Can Higher-Order Risks Explain the Credit Spread Puzzle?

Abstract

We tweak the conventional Merton model to account for the asymmetric properties of assets returns and investors asymmetric behavior toward the upside potential of gain versus the downside risk of loss. Using an asymmetric split normal distribution, we capture empirical asymmetries in the underlying return distribution, while we conserve the attractiveness of delivering closed-form pricing formulas that collapse to the basic Merton model in the symmetric Gaussian case. The asymmetric specification outperforms the symmetric one in matching high levels of historical credit spreads. We then link the residual (non-default-model-implied) spread to two illiquidity risk factors. The first factor is extracted from several measures of idiosyncratic illiquidity variables and the second factor is a systematic factor obtained from a general index common to all studied bonds. Our model explains 70% of the BBB-AAA spread and more than 72% of BBB and AAA credit spreads relative to the on-the-run Treasury rates.