (0.043)

Labor supply elasticity, η, estimated from (12) using RTRr as an instrument for the change in regional log earningspremium. The first-stage partial F-statistic (Kleibergen-Paap) for this regression is 59.14. Given the estimate of η,the agglomeration elasticity, κ, is estimated using two alternative methods. The earnings-based approach estimates(13), and the employment-based approach estimates (14), both using nonlinear least squares, and both including1986-1990 pre-liberalization outcome trends and state fixed effects. The employment-based estimates control forindustry price changes as in column (3) of Table 7, and results using other approaches are very similar. We presentestimates for three different values of ζ, specific factors’ share of non-labor inputs, based on Valentinyi and Herrendorf(2008). See text for details. Standard errors (in parentheses) bootstrapped by regional resampling. *** Significantat the 1 percent, ** 5 percent, * 10 percent level.

49


Online Appendices

(Not for publication)

A Data and Definitions 51A.1 Tariffs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51A.2 RAIS Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54A.3 Demographic Census . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54A.4 Regional Tariff Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55A.5 Local Price Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58A.6 Regional Change in Imports and Exports . . . . . . . . . . . . . . . . . . . . . . . . 59A.7 Entry, Exit, Job Creation, and Job Destruction . . . . . . . . . . . . . . . . . . . . . 59

B Supplemental Empirical Results 61B.1 Industry-Level Outcome Pre-Trends vs. Tariff Reductions . . . . . . . . . . . . . . . 61B.2 Informal Employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63B.3 Regional Earnings Premium Regressions . . . . . . . . . . . . . . . . . . . . . . . . . 66B.4 Industry-Region Earnings Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68B.5 Formal Earnings Regression Scatterplots . . . . . . . . . . . . . . . . . . . . . . . . . 70B.6 Census Earnings, Wage, and Employment Results . . . . . . . . . . . . . . . . . . . 72B.7 Robustness Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73B.8 Potential Confounders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76B.9 Earnings and Employment Sample Splits . . . . . . . . . . . . . . . . . . . . . . . . . 86B.10 Regional Change in log Imports and Exports . . . . . . . . . . . . . . . . . . . . . . 88B.11 Overall Employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94B.12 Capital Adjustment Confidence Intervals . . . . . . . . . . . . . . . . . . . . . . . . . 95B.13 Exit by Establishment Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

C Model 100C.1 Baseline Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100C.2 Agglomeration Economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

50


A Data and Definitions

A.1 Tariffs

Tariff data come from Kume et al. (2003), who report nominal tariffs and effective rates of pro-tection from 1987 to 1998 using the Brazilian industry classification Nıvel 50. We aggregate thesetariffs slightly to an industry classification that is consistent with the Demographic Census dataused to construct local tariff shock measures. The classification is presented in Table A1. In ag-gregating, we weight each Nıvel 50 industry by its 1990 industry value added, as reported in IBGENational Accounts data. Figure A1 shows the evolution of nominal tariffs from 1987 to 1998 forthe ten largest industries. The phases of Brazilian liberalization are visible (see Section 2 for adiscussion and citations). Large nominal tariff cuts from 1987-1989 had little effect on protection,due to the presence of substantial nontariff barriers and tariff exemptions. In 1990, the majority ofnontariff barriers and tariff exemptions were abolished, being replaced by tariffs providing equiva-lent protection; note the increase in tariffs in some industries in 1990. During liberalization, from1990 to 1994, tariffs fell in all industries, then were relatively stable from 1995 onward.

In Section 4.2 we calculate post-liberalization tariff changes using UNCTAD TRAINS. SeeAppendix B.8 for details.

51


Tab

leA

1:C

onsi

sten

tIn

du

stry

Cla

ssifi

cati

onA

cros

sC

ensu

ses

and

Tar

iffD

ata

Indu

stry

Indu

stry

Nam

eN

ível

50

1970

, 198

0, 1

991

Cen

sus (atividade)

2000

, 201

0 C

ensu

s (CNAE-Dom

)1

Agr

icul

ture

101

1-03

7, 0

41, 0

42, 5

8111

01-1

118,

120

1-12

09, 1

300,

140

1, 1

402,

200

1, 2

002,

50

01, 5

002

2M

iner

al M

inin

g (e

xcep

t com

bust

ible

s)2

050,

053

-059

1200

0, 1

3001

, 130

02, 1

4001

-140

043

Petro

leum

and

Gas

Ext

ract

ion

and

Coa

l Min

ing

305

1-05

210

000,

110

004

Non

met

allic

Min

eral

Goo

ds M

anuf

actu

ring

410

026

010,

260

91, 2

6092

5Ir

on a

nd S

teel

, Non

ferr

ous,

and

Oth

er M

etal

Pro

duct

ion

and

Proc

essi

ng5-

711

027

001-

2700

3, 2

8001

, 280

028

Mac

hine

ry, E

quip

men

t, C

omm

erci

al In

stal

latio

n M

anuf

actu

ring,

and

Tra

ctor

Man

ufac

turin

g8

120

2900

110

Elec

trica

l, El

ectro

nic,

and

Com

mun

icat

ion

Equi

pmen

t and

Com

pone

nts M

anuf

actu

ring

10-1

113

029

002,

300

00, 3

1001

, 310

02, 3

2000

, 330

0312

Aut

omob

ile, T

rans

porta

tion,

and

Veh

icle

Par

ts M

anuf

actu

ring

12-1

314

034

001-

3400

3, 3

5010

, 350

20, 3

5030

, 350

9014

Woo

d Pr

oduc

ts, F

urni

ture

Man

ufac

turin

g, a

nd P

eat P

rodu

ctio

n14

150,

151

, 160

2000

0, 3

6010

15Pa

per M

anuf

actu

ring,

Pub

lishi

ng, a

nd P

rintin

g15

170,

290

2100

1, 2

1002

, 220

0016

Rub

ber P

rodu

ct M

anuf

actu

ring

1618

025

010

17C

hem

ical

Pro

duct

Man

ufac

turin

g17

,19

200

2301

0, 2

3030

, 234

00, 2

4010

, 240

9018

Petro

leum

Ref

inin

g an

d Pe

troch

emic

al M

anuf

actu

ring

1820

1, 2

02, 3

52, 4

7723

020

20Ph

arm

aceu

tical

Pro

duct

s, Pe

rfum

es a

nd D

eter

gent

s Man

ufac

turin

g20

210,

220

2402

0, 2

4030

21Pl

astic

s Pro

duct

s Man

ufac

turin

g21

230

2502

022

Text

iles M

anuf

actu

ring

2224

0, 2

4117

001,

170

0223

App

arel

and

App

arel

Acc

esso

ries M

anuf

actu

ring

2325

0,53

218

001,

180

0224

Foot

wea

r and

Lea

ther

and

Hid

e Pr

oduc

ts M

anuf

actu

ring

2419

0, 2

5119

011,

190

12, 1

9020

25Fo

od P

roce

ssin

g (C

offe

e, P

lant

Pro

duct

s, M

eat,

Dai

ry, S

ugar

, Oils

, Bev

erag

es, a

nd O

ther

)25

-31

260,

261

, 270

, 280

1501

0, 1

5021

, 150

22, 1

5030

, 150

41-1

5043

, 150

50, 1

6000

32M

isce

llane

ous O

ther

Pro

duct

s Man

ufac

turin

g32

300

3300

1, 3

3002

, 330

04, 3

3005

, 360

90, 3

7000

91U

tiliti

es33

351,

353

4001

0, 4

0020

, 410

0092

Con

stru

ctio

n34

340,

524

4500

1-45

005

93W

hole

sale

and

Ret

ail T

rade

3541

0-42

4, 5

82, 5

8350

010,

500

30, 5

0040

, 500

50, 5

3010

,530

20, 5

3030

, 530

41,

5304

2, 5

3050

, 530

61-5

3068

, 530

70, 5

3080

, 530

90, 5

3101

, 53

102,

550

2094

Fina

ncia

l Ins

titut

ions

3845

1-45

3, 5

85, 6

1265

000,

660

00, 6

7010

, 670

2095

Rea

l Est

ate

and

Cor

pora

te S

ervi

ces

40, 4

146

1-46

4, 5

43, 5

52, 5

71-5

78, 5

84, 5

8963

022,

700

01, 7

1020

, 720

10, 7

4011

, 740

12, 7

4021

, 740

22,

7403

0, 7

4040

, 740

50, 7

4090

, 920

13, 9

2014

, 920

15, 9

2020

96Tr

ansp

orta

tion

and

Com

mun

icat

ions

36, 3

747

1-47

6, 4

81, 4

82, 5

8860

010,

600

20, 6

0031

, 600

32, 6

0040

, 600

91, 6

0092

, 610

00,

6200

0, 6

3010

, 630

21 ,6

4010

,640

20, 9

1010

97Pr

ivat

e Se

rvic

es39

, 43

511,

512

, 521

-523

, 525

, 531

, 533

, 541

, 542

. 544

, 54

5, 5

51, 5

77, 5

86, 5

87, 6

13-6

19, 6

22-6

24, 6

32, 9

01,

902

1500

, 500

20, 5

3111

, 531

12, 5

3113

, 550

10, 5

5030

, 630

30,

7000

2, 7

1010

, 710

30, 7

2020

, 730

00, 7

4060

, 800

11, 8

0012

, 80

090,

850

11, 8

5012

, 850

13, 8

5020

, 850

30, 9

0000

, 910

20,

9109

1, 9

1092

, 920

11, 9

2012

, 920

30, 9

2040

, 930

10, 9

3020

, 93

030,

930

91, 9

3092

, 950

0098

Publ

ic A

dmin

istra

tion

4235

4, 6

10, 6

11, 6

21, 6

31, 7

11-7

17, 7

21-7

2775

011-

7501

7, 7

5020

NontradableTradable

Consi

sten

tin

dust

rycl

ass

ifica

tion

use

din

gen

erati

ng

loca

lta

riff

shock

sfr

om

Nıv

el50

tari

ffdata

inK

um

eet

al.

(2003)

and

Dec

ennia

lC

ensu

sdata

.

52


Figure A1: Tariffs - 1987-1998

Tex$les

Auto, Transport, Vehicles

Food Processing Nonmetallic Mineral Manuf

Electric, Electronic Equip.

Machinery, Equipment Metals

Agriculture

Chemicals

Petroleum Refining

0

10

20

30

40

50

60

70

80

90

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998

Nominal tariffs from Kume et al. (2003), aggregated to the industry classification presented in Table A1. The tenlargest industries by 1990 value added are shown.

53


A.2 RAIS Data

The Relacao Anual de Informacoes Sociais (RAIS) is a high quality census of the Brazilian formallabor market. Originally, RAIS was created as an operational tool for the Brazilian governmentto i) monitor the entry of foreign workers into the labor market; ii) oversee the records of theFGTS (Fundo de Garantia do Tempo de Servico) program, a national benefits program consistingof employers’ contributions to each of its employees; iii) provide information for administeringseveral government benefits programs such as unemployment insurance; and iv) generate statisticsregarding the formal labor market. Today it is the main tool used by the government to enablethe payment of the ”abono salarial” to eligible workers. This is a government program that paysone additional minimum wage at the end of the year to workers whose average monthly wage wasnot greater than two times the minimum wage, and whose job information was correctly declaredin RAIS, among other minor requirements. Thus, workers have an incentive to ensure that theiremployer is filing the required information. Moreover, firms are required to file, and face fines untilthey do so. Together, these requirements ensure that the data in RAIS are accurate and complete.

