Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income...

79
Export markets and microenterprise performance: Evidence from Vietnam Brian McCaig Department of Economics, Wilfrid Laurier University Nina Pavcnik Department of Economics, Dartmouth College and The National Bureau of Economic Research December 2018 First draft: August 2013 Preliminary Abstract Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine how export opportunities induced by the 2001 U.S.-Vietnam Bilateral Trade Agreement affect the performance of non-farm microenterprises in Vietnam. On average, microenterprises in industries with greater declines in U.S. tariffs on Vietnamese exports are more likely to exit and surviving microenterprises expand revenue, while other outcomes are not affected. However, the responses to tariff cuts differ with the initial size of the microenterprise and the gender of the manager. Female-run microenterprises are less likely to exit if they are small and become more likely to exit with size. In contrast, male-run businesses are more likely to exit if small and exit becomes less likely as size increases. Initially small microenterprises experience a contraction of revenue, while initially larger businesses account for the observed expansion of revenue within an industry in response to tariff cuts. We also find evidence of adjustment in the prevalence of operating the business as the individual’s primary job and the incidence of hiring workers from outside the family that varies by gender and initial size.

Transcript of Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income...

Page 1: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Export markets and microenterprise performance: Evidence from Vietnam

Brian McCaig

Department of Economics,

Wilfrid Laurier University

Nina Pavcnik

Department of Economics,

Dartmouth College

and

The National Bureau of Economic Research

December 2018

First draft: August 2013

Preliminary

Abstract

Most workers in low-income countries are self-employed and work for a household-owned business

or a farm. We examine how export opportunities induced by the 2001 U.S.-Vietnam Bilateral Trade

Agreement affect the performance of non-farm microenterprises in Vietnam. On average,

microenterprises in industries with greater declines in U.S. tariffs on Vietnamese exports are more

likely to exit and surviving microenterprises expand revenue, while other outcomes are not

affected. However, the responses to tariff cuts differ with the initial size of the microenterprise and

the gender of the manager. Female-run microenterprises are less likely to exit if they are small and

become more likely to exit with size. In contrast, male-run businesses are more likely to exit if small

and exit becomes less likely as size increases. Initially small microenterprises experience a

contraction of revenue, while initially larger businesses account for the observed expansion of

revenue within an industry in response to tariff cuts. We also find evidence of adjustment in the

prevalence of operating the business as the individual’s primary job and the incidence of hiring

workers from outside the family that varies by gender and initial size.

Page 2: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

1. Introduction

Non-farm microenterprises are prevalent and account for a large share of employment in

low income countries (Gollin 2002, Banerjee and Duflo 2007, Bento and Restuccia 2018). For

example, 80% of manufacturing workers in India were employed by informal firms (Nataraj 2011).

The high prevalence of microenterprises contributes toward lower aggregate productivity in low-

income countries (Hsieh and Klenow 2009, Bento and Restuccia 2018, McCaig and Pavcnik 2018).

Understanding the factors that influence their performance is also crucial for the livelihoods of

families in low-income countries as these microenterprises contribute a sizable share of household

income. Consequently, a large literature examines the effects of various targeted interventions that

aim to relieve supply constraints facing microenterprises ranging from access to credit, business

training, and assistance with business registration (see Banerjee, Karlan, and Zinman 2015,

McKenzie and Woodruff 2014, Bruhn and McKenzie 2014 for respective surveys). Less is known

about how microenterprise performance is affected by nation-wide policies, including trade

agreements. Policy makers often motivate the use of such policies to alleviate poverty in low-

income countries, but much of the studies of the effects of trade policy on firm performance has

been confined to firms in registered manufacturing, thus not capturing the effects of trade policy on

performance of firms in which the poor are more likely to work and potentially missing an

important margin of adjustment (see Harrison and Rodriguez-Clare 2010, Goldberg and Pavcnik

2016 for recent reviews).

In this paper, we examine the effects of output-market shocks (i.e., demand-side) on the

performance of non-farm microenterprises in Vietnam. Non-farm microenterprises play an

important role in Vietnam, accounting for 66% of manufacturing workers, 68% of workers outside

of agriculture, and 23% of household income in 2001/02 (Benjamin, Brandt, and McCaig 2017). We

focus on how new export market opportunities, induced by the 2001 U.S.-Vietnam Bilateral Trade

Agreement (henceforth the BTA) influenced performance of microenterprises in Vietnam.

The BTA and the available data provide an excellent setting to study the consequences of new

export opportunities for microenterprises. The main immediate policy change in the BTA was a

large reduction on U.S. tariffs on imports from Vietnam and these tariff cuts varied substantially

across Vietnamese industries in ways not related to industry-specific political influence and

contemporaneous economic conditions (McCaig 2011, McCaig and Pavcnik 2018). Our research

design uses this differential and arguably exogenous exposure of microenterprises in different

Page 1 of 78

Page 3: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

industries to tariff changes to overcome the inherent identification concern in studying the effects

of international trade on microenterprise performance, whereby forces that influence

microenterprise performance such as urbanization and economic growth also directly impact a

country’s participation in international trade.

We link these declines in export costs with detailed data on microenterprises from the

nationally representative Vietnam Household Living Standards Surveys (VHLSS) that span the policy

change. Our analysis primarily relies on a module on microenterprises with a panel subcomponent.

The definition of a microenterprise in Vietnam is based on the registration status of a firm,

consistent with the practice in studies of microenterprise performance in development economics

in other settings (La Porta and Shleifer 2008, 2014, Nataraj 2011, Ulyssea 2018). The data enables

us to examine the effects of declines in tariffs on several measures of microenterprise performance,

including exit, entry, revenue, hiring of non-family workers, whether a microenterprise has a

business license (a step toward formality), whether the microenterprises is the manager’s primary

job, and whether the business operates for 12 months.

We find that the U.S. tariff reductions are associated with increase exit rates and revenue

growth among the surviving firms. Among female managers, initially smaller businesses are less

likely to exit in response to the tariff reductions and exit becomes more likely for larger female

businesses. In contrast, among male managers, exit is more likely for smaller businesses in response

to the tariff reductions and decreases with business size. Despite large rates of overall entry,

microenterprise entry is not strongly related to the tariff reductions. Among surviving businesses,

we find that revenue increases in response to the tariff reductions, largely due to male-run

businesses. Among female managers, there is an increase in the probability of hiring outside

workers and a decrease in the likelihood of having a license. For male managers, we find evidence

of an increase in the prevalence of having a license. We find little evidence of average effects on the

business being the manager’s primary job or the business operating full-time. These average results

hide significant variation in effects based on initial size. We find revenue contraction among the

initially smallest microenterprises and relative expansion among the larger microenterprises. For

male managers, initially small microenterprises were less likely to be the manager’s primary job

whereas the initially larger ones became more likely to be the manager’s primary job in response to

the BTA.

Page 2 of 78

Page 4: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

These results highlight the importance of looking at the heterogeneous effects of tariff cuts

across microenterprises by size and manager gender for understanding exit and growth. Underlying

business heterogeneity also features importantly in the microenterprise literature (Maloney (2004),

Rogers and Swinnerton (2004), La Porta and Shleifer (2008), De Mel, McKenzie, and Woodruff

(2013)).

Our results, combined with evidence from McCaig and Pavcnik (2018), provide insights on

how export opportunities induce the reallocation of workers between microenterprises and

employers in the formal sector in low-income countries and microenterprise performance.

Importantly, the identification of the effects of the BTA on microenterprise performance is based on

comparisons of businesses in industries that received larger tariff cuts to those in industries with

lower tariff cuts. Thus, our results suggest that export market opportunities expand revenue and

hiring of outside labor toward industries with bigger tariff reductions within the microenterprise

sector. However, it is important to put this in a broader general equilibrium perspective. Evidence

in McCaig and Pavcnik (2018), based on nationally representative labor force data that includes

employment in microenterprises and the enterprise sector, shows that the share of workers

working for microenterprises in an industry declines in response to tariff cuts, consistent with

predictions of models such as Melitz (2003), Gollin (2008), and Lucas (1978). Thus, overall export

market opportunities are expanding employment among the formal enterprises more than among

the microenterprises. While the current paper finds evidence that some microenterprises are more

likely to hire outside labor in response to expanded export market opportunities (perhaps because

they subcontract with larger firms that directly export) and the surviving firms expand revenue,

these effects appear to be dominated by an even greater expansion of employment opportunities

among the formal employers in industries with greater export opportunities. Thus, it appears that

expansion of formal jobs is not occurring because of an absolute contraction of the informal sector,

but instead by a relatively greater expansion of the jobs in the formal sector in response to export

market opportunities.

Our study is related to several literatures. We contribute to the existing literature on

microenterprise performance. The majority of this literature examines the responses of

microenterprises to changes in supply-side constraints, including credit, business training, and

registration assistance (see Banerjee, Karlan, and Zinman 2015, McKenzie and Woodruff 2014,

Bruhn and McKenzie 2014 for respective surveys). Interventions and policy changes that influence

Page 3 of 78

Page 5: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

output markets and the corresponding micro mechanisms of adjustment within and between

microenterprises have received less attention in the literature. We contribute to an emerging

literature that focuses on the effects of demand side policies for microenterprise performance.

Some of these studies have examined the removal of small product reservations on small and

medium enterprises (see Rotemberg 2018, Martin, Nataraj, and Harrison 2017), and trade policy

(Edmonds and Pavcnik 2006, Nataraj 2011, Brambilla, Porto, and Tarozzi 2012, and Atkin,

Khandelwal, and Osman 2017). Most of these studies focus on demand-side interventions that

directly affect microenterprises. Our setting examines a policy change that is expected to directly

benefit firms in the formal enterprise sector (Melitz 2003, Demidova and Rodriguez-Clare 2013).

Consequently, as we discuss in the conceptual framework in section 3, this policy will likely

influence microenterprises through the general equilibrium effects of trade either through

competition in product markets, the opportunity cost of working in a microenterprise and other

options, or through subcontracting. Indeed, McCaig and Pavcnik (2018) find that the BTA was

associated with reallocation of workers from microenterprises to formal registered firms. The

current paper examines how this influences microenterprise performance directly.

We also contribute to the literature on the effects of trade on firm performance. Both

theoretical and empirical studies clearly establish that initially larger, more productive firms expand

and upgrade technology, quality, and or productivity, while smaller firms contract and exit (Melitz

(2003), Verhoogen (2008), Iacovone and Javorcik (2008), Bustos (2011), Lileeva and Trefler (2010)).

These findings are confined to registered manufacturing firms, with Nataraj (2011) and Brambilla,

Porto, and Tarozzi (2012) being notable exceptions. Our study contributes to this literature by

showing whether and how these trade policy changes influence microenterprise performance,

which is important for understanding how trade affects the livelihoods of households in low-income

countries given that microenterprises account for a large share of household income. The nature of

trade policy change matters (Goldberg and Pavcnik 2016). Previous studies have focused on

increased import competition (Nataraj 2011) and increased costs of accessing export markets due

to temporary antidumping duties (Brambilla, Porto, and Tarozzi 2012). Our study complements the

existing work by focusing on a permanent reduction in export costs induced by a trade agreement.

This is important given the belief among policy makers about the importance of increased market

access for poverty reductions in low-income countries implicit in many recent trade agreements,

including the Comprehensive and Progressive Agreement for Trans-Pacific Partnership and the

Page 4 of 78

Page 6: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Doha round of the WTO negotiations. Importantly, combined with McCaig (2011) and McCaig and

Pavcnik (2018), it shows the importance of general equilibrium effects of trade that operate

through the labor market.

The paper proceeds as follows. Section 2 describes the key policy changes in the BTA and its

features that provide the basis for our research design. Section 3 discusses the definition of a

microenterprise in Vietnam and overviews the conceptual framework about how declines in export

costs are expected to affect microenterprise performance. Section 4 describes the data and

characteristics of microenterprises. Section 5 introduces our econometric methodology. Section 6

presents the results and Section 7 discusses the implications of our findings and concludes.

2. Background on the U.S.-Vietnam Bilateral Trade Agreement

The literature on the effects of aggregate demand shocks on microenterprise performance

is scarce. Our study focuses on the consequences of a product-level demand shocks induced by

implementation of a trade agreement. In this section, we describe the U.S.-Vietnam Bilateral Trade

Agreement (BTA) and highlight its key features that we utilize in our empirical methodology and

identification strategy in Section 5. This description is taken from McCaig and Pavcnik (2018).1

The BTA was implemented on December 10, 2001.2 The agreement led to negligible changes

in Vietnam’s import tariff commitments to the U.S. because Vietnam already applied Most Favored

Nation (MFN) tariffs on U.S. imports.3 The main trade policy change was for the U.S to immediately

grant Vietnam Normal Trade Relations (NTR) or MFN access to the U.S. market. Prior to the BTA

Vietnam was subject to tariffs according to Column 2 of the U.S. tariff schedule. With the BTA,

Vietnam became subject to MFN tariff rates. In our analysis, we use industry-level U.S. import ad

valorem equivalent tariffs applied to Vietnamese exports constructed from these two tariff

1 Future versions of the paper will provide a shorter summary of this analysis. 2 See STAR-Vietnam (2003) and McCaig (2011) for an extensive discussion of the BTA. 3 The BTA required Vietnam to reduce import tariffs on approximately 250 (out of approximately 6000) 6-digit HS agricultural and manufactured food products. As these tariff cuts were small in comparison to the U.S. tariff cuts and only affected a relatively small number of products, we do not discuss them in detail. Our results are robust to controlling for these tariff cuts. As part of the BTA, Vietnam was required to implement various regulatory and legal changes over a period of 10 years following the implementation of the BTA. These included commitments to improve market access in services such as banking and telecommunication, intellectual property rights, and protection of foreign direct investment (STAR-Vietnam (2003)).

