Nishad Kapadia

Slopes as Factors: Characteristic Pure Plays

Returns to pure play strategies, estimated as Fama-MacBeth slope coefficients on standardized size, value and momentum characteristics, have positive and significant four factor alphas. The mispricing of these characteristics-based strategies by the four factor model is due in part, but not entirely, to (1) the effect of microcap stocks on the pure play returns and (2) the effect of stale book and market capitalization information on the SMB and HML factors. Adjusting for these issues, the value and momentum pure play strategies still have positive and significant four factor alphas. We examine thirteen reported anomalies and find that five have insignificant alphas when the pure play returns are used as factors. Eight are insignificant when an interaction between size and value characteristics is included as a factor.

What Explains the Distress Risk Puzzle: Death or Glory?

Campbell, Hilscher, and Szilagyi (2008) show that firms with a high probability of default have significantly low average future returns. We show that there is a large overlap between stocks classified as high default risk, and those that are likely to produce extremely high returns over the next year (‘glory’ stocks). Predicted glory and predicted distress are highly correlated, with over 50% of firms in the top distress risk quintile also in the top quintile of predicted glory. Stocks with high predicted probabilities for glory also earn abnormally low average returns. We find evidence that low returns to high glory firms are also present in firms with zero leverage, where financial distress is unlikely, and that the low returns to high distress risk firms are large and significant in ‘speculative’ firms (with high sales growth and market-to-book ratios) that have high predicted glory; subsequent returns are small and statistically insignificant in ‘traditionally distressed’ firms. Thus, we show that, on average, firms which have a high potential for death (default) also tend to have a high potential for glory; where the two factors can be separated, the results suggest that it is glory, rather than distress, which is responsible for the low expected returns in securities.

Davids, Goliaths, and Business Cycles

We show that a simple, intuitive variable, GVD (Goliath versus David) reflects time-variation in discount rates related to changes in aggregate business conditions. GVD is the annual change in the weight of the largest 250 firms in the aggregate stock market, and is motivated by research that shows that small firms are more severely impacted than large firms by economic shocks due to differences in access to external finance. We find that GVD is the best single predictor of market returns out-of-sample among traditional predictors, predicting quarterly market returns with an out-of-sample R2 of 6.3% in the 1976--2011 evaluation period.

Are Credit Ratings Redundant When Market Prices Reflect Credit Risk

We find that Credit Rating Agencies (CRAs) see through transitory shocks to credit risk that stem from transitory shocks to equity prices, while market-based measures of credit risk do not. For a given stock return, CRAs are significantly less likely to downgrade firms with transitory shocks than those with permanent shocks. However, credit default swap spreads and model-implied default probabilities do not distinguish between such shocks. These results explain why ratings are useful despite the availability of market-based estimates of credit risk: the ability to ignore transitory shocks is valuable because rating changes have real consequences for private contracts and access to capital.

Do Idiosyncratic Jumps Matter

We show that idiosyncratic jumps are a key determinant of mean stock returns from both an ex post and ex ante perspective. Ex post, the entire annual average return of a typical stock accrues on the four days on which its price jumps. Ex ante, idiosyncratic jump risk earns a premium: a value-weighted weekly long-short portfolio that buys (sells) stocks with high (low) predicted jump probabilities earns annualized mean returns of 9.4% and four-factor alphas of 8.1%. This strategy’s returns are larger when there are greater limits to arbitrage. These results are consistent with investor aversion to idiosyncratic jump risk.

Testing Factor Models on Characteristic and Covariance Pure Plays

We test the recent Fama-French five-factor model and Hou-Xue-Zhang four-factor model using test assets from Fama-MacBeth regressions, which are pure plays on particular characteristics or covariances. Our tests resolve the errors-in-variable bias in Fama-MacBeth regressions with estimated betas. Monte Carlo evidence shows that the tests are unbiased even with time-varying stock betas and characteristics. For both factor models, characteristic pure plays generally have positive alphas, and covariance pure plays have negative alphas. The models fail especially in explaining returns to investment and when pure plays are momentum-neutral. The rejections are economically significant.

Safe Minus Risky: Do Investors Pay a Premium for Stocks that Hedge Stock Market Downturns?

Stocks that hedge against sustained market downturns — periods from peak to trough in S&P500 levels at the business cycle frequency — should have low expected returns, but they do not. We use ex-ante firm characteristics and covariances to construct a tradeable Safe Minus Risky (SMR) portfolio that hedges market downturns out-of-sample. Although downturns correspond to significant declines in GDP growth, SMR has significant positive average returns and four factor alphas (both about 0.75% per month). Risk-based models do not explain SMR’s returns, but mispricing does. Risky stocks are overpriced when sentiment is high, resulting in subsequent returns of -1% per month.

Estimating the Cost of Equity: Why Do Simple Benchmarks Outperform Factor Models?

We compare the accuracy of cost of equity estimates based on leading factor models to two simple alternatives: the asset mean and the market mean. The market mean proves to be a serious competitor to traditional implementations of factor models even if the underlying factor model is true. Pricing errors (alphas) that are negatively correlated with firm and industry market betas further improve the performance of the market mean relative to model-based approaches. We propose Bayesian estimators that address key weaknesses in standard plug-in approaches. These estimators outperform plug-in methods and the market mean in simulations and out-of-sample.