The labor market power of exporting firms: Evidence from Latin America

Amodio, Francesco; Brancati, Emanuele; de Roux, Nicolás; Di Maio, Michele Working Paper The labor market power of exporting firms: Evidence from Latin America IDB Working Paper Series, No. IDB-WP-1579 Provided in Cooperation with: Inter-American Development Bank (IDB), Washington, DC Suggested Citation: Amodio, Francesco; Brancati, Emanuele; de Roux, Nicolás; Di Maio, Michele (2024) : The labor market power of exporting firms: Evidence from Latin America, IDB Working Paper Series, No. IDB-WP-1579, Inter-American Development Bank (IDB), Washington, DC, https://doi.org/10.18235/0012855 This Version is available at: https://hdl.handle.net/10419/299412 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/3.0/igo/ The Labor Market Power of Exporting Firms: Evidence from Latin America Francesco Amodio Emanuele Brancati Nicolás de Roux Michele Di Maio WORKING PAPER No IDB-WP-1579 InterA merican Development Bank Department of Research and Chief Economist March 2024 * McGill University ** Sapienza University *** Universidad de Los Andes The Labor Market Power of Exporting Firms: Evidence from Latin America Francesco Amodio* Emanuele Brancati** Nicolás de Roux*** Michele Di Maio** InterA merican Development Bank Department of Research and Chief Economist March 2024 Cataloging-in-Publication data provided by the Inter-American Development Bank Felipe Herrera Library The labor market power of exporting firms: evidence from Latin America / Francesco Amodio, Emanuele Brancati, Nicolás de Roux, Michele Di Maio. p. cm. — (Working Paper Series ; 1579) Includes bibliographical references. 1. Labor market-Latin America. 2. Labor supply-Latin America. 3. Export trading companies-Latin America. I. Amodio, Francesco. II. Brancati, E. (Emanuele). III. Roux Uribe, Nicolás de. IV. Di Maio, Michele. V. Inter-American Development Bank. Department of Research and Chief Economist. VI. Series. IDB-WP-1579 http://www.iadb.org Copyright © 2024 Inter-American Development Bank ("IDB"). This work is subject to a Creative Commons license CC BY 3.0 IGO (https://creativecommons.org/licenses/by/3.0/igo/legalcode). The terms and conditions indicated in the URL link must be met and the respective recognition must be granted to the IDB. Further to section 8 of the above license, any mediation relating to disputes arising under such license shall be conducted in accordance with the WIPO Mediation Rules. Any dispute related to the use of the works of the IDB that cannot be settled amicably shall be submitted to arbitration pursuant to the United Nations Commission on International Trade Law (UNCITRAL) rules. The use of the IDB's name for any purpose other than for attribution, and the use of IDB's logo shall be subject to a separate written license agreement between the IDB and the user and is not authorized as part of this license. Note that the URL link includes terms and conditions that are an integral part of this license. The opinions expressed in this work are those of the authors and do not necessarily reflect the views of the Inter-American Development Bank, its Board of Directors, or the countries they represent. Abstract Using establishment-level data from the World Bank Enterprise Survey, we assess the market power of exporting firms across 16 countries in Latin America. Leveraging information on export destinations, as well as exchange rate and price data, we construct exchange rate-driven shocks to the marginal revenue product of individual firms. By examining firms’ employment and wage responses, we estimate the inverse elasticity of the labor supply they face—a direct indicator of labor market power. In our preferred specification, we estimate that workers employed in exporting firms produce on the margin 83% more than what they earn as wage. We investigate the correlations between labor market power and firm characteristics, country attributes, and labor market institutions and regulations. We find that labor market power is higher for firms in countries where unions, collective bargaining, and unemployment protection are less prevalent.1 JEL classifications: F10, F14, F16, J2, J3, J42, L10, O54 Keywords: Firms, Exports, Labor market power, Labor market institutions, Latin America 1Copyright © [2024]. Inter-American Development Bank. Used by permission. The work was financed with the support of the Latin America and the Caribbean Research Network of the Inter-American Development Bank. