THE JOURNAL OF FINANCE * VOL. XL, NO. 3 * JULY 1985 Does the Stock Market Overreact? WERNER F. M. De BONDT and RICHARD THALER* ABSTRACT Research in experimentalpsychology suggests that, in violation of Bayes' rule, most people tend to "overreact"to unexpected and dramatic news events. This study of marketefficiencyinvestigateswhethersuch behavioraffects stock prices.The empirical evidence, based on CRSP monthly return data, is consistent with the overreaction hypothesis.Substantialweak formmarketinefficienciesare discovered.The resultsalso shed new light on the Januaryreturnsearnedby prior"winners"and "losers."Portfolios of losers experience exceptionally large January returns as late as five years after portfolioformation. As ECONOMISTS INTERESTED IN both market behavior and the psychology of individual decision making, we have been struck by the similarity of two sets of empirical findings. Both classes of behavior can be characterizedas displaying overreaction. This study was undertakento investigate the possibility that these phenomena are related by more than just appearance.We begin by describing briefly the individual and market behavior that piqued our interest. The term overreaction carries with it an implicit comparison to some degree of reaction that is consideredto be appropriate.What is an appropriatereaction? One class,,of tasks which have a well-established norm are probability revision problemsfor which Bayes' rule prescribesthe correctreactionto new information. It has now been well-established that Bayes' rule is not an apt characterization of how individuals actually respond to new data (Kahneman et al. [14]). In revising their beliefs, individuals tend to overweight recent information and underweightprior (or base rate) data. People seem to make predictions according to a simple matching rule: "The predicted value is selected so that the standing of the case in the distributionof outcomes matches its standing in the distribution of impressions" (Kahneman and Tversky [14, p. 416]). This rule-of-thumb,an instance of what Kahneman and Tversky call the representativeness heuristic, violates the basic statistical principal that the extremeness of predictions must be moderatedby considerationsof predictability.Grether [12] has replicatedthis finding under incentive compatible conditions. There is also considerable evidence that the actual expectations of professionalsecurity analysts and economic forecastersdisplay the same overreactionbias (for a review, see De Bondt [7]). One of the earliest observations about overreactionin markets was made by J. M. Keynes:"... day-to-day fluctuations in the profits of existing investments, * Universityof Wisconsin at Madisonand CornellUniversity,respectively.The financial support of the C.I.M. DoctoralFellowshipProgram(Brussels, Belgium) and the Cornell GraduateSchool of Managementis gratefullyacknowledged.We receivedhelpful comments from SeymourSmidt, Dale Morse,Peter Bernstein, Fischer Black, RobertJarrow,Edwin Elton, and Ross Watts. 793 794 The Journal of Finance which are obviously of an ephemeral and nonsignificant character,tend to have an altogether excessive, and even an absurd, influence on the market" [17, pp. 153-154]. About the same time, Williams noted in this Theory of Investment Valuethat "priceshave been based too much on current earning power and too little on long-term dividend paying power" [28, p. 19]. More recently, Arrowhas concludedthat the work of Kahneman and Tversky "typifies very precisely the exessive reaction to current information which seems to characterize all the securities and futures markets" [1, p. 5]. Two specific examples of the research to which Arrowwas referringare the excess volatility of security prices and the so-called price earnings ratio anomaly. The excess volatility issue has been investigated most thoroughly by Shiller [27]. Shiller interprets the Miller-Modiglianiview of stock prices as a constraint on the likelihood function of a price-dividendsample. Shiller concludes that, at least over the last century, dividends simply do not vary enough to rationally justify observedaggregateprice movements.Combiningthe results with Kleidon's [18] findings that stock price movements are strongly correlatedwith the following year's earnings changes suggests a clear pattern of overreaction.In spite of the observedtrendiness of dividends, investors seem to attach disproportionate importanceto short-run economic developments.1 The price earnings ratio (PIE) anomaly refers to the observation that stocks with extremelylow PIE ratios (i.e., lowest decile) earn largerrisk-adjustedreturns than high