Observations in the data are indexed by a worker ID number, the Programa de Integracao So-cial (PIS), and an establishment registration number, the Cadastro Nacional da Pessoa Jurıdica(CNPJ). Both of these identifiers are consistent over time, allowing one to track workers and estab-lishments across years. Establishment industry is reported using the Subsetor IBGE classification,which includes 12 manufacturing industries, 2 primary industries, 11 nontradable industries, and1 other/ignored.60 Worker education is reported using the following 9 education categories (list-ing corresponding years of education in parentheses): illiterate (0), primary school dropout (1-3),primary school graduate (4), middle school dropout (5-7), middle school graduate (8), high schooldropout (9-10), high school graduate (11), college dropout (12-14), and college graduate (≥ 15).

In each year, and for each job, RAIS reports average earnings throughout the year, and earningsin December.61 We focus on labor market outcomes reported in December of each year. This choiceensures that earnings and formal employment status are measured at the same time for all workersand all jobs. It avoids the potential confounding effects on average yearly earnings that might arisein situations where some workers begin working in early in the year and others begin late in theyear.

A.3 Demographic Census

We utilize information from the long form of the Demographic Censuses (Censo Demografico) for1970, 1980, 1991, 2000, and 2010. The long form micro data reflect a 5 percent sample of thepopulation in 1970, 1980, and 2010, a 5.8 percent sample in 1991, and a 6 percent sample in 2000.The primary benefit of the Census for our purposes is the ability to observe those outside formalemployment, who are not present in the RAIS database.

Although our main analysis focuses on monthly earnings, following the information availablein RAIS, the Census provides weekly hours information from 1991-2010, allowing us to calculatehourly wages as monthly earnings divided by 4.33 times weekly hours. Census results for monthlyearnings and hourly wages are very similar. In 1970 and 1980, hours information is presented in

60A less aggregate industry classification (CNAE) is available from 1994 onward, but we need a consistent classifi-cation from 1986-2010, so we use Subsetor IBGE.

61From 1994 onward, RAIS reports hours, making it possible to calculate hourly wages. However, since we need aconsistent measure from 1986-2010, we focus on monthly earnings.

54


5 rough bins. Thus, when calculating pre-liberalization trends using data from 1970 and 1980, weuse monthly earnings even when examining hourly wage outcomes.

In 1991-2010, the Census asks whether each worker has a signed work card. This is the standarddefinition of formal employment, and is necessary for a worker to appear in the RAIS sample.Thus, we use this as our primary definition of formal employment. In 1980 and 1991, there is analternative proxy for formal employment, reporting whether the worker’s job includes contributionsto the national social security system. When calculating pre-liberalization outcome trends for 1980-1991, we use this alternative measure to identify formally employed workers. The social securitycontributions proxy appears to be a good one; in 1991, when both measures are available, 95.9percent of workers would be classified identically when using either measure. In 1970, there is noinformation on formality, so pre-liberalization outcome trends for 1970-1980 are calculated for allworkers.

The definition of employment changes across Census years. In 1970 it includes those reportingworking or looking for work during August 1970 (the questionnaire does not separately identifyworking vs. looking for work). In 1980 it includes those who report working during the yearprior to September 1, 1980. In 1991 it includes those reporting working regularly or occasionallyduring the year prior to September 1, 1991. In 2000 and 2010 it includes those who report paidwork, temporary leave, unpaid work, or cultivation for own consumption during the week of July23-29 in 2000 and July 25-31 in 2010. Note that the employment concept changes substantiallyacross years. This highlights yet another benefit of using RAIS as our primary data source, sincethe employment concept in RAIS is consistent throughout the sample. Yet, while the changescomplicate the interpretation of Census-based employment rates over time, there is no reasonto expect systematic differences across regions to result from the changing employment concept.Thus, our cross-region identification strategy should be valid when using the Census to measureemployment in spite of these measurement issues.

A.4 Regional Tariff Changes

Regional tariff reductions, defined in (2), are constructed using information from various sources.Tariff changes come from Kume et al. (2003), and are aggregated from the Nıvel 50 level to the in-dustry classification presented in Table A1 using 1990 value-added weights from the IBGE NationalAccounts. Figure 1 shows the resulting industry-level variation in tariff changes.

The weights, βri in (2) depend upon the initial regional industry distribution (λri) and thespecific-factor share in production (ϕi). We calculate the λri using the 1991 Census. We use theCensus because it provides a less aggregate industry definition than what is available in RAIS, andbecause the Census allows us to calculate weights that are representative of overall employment,rather than just formal employment. However, note that shocks using formal employment weightsyield very similar results (see Panel D of Table B6). We calculate the ϕi using data from the UseTable of the 1990 National Accounts from IBGE. The table “Componentes do Valor Adicionado”provides the wagebill (Remuneracoes) and gross operating surplus (Excedente Operacional BrutoInclusive Rendimento de Autonomos), which reflects the share of income earned by capital. Wedefine ϕi as capital’s share of the sum of these two components. When imposing equal cost sharesacross industries (see Section 5.4.2), we calculate ϕ using the economy-wide wagebill and grossoperating surplus, yielding a value of ϕ = 0.544.

Because Brazilian local labor markets differ substantially in the industry distribution of their

55


employment, the weights βri vary across regions. Figure A2 demonstrates how variation in industrymix leads to variation in RTRr. The figure shows the initial industry distribution of employment forthe regions facing the largest tariff reduction (Rio de Janeiro) the median tariff reduction (Alfenasin southwestern Minas Gerais state), and the smallest tariff reduction (actually a small increase,Mata Grande in northwest Alagoas state). The industries on the x-axis are sorted from the mostnegative to the most positive tariff change. Rio de Janeiro has more weight on the left side of thediagram, by virtue of specializing in manufacturing, particularly in apparel and food processingindustries, which faced quite large tariff reductions. Thus, its regional tariff reduction is quitelarge. Alfenas is a coffee growing and processing region, which also has some apparel employment,balancing the large tariff declines in apparel and food processing against the small tariff increasein agriculture. Mata Grande is located in a sparsely populated mountainous region, and is almostexclusively agricultural, leading it to experience a small tariff increase overall. Thus, although allregions faced the same set of tariff reductions across industries, variation in the industry distributionof employment in each region generates substantial variation in RTRr.

56


Figure A2: Variation Underlying Regional Tariff Reduction

0.75 0.97

0.00

0.10

0.20

0.30

0.40

0.50

Indu

stry

Wei

ght

Rubb

er

Appa

rel

Oth

er M

anuf

.

Phar

ma.

, Per

fum

es, D

eter

gent

s

Plas

tics

Auto

, Tra

nspo

rt, V

ehicl

es

Nonm

etal

lic M

iner

al M

anuf

Elec

tric,

Ele

ctro

nic

Equi

p.

Food

Pro

cess

ing

Mac

hine

ry, E

quip

men

t

Petro

leum

Refi

ning

Text

iles

Chem

icals

Woo

d, F

urni

ture

, Pea

t

Pape

r, Pu

blish

ing,

Prin

ting

Met

als

Foot

wear

, Lea

ther

Min

eral

Min

ing

Petro

leum

, Gas

, Coa

l

Agric

ultu

re

Industries sorted from most negative to most positive tariff change

Rio de Janeiro, RJ (.15) Alfenas MG (.03) Mata Grande, AL (-.01)

Industry distribution of 1991 employment in the regions facing the largest (Rio de Janeiro, RJ), median (Alfenas,MG) and smallest (Mata Grande, AL) regional tariff reduction. Industries sorted from the most negative to themost positive tariff change (see Figure 1). More weight on the left side of the figure leads to a larger regional tariffreduction, and more weight on the right side leads to a smaller regional tariff reduction.

57


A.5 Local Price Indexes

Moretti (2013) calculates local price indexes for the U.S. using the change in monthly rents for 2 or3 bedroom apartments. We adjust this approach to the Brazilian context in a few ways. First, wefocus on 1 or 2 bedroom apartments, which are far more common in the Brazilian setting, accountingfor more than 85 percent of the stock of rental units in 1991 and 2010. Many Brazilian cities includefavelas with somewhat improvised structures, and rural areas often feature less formal dwellings.We restrict the sample to include only units with modern construction materials (masonry or woodframing), with at least one bathroom, and with modern sanitation (sewer or septic tank). Theserestrictions allow us to avoid comparing modern apartments to informal dwellings. Using thissample of apartments, we calculate the change in log average monthly rent in each region. 19very sparsely populated microregions do not have observations for any rental units satisfying thesecharacteristics in either 1991 or 2010, so we have rent indexes for 456 microregions in our sample.

We then need to transform the change in rental prices into a regional price index. Given thecross-sectional nature of our analysis, we only need to be concerned with prices that vary at thelocal level, i.e. nontradables, since tradable goods prices move together across regions, and thusdo not affect this exercise. Using local Consumer Price Indexes produced by the Bureau of LaborStatistics for 23 U.S. metropolitan areas, Moretti (2013) shows that, as expected, local non-housingnontradables’ prices move with local rental prices. He estimates a slope of 0.35 for the effect ofhousing prices on non-housing nontradables’ prices. The Brazilian Consumer Price Index (Indicesde Precos ao Consumidor - IPC) system reports that in 2002-03, housing’s share of consumptionwas 16.24 percent and that the share for other nontradable goods was 39.94 percent (IBGE 2005).Together, these figures imply that the effective weight on housing prices in the consumer price indexis 0.1624 + 0.3994 · 0.35 = 0.3022. Our local price deflator is therefore 0.3022 times the change inlog rental prices in the region.

58


A.6 Regional Change in Imports and Exports

Import and export data between Brazil and the rest of the world come from Comtrade, at the6-digit HS level. We map from HS codes to the industries presented in Table A1 and calculate totalBrazilian trade flows by industry and year.

In the main text, we follow Autor et al. (2013) (ADH) by generating regional weighted averagesof changes in imports and exports per worker. For each industry, we calculate the change in imports(Mit) and exports (Xit) from 1990 to each subsequent year t. These trade flows are measured in$100,000 units. We then generate the regional change in imports and exports per worker as follows.

RegImprt =∑i

Lrit0Lrt0

∆1990−tMit

Lit0(16)

RegExprt =∑i

Lrit0Lrt0

∆1990−tXit

Lit0(17)

The rightmost ratios in these expressions measure the change in imports or exports per workerinitially employed in the industry, in year t0 = 1991. The preceding ratios represent industryweights for each region, reflecting industry i’s share of tradable employment in region r in 1991.These weights are equivalent to λri in (1). We then generate weighted averages by summing theseterms over tradable industries. Finally, we construct regional net exports as the difference inregional exports and imports.

RegNetExprt = RegExprt −RegImprt

In Appendix B.10, we present an alternative set of results based on the change in log tradeflows rather than the change in trade flows per worker.