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the Inter-American Development Bank, its Board of Directors, or the countries they represent. We thank Ricky Andr´es Carrillo for excellent research assistance. 1 Introduction When employers face limited competition for workers, they are able to suppress wages below the marginal product of labor. This has significant implications for the labor share of income, income distribution, and overall welfare. A growing body of literature has estimated significant levels of market power in both high-income and low and middle-income countries (see e.g., Azar et al., 2022b; Berger et al., 2022; Yeh et al., 2022; Bassier et al., 2022; Amodio and de Roux, 2023; Muralidharan et al., 2023; Bassier, 2023). While the determinants of labor market power are still under-explored, recent studies are beginning to bridge this gap. For instance, Felix (2021) investigates the effects of tariff reductions on wage markdowns, Amodio et al. (2022) delve into the connection between self-employment and labor market power in Peru, and Estefan et al. (2024) examine the impact of labor policies, particularly labor outsourcing in Mexico, on wage markdowns. Labor market institutions are a natural candidate for explaining variation in labor market power. These institutions have been shown to significantly affect labor market dynamics (Nickell, 1997; Heckman and Pag´es, 2003). For instance, stringent labor regulations, as discussed by Botero et al. (2004), can influence employment levels, but they may also empower workers, potentially leading to a reduction in wage markdowns. To effectively study the relationship between labor market power and institutional features, especially those that vary at the country level, it is essential to employ uniform measures of labor market power that are consistent across different countries. This paper examines the variation in labor market power in relation to firm and country-specific characteristics, particularly focusing on labor market institutions. We use establishment-level data from the World Bank Enterprise Survey (WBES) to estimate the market power of large employers in Latin American countries. By leveraging data on export destinations, exchange rates, and prices, we construct exchange rate-driven shocks to the marginal revenue product of firms in leading export sectors, enabling us to assess their impact on employment and wages. This approach yields estimates of the inverse elasticity of the labor supply curve faced by individual firms, a direct measure of their labor market power. Importantly, our estimates are consistent across Latin American firms, which allows us to relate them to country-level characteristics. Our focus on Latin America is strategic, as these countries exhibit significant uniformity in various aspects, yet display considerable diversity in their labor institution environments (Heckman and Pag´es, 2003). Additionally, unlike continents such as Africa, Latin America has a substantial presence of large firms operating in a wide array of export sectors. Using export shocks to generate exogenous variation in the marginal revenue product of labor to trace firms’ labor supply curve allows us to estimate labor market power with fewer assumptions compared 1 to methods that rely on production function estimation (Amodio and de Roux, 2023). However, since we must consider exporting establishment to construct the instrument, this measurement strategy is limited to assessing the labor market power of large firms. We construct export shocks corresponding to a devaluation of the local currency. In our panel regression analysis, which includes year-specific fixed effects for both local labor markets and country-sector pairs, we initially show that these shocks lead firms to increase sales, employment, and wages.2This result alone constitutes evidence on the existence of labor market power, as in a perfectly competitive labor market, shocks to firms’ marginal revenue product of labor would not affect wages. Employing an instrumental variable (IV) strategy, we regress the logarithmic change in wages against the logarithmic change in employment, using our export shock as the instrument to estimate wage markdowns. We estimate an inverse labor supply elasticity of 0.83, indicating that workers in large firms produce, at the margin, 83% more than their received wages.3This finding remains robust across different definitions of exporting sectors. Additionally, we observe no significant