The Optimal Mortgage Loan Portfolio in UK Regional Residential Real Estate

Abstract

In this study, we propose a method based on large deviation theory (LDT), which minimises credit risk (expected loss). We demonstrate how mortgage loan portfolios can be optimised using geographical differences in the risk characteristics of mortgage loans in the UK. Our empirical results show that credit risk can be reduced by a third when the LDT method is used instead of the benchmark portfolios that we calculate with regional-gross-value-added weights and equal weights. More importantly, the difference in the expected loss between these portfolios increases further during bearish housing markets. To see that such numbers matter, in an extreme scenario, the UK mortgage lenders could lose more than 2% a year as the consequence of mortgage defaults, which is equivalent to an annual loss of approximately 20 billion pounds in the UK. Although this extreme state would not continue for a long time, it nevertheless represents a huge potential loss for mortgage lenders and investors.

Keywords: Large deviation theory, Mortgage loans, Optimal portfolio, Regional allocation