RegLnImprt =∑i

Lrit0Lrt0

∆1990−t ln(Mit) (18)

RegLnEmprt =∑i

Lrit0Lrt0

∆1990−t ln(Xit) (19)

We emphasize that this measure is presented only for descriptive purposes and as a statisticalrobustness test since it does not have the same theoretical underpinnings as the measures followingAutor et al. (2013).

A.7 Entry, Exit, Job Creation, and Job Destruction

We calculate cumulative job creation and job destruction following Davis and Haltiwanger (1990).

job creationrt ≡∑

e∈Ert, get>0

xetXrt

get, (20)

job destructionrt ≡∑

e∈Ert, get<0

xetXrt|get|, (21)

59


where get ≡Let − Le,1991

xet, xet ≡

1

2(Let + Le,1991), Xrt ≡

∑e∈Ert

Let,

Let is employment at establishment e in year t and Ert is the set of active establishments in regionr in year t. Note that employment growth, get, is calculated from 1991 to year t. The dependentvariables for the regressions underlying Figure 9 are ln(job creationrt) and ln(job destructionrt).

Entry and exit are calculated analogously, replacing establishment employment with an indica-tor for the establishment being active in the relevant year, ιet.

entryrt ≡∑

e∈Ert, get>0

xet

Xrt

get, (22)

exitrt ≡∑

e∈Ert, get<0

xet

Xrt

|get|, (23)

where get ≡ιet − ιe,1991

xet, xet ≡

1

2(ιet + ιe,1991), Xrt ≡

∑e∈Ert

ιet,

These definitions yield intuitive results. entryrt is equivalent to the share of active firms in regionr in year t that were not active in 1991. exitrt is equivalent to the number of firms in region rthat were active in 1991 and not in year t, divided by the number of firms active in year t. Thedependent variables for the regressions underlying Figure 8 are ln(entryrt) and ln(exitrt)

60


B Supplemental Empirical Results

B.1 Industry-Level Outcome Pre-Trends vs. Tariff Reductions

Along with regional variation in the industrial composition of employment, our analysis relies onvariation in tariff cuts across industries. Here we analyze the relationship between tariff cuts duringliberalization (1990-1995) and trends in industry wages and employment before liberalization, 1980-1991. We calculate these pre-liberalization outcome trends using the Demographic Census, toprovide a longer pre-liberalization period than what is available in RAIS, which starts in 1986.

We implemented a variety of specifications, with results reported in Table B1. In all spec-ifications, the independent variable is the proportional reduction in one plus the tariff rate (-∆1990−95 ln(1 + τi)). In panels A-C the dependent variable is the change in log industry earnings.Panel A uses average log earnings; Panel B uses average log earnings residuals controlling for in-dividual age, sex, education, and formal status; and Panel C uses average log earnings residualscontrolling for these individual characteristics and region fixed effects. In Panel D, the dependentvariable is the change in industry log employment. Column (1) weights industries equally, andpresents standard errors based on pairwise bootstrap of the t-statistic, to improve small sampleproperties with only 20 tradable industry observations. Column (2) uses the same estimator, butdrops agriculture. Column (3) uses heteroskedasticity weights and presents heteroskedasticity-robust standard errors, which are likely understated in this small sample (MacKinnon 2011). Col-umn (4) uses the same estimator, but drops agriculture. In all cases, the results should be seenprimarily as suggestive, because the analysis uses only 19 or 20 observations.

Nearly all of the earnings estimates are positive, indicating larger tariff reductions in indus-tries experiencing more positive wage growth prior to liberalization. The majority of the estimatesare insignificantly different from zero, with the exception of weighted results in Panels A and B.These specifications heavily weight agriculture, which exhibited negative wage growth prior to lib-eralization and experienced essentially no tariff decline during liberalization, driving the strongnegative relationship. By dropping agriculture, Column (4) confirms that the significant rela-tionship is driven by agriculture. The employment estimates are larger, and change sign acrosscolumns. Given the diversity of findings across earnings and employment specifications, this ex-ercise is somewhat inconclusive. Tariff cuts may or may not have been substantially correlatedwith pre-liberalization outcome trends. These findings motivate us to control for pre-liberalizationoutcome trends whenever possible throughout the paper. This ensures that our results are robustto potential spurious correlation between liberalization-induced labor demand shocks and ongoingtrends.

61


Table B1: Pre-Liberalization Industry Trends - 1980-1991

unweighted, bootstrapped

unweighted, bootstrapped, omitting

agriculture

weighted weighted, omitting agriculture

1980-1991 change in log: (1) (2) (3) (4)

Panel A: average earningsIndustry tariff reduction 0.345 0.111 1.029*** 0.510

(0.322) (0.354) (0.139) (0.582)Panel B: earnings premia (with individual controls)

Industry tariff reduction 0.203 -0.017 0.610*** -0.235(0.273) (0.311) (0.157) (0.350)

Panel C: earnings premia (with individual and region controls)Industry tariff reduction 0.135 0.044 0.184 0.018

(0.177) (0.209) (0.158) (0.222)

Panel D: employmentIndustry tariff reduction -1.624 -2.696** 0.687 -1.651

(1.272) (1.361) (0.417) (1.894)

Observations 20 19 20 19

Decennial Census data. 20 industry observations (19 omitting agriculture). See text for details of dependent andindependent variable construction. Column (1) weights industries equally, and presents standard errors based onpairwise bootstrap of the t-statistic. Column (2) uses the same estimator as Column (1), but drops agriculture.Column (3) uses heteroskedasticity weights and presents heteroskedasticity-robust standard errors. Column (4) usesthe same estimator as Column (3), but drops agriculture. *** Significant at the 1 percent, ** 5 percent, * 10 percentlevel.

62


B.2 Informal Employment

The following results provide some descriptive evidence on the informal sector in Brazil. Informalityis defined as working without a signed work card (Carteira de Trabalho e Previdencia Social), whichentitles workers to benefits and labor protections afforded them by the legal employment system.Table B2 shows that the overall rate of informality increased from 1991 to 2000, before decreasingsubstantially from 2000 to 2010. Rates of informality are highest in agriculture and much lower inmanufacturing. Table B1 breaks out informality rates in the manufacturing sector into individualindustries. Finally, Table B2 focuses on the year 2000 and shows the industry distribution of formaland informal employment. There is very substantial overlap in the industry distributions of formaland informal employment. The biggest differences occur in agriculture, which comprises a muchlarger share of informal employment, and food processing and metals, which comprise larger sharesof formal employment. In contrast, the nontradable share is nearly identical for formal and informalemployment.

63


Table B2: Informal Share of Employment - 1991-2010

1991 2000 2010

Overall 0.58 0.64 0.49

Agriculture 0.89 0.86 0.83Mining 0.61 0.45 0.21Manufacturing 0.28 0.39 0.29Nontradable 0.55 0.64 0.48

Author’s calculations using Brazilian Demographic Census data for workers age 18-64. Informality defined as nothaving a signed work card.

Figure B1: Informal Share of Employment by Industry - 1991-2010

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Rubb

er

Apparel

Other M

anuf.

Pharma., Perfumes, D

etergents

PlasBcs

Auto, Transpo

rt, V

ehicles

Non

metallic M

ineral M

anuf


Food

Processing

Machine

ry, Equ

ipmen

t

Petroleu

m Refi

ning

TexBles

Chem

icals

Woo

d, Furniture, Peat

Pape

r, Pu

blish

ing, Prin

Bng

Metals

Footwear, Leathe

r

Mineral M

ining

Petroleu

m, G

as, Coal

Agriculture

Non

traded

Inform

al Sha

re of Ind

ustry Em

ploymen

t 1991

2000

2010

Authors’ calculations using Brazilian Demographic Census data for workers age 18-64. Informality defined as nothaving a signed work card. Industries sorted from most negative to most positive tariff change (with the exceptionof the nontraded sector).

64


Figure B2: Industry Distribution of Formal and Informal Employment - 2000

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.10

Rubb

er

Apparel

Other M

anuf.

Pharma., Perfumes, D

etergents

Plas8cs

Auto, Transpo

rt, V

ehicles

Non

metallic M

ineral M

anuf


Food

Processing

Machine

ry, Equ

ipmen

t

Petroleu

m Refi

ning

Tex8les

Chem

icals

Woo

d, Furniture, Peat

Pape

r, Pu

blish

ing, Prin

8ng

Metals

Footwear, Leathe

r

Mineral M

ining

Petroleu

m, G

as, Coal

Agriculture

Non

traded

Indu

stry Sha

re of S

ector E

mploymen

t

formal

informal

0.23 0.70 0.69

Authors’ calculations using year 2000 Brazilian Demographic Census data for workers age 18-64. Informality definedas not having a signed work card. Industries sorted from most negative to most positive tariff change (with theexception of the nontraded sector).

65


B.3 Regional Earnings Premium Regressions

As discussed in Section 4.1, we calculate regional earnings premia by regressing workers’ log De-cember earnings on flexible demographic and educational controls, industry fixed effects, and regionfixed effects, separately in each year. Table B3 shows the coefficient estimates from these earningspremium regressions for 1991, 2000, and 2010. The region fixed effect estimates provide average logearnings for formally employed workers in the region, controlling for the age, sex, education, andindustry composition of the region’s employment. These regional premia then form the outcomevariable in our earnings analyses.

Note that the coefficient estimates on the controls conform to expectations. Women are paidless than otherwise similar men, and this earnings gap declines over time. Workers exhibit aninverted U-shaped wage profile as they age, as is standard in Mincerian regressions. The returnsto education are monotonically positive.

66


Table B3: Regional Earnings Premium Regressions - 1991, 2000, 2010

(1) (2) (3)1991 2000 2010

Female -0.349*** -0.295*** -0.261***(0.000) (0.000) (0.000)

Age25-29 0.210*** 0.209*** 0.148***

(0.000) (0.000) (0.000)30-39 0.407*** 0.374*** 0.273***

(0.000) (0.000) (0.000)40-49 0.527*** 0.525*** 0.382***

(0.001) (0.000) (0.000)50-64 0.435*** 0.507*** 0.474***

(0.001) (0.001) (0.000)

Education (years)Primary School Dropout (1-3) 0.015*** 0.007*** 0.130***

(0.001) (0.001) (0.001)Primary School Graduate (4) 0.111*** 0.075*** 0.182***

(0.001) (0.001) (0.001)Middle School Dropout (5-7) 0.188*** 0.129*** 0.206***

(0.001) (0.001) (0.001)Middle School Graduate (8) 0.297*** 0.181*** 0.236***

(0.001) (0.001) (0.001)High School Dropout (9-11) 0.454*** 0.305*** 0.289***

(0.001) (0.001) (0.001)High School Graduate (12) 0.711*** 0.523*** 0.430***

(0.001) (0.001) (0.001)College Dropout (13-15) 0.967*** 0.902*** 0.792***

(0.002) (0.001) (0.001)College Graduate (≥16) 1.374*** 1.384*** 1.368***

(0.001) (0.001) (0.001)

Fixed EffectsIndustry (24) X X XRegion (475) X X X

Observations 13,582,443 17,733,492 30,662,075R-squared 0.858 0.842 0.759

dependent variable: log monthly earnings

Individual worker observations from RAIS. Earnings premium regressions were run for each year from 1986-2010.Here we show three years as examples. The region fixed effect estimates provide average log earnings for formallyemployed workers in the region, controlling for the age, sex, education, and industry composition of the region’semployment. These regional premia then form the outcome variable in our regional earnings analyses. The omittedcategory is a male, age 18-24, with 0 years of education (illiterate). Robust standard errors in parentheses. ***Significant at the 1 percent, ** 5 percent, * 10 percent level.

