the monetary environment and long-run reversals in stock returns

23
THE MONETARY ENVIRONMENT AND LONGRUN REVERSALS IN STOCK RETURNS Luis GarciaFeijoo Florida Atlantic University Gerald R. Jensen Northern Illinois University Abstract Previous research attributes longrun reversals to investor overreaction or taxmotivated trading; we offer an alternative explanation based on the monetary environment. Prices rebound for stocks that have performed poorly over the past several years (losers); however, the rebound occurs only during expansive monetary conditions. Winners only reverse course when monetary conditions are restrictive. Past research shows that the threefactor model explains longrun stock reversals; we show that the monetary environment plays an instrumental role in the observation. Finally, we show that reversal patterns are closely linked to both the monetary environment and a rms level of nancial constraints. JEL Classification: G12, G32 I. Introduction Over the past two decades, there has been increasing debate surrounding the longrun reversal that has been identied in stock returns. The longrun reversal pattern rst garnered signicant academic interest when DeBondt and Thaler (1985) observed that stocks with a prolonged period of poor performance, or loser stocks,subsequently outperformed winnersby an average of 31.9% over the next ve years. They attribute this observation to overreaction, whereby investors become overly pessimistic about stocks that are performing poorly and overly optimistic about stocks exhibiting superior performance. A consequence of investor overreaction is that price reversals occur for losers and winners as investors ultimately discover their opinions were too extreme. Subsequently, researchers have advanced alternative explanations for the longrun reversal pattern; these explanations rely on either rational, economic investor behavior or investor trading decisions that are based on irrational views. The authors acknowledge helpful comments received from an associate editor, an anonymous referee, and Werner DeBondt, Richard DeFusco, Sanjay Deshmukh, Art Durnev, Keith Gamble, Jon Garnkel, John Geppert, Wei Li, Jeff Madura, Amrita Nain, Manferd Peterson, Yiming Qian, Ashish Tiwari, Emre Unlu, Anand Vijh, Tong Yao, and Tom Zorn. In addition, the paper beneted from comments received from seminar participants at the University of Iowa, DePaul University, the University of NebraskaLincoln, Northern Illinois University, and at the meetings of the 2012 Financial Management Association (Atlanta, GA). All errors and omissions are our own. The Journal of Financial Research Vol. XXXVII, No. 1 Pages 325 Spring 2014 3 © 2014 The Southern Finance Association and the Southwestern Finance Association RAWLS COLLEGE OF BUSINESS, TEXAS TECH UNIVERSITY PUBLISHED FOR THE SOUTHERN AND SOUTHWESTERN FINANCE ASSOCIATIONS BY WILEY-BLACKWELL PUBLISHING

Upload: gerald-r

Post on 23-Dec-2016

227 views

Category:

Documents


10 download

TRANSCRIPT

Page 1: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

THE MONETARY ENVIRONMENT AND LONG‐RUN REVERSALSIN STOCK RETURNS

Luis Garcia‐Feijoo

Florida Atlantic University

Gerald R. Jensen

Northern Illinois University

Abstract

Previous research attributes long‐run reversals to investor overreaction or tax‐motivatedtrading; we offer an alternative explanation based on the monetary environment. Pricesrebound for stocks that have performed poorly over the past several years (losers);however, the rebound occurs only during expansive monetary conditions. Winners onlyreverse course when monetary conditions are restrictive. Past research shows that thethree‐factor model explains long‐run stock reversals; we show that the monetaryenvironment plays an instrumental role in the observation. Finally, we show that reversalpatterns are closely linked to both the monetary environment and a firm’s level offinancial constraints.

JEL Classification: G12, G32

I. Introduction

Over the past two decades, there has been increasing debate surrounding the long‐runreversal that has been identified in stock returns. The long‐run reversal pattern firstgarnered significant academic interest when DeBondt and Thaler (1985) observed thatstocks with a prolonged period of poor performance, or “loser stocks,” subsequentlyoutperformed “winners” by an average of 31.9% over the next five years. They attributethis observation to overreaction, whereby investors become overly pessimistic aboutstocks that are performing poorly and overly optimistic about stocks exhibiting superiorperformance. A consequence of investor overreaction is that price reversals occur forlosers and winners as investors ultimately discover their opinions were too extreme.Subsequently, researchers have advanced alternative explanations for the long‐runreversal pattern; these explanations rely on either rational, economic investor behavior orinvestor trading decisions that are based on irrational views.