PIE stocks (Basu [3]). Most financial economists seem to regardthe anomaly as a statistical artifact. Explanations are usually based on alleged misspecificationof the capital asset pricing model (CAPM). Ball [2] emphasizes the effects of omitted risk factors. The PIE ratio is presumed to be a proxy for some omitted factor which, if included in the "correct"equilibrium valuation model, would eliminate the anomaly. Of course, unless these omitted factors can be identified, the hypothesis is untestable. Reinganum [21] has claimed that the small firm effect subsumes the PIE effect and that both are related to the same set of missing (and againunknown) factors. However,Basu [4] found a significant PIE effect after controlling for firm size, and earlier Graham [11] even found an effect within the thirty Dow Jones Industrials,hardly a group of small firms! An alternative behavioral explanation for the anomaly based on investor overreactionis what Basu called the "price-ratio"hypothesis (e.g., Dreman [8]). Companies with very low PIE's are thought to be temporarily "undervalued" because investors become excessively pessimistic after a series of bad earnings reports or other bad news. Once future earnings turn out to be better than the unreasonablygloomy forecasts, the price adjusts. Similarly, the equity of companies with very high PIE's is thought to be "overvalued,"before (predictably) falling in price. While the overreactionhypothesis has considerablea priori appeal,the obvious question to ask is: How does the anomaly survive the process of arbitrage?There ' Of course,the variabilityof stock prices may also reflect changes in real interest rates. If so, the price movementsof other assets-such as land or housing-should match those of stocks. However, this is not actually observed. A third hypothesis, advocated by Marsh and Merton [19], is that Shiller'sfindingsare a result of his misspecificationof the dividendprocess. Does the Stock Market Overreact? 795 is really a more general question here. What are the equilibria conditions for marketsin which some agents are not rational in the sense that they fail to revise their expectations accordingto Bayes' rule? Russell and Thaler [24] addressthis issue. They conclude that the existence of some rational agents is not sufficient to guarantee a rational expectations equilibrium in an economy with some of what they call quasi-rationalagents. (The related question of market equilibria with agents having heterogeneous expectations is investigated by Jarrow [13].) While we are highly sensitive to these issues, we do not have the space to address them here. Instead, we will concentrate on an empiricaltest of the overreaction hypothesis. If stock prices systematically overshoot, then their reversal should be predictable from past return data alone, with no use of any accounting data such as earnings. Specifically, two hypotheses are suggested:(1) Extreme movements in stock prices will be followed by subsequent price movements in the opposite direction. (2) The more extreme the initial price movement, the greater will be the subsequent adjustment. Both hypotheses imply a violation of weak-form market efficiency. To repeat, our goal is to test whether the overreactionhypothesis is predictive. In other words, whether it does more for us than merely to explain, ex post, the P/E effect or Shiller's results on asset price dispersion. The overreactioneffect deserves attention because it represents a behavioralprinciple that may apply in many other contexts. For example, investor overreactionpossibly explains Shiller's earlier [26] findings that when long-term interest rates are high relative to short rates, they tend to move down later on. Ohlson and Penman [20] have further suggestedthat the increasedvolatility of security returns following stock splits may also be linked to overreaction.The present empiricaltests are to our knowledgethe first attempt to use a behavioralprinciple to predict a new market anomaly. The remainderof the paper is organizedas follows. The next section describes the actual empirical tests we have performed. Section II describes the results. Consistent with the overreaction hypothesis, evidence of weak-form market inefficiency is found. We discuss the implications for other empirical work on asset pricing anomalies. The paper ends with a brief summaryof conclusions. I. The Overreaction Hypothesis: Empirical Tests The empirical testing proceduresare a variant on a design originally proposed by Beaver and Landsman [5] in a different context. Typically, tests of semistrong form market efficiency start, at time t = 0, with the formation