impact of export shocks on firm upgrading indicators, such as the proportion of skilled workers or the introduction of new processes. This mitigates concerns that our results are influenced by quality upgrading in firms exposed to shocks, leading to higher wages. We then shift our focus to examine how labor market power varies across different dimensions. We find that that firms located in smaller cities, outside the country’s capital, which are foreign-owned (as opposed to state-owned) and larger according to their sales levels, exhibit higher levels of labor market power. Additionally, we observe that firms in countries with lower overall income and a lower labor share of income tend to have more labor market power. More importantly, regarding labor market institutions, we find that in countries where unions, collective bargaining, and unemployment protection are less prevalent, firms possess greater labor market power. These findings carry substantial implications, suggesting that while pro-worker labor market institutions may adversely impact employment and informality, they may also mitigate the labor market power of large firms. This, in turn, can support workers, enhance welfare, and contribute to reducing inequality. This brief paper contributes to the burgeoning literature that quantifies labor market power. In the context of developed countries, studies such as Azar et al. (2022b), Bassier et al. (2022), and Yeh et al. (2022) report wage markdowns in the US ranging from 1.14 to 1.53. In the sphere of developing countries, Tortarolo and Z´arate (2020) and Amodio and de Roux (2023) estimate wage markdowns of 1.12 and 1.4, respectively, in Colombian manufacturing plants. Muralidharan et al. (2023) find a markdown of 1.33 among landowners in India, while Felix 2We have access to a confidential version of the WBES data with the geolocation of the firm that allows us to identify its local labor market. 3Alternatively, workers earn about 55 cents for every marginal dollar they produce. 2 (2021) reports a markdown of 2 in the same country. Our research contributes to this literature by offering new estimates that are consistent across a range of Latin American countries. These estimates do not require the use of production function estimation and reveal a wage markdown for large Latin American firms of 1.83. We further contribute to the growing body of research that investigates the determinants of labor market power. This includes Azar et al. (2022a), who establish a link between labor market power and market concentrationAmodio et al. (2022), who explore the relationship between self-employment and labor market power, Felix (2021) who examines the connection between labor market power and trade reforms in Brazil, and Estefan et al. (2024) who analyze the impact of outsourcing policies on wage markdowns. Our study adds to this line of work by showing that pro-worker labor market institutions, including labor unions and unemployment protection, are associated with lower wage markdowns. This finding is facilitated by our uniform estimates across a diverse set of countries, which we derive using a plausible source of exogenous variation. The rest of the paper is organized as follows. The next section describes the data. Section 3 presents the empirical strategy and summary statistics. Sections 4 discusses estimates of labor market power across large firms in Latin America. Section 5 relates these estimates to firm and country-level characteristics. Finally, Section 6 concludes. 2 Data Our main data source is the World Bank Enterprise Survey (WBES). The WBES is administered to owners and top managers of formal sector firms in the manufacturing, retail, and service sectors in many developing countries. The survey provides information on, among other features, annual sales, cost of inputs and labor, employment and workforce composition, exports and imports, capacity utilization, innovation, and performance. Information is available for about 180,000 firms in 148 countries between 2006 and 2020. The survey is representative at the country level of the population of privately-owned firms with at least 5 employees operating in the formal (non-agricultural) sector. The administration of the questionnaire is conducted face-to-face in different years and at different time intervals. In terms of coverage, most of the countries in the sample are lowand middle-income economies. Firms are selected using random sampling techniques with three