We estimate a structural model of endogenous search and loan size choice to quantify how asymmetric information and search frictions jointly distort the auto loan market. 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. Empirically, we find evidence of adverse selection on the intensive margin: borrowers who demand more credit are significantly riskier. Using instrumental variables, we show that this correlation is not driven by moral hazard, but by adverse selection. Furthermore, borrowers who search more are riskier and face higher interest rates, indicating that riskier borrowers have more incentives to search. On the demand side of our model, borrowers choose loan size and search intensity, with private information over their search costs and willingness to pay. Car dealers can supply quotes at no search cost to the borrower, which shapes how much they search on their own. Riskier borrowers select into search because their higher reservation rates expose them more to rate dispersion. On the supply side, we allow asymmetric lenders to compete over rates with non-linear menus depending on the loan size. Our estimates show that allowing lenders to perfectly observe consumers’ private information would increase originations and borrowing. Larger loans make search more attractive, reducing average rates in equilibrium, given that search costs are high. Borrowers also derive significant surplus from the intermediation provided by car dealers, as most of the spread over a frictionless benchmark is the cost of a thin consideration set.