of portfolios on the basis of some event that affects all stocks in the portfolio, say, an earnings announcement. One then goes on to investigate whether later on (t > 0) the estimated residual portfolio return ut' -measured relative to the single-period CAPM-equals zero. Statistically significant departuresfrom zero are interpreted as evidence consistent with semistrong form market inefficiency, even though the results may also be due to misspecification of the CAPM, misestimation of the relevant alphas and/or betas, or simply market inefficiency of the weak form. 796 The Journal of Finance In contrast, the tests in this study assess the extent to which systematic nonzero residual return behavior in the period after portfolio formation (t > 0) is associated with systematic residual returns in the preformationmonths (t < 0). We will focus on stocks that have experiencedeither extreme capital gains or extreme losses over periods up to five years. In other words, "winner"(W) and "loser" portfolios (L) are formed conditional upon past excess returns, rather than some firm-generatedinformationalvariable such as earnings. Following Fama [9], the previous argumentscan be formalizedby writing the efficient market'scondition, 1IF-) (Rjt-Em(Rjt I Fm 1)I Ft-,) = E(li2t = 0 where Ft-1 represents the complete set of information at time t - 1, Rjt is the return on securityj at t, and Em(RjtIF m1) is the expectation of Rjt, assessed by the market on the basis of the information set Fml1.The efficient market hypothesis implies that E(twt IFt-,) = E(i2Lt IFt-) = 0. As explained in the introduction, the overreaction hypothesis, on the other hand, suggests that E(twt IFt1) < 0 and E(kLt IFt-1) > 0. In order to estimate the relevant residuals, an equilibrium model must be specified.A commonprocedureis to estimate the parametersof the market model (see e.g., Beaver and Landsman [5]). What will happen if the equilibriummodel is misspecified?As long as the variation in Em(R?tIF m1) is small relative to the movements in uIjt,the exact specification of the equilibriummodel makes little difference to tests of the efficient market hypothesis. For, even if we knew the "correct"model of Em(Rjt IFml1),it would explain only small part of the variation in Pit.2 Since this study investigates the return behavior of specific portfolios over extended periods of time (indeed, as long as a decade), it cannot be merely assumed that model misspecification leaves the conclusions about market efficiency unchanged. Therefore, the empirical analysis is based on three types of return residuals: market-adjustedexcess returns; market model residuals; and excess returns that are measured relative to the Sharpe-Lintnerversion of the CAPM. However, since all three methods are single-index models that follow from the CAPM, misspecification problems may still confound the results. De Bondt [7] formally derives the econometric biases in the estimated marketadjusted and market model residuals if the "true"model is multifactor, e.g., RPt = Aj + BjRmt + CjXt + ejt. As a final precaution, he also characterizes the securities in the extreme portfolios in terms of a number of financial variables. If there were a persistent tendency for the portfolios to differ on dimensions that may proxy for "risk,"then, again, we cannot be sure whetherthe empiricalresults support market efficiency or market overreaction. It turns out that, whicheverof the three types of residualsare used, the results 2 Presumably,this same reasoningunderliesthe commonpracticeof measuringabnormalsecurity price performanceby way of easily calculablemean-adjustedexcess returns [where,by assumption, E(R,) equals a constant Kj], market-adjustedexcess returns (where,by assumption,aj = 0 and f3,= 1 for all j), ratherthan more complicatedmarketmodelresiduals,let along residualsrelativeto some multifactormodel. Does the Stock Market Overreact? 797 of the empiricalanalysis are similar and that the choice does not affect our main conclusions. Therefore,we will only report the results based on market-adjusted excess returns. The residuals are estimated as Ujt = Rjt- Rmt.There is no risk adjustmentexcept for movements of the market as a whole and the adjustment is identical for all stocks. Since, for any period t, the same (constant) market return Rlmtis subtracted from all Rjt's, the results are interpretablein terms of raw (dollar)returns.As shown in-De Bondt [7], the use of market-adjustedexcess returns has the further advantage that it is likely to bias the research design against the overreaction hypothesis.3Finally, De Bondt shows that winner and loser portfolios, formed on the