Inferring bank-to-bank competition from dynamic time series analysis of price correlations

Inferring bank to bank rivalry and competition generally requires the estimation of a full demand model, with high data requirements, unavailable to most researchers. We suggest dynamic time series analysis of price correlations to infer about bank to bank competition, taking into account the well-known criticisms to price correlations for delimiting relevant markets. The method is applied for credit markets in Brazil, where bank monthly loan interest rates time series are available. We conclude that there is little rivalry between large banks in most of the credit markets studied.