Downside Risk and the Size of Credit Spreads

by Gordon Gemmill of the University of Warwick, and
Aneel Keswani of the City University, London

April 2010

Abstract: We investigate why spreads on corporate bonds are so much larger than expected losses from default. Systematic factors make very little contribution to spreads, even if higher moments or downside effects are taken into account. Instead we find that sizes of spreads are strongly related to idiosyncratic-risk factors: not only to idiosyncratic equity volatility, but even more to idiosyncratic bond volatility and idiosyncratic bond value-at-risk. Idiosyncratic bond volatility helps to explain spreads because it reflects not just the distribution of firm value but is also a proxy for liquidity risk. Idiosyncratic bond value-at-risk adds to this by capturing the left-skewness of the firm-value distribution. We confirm our results both for the initial 1997-2004 sample period and also out of sample for 2005-2009, which includes the sub-prime crisis. Overall, credit spreads are large because they incorporate a large risk premium related to investors' fears of extreme losses.

JEL Classification: G12.

Keywords: Bond, Idiosyncratic risk, Downside risk, Credit spread puzzle, Pricing kernel, Liquidity, Sub-prime crisis

Published in: Journal of Banking & Finance, Vol. 35, No. 8, (August 2011), pp. 2021-2036.

Previously titled: Downside Risk, Liquidity and the Size of Credit Spreads --and-- Idiosyncratic Downside Risk and the Credit spread Puzzle

Download paper (239K PDF) 35 pages

[