stratification levels to ensure representativeness across firm size (5-19 employees; 20-99 employees; and 100+ employees), sector (manufacturing, retail, and other services, with further sub-sectors in selected economies), and subnational region. Importantly, the WBES follows a standardized methodology that allows for comparable information on an extensive set of firms’ activities. The original dataset is a repeated cross-section, but several firms are interviewed in multiple waves. In this paper, we 3 focus on the panel dimension of the WBES. We have access to a confidential version of the WBES dataset with information on firms’ geo-localization. Starting from 2010, geo-reference information on each firm is tracked by the devices used in the Computer-Assisted Personal Interviews (CAPI). The geo-localization of the firm allows us to find its relevant local labor market. For firms observed before 2010, we impute the geo-localization from 2010 or onwards and therefore assume that the firm location is stable over time. We construct a panel of establishments using data from 16 countries in Latin America and combine it with data on exchange rates and inflation. We obtain nominal exchange and inflation rates from the IMF, the Bank of Italy, the World Bank and the OECD. Finally, we use information from WITS that contains data on bilateral export flows by sector (2-digit ISIC Rev 3.1). 3 Empirical Strategy The wage-setting power of employers is measured by the elasticity of the labor supply they face (Manning, 2003). Our strategy builds on the one that Amodio and de Roux (2023) implement for Colombia. We leverage pre-determined variation on the export destinations of a firm’s sector combined with variation in real exchange rates to generate firm-specific shocks to the marginal revenue product and thus variation in labor demand. These exchange rate-driven export shocks act as a labor demand shifters and can be used to trace the slope of the labor supply curve faced by individual firms. Following a positive export shock, if the labor market is perfectly competitive, the equilibrium number of hired workers will increase, but the wage paid will not. This is because the firm takes the price of labor as given and equal to the ongoing market wage. If the firm has labor market power, both the equilibrium number of hired workers and the wage paid will increase. We can thus identify the inverse elasticity of the labor supply curve by taking the ratio between the log change in wage and the log change in employment. To operationalize this theory, we start by deriving the share of exports of sector sin country cto destination din each year t, given by: Sscdt =Expscdt PdExpscdt where Expscdt is the total export value from sector sin country cto destination dfrom the WITS data. We then obtain the real exchange rate between country cand destination din year tas follows: e Rcdt =Rn cdt CP Idt CP Ict  Here, Rn cdt is the nominal exchange rate in units of country c’s currency for one unit of the 4 References Amodio, F. and de Roux, N. (2023). Measuring Labor Market Power in Developing Countries: Evidence from Colombian Plants. forthcoming Journal of Labor Economics. Amodio, F., Medina, P., and Morlacco, M. (2022). Labor Market Power, Self-Employment, and Development. IZA Discussion Papers 15477, Institute of Labor Economics (IZA). Azar, J., Marinescu, I., and Steinbaum, M. (2022a). Labor market concentration. Journal of Human Resources, 57(S):S167–S199. Azar, J. A., Berry, S. T., and Marinescu, I. (2022b). Estimating labor market power. Technical report, National Bureau of Economic Research. Bassier, I. (2023). Firms and inequality when unemployment is high. Journal of Development Economics, 161:103029. Bassier, I., Dube, A., and Naidu, S. (2022). Monopsony in movers: The elasticity of labor supply to firm wage policies. Journal of Human Resources, 57(S):S50–s86. Berger, D., Herkenhoff, K., and Mongey, S. (2022). Labor market power. American Economic Review, 112(4):1147–93. Botero, J. C., Djankov, S., Porta, R. L., Lopez-de Silanes, F., and Shleifer, A. (2004). The regulation of labor. The Quarterly Journal of Economics, 119(4):1339–1382. Donaldson, D. and Hornbeck, R. (2016). Railroads and american economic growth: A “market access” approach. The Quarterly Journal of Economics, 131(2):799–858. Estefan, A., Gerhard, R., Kaboski, J. P., Kondo, I. O., and Qian, W. (2024). Outsourcing policy and worker outcomes: Causal evidence from a mexican ban. Technical