This is a joint blog with Cesar Rodriguez.
The need for public intervention to expand access to credit for firms in emerging markets has been debated for long by policy makers and scholars. Even more controversial has been the role that government-owned Banks should play in these economies.
We focus our attention on the credit lines managed by the Brazilian publicly owned Bank BNDES, and the government’s innovation agency FINEP, and their potential effects on employment creation, labor productivity and export.
The results are quite interesting:
We found that access to public credit lines has a significant and robust positive impact on employment creation and exports. We do not find evidence of a significant effect on labor productivity (though, if you read the paper, you’ll see that this can also be due to the specific measure of labor productivity we use).
Interestingly enough, our findings show that impact on exports is driven by the increase in export volumes among exporting firms, while we do not detect any significant effect on the probability of becoming an exporter, a result that seems to confirm the position of those saying that public credit benefits mostly established firms, rather than new ones.
In addition to these findings, another very interesting feature of the paper is the data we use for the analysis. As mentioned before in this blog, much evaluation work can be done with datasets originally collected for other purposes.
In this particular case, the evaluation drawn upon two administrative datasets: the Relação Anual de Informações Sociais (Annual Report on Social information) which is an administrative file maintained by the Brazilian Ministry of Employment and Labor (Ministério do Trabalho e Emprego), and the Foreign Trade Dataset from the Secretariat of Foreign Trade (SECEX) of the Ministry of Development, Industry and Foreign Trade. These data were merged among them and with data on access to public credit line by the Instituto de Pesquisa Econômica Aplicada (IPEA).
As a result, we constructed a unique panel dataset which includes information on both firm-level performance and access to public credit lines.
Given this particular data set, we opted for a methodological approach based on a combination of matching and difference-in-difference to control for selection bias when estimating the impact of the credit programs.
Actually, the core of our estimation strategy is based on a difference-in-differences technique, which we complement with matching methods for robustness check.
We hope you’ll enjoy reading!