Sample Selection Bias in Mortgage Market Credit Risk Modeling
Abstract The mortgage crisis that started in the U. S. in 2007 and lasted until 2009 was characterized by an unusually large number of defaults on the subprime mortgage market. As a result, it developed into a global economic recession and placed the stability of the world banking system in jeopardy. Therefore, the issues of credit risk modeling showed the shortcomings of the current credit risk practice. Truncation, or partial observability, and simultaneous equations bias causes sample selection bias. As a result, parameter estimates are biased and inconsistent. Firstly, we provide an overview of current approaches in the mortgage literature to control for the sample selection bias correction, such as the Heckman model and bivariate probit model with selection. Secondly, a review of the most significant mortgage studies discussing this problem is introduced. Specifically, different structural models, specific datasets and empirical results are regarded. In addition, we discuss such key credit risk determinants as borrower characteristics, terms of the mortgage contract, mortgage characteristics, and macroeconomic conditions. Finally, we conclude the discussion with possible research questions.
Keywords Credit risk • Default • Mortgage • Sample selection bias
JEL Classification C10, C34, G21
Different concepts are used to measure credit risk, such as probability of default (PD); loss given default (LGD); exposure at default (EAD); maturity (M) and correlated defaults. Default is arguably more relevant to the recent subprime mortgage market collapse and related spillover effects. During the financial crisis, almost one out of ten mortgages was delinquent.
Default imposes enormous costs on all market participants. First, there are credit organizations and the Institute of the Mortgage Insurance Development (Russian
A. Lozinskaia (H)
National Research University Higher School of Economics, Perm, Russia e-mail: AMPoroshina@gmail. com
© Springer International Publishing Switzerland 2015
A. K. Bera et al. (eds.), Financial Econometrics and Empirical Market
Microstructure, DOI 10.1007/978-3-319-09946-0_____ 17
stock life insurance company “AIGK”). The latter company insures a borrower’s liability and financial risks of creditors. Second, a defaulted borrower at a minimum meets the cost of moving and damages the borrower’s credit score, making it difficult to buy another house and forcing a period of rental occupancy. In addition, a lower credit score seriously restricts access to credit approval in the near future. Finally, default is associated with additional psychic costs (Guiso et al. 2009).
Therefore, default modeling is an essential element of a risk management system in any credit organization. However, the notion of mortgage default has not yet been incorporated in the Russian legislation.
Usually, estimation results of default are obtained from a single-equation model, which allows for an important inference about the credit risk and key determinants of it. Moreover, test discrimination in the credit underwriting process plays a significant role in the mortgage supply decision. However, such estimates could be biased and inconsistent due to a sample selection bias. It leads to misunderstanding and misinterpretation of the obtained results.
In the first section, we analyze some widely used econometric models for credit underwriting and default processes, and focus on the sample selection bias problem. The second part reviews mortgage literature that discusses the problems of credit risk modeling and the sample selection bias. Then we discuss key credit risk determinants and conclude with main research questions and suggestions for further empirical work.