67


B.4 Industry-Region Earnings Results

In this appendix, we pursue an alternative research design for studying the earnings effects ofliberalization, with the unit of analysis at the industry × region level. Equations (13) and (14)suggest a similar industry × region research design, but in the context of our model this approachonly applies in the long run. The results in this section are therefore supplementary. Similarapproaches appear in Hakobyan and McLaren (forthcoming) and Acemoglu, Autor, Dorn, Hansonand Price (2016).

We must first estimate earnings premia at the industry × region level. Because there are 4,773such combinations, it is not feasible to directly estimate industry × region fixed effects (and obtaintheir standard errors) in an earnings regression like (4). Instead, we estimate the following earningsregression separately in each year t, absorbing the industry × region fixed effects, µirt

ln(earnjrit) = XjtΓt + µirt + ejrit (24)

We then calculate residuals ujrit ≡ ln(earnjrit)−XjtΓt, and average these residuals within industry× region bins to recover the fixed-effect estimates, µirt. Because this procedure does not directlyyield standard errors for these fixed-effect estimates, we implement 100 bootstrap repetitions andcalculate bootstrap standard errors. We then calculate the change in these industry × regionearnings premia and use them as the dependent variable in the following regression specification.

µirt − µir,1991 = θtRTRr + δtd ln(1 + τi) + αst + γt(µir,1990 − µir,1986) + εirt (25)

We estimate this regression separately in each year t > 1991. Note that the unit of observationis the industry × region, and we include both the region-level tariff reduction, RTRr, and theindustry-level tariff change, d ln(1 + τi). In the most stringent specifications, we control for statefixed effects and both region and industry earnings pre-trends.

The results for 2000 and 2010 appear in Panel A of Table B4. First, note that the estimatedregional and industry-level effects of liberalization are quite consistent across specifications. Wecontinue to find large increases in the regional effects of liberalization between 2000 and 2010,while the industry-level results are quite constant over time. The industry-level tariff changeestimates reflect the direct effect of liberalization on workers in the affected industry, irrespectiveof their region of residence. The RTRr coefficients capture the regional general-equilibrium effectsoperating across industries. In 2010, the coefficient on RTRr is more than three times larger thanthe industry effects. This finding makes clear the central role of regional labor market equilibriumin affecting workers outcomes following trade liberalization. Panel B replaces the controls for theindustry-level tariff change with industry fixed effects, finding similar results.

68


Table B4: Industry × Region log Formal Earnings 2000, 2010

Change in industry × region earnings(1) (2) (3) (4) (5) (6)

Panel A: Industry Tariff-Change ControlsRegional tariff reduction (RTR) -0.618** -0.878*** -0.762*** -2.567*** -2.485*** -2.252***

(0.270) (0.276) (0.254) (0.582) (0.266) (0.216)Industry tariff reduction (-d ln(1+τi)) -0.683*** -0.654*** -0.661*** -0.684*** -0.573** -0.590***

(0.215) (0.216) (0.215) (0.254) (0.251) (0.215)Formal earnings pre-trend (86-90) -0.196** -0.393***


R-squared 0.082 0.116 0.143 0.212 0.323 0.390

Panel B: Industry Fixed EffectsRegional tariff reduction (RTR) -0.528** -0.777*** -0.724*** -2.199*** -2.109*** -2.036***

(0.229) (0.217) (0.212) (0.538) (0.248) (0.244)Formal earnings pre-trend (86-90) -0.243*** -0.277***


R-squared 0.313 0.345 0.375 0.437 0.530 0.554

1991-2000 1991-2010

Negative coefficient estimates for the regional tariff reduction imply larger declines in formal earnings in regions facinglarger tariff reductions. Negative coefficients for the industry tariff reduction imply larger declines in formal earningsin industries facing larger tariff reductions. 4,733 industry × region observations in each year. Earnings premiacalculated controlling for age, sex, and education. Panel A controls for industry tariff reductions, while Panel B usesindustry fixed effects to capture liberalization’s direct effect on each industry. Efficiency weighted by the inverseof the squared standard error of the estimated change in log formal earnings premium (see text for description ofbootstrap standard error calculation). Pre-trends computed for 1986-1990. Standard errors (in parentheses) adjustedfor 112 mesoregion clusters. *** Significant at the 1 percent, ** 5 percent, * 10 percent level.

69


B.5 Formal Earnings Regression Scatterplots

Figure B3 shows scatter plots underlying the formal earnings regression estimates in Figure 3 for1995, 2000, 2005, and 2010. Each marker represents a microregion, and microregions in each majorregion are shown with a separate type of marker. The size of each marker is proportional tothe weight the relevant microregion receives in the estimation. The mean value of the dependentvariable is normalized to zero in each year to focus attention on the slope.

These scatter plots make clear three important points about the earnings estimates. First,as shown in Figure 3, the magnitude of the slope increases substantially and steadily as timepasses following liberalization. Second, the relationship between changes in formal earnings premiaand regional tariff reductions is approximately linear in all time periods, justifying our choice offunctional form. Third, the increasing magnitude slope is driven by shifts in earnings across largenumbers of microregions in various parts of the country, rather than by a few outliers.

70


Fig

ure

B3:

For

mal

Ear

nin

gsP

rem

iavs.

Reg

ion

alT

ariff

Red

uct

ion

s-

1995

,20

00,

2005,

2010

-1-.50.51

0.05

.1.15

1995

-1-.50.51

0.05

.1.15

2000

-1-.50.51

0.05

.1.15

2005

-1-.50.51

0.05

.1.15

2010

Sca

tter

plo

tsof

regre

ssio

ns

beh

ind

Fig

ure

3es

tim

ate

s.475

mic

rore

gio

nobse

rvati

ons.

Y-a

xis

isth

ech

ange

inlo

gfo

rmal

earn

ings

pre

miu

min

agiv

enre

gio

nfr

om

1991

toth

eyea

rlist

edin

the

titl

e.D

epen

den

tva

riable

dem

eaned

inea

chyea

r.X

-axis

isth

ere

gio

nal

tari

ffre

duct

ion,RTRr.

The

size

of

each

mark

erre

pre

sents

the

inver

seof

the

vari

ance

of

the

dep

enden

tva

riable

esti

mate

,use

dto

wei

ght

the

regre

ssio

ns.

Majo

rre

gio

ns

show

nw

ith

diff

eren

tsy

mb

ols

.South

east

:dark

blu

eci

rcle

s,N

ort

hea

st:

ora

nge

dia

monds,

Cen

ter-

Wes

t:re

dtr

iangle

s,South

:light

blu

eX

s,N

ort

h:

gre

ensq

uare

s.

71


B.6 Census Earnings, Wage, and Employment Results

Table B5 estimates versions of equation (3) using formal sector outcomes calculated using Censusdata. Because the Census includes hours information, we are able to examine both earnings andwage premia. In all cases, we find negative coefficients onRTRr, indicating that regions facing largertariff reductions experienced relative declines in monthly earnings, hourly wages, or employment.We also find substantial growth in the magnitude of these effects, corroborating the results in Table2, which uses RAIS outcomes.

Table B5: Census Regional log Formal Earnings, Wages, and Employment - 2000, 2010


Panel A: log Formal Earnings PremiaRegional tariff reduction (RTR) -‐0.397 -‐0.293** -‐0.261** -‐1.384** -‐0.890*** -‐0.855***

(0.335) (0.120) (0.116) (0.572) (0.198) (0.186)Formal earnings pre-‐trend (86-‐90) -‐0.0896* -‐0.0994


R-‐squared 0.031 0.579 0.583 0.156 0.718 0.720

Panel B: log Formal Wage PremiaRegional tariff reduction (RTR) -‐0.630* -‐0.533*** -‐0.495*** -‐1.320** -‐0.765*** -‐0.721***

(0.355) (0.124) (0.118) (0.525) (0.173) (0.163)Formal earnings pre-‐trend (86-‐90) -‐0.108* -‐0.124*


R-‐squared 0.071 0.605 0.609 0.136 0.718 0.721

Panel B: log Formal EmploymentRegional tariff reduction (RTR) -‐2.478*** -‐1.756*** -‐1.619*** -‐3.913*** -‐2.865*** -‐2.725***

(0.487) (0.281) (0.258) (0.758) (0.443) (0.417)Formal employment pre-‐trend (86-‐90) 0.211*** 0.227**


R-‐squared 0.317 0.612 0.630 0.319 0.629 0.638

1991-‐2000 1991-‐2010

Outcomes calculated using Census data. Negative coefficient estimates for the regional tariff reduction imply largerdeclines in formal earnings, wages, or employment in regions facing larger tariff reductions. 475 microregion observa-tions. Regional earnings premia calculated controlling for age, sex, education, and industry of employment. Efficiencyweighted by the inverse of the squared standard error of the estimated outcome. RAIS pre-trends computed for 1986-1990. Standard errors (in parentheses) adjusted for 112 mesoregion clusters. *** Significant at the 1 percent, ** 5percent, * 10 percent level.

72


B.7 Robustness Tests

Table B6 presents the various earnings robustness tests summarized in Section 4.2. For comparison,Panel A shows the main specification, corresponding to the estimates in Table 2 and Figure 3.

All panels in the table control for pre-liberalization earnings growth from 1986-1990, calculatedusing RAIS. Panel B additionally controls for longer pre-liberalization earnings trends calculatedusing the Census. The 1980-1991 control reflects the growth in formal earnings premium, whereformality is defined based on whether the worker’s job included social security contributions. The1970-1980 control reflects growth in the earnings premium for all workers, since there is no formalityinformation in the 1970 Census. See Appendix A.3 for more detail on Census data. A potentialproblem with Panel B is mechanical endogeneity, because the 1980-1991 pre-trend and the 1991-tearnings growth dependent variable both include the 1991 earnings premium. Panel C resolves thisissue by using earnings growth from 1992 to year t as the dependent variable, while including theCensus pre-trend controls.

Panel D calculates the regional tariff reduction (RTRr) in (2) using weights based on the initialindustry distribution of regional formal employment, rather than overall employment. Panel Ecalculates RTRr using effective rates of protection rather than nominal tariffs. Effective rates ofprotection capture the overall effect of liberalization on producers in a given industry, accountingfor tariff changes on industry inputs and outputs. Kume et al. (2003) provide effective rates ofprotection along with the nominal tariffs used in our main analysis. The magnitude of the changesin effective rates of protection is larger than for nominal tariffs, so the coefficients in Panel 3smaller by the same proportion. Since versions of RTRr based on effective rates of protection andnominal tariffs are nearly perfectly correlated (correlation = 0.993), the variation in earnings growthexplained by both versions is nearly identical. Panel F calculates RTRr including the nontradablesector with a tariff change value of 0. This measure ignores the fact that nontradable prices movewith tradable prices (see Appendix A.5 and Panel B of Table 2), and in doing so underestimatesthe magnitude of the average liberalization-induced price change faced by each region. Because themagnitude of RTRr is reduced, the coefficient estimates are inflated by the same proportion.

