This article estimates default intensities within the continuous-time Jarrow and Turnbull model for German bank and corporate bond prices. It is shown that a joint implicit estimation of the default intensity and the recovery rate is numerically unstable. In addition to cross-sectional estimations, separate estimations (for each bond individually) are performed. Results strongly support separate estimation over the building of any cross-sections. In contrast to preceeding literature, the optimum volume of data required to provide reasonable estimates of the default intensity is also investigated. It is shown that calibration based on daily data as a rule does not minimize the ex ante mean squared pricing errors. Finally, it is shown that the constant default intensity assumption is not sound with the underlying data and the determinants of the default intensity are investigated. Regressions show that the lagged default intensity estimate, the level of the default-free term structure and liquidity proxies affect the estimated default intensity via joint parameters.
- Credit risk
- Intensity-based models
- Jarrow/Turnbull model
- Term structure of interest rates
ASJC Scopus subject areas
- Economics, Econometrics and Finance (miscellaneous)