The authors acknowledge helpful comments received from an associate editor, an anonymous referee, andWerner DeBondt, Richard DeFusco, Sanjay Deshmukh, Art Durnev, Keith Gamble, Jon Garfinkel, John Geppert,Wei Li, Jeff Madura, Amrita Nain, Manferd Peterson, Yiming Qian, Ashish Tiwari, Emre Unlu, Anand Vijh, TongYao, and Tom Zorn. In addition, the paper benefited from comments received from seminar participants at theUniversity of Iowa, DePaul University, the University of Nebraska–Lincoln, Northern Illinois University, and at themeetings of the 2012 Financial Management Association (Atlanta, GA). All errors and omissions are our own.

The Journal of Financial Research � Vol. XXXVII, No. 1 � Pages 3–25 � Spring 2014

3

© 2014 The Southern Finance Association and the Southwestern Finance Association

RAWLS COLLEGE OF BUSINESS, TEXAS TECH UNIVERSITYPUBLISHED FOR THE SOUTHERN AND SOUTHWESTERN

FINANCE ASSOCIATIONS BY WILEY-BLACKWELL PUBLISHING

Page 2: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

The explanation we offer falls most clearly into the category of rational,economic behavior by investors. Departing from the traditional literature, our explanationfocuses on time variation in risk and risk premia. Recent theoretical and empiricalevidence suggests that stock return expectations are affected by time variation in thefunding conditions for investors and firms (e.g., Brunnermeier and Pedersen 2009; Jensenand Moorman 2010). This research suggests that improved funding reduces risk aversionfor market makers and speculators and results in greater market liquidity.

Gertler and Gilchrist (1994) contend that small firms, due to their limited accessto funds, face greater financial constraints, which makes themmore sensitive to variationsin credit conditions. Thorbecke (1997) and Jensen and Moorman (2010) find evidencethat small firms are more sensitive to shifts in monetary policy, which offers empiricalsupport for Gertler and Gilchrist’s claim. We extend this line of research and argue thatfavorable monetary environments are most beneficial for firms that are most deficient infunding (small firms and financially constrained firms). Likewise, restrictive monetaryenvironments have the most negative implications for firms that have the least access to,and are most reliant on, external sources of funding, which again are firms of relativelysmall size and high levels of financial constraints.

We investigate the relation between monetary conditions and the long‐termreversal in stock prices for both winner and loser stocks. Our evidence supports thecontention that the prominence of the reversal pattern is conditional on the monetaryenvironment. In particular, a strong reversal pattern exists for losers during periods whenmonetary conditions are expansive, while there is no significant reversal for losers whenmonetary conditions are restrictive. We find that this observation holds even afterexcluding January returns from the analysis. Furthermore, we find the most substantialprice rebound occurs for firms that are small, financially constrained, and financiallydistressed, that is, firms that have characteristics that suggest they are funding deficient.1

A comparable relation exists with winner stocks, as we find that when monetaryconditions are restrictive, small, financially constrained winners suffer the most. Thisfinding is consistent with the contention that the availability of financing is moreimportant for firms that have relatively limited access to funding sources.

Overall, our evidence supports the contention that the monetary environment isan underlying factor that is at least partially responsible for the reversal pattern in loser andwinner stocks. Furthermore, our results provide an explanation for Fama and French’s(1996) observation that the three‐factor model captures the long‐run reversal pattern. Weargue that small firms and financially unstable firms, which according to Fama and Frenchcomprise a large part of the loser portfolio, are more sensitive to monetary conditions. Forsuch firms, the availability of financing is a prominent concern so the firms are moredependent on the monetary environment. Based on this premise, periods evidenced bygreater availability of financing are especially beneficial for these financially constrainedfirms as it provides access to capital to finance firm operations and for investors toaccumulate the stock.

1The discussion in Fama and French (1996) describes small firms and value firms as being more susceptible tofinancing problems.

4 The Journal of Financial Research

Page 3: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

In contrast, we report evidence that a restrictive monetary environment isespecially problematic for stocks that have appreciated markedly over the past few years(winners). The behavioral finance literature links winners with investor optimismregarding future firm prospects. A restrictive monetary environment may be viewed asproblematic for winners if it raises concerns about firms’ ability to maintain theirfavorable performance. The concerns would be heightened for small firms and financiallyconstrained firms, which have relatively few financing alternatives available.

