![Sample our Urban Studies journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110826/TOC-Subject-Sample-2021-Urban-Studies.jpg"})
Abstract
Using a screening model with asymmetric information, I evaluate the positive and normative effects of the subsidized default insurance policy in the U.S. mortgage market. The model implies that the subsidy raises interest rates for eligible mortgages, which is contrary to conventional wisdom but is consistent with the empirical evidence in Zhao (Citation2019). Moreover, the model implies that the subsidy hurts borrowers it was intended to help, as well as raises the aggregate mortgage default rate. My article highlights the adverse impact of the subsidy on welfare and financial stability, and sheds light on the root cause of the global financial crisis. It also provides potentially useful reference to other countries that have (or are considering adopting) a mortgage subsidy mechanism similar to that in the US.
Acknowledgment
The article is based on the theoretical section of Chapter 1 of my PhD dissertation at New York University in May 2016, and was completed prior to my IMF employment. I am deeply indebted to my dissertation committee members Viral Acharya, Hunt Allcott, Tim Christensen, Kris Gerardi, William Greene, and Ennio Stachetti. I am also extremely grateful for the helpful discussions with Deepal Basak, Alessandro Bonatti, Rodrigo Cubero, Yeon-Koo Che, Udaibir Das, Eduardo Davila, Joyee Deb, Anthony DeFusco, Kfir Eliaz, Eduardo Faingold, Alan X. Feng, Chris Flinn, Scott Frame, Dan Greenwald, Joseph Gyourko, Martin Hackmann, Thomas Holmes, Ali Hortacsu, Przemek Jeziorski, Tumer Kapan, Anastasios Karantounias, Richard A. Koss, Sam Kruger, Alessandro Lizzeri, Federico Mandelman, Niko Matouschek, James Morsink, Isabelle Perrigne, Jose Daniel Rodriguez Delgado, Paul Scott, Bowen Shi, Bruno Strulovici, Petra Todd, Miguel Urquiola, Felix Vardy, Quang Vuong, Lawrence White, Tao Zha, Jidong Zhou, participants at the IMF MCM Policy Forum in November 2016, and especially Stijn Van Nieuwerburgh and Alex Murray. All remaining errors are my own. The views expressed here are those of the author and do not necessarily represent the views of the IMF, its Executive Board, or IMF management.
Notes
1 Throughout the article, the terms “bank” and “lender” are used interchangeably, so are the terms “borrower,” “consumer,” and “customer.”
2 Note that, with an underpriced default insurance scheme, the bank will optimally choose to over insure by taking excessive risk. There are two ways to do so. First, the bank can take excessive risk through an extensive margin, i.e., by relaxing the lending criteria and extending mortgage loans to more borrowers. This is the margin studied by Keys et al. (Citation2010). Second, the bank can take excessive risk through an intensive margin, i.e., by raising the interest rate and/or leverage ratio of the same borrower (for the same borrower, a higher interest rate and/or a higher leverage ratio means a larger repayment burden and thus a higher default risk). It is the intensive margin that is the focus of this article.
3 The paper also implies that the GSE subsidy could actually lead to a lower homeownership rate among low-income households rather than a higher one (as claimed by GSEs). This is because, in the model, the subsidy increases the mortgage rate and makes it harder for low-income people to own houses. This could have some important welfare implications because a higher homeownership rate is associated with higher crime-prevention incentives by residents (as in Hoff and Sen, Citation2005), higher self-employment rates when home prices are rising rapidly (as in Harding and Rosenthal, Citation2017), etc. But as noted earlier, the focus of the article is on the intensive margin rather than the extensive margin.
4 A more detailed review of this literature is provided in Zhao (Citation2019).
5 Of course, this also means that my article does not capture the systematic impact of GSEs on the macroeconomy manifested in the risk-free rate, as in any partial equilibrium model.
6 I choose to model strategic default because the focus of the article is on how the strategic behavior of borrowers (and the lender) could fundamentally reverse the conventional wisdom (that the GSE subsidy decreases the mortgage rate). If we choose to model illiquidity default instead, then the focus would be on macroeconomic/employment shocks that are not under the control of the borrowers’ strategic decisions.
7 Note that the revelation principle is valid because the principal (lender) has full commitment to his/her decisions. Without this full commitment, there would be some “cheap talk” from the borrowers; the revelation principle would not be applicable, and the IC constraint would be invalid.
8 By the revelation principle, this is equivalent to saying that all other borrowers with will have no incentive to accept the contracts designed for type θ0 and higher types, where θ0 is the lowest participating type.
9 Note that the “default threshold effect” is zero in equilibrium because the borrower is indifferent between repayment and default at the threshold housing price that is,
10 There are two differences: first, we need to ensure there is no across-group deviations; second, we need to deal with the discontinuity of at θ1 introduced by the subsidy, so we cannot directly use the elasticity argument as in Lemma 2.
11 The intuition is as follows: b and θ are inversely related in this model, so all loans in the neighborhood around θ0 are jumbo loans and are ineligible for the subsidy. Therefore, when the bank makes the decision of decreasing or increasing θ0 by ϵ, the bank cannot get the subsidy from loans in the neighborhood of θ0 anyway, so the introduction of the subsidy has no effect on the bank’s optimality condition for choosing θ0, hence, no effect on the equilibrium θ0. Appendix H also shows that the subsidy increases both the loan size (household leverage) b and interest rate R in the jumbo group (for approved borrowers).