A MODIFIED METHOD FOR INFERRING THE EFFECTIVE BID-ASK SPREAD FROM SECURITY RETURNS

This paper presents a modified method for inferring the effective bid-ask spread from security returns in an eMicient market. The Modified Method removes from security returns the systematic effect of market movements and makes use of the equivalence properties of the moving average process and serial covariance function. The Modified Method is tested with the CRSP daily, weekly, and monthly returns data. The results show that the spread estimates are non-negative and sample time-interval independent. The results are compared with those of Roll (1984) and Amihud and Mendelson (1986).     

Nonlinear mean reversion in stock prices

This paper provides new evidence on the time-series predictability of stock market returns by introducing a test of nonlinear mean reversion. The performance of extreme daily returns is evaluated in terms of their power to predict short- and long-horizon returns on various stock market indices and size portfolios. The paper shows that the speed of mean reversion is significantly higher during the large falls of the market. The parameter estimates indicate a negative and significant relation between the monthly portfolio returns and the extreme daily returns observed over the past one to eight months. Specifically, in a quarter in which the minimum daily return is −2% the expected excess return is 37 basis points higher than in a month in which the minimum return is only −1%. This result holds for the value-weighted and equal-weighted stock market indices and for each of the size decile portfolios. The findings are also robust to different sample periods, different indices, and investment horizons.     

Testing multi-beta asset pricing models

This paper presents a complete solution to the estimation and testing of multi-beta models by providing a small sample likelihood ratio test when the usual normality assumption is imposed and an almost analytical GMM test when the normality assumption is relaxed. Using 10 size portfolios from January 1926 to December 1994, we reject the joint efficiency of the CRSP value-weighted and equal-weighted indices. We also apply the tests to analyze a new version of Fama and French's [Fama, E.F., French, K.R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics. 33, 3–56.] three-factor model in addition to two standard ones, and find that the new version performs the best.     

Equal-weighting and value-weighting: which one is better?

Review of Quantitative Finance and Accounting - Prior research shows that noisy prices can introduce biases in returns causing equal-weighted portfolios to outperform value-weighted portfolios. In...     

Structural Breaks and Predictive Regression Models of Aggregate U.S. Stock Returns

In this article we examine the structural stability of predictiveregression models of U.S. quarterly aggregate real stock returnsover the postwar era. We consider predictive regressions modelsof S&P 500 and CRSP equal-weighted real stock returns basedon eight financial variables that display predictive abilityin the extant literature. We test for structural stability usingthe popular Andrews SupF statistic and the Bai subsample procedurein conjunction with the Hansen heteroskedastic fixed-regressorbootstrap. We also test for structural stability using the recentlydeveloped methodologies of Elliott and Müller, and Baiand Perron. We find strong evidence of structural breaks infive of eight bivariate predictive regression models of S&P500 returns and some evidence of structural breaks in the threeother models. There is less evidence of structural instabilityin bivariate predictive regression models of CRSP equal-weightedreturns, with four of eight models displaying some evidenceof structural breaks. We also obtain evidence of structuralinstability in a multivariate predictive regression model ofS&P 500 returns. When we estimate the predictive regressionmodels over the different regimes defined by structural breaks,we find that the predictive ability of financial variables canvary markedly over time.     

A First Look at the Accuracy of the CRSP Mutual Fund Database and a Comparison of the CRSP and Morningstar Mutual Fund Databases

This paper examines problems in the CRSP Survivor Bias Free U.S. Mutual Fund Database (CRSP, 1998) and compares returns contained in it to those in Morningstar. The CRSP database has an omission bias that has the same effects as survivorship bias. Although all mutual funds are listed in CRSP, return data is missing for many and the characteristics of these funds differ from the populations. The CRSP return data is biased upward and merger months are inaccurately recorded about half the time. Differences in returns in Morningstar and CRSP are a problem for older data and small funds.     

Does the Stock Market Overreact?

Research in experimental psychology suggests that, in violation of Bayes' rule, most people tend to “overreact” to unexpected and dramatic news events. This study of market efficiency investigates whether such behavior affects stock prices. The empirical evidence, based on CRSP monthly return data, is consistent with the overreaction hypothesis. Substantial weak form market inefficiencies are discovered. The results also shed new light on the January returns earned by prior “winners” and “losers.” Portfolios of losers experience exceptionally large January returns as late as five years after portfolio formation.     

