Robert Savickas

Event-Induced Volatility and Tests for Abnormal Performance

I analyze a simple test statistic for mean abnormal returns in the presence of stochastic volatility during both event and nonevent windows and in the presence of event-induced variance increases. Unlike previous tests, the parametric test evaluated here does not require that the volatility effect of the event be the same across all securities. Simulations show that the test exhibits nontrivial gains in power over previously developed parametric and nonparametric tests, and the true null hypothesis is rejected at appropriate levels.

Is the Value Premium a Proxy for Time-Varying Investment Opportunities? Some Time-Series Evidence

We uncover a positive stock market risk-return tradeoff after controlling for the covariance of market returns with the value premium. Fama and French (1996) conjecture that the value premium proxies for investment opportunities; therefore, by ignoring it, early specifications suffer from an omitted variable problem that causes a downward bias in the risk-return tradeoff estimation. We also document a positive relation between the value premium and its conditional variance, and the estimated conditional value premium is strongly countercyclical. The latter evidence supports the view that value is riskier than growth in bad times, when the price of risk is high.

Foreign Exchange Rates Don't Follow a Random Walk

We find that a relatively high level of average U.S. industry- or firm-level idiosyncratic stock volatility is usually associated with a future appreciation in the U.S. dollar. This relation is highly significant for most foreign currencies in both in-sample and out-of-sample forecasting tests. We also document a positive and significant relation between a countrys idiosyncratic volatility and the future U.S. dollar price of its currency - in France, Germany, and Japan. Given that idiosyncratic volatility forecasts GDP growth in both U.S. and foreign countries, our results appear to be consistent with monetary models of exchange rates.

Cross-Sectional Estimation Biases in Risk Premia and Zero-Beta Excess Returns

This paper shows that the classic cross-sectional asset pricing tests tend to suffer from severe risk-premium estimation errors because of small variation in betas. We explain how the conventional approach uses low criteria to validate an asset-pricing model and suffers from the model-misspecification issue because of the complication associated with the zero-beta excess return. We show that the resulting biases in estimates of risk premia and their standard errors are severe enough to lead researchers into inferring incorrect implications about some asset-pricing theories being tested. Further, we suggest that one simple method of mitigating these issues is to restrict the zero-beta excess returns to their theoretical values in the crosssectional regressions and to conduct the straightforward test of whether the estimated ex-ante risk premia are consistent with the observed ex-post ones. The empirical testing results not only further affirm the higher efficiency of the estimates produces by the suggested method, but also show, contrary to some prior evidence, that the market factor is priced consistently.

Asset Pricing Anomalies: Two Hedge Factors with Negative Risk Premia Embedded in Portfolios!

The primary purpose of this research is to empirically test a new asset pricing model, the Relative Asset Pricing Model (RAPM), and to confirm whether hedge portfolios on two new risk factors highlighted in that model, and embedded in all portfolios, have negative and significant risk premia. In a number of specifications, this first test of RAPM appears to validate the importance of these two ubiquitous risk factors: (i) a liability proxy or Strategic Asset Allocation (SAA); and (ii) an index of traditional rebalancing back to the SAA (REBAL). A secondary goal of the research is to then examine what RAPM and the inclusion of these two risk factors mean for other commonly used factors (like Value Premium, Capitalization Premium, and Momentum Premium) that are increasingly being added to portfolios in a new trend called “factor-based investing”. In many regressions, adding these two new factors either individually or together causes traditional factors like Capitalization and – in fewer cases – Momentum to lose their significance. Valuation continues to be significant even in tests where the SAA is heavily biased to include Value indices. These findings have interesting implications for improving portfolio performance especially in a low-yielding environment. Investors would do well to consider an intelligent approach to managing portfolio drift. With respect to factor-based investing, two results stand out: (i) Valuation seems to be a very robust strategy; and (ii) early movers probably benefit before these strategies get incorporated into the industry-wide SAA and then lose significance – a seemingly obvious result but validated in the context of a robust asset pricing model, thereby also serving as a first test of (a special case of) the Adaptive Markets Hypothesis.

Return Decomposition: A Bayesian State-Space Model Approach

We test the implications of the return decomposition of Campbell (1991), in which the unexpected market return is decomposed into cash-flow and discount-rate news. Unlike most of the previous literature, which uses VAR models to implement the return decomposition, we propose a state-space model with parameter restrictions, which is a more systematic and direct than the VAR approach on two aspects. First, the state-space model approach simultaneously models returns, cash-flow news, and discount-rate news, whereas the VAR approach models discount-rate news and uses the return decomposition to back out cash flow news. Second, the state-space model allows us to use information directly related to market and portfolio returns to study the return decomposition, whereas the VAR approach utilizes different predictive variables. We find that discount-rate news has larger variance than does cash-flow news and that the two types of news are nearly perfectly correlated, which can be economically intuitive. We also find that small stocks have higher cash-flow betas than do large stocks, but do not find that value stocks have higher ashflow betas than do growth stocks.

House Price Variation and the Convenience Yield to Owning One’s Home

Existing models of mortgage default cannot explain variation in mortgage performance across origination years because they do not account for the “convenience yield” households derive from owning, rather than renting, a home. The convenience yield is the flow of services from owning rather than renting an otherwise identical residence. This includes the ability to enter into a longer-term housing services contract, customize the residence, and pledge the house as collateral for external finance to increase current period non-housing consumption. We use house price and owner’s equivalent rent data to estimate the stochastic convenience yield of Gibson and Schwartz (1990) and Schwartz (1997) parameterized through the Kalman filter. We find that the national convenience yield is closely related to the rate of home equity withdrawal, while cross-sectional variation in the average convenience yield across metropolitan areas is explained by population churn rates. Both findings are consistent with the theory of an endogenous convenience yield to home ownership.

Market Sentiment and Dividend Policy: Cross Sectional Evidence

Existing empirical findings on the validity of catering theory of dividend policy are mixed at the aggregate-market level, while there is little cross-sectional evidence based on individual firms, which is necessary for studying investor preferences for dividend policy with greater resolution. To fill this gap in the literature, we provide cross-sectional evidence of the explanatory ability of market sentiment on dividend policy decisions. Sentiment sensitivity of individual stock returns (SS), measured by the slope coefficient in the regression of stock returns on market sentiment index changes, is an important cross-sectional determinant on dividend policy. Our empirical results suggest that: firms with high SS pay lower dividends or have lower propensity to pay than do otherwise identical firms; cumulative abnormal stock returns associated with dividend initiations (omissions) are significantly related to the SS. These results remain quantitatively the same after controlling for firm characteristics and other risk factors. Our paper highlights the importance of investor behavior on dividend decisions.

Distribution Effect and Idiosyncratic Volatility Discount

Based on information signaling theory, this paper empirically shows that distribution effect matters for idiosyncratic volatility discount. Idiosyncratic volatility discount arises primarily among stocks with non-cash distributions. Such stocks show a significant idiosyncratic volatility discount, measured by the return difference between high and low idiosyncratic volatility portfolios. For stocks with cash distributions, this discount is statistically insignificant. We attribute these effects to the fact that the information content of cash distributions reduces the uncertainty of future capital gains, which in turn drives out the idiosyncratic volatility discount. We test several explanations of idiosyncratic volatility discount under the framework of distribution effect and find evidence contrary to those explanations. In addition, our empirical results suggest that extreme values may obscure some of the extant explanations on idiosyncratic volatility discount.