II. Related Literature

Explanations for the Long‐Run Reversal in Returns

According to the overreaction theory, advanced byDeBondt and Thaler (1985), investors’excessive pessimism regarding the prospects of poorly performing stocks tends to drivetheir prices too low, but ultimately, investors recognize the stocks are underpriced andprices rebound. Likewise, investors become overly optimistic regarding the potential forstrongly performing stocks and tend to drive stock prices beyond their intrinsic values;again a reversal occurs when investors eventually realize their error. Subsequent researchby Chopra, Lakonishok, and Ritter (1992) and Lakonoshok, Shleifer, and Vishny (1994)confirm a long‐run reversal in stock returns and find evidence supporting the overreactiontheory advanced by DeBondt and Thaler.

Later studies by Barberis, Shleifer, and Vishny (1998), Daniel, Hirshleifer, andSubrahmanyam (1998), and Hong and Stein (1999) advance the view that long‐termreversals are an outcome of the same investor behavioral biases that are responsible forshort‐term momentum. Thus, in this literature, the two are considered inseparablephenomena. In contrast, George and Hwang (2004) find strong evidence indicating thatthe two anomalies are separate. Specifically, the authors find that a stock’s nearness to its52‐week high explains the short‐term trending in prices (price momentum) but does notexplain the long‐run price reversal. George and Hwang conclude that the two returnpatterns are separate phenomena and are induced by different investor behaviors. Theauthors propose two investor behavioral biases, anchoring and reference points, asexplanations for price momentum. The authors, however, find that these behavioral biasesdo not explain long‐run price reversals.

The focus of our research is to provide evidence regarding the underlying factorsthat drive long‐run price reversals. Our research is most consistent with the studiesdiscussed below, which advocate rational, economic behavior to explain the long‐runreversal pattern. In an early effort to link reversals with risk factors, Fama and French(1996) examine the long‐run reversal pattern relative to the three‐factor model. Theauthors find that after controlling for the three factors, the long‐run reversal patternbecomes insignificant. This finding suggests that the pattern is subsumed by the factorsincluded in the three‐factor model. However, consistent with the contention of Georgeand Hwang (2004) that momentum and long‐term reversals are separate effects, Famaand French find that the three‐factor model explains long‐run reversals but does notexplain momentum returns. Thus, Fama and French provide evidence supporting the

The Monetary Environment 5

Page 4: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

independence of momentum and long‐run reversals, and furthermore, their findingssupport a risk‐based explanation for long‐run reversals.

Another area of research attributes long‐run reversals to investor reactions tonews events (e.g., Berk, Green, and Naik 1999; Brav and Heaton 2002; Lewellen andShanken 2002). These studies differ from ours as they do not examine long‐run reversalsrelative to economic conditions. Cooper, Gutierrez, and Hameed (2004) relate economicconditions to momentum in stock returns but do not relate economic conditions and long‐run reversals

More recently, George and Hwang (2007) propose a tax‐based explanation forthe long‐run reversal in stock prices. Specifically, they contend the reversal is due to tax‐based trading, rather than the traditional explanation of investor overreaction. The authorsnote that DeBondt and Thaler (1987) and Grinblatt and Moskowitz (2004) find that long‐run reversals for “losing stocks” are concentrated in January, which suggests that end‐of‐year, tax‐based trading decisions may be responsible for the reversal pattern exhibited bylosers. George andHwang confirm that the long‐run reversal of loser stocks is exclusive toJanuary. Based on this evidence, they conclude that the reversal of losers is attributed toend‐of‐year, tax‐based trading.

George and Hwang (2007) find no reversal for losers outside of January, yet theyobserve a consistent reversal for winners. They argue that this evidence is consistent withthe capital gains lock‐in hypothesis. According to this hypothesis, stocks that havetrended up over a prolonged period have accumulated locked‐in capital gains.2

Subsequent return patterns prevail for the winners because investors holding the stocksrequire a premium to trade due to the adverse tax treatment associated with realizing thegains, whereas no premium is required by investors holding stocks with locked‐in losses(losers). Because capital gains are taxedwhen realized, the authors argue that investors arereluctant to sell winners unless they are compensated for the tax that will be assessed whenthe gain is realized. The authors evaluate returns in an environment that taxes capital gainson stocks, the U.S. market, versus a market that does not tax capital gains, Hong Kong.Based on their cross‐market analysis, they find evidence consistent with the tax‐basedexplanation and inconsistent with the overreaction hypothesis.

