This dissertation first investigates the possible house price trend and the relationship with the mortgage market, from the perspective of risk management; then it chooses the angle from bond insurers and figures out possible methods to avoid capital procyclicality. In Chapter I, we apply vector auto regression models (VAR) and simultaneous equations models (SEM) to estimate the dynamic relations among house price returns, mortgage rates and mortgage default rates, using historical data during the time period of 1979 through second quarter 2008. We find that house prices would be better estimated and predicted with the consideration of the mortgage market. In Chapter II, following the methodology of co-integration, we first construct several succinct measures to display the possible intrinsic values of house prices. In the short run, house price return dynamics are investigated by dynamic adjustments following Capozza et al (2002) and error correction models. We examine the possible overshooting problem of house price returns. By analytical derivations and simulations, we demonstrate the effects of the coefficients on overshooting. In Chapter III, we adopt a structural model with time-varying correlations for bond insurers. We consider losses due to bond insurers’ downgrading and losses from both insurance contracts and investment portfolio. On that basis, we propose forward-looking smoothing rules of capital over a full business cycle, instead of only based on a short-term horizon, to avoid the procyclicality. With the smoothed capital, a bond insurer can actually establish some capital buffer in good times to support the potential losses in crisis.
| Identifer | oai:union.ndltd.org:GEORGIA/oai:scholarworks.gsu.edu:rmi_diss-1023 |
| Date | 08 August 2010 |
| Creators | Wei, Xiangjing |
| Publisher | ScholarWorks @ Georgia State University |
| Source Sets | Georgia State University |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Risk Management and Insurance Dissertations |
Page generated in 0.0018 seconds