Panel G omits industry fixed effects when calculating regional earnings premia. This maintainsthe national industry-level variation in earnings in the outcome measure, rather than restrictingattention to the regional equilibrium earnings used in our main specifications. Panel H omits allcontrols from the earnings premium regressions, using simple average log earnings for workers inthe relevant region. While the main analysis weights by the inverse of the squared standard errorof the estimated growth in regional wage premium, Panel I weights all regions equally, and PanelJ weights by 1991 formal employment.

In all cases, the effects grow substantially over time, as in our main specification. In fact, in allbut two of these robustness tests, the long run effect of liberalization on earnings is larger than itis in the main specification.

Table B7 shows that the formal employment results in Table 2 and Figure 4 are similarly robust.The panel labels correspond to Table B6, so see above for descriptions of each specification. Notethat Panels G and H do not apply to employment, since they relate to earnings premia. Panel Ialso does not apply, because the main employment specification is unweighted, since RAIS containsthe population of formally employed workers.

73


Table B6: Robustness: Regional log Formal Earnings Premia - 1995, 2000, 2005, 2010



(0.120) (0.141) (0.139) (0.169)Panel B: Long pre-trends (Census: 1970-80, 1980-91, RAIS: 1986-90)


Panel C: Long pre-trends, earnings growth from 1992 to tRegional tariff reduction (RTR) -0.478*** -1.009*** -1.737*** -2.039***

(0.115) (0.197) (0.195) (0.215)Panel D: RTR using formal employment industry weights

Regional tariff reduction (RTR) -0.089 -0.358 -1.270*** -1.665***(0.189) (0.217) (0.246) (0.298)

Panel E: RTR using effective rates of protectionRegional tariff reduction (RTR) -0.047 -0.328*** -0.823*** -1.017***

(0.076) (0.091) (0.090) (0.107)Panel F: RTR including zero nontradable price change

Regional tariff reduction (RTR) -0.798 -1.758*** -3.350*** -4.625***(0.489) (0.577) (0.643) (0.696)

Panel G: Earnings premium without industry fixed effectsRegional tariff reduction (RTR) 0.131 -0.422*** -1.420*** -1.895***

(0.148) (0.151) (0.163) (0.209)Panel H: Earnings premium with no controls (mean log earnings)

Regional tariff reduction (RTR) 0.317* 0.046 -1.192*** -1.905***(0.171) (0.214) (0.147) (0.182)

Panel I: Unweighted (equally weighted)Regional tariff reduction (RTR) -0.244 -0.490** -1.074*** -1.546***

(0.169) (0.197) (0.215) (0.224)Panel J: Weighted by 1991 formal employment

Regional tariff reduction (RTR) -0.020 -0.345** -1.217*** -1.631***(0.128) (0.147) (0.143) (0.175)


Negative coefficient estimates for the regional tariff reduction (RTR) imply larger declines in formal earnings in regionsfacing larger tariff reductions. 475 microregion observations, except Panels B and C, which use a more aggregateregion definition with 405 observations for consistency with 1970 and 1980 Census data. Regional earnings premiacalculated controlling for age, sex, education, and industry of employment except in Panels G and H. Standard errors(in parentheses) adjusted for 112 mesoregion clusters, except Panels B and C with 90 mesoregion clusters. Efficiencyweighted by the inverse of the squared standard error of the estimated change in log formal earnings premium exceptin Panels I and J. See text for detailed description of each panel. *** Significant at the 1 percent, ** 5 percent, * 10percent level.

74


Table B7: Robustness: Regional log Formal Employment - 1995, 2000, 2005, 2010

Change in log Formal Employment: 1991-1995 1991-2000 1991-2005 1991-2010(1) (2) (3) (4)

Panel A: Main specificationRegional tariff reduction (RTR) -1.900*** -3.533*** -4.517*** -4.663***

(0.422) (0.582) (0.685) (0.679)Panel B: Long pre-trends (Census: 1970-80, 1980-91, RAIS: 1986-90)


Panel C: Long pre-trends, earnings growth from 1992 to tRegional tariff reduction (RTR) -0.722 -2.957*** -4.252*** -4.102***

(0.804) (0.972) (1.084) (1.070)Panel D: RTR using formal employment industry weights

Regional tariff reduction (RTR) -1.728*** -2.690*** -4.491*** -4.362***(0.598) (0.793) (0.782) (0.789)

Panel E: RTR using effective rates of protectionRegional tariff reduction (RTR) -1.201*** -2.336*** -2.959*** -3.074***

(0.274) (0.369) (0.438) (0.430)Panel F: RTR including zero nontradable price change


Panel G: Not applicablePanel H: Not applicablePanel I: Not applicable

Panel J: Weighted by 1991 formal employmentRegional tariff reduction (RTR) -1.195*** -2.119*** -3.406*** -2.842***


Negative coefficient estimates for the regional tariff reduction (RTR) imply larger declines in formal employmentin regions facing larger tariff reductions. 475 microregion observations, except Panels B and C, which use a moreaggregate region definition with 405 observations for consistency with 1970 and 1980 Census data. Panel labelscorrespond to Table B6, so Panels G and H, which relate to earnings premia, are not applicable here, nor is PanelI, since the main specification is unweighted. Standard errors (in parentheses) adjusted for 112 mesoregion clusters,except Panels B and C with 90 mesoregion clusters. See text for detailed description of each panel. *** Significantat the 1 percent, ** 5 percent, * 10 percent level.

75


B.8 Potential Confounders

B.8.1 Post-Liberalization Tariff Reductions

We calculate post-liberalization regional tariff reductions as in (2), but use tariff reductions between1995 and year t > 1995. Because the Kume et al. (2003) data end in 1998, we use UNCTADTRAINS to construct post-liberalization tariff reductions. The TRAINS data are reported by 6-digit HS codes. In order to maintain as much industry variation as possible, we created an industrymapping from HS codes to Census industry codes, which yields 44 consistently identifiable tradableindustries. This provides more industry detail than the main industry definition in Table A1.The concordance is available upon request. Panel B of Table B8 includes these post-liberalizationtariff reduction controls in the regional earnings growth regression. The post-liberalization controlhas the expected negative coefficient, but its inclusion has very little effect on the liberalizationcoefficient.

B.8.2 Real Exchange Rates

We construct regional real exchange rate shocks as follows. We begin with real exchange ratesbetween Brazil and its trading partners, calculated from Revision 7.1 of the Penn World Tables.We then calculate each country’s 1989 shares of Brazil’s imports and exports in each industryusing Comtrade. As in the prior section, we use the industry definition mapping from HS codes toCensus industries. Industry-specific real exchange rates are weighted averages of country-specificreal exchange rates, weighting either by the 1989 import share or export share. We define industry-level real exchange rate shocks as the change in log industry real exchange rate from 1990 toeach subsequent year. Finally we create regional real exchange rate shocks as weighted averagesof industry real exchange rate shocks, where the region’s industry weights are given by the 1991industry distribution of employment. Panel C of Table B8 includes both the import-weighted andexport-weighted real exchange rate controls. With these controls, the earnings effects grow evenmore than in the main specification.

B.8.3 Privatization

Substantial privatization in Brazil began in 1991 with the administration of President Collor, butsignificantly increased during President Cardoso’s administration (1995-2002). Beginning in 1995,the RAIS data allow us to identify as state-owned any firm at least partly owned by the government.In panels D and E of Table B8, we include different controls for the regional effects of privatization.Panel D includes quartile indicators for the 1995 share of regional employment in state-owned firms,controlling flexibly for the initial share of employment subject to potential privatization. Panel Econtrols for the change in state-owned firm employment share from 1995 to subsequent year t. Inboth cases, the privatization controls have no meaningful effect on the RTRr coefficients.

76


Table B8: Potential Confounders: Regional log Formal Earnings Premia - 1995, 2000, 2005, 2010



(0.120) (0.141) (0.139) (0.169)Panel B: Post-liberalization tariff reductions


Post-liberalization (1995 to t) regional n/a -3.705 -2.415 -2.124tariff reductions (3.273) (3.872) (1.425)

Panel C: Exchange ratesRegional tariff reduction (RTR) -0.113 -0.482*** -1.475*** -1.728***

(0.118) (0.161) (0.184) (0.207)Import-weighted real exchange rate 0.136 0.570* 0.243* 0.374*

(0.133) (0.327) (0.141) (0.201)Export-weighted real exchange rate 0.051 -0.164 0.084 -0.166

(0.160) (0.280) (0.384) (0.355)Panel D: Privatization: initial state-owned employment share


Quartile indicators, 1995 state-owned ✓ ✓ ✓ ✓employment share distirbution

Panel E: Privatization: change in state-owned employment share, 1995 to tRegional tariff reduction (RTR) -0.096 -0.514*** -1.243*** -1.558***

(0.120) (0.149) (0.146) (0.182)Change in state-owned employment share 0.095 0.286 0.176

(0.178) (0.214) (0.227)Formal earnings pre-trend (86-90) ✓ ✓ ✓ ✓State fixed effects (26) ✓ ✓ ✓ ✓

Negative coefficient estimates for the regional tariff reduction (RTR) imply larger declines in formal earnings inregions facing larger tariff reductions. 475 microregion observations. Regional earnings premia calculated controllingfor age, sex, education, and industry of employment. Standard errors (in parentheses) adjusted for 112 mesoregionclusters. Efficiency weighted by the inverse of the squared standard error of the estimated change in log formalearnings premium. See text for detailed description of each panel and for control construction. *** Significant at the1 percent, ** 5 percent, * 10 percent level.

77


B.8.4 Commodity Price Boom

Figures B4 - B6 show price indexes for major Brazilian export commodities from 1991 to 2010, usingdata from the Primary Commodity Price Series produced by the International Monetary Fund.Although there is some variability across commodities, there is little evidence of the commodityprice boom prior to 2004. For example, in Figure B4, aggregate agricultural commodity prices werenearly identical in 1991 and 2003, and Brazil’s main agricultural exports (coffee, sugar, soy, andcotton) all have lower average prices in 2003 than in 1991. Natural resource and meat commoditiesin Figures B5 and B6 exhibit similar patterns, with relatively flat series prior to 2003 and substantialgrowth for many commodities starting in 2004.

Note the contrast between this time series pattern and the effects of regional tariff reductions onformal earnings and employment growth in Figures 3 and 4. The earnings effects grow steadily from1996 to 2003, in spite of the fact that most commodity prices actually fell a bit during that timespan. Commodity prices start growing very quickly in 2004 and later, during which the earningseffects start to level off. A similar argument applies to the employment effects, which grow from1994 to 2004 and level off subsequently. Thus, the timing of the commodity price boom doesnot conform with the timing of the earnings and employment effects, making it very unlikely thatcommodity prices drive our results, particularly before 2004.