Page 5 of 78

Page 7: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

schedules by McCaig (2011) as the main policy variable to measure the industry-level policy cost of

accessing export markets.4

Our identification strategy in Section 5 relies on several features of the U.S. tariff declines.

Table 1 summarizes industry tariff levels and changes overall and for broad sectors. First, the U.S.

tariff cuts were large, as the BTA on average reduced tariffs by 20.9 percentage points, from 23.4 to

2.5 percent. The large magnitude of tariff cuts makes it ex ante plausible to separate the effects of

changes in tariffs from confounding changes in the Vietnamese economy. Our empirical

methodology in Section 5 relies on the heterogeneity of tariff declines across industries to identify

the effects of lower exporting costs on performance of microenterprises. Thus, a second useful

feature of the BTA is that the tariff cuts varied widely across industries. As Table 1 suggests, the

standard deviation of the industry tariff decline is 17.9 percentage points. Industries within

manufacturing experienced the largest average tariff cut of 30.2 percentage points, with the

average tariff falling from 33.8 to 3.6 percent.

Importantly, McCaig and Pavcnik (2018) show that these tariff declines significantly affected

the volume and structure of Vietnamese exports to the U.S. and worldwide. During this period,

Vietnam’s aggregate worldwide exports were expanding, but the exports to the U.S. grew even

more. Figures 1 and 2, also reported in Fukase (2013), show the value and the share, respectively,

of Vietnamese exports to the U.S. from 1997 through 2006. The implementation of the BTA led to a

significant surge in exports, which is evident from the break in trend in 2001 in Figure 1. This break

is especially pronounced for manufactured exports, which experienced substantially larger BTA

tariff cuts than primary sector exports.5 Figure 2 indicates that the share of Vietnamese exports

going to the U.S. grew rapidly from 5.1 percent in 2000 to 19.0 percent in 2004 and this increase

was primarily driven by manufacturing, where U.S. exports accounted for 26.1 percent of

Vietnamese exports by 2004.6 The top eight exports to the U.S. according to 2004 value by industry

4 McCaig (2011) uses detailed information on U.S. tariffs for both of these tariff schedules from the U.S. International Trade Commission’s online Tariff Information Center and computes the ad valorem equivalent of any specific tariffs. He then matches the tariff lines to industries by the concordance provided by the World Bank via the World Integrated Trade Solution database to construct industry-level tariffs according to 2-digit ISIC industry nomenclature. This classification closely matches the industry classification in the VHLSSs. 5 Total manufacturing exports also increased following the BTA, as they grew at an annual rate of 23.4% between 2001 and 2006 as compared to 12.8% between 1997 and 2001. The corresponding figures for total exports are 13.1% between 1997 and 2001 and 21.5% between 2001 and 2006. 6 As a non-member of GATT and the WTO, Vietnam was not subject to the Multi Fibre Agreement and did not initially face any export quotas for textile and apparel products destined for the U.S. In July 2003, a bilateral textile

Page 6 of 78

Page 8: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

were apparel; footwear; textiles; food products and beverages; furniture; agriculture; refined

petroleum; and office, accounting and computing machinery.

Figure 3 shows the relationship between growth in exports to the U.S. between 2001 and

2004 and tariff changes across 2-digit ISIC industries. A strong negative relationship suggests that

industries with greater tariff cuts experienced faster export growth. McCaig and Pavcnik (2018) find

that the industry-level regression of the change in log exports to the U.S. between 2001 and 2004

on the change in U.S. tariffs yields a statistically significant estimate of the coefficient on the change

in U.S. tariffs for traded industries and for manufacturing.

Importantly, this BTA-related expansion of U.S. exports is not driven by industry-specific

global demand shocks. Unlike exports to the U.S., Vietnamese exports to the E.U. were already

subject to MFN tariffs prior to the implementation of the BTA (STAR-Vietnam (2003)). As a high-

income export market destination, the E.U. likely faces similar industry-specific demand for low-

income country exports as the U.S. market. To the extent that U.S. tariff changes are correlated

with these shocks, BTA-induced tariff changes would also be spuriously correlated with Vietnamese

exports to the E.U. However, analysis suggests no association between the changes in U.S. tariffs

and changes in Vietnamese exports to the E.U. (see Appendix Table A.1 in McCaig and Pavcnik

(2018)). It is therefore unlikely that BTA-induced tariff changes are spuriously correlated with

industry-specific global demand shocks for Vietnamese goods.7

A fourth useful feature of the U.S. tariff cuts induced by the BTA is that the usual concern

about the political economy of protection and the endogeneity of tariff changes are potentially less

severe. Industry-specific tariff cuts occurred by the U.S. reassigning Vietnam from one pre-existing

tariff schedule to another. Prior to the BTA, imports from Vietnam were covered by Column 2 of the

U.S. tariff schedule, whereas after the BTA they were covered by Most Favored Nation tariffs, also

known as Normal Trade Relations. The Column 2 and MFN tariffs began to diverge in 1951 when the

agreement came into force that imposed quotas on Vietnamese textile and apparel exports to the U.S. This agreement is likely responsible for the reduction in the rate of growth of the share of U.S.-bound Vietnamese manufacturing exports following 2003. In the analysis below, this is one of the reasons why we restrict our period to the two years immediately following the implementation of the BTA. To the extent these quotas affected Vietnamese households in 2003 they would likely attenuate our findings. 7 We obtain qualitatively similar results when we exclude industries whose exports accounted for less than 0.5% of total Vietnamese exports in 2001. We also find qualitatively similar results when we use growth rates as in Davis

and Haltiwanger (1992) as a dependent variable. These growth rates are defined as

and accommodate zero exports in an industry at either the start or end of the period.

1 10.5t t t tg y y y y

Page 7 of 78

Page 9: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

U.S. assigned Vietnam and twenty other communist countries to a list of countries without normal

trade relations. These countries became subject to substantially higher Column 2 tariffs, which were

based on tariffs levels legislated by the Tariff Act of 1930 (Pregelj (2005)). The Column 2 tariff rates

have remained relatively unchanged over time (Pregelj (2005)). Immediately prior to the BTA, the

mean Column 2 tariff across 4-digit HS products remained essentially unchanged, at 31.2 and 31.5

percentage points, respectively in 1997 and 2001, and the correlation was 0.991 (McCaig (2011)).

While the U.S. MFN tariffs have fallen over time, Vietnam was not part of the negotiation process as

a non-member of the GATT and the WTO.

The U.S. tariff cuts were presented as an all-or-nothing package whereby exports from

Vietnam into the U.S. would immediately be covered by MFN tariff rates (negotiated among the

WTO members in a round that concluded by 1995) instead of Column 2 tariffs. The movement of

Vietnam from one pre-existing U.S. tariff schedule to a second pre-existing U.S. tariff schedule

implies that neither U.S. nor Vietnamese industries had an opportunity to influence the tariff cuts

faced by specific industries at the time of the implementation of the BTA.

McCaig and Pavcnik (2018) further confirm the lack of correlation between BTA-induced

tariff changes and pre-existing industry trends and levels. In particular, BTA-induced tariff changes

do not appear to be related to pre-existing trends in Vietnamese exports to the U.S nor to other

high-income destinations such as the E.U. A falsification check of the growth of exports to the U.S.

between 1997 and 2000, where the industry-level pre-BTA tariffs are matched with exports in 1997

and the post-BTA tariffs are matched with exports in 2000, yields a coefficient substantially smaller

in magnitude that is statistically insignificant (see McCaig and Pavcnik (2018) Appendix Table A.1,

Panel B, columns 1 and 2). They obtain a similar finding for growth of exports to the E.U. between

1997 and 2000 (see Appendix Table A.1, Panel B, columns 3 and 4).8 Thus, the export growth to the

U.S. following the BTA is not simply the continuation of pre-existing trends. In addition, McCaig and

Pavcnik (2018) regressed the change in U.S. tariffs on a measure of the unskilled labor intensity of

an industry (measured by the share of workers that completed grade 9 or less) and the share of

workers within the industry working in household businesses prior to the implementation of the

BTA. Across traded, all, and manufacturing industries they find partial correlations of 0.155, -0.120,

and 0.030 for the share of unskilled labor and 0.207, 0.047, and 0.056 for the share of informal

8 A similar regression for worldwide exports between 1997 and 2000 also yields statistically insignificant findings.

Page 8 of 78

Page 10: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

workers. None of the correlations are statistically significant. Overall, neither contemporaneous

growth in demand for Vietnamese exports from other high-income countries, nor pre-existing

trends in industry exports, nor baseline industry characteristics are statistically correlated with the

BTA-induced industry tariff changes.

3. Conceptual Framework

In this section, we first discuss the empirical definition of a microenterprise in Vietnam and how this

definition relates to others used in the literature. We then discuss the channels through which

international trade could affect performance of microenterprises.

3.1. What is a microenterprise in Vietnam

Our definition of a microenterprise is based on a registration status of a firm. This is a

common practice in development economics (La Porta and Shleifer (2008, 2014), Nataraj (2011),

Ulyssea (2018)). Vietnamese law distinguishes whether a firm is a household business or an

enterprise. In Vietnam all state, foreign and collective businesses are legally required to register as

enterprises under Vietnam’s Enterprise Law.9 However, private businesses can legally operate as a

household business or a private enterprise.10 Thus, any private business which is not registered as

an enterprise is, broadly speaking, considered to be a household business. This is our definition of a

microenterprise.

Rules describing in which of the modes the business should be operating are at times vague.

However, businesses that regularly employ workers, employ more than 10 workers, or operate in

more than one location are consistently required to register.11 Thus, while small, single-location

businesses may operate as household businesses or enterprises, all larger businesses are required

to operate as enterprises. The average household business in manufacturing has only 1.5 workers

(including the owner), well below the enterprise employment threshold, and being a household

business does not imply that a business operates illegally. Household businesses can operate in the

9 During our study period, the relevant version of the Enterprise Law is the Law on Enterprises passed in 1999. 10 See Decree No. 02/2000/ND-CP and Decree No. 109/2004/ND-CP. 11 Decrees No. 02/2000/ND-CP of 3 February 2000 and No. 109/2004/ND-CP of 2 April describe household business and enterprise registration requirements during our study period, with the first decree focusing on regular employment and the second on the 10 worker threshold.

Page 9 of 78

Page 11: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

physical premise of a household (or farm), market stalls, industrial zones, trade centers, and in

variable locations (e.g., street vendors).

What are the benefits and costs of a firm being an enterprise as opposed to a household

business? Enterprises, relative to household businesses, have easier access to export licenses,

customs certificates, opportunities to bid on government contracts, the right to open branches and

to operate outside their home district (Malesky and Taussig (2009)). At the same time, running an

enterprise (as opposed to a household business) entails the registration cost and more rigorous

accounting. Taussig and Hang (2004) reports benefits of being an enterprise (relative to household

business) as greater ability to trade beyond home district, ability to expand, value added tax

receipts, legal ability to establish branch locations, a stamp for making transactions more official,

more predictable, law based interactions with government, ability to access equity for limited and

joint stock companies, and greater access to government investment incentives. Costs of

formalization include registration costs, annual registration fee, certified chief accountant, greater

reporting requirements, potential for increased attention from local authorities, and potential for

increased taxes with movement from lump sum to standard tax calculations. They also report that

many laws governing household businesses are the same as those for sole proprietorships, the

simplest form of a company.12

Importantly, our definition of microenterprise in Vietnam is based on a registration status of

a firm, which is a common practice in development economics (La Porta and Shleifer (2008, 2014),

Nataraj (2011), de Mel, McKenzie, and Woodruff (2013)). As a result, out definition is consistent

with other studies of microenterprises in the literature. Importantly, as we discuss in the data

section, our study is rare in as much that it relies on a nationally representative sample of

households (rather than focus on microenterprises in an industry or a small geographic area).

3.2 Performance of microenterprises and declines in export costs

We briefly discuss why tariff reductions on exports from a low-income country

(corresponding to the main trade policy change in the BTA) could affect the microenterprise

12 See CIEM, Assessment of the Strengths and Weaknesses of the Enterprise Law: Recommendations for Amendments and Additions (Draft), 2004, p. 62) for more details. The information on the costs of registering as a private enterprise in Vietnam is further summarized by the World Bank’s Doing Business Survey.

Page 10 of 78

Page 12: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

performance within an industry. This discussion guides the empirical framework and analysis in

Section 5.

A reduction in tariffs on exports from a low-income country will increase product demand

and labor demand in the country. In order to understand the effects these lower tariffs have on

microenterprises, it is useful to first discuss the effects these tariffs in an economy-wide

perspective.

