City Research Online

Abstract

This dissertation proposes an integrated measure of credit and market risk for interest rate swap portfolios. Our research is based upon the use of EUR interest rate and credit spread data for the period 2001 - 2004. There are three self-contained but seemingly related projects in this dissertation. The objectives of this research are: 1) to price interest rate sensitive and credit spread options under the Longstaff & Schwartz 1992 framework; 2) to devise an integrated measurement approach of credit and market risk; 3) to extent the proposed integrated approach in measuring economic loss and compare it with the current standard approach.

The mean reverting and GARCH characteristics of EUR credit spread indices were investigated between 2001 and 2004. We find evidence of significant GARCH effects in the EUR credit spread indices and mean reversion which is dependent on the frequency of the time series. These properties of the EUR credit spread indices suggest a stochastic term structure volatility model would be suitable to model their evolution. The model of choice was used to estimate discount curves based upon the observed interest rate term structure. The range of yield curve shapes fitted accurately was extensive suggesting the model would be suitable in fitting the more complex credit spread curves. The estimation of yield curves over a period of time suggested that the volatility of the model parameters is reduced substantially with the use of weekly data. The model was able to match the market implied volatility in pricing interest rate options with greater accuracy in the pricing of short-term options. The LS model was quite successful in the fitting of various credit spread curves Although the pricing of credit spread options using the LS model is internally inconsistent evidence suggested that it prices short term spread options with good accuracy. The direct link between credit spreads and default probabilities was fully exploited by estimating implied default probabilities. Evidence suggested that the implied default probabilities did not violate the no-arbitrage conditions of credit risk pricing. A time series examination showed only in one occasion that a lower rating had a lower probability of default than its immediate higher rating. Also the historical transition matrix of S&P proved to be quite far from the expectations of the credit markets.

A dynamic approach to manage the risks associated with 10 different rated hypothetical interest rate swap portfolios was proposed based upon a hedging methodology. The proposed dynamic hedging of the swaps default risk is done by taking offsetting exposure related positions in respective credit spread index options. The efficiency of the hedging methodology shows strong linkages between the swap exposures and the credit spread index options. The integrated measure is proved to be higher at all times than the market VaR of swaps. Evidence suggest that the credit risk part of the integrated measure is not correlated with its respective market risk. The approach illustrates in a single overall market VaR measure both the market and the implicit credit risk run by a portfolio of swaps over a specified time horizon and confidence level.

The proposal of the integrated measure was put further to the test by performing a comparison between the existing methodology of integrated credit risk measurement and the proposed analytic “integrated” methodology. The comparison was performed on an actual swap portfolio taken from a medium sized European Bank. The comparison yielded similar results with the integrated approach measuring higher economic loss. The link of the expected credit exposure to the credit spread index option was evident. The use of historical simulation (HS) over a multi-step Monte Carlo (MC) simulation to measure expected credit exposures over a 3 month period is proved to be less accurate but not substantially different across all cases suggesting that over small time intervals the HS method is a fast and efficient way of measuring expected credit exposures.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HF Commerce H Social Sciences > HG Finance
Departments:	Bayes Business School Bayes Business School > Bayes Business School Doctoral Theses Doctoral Theses

Export

Downloads