Default contagion among credit modalities: evidence from Brazilian data

The aim of this paper is to assess the impact of defaulting on one personal credit modality on future defaults on other modalities. Using Brazilian microdata, we run a logistic regression to estimate the probability of default on a given credit modality. We do so by including personal overdue exposures in other credit modalities among the explanatory variables. Our results show that the effect of this is positive and significant although not uniform across the different modalities. Specifically, it was found that defaults on collateralized credit modalities (vehicle and real estate financing) have a greater impact on future defaults on other credit modalities. Moreover, defaults on riskier loan categories (overdraft and nonpayroll-deducted personal credit) are more greatly affected by defaults on other modalities. We also discuss argumentatively the rationale behind these results.