To begin with, reductions in trade costs influence the relative size of industries, as

emphasized in neoclassical trade models. In this particular case, one expects to observe greater

expansions of output and employment in industries with larger tariff cuts relative to industries

subject to smaller declines in trade frictions. In addition, the effects of tariff reductions could differ

across firms within an industry. If firms differ in underlying profitability due to heterogeneity in

marginal costs of production and face a fixed cost of exporting, the reduction in variable export

costs disproportionately raises the profitability of firms with a lower marginal cost of production

(Melitz (2003), Demidova and Rodriguez-Clare (2013)). Firm-specific marginal cost differences might

stem from differences in entrepreneurial ability of the owner/manager (Lucas (1978), Gollin (2008))

or underlying productivity (Melitz (2003)). Microenterprises differ from firms in the enterprise

sector in many dimensions and exhibit substantially lower productivity, owing in part to lower

education or managerial ability of owners and in part due to labor productivity differences (McCaig

and Pavcnik (2018)).13 In this setting, only initially more productive firms benefit from declines in

policy-induced variable export costs because only they earn high enough variable profits from

increased exports to cover the fixed cost of exporting. Declines in tariffs increase product and labor

demand (and profitability) among these more productive firms, while increasing the labor costs and

reducing the profitability of inefficient firms that only serve the domestic market. This is predicted

to shift the composition of market share and employment away from less productive employers

(such as microenterprises) toward more productive firms in the enterprise sector.14

13 McCaig and Pavcnik (2018) show that microenterprises have lower labor productivity than registered enterprises in Vietnam. See Gollin (2008), La Porta and Shleifer (2008, 2014), de Mel, McKenzie, and Woodruff (2013), and Nataraj (2011) for evidence from other countries. 14 Mrazova and Neary (forthcoming) show that the selection effects in Melitz style models are very robust to functional form assumptions and market structure, requiring supermodularity of the profit function in marginal production costs and market access costs (export).

Page 11 of 78

Page 13: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

These models predict that within an industry, trade will expand the employment in initially

bigger, better performing firms and contract employment in initially less efficient firms. Recent

papers using firm-level data show that increased export market access is associated with a

reallocation of market share from less to more productive firms as well as increases in wage

inequality, quality upgrading and technology upgrading in formal urban manufacturing in middle-

income countries (Verhoogen (2008), Iacovone and Javorcik (2008), Bustos (2011)). These benefits

are particularly pronounced for initially more productive, larger firms.

However, these studies do not provide any guidance on the effects of exporting on

microenterprises because they are not observed in their setting. Indeed, McCaig and Pavcnik (2018)

find that labor reallocates toward the enterprise sector, providing direct evidence on the

importance of linkages between these two sectors. Their findings suggest that the probability that a

worker works in a household business (as opposed to a more formal employer) has declined in

response to export opportunities, so that the aggregate share of informal sector employment has

been shrinking in Vietnam in response to the BTA. These effects are most evident among younger

cohorts of workers and in provinces that are more integrated into international markets (as

measured by a province’s proximity to a major seaport).

The evidence in McCaig and Pavcnik (2018) suggests that export opportunities expand the

share of workers working for formal establishments. This appears to largely be happening because

of increasing demand for labor in the formal sector through expansion of existing, larger, formal

firms and the entry of new formal firms and not due to the most successful microenterprises

expanding and transitioning to the formal sector. The current paper provides further evidence on

this process by directly examining how microenterprises adjust their performance to changes in

export costs. Microenterprise performance is of direct interest given their contribution to

employment and household income in a low-income country setting.

Several caveats are worth keeping in mind for interpretation of empirical results. First, a

framework such as Melitz (2003) assumes product-market competition among the firms, implying,

in our context, that microenterprise products are imperfect substitutes for varieties produced by

firms in the enterprise sector, including exported varieties. This is clearly a strong assumption. In

fact, only 1% of household businesses report exporting and they view other non-state private

businesses as their main source of competition (Kokko and Sjoholm (2005)). Hence, the expansion

Page 12 of 78

Page 14: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

of larger, more formal firms in response to increased foreign market access may not increase

product competition faced by household businesses.

Second, even if microenterprises and formal enterprises do not compete in product

markets, exporting could affect employment in microenterprises through the general equilibrium

effects of trade on labor demand. In fact, evidence from Vietnam suggests that exporting

opportunities from the BTA raise wages (McCaig (2011), Fukase (2013)). If microenterprises

compete for labor with firms in the enterprise sector, which disproportionately benefit from

declines in export costs (Melitz (2003)), the increased labor demand among firms in the enterprise

sector increases the opportunity cost of working for a household business, resulting in a relative

contraction of employment in microenterprises (see also Lucas (1978), Gollin (2008)).

A large number of household businesses do not employ outside labor, instead relying on

labor supplied by household members. For example, in Vietnam, only about 10 percent of

microenterprises hire outside labor. Schoar (2010) and Woodruff (2007) suggest that

microenterprises mainly employ household labor in most less developed countries, including

Mexico, Colombia, and Sri Lanka. As such, Melitz-style models, which assume perfectly competitive

labor markets and labor mobility, may not accurately depict the opportunity cost of labor for these

businesses. If the wage rate is not an accurate reflection of the opportunity cost of labor for the

business, the predictions of exit from these models may not apply in this context. Furthermore,

Schoar (2010) and Woodruff (2007) suggest that the existing literature finds that very few

microenterprises create new jobs in the economy through expansion of employment beyond

household members.

Of course, this discussion abstracts from frictions that might impede the mobility of

individuals from microenterprises to the enterprise sector. To the extent that such frictions exist,

they dampen the reallocation in response to declines in export costs, making it more difficult to

detect empirically reallocation across this margin of employment after tariff declines. Likewise,

firms might face different distortions across the two sectors (see Hsieh and Klenow (2009)). This

would lead to lower employment in a sector facing greater distortions than in Melitz (2003) without

distortions. Nonetheless, McCaig and Pavcnik (2018) find evidence consistent with reallocation of

workers toward more formal, productive establishments in response to declines in export costs.

Third, household businesses may operate as subcontractors for larger, more formal

businesses. In Hanoi and Ho Chi Ming City, over 90% of output from household businesses is sold

Page 13 of 78

Page 15: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

either directly to households or to another household business or small enterprise (World Bank

(2010)), although sales to larger enterprises might account for about a quarter of output in

manufacturing household businesses (World Bank (2010)).15 About 10% of household businesses

report subcontracting relationships in 2002 (Kokko and Sjöholm (2005)). While these figures suggest

subcontracting might not play a large role for non-farm household businesses during our sample

period, some household businesses may indirectly benefit from increased demand for products

produced by large firms that benefit directly from the export opportunities.

The above discussion of the theoretical predictions on how expanded export opportunities

will affect performance of household businesses suggests that how informal household businesses

respond to export opportunities is an empirical question.

3. Data

In this section, we describe the two household surveys we use in our analysis, our

procedure for matching businesses across the two surveys, and the key business variables used in

the analysis. Lastly, we discuss the tariff data we match to the household surveys.

We use two nationally representative household surveys, the 2002 and 2004 Vietnam

Household Living Standard Surveys (VHLSS), which were conducted by the General Statistics Office

(GSO) of Vietnam. The surveys were conducted throughout 2002 and 2004 and feature a one-year

recall period, covering 2001/2002 and 2003/2004.16 These surveys cover rural and urban areas, and

contain a household and (thus individual)-level panel subsample that allows one to study individual

employment transitions and better control for unobserved business heterogeneity. Of the 74,350

15 This information is based on household businesses operating in Hanoi and Ho Chi Minh City as covered by the 2007 and 2009 Household Business and Informal Sector surveys conducted by the GSO. 16 The BTA was implemented on December 10, 2001. The 2002 survey interviewed households throughout the year. With a recall period of 12 months, individuals interviewed at the start of 2002 have a recall period that almost entirely precedes the BTA, while individuals interviewed at the end of 2002 have a recall period almost exclusively after the implementation of the BTA. Our results thus potentially underestimate the full impact of the BTA.

Page 14 of 78

Page 16: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

and 45,928 households surveyed in 2002 and 2004 respectively, we can track 22,415 households

and their respective members between the two surveys (McCaig, 2009).17, 18

Each survey contains modules related to household demographics, education, health,

employment, income generating activities, and expenditures. We obtain information on

microenterprises from a detailed module on non-farm private businesses run by the household.

These include household businesses and private enterprises.19 Most of these businesses are not

covered in the more widely available firm-level data of registered firms in Vietnam.

The business module collects information on whether a household operates a business, the

industry in which the business operates, the number of months it operated during the past 12

months, the wage bill, revenue and expenditures, who is the most knowledgeable person (hereafter

referred to as the manager) and whether the business has a license. A household business can be

either licensed or unlicensed.20

Although the household surveys were not directly designed to track businesses over time,

we can do so taking advantage of information on the business that is not likely to change in a short

period. In particular, we use information on the industry of the business and the manager of the

business to match businesses between the 2002 and 2004 surveys. Unfortunately, the 2004 VHLSS

did not report the manager of the business and thus before we can match businesses over time we

17 The decline in the sample size between the 2002 and 2004 VHLSSs is primarily due to a reduction in the number of households surveyed within an enumeration area. The 2002 VHLSS surveyed households within 3,001 enumeration areas averaging approximately 25 households per enumeration area. The 2004 VHLSS surveyed 3,062 enumeration areas, but only 15 households per enumeration area. 18 The VHLSSs feature a rotating panel by enumeration area. Thus, not all enumerations areas surveyed in 2002 were intended to be resurveyed in 2004. This accounts for why the number of panel households is noticeably lower than the total number of households surveyed in the 2004 VHLSS, as only about half of the enumeration areas surveyed in 2004 were surveyed in 2002. 19 The business modules do not distinguish whether a business run by the household is a household business or a private enterprise. In 2004 we can use information on whether the business manager/owner reports working in the private sector (as opposed to the household business sector) to gauge the prevalence of businesses that are private enterprises. Only 1.5 percent of panel businesses could be considered private enterprises by this definition. As a result, we do not explore this margin further. 20 Most of the studies using the 1993 and 1998 Vietnam Living Standards Surveys and the 2002 and 2004 Vietnam Household Living Standards Surveys simply refer to the businesses included in the datasets as non-farm household enterprises (NFHE). In 1993 Vietnam did not distinguish between household businesses and private enterprises. This began sometime later, but the distinction did exist by the 1998 survey. Neither the 2002 nor 2004 survey distinguishes between household businesses and private enterprises. This is probably why researchers use the term non-farm household enterprise since it incorporates both types of businesses covered in the business modules. However, data from the 2006 VHLSS, which distinguishes between household businesses and private enterprises, suggest that even in 2006, a very small share of private businesses, 2.2%, are private enterprises.

Page 15 of 78

Page 17: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

begin by predicting the manager of the business by matching information reported about the

business with information reported in the labor module about workers within the household. A full

description of the process is available in the Data Appendix. We test the procedure using the 2006

VHLSS, which contained the same business information as the 2004 VHLSS but also reported the

manager, and find that our procedure correctly predicts the manager for 92.4 percent of businesses

in the 2006 VHLSS.21 Thus, we feel very confident about our ability to accurately predict the

manager for businesses reported in the 2004 VHLSS. With information on the manager and industry

of operation in both 2002 and 2004, we construct a business panel by first matching businesses

over time within a household by industry and manager. Subsequently, among remaining businesses

we match by either manager or industry within the household. Full details of the procedure can be

found in the Data Appendix. We match 3,821 businesses by industry and manager, 1,272 businesses

by industry only, and 1,038 businesses by manager only leading to a panel of 6,131 businesses. This

represents 84.4 percent of all possible businesses that could be matched over time within panel

households.

We link the business and manager data to detailed information on U.S. tariffs on

Vietnamese exports in a given time period based on the business' industry affiliation. This detailed

tariff data has been collected and previously used by McCaig (2011) and McCaig and Pavcnik (2018).

4. Microenterprise characteristics

Microenterprises employ the majority of workers in low-income countries, but substantially

less is known about their performance relative to the performance of formal firms usually captured

in conventional firm-level data sources. We therefore examine basic microenterprise

characteristics, how they relate to performance, and compare these patterns to those observed for

more formal firms in the existing literature and to the patterns noted by the existing literature on

the topic reviewed above in less developed countries and in Vietnam.

In Table 2, we provide a summary of the number of households operating microenterprises

and the number of microenterprises. 37% of households own a business, with almost 80% of these

households owning only one business, and 19% of these households owning two businesses.

21 We did not use the 2002 VHLSS for testing the algorithm because it did not collect as much information about an individual’s secondary job as the 2004 and 2006 VHLSSs did. Since many businesses are run as a second job, testing the algorithm using the 2006 VHLSS is more appropriate.

Page 16 of 78

Page 18: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Households with more than two businesses are rare. In Table 3, we report summary statistics on

microenterprises. About 70% of businesses operate in services (i.e. tertiary activities), while

manufacturing accounts for the vast majority of the remaining microenterprises. Within

manufacturing, microenterprises are most common in food and beverage production, wood

processing, clothing and apparel, furniture manufacturing, and textiles. 20% of microenterprises

report having a business license and 11% hire outside labor. Microenterprises employ on average

only 1.7 employees and 0.3 paid workers.22 The low number of employees per microenterprise is

consistent with employment patterns in microenterprises from other less developed countries

surveyed in Woodruff (2007) and Schoar (2010).

A large literature on firms operating in the formal sector documents a high degree of

heterogeneity in underlying performance (see Melitz and Redding (2014) for a survey). These

studies do not capture microenterprises. Vietnamese household businesses tend to be smaller (as

measured by revenue or employment) and have lower labor productivity than firms in the

enterprise sector (McCaig and Pavcnik (2018)). Household businesses also exhibit a large degree of

heterogeneity in performance. Figure 4 plots the density of log revenues for household businesses

in 2002 and 2004 and shows large differences among microenterprises. This heterogeneity in

performance motivates the empirical strategy we employ in Section 5, where we explore

differential effects of export market opportunities on performance by household business size.