basis of market-adjustedexcess returns, do not systematically differ with respect to either market value of equity, dividendyield or financial leverage. We will now describe the basic research design used to form the winner and loser portfolios and the statistical test proceduresthat determine which of the two competing hypotheses receives more support from the data. A. Test Procedures: Details Monthly return data for New York Stock Exchange (NYSE) common stocks, as compiled by the Center for Research in Security Prices (CRSP) of the University of Chicago, are used for the period between January 1926 and December 1982. An equally weighted arithmetic average rate of return on all CRSP listed securities serves as the market index. 1. For every stockj on the tape with at least 85 months of returndata (months 1 through 85), without any missing values in between, and starting in January 1930 (month 49), the next 72 monthly residualreturns ujt (months 49 through 120) are estimated. If some or all of the raw return data beyond month 85 are missing, the residual returns are calculated up to that point. The procedure is repeated 16 times starting in January 1930, January 1933,..., up to January 1975. As time goes on and new securities appear on the tape, more and more stocks qualify for this step. 2. For every stock j, starting in December 1932 (month 84; the "portfolio formation date") (t = 0), we compute the cumulative excess returns CUj = t_35 ujt for the prior 36 months (the "portfolioformation"period, months 49 through 84). The step is repeated 16 times for all nonoverlappingthreeyear periods between January 1930 and December 1977. On each of the 16 relevant portfolio formation dates (December 1932, December 1935,..., December 1977), the CUj's are ranked from low to high and portfolios are formed. Firms in the top 35 stocks (or the top 50 stocks, or the top decile) are assigned to the winner portfolio W; firms in the bottom 35 stocks (or the bottom 50 stocks, or the bottom decile) to the loser portfolio L. Thus, the portfolios are formed conditional upon excess return behaviorprior to t = 0, the portfolio formation date. 3. For both portfolios in each of 16 nonoverlappingthree-year periods (n = 'We will come back to this bias in Section II. 798 The Journal of Finance 1,..., N; N = 16), starting in January 1933 (month 85, the "starting month") and up to December 1980, we now computethe cumulativeaverage residual returns of all securities in the portfolio, for the next 36 months (the "test period,"months 85 through 120), i.e., from t = 1 through t = 36. We find CARWn,t and CARL,n,t. If a security's return is missing in a month subsequentto portfolio formation,then, from that moment on, the stock is permanently droppedfrom the portfolio and the CAR is an average of the availableresidualreturns.Thus, whenevera stock dropsout, the calculations involve an implicit rebalancing.4 4. Using the CAR's from all 16 test periods, averageCAR's are calculated for both portfolios and each month between t = 1 and t = 36. They are denoted ACARW,tand ACARL,t. The overreaction hypothesis predicts that, for t > 0, ACARW,t < 0 and ACARL,t > 0, so that, by implication, [ACARL,t ACARw,t]> 0. In order to assess whether, at any time t, there is indeed a statistically significant difference in investment performance, we need a pooled estimate of the population variance in CARt, = [n=(CARwnt - ACARw,t)2 + n=](CARL, n, t- ACARL,t)2]/2(N - 1). With two samples of equal size N, the variance of the difference of sample means equals 2S2/N and the t-statistic is therefore Tt = [ACARL, t- ACARw,J]/ 2St/N. Relevant t-statistics can be found for each of the 36 postformation months but they do not represent independent evidence. 5. In order to judge whether, for any month t, the average residual return makes a contribution to either A CARw,t or ACARL,t, we can test whether it is significantly different from zero. The sample standard deviation of the winner portfolio is equal to St = V>N =1(ARw,n,t- ARw,t)2/N - 1. Since St/ v'IN representsthe sample estimate of the standarderrorof AR wt, the t-statistic equals Tt = ARw,t/(st/ViN). Similar proceduresapply for the residuals of the loser portfolio. B. Discussion Several aspects of the research design deserve some further comment. The choice of the data base, the CRSP Monthly Return File, is in part justified by 4Since this study concentrateson companiesthat experienceextraordinaryreturns,either positive or negative, there may be some concern that their attrition rate sufficiently deviates from the "normal"rate so as to cause a survivorshipbias. However,this concernis unjustified.When a security is delisted, suspendedor halted, CRSP determineswhether or not it is possible to trade at the last listed price. If no trade is possible, CRSP tries to find a subsequentquote and uses it to computea returnfor the last period. If no such quote is availablebecause the stockholdersreceive nothing for their shares, the return is entered as minus one. If trading continues, the last return ends with the last listed price. Does the Stock Market Overreact? 