An Empirical Study of Hedge-Ratios for Corporate Bonds

Investments in corporate bonds, or credits, are subject to two important sources of market risk: interest rate risk and credit risk. 1. Interest rate risk: uncertainty in risk-free interest rate (government bonds) 2. Credit risk: uncertainty of potential downgrades and default These two market risks provide an opportunity for the active investor to gain additional returns or avoid losses. Active credit investors are specialized in evaluating and comparing the credit risk of corporate bonds. Given their analysis of credit risk these investors buy a portfolio of bonds of which they expect higher returns than the benchmark which usually consists of all available bonds in the market. However, an investment in credits may also lead to an unintended exposure to interest rate risk. Since the specialized credit investor does not have the skill to predict the risk-free interest rate he or she does not want to have such an exposure. Therefore, the investor wants to hedge unintended interest rate risk in the portfolio. The duration is a well known measure for interest rate sensitivity and is used to set up a hedge against the unintended interest rate risk in the portfolio. This hedge should immunize the loss and profit resulting from interest rate changes. De Backer (2006) [5] shows that this method of hedging is accurate for bonds with a low spread, but for bonds with higher spreads this is not an accurate way to immunize the interest rate risk as is shown by Ben Dor et al (2004) [1]. For high yielding bonds the hedge does not reduce the risk on the contrary it has even increased it during specific sub-periods in history. Therefore the hedging of corporate bonds to interest rate risk is an interesting field of investigation. Nevertheless, not much academic research has been published in this area which is further underlining the need to do new research. De Backer (2006) [5] proposes a new measure, Empirical Duration. Empirical Duration (ED) is the sensitivity of a credit investment to the interest rate risk. De Backer finds that the ED is depending on the credit spread and the maturity of the corporate bond. By combining 5 different intervals of the maturity and 6 different intervals of the credit spread a so called Maturity-Spread Table (MS-Table) is created with 30 distinct Maturity-Spread buckets. Each bucket has a different ED value, the ED of a specific corporate bond is determined by matching its maturity and spread to the buckets. The ED of a bond portfolio is the market value weighted average of the individual ED values of 4 the bonds. When the ED of the portfolio is different from the ED value of the benchmark the credit investor is able to hedge out the undesired interest rate risk exposure. Within the initial MS-Table standard intervals were used without very well thought-out considerations. The interval values of the MS-Table have a very big influence on the ultimate table of Empirical Durations which will be used to hedge against unintended interest rate risk. Therefore the interval decision of the MS-Table is highly important. We use a genetic algorithm to evaluate and improve the quality of the MS-Table that has been developed by De Backer. We are able to reduce the number of maturityspread buckets from 30 to 15. In addition the ED values of the new buckets are optimized. The new MS-Table has a 52% better fitness function than the MS-Table from De Backer. The new MS-Table delivers a better hedge of undesired interest rate to a credit investor.