Predictive Strength of Lending Technologies in Funding SMEs

Using a proprietary database of lending decisions (N = 9,898) for small and medium‐sized enterprises (SMEs), the paper investigates how banks cope with the adverse selection dilemma. Based on an intertemporal framework, we qualify incorrect and correct lending decisions of banks and investigate the power of lending technologies to predict errors and correct choices. Findings suggest that adverse selection can be better controlled by a durable bank–firm relationship, as well as by an atomistic loan decision process, at the local level. By contrast, a loan decision‐making process based exclusively on hard financial information about SMEs may lead to adverse selection errors.