The 'detrimental effect' of corruption in Brazil: 'who pays more?' Applications using instrumental variable quantile regression

Abstract This study analyzes the impact of corruption on the dynamics of investment by Brazilian companies. For this, was employed a sample of 3,444 companies according to microdata available by the World Bank Enterprise Surveys. Using instrumental variable quantile regression estimator, the model results indicate that, in the absence of control due to endogeneity problems, the traditional estimator Koenker and Basset tends to underestimating the impact of corruption in reducing the elasticity coefficients of business investment. Firms located in the upper quantile of the conditional distribution of sales have better results in the implementation of investments even after the perceptual effect of corruption, unlike firms in lower quantiles. This suggests that firms with high performance are less "impacted" with the presence of corruption in relation to firms with poor performance.