To reinforce this time-series evidence, in Table B9 we implement a wide variety of tests to ruleout the commodity price boom as a potential confounder. Panel A reproduces the main specifica-tion for comparison. Panels B and C respectively restrict the sample of regions to those with belowmedian and bottom quartile employment shares in agriculture and mining, the sectors affected bythe commodity price boom. Note that mining includes fuel extraction. When focusing on regionswith minimal exposure to commodity sectors, we find even larger growth in the effects of liberaliza-tion on earnings than in the entire sample. In Panel D, we maintain all regions, but only calculateearnings premia for workers employed in the manufacturing sector, omitting workers in commodityand nontradables sectors. Once again, the earnings effects continue to grow substantially over timegiven this restriction. As an aside, note that earnings in the manufacturing sector, which mostdirectly experienced the effects of trade liberalization, exhibit significant effects on impact, in 1995.This finding suggests that the very short-run effects of liberalization were concentrated in the in-dustries facing the largest tariff cuts, but that the earnings effects spread out to other sectors overtime through labor market equilibrium.

In Panels E and F, rather than restricting the sample of regions or workers, we control for thecommodity price boom directly. We utilize the regional commodity price shocks constructed byAdao (2015). Special thanks to Rodrigo Adao for sharing his data and code. See Appendix C inAdao (2015) for details on the data source and his equation (16) for the shock construction. Tosummarize, he calculates commodity-specific changes in log price from 1991 to each subsequentyear using data from the Commodity Research Bureau, and constructs regional weighted averagesof these commodity price shocks. The weights reflect each commodity sector’s share of total laborpayments in all commodity sectors in the region in 1991. Because this measure does not incorporateregional variation in overall exposure to commodity price changes, we also flexibly control for theregional importance of commodities by including quartile indicators for the region’s 1991 share ofregional employment in agriculture and mining. As seen in Panel E, controlling directly for thesecommodity price movements has little influence on the increasing profile of earnings effects, eitherbefore or after the beginning of commodity price boom.

78


We implement a similar exercise in Panel F of Table B9 using more detailed commodity priceinformation from the IMF Primary Commodity Price Series, a subset of which is shown in FiguresB4 - B6. While Adao’s measure uses only 6 aggregate commodity indexes, our alternative measureuses 19 separate indexes. As an example, this detail allows us to distinguish between the commodityprice changes faced by regions specialized in coffee vs. sugar, which are grouped together underAdao’s classification. Table B10 shows our mapping from commodity industry codes in the 1991Census to the IMF price indexes. We calculate the change in log price index from 1991 to eachsubsequent year for each IMF commodity and then generate regional weighted averages of theseprice changes, where weights reflect the relevant commodity’s share of regional employment in1991. As seen in Panel F, including these commodity price controls has no substantive effect uponthe relationship between regional tariff reductions and regional earnings, either before or afterthe beginning of commodity price boom. Because our regional commodity price shock measurealready incorporates the overall share of commodity industries in regional employment, we do notadditionally control for that share, though doing so has no substantive effect on the results.

The rise of China appears to have played a substantial role in driving up commodity prices inthe late 2000s. As a final test of the commodity price boom hypothesis, we follow Costa et al.(2016), who study the regional effects of import competition from China and Chinese demand forexports. Rather than focusing on commodity prices, Costa et al. study the effects of import andexport quantity shocks, along the lines of Autor et al. (2013). They construct industry-level Chineseimport supply (IS) and export demand (XD) shocks as the growth in industry imports from orexports to China from 2000 to 2010, divided by Brazilian employment in the industry in 2000.They then generate regional weighted average shocks using the year 2000 industry distribution ofemployment in each region. Finally, they instrument for these shocks using similar measures basedon the growth in Chinese trade to countries other than Brazil. Special thanks to Francisco Costafor providing us with their shock and instrument measures.

Because Costa et al. examine shocks and outcomes between 2000 and 2010, in Panel A of TableB11 we provide our baseline earnings estimate for this time period, with a base year of 2000 ratherthan 1991. We use a slightly more aggregate region definition to match theirs, yielding 405 regionobservations. The coefficient estimate of -1.068 in column (1) is nearly identical to the differencebetween the estimates for 2000 and 2010 in columns (3) and (6), respectively, of Panel A in Table2 of -1.065. The slight difference results from the difference in region definitions between the twotables. In columns (2)-(4) of Table B11, we introduce the Chinese import supply and exportdemand shocks, instrumented following Costa et al.. The two shocks have the expected sign, withincreased import competition lowering regional earnings and increased export demand increasingthem (very slightly), though only the import supply shock is statistically different from zero. Thisresult might seem surprising, given that Costa et al. find significant effects of export demand onwages. However, when they control for the regional composition of workers and for outcome pre-trends, as we do here, the export result loses statistical significance (see their Table 2, Panel B,column (5) in their paper). When we include these controls, they have only a very small effecton our coefficient of interest, further confirming that the divergence in earnings growth betweenregions facing larger and smaller tariff reductions was not driven by China’s effects on commoditymarkets.

79


Figure B4: Agricultural Commodity Prices - 1991-2010

0

0.5

1

1.5

2

2.5

3

3.5

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Coffee Sugar Soy Co7on Agriculture

Monthly price series from IMF Primary Commodity Price Series, rescaled to equal 100 in January 1991.

80


Figure B5: Resource Commodity Prices - 1991-2010

0

1

2

3

4

5

6

7

8

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Wood Fuel Metals

Monthly price series from IMF Primary Commodity Price Series, rescaled to equal 100 in January 1991, except thefuel index which begins in 1992 and is rescaled to 1 in January 1992.

81


Figure B6: Meat Commodity Prices - 1991-2010

0

0.5

1

1.5

2

2.5

3

3.5

4

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Beef Fish Poultry

Monthly price series from IMF Primary Commodity Price Series, rescaled to equal 100 in January 1991.

82


Table B9: Commodity Price Boom: Regional log Formal Earnings Premia - 1995, 2000, 2005, 2010



(0.120) (0.141) (0.139) (0.169)Panel B: Below-median agriculture/mining employment share (238 obs)


Panel C: Bottom quartile agriculture/mining employment share 118 obs)Regional tariff reduction (RTR) 0.006 -0.347 -1.379*** -2.153***

(0.267) (0.262) (0.280) (0.373)Panel D: Manufacturing sector earnings


Panel E: Direct commodity price controls per Adao (2015)Regional tariff reduction (RTR) -0.052 -0.290 -1.269*** -1.926***

(0.259) (0.257) (0.276) (0.372)Regional commodity price shocks 0.033 -0.039 0.118 0.045

(Adao 2015) (0.207) (0.167) (0.092) (0.127)Quartile indicators, 1991 agriculture/mining ✓ ✓ ✓ ✓

employment share distributionPanel F: Direct commodity price controls using detailed commodity price data (IMF)

Regional tariff reduction (RTR) 0.023 -0.591*** -1.210*** -1.718***(0.143) (0.188) (0.137) (0.330)

Regional commodity price shocks 0.179 0.160 0.421 -0.069(IMF data) (0.120) (0.266) (0.277) (0.149)


Negative coefficient estimates for the regional tariff reduction (RTR) imply larger declines in formal earnings in regionsfacing larger tariff reductions. 475 microregion observations unless otherwise noted (Panels B and C). Regionalearnings premia calculated controlling for age, sex, education, and industry of employment. Standard errors (inparentheses) adjusted for 112 mesoregion clusters. Efficiency weighted by the inverse of the squared standard error ofthe estimated change in log formal earnings premium. See text for detailed description of each panel and for controlconstruction. *** Significant at the 1 percent, ** 5 percent, * 10 percent level.

83


Table B10: Mapping from Commodity Industries to IMF Price Indexes

1991 Census Industry (atividade) IMF Index11 Agave PRAWM Agricultural Raw Materials12 Cotton PCOTTIND Cotton13 Rice PRICENPQ Rice14 Banana PBANSOP Bananas15 Cocoa PCOCO Cocoa beans16 Coffee PCOFFOTM Coffee17 Sugar cane PSUGAISA Sugar18 Tobacco PRAWM Agricultural Raw Materials19 Cassava PRAWM Agricultural Raw Materials20 Corn PMAIZMT Maize (corn)21 Soybeans PSOYB Soybeans22 Wheat PWHEAMT Wheat23 Horticulture and floriculture PRAWM Agricultural Raw Materials24 Forestry PSAWORE Soft Sawnwood25 Other agricultural products PRAWM Agricultural Raw Materials26 Livestock PBEEF Beef27 Aviculture PPOULT Poultry (chicken)28 Beekeeping and Sericulture PRAWM Agricultural Raw Materials29 Other livestock PRAWM Agricultural Raw Materials31 Rubber PRUBB Rubber32 Yerba mate PRAWM Agricultural Raw Materials33 Plant fibres PRAWM Agricultural Raw Materials34 Fruits, oilseeds, and waxes PRAWM Agricultural Raw Materials35 Wood PSAWORE Soft Sawnwood36 Charoal PCOALAU Coal37 Other harvesting activities PRAWM Agricultural Raw Materials41 Fishing PFISH Fishmeal42 Aquaculture PFISH Fishmeal51 Oil and natural gas mining PNRG Fuel (Energy)52 Coal mining PCOALAU Coal55 Metallic mineral panning and deposition PMETA Metals56 Radioactive mineral mining PURAN Uranium58 Metallic mineral mining (except those in other categories) PMETA Metals581 Agriculture and livestock services PRAWM Agricultural Raw Materials

84


Table B11: Regional log Formal Earnings Premia with Costa et al (2015) controls

Change in log Formal Earnings Premia, 2000-2010: (1) (2) (3) (4)OLS IV IV IV


Formal earnings pre-trend (86-90) -0.077 -0.064 -0.076 -0.063(0.053) (0.055) (0.051) (0.055)

China import supply (Costa et al. 2015) -0.034*** -0.034***(0.010) (0.010)

China export demand (Costa et al. 2015) 0.001 0.001(0.002) (0.002)

State fixed effects (26) ✓ ✓ ✓ ✓

R-squared 0.733 0.738 0.733 0.7391st stage F (Kleibergen-Paap) 22.16 441.6 11.31

Negative coefficient estimates for the regional tariff reduction (RTR) imply larger declines in formal earnings inregions facing larger tariff reductions. See text for description of China import supply and export demand controlsand associated instruments from Costa et al. (2015). First stage partial F-statistics reported in brackets. 405microregion observations. Regional earnings premia calculated controlling for age, sex, education, and industry ofemployment. Standard errors (in parentheses) adjusted for 90 mesoregion clusters. Efficiency weighted by the inverseof the squared standard error of the estimated change in log formal earnings premium. *** Significant at the 1percent, ** 5 percent, * 10 percent level.

85


B.9 Earnings and Employment Sample Splits

Tables B12 and B13 present earnings results splitting the sample of workers into those employed inthe tradable sector (Panel B), those employed in the nontradable sector (Panel C), more educated(Panel D), and less educated (Panel E). Note that earnings and employment effects grow for allsubsamples. The employment effects are concentrated in the tradable sector and among less skilledworkers, though panels D and E in both tables should be interpreted with care, as the regionaltariff reduction shocks are derived from a model with a single type of labor. For a more generalmodel with two skill types, see Dix-Carneiro and Kovak (2015a).

