Modelling demand for mortgage loans using loan-level data
This paper is concerned with modeling the demand for mortgage loans. The demand for loans can be represented as two functions: probability of borrowing and the loan amount, depending on borrower-specific characteristics, contract terms and set of macrovariables. The decision-making process for borrowing can be described as the sequence of decisions on: 1) choosing the credit program; 2) approving of a borrower; 3) choosing contract terms from a feasible set; 4) and loan performance. Following Philips and Yezer (1996) and Attanasio, Goldberg and Kyriazidou (2008) the author proposes an econometric approach that deals with endogeneity and self-selection of borrowers when estimating the demand-for-loan equations and specifies the structure of data that is required for implementation.