In addition to heterogeneity in revenue, microenterprises exhibit substantial heterogeneity

in holding a license and hiring workers, two business characteristics associated with more formal

interactions with government officials and labor markets. Table 4 divides microenterprises in 2002

into three size bins, based on whether their revenue lies in the bottom, middle, or the upper third

of the industry revenue distribution in 2002. Larger microenterprises are more likely to hold a

license (37% of large businesses hold a business license, while 6% of small businesses do) and hire

workers from outside of the household (23% of large businesses hire outside labor versus only 3%

of small household businesses). Figures 5 and 6 show nonparametrically that the probability that a

household business holds a business license or hires outside labor increases with revenue. Low-

revenue businesses are very unlikely to hold a business license or hire outside labor, but the

probability of holding a license (or hire outside labor) increases rapidly with household business

22 The information on the number of employed individuals is only available in 2004.

Page 17 of 78

Page 19: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

size. Initially more successful household businesses that do not have a license are also more likely to

obtain a business license in the future. Figure 4 plots the relationship between obtaining a business

license between 2002 and 2004 and revenue in 2002 for household businesses without a business

license in 2002 and shows the positive relationship that is particularly steep for large businesses.

Similarly, Figure 5 plots the probability of hiring outside labor in 2004 versus initial revenue for

businesses that did not hire outside labor in 2002. The relationship is strongly positive, except

among the few businesses with initially high revenues.23 In sum, initially more successful

microenterprises display or eventually adapt more of the characteristics associated with firms in the

registered enterprise sector, but this is nonetheless relatively rare.

The microenterprise sector features large rates of entry and exit (Table 5).24 42 percent of

microenterprises exited between 2002 and 2004 and 42 percent of businesses operating in 2004

were new since 2002. These rates of entry and exit exceed the rates reported for formal firms

(Roberts and Tybout (1996)). High rates of entry and exit of Vietnamese household businesses are

also consistent with evidence in Vijverberg et al. (2006), World Bank (2010) and McCaig and Pavcnik

(2017). Firm exit is highly correlated with firm performance. Exit rates are substantially higher for

smaller firms than for larger firms. Figure 9 plots the propensity that a firm exits against revenue.

Like previous literature on formal firms, we find that the probability of exit declines with firm

revenue. Low-revenue firms face a high probability of exit, which diminishes with size. There is a

slight increase in the probability of exit for the largest firms (followed by a decline).25 Our evidence

is consistent with World Bank (2010) and McKenzie and Paffhausen (forthcoming), which finds that

initially better performing microenterprises and older businesses are less likely to exit.

5. Empirical Implementation

5.1 Business exit

We examine the effect of declines in U.S. tariffs on Vietnamese exports on the performance

of Vietnamese household businesses by linking industry-level tariffs to micro-level data on

23 Almost all microenterprises at such high level of revenue hire outside labor (see Figure 3), so the observed decline could simply reflect a low number of observations and measurement error at high levels of revenue. 24 Our procedure for defining panel businesses (described in the data appendix) is likely overly cautious. This implies that we somewhat overestimate entry and exit. 25 Very few microenterprises have revenue that exceeds 106 million dong so those results might be affected by measurement error.

Page 18 of 78

Page 20: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

household businesses in Vietnam that spans the period of the U.S.-Vietnam BTA. We begin by

exploring whether business exit is related to the tariff reductions. To do so, we use the 2002

business data for panel households and construct an indicator variable for being a business that

exits. We then regress this indicator on the change in tariff within the industry according to the

following framework:

ij j ij ijy tariff X u (1)

where yij is an indicator variable for business i in industry j exiting between 2002 and 2004, jtariff

is the change in the tariff in industry j due to the BTA, and Xij is a vector of business control variables

(an indicator for whether a the business is in an urban area, the gender, age, and education of the

manager, an indicator for whether a manager is an ethnic minority, and province fixed effects). We

perform these regressions for all managers and then separately for female and male managers

given the literature on the existence on performance differences between female and male-owned

businesses.

We report the results in Table 6. We provided separate results for three sets of industries:

traded industries (excluding agriculture, forestry, and aquaculture as the business module covers

non-farm businesses), traded manufacturing, and all industries (we assign a tariff change of 0 to

non-traded industries). On average, within traded industries there is little relationship between the

size of U.S. tariff cuts and the probability of exit. This is the case for all managers, female managers,

and male managers. However, within manufacturing there is evidence of an increase in the

probability of exiting in response to the tariff reductions, particularly for businesses run by males.

The average tariff reduction, of 30.4 percentage points, is associated with an 8.1 percentage point

increase in the probability of exit for manufacturing businesses run by males. The aggregate exit

rates for male-run manufacturing businesses was 36 percent. For all industries, where the tariff

change has been set to 0 for businesses in non-traded industries, the results suggest that exit is

higher in response to larger tariff reductions for female-run businesses. This is largely due to a

lower exit rate among female-run retail businesses than for most traded industries.

The models in Section 3 suggest that the tariff cuts should affect firms differentially

depending on their initial performance. We expand the baseline specification to consider whether

there are differential effects of new export opportunities (as measured by U.S. tariff cuts) on

microenterprise exit depending on their initial size, as Melitz (2003) would predict. We augment the

Page 19 of 78

Page 21: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

specification in (1) to allow for the effect of tariffs to differ across initially small, medium, and large

household businesses in a method similar to Bustos (2011):

3 3

2 2

r r r r

ij j ij j ij ij ij

r r

y tariff D tariff D X u

(2)

where is an indicator for whether a firm is in the bottom third (small), middle (medium), or

upper third (large) of the 2002 firm revenue distribution within an industry. The interaction terms

allow for differential effects within an industry and correspondingly and measure the effect

of tariffs on medium and large businesses, respectively, relative to small businesses. Thus, a positive

value for , for example, implies a decrease in the probability of exit for large microenterprises

relative to small microenterprises in response to the U.S. tariff reductions. The overall effect for

large businesses would then be 3 . In addition, the specification controls for potential

differential exit across small, medium, and large household businesses by the inclusion of indicators

for the initial microenterprise size.

We use three different specifications based on heterogeneity in initial size. Within an

industry, we divide firms into three equal sized bins of small, medium, and large firms based on

initial revenue and separately based on initial expenses. As both revenue and expenses are likely

reported with error, our size bins will have non-classical measurement error that will lead to

attenuation bias (Pischke, 2007). Our third specification uses the expense based size bins as

instrumental variables for the revenue based size bins. The IV estimator is not consistent in this

context, but its properties are such that the OLS and IV results can bound the true coefficient

(Pischke, 2007).

We report the results for female and male managers combined in Table 7a. For traded

industries, we find little evidence of differential exit rates by size in response to the tariff

reductions. Within manufacturing, we find evidence that large businesses are more likely to exit in

response to tariff reductions than small and medium businesses when size is based on initial

revenue. The estimate suggests that exit was 7 percentage points higher in initially large

microenterprises than in initially small microenterprises in response to the average manufacturing

tariff reduction. However, the magnitudes of the estimates are much smaller using expense based

size bins and the IV approach and are not statistically different from 0. Across all industries, we find

some evidence suggestive of exit being less likely among initially large businesses relative to initially

r

iD

2 3

3

Page 20 of 78

Page 22: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

small businesses. The differential pattern is more noticeable using expense-based size definition

than for the revenue-based size definition.

Again, we see signs of important differences across female and male managers in terms of

exit (Tables 7b and 7c). Based on businesses in manufacturing, small, female-run businesses are less

likely to exit in response to tariff reductions whereas small, male-run businesses are more likely to

exit. For female-run manufacturing businesses, they become more likely to exit in response to the

tariff reductions as the business gets larger whereas male-run businesses are less likely to exit as

the business increases in size, particularly for medium sized businesses.26

Very few existing studies on microenterprises examine exit as an outcome variable,

especially relating it to plausibly exogenous variation to demand shocks (see McKenzie and

Pauffhausen (forthcoming) for discussion of exit in this literature). The above results are thus

informative about market-level factors (in our case, changes in product demand induced by changes

in trade policy) in influencing microenterprise exit. Moreover, the differential effects of tariff

reductions on exit by manager gender and initial size of the business imply that the sample of

surviving firms will differ based on the gender and initial performance of the business. This will have

implications for how to interpret evidence of whether the tariff reductions influenced firm growth.

However, before we turn to focusing on firm growth among surviving firms, we examine firm entry

in response to the tariff reductions.

26 In future, we will expand the exit analysis in three important ways. First, we will provide details on what the managers of exiting businesses are doing subsequent to closing the business. Using data from the 2004, 2006, and 2008 VHLSSs, McCaig and Pavcnik (2017) report that among former managers, 31% are self-employed in agriculture and 18% have left the household. Second, we will explore the sensitivity of our exit results to our business panel by aggregating over all businesses within a household and subsequently examining business exit at the household level. Third, using a special module in the 2004 VHLSS, we will provide some descriptive evidence on the reasons provided by managers as to why they closed their business. The most common reason given for closing a business is other activities creating more income.

Page 21 of 78

Page 23: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

5.2 Entry

To empirically examine whether entry is influenced by the tariff reductions, we employ

equation (1) using the sample of businesses operated by panel households in 2004. The dependent

variable is an indicator variable for whether the business entered between 2002 and 2004 and the

change in tariff is based on the industry of operation in 2004.

Across traded and manufacturing industries, we see little evidence of entry in response to

the BTA tariff reductions (Table 8). The estimates are small in magnitude and statistically

insignificant for all managers as well as for female and male managers separately. For female

managers, we see some evidence of an entry response to tariffs for all industries (where non-traded

industries have been assigned a tariff change of 0). Together with the results for traded and

manufacturing, this implies lower entry rates in non-traded industries for female managers.

Together, the results for exit and entry suggest that industries that received larger tariff cuts

experienced net exit, but with different patterns of net exit between female and male managers.

Among male managers, initially smaller businesses were more likely to exit whereas for female

managers, initially smaller businesses were less likely to exit in response to the BTA tariff

reductions.

5.3 Business growth and transitions

Among surviving businesses, we relate outcome y of business i in industry j at time t to

industry tariffs using the following framework:

ijt jt i t ij ijty tariff D X u (3)

where yijt is the outcome (ln revenue, and indicators for hiring outside labor, having a license, being

the manager’s primary job, and operating for 12 months) and tariffjt is in tariff in industry j with the

Column 2 tariffs assigned to 2002 and the MFN tariffs assigned to 2004 survey year and industry is

based on the business’s initial industry. We include Xij, a vector of initial business characteristics,

which includes an indicator for the business being in an urban area, the gender, age, and education

of the manager, and an indicator for whether a manager is an ethnic minority, as well as province

fixed effects, interacted with a 2004 dummy, Dt. Lastly, we include business fixed effects. The

business fixed effects in equation (3) imply that all time invariant factors for business outcomes in

level have been controlled for. Furthermore, the inclusion of initial business characteristics and

Page 22 of 78

Page 24: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

province fixed effects interacted with a 2004 dummy allow for differential changes over time in

business outcomes in relation to these characteristics.

We find evidence of revenue growth in response to the tariff reductions, but little evidence

of other changes for all managers, as reported in Table 9a. In manufacturing industries, the

coefficient is negative, large in magnitude, and statistically significant. A surviving business in an

industry that experienced the average tariff change of -0.304 experienced faster revenue growth of

8% compared to no tariff change. In contrast, the estimates for almost all other outcomes are small

and statistically insignificant. The differences across female and male managers (Tables 9b and 9c)

continue among surviving businesses. Revenue growth in response to the tariff reductions are

larger in manufacturing for male managers (-0.466 versus -0.128).

In addition to revenue, we consider two outcomes that are in general associated with steps

toward formalization—hiring outside labor and having a business license. Although licensed

household businesses are not registered enterprises (i.e., formal), obtaining a business license is

often viewed as a step toward greater formality for a firm. Indeed some earlier studies consider

only household business without a license as informal (Cling et al. (2010), p. 6, 2012).27

We also see differences in responses between male and female managers for hiring outside

labor and having a license. Female managers increased the prevalence of hiring outside labor. The

estimate in manufacturing suggests an increase of 2.6 percentage points in the incidence of hiring

outside labor in response to the average manufacturing tariff reduction. Although this may seem

small, recall that only about 10% of businesses report hiring outside workers. Female managers

27 Information on the costs of obtaining a business license is not readily available, but costs do not seem to be a significant barrier. Only a small percentage of microenterprises without a license consider the lack of license being related to expense or complicated nature of the licensing process (World Bank (2010)) and World Bank (2009)). Among those without a license, the vast majority reports they are not registered either because it is not compulsory or because they don’t know if they need to register (World Bank (2010)). A web site aimed at the business community (http://www.vietnam-briefing.com/news/vietnams-taxes-business.html/) suggests that the fee varies by monthly income of the business from 50,000 to 1,000,000 dong per year, but the site does not list the original source for this information nor the relevant period. Household businesses that hold a license report the main advantage of having a license as less corruption, followed by better access to market places, and easier loan access (World Bank (2010)). One possible explanation for not obtaining a business license is that a requirement of being licensed is to pay taxes. However, Cling et al. (2012) find that some unlicensed businesses report paying taxes and most household businesses also make additional payments to public officials. Therefore avoiding taxes by being unlicensed may not lead to an overall decrease in payments to government agencies. Despite imperfect adherence to the law obtaining a business license is considered as a first move toward formality (Cling et al. (2010, 2012)).

