Reply to “Comment on: ‘Why is bank credit in Brazil the most expensive in the world?”’

To explore these arguments and improve the context of the discussion of low creditor rights in the original article, we run a series of interviews with market agents of the Brazilian lending industry, including employees of commercial banks, fintechs’ entrepreneurs and a former banker. Low creditor protection is included in the original manuscript, illustrated by a figure that contrasts the strength of creditors’ legal rights (replicated here as Figure 1), but it is true that the referred articles in the original study do not explicitly include low creditor’s protection as a dependent variable (in the original manuscript, we ponder that this is most likely due to the lack of data). But first, let us delve into the reasons for which market concentration is a fundamental determinant of the extremely high interest rates in Brazil, even though concentration is just a proxy for competition. Let us consider the counterfactual: market concentration would not contribute to higher interest rates only if, as Barbosa (2020) points out, we observe competitive behavior between the leading banks. Yet, price competition is almost unobserved in the Brazilian credit markets, either anecdotally or empirically. After all, that is what most authors cited in the original article are trying to test: what

1. The high concentration in the Brazilian banking industry may not significantly lead by itself to structurally higher lending rates.
2. Low creditor's protection has not been adequately addressed, in the original article. The interaction of low creditor protection with high concentration may better explain why lending rates are so high in Brazil.
To explore these arguments and improve the context of the discussion of low creditor rights in the original article, we run a series of interviews with market agents of the Brazilian lending industry, including employees of commercial banks, fintechs' entrepreneurs and a former banker. Low creditor protection is included in the original manuscript, illustrated by a figure that contrasts the strength of creditors' legal rights (replicated here as Figure 1), but it is true that the referred articles in the original study do not explicitly include low creditor's protection as a dependent variable (in the original manuscript, we ponder that this is most likely due to the lack of data).
But first, let us delve into the reasons for which market concentration is a fundamental determinant of the extremely high interest rates in Brazil, even though concentration is just a proxy for competition. Let us consider the counterfactual: market concentration would not contribute to higher interest rates only if, as Barbosa (2020)   drives lending rates. Potentially, we could observe strong competition in a concentrated market, but that does not seem to be the case in Brazilian credit markets. In fact, in Coelho et al. (2017) the presence of a public bank does not necessarily affect the conduct of private banks, given that banks may not compete at all on prices. Thus, lack of competition may not lead to higher lending rates. There is no need to distinguish the types of clients that public and private banks cater to; a collusive market can easily accommodate a new entrant if long-run profits are not strongly affected. Conduct and concentration, at least regarding banking competition in Brazil, seem to be correlated. Collusion also does not need to be perfect to engender social outcomes far from optimal. Partial collusion would be more than enough for that. Lucinda (2010) shows that collusion is found in Brazilian credit markets, for some firms and in some time periods. As the market has become more concentrated over the years, there is no compelling mechanism that would disentangle market power and concentration in Brazilian credit markets, given the original description of that context, and the breadth of empirical evidence about it.
Regarding creditors' rights, given that there is insufficient variation for the variable to be satisfactorily included in empirical specifications of studies that estimate the drivers of lending rates in Brazil, we have chosen to collect primary data to investigate whether market agents believe it is a fundamental driver of high spreads in Brazil. We cannot directly measure the effect of creditors' rights on the level or change of lending rates, but we should expect a positive correlation between market beliefs and high interest rates in Brazil. For that purpose, we surveyed 10 market agents, from August 24 to September 01, 2020. Interviews were conducted via Google Meeting. The goal is to crosscheck the variables in the academic literature with the perception of market participants. Interviews are divided in two sections: first, the selection of the most relevant variable that, in the opinion of the interviewees, explains the perennially high interest rates in Brazil, and then an unstructured conversation about the topic. Initial options are: Lack of competition between financial companies; Selic rate; Earmarked credit; Operational costs; Taxes; Low creditors' protection rights; Others. We ask the interviewers to rank these options. All interviewees but one chose low creditors' protection rights as the main explanatory variable. The one who did not -the former banker -picked lack of competition as the main explanatory variable to high interest rates in Brazil. Table 1 summarizes the results. By far, the most selected variables are lack of competition/concentration and low creditors' rights.
In the open-ended part of the interviews, three points were raised by market agents on how low creditors' protection rights influence lending rates: the default rate is high, the recovery rate is low, and the regulatory and tax treatments, such as rules for provisions, increase opportunity costs for lending. One fintech entrepreneur stated that his credit recovery rate is literally 0%. Another interviewee mentioned that his company has difficulties accessing the credit history of potential clients. Although mechanisms such as the Central Bank's Credit Information System (SCR) are designed to facilitate banks' access to credit records, interviewees note that the quality of information on non-banking credit and overall earnings capacity is poor, which results in models of credit risk assessment that are lacking. Interviewees also mentioned that, according to a 2016 World Bank study, Brazilian financial institutions recover, in bankruptcy proceedings, approximately 16% of the value of assets offered as collateral, whereas the median in other countries is 69%. Recovery also takes longer, around 4 years versus 2 years in the rest of the world.
Summing up, low creditors' protection rights and lack of competition, as proxied by market concentration, seem to be the main drivers of high lending rates in Brazil. The original article should have delved more deeply into the role of creditors' inability to recover credit as another fundamental driver of high lending rates in Brazil. Barbosa's comments expertly complement the ones in the original article, and his note on Alencar et al. (2020), who find that market concentration hampers 27.5% of the potential reducing effect of the law in the interest rate of new corporate credit operations, is illuminating. Unfortunately, there is no indication that either market concentration or creditors' rights are going to change in the next few years. Thus, we should expect credit to remain expensive and scarce, especially for small and medium size enterprises, an obstacle for Brazil to escape the middle-income trap.