Did the 2017 labor reform reduce unemployment in Brazil? An analysis based on the dynamic multilevel latent factor model

This study evaluates the impact of the 2017 Brazilian labor reform on the unemployment rate. Implemented through Law No. 13,467/2017, the reform aimed to increase labor market flexibility as a strategy to address the rise in unemployment that began in 2014 following the economic contraction. Empirically, we employ the Dynamic Multilevel Latent Factor Model (DM-LFM), a Bayesian approach that offers methodological advantages over the traditional Synthetic Control Method (SCM), being particularly suitable for contexts with a limited number of control units and capable of incorporating multiple sources of heterogeneity in the data. The results are consistent with an adverse short-run effect of the reform on the labor market, suggesting an increase in the unemployment rate of approximately 3 percentage points in the first years following its implementation, although the magnitude and statistical significance of the effects are sensitive to the assumed treatment onset year. The analysis of transmission mechanisms suggests that this effect may have operated through a reduction in formal wage employment and industrial employment, accompanied by an expansion of self-employment and vulnerable employment. This pattern of occupational reallocation is consistent with the predictions of dual labor market theory and with evidence from economies characterized by high levels of informality. Robustness analyses, including placebo tests over time and across units, support the direction of the findings from the preferred specification, although the magnitude and statistical significance of the estimated effects are sensitive to the assumed treatment timing.