George and Hwang (2007) conclude that the reversal of loser stocks is isolated toJanuary and is due to tax‐based, end‐of‐year trading. They further note that the capitalgains lock‐in hypothesis predicts that, with the exception of January, only winners willreverse. The authors argue that their evidence for the U.S. market provides strong supportfor the capital gains lock‐in hypothesis.

The Monetary Environment and Expected Stock Returns

There are a number of studies that establish a relation between Federal Reserve (Fed)monetary policy changes, fund availability, and security returns. As outlined below, thisresearch theoretically and empirically motivates a potential link between monetaryconditions and long‐run reversals in stock prices.

2The locked‐in gain hypothesis advocated by George and Hwang (2007) is based on the model developed byKlein (1999).

6 The Journal of Financial Research

Page 5: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

Several studies identify a systematic relation between changes in Fed monetarypolicy and expected security returns (e.g., Jensen, Mercer, and Johnson 1996;Patelis 1997; Thorbecke 1997). These studies report evidence supporting the view thatexpansive monetary environments are associated with higher equity returns, whilerestrictive monetary conditions tend to coincide with lower returns. This evidencesupports the view that monetary conditions represent a risk factor that influences thesecurity pricing decisions of investors.

Related research by Fujimoto (2004) and Chordia, Sarkar, and Subrahmanyam(2005) link changes in Fed policy with variation in aggregate market liquidity.Specifically, the authors show that relatively tight (relaxed) monetary policy correspondsto less (more) aggregate liquidity. Brunnermeier and Pedersen (2009) contend thatchanges in aggregate liquidity correspond to changes in funding conditions for marketparticipants. Finally, Jensen and Moorman (2010) find evidence that links the threevariables together. In particular, they report a systematic relation among shifts in Fedpolicy, aggregate funding/liquidity conditions, and investor pricing decisions. Theauthors report empirical evidence that stock prices are influenced significantly by shifts inmonetary conditions (as identified by changes in Fed policy). They find that illiquid stocksexperience a substantial price rebound when funding conditions improve. They argue thatwhen monetary conditions are restrictive, investors are reluctant to hold illiquid stocks intheir portfolios. However, a shift to more favorable monetary conditions allows investorsto increase their allocation to illiquid stocks, which increases the demand for the stocks,and ultimately their prices.

Finally, our study is motivated by several studies that separately link individualfirm characteristics to the prominence of long‐run price reversals and the sensitivity tomonetary conditions (e.g., Fama and French 1996; George and Hwang 2004, 2007).There is also considerable evidence (e.g., Chan and Chen 1991; Hadlock and Pierce 2010)suggesting that firms that are relatively small and young have characteristics that makethem more capitally constrained. Consistent with these studies, Thorbecke (1997) andJensen and Moorman (2010) report evidence that small firms are more sensitive to shiftsin monetary policy. For example, Table 1 of Thorbecke shows that stock returns of smallfirms have a greater response to monetary policy shocks, which supports Gertler andGilchrist’s (1994) contention that small firms face greater financial constraints. Given thatsmall firms exhibit larger price reversals and show greater sensitivity to both changingcredit conditions and monetary policy shocks, we believe it is reasonable to ask whetherthe reversal pattern is influenced by a firm’s access to financing and the monetaryenvironment.

III. Sample and Variable Descriptions

Sample and Description of Winners and Losers

We evaluate monthly stock returns for winners and losers from 1963 through 2010. Firmsare identified as winners (losers) based on their prior long‐term stock performance;winners represent the firms ranked in the top 20% of past performance, and losers are

The Monetary Environment 7

Page 6: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

ranked in the bottom 20%. We base our selection of the portfolio‐formation period onthe findings reported in Fama and French (1996). In particular, Fama and Frenchevaluate a number of alternative formation periods and determine that the mostprominent long‐run reversal exists when the formation period is based on stock returnsfrom month –60 through month –13 relative to the evaluation period, that is, a five‐yearperformance interval with a one‐year skip period. This formation period is shown toreveal the strongest reversal pattern in both their 1931–1963 and 1963–1993 samples.Fama and French find, however, that the three‐factor model subsumes the long‐runreversal pattern associated with this formation period. They attribute this finding to thefact that the three‐factor model captures the economic essence of winners and losers sincewinners tend to be strong stocks, and losers are relatively small, distressed stocks. Usingthe same portfolio‐formation approach as Fama and French allows us to directlyinvestigate the extent to which differences in the monetary environment explain theirfindings.

