Job Market Paper

Page

A Disaster Explanation for the Term Structure of Returns

This paper incorporates recovery of dividend in addition to the drop in dividend and the rise in inflation in the disaster framework, so that the cash flow effect of the disaster is contemporary for realized variance, transitory for dividends, and accumulative for nominal bonds. The new framework accounts for the following facts (1) the term structure of holding period return for nominal bonds is upward sloping while its Sharpe ratio is downward sloping, as found by Duffee (2010), (2) the term structure of holding period return for dividend strips and its Sharpe ratio are both downward sloping, as found by van Binsbergen, Brandt, and Koijen (2012), and (3) the Sharpe ratio of holding period return is significantly negative for one month variance forward, and close to zero for variance forward with maturity longer than two months, as found by Dew-Becker, Giglio, Le, and Rodriguez (2017). An international extension explains the fact that (4) either sorted by short rate or negative term premia, the carry trade using short term bonds is profitable while the carry trade with long term bonds is not, as found by Lustig, Stathopoulos, and Verdelhan (2017). The model suggests that the following new trading strategy should have high return: sorting countries by a linear combination of forward rates that measures the country’s exposure to the jump in inflation induced by the disaster, and trading a synthetic slope that longs/shorts two bonds with different maturity which is directly exposed to this jump. The Sharpe ratio is 1.34, consistent with the notion that asset highly exposed to disaster risk earns high return.