Table B12: Sample Splits: Regional log Formal Earnings Premia - 1995, 2000, 2005, 2010


Panel A: Full sampleRegional tariff reduction (RTR) -0.096 -0.529*** -1.294*** -1.594***

(0.120) (0.141) (0.139) (0.169)Panel B: Tradable sector workers


Panel C: Nontradable sector workersRegional tariff reduction (RTR) -0.060 -0.389** -1.143*** -1.401***

(0.150) (0.179) (0.169) (0.183)Panel D: More educated workers (high school or more)

Regional tariff reduction (RTR) 0.310* -0.539*** -1.611*** -2.053***(0.160) (0.173) (0.192) (0.224)

Panel E: Less educated workers (less than high school)Regional tariff reduction (RTR) -0.218* -0.626*** -1.354*** -1.758***


Negative coefficient estimates for the regional tariff reduction (RTR) imply larger declines in formal earnings inregions facing larger tariff reductions. 475 microregion observations. Regional earnings premia calculated controllingfor age, sex, education, and industry of employment. Standard errors (in parentheses) adjusted for 112 mesoregionclusters. Efficiency weighted by the inverse of the squared standard error of the estimated change in log formalearnings premium. *** Significant at the 1 percent, ** 5 percent, * 10 percent level.

86


Table B13: Sample Splits: Regional log Formal Employment - 1995, 2000, 2005, 2010

Change in log Formal Employment: 1991-1995 1991-2000 1991-2005 1991-2010(1) (2) (3) (4)

Panel A: Full sampleRegional tariff reduction (RTR) -1.900*** -3.533*** -4.517*** -4.663***

(0.422) (0.582) (0.685) (0.679)Panel B: Tradable sector workers


Panel C: Nontradable sector workersRegional tariff reduction (RTR) 0.726 -0.733 -1.500* -1.600**

(0.455) (0.664) (0.778) (0.749)Panel D: More educated workers (high school or more)

Regional tariff reduction (RTR) 0.141 -1.219* -1.637** -2.195***(0.450) (0.644) (0.772) (0.733)

Panel E: Less educated workers (less than high school)Regional tariff reduction (RTR) -2.507*** -4.556*** -6.328*** -6.910***


Negative coefficient estimates for the regional tariff reduction (RTR) imply larger declines in formal employment inregions facing larger tariff reductions. 475 microregion observations. Standard errors (in parentheses) adjusted for112 mesoregion clusters. *** Significant at the 1 percent, ** 5 percent, * 10 percent level.

87


B.10 Regional Change in log Imports and Exports

This section presents versions of Figure 5 and Table 6 using alternative trade quantity measuresreflecting regional weighted averages of the change in log imports or exports rather than the changein imports or exports per worker. These alternative measures are presented only for descriptivepurposes and as a statistical robustness test.

We use the change in trade per worker in the main text both because it is theoretically justified,as shown by Autor et al. (2013), and because it intuitively captures the effects of changing tradequantities on local labor market outcomes. Figures B7 and B8, which show scatter plots relatingthe industry-level change in trade per worker and the change in log trade for imports and exportsin 2000 and 2010. Markers are proportional to 1991 employment, and industry labels correspondto Table A1. These plots show that the change in log trade often deviates substantially from thechange in trade per worker. This deviation occurs primarily in industries with relatively small tradeflows and relatively large employment. As an example, consider the Wood, Furniture, and Peatindustry (code 14) in Panel A of Figure B7. In this industry, Brazil imported R$71 million in 1990and R$249 million in 2000 (all values in year 2005 Reais). This very large proportional growthin imports corresponds to the large value for the change in log imports of 1.25. However, initialemployment in this industry was also quite large, 822,579, so the change in imports per worker wasonly R$216, much smaller than the values in the thousands or tens of thousands in other industries.Therefore, although the amount of imports increased very much in proportional terms, it was stillinsignificant compared to the number of workers in the industry. The change in trade per workercaptures the relative scale of trade and employment, while the change in log trade does not.

Nonetheless, figure B9 shows the relationships between RTRr and the regional change in logimports and exports (paralleling Figure 5). See (18) and (19) in Appendix A.6 for details onconstructing the change in log trade measures. Regions facing larger tariff reductions experiencelarger increases in log imports and larger declines in log exports. The magnitude of each effect growsover time, suggesting that perhaps slow trade quantity responses could explain the slow growth ofregional earnings and employment effects in Figures 3. We demonstrate that this is not the caseby directly controlling for the regional change in log import and export measures when estimatingthe effect of RTRr on regional earnings growth. The specifications in Table B14 parallel those inTable 6, using the alternative change-in-log measures of trade flows for both the controls and theinstruments. In all cases, the earnings effects of liberalization grow even more when controlling forimport and export quantity growth than in the baseline specification in Panel A. The relevant Stockand Yogo (2005) critical value for the first-stage F-statistic is 21, so Panel C exhibits a potentialweak instruments problem. We therefore present two additional sets of results in Table B15, usingthe change in trade per worker measure when calculating the instruments rather than the changein log trade flow measure. The weak instrument issue is not longer present, and the effects of RTRron regional earnings still increase substantially over time.

As with the standard trade-per-worker trade flow measures considered in the main text, theanalysis using the alternative change-in-log trade flow measure rules out slow import or exportresponses as the mechanism driving the slowly growing earnings effects. This is to be expectedgiven the discussion at the beginning of this section, since the change in log trade measure doesnot well capture the effects of changing trade flows on workers’ labor market outcomes.

88


Figure B7: Change in Trade Per Worker vs. Change in log Trade - 1990-2000

Panel A: Imports

12

4 5

810

12

1415

17

21

222324 25

320

1000

020

000

3000

040

000

chan

ge in

impo

rts p

er w

orke

r, 19

90-2

000

-.5 0 .5 1 1.5change in log imports, 1990-2000

Omits the 4 smallest industries, with less than 150,000 employees in 1991.

Panel B: Exports

12

4

5

810

12

1415

17

21

22 23

24

25

32

050

0010

000

1500

020

000

chan

ge in

exp

orts

per

wor

ker,

1990

-200

0

-.5 0 .5 1 1.5change in log exports, 1990-2000


Each point is an industry, with labels corresponding to Table A1. The y-axis measures the change in trade perworker initially employed in the industry (in 1991) and the x-axis measures the change in log trade. The figure omitsthe 4 smallest industries in terms of 1991 employment, which often fall well outside the scale shown. Because theyare small, they receive very little weight in the regional analysis that forms this paper’s main analysis. Marker sizeproportional to 1991 employment.

89


Figure B8: Change in Trade Per Worker vs. Change in log Trade - 1990-2010

Panel A: Imports

12 4

5

8

10

12

1415

17

21

222324 25

320

2000

040

000

6000

080

000

1000

00ch

ange

in im

ports

per

wor

ker,

1990

-201

0

.5 1 1.5 2 2.5change in log imports, 1990-2010


Panel B: Exports

1

2

4

5

8

10

12

14

15

17

21

222324

25

32

020

000

4000

060

000

chan

ge in

exp

orts

per

wor

ker,

1990

-201

0

-1 0 1 2change in log exports, 1990-2010


Each point is an industry, with labels corresponding to Table A1. The y-axis measures the change in trade perworker initially employed in the industry (in 1991) and the x-axis measures the change in log trade. The figure omitsthe 4 smallest industries in terms of 1991 employment, which often fall well outside the scale shown. Because theyare small, they receive very little weight in the regional analysis that forms this paper’s main analysis. Marker sizeproportional to 1991 employment.

90


Figure B9: Regional Imports and Exports, Change in log Measure - 1992-2010

-‐20

-‐15

-‐10

-‐5

0

5

10

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Imports

Exports


Each point reflects an individual regression coefficient, θt, following (3), where the dependent variable is the change inregional imports (blue circles) or exports using the change in log measures described in (18) and (19) in Appendix A.6.The independent variable is the regional tariff reduction (RTR), defined in (2). Note that the RTR always reflectstariff reductions from 1990-1995. All regressions include state fixed effects, but do not include pre-liberalizationtrends due to a lack of Comtrade trade data before 1989. Positive (negative) estimates imply larger increases in tradeflow in regions facing larger (smaller) tariff reductions. Vertical bar indicates that liberalization was complete by1995. Dashed lines show 95 percent confidence intervals. Standard errors adjusted for 112 mesoregion clusters.

91


Table B14: Slow Response of Imports or Exports, Change in log Measure - 1995, 2000, 2005, 2010(Part 1 of 2)



(0.120) (0.141) (0.139) (0.169)

Panel B: Controls for trade quantitities (OLS)Regional tariff reduction (RTR) -0.142 -0.581*** -1.476*** -1.887***

(0.124) (0.142) (0.151) (0.212)Import quantity control (change in log)

Export quantity control (change in log)

Panel C: Latin America IVRegional tariff reduction (RTR) -0.236 -0.608*** -1.479*** -2.158***



First-‐stage F (Kleibergen-‐Paap)

Panel D: Colombia IV Regional tariff reduction (RTR) -0.227 -0.570*** -1.316*** -1.985***





(0.062)

0.025**(0.012)-0.009(0.008)

-0.009

(0.032)61.45

-0.052(0.039)19.71

-0.044(0.049)-0.059*

Negative coefficient estimates for the regional tariff reduction (RTR) imply larger declines in formal earnings inregions facing larger tariff reductions. Panel A replicates the earnings results in columns (3) and (6) of Table 2and in Figure 3. Panels B-D control for regional import and export quantities, calculated using the change in logtrade flows; see Appendix A.6 for details. These panels stack the data across years, allowing the effect of RTRr tovary over time but fixing the import and export quantity coefficients over time. We instrument for the potentiallyendogenous import and export controls using regional measures of commodity price growth from Adao (2015) andwith regional trade flows for other countries. We consider the combination of Argentina, Chile, Colombia, Paraguay,Peru, and Uruguay (“Latin America”) Colombia alone. In each case, we measure imports and exports between thesecountries and the rest of the world, excluding Brazil. This gives us 2 endogenous variables and 57 instruments (= 3instruments × 19 years). First-stage Kleinbergen-Paap F statistics are shown, for comparison to the Stock and Yogo(2005) critical value of 21 to reject 5 percent bias relative to OLS. Standard errors (in parentheses) adjusted for 112mesoregion clusters. Efficiency weighted by the inverse of the squared standard error of the estimated change in logformal earnings premium. *** Significant at the 1 percent, ** 5 percent, * 10 percent level.

92


Table B15: Slow Response of Imports or Exports, Change in log Measure - 1995, 2000, 2005, 2010(Part 2 of 2)


Panel E: Latin America IV (change in trade per worker)Regional tariff reduction (RTR) -0.063 -0.487*** -1.143*** -1.364**




Panel F: Colombia IV (change in trade per worker)Regional tariff reduction (RTR) -0.096 -0.529*** -1.294*** -1.594***





(0.055)-0.026(0.026)61.58

-0.022(0.050)0.005

(0.031)31.44

-0.045

See Table B14 for general notes. Because the instrument in Panel C of Table B14 was marginally weak, Panels E andF present versions using instruments based on the change in imports per worker, while the trade quantity controlsare calculated using the change in log. Both specifications reject the weak instrument concern.

