34 Pages Posted: 13 May 2005
Date Written: May 10, 2005
One of the primary results of studying asymmetric information as part of information economics over the last 30 years has been a theory of markets for lemons - that is, markets for goods that are likely to be of poor quality. Unfortunately, it has proven difficult to directly test the lemons theory because the causative factor - asymmetric information - implies that key quality variables are not observable. In this paper, we avoid this problem by studying the market for mortgage backed securities (MBS), a market which exhibits significant asymmetric information on an ex ante basis, but ex post reveals individual security qualities. In this market, some of the MBS produced in any given month are repackaged into multi-class security structures where certain classes enjoy higher priority claims on the underlying mortgage cash flows than others. As shown in DeMarzo and Duffie (1999), these security structures are optimal when the market values of the underlying assets are sensitive to private information held by the issuer of the securities. Utilizing a comprehensive dataset of all Freddie Mac Gold Participation Certificates (PCs) issued between 1991 and 2002, we show that, consistent with lemons theory, the PCs used to construct multi-class MBS tend to be of lower quality than other PCs. Our results suggest that structured multi-class securities, which after years of explosive growth now surpass traditional single-class securities in importance, are a market response to the problem of efficiently allocating capital when problems of asymmetric information are acute. Our results also have implications for public policies related to government sponsored enterprises.
Keywords: Asymmetric Information, Mortgage-Backed Securities
JEL Classification: D82, G12, G14, G18
Suggested Citation: Suggested Citation
Downing, Chris and Jaffee, Dwight M. and Wallace, Nancy, Information Asymmetries in the Mortgage-Backed Securities Market (May 10, 2005). AFA 2006 Boston Meetings Paper. Available at SSRN: https://ssrn.com/abstract=722423 or http://dx.doi.org/10.2139/ssrn.722423