799 our concern to avoid certain measurement problems that have received much attention in the literature. Most of the problems arise with the use of daily data, both with respect to the risk and return variables. They include, among others, the "bid-ask"effect and the consequences of infrequenttrading. The requirementthat 85 subsequent returns are available before any firm is allowed in the sample biases the selection towards large, established firms. But, if the effect under study can be shown to apply to them, the results are, if anything, more interesting. In particular,it counters the predictablecritiquethat the overreactioneffect may be mostly a small-firm phenomenon. For the experiment described in Section A, between 347 and 1,089 NYSE stocks participate in the various replications. The decision to study the CAR's for a period of 36 months after the portfolio formation date reflects a compromise between statistical and economic considerations, namely, an adequatenumberof independent replications versus a time period long enough to study issues relevant to asset pricing theory. In addition, the three-yearperiod is also of interest in light of Benjamin Graham'scontention that "the interval required for a substantial underevaluation to correct itself averages approximately 11/2 to 21/2 years" [10, p. 37). However, for selected experiments,the portfolio formation (and testing) periods are one, two, and five years long. Clearly,the numberof independent replicationsvaries inversely with the length of the formationperiod. Finally, the choice of December as the "portfolio formation month" (and, therefore, of January as the "starting month") is essentially arbitrary.In order to check whether the choice affects the results, some of the empirical tests use May as the portfolio formation month. II. The Overreaction Hypothesis: Empirical Results A. Main Findings The results-of the tests developed in Section I are found in Figure 1. They are consistent with the overreaction hypothesis. Over the last half-century, loser portfolios of 35 stocks outperformthe market by, on average, 19.6%,thirty-six months after portfolio formation. Winner portfolios, on the other hand, earn about 5.0% less than the market, so that the difference in cumulative average residual between the extreme portfolios, [ACARL,36- ACARW,36] equals 24.6% (t-statistic: 2.20). Figure 1 shows the movement of the ACAR's as we progress through the test period. The findings have other notable aspects. First, the overreaction effect is asymmetric;it is much larger for losers than for winners. Secondly, consistent with previous work on the turn-of-the-year effect and seasonality, most of the excess returns are realized in January. In months t = 1, t = 13, and t = 25, the loser portfolio earns excess returns of, respectively, 8.1% (t-statistic: 3.21), 5.6% (3.07), and 4.0%(2.76). Finally, in surprisingagreementwith Benjamin Graham's claim, the overreactionphenomenon mostly occurs during the second and third year of the test period. Twelve months into the test period, the difference in performancebetween the extreme portfolios is a mere 5.4% (t-statistic: 0.77). The Journal of Finance 800 Test Periods of 16 Three-Year January 1933 and December 1980 Average Between Three Years Length of Formation Period: 0.20- 0. 15-] Loser Portfolio '.10 C A 0.05-i 0.00 - ~ ~ - ~ - - - -- - - - . - ~ - - ~ -0.05-r Winner Portfol io -o so,-f 0 req p. e e~?~pvi-e s,'u-vq.ej 5 10 r sr 9., I5 v-, i.e-s 20 e s-r r-t 25 30 35 MON4TH1S AFTEn PORTFOLID FOIRATION Figure 1. Cumulative Average Residuals for Winner and Loser Portfolios of 35 Stocks (1-36 months into the test period) While not reported here, the results using market model and Sharpe-Lintner residualsare similar. They are also insensitive to the choice of December as the month of portfolio formation (see De Bondt [7]). The overreaction hypothesis predicts that, as we focus on stocks that go throughmore (or less) extreme return experiences (duringthe formationperiod), the subsequentprice reversals will be more (or less) pronounced.An easy way to generate more (less) extreme observations is to lengthen (shorten) the portfolio