Empirics prove decreased payroll taxes increase wages and employment

This study attempted to estimate the effect of a change in labour taxation in Brazil on employment and wages. Starting in January 2012, firms not under the special tax regime for small and micro enterprises (Simples) that operated in four economic sectors experienced a cut in the employers’ contribution to social security of 20 percent of the wage bill. Relying on administrative data that allowed for a firm level analysis, the study explored information on the tax regime to build a potentially credible counterfactual necessary for the measurement of the policy effects. The main results suggest that treated firms increased the number of formal employment contracts by 14-16 percent, total working hours by 8-10 percent, and wages by 2-2.3 percent, in relation to untreated firms in the same sector and size range. These figures indicate that the payroll tax reduction resulted in employment generation, increase in labour demand, and in a small increase in wages in its first year of implementation. These estimates are statistically significant in almost all specifications. They seem robust to the exclusion of the textile sector, which was only covered later, as the effects become stronger if the analysis is limited to firms in the sectors covered earlier. The study also confirmed that employment and labour demand were not anticipated by firms in 2011, as the bulk of the policy effects seem to have happened after the policy adoption. Besides, weaker effects on employment and labour input were found when larger firms are included in the analysis. When the analysis is disaggregated by sector, the largest coefficients of variation in employment and wages were estimated in the leather and shoes industry, while the textile sector shows the smallest. The garment industry and services sectors fell in between the two. On the one hand, the leather and shoes industry is the most export-oriented sector, thus may have captured the full benefit of a payroll tax reduction designed to stimulate exports. On the other, the textile sector was only fully covered by the policy at a latter moment in time, which seems to explain the smaller effects in 2012. Empirical studies carried out in Latin-America and elsewhere tend to show small employment effects and partial shifting of payroll tax changes on to wages (Gruber 1997, Cruces et al. 2010, Bennmarker et al. 2009). This study points in the opposite direction, suggesting a substantial employment effect. The reasons might be that the analysis here is limited to a few sectors and does not attempt to evaluate the overall impact on the aggregate employment level. The employment effect is consistent with the theory that a decrease in labour costs results in an expansion in the demand for labour at firm or sector levels. The shift in demand interacts with an elastic labour supply to the formal labour market, determined by the size of the informal economy and the high unemployment rate, resulting in the adjustment by the employment with smaller effect on wages. 48 Results presented here contrast with two studies about the Brazilian payroll tax reform that found small positive effects on employment, though not across all covered sectors (Dallava 2014, FGV Projetos 2013). It is believed that the empirical approach adopted in this paper, and the control for firm fixed effects, offers a better solution to the identification problem that comes from the evaluation of such policy, as seen by more intuitive results and highly significant estimates. Once data for 2013 and 2014 are available, a broader policy evaluation using this approach could include sectors that were covered in the second half of 2012 onwards. It is also possible to undertake further evaluations using other data sources that contain information about relevant covariates such as capital intensity, and gross revenue. Moreover, the aggregate impact of the payroll tax reduction and its substitution by a revenue tax could be evaluated more broadly by use of an adequate methodology and including fiscal and inter-sectorial interactions. The policy seems to have achieved the expected increase in formal employment and perhaps an enhanced competitiveness, at least for the covered industries and in its first year. However, given the limitations of this study, these conclusions do not imply that aggregate effects will be the same. Besides, reductions in other non-wage labour costs in Brazil may not have similar effects, as changes in each component of labour costs will trigger specific effects on the labour market. Instead of a simplistic generalization, it seems also necessary to consider the implications of changes in social security, training, and the financing of other social programmes