House Price Risk Models for Banking and Insurance Applications

Abstract

The recent international credit crisis has highlighted the significant exposure that banks and insurers, especially mono-line credit insurers, have to residential house price risk. This paper provides an assessment of risk models for residential property for applications in banking and insurance including pricing, risk management, and portfolio management. Risk factors and heterogeneity of house price returns are assessed at a postcode-level for house prices in the major capital city of Sydney, Australia, over the period 01-1979 to 03-2011. The paper shows how a significant proportion of house price variability is due to heterogeneity requiring broader risk assessment than market-wide house price indices. Although time series models of market price indices capture the temporal risks of house prices, panel data models with random effects and variable slopes are required to capture cross-sectional heterogeneity and to quantify the risk of postcode-level house prices compared to the market price index. Macroeconomic and financial variables, as well as geographic and socio-demographic postcode characteristics are shown to be important house price risk factors.

Keywords: house price risk, statistical models, risk management

JEL Classification: G21, G22, G32, R31, L85