The major focus of our research is to investigate the performance of winnersand losers relative to alternative monetary conditions. Following Jensen and Moor-man (2010), we rely on a composite measure of the monetary environment, which isbased on two monetary policy indicators that identify shifts in Fed policy. Jensen andMoorman provide evidence supporting the efficacy of their methodology in classifyingconditions as either expansive or restrictive. The following section provides a briefoverview of the approach advocated by Jensen and Moorman, and adopted in ourresearch.

Differentiating the Monetary Environment

Jensen and Moorman (2010) use two alternative monetary policy indicators ascomponents in developing a composite measure designed to differentiate the monetaryenvironment. The composite measure is intended to reflect the current environment andexpectations regarding future fund availability. Shifts in Fed policy influence the financialmarkets by affecting the expectations of market participants regarding fund availability,and also by ultimately leading to differences in monetary and reserve aggregate levels andgrowth rates.3

The first monetary policy component is based on changes in the federal fundsrate, which has a long history as an indicator of monetary policy stringency (e.g.,Bernanke and Blinder 1992; Patelis 1997; Perez‐Quiros and Timmermann 2000). Thesecond component is based on changes in the Fed discount rate and has been usedpreviously to identify periods in which the Fed’s “broad” monetary policy objectivesdiffer significantly (e.g., Booth and Booth 1997; Fujimoto 2004; Jensen and Moorman2010).

3Several studies model the influence of monetary policy on financial market participants via the impact onbank lending or the availability of money (e.g., Bernanke and Blinder 1988). Alternatively, monetary policy hasbeen linkedwith aggregate liquidity, which has been argued to affect market participants’willingness to hold illiquidassets (e.g., Brunnermeier and Pedersen 2009; Jensen and Moorman 2010). The monetary‐conditions measure weuse is consistent with both motivations.

8 The Journal of Financial Research

Page 7: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

Figure I illustrates the relation between the two monetary policy components thatcomprise the monetary‐conditions measure. As an indicator of policy stringency(tightness in the short‐term market), the federal funds rate is plotted on the graph, and theshaded areas indicate shifts in the Fed’s broad policy (as identified by directional changesin the Fed discount rate). Shaded areas represent periods where broad policy is restrictive(an increasing Fed discount rate environment), and unshaded areas denote a broad policythat is expansive. The data reported in the figure confirm the close association between thetwo policy variables. In particular, shifts in the Fed’s broad policy generally correspondwith subsequent like‐direction changes in the Fed’s short‐term policy rate (federal fundsrate). Although the two variables align fairly closely, it is obvious that they are notredundant as the two diverge occasionally, with the divergences generally occurringshortly before a broad policy shift. The graph is consistent with the view that a shift in theFed’s broad policy stance can be viewed as a signal that recent developments in monetarystringency will continue.

Our focus is to identify environments that represent generally different monetaryenvironments. Thus, following Jensen andMoorman (2010), we consider binary versionsof the two policy indicators in developing the monetary‐conditions measure.4 Thisapproach is supported by Patelis (1997) who claims that the influence of shifts in

Figure I. Monetary Conditions, 1963–2010.

4The monetary‐conditions measure is intended to identify periods that reflect broadly different Fed policies.However, the measure is not likely to capture differences in the environment that are unrelated to interest rates, nordoes the measure differentiate rate movements due to shifts in demand from those associated with changes in supply.Finally, the interdependent nature of business conditions and monetary conditions means the monetary‐conditionsmeasure necessarily incorporates elements of both.