Page 23 of 78

Page 25: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

decreased the likelihood of having a license and male managers increased the likelihood, although

the latter estimate is a similar magnitude but not precisely estimated.

Similar to the exit analysis, we expand the specification in equation (3) to consider whether

there are differential effects of new export opportunities (as measured by U.S. tariff cuts) on

business performance depending on their initial size:

3

2

r r

ijt jt i jt i t ij ijt

r

y tariff D tariff D X u

(4)

where is an indicator for whether a firm is in the bottom third (small), middle (medium), or

upper third (large) of the 2002 firm revenue distribution.28 Additionally, the vector of control

variables is expanded to included interactions between initial size and a 2004 dummy to control for

differentials changes across businesses of different sizes. For brevity, we focus on heterogeneous

results for ln revenue as the outcome. Tables for the other outcomes are reported in the appendix

(Tables B1 to B4).

We find strong evidence of differential responses in revenue based on initial size. We

display the results for all managers in Table 10a. As in previous tables, we report the results for

traded industries, manufacturing industries, and all industries. The tariffs are based on the initial

industry of operation, as is the sample inclusion into the three industry samples. Across traded and

manufacturing industries, we consistently find statistically significant evidence of heterogeneous

responses in revenue to the U.S. tariff reductions. Small businesses experienced a greater decline in

revenue in industries with larger tariff cuts when size is based on initial revenue or initial expenses.

Medium sized and large businesses experienced faster revenue growth in response to the tariff

reductions than initially small businesses within their industry. The total effect for medium size

business is not statistically different from 0 using either initial revenue or initial expenses to define

the size. However, the total effect for large businesses is statistically different from 0 using initial

expenses to define the size. The average manufacturing tariff reduction (-0.304) is associated with a

26% reduction in revenue for initially small businesses and a 14% increase in revenue in initially

large businesses using the expense based size definition.

28 These indicators are based on the position of a household business in the industry’s revenue distribution in 2002. We also construct similar indicators based on initial expenses in 2002.

r

iD

Page 24 of 78

Page 26: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Finally, the magnitude of the coefficients is larger in traded and manufacturing industries,

which are more directly impacted by trade, than in the all industry sample that also includes

businesses in non-traded industries. A large existing literature has found evidence of the

importance of heterogeneity in responses of firms in the formal sector to export market

opportunities (Melitz and Redding (2014)). Our evidence suggests that heterogeneity in responses

to export market opportunities plays some role even among microenterprises. The increase in

revenue in industries with bigger tariff cuts found in the previous section is driven by revenue

growth in initially larger microenterprises, which corresponds to a reallocation of revenue from

small to larger firms within an industry. Initially small microenterprises are contracting revenue with

tariff cuts, while initially larger microenterprises appear to be growing relative to smaller

microenterprises.

By gender, we see strong evidence of heterogeneous responses by size for both female and

male managers (see Tables 10b and 10c). The results are very similar across female and male

managers when size is based on initial expenses. When sized is based on initial revenue, the

heterogeneity between small and large microenterprises is greater for females than for males.29

In Appendix Tables B1 through B4, we display regression results for the other business

outcomes (hiring outside labor, having a license, being the manager’s primary job, and operating for

12 months) by initial size and gender. We find evidence of female managers of medium and large

businesses becoming more likely to hire outside labor, female and male managers of medium

microenterprises being less likely to have a license, some evidence of small businesses being less

likely to be the manager’s primary job and of large businesses being more likely to be the manager’s

primary job, and some (imprecise) evidence of an increase in the likelihood of operating for 12

months among medium sized microenterprises. However, it should be noted that the results are

not always consistent across the revenue and expense based size definitions.

6. Discussion of results and conclusion

Using nationally representative surveys of microenterprises in Vietnam, we document large

differences in revenue across microenterprises in Vietnam and find that initially larger businesses

29 In future analysis, we will also explore whether the results are different between provinces that are close versus far away from major seaports, as in McCaig and Pavcnik (2018).

Page 25 of 78

Page 27: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

are more likely to hire outside (i.e. non-household member) labor, hold a business license, and

survive. This underlying heterogeneity matters for how these businesses respond to expanded

export market opportunities induced by the U.S.-Vietnam Bilateral Trade Agreement.

Our results suggest that microenterprises exited in response to the tariff cuts, particularly

those run by males. However, the pattern of exit differs across female and male managers. Small

female-managed businesses were less likely to exit whereas small male-managed businesses were

more likely to exit and large female-managed businesses were more likely to exit in response to the

tariff reductions. The increased exit among larger female-run businesses may be indicative of these

managers experiencing the greatest increase in their outside opportunities as better educated

managers run larger businesses and better educated individuals are more likely to transition out of

the informal sector (McCaig and Pavcnik, 2015).

Among surviving microenterprises, the tariff reductions are associated with an expansion of

revenue, particularly among male managers. We also find some evidence that female managed

businesses are more likely to hire outside workers, but less likely to hold a license in response to

declines in tariff cuts. As for exit, the initial size of the business matters as revenue growth was

negative for initially smaller businesses and either less negative or positive for larger businesses.

These results highlight the importance of looking at heterogeneous effects of demand

shocks across microenterprises for understanding how the sector responds. At the same time, our

results also suggest that the changes in aggregate demand do not necessarily influence all examined

outcomes of microenterprises. Part of the lack of findings might be due to imprecise estimates even

though we are relying on a large, nationally representative sample of households, which is rare in

this literature (see McKenzie and Paffhausen (forthcoming) for a review of sample sizes in recent

RCT studies of microenterprises).

However, absence of formal firms in the analysis has to be taken into account when

interpreting the findings. In our analysis above, we, like existing literature on microenterprises,

focuses on the microenterprise sector in isolation from the rest of the economy. This is due to the

lack of surveys that simultaneously cover firm-level outcomes for both sectors. In addition, it might

be sensible to focus on the microenterprise sector on its own, to the extent that these firms tend to

operate in a more similar market environment than their formal counterparts. Detailed surveys of

microenterprises enable one to examine how these microenterprises adjust to aggregate demand

Page 26 of 78

Page 28: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

shocks, which is important given their large contribution to household income in low-income

country settings.

Nonetheless, the focus on microenterprises alone has to be taken into account when

interpreting the findings. In particular, our evidence on expanding revenue and increased hiring of

outside labor in industries with larger tariff cuts might be at first surprising given that Melitz style

models would predict that these microenterprises should contract. However, one has to be careful

in interpreting our results. The identification of the effects of the BTA on microenterprise

performance is based on comparisons of performance of these businesses in industries that

received larger tariff cuts to those in industries with lower tariff cuts. Thus, our results suggest that

export market opportunities expand revenue and hiring of outside labor toward industries with

bigger tariff within the informal sector. However, we cannot rule out that household business

revenue and hiring of outside labor is declining in response to export opportunities relative to

formal firms within the industry. One could examine this claim more directly for revenue and exit by

using data that would simultaneously cover formal firms in the enterprise sector and household

businesses, but this is a topic for future work.

However, our results combined with evidence from McCaig and Pavcnik (2018) provide

insights on how export opportunities induce the reallocation of workers between household

businesses and employers in the formal sector in low-wage countries. Evidence in McCaig and

Pavcnik (2018), which is based on labor force data from the VHLSSs, which is representative of

employment in household businesses and the enterprise sector, clearly shows that the share of

workers working for household businesses in an industry declines in response to tariff cuts. Thus,

for the case of labor, export market opportunities are expanding employment among the formal

enterprises more than among the household businesses. These effects were particularly

pronounced for younger workers and for workers in provinces proximate to major seaports. While

the evidence in the current paper suggests that microenterprises in industries with larger tariff cuts

are more likely to hire outside labor and increase revenue, the evidence from McCaig and Pavcnik

(2018) suggest that new export opportunities increased employment opportunities in the formal

sector proportionally more.

Overall, this evidence is in line with La Porta and Shleifer (2008) who show that the level of

informality in an economy generally declines with economic development through the growth of

Page 27 of 78

Page 29: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

existing formal firms and the decline of informal firms rather than formalization of firms in the

informal sector. 30

30 McCaig and Pavcnik (2018) note that worker switching from working for a household business to working for an employer in the formal sector often coincides with a change in industry affiliation. In addition, entrants into the labor force are more likely to begin working for a formal employer, while those that exit the labor force are more likely to have previously worked for a household business. All these facts are inconsistent with the aggregate share of informal workers declining due to formalization of previously unregistered household businesses.

Page 28 of 78

Page 30: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

References

Atkin, David, Amit Khandelwal, and Adam Osman. 2017. “Exporting and Firm Performance: Evidence

from a Randomized Trial,” Quarterly Journal of Economics, 132 (2): 551-615.

Banerjee, Abhijit V. and Esther Duflo. (2007). “The Economic Lives of the Poor” Journal of Economic

Perspectives, 21(1), pp. 141-168.

Banerjee, Abhijit V., Dean Karlan, and Jonathan Zinman. 2015. “Six randomized evaluations of

microcredit: Introduction and further steps,” American Economic Journal: Applied Economics, 7(1),

p. 1-21.

Benjamin, Dwayne, Loren Brandt and Brian McCaig. 2017. "Growth with equity: Income inequality

in Vietnam, 2002-14" Journal of Economic Inequality, 15 (1): 25-46.

Bento, Pedro and Diego Restuccia. 2018. "On Average Establishment Size across Sectors and

Countries," NBER Working Papers 24968.

Brambilla, I., G. Porto, and A. Tarozzi, 2012. "Adjusting to Trade Policy: Evidence from U.S.

Antidumping Duties on Vietnamese Catfish," The Review of Economics and Statistics, Vol. 94, No. 1,

pp. 304-319.

Bruhn, Miriam and David McKenzie. 2014. “Entry regulation and the formalization of

microenterprises in developing countries,” The World Bank Research Observer, 29 (2): 186-201.

Bustos, P. (2011). "Trade liberalization, exports, and technology upgrading: Evidence on the Impact

of MERCOSUR on Argentinean Firms," American Economic Review, 101(1), 304-40.

Cling, Jean Pierre, Mireille Razafindrakoto, and François Roubaud. 2010. “The informal economy in

Vietnam.”

Page 29 of 78

Page 31: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Cling, Jean-Pierre, Mireille Razafindrakoto, and François Roubaud. 2012. “To be or not to be

registered? Explanatory factors behind formalizing non-farm household businesses in Vietnam,”

Journal of the Asia Pacific Economy, 17(4), 632-652.

Davis, Steven J. and John Haltiwanger. (1992). “Gross job creation, gross job destruction, and

employment reallocation,” The Quarterly Journal of Economics, 107(3), pp. 819-863.

De Mel, Suresh, D. McKenzie, and C Woodruff (2013). “The Demand for, and Consequences of,

Formalization among Informal Firms in Sri Lanka,” American Economic Journal: Applied Economics

5(2): 122-150.

Demidova, S. and A. Rodríguez-Clare. (2013). “The simple analytics of the Melitz model in a small

open economy,” Journal of International Economics, 90(2), 266-72.

Edmonds, Eric and Nina Pavcnik. 2006. “Trade Liberalization and the Allocation of Labor between

Households and Markets in a Poor Country,” Journal of International Economics 69 (2): 257-295.

Fukase, Emiko. 2013. “Export liberalization, job creation and the skill premium: Evidence from the

U.S.-Vietnam Bilateral Trade Agreement,” World Development, 41, 317-337.

Goldberg, Pinelopi and Nina Pavcnik. 2016. “The effects of trade policy,” in Handbook of

Commercial Policy.

Gollin, Douglass. 2008. “Nobody’s business by my own: Self-employment and small enterprise in

economic development,” Journal of Monetary Economics 55, 219-233.

Hsieh, Chang-Tai and Peter J. Klenow (2009). “Misallocation and manufacturing TFP in China and

India,” Quarterly Journal of Economics, Vol. 124, No. 4, pp. 1403-1448.

Hsieh, Chang-Tai and Peter J. Klenow. 2014. “The life cycle of plants in India and Mexico,” Quarterly

Journal of Economics, 129(3): 1035-1084.

Page 30 of 78

Page 32: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Iacovone, L. and B. Smarzynska Javorcik (2008). "Shipping Good Tequila Out: Investment, Domestic

Unit Values and Entry of Multi-product Plants into Export Markets" University of Oxford, mimeo

Kokko and Sjöholm. 2005. “The internationalization of Vietnamese small and medium-sized

enterprises,” Asian Economic Papers, 4(1).

La Porta, R. and A. Shleifer. 2008. “The unofficial economy and economic development,” Brookings

Papers on Economic Activity, Economic Studies Program, The Brookings Institution, Vol. 39, No. 2,

pp. 275-363.

La Porta, R. and A. Shleifer. (2014). “Informality and development,” Journal of Economic

Perspectives, 28, 109-26.

Lucas, Robert (1978). “On the Size Distribution of Business Firms,” The Bell Journal of Economics,

9(2), pp. 508-523.

Malesky, Edmund and Markus Taussig. 2009. “Out of the gray: The impact of provincial institutions

on business formalization in Vietnam,” Journal of East Asian Studies, 9, 249-290.

Martin, Leslie A., Shanthi Nataraj, and Ann E. Harrison. 2017. "In with the Big, Out with the Small:

Removing Small-Scale Reservations in India." American Economic Review, 107 (2): 354-86.

McCaig, Brian. 2009. “The reliability of matches in the 2002-2004 Vietnam Household Living

Standards Survey Panel,” The Australian National University Centre for Economic Policy Research

Discussion Paper No. 622.