formationperiod;alternatively, for any given formation period (say, two years), we may compare the test period performance of less versus more extreme portfolios, e.g., decile portfolios (which contain an average 82 stocks) versus portfolios of 35 stocks. Table I confirms the prediction of the overreaction hypothesis. As the cumulative average residuals (during the formation period) for various sets of winner and loser portfolios grow larger, so do the subsequent price reversals, measured by [ACARL,t - ACARw,,] and the accompanying t-statistics. For a formation period as short as one year, no reversal is observed at all. Table I and Figure 2 further indicate that the overreaction phenomenon is qualitativelydifferent from the January effect and, more generally,from season- 801 Does the Stock Market Overreact? Table I Differencesin CumulativeAverage(Market-Adjusted)ResidualReturnsBetween the Winnerand Loser Portfolios at the End of the FormationPeriod, and 1, 12, 13, 18, 24, 25, 36, and 60 Months into the Test Period Portfolio Selection Procedures:Length of the Formation Period and No. of Independent Replications CAR at the End of the Formation Period Average No. of Winner Loser Stocks Portfolio Portfolio Difference in CAR (t-Statistics) Months After Portfolio Formation 1 12 13 18 24 25 36 60 10 five-year periods 50 1.463 -1.194 0.070 (3.13) 0.156 (2.04) 0.248 0.256 (3.14) (3.17) 0.196 0.228 0.230 0.319 (2.15) (2.40) (2.07) (3.28) 16 three-yearperiods 35 1.375 -1.064 0.105 (3.29) 0.054 (0.77) 0.103 0.167 (1.18) (1.51) 0.181 0.234 0.246 (1.71) (2.19) (2.20) NA* 24 two-year periods" 35 1.130 -0.857 0.062 -0.006 (2.91) (-0.16) 0.074 0.136 (1.53) (2.02) 0.101 (1.41) NA NA NA 25 two-year periods' 35 1.119 -0.866 0.089 (3.98) 0.011 (0.19) 0.092 0.107 (1.48) (1.47) 0.115 (1.55) NA NA NA 24 two-year periods" (deciles) 82 0.875 -0.711 0.051 (3.13) 0.006 (0.19) 0.066 0.105 (1.71) (1.99) 0.083 (1.49) NA NA NA 25 two-year periods' (deciles) 82 0.868 -0.714 0.068 (3.86) 0.008 (0.19) 0.071 0.078 (1.46) (1.41) 0.072 (1.29) NA NA NA 49 one-year periods 35 0.774 -0.585 0.042 -0.076 (2.45) (-2.32) NA NA NA -0.006 (-0.15) 0.007 -0.005 (0.14) (-0.09) The formation month for these portfolios is the month of Decemher in all uneven years hetween 1933 and 1979. 'The formation month for these portfolios is the month of December in all even years between1932 and 1980. * NA, not applicable. ality in stock prices. Throughoutthe test period, the difference in ACAR for the experiment with a three-year formation period (the upper curve) exceeds the same statistic for the experiments based on two- and one-year formationperiods (middle and lower curves). But all three experiments are clearly affected by the same underlyingseasonal pattern. In Section I, it was mentioned that the use of market-adjustedexcess returns is likely to bias the researchdesign against the overreactionhypothesis. The bias can be seen by comparingthe CAPM-betasof the extreme portfolios. For all the experiments listed in Table I, the average betas of the securities in the winner portfolios are significantly larger than the betas of the loser portfolios.5 For example, for the three-year experiment illustrated in Figure 1, the relevant numbers are respectively, 1.369 and 1.026 (t-statistic on the difference: 3.09). Thus, the loser portfolios not only outperformthe winnerportfolios;if the CAPM is correct, they are also significantly less risky. From a different viewpoint, therefore, the results in Table I are likely to underestimate both the true magnitudeand statistical significance of the overreactioneffect. The problem is particularlysevere with respect to the winner portfolio. Rather than 1.369, the residual return calculations assume the CAPM-beta of that portfolio to equal 5The CAPM-betasare found by estimatingthe marketmodel over a period of 60 months prior to portfolioformation. The Journal of Finance 802 Average of 49 One-Year Periods, 24 Two-Year Periods, 16 Three-year Periods Between January 1931 and December 1982 0.201 _ O. 15- 0.0 I 1? 