This paper studies the joint impact of asymmetric information and search frictions in the market for auto loans. Using administrative data from the Central Bank of Brazil, we observe the universe of loan contracts and, crucially, the specific set of lenders contacted by each borrower. We document two key empirical findings. First, we find evidence of adverse selection on the intensive margin: after controlling for observables using random forests, we estimate a positive correlation between loan size and default risk, which is consistent with unobservably riskier individuals self-selecting into larger loans. Second, we show that search frictions are a significant source of market power. Using the distance to bank branches as an instrument for search intensity, we find that borrowers who search more obtain significantly lower interest rates. To quantify the welfare implications of these frictions, we develop a structural model endogenizing borrowers’ search and loan size choices. On the supply side, we allow lenders to offer non-linear pricing menus. In next steps, we use this framework to evaluate the welfare effects of the introduction of credit scores in Brazil, specifically analyzing how search frictions can limit the consumer gains.