McCaig, Brian (2011). “Exporting out of poverty: Provincial poverty in Vietnam and U.S. market

access.” Journal of International Economics, 85(1), pp. 102-113.

Page 31 of 78

Page 33: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

McCaig, Brian and Nina Pavcnik. 2015. “Informal employment in a growing and globalizing low-

income country,” American Economic Review, 105(5), pp. 545-50.

McCaig, Brian and Nina Pavcnik. 2018. “Export markets and labor allocation in a low-income

country,” American Economic Review, 108(7), pp. 1899-1941.

McCaig, Brian and Nina Pavcnik. 2017. “Out with the old and unproductive, in with the new and

similarly unproductive: Microenterprise dynamics in a growing low-income economy,” unpublished

manuscript.

McKenzie, David and Anna Luisa Paffhausen. forthcoming. “Small firm death in developing

countries,” Review of Economics and Statistics.

McKenzie, David and Christopher Woodruff. 2014. “What are we learning from business training

and entrepreneurship evaluations around the developing world?” The World Bank Research

Observer, 29(1), p. 48-82.

Melitz, M. (2003): “The Impact of Trade on Intra-industry Reallocations and Aggregate Industry

Productivity,” Econometrica 71, 1696-1725.

Melitz, Marc J. and Stephen Redding. 2014. “Heterogeneous Firms and Trade,” Handbook of

International Economics, 4th ed., 4: 1-54. Elsevier.

Nataraj, S. (2011). “The impact of trade liberalization on productivity: Evidence from India’s formal

and informal manufacturing sectors,” Journal of International Economics, 85(2), 292-301.

Pages, Carmen. (2012). “What are the most important research questions on the economics of

growth and labour market outcomes in low-income countries?” First GLM|LIC workshop: Setting

the research agenda.

Page 32 of 78

Page 34: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Pregelj, V. (2005). “Normal-Trade-Relations (Most-Favored-Nation) Policy of the United States,”

Congressional Research Service Report RL31558.

Roberts, M. and J. Tybout (1996). Industrial Evolution in Developing Countries: Micro Patterns of

Turnover, Productivity, and Market Structure. Oxford University Press.

Rotemberg, Martin. 2018. “Equilibrium effects of firm subsidies,” unpublished manuscript.

Schoar, A. (2010). “The Divide between Subsistence and Transformational Entrepreneurship.”

Innovation Policy and the Economy, Volume 10 (pp. 57-81). University of Chicago Press.

STAR-Vietnam. (2003). “An Assessment of the Economic Impact of the United States – Vietnam

Bilateral Trade Agreement.” The National Political Publishing House: Hanoi, Vietnam.

Taussig, M. and Hang, P. T. T. (2004). Private enterprise formality and the role of local government.

Asian Development Bank Discussion Paper No. 2.

Ulyssea, Gabriel. 2018. "Firms, Informality, and Development: Theory and Evidence from Brazil."

American Economic Review, 108(8):2015-47.

Verhoogen, E. (2008). "Trade, Quality Upgrading, and Wage Inequality in the Mexican

Manufacturing Sector," Quarterly Journal of Economics, vol. 123, no.2, pp. 489-530.

Vijverberg, Wim, Hoang, T. T. H., Nguyen, C.T., Nguyen, N.Q., Nguyen, T.Q., Phung, D.T., and Vu,

T.K.M. 2006. “Non-farm household enterprises in Vietnam: A research project using data from

VHLSS 2004, VHLSS 2002 and AHBS 2003,” Research report. Hanoi: The World Bank.

Woodruff, C. (2007). “Self-employment: Engine of Growth or Self-help Safety Net?” Employment

and Shared Growth: Rethinking the Role of Labor Mobility for Development. Washington, DC: The

World Bank, 53-68.

Page 33 of 78

Page 35: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Woodruff, Christopher. 2012. “The other half: What we do(n’t) know about self-employment in

LICs?” First GLM|LIC workshop: Setting the research agenda.

World Bank (2009). “Vietnam development report 2010: Modern institutions.” Washington, D.C.:

The World Bank.

World Bank (2010). “Dynamics of the informal sector in Hanoi and Ho Chi Minh City 2007-2009:

Main findings of the Household Business & Informal Sector survey (HB&IS),” Washington, D.C.: The

World Bank.

Page 34 of 78

Page 36: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Appendix A: Data

In this appendix, we describe in detail two data preparation steps. First, we describe how

we predict a manager for all businesses in the 2004 VHLSS. Unlike the 2002 VHLSS, respondents in

2004 were not asked to identify the most knowledgeable individual for the business (hereafter

referred to as the manager for brevity). Knowing the manager of the business in 2004 has three

advantages: (1) in the employment module the individual reports whether they were self-employed

in an enterprise or self-employed in a household business and thus we can use information from

the employment module to identify whether a business is a household business or a private

enterprise; (2) in the employment module the individual also reports how long they have been

doing the job and thus we can infer a possible year of start for the business; and (3) it provides

additional information for the business which can be used to help create a panel at the business

level. To test the accuracy of the manager prediction algorithm we also run it for the 2006 VHLSS

which contains the same individual and business information that we use for the 2004 VHLSS. The

algorithm correctly predicts the manager for 91.2% of businesses in the 2006 VHLSS.

Second, we explain how we match businesses between the 2002 and 2004 VHLSSs. The

surveys were not designed to directly create a panel of businesses. However, we use a combination

of information on the manager and the industry of operation of the business to match them over

time within a panel household. In total, we match 6,131 businesses out of a maximum possible

number of matches of 7,261.

A.1 Predicting the manager for businesses in the 2004 VHLSS

In this subsection, we provide a detailed description of the data available in the

employment and business modules of the 2004 VHLSS, the algorithm used for matching, a summary

of how the matches were made, and the percentage of successful predictions from using the same

algorithm on data from the 2006 VHLSS.

We combine data from the employment and business modules of the 2004 VHLSS that can

be matched. In particular, from the employment module we identify individuals that reported being

self-employed in a household business for either their primary or secondary job during the past

year. For these jobs, we use information on the industry, the number of months worked during the

past years, the number of days per month usually worked, and the number of years the individual

Page 35 of 78

Page 37: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

has been doing the job. From the business module, we use information on the industry, the number

of months operating during the past year, the average number of days per month operating, and

the year the business started.31

In Table A1 we provide a summary of the matches by the step within the manager

prediction algorithm at which the match was made. The table is organized sequentially such that

the first step of the algorithm was to identify the manager for businesses in which only one

household member reported being self-employed in the industry of the business and then only

businesses without a predicted manager would proceed to the next row. The first step of the

algorithm matches an individual as the manager for the business for 70.5% of all businesses in the

2004 VHLSS. The corresponding rate of success using the 2006 VHLSS is 99.3%. Thus, for a large

share of businesses we have a very high degree of confidence in our predicted manager. Next, we

identified a manager for any remaining businesses when there was only one household member for

whom the number of years in the job, the number of months worked in the past year, and the

number of days per month matched. And so on down the rows of the table.32 In sum, the algorithm

correctly identified the manager for 92.4% of businesses in the 2006 VHLSS. Thus, our manager

prediction algorithm is doing a very good job of identifying the manager of the business.33

A.2 Matching businesses between the 2002 and 2004 VHLSSs

Not all businesses run by a panel household should be matched over time. For example, any

household that reports running a different number of businesses across the two years has

experienced net entry or exit of businesses and thus at least one business within the household

31 The year the business started is only available for about 1/5th of the sample since this question was not asked of all businesses, but instead was part of an extra module on businesses that only 1/5th of households were asked. 32 Note that the percentage of successfully identified managers in the 2006 VHLSS for “Only household member with matching months and days per month” is likely an underestimate of the rate for the 2004 VHLSS. This is because only about 1/5th of businesses in the 2004 VHLSS have information on the year when the business started whereas all businesses in the 2006 VHLSS have this information. Thus some 2004 businesses for which the year was not reported, but the number of years, months, and days all matched would only be matched in the row “Only household member with matching months and days per month”. Indeed, in the 2006 VHLSS 11.3% of businesses are matched in the step “Only household member with matching years, months, and days per month” as compared to only 1.8% in the 2004 VHLSS and 3.2% of 2006 VHLSS businesses were matched to a manager in the step “Only household member with matching months and days per month” as compared to 9.5% in the 2004 VHLSS. 33 Our algorithm does not predict a manager for 595 out of 21,458 (2.8 percent) businesses. This could be due to the business being managed by an individual as their third job, which our algorithm currently does not include, or due to measurement error either in the industry of the business or the industry of the job.

Page 36 of 78

Page 38: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

should not be matched. Thus, for any given the household the maximum number of matched

businesses is the minimum of the number of businesses run in either year. Table A2 summarizes the

number of businesses run by panel households in 2002 and 2004. There are 22,415 panel

households in our dataset. A little over half of the households did not operate a business in 2002 or

2004. The number of businesses that can potentially be matched is 7261.34

The most valuable information that we have for matching businesses over time is the

manager and the industry of operation. We begin by matching businesses within a household by

industry-manager and find 3,821 matches. These represent businesses that have a unique industry-

manager combination within the household in both years and the combination existed in both years

(e.g., the same manager operated a business in the same industry in both years and did not manage

any other businesses in the same industry in either year). Note that this will include instances in

which a manager closed one business and opened a new business in the same industry, but 97.2

percent or predicted managers in 2004 report doing the job for at least 2 years, suggesting that

most of these businesses are indeed continuing businesses.

Next, among the remaining businesses we relax the matching criteria to be (1) matched just

by manager, which allows for industry switching and (2) matched just by industry, which allows for

the manager within the household to change. Table A.3 summarizes the outcomes from all three

steps for matching businesses.

34 This is derived by summing over min(i,j)*aij where i represents the number of businesses run by the household in 2002, j is the number of businesses run by the household in 2004, and aij is the number of households operating i businesses in 2002 and j businesses in 2004.

Page 37 of 78

Page 39: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 1: Value of Vietnamese exports to the U.S., 1997 to 2006

Notes: Authors' calculations from COMTRADE.

Page 38 of 78

Page 40: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 2: Share of the United States in Vietnam's Exports

Notes: Authors' calculations from COMTRADE.

Page 39 of 78

Page 41: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 3: Growth of Vietnamese exports to the US versus US tariff cuts by industry

Page 40 of 78

Page 42: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 4: Density of ln revenue for 2002 and 2004

Notes: All values are expressed in 2004 prices. The sample includes all businesses in the 2002 and

2004 cross sections.

Page 41 of 78

Page 43: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 5: Probability that a business holds a business license versus revenue

Notes: All values are expressed in 2004 prices. The sample includes all businesses in the 2002 cross

section.

Page 42 of 78

Page 44: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 6: Probability that a business hires outside labor versus revenue

Notes: All values are expressed in 2004 prices. The sample includes all businesses in the 2002 and

2004 cross sections which are pooled together.

0.2

.4.6

.81

Pro

pe

nsity to

hire

outs

ide la

bo

r

10^2 10^3 10^4 10^5 10^6 10^7Real revenue (thousands dong)

Page 43 of 78

Page 45: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 7: Probability that a business obtains a license between 2002 and 2004 versus revenue in

2002

Notes: All values are expressed in 2004 prices. The sample includes all panel businesses that did not

have a license in 2002.

0.1

.2.3

.4

Pro

pe

nsity to

obta

in a

lic

en

se b

etw

ee

n 2

002

and

200

4

10^2 10^3 10^4 10^5 10^6Real revenue in 2002 (thousands dong)

Page 44 of 78

Page 46: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 8: Probability that a business hires outside labor in 2004 versus revenue in 2002 for

businesses that did not hire labor in 2002

Notes: All values are expressed in 2004 prices. The sample includes all panel businesses that did not

hire outside labor in 2002.

0

.05

.1.1

5.2

Pro

pe

nsity to

hire

outs

ide la

bo

r in

20

04

10^2 10^3 10^4 10^5 10^6Real revenue in 2002 (thousands dong)

Page 45 of 78

Page 47: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 9: Probability that a business exits between 2002 and 2004 versus revenue in 2002

Notes: All values are expressed in 2004 prices. The sample includes all businesses operated by panel

households in 2002.

0.2

.4.6

.8

Pro

pe

nsity to

exit b

etw

een

200

2 a

nd 2

004

10^2 10^3 10^4 10^5 10^6Real revenue (thousands dong)

Page 46 of 78

Page 48: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Figure 10: Mean revenue in 2004 by age of business

Notes: All values are expressed in 2004 prices. The sample includes all 2004 businesses operated by

households that were asked the extended business module of the 2004 VHLSS, which is

approximately 1/5th of all households in the cross section.

Page 47 of 78

Page 49: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 1: Summary of U.S. tariffs applied to imports from Vietnam

IndustryNumber of industries

Mean pre-BTA tariff

(Column 2)

Mean post-BTA tariff

(MFN)

Mean change in

tariff

Standard deviation of tariff change

Traded industries 34 0.234 0.025 -0.209 0.179All industries 60 0.133 0.014 -0.119 0.170Manufacturing 22 0.338 0.036 -0.302 0.153Notes: The tariffs reported are simple averages across the indicated set of industries. Non-traded industries, which are included in "All industries" have been assigned a tariff of 0 both before and after the BTA.