Liquidity biases in asset pricing tests

Microstructure noise in security prices biases the results of empirical asset pricing specifications, particularly when security-level explanatory variables are cross-sectionally correlated with the amount of noise. We focus on tests of whether measures of illiquidity, which are likely to be correlated with the noise, are priced in the cross-section of stock returns, and show a significant upward bias in estimated return premiums for an array of illiquidity measures in Center for Research in Security Prices (CRSP) monthly return data. The upward bias is larger when illiquid securities are included in the sample, but persists even for NYSE/Amex stocks after decimalization. We introduce a methodological correction to eliminate the biases that simply involves weighted least squares (WLS) rather than ordinary least squares (OLS) estimation, and find evidence of smaller, but still significant, return premiums for illiquidity after implementing the correction.     

Estimation of expected return: CAPM vs. Fama and French

Most practitioners favour a one-factor model (CAPM) when estimating expected return for an individual stock. For estimation of portfolio returns, academics recommend the Fama and French three-factor model. The main objective of this paper is to compare the performance of these two models for individual stocks. First, estimates for individual stock returns based on CAPM are obtained using different time frames, data frequencies, and indexes. It is found that 5 years of monthly data and an equal-weighted index, as opposed to the commonly recommended value-weighted index, provide the best estimate. However, performance of the model is very poor; it explains on average 3% of differences in returns. Then, estimates for individual stock returns are obtained based on the Fama and French model using 5 years of monthly data. This model, however, does not do much better; independent of the index used, it explains on average 5% of differences in returns. These results therefore bring into question the use of either model for estimation of individual expected stock returns.     

The E-V stationarity of secure returns: Some empirical evidence

The joint hypothesis that the expected return vector and the variance-covariance matrix of returns are intertemporally stationary is rejected for the monthly return data for securities included on the CRSP file. This finding is robust in that it is invariant to the type of sampling procedure used and to whether nominal or real security returns are used. In contrast, the hypothesis that the correlation matrix of returns is intertemporally stationary could not be rejected for the same samples of securities.     

Testing for persistence in stock returns with GARCH-stable shocks

We investigate persistence in CRSP monthly excess stock returns, using a state space model with stable disturbances. The non-Gaussian state space model with volatility persistence is estimated by maximum likelihood, using the optimal filtering algorithm given by Sorenson and Alspach (1971 Automatica 7 465–79). The conditional distribution has a stable α of 1.89, and normality is strongly rejected even after accounting for GARCH. However, stock returns do not contain a significant mean-reverting component. The optimal predictor is the unconditional expectation of the series, which we estimate to be 9.8% per annum.     

Return reversals in the bond market: Evidence and causes

The finance literature has shown that equity returns are predictable using past returns. This study extends that literature by examining bond return predictability. Using returns constructed from dealer bid prices, we find short- to intermediate-term reversals in investment grade corporate bond returns. These reversals are larger in the first half of the sample period and consistent with the predictions of dealer inventory cost models. This supports Jegadeesh and Titman’s [J. Financ. Intermed. 4 (1995) 116] assertion that daily, weekly, and monthly reversals in equity returns come from dealer inventory considerations, not behavioral biases. Finally, unlike equity returns, we find no evidence of momentum in bond returns.     

Is Stock Market Overreaction Persistent Over Time?

This paper examines the stability and persistence of the market overreaction hypothesis as posited by DeBondt and Thaler (1985 and 1987), and reinforced by Chopra, Lakonishok, and Ritter (1992). Using monthly CRSP data for the period 1926 through 1992, we find that returns obtained from a contrarian investment strategy are not time-stationary. Specifically, there is no winner-loser portfolio relationship during the post-war period of 1940_50s. The relationship resumes during the pre-energy-crisis subperiod, but weakens again during the post-energy-crisis subperiod. The effectiveness of trading based upon the overreaction hypothesis is, therefore, suspect.     

Seasonal variations in the U.S. Stock Market returns: 1927–1984

When the seasonal components of the monthly returns as opposed to the returns themselves, are examined over the 1927–1984 period, the Standard & Poor's 500 Composite Index (S&P 500) and the Center for Research in Security Prices (CRSP) value-weighted portfolio exhibit significant seasonality. Their seasonal behavior is quite similar to that of the smallest quintile of New York Stock Exchange (NYSE) stocks and the CRSP equally weighted portfolio during March through October. While January is strong for the two latter portfolios, December, November, and January appear to be consistently strong for the two former portfolios. The seasonal pattern has, however, changed substantially over time. While June and July returns experienced a significant drop in seasonal strength, March and April returns gained seasonal strength for all four portfolios from 1927–1958 to 1959–1984. These changes coincide in an inverse fashion with the shifts in interest rate seasonality.     