The Monetary Environment 9

Page 8: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

monetary policy stringency varies based on the Fed’s broad policy stance. The approach isfurther supported by Figure I and anecdotal evidence that the Fed frequently modifies itspolicy “bias” or “tilt” without changing its fundamental policy stance.5

Following Jensen and Moorman (2010), we classify the monetary environ-ment as expansive if both monetary policy measures are consistent in signalingexpansive conditions. In contrast, we identify a restrictive environment when eithermeasure signals restrictive conditions. For example, conditions are expansive whenthe effective federal funds rate decreases from month t–1 to month t and the previouschange in the Fed discount rate was also a decrease. In contrast, monetary conditionsare considered restrictive if either monetary indicator signals a restrictive policy.Restrictive environments result whenever market participants have reason to believe thatfinancing may be less readily available in the future, which is the case when either of themonetary indicators suggests the Fed is applying a restrictive policy. For months inwhich neither of the policy rates change, the existing monetary environment ismaintained.6

IV. Results

Evidence Regarding Long‐Run Reversals

To confirm the existence of a long‐run reversal in stock returns, Panel A of Table 1 reportsequally weighted average returns for past‐performance quintiles. We follow Fama andFrench (1996) in establishing a firm’s long‐run past performance.7 Specifically, we adoptthe Fama–French formation period, which is composed of months –60 through –13. Atthe beginning of each month, firms are allocated to quintiles based on continuouslycompounded returns during the portfolio‐formation period. The lowest (highest) past‐performance quintile is denoted as loser (winner) portfolio. The sample period is fromJune 1963 to December 2010 (i.e., the first reversal is based on performance measuredfrom 1958 to 1962). Returns are adjusted for delisting by using the delisting return fromthe Center for Research in Security Prices (CRSP); however, if the delisting is forperformance‐related reasons, the delisting return is –55% if trading onNASDAQor –30%if on NYSE/AMEX. The portfolios are reformed monthly.

Following Fama and French (1996), we measure portfolio return for month t,which is 12 months after the month –60 through –13 formation period. Fama and Frenchcontend that a 12‐month skip period is necessary to clearly separate the momentum effect

5See Bernanke, Reinhart, and Sack (2004) for a discussion of Fed policy shifts and Fed communications.6There are 367 months in which monetary conditions (using both the broad and short‐term policy indicators)

do not change from the previous month: 154 months in which both remain restrictive, 113 in which both remainexpansive, 59 in which the broad policy remains expansive and the short‐term indicator remains restrictive, and 41 inwhich the broad indicator remains restrictive and the short‐term indicator remains expansive.

7We repeat the analysis using two alternative formation periods (months –36 through –7 and months –36through –2). Our findings confirm the evidence of Fama and French (1996) that the month –60 through –13formation period produces the most prominent reversal; however, the reversal phenomenon also exists with otherformation periods.

10 The Journal of Financial Research

Page 9: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

from the long‐run reversal effect. We rely on the month –60 through –13 formation periodfor several reasons. First, this formation period most clearly distinguishes momentumfrom long‐run reversals. Second, the approach is used in calculating the long‐run reversalreturns reported on Kenneth French’s website (http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/index.html). Third, one of our primary objectives is to determinewhether monetary conditions explain the long‐run reversal phenomenon documented inprevious research. Thus, we evaluate the most prominent pattern established in widelyacknowledged research.

The returns in Panel A clearly show evidence of a long‐run reversal pattern inreturns as losers outperform winners by 90 basis points per month. The 90 basis pointdifference in performance corresponds fairly closely with the corresponding differencereported by Fama and French (1996). In particular, Fama and French report a 74 basispoint difference in monthly returns between their loser and winner deciles over the periodfrom 1963 through 1993. Our results are reassuring as they indicate that the reversalpattern identified by Fama and French is maintained in our sample, which extends the

TABLE 1. Mean Monthly Returns and Regression Results for Past Performance Portfolios: June 1963 toDecember 2010.

