Two Essays in Asset-Pricing
Past research documents a positive link between momentum and firm-level default risk, yet this anomaly is not connected to default risk at the macro level. Namely, there is no documented momentum during recessions, when default is higher on average. In the first essay, "Momentum and Aggregate Default Risk," we attempt to resolve this puzzle by analyzing momentum pro ts over time, conditional on both business cycles and unexpected changes in aggregate default risk. First, we show that momentum is driven by shocks to aggregate default, rather than general economic conditions such as expansions and recessions. Using the Fama and MacBeth procedure, we find that a conditional default shock factor is priced and can explain a large portion of the total momentum returns. Second, we provide a risk-based explanation for this anomaly by linking the returns of momentum portfolios to shareholder recovery during financial distress. We find that losers have higher recovery (i.e., shareholders have high bargaining power) on average, and, as a result, have relatively lower risk in high default states of the world. Therefore, loser stocks have a lower risk premium and lower expected returns in worsening aggregate default conditions, leading to the observed momentum. This effect is more pronounced among stocks of firms with low credit ratings. Our results help to reconcile the seemingly contradictory evidence documented by previous studies and o er a rational explanation for the momentum anomaly.
In the second essay, "Sources of Momentum in Bonds," we study the relationship between momentum in bond returns and aggregate default. We document that momentum in corporate bonds occurs mainly during periods of high default shocks and is driven by losers. Supporting this result, we find that conditional default risk is priced in the cross-section of corporate bond portfolios. Motivated by these findings, we develop a theoretical model connecting bond momentum returns to the ability of bondholders to recover value in financial distress. Specifically, we find that losers have relatively higher recovery potential and, therefore, become less risky when high default shocks occur. Thus, losers have lower expected returns in high default shocks, leading to the observed conditional momentum. Further, US government bonds, with default risk approaching zero, feature no momentum, however this anomaly prevails in sovereign bonds with positive default risk, consistent with our main results.