report, National Bureau of Economic Research. Felix, M. (2021). Trade, labor market concentration, and wages. Mimeo. Frias, J. A., Kaplan, D. S., and Verhoogen, E. (2012). Exports and Within-Plant Wage Distributions: Evidence from Mexico. American Economic Review Papers and Proceedings, 102(3):435–440. Heckman, J. J. and Pag´es, C. (2003). Law and employment: Lessons from latin america and the caribbean. Manning, A. (2003). Monopsony in Motion: Imperfect Competition in Labor Markets. Princeton University Press. Muralidharan, K., Niehaus, P., and Sukhtankar, S. (2023). General equilibrium effects of (improving) public employment programs: Experimental evidence from india. Econometrica, 91(4):1261–1295. Nickell, S. (1997). Unemployment and labor market rigidities: Europe versus north america. Journal of Economic perspectives, 11(3):55–74. 11 Tortarolo, D. and Z´arate, R. (2020). Imperfect competition in product and labor markets. a quantitative analysis. Mimeo. Verhoogen, E. A. (2008). Trade, Quality Upgrading, and Wage Inequality in the Mexican Manufacturing Sector. The Quarterly Journal of Economics, 123(2):489–530. Yeh, C., Macaluso, C., and Hershbein, B. (2022). Monopsony in the us labor market. American Economic Review, 112(7):2099–2138. 12 Tables and Figures Table 1: Descriptive Statistics Observations Mean St. Dev. Median Panel A: Full Sample Year 6905 2009.1 3.8 2009 Employment 6905 127.9 535.3 30 Wage (annual) 6905 6344 6094 4605 Sales (millions) 6611 4.7 11.9 .83 Sales per worker 6611 51174 111740 26283 Expoter 6902 .302 .459 0 Importer 4563 .751 .432 1 Export shock 2083 .114 .318 0 Import Shock 2073 .464 .499 0 ∆ log Employment 3779 .088 .695 .069 ∆ log Wage 3726 .177 1.076 .199 ∆ log Sales 3489 .127 1.077 .158 ∆ log Sales per Worker 3487 .046 .977 .064 Panel B: Export Shock = 0 Year 1845 2010.23 2.6 2009 Employment 1845 118.3 430.0 28 Wage (Annual) 1845 6536 6155 4668 ∆ log Employment 1845 .063 .657 .041 ∆ log Wage 1811 .231 1.074 .241 ∆ log Sales 1719 .137 1.04 .173 ∆ log Sales per Worker 1719 .079 .958 .095 Panel C: Export Shock = 1 Year 238 2011.6 3.2 2009 Employment 238 231.3 428.3 89.5 Wage (Annual) 238 9206 7854 6996 ∆ log Employment 238 .047 .655 .054 ∆ log Wage 236 .224 1.046 .256 ∆ log Sales 223 .23 1.155 .161 ∆ log Sales per Worker 223 .164 1.079 .091 Notes: The unit of observation is an establishment-year. Wages and sales are in are in constant 2000 USD. 13 Table 2: Effect of Export Shocks on Sales and Employment ∆ Log Sales ∆ Log Employment ∆ Log S/E (1) (2) (3) (4) (5) (6) (7) Export Shock 0.311 0.466** 0.490** 0.191*** 0.224*** 0.259*** 0.133 (0.191) (0.195) (0.210) (0.066) (0.067) (0.076) (0.171) Import Shock 0.078 0.179 0.177 -0.025 -0.009 -0.009 0.137 (0.202) (0.163) (0.165) (0.062) (0.063) (0.063) (0.155) Log Employmentt−10.419*** 0.420*** -0.082*** -0.076*** 0.106*** (0.102) (0.103) (0.016) (0.016) (0.024) Log Waget−10.043 0.044 0.106*** 0.109*** -0.269*** (0.046) (0.047) (0.016) (0.016) (0.044) Log Salest−1-0.369*** -0.367*** (0.084) (0.084) Share of exportst−1-0.001 -0.001* -0.001 (0.002) (0.001) (0.001) Sector ×Ctr. ×Year FEs Yes Yes Yes Yes Yes Yes Yes LLM ×Year FEs Yes Yes Yes Yes Yes Yes Yes Observations 1035 1035 1035 1101 1101 1101 1035 R20.258 0.345 0.345 0.191 0.226 0.228 0.324 Notes: Sample is restricted to firms in the top 12 export sectors in each country. Standard errors clustered at the sector ×country and local labor market level. * p-value <0.1; ** p-value <0.05; *** p-value <0.01. 14 Table 3: Estimates of Labor Market Power, ϵ Top 15 Sectors Top 12 Sectors Top 10 Sectors ∆ log N∆ log w∆ log N∆ log w∆ log N∆ log w (1) (2) (3) (4) (5) (6) Export Shock 0.183** 0.249** 0.259** 0.214** 0.275** 0.251** (0.072) (0.093) (0.076) (0.107) (0.074) (0.115) Implied ϵ1.365** 0.828** 0.913** (0.582) (0.397) (0.421) F-statistic 6.433 11.72 13.806 Controls Yes Yes Yes Yes Yes Yes Sector ×Ctr. ×Year FEs Yes Yes Yes Yes Yes Yes LLM ×Year FEs Yes Yes Yes Yes Yes Yes Observations 1311 1311 1101 1101 998 998 R20.241 0.559 0.228 0.549 0.232 0.559 Notes: The first row reports the estimated coefficient of a regression of ∆ log Nor ∆ log won the export shock for firms in the top 15, top 12 and top 10 export sectors. The second row reports the estimated coefficient of an IV regression of ∆ log Non ∆ log wusing the export shock as instrument. The set of controls includes Log Employmentt−1, Import Shock, Log Waget−1, Log Salest−1, and Share of exportst−1. Standard errors clustered at the sector ×country and local labor market level. * p-value <0.1; ** p-value <0.05; *** p-value <0.01. 