Abstract/Summary

Asset value volatility is at the heart of the capital structure optimisation theory as proposed by Leland (1994). Asset value and volatility are two key unobservable inputs required for pricing equity and debt using structural models that are based on option pricing theory. These inputs are normally calculated using equity prices, book value of debt and leverage. However, when tested empirically, the prediction errors of structural models suggest that such calculations of asset value and volatility may be inaccurate. Furthermore, in all structural models it is assumed that asset value volatility is constant over time, which may also lead to inaccurate credit spread predictions. This is the first empirical study into bond-implied asset values and volatility. This study uses three-year time series of daily bond- and equity prices for 36 Western-European companies to derive bond-implied asset values and asset value volatilities. Bond-implied values were calculated using a structural credit model developed by Leland and Toft (1996). The values were obtained by simultaneously solving a system of two equations and two unknowns. The two equations were the value of a bond and the value of equity; the .two unknowns were asset value and asset value volatility. An equity-based method for calculating asset value and volatility was used to generate an alternative data set for the same daily observations. The two sets of data for each firm were compared to determine whether there are statistically significant differences between bond-implied and equity-based asset values and volatility, and whether bond-implied volatility is stable over time. Additionally, the correlation between bond-implied volatility and the bond's time to maturity was tested to determine whether there is a constant relationship between these two variables. This study shows that there are significant differences between asset values and asset value volatilities derived from bond prices and those derived using equity values and leverage. Additionally, it is shown that bond-implied asset value volatility is not constant over time. The relationship between bond-implied asset value volatility and a bond's time to maturity as measured by their correlation coefficient was not uniform across the sample, in some cases demonstrating strong positive correlation, in other cases strong negative correlation and in some cases no strong relationship at all.