The approach to the measurement of credit risk recommended by the new Basel Capital Accord (Basel II) gives a wide choice of basic risk estimators. However, the rules for estimating asset correlations are defined in an ambiguous manner. The main aim of this paper is to verify the assumptions made by the Basel Committee on asset correlation. The authors thus refer to results of other researchers and present the results of research conducted based on a portfolio of automobile loans granted to individuals. They verify the hypothesis that the level of correlation established by the Basel Committee is overestimated for retail loans and show that, for the auto portfolios analyzed, the real level of correlation is much lower than that recommended in the new Basel Capital Accord. The results and their practical implications are presented together with a review of other studies.