Determinants of mortgage pricing: A quantile regression analysis.

This paper examines the impact of consumers’ creditworthiness, loan terms, loan amount, loan-to-value ratio, and loan purpose on mortgage pricing while controlling for market condition, macroeconomic performance, and loan sources. Applying quantile regression methods based on a lender-provided dataset covering the period of 2002–2007, this paper is able to differentiate these effects with respect to distribution of mortgage pricing. The regression results indicate that loan applicants with higher levels of creditworthiness in terms of credit score receive a lower mortgage premium above default and interest rate risk. For every 10 point increases in credit scores, the mortgage premium above default and interest rate risk declines consistently from by 1 basis point at the 10th quantile to 5 basis points at the 90th quantile. Loan applicants receive discounts ranging from 24 to 117 basis points, 22 to 115 basis points, 13 to 87 basis points for mortgage loans with LTV ratio below 60percent, between 60 and 70 percent, and between 70 percent and 80 percent, respectively. [ABSTRACT FROM AUTHOR]