The subprime mortgage market is in free fall. Since the end of 2005, default rates on subprime mortgages have soared from 6.5% to over 30%, while foreclosure rates have jumped. Future increases are seemingly inevitable, and the prospects for recovery are threatened by the ongoing turmoil in financial markets, and the curtailment of credit to distressed homeowners.
The precipitous nature of this deterioration, and its potential spillover to the broader economy, has provoked a public outcry for swift and decisive policy responses. Accordingly, Congress, the Executive Offi ce of the President (EOP), the Treasury Department, and various federal and state mortgage regulators are all in the process of reviewing the rules under which new subprime mortgages may be issued, and renegotiating the terms under which previously&issued mortgages can default or be repaid. The EOP, in particular, has instructed the Federal Housing Administration to expand its insurance of mortgages in order to help creditworthy borrowers secure more favorable refinancing. Fannie Mae and Freddie Mac, on the other hand, have both increased their exposure to mortgage assets in an effort to shore up the housing market. This policy response assumes that subprime defaults are a result of temporary failures in both origination and securitization markets, and that government intervention is needed to correct and mediate these failures.
This paper provides an alternative story for the subprime crisis and hence differs in its policy prescriptions. Rather than focusing on correctable market failures, we characterize subprime developments as outcomes of rational errors in the risk perceptions of market participants. The model we develop rationally explains the pattern of mortgage default rates (both prime and subprime) from 2001 through 2007, and captures both the boom and the bust of the subprime market. While investors have long agreed on the riskiness of individual subprime mortgages, no such agreement exists over the risk in pools of subprime assets. Accordingly, the key element in our model is the evolution of investor beliefs over the importance of aggregate risk to subprime portfolios.
In our model, investors view mortgage borrowers as a dividend process, which pays a contractual stream of income while the mortgage remains current and a final payment when the mortgage either defaults or is paid off. The properties of the dividend process are governed by the evolution of the mortgage borrowerms credit quality. Investors have knowledge over the initial value of the householdms credit quality and understand the factors which determine its evolution. They do not, however, know how important aggregate shocks are to this evolution, and cannot directly observe credit quality after the initial date.
Initially, in 2002, investors draw random beliefs of the importance of aggregate risk in subprime portfolios. Investors who draw low&weight beliefs infer a very high Sharpe ratio for subprime assets. In contrast, investors who draw high&weight beliefs infer a low Sharpe ratio, and purchase very few, if any, subprime assets.
Investors subsequently use incoming information on aggregate defaults and early mortgage payoffs to update their beliefs. Low and stable default and prepayment rates in 2003 and 2004 lead investors to believe the weight on aggregate risk is lower than previously assumed. As a result, over time, the pool of potential subprime investors broadens, increasing the size of the subprime market, pushing down spreads and degrading underwriting standards.
We calibrate the model to replicate the aggregate default and prepayment rates observed in 2003. By varying only the initial average credit quality, we are able to match both subprime and prime mortgage default rates, as well the prepayment rates. In calibrating the model, we assign a high weight to the aggregate shock. However, model investors, who observe the low and stable defaults, move toward beliefs of a low weight on the aggregate shock.
Contents
1 Introduction
2 The Rise and Fall of the Subprime Market
3 Literature Review
4 The Model
4.1 The Household Problem
- 4.1.1 Renterms Problem
4.1.2 Ownerms Problem
4.2 The Mortgage Market
4.3 Individual Mortgages
4.4 The Investor Problem
5 Data
6 Calibration and Parameterization
7 Simulation: The Subprime Market 2002 to 2007
7.1 Inferring the Value of ? in 2003 and 2004
7.2 The Explosion of Subprime Lending: 2002&2005
7.3 The Collapse of the Subprime Market: 2006&2007
- 7.3.1 The Probability of Observing 2006 Defaults using 2003 Parameter Values
7.3.2 Increasing the Variance and Mean of the Aggregate Shock
7.3.3 Deteriorating Initial Credit Quality Cannot Explain 2006
8 Evaluating Policy Options
8.1 A Moratorium on Foreclosures
8.2 Government Guarantee of Mortgages
8.3 A Decrease in Mortgage Interest Rates
8.4 Cash Transfers to Households
9 Conclusion
Download
PDF Ebook The Subprime Mortgage Crisis: Irrational Exuberance or Rational Error?
