Asymmetric information in loan contracts: New evidence from Italian big data

We provide new empirical evidence for an open problem regarding asymmetric information in loan contracts: the effect of higher collateral requirements on the interest rates applied by banks to borrowers is not clear under such asymmetry. Previous literature has argued both for positive and negative links, based on different models and econometric analyses. Our purpose is to examine recent Italian data, analyzing for the first time big data for thousands of borrowers collected by means of a European Central Bank project. First, we apply an unsupervised analysis to this unique microeconomic data set with a focus on the link between interest rate and loan to value. We do not find evidence of a strong link between those variables. Then, we develop a game-theoretic model, which supports the empirical results, based on the principal-agent problem adapted to the specific case of loans. Finally, we analyze loan data in a supervised setting, controlling for different borrowers’ categorical variables related to the interest rate determination. We conclude that the interest rate in loan contracts is influenced by asymmetric information and that higher collateral is not necessarily associated with a lower interest rate.