Can Wage Subsidies Boost Employment in the Wake of an Economic Crisis? Evidence from Mexico

Abstract

This paper measures the employment effect of a programme in Mexico that granted firms wage subsidies during the recent economic crisis. I use monthly administrative data at the industry level, along with Euclidean distance matching to construct groups of eligible and ineligible durable goods manufacturing industries that display statistically identical preprogramme trends in employment. Difference-in-difference results show a positive but not statistically significant effect of the wage subsidies on employment during the programme’s eight-month duration. The size of the effect increases to 18 per cent after the programme ended and the results indicate that employment after the programme recovered faster in eligible industries than in ineligible industries. Additional analysis suggests that the programme did not incentivise firms to retain workers with job-specific skills as originally intended. Instead, the payment of subsidy funds, which only happened towards the end of the programme, seems to have provided liquidity for hiring back workers.

Acknowledgment

I thank David McKenzie and Will Wiseman for valuable comments and the Mexican Social Security Institute (IMSS) for sharing their data on employment in manufacturing industries. Nathaniel Russell, Christian Salas, and Lucía Juárez provided excellent research assistance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary Material

Supplementary Materials are available for this article which can be accessed via the online version of this journal available at https://doi.org/10.1080/00220388.2020.1715941.

Notes

1. Based on the ILO/World Bank Inventory of policy responses to the global financial and economic crisis of 2008 (http://www.ilo.org/empelm/projects/WCMS_158875/lang–en/index.htm). However, note that a wage subsidy programme may not be an optimal policy to relax credit constraints in a crisis. Policies that provide financing to firms may achieve similar goals without distorting other firm decisions, for example how much capital to sell and whether to spend on advertising.

2. A related paper by Behaghel, Crépon, and Sédillot (2008) studies the effects of dismissal costs for older workers in France and finds these to have little effect on layoffs, but a more sizeable effect on the transition rate from unemployment to employment. Kugler and Pica (2008) show that an increase in dismissal costs for small firms in Italy led to a decrease of accessions and separations by about 13 per cent and 15 per cent in small relative to large firms.

3. Part A provides matching grants for up to 70 per cent of the cost for projects related to technology transfer that address market failures, including training, consulting, licencing, and patents, for any firms or organisations that engage in high technology activities. PRODIAT did not disburse any of these matching grants in 2009 and only disbursed two matching grants with a total value of 3.9 million pesos (US$250,000) in 2010 and eight matching grants with a total value of 9.8 million pesos (US$680,000) in 2011. The total amount disbursed for 2010 and 2011 corresponds to only about 1.5 per cent of the funding disbursed under PRODIAT Part B.

4. Throughout the paper, I use an exchange rate of 15.82 pesos per US Dollar.

5. The average daily minimum wage in 2009 was 53 pesos, or US$3.35 (Comisión Nacional de los Salarios Mínimos, 2015).

6. Calculations based on data from IMSS for December 2008. IMSS reports number of employers, which may either be plants or firms.

7. The data suggest that a few very large firms received subsidies.

8. The partial information available on individual subsidies suggests that some applicants were paid in full in 2009, while others were paid over the course of two or three years. There are also some firms that received payments for the first time in 2010 or 2011.

9. Some firms producing nondurable goods, such as rubber and plastic, were also declared eligible for the programme in the 10 April 2009 revision of the operating rules, under the condition that 90 per cent of their production is used as an input in other eligible industries.

10. As posted on the Bureau of Labour Statistics website: https://www.bls.gov/jlt/jltnaics.htm.

11. This approach of matching eligible and ineligible industries differs slightly from the traditional matching method that matches treated and untreated firms or individuals. It is in this sense related to Angrist and Kuersteiner (2011) who use matching to determine a policy decision.

12. I control for US imports from all countries other than Mexico since this variable is likely not affected by the Mexican wage subsidy programme, that is it is likely exogenous to the programme. About three per cent of the observations for US imports are zeros that are dropped when using the log of US imports. Using the fourth root of US imports instead, which mimics a log function but does not drop zeros, gives very similar results.

13. The US car buyback programme was called the Car Allowance Rebate System. It was originally a one billion dollar programme slated to run from 1 July 2009 to 25 August 2009. However, the programme was so successful that the funds were exhausted by June 30th. An additional $2 billion was approved, and those funds lasted until August 24^th. In Mexico, the Programa de Renovacion Vehicular began in August 2009 with a budget of 250 million pesos. It was suspended around March 2010 because of low demand.

14. When controlling for industry specific time trends, the effect is slightly smaller and statistically significant at the 15-per cent level. Although the result here is weaker, the effects of the programme over time, as estimated in Table 4, are statistically significant at the 10-per cent level when controlling for industry specific time trends.

15. See also Lalive, Landais, and Zweimuller (2015) for a discussion of market externalities of labour market programmes.

16. Another potential source of bias is that the programme may have spill over effects across product markets, if firms in ineligible industries supply goods to firms in eligible industries. In this case, the impact estimates here would underestimate the programme’s true effect. I do not have information on industry linkages. However, looking at the product descriptions (Table A2 in Appendix A), many of the goods produced by ineligible industries are unlikely to be inputs in the production process of eligible industries.

17. I also examine whether eligible industries are increasingly hiring individuals who were laid off in ineligible industries (see discussion of Figure A4 in Appendix A). Although based on very few individuals, Figure A4 suggests that less than 10 per cent of individuals who lose their job with access to IMSS in an ineligible industry get hired in an eligible industry in subsequent quarters and that this small percentage does not increase in quarters after the wage subsidy programme.

18. Another possible explanation is that the effects on employment during the programme are not statistically significant due to lack of power. The size of the estimates is not small, ranging from 2.7 to 8.7 per cent.

19. Another reason of laying off employees with low job tenure is that government mandated severance payments increase with number of years on the job.

20. The average firm size ranges from nine to 516 employees in the sample of 42 industries matched with a 0.3 caliper, with a median of 60.

21. This conclusion seems reasonable even though the wage subsidies tended to go to the largest firms in each industry (see Section 2). I do not have data on the firm size distribution within industries, but it seems plausible that a relatively large firm in an industry with 32 employees on average (the 25th percentile of average firm size in my sample) has more than 100 employees. A relatively large firms in an industry with 117 employees on average (the 75th percentile) may have more than 500 employees. Data from the 2010 World Bank Enterprise Survey suggests that, in manufacturing industries eligible for the wage subsidy programme, 23 per cent of firms with more than 100 employees faced credit constraints. The Enterprise Survey is not representative for very large firms, but among the 20 manufacturing firms in eligible sectors with more than 500 employees in the survey, only 10 per cent are credit constrained. It thus makes sense that relatively large firms in industries with large firm sizes may not benefit as much from the subsidies as large firms in industries with small firm sizes.