15 Table 4: Heterogeneity Panel A: Firm Characteristics Country Foreign State Large Market Capital Owned Owned Firm Access (1) (2) (3) (4) (5) Yes 0.048 1.944*** -1.222 1.043*** 0.802 (0.689) (0.608) (1.010) (0.274) (0.644) No 1.135** 0.670 0.802** 0.513 0.906 (0.505) (0.458) (0.375) (0.735) (0.800) Observations 1101 1098 1098 1101 1101 R2-0.071 0.089 0.118 0.098 0.097 Panel B: Country Characteristics and Institutions High High High Collect. Unemp. Income Lab. Share Union Bargain Protec. Yes 0.968 -0.204 0.150 -0.413 -0.185 (0.628) (1.018) (0.710) (0.845) (0.778) No 0.648** 1.186** 1.267** 1.427*** 1.296** (0.305) (0.518) (0.603) (0.488) (0.593) Observations 1184 1184 1144 974 1101 R20.049 0.023 -0.057 -0.233 0.006 Notes: Sample is restricted to firms in the top 12 export sectors in each country. Each column of each panel shows the result of an IV regression of ∆ log Non ∆ log wusing the export shock as instrument and where all the right hand variables are interacted with an indicator variable listed in the column header. Each column reports the estimated coefficients of the interaction with ∆ log w. In panel A, the indicator variable in column 1 is equal to one if the firm is located in the country capital, in column 2 if its ownership is less than 10% foreign, in column 3 if its ownership is more than 10% public, in column 4 if its total sales are above the median of total sales across firms, in column 5 if the firm’s market access is above the median. In panel B, the indicator variable in column 1 is equal to one if the firm belongs to upper-middle and high-income countries as opposed to low-middle income countries according to the World Bank classification, in column 2 if the firm is located in a country with a labor share above the median labor share as reported by the ILO, in column 3 if the firm is located in a country with share of workers who belong to a union above the median as reported by the ILO, in column 4 if the firm is located in a country with a share of employees covered by collective agreements above the median as reported by the ILO, in column 5 if the firm is located in a country with unemployment protection available after one year of tenure on the job. Standard errors clustered at the sector ×country and local labor market level. * p-value <0.1; ** p-value <0.05; *** p-value <0.01. 16 Figure 1: Distribution of Employment and Wage Changes 0 .2 .4 .6 .8 1 -2 -1 0 1 2 Change in log employment, residual Export Shock = 0 Export Shock = 1 0 .2 .4 .6 .8 -2 -1 0 1 2 Change in log wages, residual Export Shock = 0 Export Shock = 1 Notes: This figure shows the distribution of employment changes (left) and wages (right) across waves for firms with and without an export shock. We plot the residuals of a regression of the change of employment on the set of baseline controls, a set of sector-country-year fixed effects, and a set of local labor market year fixed effects. 17 APPENDIX: Additional Figures and Tables (for online publication) Figure A1: Variation in Export Shocks 0 2 4 6 -.5 0 .5 1 Effective Exchange Rate Variation Across Countries and Sectors Within Countries and Sectors Notes: This figure shows the distribution of the exchange rate shocks for two level of variation. The blue line shows the variation across countries and sector. The red line shows the variation within countries and sectors. Figure A2: Labor Demand Shift -.3 -.2 -.1 0 .1 .2 -1.5 -1 -.5 0 .5 1 Residuals Export Shock = 0 Export Shock = 1 Notes: This figure is a binned plot of changes in wages (y-axis) on changes in employment (x-axis) after controls and fixed effects have been partialled out, for firms that experienced an export shocks and firms that did not. Online Appendix p.1 Figure A3: Robustness to Dropping Individual Countries 0 .5 1 1.5 2 Argentina Bolivia Chile Colombia DominicanRepublic Ecuador ElSalvador Guatemala Honduras Mexico Nicaragua Panama Paraguay Peru Suriname Uruguay Notes: This figure shows the estimate of εobtained in a sample that excludes the country listed in the x-axis. The vertical lines represent 95% confidence intervals. Online Appendix p.2 Table A1: Potential Sample Country Establishments LLMs Argentina 631 7 Bolivia 265 3 Chile 380 5 Colombia 499 4 Dominican Republic 98 4 Ecuador 214 3 El Salvador 273 6 Guatemala 331 5 Honduras 161 6 M´exico 199 8 Nicaragua 215 6 Panama 98 3 Paraguay 208 2 Peru 537 4 Suriname 55 2 Uruguay 293 2 Total 4,457 70 Online Appendix p.3