The Influence of Collateral on Capital Requirements in the Brazilian Financial System: an approach through historical average and logistic regression on probability of default

Using data drawn from the Brazilian Central Bank Credit Information System, this paper evaluates the impact of the use of collateral on the probability of default and, consequently, on capital requirement levels in the Brazilian financial system. Literature suggests that the existence of collateral in some credit operations increases the debtor's readiness to honor its commitment and, therefore, could result in a lower probability of default. The methodology used to calculate capital requirements is based on the Basel II IRB-Foundation Approach, although the probabilities of default have been estimated by historical averages following Basel II orientation, and corroborated by a logistic regression model. The test of hypothesis about difference between collateralized and uncollateralized probabilities of default for each risk class indicates that they are statistically different. This result was obtained both from historical average probability of default as from logistic regression model.