Are Lemons Sold First? Dynamic Signaling In The Mortgage Market

Are Lemons Sold First? Dynamic Signaling In The Mortgage Market

Manuel Adelino
Duke University; Duke Innovation & Entrepreneurship Initiative

Kristopher Gerardi
Federal Reserve Bank of Atlanta

Creating Strategic Value With Joseph Calandro Jr.

InvestingValueWalk's Raul Panganiban interviews Joseph Calandro Jr., Managing Director of a global consulting firm and fellow of the Gabelli Center for Global Security Analysis at Fordham University. Q2 2020 hedge fund letters, conferences and more Interview with Joseph Calandro Jr. ValueWalk's . . . SORRY! This content is exclusively for paying members. SIGN UP HERE Read More

Barney Hartman-Glaser
University of California, Los Angeles (UCLA) – Anderson School of Management


FRB Atlanta Working Paper No. 2016-8


A central result in the theory of adverse selection in asset markets is that informed sellers can signal quality by delaying trade. This paper uses the residential mortgage market as a laboratory to test this mechanism. Using detailed, loan-level data on privately securitized mortgages, we ?nd a strong relation between mortgage performance and time-to-sale. Importantly, this ?nding is conditional on all observable information about the loans. This effect is strongest in the “Alt-A” segment of the market, where loans are often originated with incomplete documentation. The results provide some of the ?rst evidence of a signaling mechanism through delay of trade.

Are Lemons Sold First? Dynamic Signaling In The Mortgage Market – Introduction

One of the most widely studied market settings in economics is that of a seller with private information about the quality of an asset facing less informed buyers. In the presence of such an adverse selection problem, sellers can take actions that reveal their private information as in the classic signaling model of Spence (1973). This notion of signaling has been successfully applied in theoretical models of financial markets to explain a variety of phenomena from the optimality of debt (DeMarzo and Duffie (1999)) to the fragility of over-the-counter markets (Daley and Green (2012)). However, there is remarkably little empirical evidence that agents actually engage in costly signaling to overcome informational asymmetries. This paper uses data on the U.S. mortgage market to test a central prediction of dynamic signaling models, namely that sellers signal asset quality by delaying the sale of higher quality mortgages.

In dynamic environments with adverse selection and durable assets, sellers can signal private information through the timing of their trades. Sellers of high quality assets face a lower cost of waiting because assets produce a higher interim dividend stream, or, equivalently, have a lower probability of a negative event (e.g., default). Thus, a central prediction of the costly signaling framework is that there is a positive relationship between unobserved asset quality and time-to-sale, often referred to as the skimming property.1 We provide evidence of the skimming property in the originate-to-distribute mortgage market by showing that mortgages that take longer to sell default at a lower rate after controlling for observable characteristics.

To motivate our empirical tests, we first present a simple model of mortgage sales. This model illustrates the skimming property and shares several features with many models in the literature. In our model, a mortgage originator (the seller) faces a competitive market of buyers for the mortgage. The seller privately observes the quality of the mortgage (measured by its probability of default) and we assume that default is publicly observable and extinguishes the possibility of sale. The fact that the gains from trade are lost if the mortgage defaults before it is sold implies a higher cost of waiting for sellers of lower quality mortgages. A separating equilibrium emerges in which time-to-sale perfectly reveals mortgage quality. This equilibrium provides the key prediction that mortgages that take longer to sell should default at a lower rate.

There are two central requirements for a test of any signaling equilibrium, including tests of the skimming property in asset markets. First, there has to be a plausible source of unobserved heterogeneity in asset quality that is (i) known (at least partially) by the seller, (ii) unknown to potential buyers, and (iii) known to the econometrician. That the private information of the seller is also known to the econometrician is a fundamental challenge in testing models of adverse selection. The second requirement is that at least a subset of observable characteristics that are available to the agents is also observable by the econometrician. The distinction between observable and unobservable asset characteristics is key to any such test, as the predictions of signaling models only apply to unobserved heterogeneity. In fact, most models predict that assets that are observably better should trade faster, not slower. As a result of these requirements, this core prediction of dynamic adverse selection models has yet to be empirically tested.

The mortgage market serves as a unique laboratory for such a test. First, mortgages are durable assets that are characterized by an objective measure of quality based on the probability of repayment (i.e. credit risk). While outcomes were not known at the time of sale, they are known to the econometrician ex post. Second, during the middle of the last decade there was an active secondary market for mortgages, where issuers of mortgagebacked securities (the buyers) purchased large portfolios of mortgages from originators (the sellers). Third, there is detailed micro data available on the observable characteristics of borrowers and mortgage contracts known to issuers (the buyers), the originators (the sellers), and the econometrician. This provides us a good proxy of observable mortgage quality at the time of the sale. Finally, there is evidence from previous studies that the originators of mortgages have private information that is correlated with ex post mortgage performance but is not available to the buyers. By conditioning on observables and tracking mortgage performance, we are able to distinguish between variation in mortgage quality that is due to observable characteristics from “excess” variation in default behavior. The central test in this paper is whether excess default is related to time-to-sale.

Using data on mortgages securitized in the non-agency, private-label securitization (PLS) market, we find a clear negative relationship between time-to-sale and the component of mortgage performance that is not explained by mortgage characteristics. In our baseline specifications we find that, after conditioning on all underwriting characteristics, PLS loans sold five months or more after origination are approximately 5 percentage points less likely to default relative to loans sold immediately after origination. This is an economically meaningful difference, as it is approximately 30 percent of the average default rate in our sample (16 percent). Our empirical results are robust to different ways of defining default, alternative default horizons, different specifications, and, most importantly, to alternative data sources that differ in their representativeness of the PLS market. These results are consistent with previous studies that have found an important role for private information in the PLS market (Demiroglu and James (2012a), Jiang et al. (2014b), Griffin and Maturana (ming), and Piskorski et al. (2015)).

The results on ex post default are in contrast to those using ex ante measures of risk. In fact, we find no relation between summary measures of credit risk at origination and time-to-sale, even though credit risk at origination is strongly related to performance. Put differently, while unobserved quality is related to delay of trade, observable risk measures are not.

Mortgage Market, Dynamic Signaling, PLS Market

See full PDF below.