Page 48 of 78

Page 50: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 2: Number of households operating businesses2002 2004

Number of households 74,350 45,928Number of households operating businesses 27,815 17,293

1 business 21,740 13,6842 businesses 5,308 3,1203 businesses 665 4224 businesses 102 67

Number of businesses 34,759 21,458Notes: These results are based on the repeated cross sections of the 2002 and 2004 Vietnam Household Living Standards Surveys

Page 49 of 78

Page 51: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 3: Microenterprise summary statistics

Mean St. Dev. Mean St. Dev.Indicator for primary industry 0.015 0.121 0.009 0.092Indicator for secondary industry 0.286 0.452 0.285 0.451Indicator for tertiary industry 0.699 0.459 0.707 0.455Indicator for urban 0.334 0.471 0.334 0.472Household size 4.751 1.761 4.682 1.717Manager characteristics

Female 0.571 0.495 0.583 0.493Head of household 0.449 0.497 0.427 0.495Age 15-24 0.083 0.276 0.072 0.259Age 25-34 0.279 0.449 0.236 0.424Age 35-44 0.345 0.475 0.354 0.478Age 45-54 0.184 0.387 0.221 0.415Age 55-64 0.069 0.253 0.073 0.261Age 65 and older 0.038 0.192 0.040 0.195Ethnic minority 0.070 0.256 0.078 0.268

Microenterprise characteristicsIndicator for business license 0.195 0.396 0.216 0.412Indicator for hiring outside labor 0.108 0.310 0.091 0.287Revenue 18855 122903 30231 283587Expenses 9755 81175 20008 277671Share of expenses on labor 0.032 0.118 0.028 0.113Number of workers 1.67 2.74Number of paid workers 0.32 2.19Age of business (years) 7.69 6.90

2002 2004

Note: Authors's calculations based on the repeacted cross sections of the 2002 and 2004 VHLSSs. Information on number of workers, paid workers, and age of business is not available in the 2002 VHLSS.

Page 50 of 78

Page 52: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 4: Summary statistics of key microenterprise characteristics based on size in 2002

Variable Mean # obs Mean # obs Mean # obs Mean # obsRevenue 18913 34743 2651 11643 8178 11624 46284 11476Indicator for having a license 0.19 34740 0.06 11642 0.15 11624 0.37 11474Indicator for hiring outside labor 0.11 33544 0.03 10991 0.06 11252 0.23 11301

Labor expenses 1102 33544 14 10991 117 11252 3142 11301Labor expenses conditional on being positive 10231 3614 510 307 1887 698 13608 2609

Large

Notes: Revenue and labor expenses are reported in 000s of dong in 2004 prices. Businesses are defined as small, medium, and large based on whether their revenue lies in the bottom, middle, or the upper third of the revenue distribution within their industry in 2002.

All Small Medium

Page 51 of 78

Page 53: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 5: Exit, entry and survival ratesYear Entry Surviving Exiting Total

2002 n.a. 6571 4034 106052004 3944 6571 n.a. 10515

2002 n.a. 0.62 0.38 1.002004 0.38 0.62 n.a. 1.00

Number of microenterprises

Share of microenterprises

Notes: The table is based on all businesses operated by panel households in the 2002 and 2004 VHLSSs.

Page 52 of 78

Page 54: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 6: Exit regression resultsDependent variable: Indicator for exiting between 2002 and 2004

(1) (2) (3)

Traded industriesManufacturing

industries All industries

-0.030 -0.173 -0.130**(0.096) (0.101) (0.055)

Observations 3,563 2,909 10,589R-squared 0.065 0.083 0.043

-0.032 -0.094 -0.220***(0.082) (0.099) (0.066)

Observations 1,807 1,568 6,107R-squared 0.095 0.106 0.057

-0.017 -0.268** 0.031(0.140) (0.104) (0.079)

Observations 1,756 1,341 4,482R-squared 0.067 0.092 0.043

All managersChange in industry tariff

Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The change in industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include province fixed effects and controls for business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, and an indicator for whether the manager is an ethnic minority).

Female managersChange in industry tariff

Male managersChange in industry tariff

Page 53 of 78

Page 55: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 7a: Exit heterogeneity regression results, female and male managersDependent variable: Indicator for exiting between 2002 and 2004

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Change in industry tariff 0.073 0.041 -0.165(0.115) (0.141) (0.107)0.052 -0.000 0.019

(0.098) (0.094) (0.077)-0.091 -0.232** 0.077(0.122) (0.096) (0.114)

Observations 3,563 2,909 10,589R-squared 0.114 0.135 0.080

Change in industry tariff 0.059 -0.011 -0.231**(0.095) (0.123) (0.092)-0.048 -0.030 0.082(0.082) (0.079) (0.076)-0.007 -0.091 0.229*(0.096) (0.086) (0.118)

Observations 3,492 2,870 10,200R-squared 0.112 0.138 0.068

Change in industry tariff 0.075 -0.035 -0.214**(0.109) (0.106) (0.097)0.007 0.158 0.011

(0.189) (0.123) (0.172)0.028 -0.026 0.243**

(0.087) (0.071) (0.107)

Observations 3,492 2,870 10,200R-squared 0.109 0.133 0.072

Endogenous regressorIndustry tariff * medium 1187 3106 1653Industry tariff * large 2725 4669 22212004*medium 4576 6304 19522004*large 389.4 5778 826.1

Change in industry tariff * large

Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The change in industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include province fixed effects and controls for business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, and an indicator for whether the manager is an ethnic minority).

F-statistic of excluded instruments

Size based on initial revenue

Change in industry tariff * medium

Change in industry tariff * large

Size based on initial expenses

Change in industry tariff * medium

IV; size based on initial revenue

Change in industry tariff * medium

Change in industry tariff * large

Page 54 of 78

Page 56: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 7b: Exit heterogeneity regression results, female managersDependent variable: Indicator for exiting between 2002 and 2004

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Change in industry tariff 0.176* 0.211* -0.211(0.095) (0.109) (0.135)-0.065 -0.144* -0.005(0.103) (0.071) (0.084)

-0.275** -0.392*** 0.001(0.122) (0.084) (0.095)

Observations 1,807 1,568 6,107R-squared 0.141 0.152 0.089

Change in industry tariff 0.163*** 0.141* -0.268***(0.056) (0.073) (0.099)

-0.170*** -0.153** 0.012(0.056) (0.055) (0.057)

-0.168** -0.210*** 0.156*(0.070) (0.055) (0.078)

Observations 1,760 1,541 5,866R-squared 0.140 0.154 0.076

Change in industry tariff 0.195*** 0.160*** -0.231**(0.062) (0.053) (0.109)-0.166 -0.076 -0.091(0.132) (0.111) (0.194)-0.111 -0.144** 0.192**(0.081) (0.065) (0.094)

Observations 1,760 1,541 5,866R-squared 0.141 0.156 0.083

Endogenous regressorIndustry tariff * medium 190 743.7 178.9Industry tariff * large 3949 4036 58872004*medium 6008 5272 43692004*large 194.1 3593 166.5

F-statistic of excluded instruments

Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The change in industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include province fixed effects and controls for business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, and an indicator for whether the manager is an ethnic minority).

Change in industry tariff * medium

Change in industry tariff * large

IV; size based on initial revenue

Change in industry tariff * medium

Change in industry tariff * large

Size based on initial revenue

Change in industry tariff * medium

Change in industry tariff * large

Size based on initial expenses

Page 55 of 78

Page 57: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 7c: Exit heterogeneity regression results, male managersDependent variable: Indicator for exiting between 2002 and 2004

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Change in industry tariff -0.115 -0.253 -0.084(0.186) (0.160) (0.121)0.367** 0.360** 0.101(0.151) (0.156) (0.119)0.157 0.013 0.167

(0.142) (0.117) (0.126)

Observations 1,756 1,341 4,482R-squared 0.121 0.158 0.087

Change in industry tariff -0.162 -0.393** -0.152(0.193) (0.139) (0.130)0.260 0.415** 0.177

(0.203) (0.162) (0.119)0.299* 0.232* 0.310**(0.164) (0.128) (0.147)

Observations 1,732 1,329 4,334R-squared 0.120 0.163 0.080

Change in industry tariff -0.156 -0.476*** -0.113(0.268) (0.163) (0.147)0.385 0.787*** 0.047

(0.424) (0.289) (0.231)0.305 0.305* 0.266*

(0.207) (0.174) (0.157)

Observations 1,732 1,329 4,334R-squared 0.110 0.147 0.072

Endogenous regressorIndustry tariff * medium 841.5 1598 787.6Industry tariff * large 1625 1696 29432004*medium 1861 3131 43402004*large 108.3 629.3 292.4

IV; size based on initial revenue

Change in industry tariff * medium

Change in industry tariff * large

F-statistic of excluded instruments

Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The change in industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include province fixed effects and controls for business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, and an indicator for whether the manager is an ethnic minority).

Change in industry tariff * large

Size based on initial revenue

Change in industry tariff * medium

Change in industry tariff * large

Size based on initial expenses

Change in industry tariff * medium

Page 56 of 78

Page 58: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 8: Entry regression resultsDependent variable: Indicator for entering between 2002 and 2004

(1) (2) (3)

Traded industriesManufacturing

industries All industries

0.050 0.011 -0.094(0.061) (0.066) (0.072)

Observations 3,507 2,962 10,453R-squared 0.088 0.102 0.052

0.054 0.048 -0.199*(0.074) (0.086) (0.112)

Observations 1,942 1,712 6,167R-squared 0.113 0.130 0.070

0.058 -0.071 0.098(0.119) (0.113) (0.081)

Observations 1,565 1,250 4,286R-squared 0.074 0.088 0.052Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The change in industry tariff is based on the 2004 industry of the business as is the sample selection by industry. All regressions include province fixed effects and controls for business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, and an indicator for whether the manager is an ethnic minority).

All managersChange in industry tariff

Female managersChange in industry tariff

Male managersChange in industry tariff

Page 57 of 78

Page 59: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 9a: Business growth and transitions, all managers(1) (2) (3)

Traded industriesManufacturing

industries All industries

Dependent variable: ln(revenue)Industry tariff -0.231** -0.264** 0.162

(0.101) (0.106) (0.108)

Observations 3,984 3,354 13,062R-squared 0.871 0.876 0.849Dependent variable: indicator for hiring outside laborIndustry tariff -0.043 -0.040 -0.037

(0.046) (0.052) (0.030)

Observations 4,020 3,406 12,788R-squared 0.805 0.814 0.749Dependent variable: indicator for having a licenseIndustry tariff -0.012 0.018 0.054**

(0.045) (0.041) (0.023)

Observations 4,062 3,430 13,138R-squared 0.778 0.787 0.758Dependent variable: indicator for primary jobIndustry tariff 0.073 0.019 -0.041

(0.051) (0.037) (0.035)

Observations 4,044 3,416 13,112R-squared 0.781 0.782 0.751Dependent variable: indicator for operating 12 monthsIndustry tariff -0.060 -0.026 -0.154***

(0.057) (0.039) (0.047)

Observations 4,062 3,430 13,142R-squared 0.701 0.698 0.683Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Page 58 of 78

Page 60: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 9b: Business growth and transitions, female managers(1) (2) (3)

Traded industriesManufacturing

industries All industries

Dependent variable: ln(revenue)Industry tariff -0.158 -0.128 0.223*

(0.138) (0.129) (0.122)

Observations 1,914 1,670 7,634R-squared 0.857 0.860 0.840Dependent variable: indicator for hiring outside laborIndustry tariff -0.085** -0.085** -0.034

(0.034) (0.034) (0.039)

Observations 1,920 1,686 7,432R-squared 0.749 0.755 0.706Dependent variable: indicator for having a licenseIndustry tariff 0.088*** 0.119*** 0.060***

(0.030) (0.021) (0.010)

Observations 1,950 1,706 7,668R-squared 0.767 0.766 0.753Dependent variable: indicator for primary jobIndustry tariff 0.091 0.041 -0.023

(0.063) (0.079) (0.030)

Observations 1,942 1,700 7,654R-squared 0.792 0.790 0.753Dependent variable: indicator for operating 12 monthsIndustry tariff -0.155* -0.069 -0.242***

(0.086) (0.071) (0.057)

Observations 1,950 1,706 7,670R-squared 0.700 0.693 0.681Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Page 59 of 78

Page 61: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 9c: Business growth and transitions, male managers(1) (2) (3)

Traded industriesManufacturing

industries All industries

Dependent variable: ln(revenue)Industry tariff -0.423* -0.466* 0.013

(0.228) (0.259) (0.151)

Observations 2,070 1,684 5,428R-squared 0.881 0.886 0.856Dependent variable: indicator for hiring outside laborIndustry tariff -0.002 0.015 -0.054

(0.053) (0.072) (0.032)

Observations 2,100 1,720 5,356R-squared 0.833 0.842 0.777Dependent variable: indicator for having a licenseIndustry tariff -0.172* -0.143* 0.027

(0.089) (0.075) (0.057)

Observations 2,112 1,724 5,470R-squared 0.790 0.801 0.767Dependent variable: indicator for primary jobIndustry tariff 0.100 0.076 -0.066

(0.096) (0.109) (0.073)

Observations 2,102 1,716 5,458R-squared 0.783 0.788 0.751Dependent variable: indicator for operating 12 monthsIndustry tariff 0.066 0.066 -0.031

(0.074) (0.062) (0.045)

Observations 2,112 1,724 5,472R-squared 0.727 0.725 0.693Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Page 60 of 78

Page 62: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 10a: Revenue and tariffs by business size, all managersDependent variable: ln(revenue)

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.722*** 0.711*** 0.201(0.183) (0.150) (0.135)