0.05- -a. 05ji -0.10 0 2 4 0 8 10 12 14 10 18 20 22 24 MONTIIS AFTErlPORTFOLIOFOntRATIOtI Figure 2. Differencesin CumulativeAverageResidualBetween Winner and Loser Portfolios of 35 Stocks (formedover the previousone, two, or three years;1-24 months into the test period) 1.00 only. This systematic bias may be responsible for the earlier observed asymmetryin the return behavior of the extreme portfolios. To reiterate, the previous findings are broadlyconsistent with the predictions of the overreactionhypothesis. However, several aspects of the results remain without adequateexplanation. Most importantly,the extraordinarilylarge positive excess returns earned by the loser portfolio in January. One method that allows us to further accentuate the strength of the January effect is to increase the number of replications. Figure 3 shows the ACAR's for an experiment with a five-year-longtest period. Every Decemberbetween 1932 and 1977, winner and loser portfolios are formed on the basis of residual return behaviorover the previous five years. Clearly,the successive 46 yearly selections are not independent.Therefore,no statistical tests are performed.The results in Figure 3 have some of the properties of a "trading rule." They represent the average (cumulative) excess return (before transaction costs) that an investor, aware of the overreaction phenomenon, could expect to earn following any 803 Does the Stock Market Overreact? Replications Average of 46 Yearly 1933 and 1978 Starting Every January Between Five Years Period: Length of Formation 0.4l 0.3. "st- 1 W " C n K .1 02-1 oK 4 Loser Portfolio 0.0 A ~ 1l ~ *A J I Winner Portfolio -0.2-1 -0.20 3 0 8 12 15 10 21 24 21 MUN[IS AFIII 30 U 33 3U 30 42 4t5 48 bl 54 51 UU FFOLIOFUX)AT10t Figure 3. CumulativeAverage Residuals for Winner and Loser Portfolios of 35 Stocks (1-60 months into the test period) December in which he chose to try the strategy. The effect of multiplying the numberof replications is to remove part of the random noise. The outstanding feature of Figure 3 is, once again, the January returns on the loser portfolio. The effect is observed as late as five Januaries after portfolio formation!Careful examination of Figure 3 also reveals a tendency, on the part of the loser portfolio, to decline in value (relativeto the market)between October and December. This observation is in agreement with the naive version of the tax-loss selling hypothesis as explained by, e.g., Schwert [25]. The winner portfolio, on the other hand, gains value at the end of the year and loses some in January (for more details, see De Bondt [7]). B. Implicationsfor OtherEmpirical Work The results of this study have interesting implications for previous work on the small firm effect, the January effect and the dividend yield and PIE effects. Blume and Stambaugh [6], Keim [16], and Reinganum [21] have studied the 804 The Journal of Finance interaction between the small firm and January effects. Their findings largely redefine the small firm effect as a "losing firm" effect around the turn-of-theyear.6 Our own results lend further credence to this view. Persistently, losers earn exceptionally large January returns while winners do not. However, the companies in the extreme portfolios do not systematically differ with respect to market capitalization. The January phenomenon is usually explained by tax-loss selling (see, e.g., Roll [23]). Our own findings raise new questions with respect to this hypothesis. First, if in early January selling pressure disappears and prices "rebound"to equilibriumlevels, why does the loser portfolio-even while it outperformsthe market-"rebound" once again in the second January of the test period? And again,in the third and fourthJanuaries?Secondly,if prices "rebound"in January, why is that effect so much larger in magnitude than the selling pressure that "caused"it during the final months of the previous year? Possible answers to these questions include the argument that investors may wait for years before realizing losses, and the observedseasonality of the market as a whole. With respect to the PIE effect, our results support the price-ratio hypothesis discussed in the introduction,i.e., high PIE stocks are "overvalued"whereas low PIE stocks are "undervalued."However, this argument implies that the PIE effect is also, for the most part, a January phenomenon. At present, there is no evidence to support that claim, except for the persistent positive relationship between dividend yield (a variable that is correlated with the PIE ratio) and January excess returns (Keim [15]). III. Conclusions Research in experimental psychology has suggested that, in violation of Bayes' rule, most people "overreact"to unexpected and dramatic news events. The question then arises whether such behavior matters at the market level. Consistent with the predictions of the overreaction hypothesis, portfolios of prior "losers",are found to outperformprior "winners."Thirty-six months after portfolio formation, the losing stocks have earned about 25% more than the winners, even though the latter are significantly more risky. Several aspects of the results remain without adequate explanation; most importantly,the large positive excess returns earned by the loser portfolio