93


B.11 Overall Employment

Table B16 shows the effect of liberalization on overall regional employment, including both formallyand informally employed workers. We use Census data to capture informally employed individuals,and control for 1980-1991 and 1970-1980 outcome pre-trends. The estimates vary substantiallyacross specifications and all but one are insignificantly different from zero. These results providelittle evidence in favor of overall employment as a potential source of agglomeration economies.

Table B16: Regional log Overall Employment - 2000, 2010

Change in log Overall Employment: (1) (2) (3) (4) (5) (6)

Regional tariff reduction (RTR) 0.203 -0.419 -0.272 0.657** -0.478 -0.265(0.209) (0.450) (0.260) (0.314) (0.683) (0.410)

Overall employment pre-trend (80-91) 0.329** 0.281** 0.538*** 0.454**(0.136) (0.130) (0.202) (0.189)

Overall employment pre-trend (70-80) 0.221*** 0.120*** 0.385*** 0.230***(0.071) (0.035) (0.093) (0.052)

State fixed effects (26) ✓ ✓ ✓ ✓ ✓ ✓

R-squared 0.563 0.479 0.585 0.574 0.484 0.609

1991-2000 1991-2010

Positive (negative) coefficient estimates for the regional tariff reduction imply larger increases (decreases) in overallemployment in regions facing larger tariff reductions. Outcomes calculated using Census data. 405 microregionobservations. Efficiency weighted by the inverse of the squared standard error of the dependent variable estimate.Pre-trends computed for 1980-1991 and 1970-1980. Standard errors (in parentheses) adjusted for 112 mesoregionclusters. *** Significant at the 1 percent, ** 5 percent, * 10 percent level.

94


B.12 Capital Adjustment Confidence Intervals

Figures B10 - B12 show the capital adjustment profiles in Figure 6, including 95-percent confidenceintervals, which were omitted from Figure 6 for clarity.

Figure B10: Capital Adjustment Quantification - Low ζ - 1992-2010

-‐3.0

-‐2.5

-‐2.0

-‐1.5

-‐1.0

-‐0.5

0.0

0.5

1.0

1.5

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010


(chg. from 1991)

Capital (establishments) adjustment, ζ = 0.152

Each point reflects an individual regression coefficient, θt, following (3). The dependent variable is capital’s con-tribution to overall adjustment, using the number of regional formal establishments as a proxy for regional capital.This figure shows the profile using the low estimate of ζ = 0.152. The independent variable is the regional tariffreduction (RTR), defined in (2). Note that the RTR always reflects tariff reductions from 1990-1995. All regressionsinclude state fixed effects, and post-liberalization regressions control for the 1986-1990 outcome pre-trend. Negativeestimates imply larger declines in the number of establishments in regions facing larger tariff reductions. Verticalbar indicates that liberalization was complete by 1995. Dashed lines show 95 percent confidence intervals. Standarderrors adjusted for 112 mesoregion clusters.

95


Figure B11: Capital Adjustment Quantification - Mid ζ - 1992-2010

-‐3.0

-‐2.5

-‐2.0

-‐1.5

-‐1.0

-‐0.5

0.0

0.5

1.0

1.5

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010


(chg. from 1991)



96


Figure B12: Capital Adjustment Quantification - High ζ - 1992-2010

-‐3.0

-‐2.5

-‐2.0

-‐1.5

-‐1.0

-‐0.5

0.0

0.5

1.0

1.5

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010


(chg. from 1991)



97


B.13 Exit by Establishment Size

Here we examine the relationship between establishment exit and RTRr, separately by initialestablishment size. We run the following specification at the establishment-year level, using thesample of all active establishments in 1991.

Exitirt =6∑

k=1

βkt Sizeki ·RTRr +

6∑k=1

φkSizeki + γtNT i + ϑtPreExit

1986−1990r + εirt (26)

where Sizeki is an indicator for whether establishment i fell into size bin k in 1991, NT i is anindicator for establishments in the nontradable sector, and PreExit1986−1990r is a pre-trend controlfor the share of regional establishments in 1986 that shut down between 1986 and 1990.

Figure B13 plots the βkt coefficients, with the relevant initial employment bin definitions shownon the right side. Although there is some variation across establishment sizes, with more exit amonglarger establishments than smaller establishments, it is clear that exit rates increased throughoutthe size distribution for establishments whose regions faced larger tariff declines.

98


Figure B13: Regional log Cumulative Formal Establishment Exit, by Establishment Size - 1992-2010

-‐0.1

0.1

0.3

0.5

0.7

0.9

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010


1-‐4

5-‐9

10-‐19

100+

20-‐49

50-‐99

Plots the βkt coefficients in (26), estimated using the sample of all active establishments in 1991. The size range,indexed by k, is reported at the right side of each profile. Vertical bar indicates that liberalization was complete by1995.

99


C Model

C.1 Baseline Model

This section generalizes the specific-factors model of regional economies from Kovak (2013) to allowfor changes in regional productivity (agglomeration economies) and changes in labor and capitalinputs.

The economy consists of many regions, indexed by r, which may produce goods in many in-dustries, indexed by i. Production in each industry uses Cobb-Douglas technology with constantreturns to scale and three inputs: labor, a fixed industry-specific factor, and capital. Labor, Lr,is assumed to be perfectly mobile between industries within a region. The industry-specific factor,Tri, is usable only in its respective region and industry and is fixed over time. Capital, Kri, isusable only in its respective region and industry but may change over time. Output of industry iin region r is

Yri = AriL1−ϕiri

(T ζiriK

1−ζiri

)ϕi, (C1)

where ϕi, ζi ∈ (0, 1). To allow for the possibility of agglomeration economies and factor adjustment,we allow Ari, Lr, and Kri to change over time. Goods and factor markets are perfectly competitive,and producers face exogenous prices Pi, common across regions and fixed by world prices and tariffs.

Consider a particular region r, and suppress the region subscript. Let aLi, aT i, and aKi be therespective amounts of labor, specific factor, and capital used in producing one unit of Yi. Regionalfactor market clearing implies ∑

i

aLiYi = L, (C2)

aT iYi = Ti ∀i, (C3)

aKiYi = Ki ∀i. (C4)

Perfect competition implies that the price equals factor payments,

aLiw + aT isi + aKiRi = Pi ∀i, (C5)

where w is the wage, si is the specific-factor price, and Ri is the price of capital. Define x as theproportional change in x, and differentiate (C5).

(1− ϕi)w + ϕiζisi + ϕi(1− ζi)Ri = Pi + Ai ∀i, (C6)

which uses the fact that, from cost minimization,

(1− ϕi)aLi + ϕiζiaT i + ϕi(1− ζi)aKi = −Ai ∀i. (C7)

Differentiate the factor market clearing conditions.∑i

λi(aLi + Yi) = L, (C8)

Yi = −aT i ∀i, (C9)

Yi = Ki − aKi ∀i, (C10)

100


where λi ≡ LiL is the share of regional labor allocated to industry i, and we use the fact that Ti = 0.

With Cobb-Douglas production, the elasticity of substitution is one, so

aKi − aT i = si − Ri, (C11)

aLi − aKi = Ri − w, (C12)

Combining (C8), (C10), and (C12) yields∑i

λiRi − w = L−∑i

λiKi. (C13)

Combine (C9), (C10), and (C11) to yield

si = Ri + Ki ∀i. (C14)

Plug this into (C6) and simplify.

Ri =Pi + Ai − (1− ϕi)w − ϕiζiKi

ϕi(C15)

Finally, plug this into (C13), solve for w, and restore regional subscripts to yield the equilibriumrelationship for regional wage changes, equation (9) in the main text.

wr =∑i

βriPi +∑i

βriAri − δr

(Lr −

∑i

λri(1− ζi)Kri

)(C16)

where βri ≡λri

1ϕi∑

j λrj1ϕj

and δr ≡1∑

j λrj1ϕj

.

C.2 Agglomeration Economies

As discussed in the main text, when examining agglomeration economies and quantifying the long-run effects of slow capital adjustment and agglomeration, we assume perfectly mobile capital in thelong run (Rr = R ∀r), and identical technology across industries (ϕi = ϕ ∀i and ζi = ζ ∀i). Theassumption of perfectly mobile capital allows us to substitute out the change in capital, Kri for thechange in its price, R, which is constant across industries and regions.

Start with the labor market clearing condition in (C8), substitute in the specific-factors clearingcondition in (C9) and the Cobb-Douglas conditions in (C11) and (C12) to yield∑

i

λisi − w = L. (C17)

Rearrange the zero-profit condition in (C6) to solve for si,

si =1

ϕζ(Pi + Ai)−

ϕ(1− ζ)

ϕζR− 1− ϕ

ϕζw, (C18)

101


and plug it into (C17). Solving for w and restoring regional subscripts yields the following expres-sion.

wr =1

1− ϕ(1− ζ)

∑i

βri(Pi + Ari)−ϕζ

1− ϕ(1− ζ)Lr −

ϕ(1− ζ)

1− ϕ(1− ζ)R (C19)

where βri ≡λri

1ϕ∑

j λrj1ϕ

= λLri

To incorporate agglomeration economies, we assume a constant elasticity agglomeration func-tion, (11), and a constant labor supply elasticity, (12). Substituting these into (C19) and simplifyingyields the following expression for the regional wage change, equation (13) in the main text, whichwe use to estimate the agglomeration elasticity, κ.

wr =η

η[1− ϕ(1− ζ)]− κ+ ϕζ

∑i

βriPi −ϕ(1− ζ)η

η[1− ϕ(1− ζ)]− κ+ ϕζR (C20)

We also use an alternative employment-based approach to estimate κ. Start by noting thatemployment in a region × industry pair is given by Lri ≡ aLriYri. Differentiating this definition andplugging in the specific-factor market clearing condition, (C9), and the Cobb-Douglas substitutionconditions, (C11) and (C12), we have

Li = si − w. (C21)

Substitute in si from (C18) and simplify.

Li =1

ϕζ(Pi + Ai)−

1− ϕ(1− ζ)

ϕζw − ϕ(1− ζ)

ϕζR (C22)

Plug in the labor supply and agglomeration equations, (12) and (11).

Li =1

ϕζPi −

η[1− ϕ(1− ζ)]− κηϕζ

w − ϕ(1− ζ)

ϕζR (C23)

Finally, plug in the equilibrium wage change in (C20), combine terms, and restore regional sub-scripts to yield equation (14) in the main text.

Lri =1

ϕζPi −

1

ϕζ· η[1− ϕ(1− ζ)]− κη[1− ϕ(1− ζ)]− κ+ ϕζ

∑i

βriPi −ϕ(1− ζ)

η[1− ϕ(1− ζ)]− κ+ ϕζR (C24)

102

Trade Liberalization and Regional Dynamics · 2017-08-25 · Trade Liberalization and Regional...

Documents

Transcript of Trade Liberalization and Regional Dynamics · 2017-08-25 · Trade Liberalization and Regional...