Lenders either sell or obtain insurance for many of the mortgages they originate to reduce credit risk and enhance liquidity. An overwhelming majority of the mortgages sold are purchased by government-sponsored enterprises. The prevailing view is that government-sponsorship of mortgage securitization causes mortgage rates to be lower than they would otherwise be. Using a model that incorporates asymmetric information and adverse selection, we provide an example in which government-sponsored mortgage securitization raises the mortgage rate.Lenders either sell or obtain insurance for many of the mortgages they originate to reduce credit risk and enhance liquidity. An overwhelming majority of the mortgages sold are purchased by government-sponsored enterprises (GSEs), The prevailing view is that government sponsorship of mortgage securitization causes mortgage rates to be lower than they would otherwise be. As the argument goes, government sponsorship provides a cost advantage that is passed on to borrowers in the form of lower mortgage rates. This conclusion follows from standard economic models under the assumption of symmetric information.In our paper, we follow the recent literature on insurance markets by making the more realistic assumption of asymmetric information. We assume that the originator of the mortgage (which we will refer to as a bank) has an information advantage over the mortgage securitizer. 2 The bank may acquire, at some cost, information on a mortgage applicant's default risk, while the securitizer is unable to acquire such information. In other words, the bank may screen the applicant but the securitizer cannot. Under this assumption, our model yields the unconventional result that mortgage rates may rise for realistic values of the model's parameters. We provide an example in which government-sponsored mortgage securitization raises the mortgage rate.The intuition for this result derives from the strategic behavior of the mortgage securitizer. The securitizer knows that banks have incentives to screeen applicants and then sell the high-risk mortgages (lemons) to the securitizer while keeping the profitable low-risk mortgages (cherries) in their own portfolios. This is a classic lemons problem. Yet, in our model, the securitizer may alter the proportion of lemons sold by manipulating the guaranteed rate of return provided to banks that sell mortgages.