every January. Much to our surprise, the effect is observed as late as five years after portfolio formation. ' Even after purging the data of tax-loss selling effects, Reinganum [22] finds a (considerably smaller) January seasonal effect related to company size. This result may be due to his particular definition of the tax-loss selling measure. The measure is related to the securities' relative price movementsover the last six monthspriorto portfolioformationonly. Thus, if many investorschoose to wait longer than six months before realizinglosses, the portfolio of small firms may still contain many "losers." REFERENCES 1. K. J. Arrow. "Risk Perception in Psychology and Economics."EconomicInquiry 20 (January 1982), 1-9. Does the Stock Market Overreact? 805 2. R. Ball. "Anomaliesin RelationshipsBetween Securities'Yields and Yield-Surrogates."Journal of Financial Economics 6 (June-September1978), 103-26. 3. S. Basu. "Investment Performanceof Common Stocks in Relation to Their Price-Earnings Ratios:A Test of the Efficient MarketHypothesis."Journalof Finance 3 (June 1977), 663-82. . "The Relationship Between Earnings' Yield, Market Value and Return for NYSE 4. CommonStocks:FurtherEvidence."Journal of FinancialEconomics12 (June 1983), 129-56. 5. W. Beaverand W. R. Landsman."Noteon the Behaviorof ResidualSecurityReturnsfor Winner and Loser Portfolios."Journal of Accounting and Economics 3 (December1981), 233-41. 6. M. Blume and R. Stambaugh."Biasesin ComputedReturns:An Applicationto the Size Effect." Journal of Financial Economics 12 (November1983), 387-404. 7. W. F. M. De Bondt. "Doesthe Stock MarketOverreactto New Information?"UnpublishedPh.D. dissertation,CornellUniversity, 1985. 8. D. N. Dreman. The New Contrarian Investment Strategy. New York: Random House, 1982. 9. E. F. Fama.Foundationsof Finance. New York:Basic Books, Inc., 1976. 10. B. Graham. The Intelligent Investor, A Book of Practical Counsel, 3rd ed. New York: Harper & BrothersPublishers,1959. 11. . The Intelligent Investor, A Book of Practical Counsel, 4th rev. ed. New York: Harper & BrothersPublishers,1973. 12. D. M. Grether."BayesRule as a DescriptiveModel:The RepresentativenessHeuristic."Quarterly Journal of Economics 95 (November1980), 537-57. 13. R. Jarrow."Beliefs,Information,Martingales,and ArbitragePricing."WorkingPaper, Johnson GraduateSchool of Management,CornellUniversity, November1983. 14. D. Kahnemanand A. Tversky. "IntuitivePrediction:Biases and CorrectiveProcedures."In D. Kahneman, P. Slovic, and A. Tversky, (eds.), Judgment Under Uncertainty: Heuristics and Biases. London:CambridgeUniversity Press, 1982. 15. D. Keim. "FurtherEvidenceon Size Effects and Yield Effects:The Implicationsof Stock Return Seasonality."WorkingPaper,GraduateSchool of Business, University of Chicago,April 1982. . "Size-RelatedAnomalies and Stock Return Seasonality:Further EmpiricalEvidence." 16. Journal of Financial Economics 12 (June 1983), 13-32. 17. J. M. Keynes. The General Theory of Employment, Interest and Money. London: Harcourt Brace Jovanovich,1964 (reprintof the 1936 edition). 18. A. W. Kleidon. "Stock Prices as Rational Forecastersof Future Cash Flows."WorkingPaper, GraduateSchool of Business, University of Chicago,November1981. 19. T. A. Marsh and R. C. Merton. "AggregateDividend Behaviorand Its Implicationsfor Tests of Stock MarketRationality."WorkingPaper No. 1475-83, Sloan School of Management,MIT, September1983. 20. J. A. Ohlson and S. H. Penman. "VarianceIncreasesSubsequentto Stock Splits: An Empirical Aberration."WorkingPaper, GraduateSchool of Business, ColumbiaUniversity, September 1983. 21. M. R. Reinganum."Misspecificationof Capital Asset Pricing: Empirical Anomalies Based on Earnings'Yields and MarketValues."Journal of FinancialEconomics9 (March1981), 19-46. 22. . "TheAnomalousStock MarketBehaviorof SmallFirmsin January."Journalof Financial Economics12 (June 1983), 89-1-04. 23. R. Roll. "Vasist das?".Journal of PortfolioManagement10 (Winter 1983), 18-28. 24. T. Russell and R. Thaler. "The Relevance of Quasi-Rationality in Competitive Markets." American Economic Review 75 (1985), forthcoming. 25. G. W. Schwert."Sizeand Stock Returns,and OtherEmpiricalRegularities."Journalof Financial Economics12 (June 1983), 3-12. 26. R. J. Shiller. "The Volatilityof Long-TermInterest Rates and ExpectationsModelsof the Term Structure." Journal of Political Economy 87 (December 1979), 1190-1219. 27. . "DoStock Prices MoveToo Muchto be Justifiedby SubsequentChangesin Dividends?" American Economic Review 71 (June 1981), 421-36. 28. J. B. Williams. The Theoryof Investment Value.Amsterdam:North-Holland, 1956 (reprint of 1938 edition).