The pricing of financial distress or default risk is one of the fundamental questions in financial economics. In a recent study, Vassalou and Xing (2004) measure default risk using a default likelihood indicator (DLI) computed according to the Black Scholes (1973) and Merton (1974) option pricing frame work and show that stocks more likely to default earn a higher return than otherwise similar stocks.
Their finding represents a puzzle for the literature on financial distress or default risk, as most recent research documents the opposite relation (see Dichev (1998), Griffi n and Lemmon (2002), Garlappi, Shu, and Yan (2007), Campbell, Hilscher, and Szilagyi (2007) and George and Hwang (2007)). We resolve this puzzle by relating Vassalou and Xingms finding to the short term return reversal first documented by Jegadeesh (1990) and Lehman (1990). In addition, we analyze a concrete channel through which a liquidity shock might occur on a stock which in turn causes the return reversal.
The default risk premium documented by Vassalou and Xing (2004) appears rather high: the stocks in the highest default risk decile earn about 90 basis points (bps) more per month than otherwise similar stocks, with a monthly Sharpe ratio of around 0.25 between 1970 and 1999. This high default risk premium represents another puzzle, as Hansen and Jagannathan (1991) point out, the associated high Sharpe ratio cannot be easily explained in perfect and complete markets. For comparison, during the same period, the monthly return on the Fama French HML factor (the return difference between high and low book to market stocks) is only 35 bps with a monthly Sharpe ratio of 0.13. In addition, this high default risk premium cannot be fully explained by the standard Fama French (1993) three factor model and a separate aggregate default risk factor seems to be needed.
To reconcile these two puzzling findings by Vassalou and Xing with the literature, our investigation first reveals that stocks in the highest DLI decile earn abnormal returns only in the first month after portfolio formation. The return on these stocks immediately declines by more than a quarter, from 2.10% in the first month to 1.52% in the second month, and stabilizes afterward. If we skip a month and use the second month returns in various asset pricing tests, we find that the returns of high default risk stock can be fully explained by the Fama French three factor model, and the additional default risk factor is no longer needed. We also verify that characteristics such as size, book to market ratio, default likelihood and loadings on risk factors barely change from the first to the second month after portfolio formation. Second, we show that abnormal returns on the highest DLI decile are confined to a small subset of stocks with similar DLIs that recently experienced large negative returns and sharp increases in their DLI measure (the high DLI losers). Thus the abnormal return on high default risk stocks documented in Vassalou and Xing (2004) is temporary, and clearly does not represent compensation for bearing systematic default risk.
Empirically, high default risk stocks are recent losers on average during the portfolio formation month, so their abnormal return in the subsequent month constitutes a short term return reversal, a robust empirical regularity first uncovered by Jegadeesh (1990) and Lehman (1990). In a cross sectional regression framework, we confirm that the past one month return drives out DLI in predicting the next month stock return.
Download
PDF Ebook Clientele Change, Liquidity Shock, and the Return on Financially Distressed Stocks
