ABSTRACT
This study combines information in the Trade Reporting and Compliance Engine (TRACE), the most extensive bond database, with other fixed income data sets in order to quantify the default and liquidity premia in a reduced form framework at the bond level. The results show that the default premium is two to three times as large as the liquidity premium, and the liquidity term structure is not decreasing in the time to maturity.
KEYWORDS:
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 Longstaff, Mithal, and Neis (Citation2005) report that EquationEquation (2)(2) (2) performs reasonably well and that alternative specifications with a mean-reverting drift do not work better than Equation (2).
2 Andrade and Kaplan (Citation1998) estimate default losses of 10% to 23% of the asset value. Consistent with this empirical evidence, I set the default costs () to 20%. Kuehn and Schmid (Citation2014) also assume that the default losses are 20% in their model.
3 Refer to Duffie and Singleton (Citation1999) and Longstaff, Mithal, and Neis (Citation2005) and for further details.
4 Edwards, Harris, and Piwowar (Citation2007) describes the TRACE dataset in detail.
5 The Enhanced TRACE is lagged by 18 months.
6 I report the median numbers in order to avoid the impact of extreme values.
7 AA- and A- are exceptions.
8 Modigliani and Sutch (Citation1966) originally developed the preferred habitat hypothesis for liquidity to explain twists in the term structure. Ogden (Citation1987) and Kotomin, Smith, and Winters (Citation2008) furnish a review on this literature.