Predictable Stock Returns: The Role of Small Sample Bias

Predictive regressions are subject to two small sample biases: the coefficient estimate is biased if the predictor is endogenous, and asymptotic standard errors in the case of overlapping periods are biased downward. Both biases work in the direction of making t-ratios too large so that standard inference may indicate predictability even if none is present. Using annual returns since 1872 and monthly returns since 1927 we estimate empirical distributions by randomizing residuals in the VAR representation of the variables. The estimated biases are large enough to affect inference in practice, and should be accounted for when studying predictability.     

Yet another look at mutual fund tournaments

We revisit the empirical evidence on the tournament hypothesis for the behavior of mutual fund managers provided by Busse [J. Financ. Quant. Anal. 36 (2001) 53]. First, we give analytical expressions for the biases arising in volatility estimates (based on both daily and monthly data) due to first-order autocorrelation effects in daily fund returns. These calculations show that tests of the tournament hypothesis based on monthly data are more robust to autocorrelation effects than tests based on daily data. Second, to address the impact of cross-correlated fund returns on these tests, we provide explicit conditions under which the tests proposed in the literature have appropriate size properties.     

Evidence of Predictable Behavior of Security Returns

This paper presents new empirical evidence of predictability of individual stock returns. The negative first-order serial correlation in monthly stock returns is highly significant. Furthermore, significant positive serial correlation is found at longer lags, and the twelve-month serial correlation is particularly strong. Using the observed systematic behavior of stock returns, one-step-ahead return forecasts are made and ten portfolios are formed from the forecasts. The difference between the abnormal returns on the extreme decile portfolios over the period 1934–1987 is 2.49 percent per month.     

Are the gains from international portfolio diversification exaggerated? The influence of downside risk in bear markets

The fundamental rationale for international portfolio diversification is that it expands the opportunities for gains from portfolio diversification beyond those that are available through domestic securities. However, if international stock market correlations are higher than normal in bear markets, then international diversification will fail to yield the promised gains just when they are needed most. We evaluate the extent to which observed correlations to monthly returns in bear, calm and bull markets are captured by three popular bivariate distributions: (1) the normal, (2) the restricted GARCH(1,1) of J. P. Morgan’s RiskMetrics, and (3) the Student-t with four degrees of freedom. Observed correlations during calm and bull markets are unexceptional compared to these models. In contrast, observed correlations during bear markets are significantly higher than predicted. Higher-than-normal correlations during extreme market downturns result in monthly returns to equal-weighted portfolios of domestic and international stocks that are, on average, more than two percent lower than those predicted by the normal distribution. If the extent of non-normality during bear markets persists over time, then a US investor allocating assets into foreign markets might want to allocate more assets into foreign markets with near-normal correlation profiles and avoid markets with higher-than-normal bear market co-movements.     

An empirical analysis of intertemporal asset pricing models with transaction costs and habit persistence

In intertemporal asset pricing models, transaction costs are usually neglected. In this paper we explicitly incorporate transaction costs in these models and analyze to what extent this extension is helpful in explaining the cross-section of expected returns. An empirical analysis using CRSP data on size-based portfolios examines the role of the transaction costs and shows that incorporating such costs in the consumption-based model with power utility does not yield very satisfactory results. However, the introduction of habit persistence substantially improves the model. We find rather strong evidence of habit persistence in monthly consumption data. The plots of the models' pricing errors indicate that an intertemporal asset pricing model with transaction costs and habit persistence explains the cross-sectional variation in the portfolio returns quite accurately.     

Are investors moonstruck? Lunar phases and stock returns

This paper investigates the relation between lunar phases and stock market returns of 48 countries. The findings indicate that stock returns are lower on the days around a full moon than on the days around a new moon. The magnitude of the return difference is 3% to 5% per annum based on analyses of two global portfolios: one equal-weighted and the other value-weighted. The return difference is not due to changes in stock market volatility or trading volumes. The data show that the lunar effect is not explained away by announcements of macroeconomic indicators, nor is it driven by major global shocks. Moreover, the lunar effect is independent of other calendar-related anomalies such as the January effect, the day-of-week effect, the calendar month effect, and the holiday effect (including lunar holidays).