-0.730*** -0.587** -0.095(0.233) (0.219) (0.219)

-1.001*** -0.884*** -0.158(0.319) (0.278) (0.189)

Observations 3,980 3,350 13,056R-squared 0.886 0.889 0.867

Industry tariff 0.718*** 0.842*** 0.117(0.235) (0.234) (0.170)

-0.800*** -0.916*** 0.015(0.199) (0.216) (0.254)

-1.231*** -1.304*** -0.083(0.293) (0.300) (0.338)

Observations 3,940 3,328 12,704R-squared 0.883 0.888 0.860Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Industry tariff * medium

Industry tariff * large

Size based on initial revenue

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 61 of 78

Page 63: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 10b: Revenue and tariffs by business size, female managersDependent variable: ln(revenue)

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.836*** 0.874*** 0.388***(0.203) (0.192) (0.082)

-0.600*** -0.446*** -0.260**(0.115) (0.080) (0.102)

-1.092*** -0.982** -0.201(0.377) (0.344) (0.210)

Observations 1,914 1,670 7,634R-squared 0.878 0.878 0.862

Industry tariff 0.792*** 0.804*** 0.176(0.212) (0.215) (0.127)

-0.683*** -0.658*** 0.046(0.213) (0.190) (0.193)

-1.215*** -1.158*** -0.114(0.154) (0.154) (0.327)

Observations 1,884 1,650 7,396R-squared 0.873 0.875 0.855

Industry tariff * medium

Industry tariff * large

Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Page 62 of 78

Page 64: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table 10c: Revenue and tariffs by business size, male managersDependent variable: ln(revenue)

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.261 0.244 -0.213(0.347) (0.367) (0.226)-0.535 -0.383 0.247(0.541) (0.597) (0.359)-0.625 -0.530 0.086(0.499) (0.502) (0.288)

Observations 2,070 1,684 5,428R-squared 0.893 0.897 0.870

Industry tariff 0.512 0.953** -0.137(0.413) (0.405) (0.300)

-0.934** -1.352** 0.050(0.446) (0.508) (0.440)

-1.164** -1.515** 0.116(0.531) (0.562) (0.418)

Observations 2,060 1,682 5,314R-squared 0.890 0.896 0.864

Industry tariff * medium

Industry tariff * large

Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Page 63 of 78

Page 65: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table A1 - Summary of how the manager was predicted for 2004 businesses

How manager prediction was made

# of businesses

in 2004

share of businesses

in 2004

Share of correct

predictions in 2006

Only household member matched by industry to the business

14701 0.705 0.993

Only household member with matching years, months, and days per month 384 0.018 0.918

Only household member matched by industry to the business 31 0.001 0.912

Only household member with matching months and days per month 1979 0.095 0.802

Only household member matched by industry to the business 44 0.002 0.833

Only household member with matching months 250 0.012 0.749Only household member matched by industry to the business 4 0.000 0.667

Worked most years out of matched household members 1026 0.049 0.751Worked most days out of matched household members 307 0.015 0.663Worked most hours per day out of matched household members 556 0.027 0.685

Only head or spouse working in the business 175 0.008 0.852Head working in the business 1108 0.053 0.691Highest ranked child working in the business 278 0.013 0.710Highest ranked individual working in the business 16 0.001 0.875Highest ranked individual-job working in the business 4 0.000 0.000Total 20863 1.000 0.924

Page 64 of 78

Page 66: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table A2 - Number of households by number of businesses run in 2002 and 2004

0 1 2 3 4 Total0 11801 1960 171 6 1 139391 1934 3932 645 74 7 65922 205 712 629 94 8 16483 12 53 102 32 6 2054 1 12 10 6 2 31

Total 13953 6669 1557 212 24 22415

Number of businesses operated by the household in 2004

Number of businesses operated by the household in 2002

Page 65 of 78

Page 67: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table A3 - Number of businesses by method of matchingNumber of

businesses …Share of

businesses …… that can potentially be matched 7261 1.000… matched by industry and manager 3821 0.526… only matched by industry 1272 0.175… only matched by manager 1038 0.143… unmatched by industry or manager 1130 0.156

Page 66 of 78

Page 68: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B1a: Hiring outside labor regression results, all managersDependent variable: Indicator for hiring labor

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.021 0.011 0.036(0.054) (0.059) (0.040)-0.089* -0.103* -0.072*(0.051) (0.058) (0.036)-0.078* -0.042 -0.130***(0.040) (0.027) (0.041)

Observations 4,020 3,406 12,788R-squared 0.806 0.814 0.750

Industry tariff -0.024 -0.074 -0.020(0.064) (0.053) (0.023)-0.042 -0.001 -0.016(0.111) (0.109) (0.063)0.012 0.105 -0.040

(0.125) (0.121) (0.069)

Observations 4,020 3,406 12,788R-squared 0.806 0.814 0.750Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Industry tariff * medium

Industry tariff * large

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Page 67 of 78

Page 69: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B1b: Hiring outside labor regression results, female managersDependent variable: Indicator for hiring labor

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff -0.026 -0.004 0.002(0.030) (0.043) (0.039)-0.084 -0.143 -0.042(0.073) (0.088) (0.043)

-0.069*** -0.070** -0.063*(0.020) (0.025) (0.036)

Observations 1,920 1,686 7,432R-squared 0.750 0.756 0.706

Industry tariff -0.063** -0.048 -0.035(0.028) (0.038) (0.022)-0.025 -0.065 0.012(0.071) (0.082) (0.055)-0.002 -0.006 -0.018(0.119) (0.120) (0.077)

Observations 1,920 1,686 7,432R-squared 0.750 0.756 0.707Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 68 of 78

Page 70: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B1c: Hiring outside labor regression results, male managersDependent variable: Indicator for hiring labor

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.106 0.046 0.074(0.098) (0.089) (0.073)-0.118 -0.074 -0.110(0.122) (0.099) (0.121)-0.148 -0.030 -0.212***(0.116) (0.116) (0.076)

Observations 2,100 1,720 5,356R-squared 0.833 0.842 0.778

Industry tariff 0.042 -0.127 -0.012(0.174) (0.134) (0.074)-0.087 0.076 -0.044(0.232) (0.235) (0.110)-0.019 0.256 -0.066(0.276) (0.239) (0.148)

Observations 2,100 1,720 5,356R-squared 0.833 0.842 0.778Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 69 of 78

Page 71: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B2a: License regression results, all managersDependent variable: Indicator for having a license

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff -0.041 -0.018 0.075**(0.060) (0.067) (0.029)

0.155*** 0.186*** -0.007(0.048) (0.053) (0.048)-0.016 -0.009 -0.065*(0.041) (0.058) (0.038)

Observations 4,062 3,430 13,138R-squared 0.779 0.788 0.759

Industry tariff 0.011 0.015 0.082**(0.092) (0.092) (0.040)0.092 0.129 -0.030

(0.092) (0.092) (0.049)-0.085 -0.047 -0.066(0.087) (0.093) (0.053)

Observations 4,020 3,406 12,784R-squared 0.781 0.789 0.761Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 70 of 78

Page 72: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B2b: License regression results, female managersDependent variable: Indicator for having a license

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.014 0.007 0.096***(0.053) (0.069) (0.030)

0.222*** 0.262*** -0.023(0.049) (0.042) (0.075)0.042 0.106 -0.083*

(0.082) (0.115) (0.045)

Observations 1,950 1,706 7,668R-squared 0.769 0.768 0.754

Industry tariff 0.117 0.091 0.100***(0.078) (0.100) (0.031)0.043 0.098 -0.051

(0.056) (0.062) (0.042)-0.050 0.031 -0.086(0.146) (0.151) (0.074)

Observations 1,920 1,686 7,430R-squared 0.772 0.769 0.757Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 71 of 78

Page 73: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B2c: License regression results, male managersDependent variable: Indicator for having a license

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff -0.223* -0.169 0.018(0.116) (0.113) (0.085)0.165 0.230 0.047

(0.172) (0.173) (0.097)0.033 -0.033 -0.038

(0.098) (0.093) (0.085)

Observations 2,112 1,724 5,470R-squared 0.791 0.803 0.768

Industry tariff -0.263 -0.274 0.028(0.194) (0.181) (0.118)0.294 0.409* 0.009

(0.208) (0.195) (0.127)0.001 0.016 -0.032

(0.173) (0.158) (0.130)

Observations 2,100 1,720 5,354R-squared 0.793 0.804 0.770Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 72 of 78

Page 74: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B3a: Primary job regression results, all managersDependent variable: Indicator for the business being the manager's primary job

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.359*** 0.288*** 0.026(0.114) (0.091) (0.102)-0.168 -0.113 -0.107(0.130) (0.127) (0.089)

-0.470*** -0.426*** -0.096(0.154) (0.147) (0.139)

Observations 4,044 3,416 13,112R-squared 0.786 0.786 0.754

Industry tariff 0.210** 0.150* 0.065(0.095) (0.084) (0.064)-0.055 -0.023 -0.189**(0.117) (0.115) (0.074)

-0.259** -0.253** -0.142(0.114) (0.109) (0.086)

Observations 4,002 3,392 12,758R-squared 0.784 0.786 0.753Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 73 of 78

Page 75: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B3b: Primary job regression results, female managersDependent variable: Indicator for the business being the manager's primary job

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.311*** 0.223* -0.003(0.103) (0.121) (0.060)

-0.170** -0.126* -0.011(0.078) (0.067) (0.077)-0.307* -0.218 -0.051(0.176) (0.184) (0.126)

Observations 1,942 1,700 7,654R-squared 0.796 0.793 0.756

Industry tariff 0.139* 0.072 0.035(0.077) (0.099) (0.027)-0.035 -0.033 -0.140**(0.082) (0.058) (0.057)-0.071 -0.024 -0.086(0.122) (0.128) (0.088)

Observations 1,912 1,680 7,416R-squared 0.795 0.794 0.757Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 74 of 78

Page 76: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B3c: Primary job regression results, male managersDependent variable: Indicator for the business being the manager's primary job

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.615** 0.611* 0.128(0.296) (0.344) (0.204)-0.327 -0.208 -0.311(0.350) (0.379) (0.196)

-0.769** -0.834** -0.234(0.320) (0.371) (0.210)

Observations 2,102 1,716 5,458R-squared 0.790 0.796 0.757

Industry tariff 0.437 0.509 0.178(0.295) (0.357) (0.180)-0.156 -0.195 -0.354*(0.360) (0.417) (0.190)-0.582* -0.706* -0.283(0.297) (0.344) (0.174)

Observations 2,090 1,712 5,342R-squared 0.785 0.793 0.753Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 75 of 78

Page 77: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B4a: Operating for 12 months regression results, all managersDependent variable: Indicator for operating for 12 months

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.093 0.120 -0.074(0.080) (0.087) (0.053)

-0.330** -0.245** -0.164**(0.144) (0.108) (0.070)0.007 -0.011 -0.080

(0.102) (0.116) (0.067)

Observations 4,062 3,430 13,142R-squared 0.704 0.700 0.686

Industry tariff -0.007 0.038 -0.110**(0.056) (0.078) (0.047)-0.081 -0.068 -0.129**(0.144) (0.144) (0.056)0.002 -0.021 -0.027

(0.059) (0.053) (0.066)

Observations 4,020 3,406 12,788R-squared 0.700 0.699 0.683Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 76 of 78

Page 78: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B4b: Operating for 12 months regression results, female managersDependent variable: Indicator for operating for 12 months

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.060 0.052 -0.153**(0.136) (0.133) (0.057)-0.377* -0.226 -0.165*(0.200) (0.137) (0.086)-0.077 0.029 -0.081(0.167) (0.134) (0.066)

Observations 1,950 1,706 7,670R-squared 0.703 0.695 0.684

Industry tariff -0.029 -0.071 -0.203***(0.144) (0.142) (0.068)-0.072 0.124 -0.078(0.256) (0.131) (0.101)-0.154 -0.028 -0.058(0.194) (0.139) (0.079)

Observations 1,920 1,686 7,432R-squared 0.700 0.695 0.679Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 77 of 78

Page 79: Export markets and microenterprise performance: Evidence ...€¦ · Most workers in low-income countries are self-employed and work for a household-owned business or a farm. We examine

Table B4c: Operating for 12 months regression results, male managersDependent variable: Indicator for operating for 12 months

(1) (2) (3)Traded

industriesManufacturing

industries All industries

Industry tariff 0.049 0.168 0.085(0.214) (0.177) (0.091)-0.153 -0.166 -0.189(0.245) (0.231) (0.156)0.201 0.005 -0.152

(0.309) (0.283) (0.165)

Observations 2,112 1,724 5,472R-squared 0.729 0.727 0.696

Industry tariff 0.027 0.181 0.093(0.208) (0.179) (0.111)-0.042 -0.189 -0.289*(0.285) (0.312) (0.172)0.151 -0.045 -0.082

(0.244) (0.203) (0.152)

Observations 2,100 1,720 5,356R-squared 0.726 0.725 0.695Notes: Robust standard errors in parentheses. Clustered by industry. *** p<0.01, ** p<0.05, * p<0.1. The industry tariff is based on the 2002 industry of the business as is the sample selection by industry. All regressions include business fixed effects and initial business characteristics (an indicator for whether a business is in an urban area, gender, age, and education of the manager, an indicator for whether the manager is an ethnic minority, and province) interacted with a 2004 dummy.

Size based on initial revenue

Industry tariff * medium

Industry tariff * large

Size based on initial expenses

Industry tariff * medium

Industry tariff * large

Page 78 of 78