Reversal Portfolio

LMWLoser P2 P3 P4 Winner

Panel A. Mean Monthly Return for Past‐Performance Portfolios

Return 1.85 1.36 1.27 1.18 0.95 0.90t‐statistic 5.99��� 5.77��� 5.99��� 5.58��� 3.75��� 5.12���

Panel B. Regression Results for Explaining Past‐Performance Portfolio Returns

a �0.02 �0.06 0.06 0.06 �0.15 0.14b 1.04 0.95 0.94 0.95 1.10 �0.05s 1.18 0.78 0.58 0.50 0.55 0.63h 0.47 0.42 0.41 0.31 0.02 0.45j 5.50 1.73 0.64 0.18 0.32 5.17t(a) �0.09 �1.04 1.06 0.91 �1.99�� 0.91t(b) 24.28��� 43.44��� 47.45��� 51.93��� 41.21��� �1.25t(s) 16.60��� 20.24��� 12.71��� 11.66��� 8.26��� 8.08���

t(h) 5.44��� 10.66��� 11.46��� 7.87��� 0.42 4.76���

t(j) 6.60��� 4.53��� 2.53�� 0.83 0.84 6.73���

R2 0.83 0.93 0.94 0.93 0.92 0.40

Note: Panel A shows equally weighted average monthly returns (in %) for portfolios (quintiles) formed on pastperformance. Following Fama and French (1996), portfolios are formed based on returns between months –60 and–13 relative to measurement month t. Panel B reports results for regressions of monthly excess returns for pastperformance portfolios on the Fama–French three factors and a dummy variable for the month of January, as follows:

Ri � Rf ¼ ai þ biðRm � Rf Þ þ siSMBþ hiHMLþ jiJanuary Dummyþ ei:

The final column shows results for the loser minus winner portfolio (LMW). The table t‐statistics [t(•)] are based onNewey–West adjusted standard errors.���Significant at the 1% level.��Significant at the 5% level.

The Monetary Environment 11

Page 10: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

Fama–French sample period by 17 years. Furthermore, our sample includes NYSE,AMEX, and NASDAQ stocks, whereas Fama and French limited their sample to NYSEstocks.

Fama and French (1996) find evidence suggesting that the three‐factormodel captures the economic essence of losers and winners. Specifically, they contendthat losers behave similarly to small distressed stocks, and thus, losers load moreheavily on the small minus big (SMB) and high minus low (HML) factors. Inaddition, several authors produce evidence indicating that the loser minus winner(LMW) premium is strongly influenced by returns in January (e.g., DeBondtand Thaler 1987; Grinblatt and Moskowitz 2004; George and Hwang 2007). Basedon this evidence, we investigate the relation between reversal returns and the Fama–French three factors (Rm – Rf, SMB, and HML), while also considering the influence ofJanuary. The results of our time‐series regression analysis are reported in Panel B ofTable 1.

The findings in Panel B support the research referenced above with respect toboth the three‐factor model and the uniqueness of January returns. Specifically, consistentwith Fama and French’s contention, the loser portfolio loads relatively heavily on theSMB factor (coefficient¼ 1.18 with t‐statistic¼ 16.60) and HML factor (coefficient¼0.47 with t‐statistic¼ 5.44). In addition, both factors are highly significant in explainingthe return on the LMWportfolio (t‐statistics¼ 8.08 and 4.76). Furthermore, the importantrole that January plays in the reversal return is confirmed by the strong significanceidentified for the January dummy in explaining the return for both the loser portfolio(coefficient¼ 5.50with t‐statistic¼ 6.60) and the LMWportfolio (coefficient¼ 5.17witht‐statistic¼ 6.73). Finally, after controlling for the three factors and January, the abnormalreturn (the intercept or a) to the LMW portfolio is only 14 basis points and is insignificant(t¼ 0.91), which confirms Fama and French’s contention that the three‐factor modelexplains the reversal return.

As noted above, DeBondt and Thaler (1987) andGrinblatt andMoskowitz (2004)find that the size of the long‐run reversal return is strongly influenced by returns inJanuary. Our findings in Table 1 confirm this claim but indicate that the significance ofJanuary is limited to the three worst past‐performance portfolios (loser, P2, and P3).George and Hwang (2007) make a stronger argument claiming that while winners reverseconsistently, losers only reverse significantly during January. To directly investigateGeorge and Hwang’s claim, in Table 2 we report the size of the loser and winner reversalwith and without returns for January.

Following George and Hwang (2004, 2007) and Grinblatt and Moskowitz(2004), we use Fama–MacBeth (1973; henceforth FM) regressions to evaluate the relativeperformance of past losers and past winners when January returns are included (Panel A)and when January returns are excluded (Panel B). Consistent with George and Hwang(2007), we define two separate dummy variables to identify stocks that are in the bottomquintile (losers) and top quintile (winners) of past performance. Panel A results confirmthe existence of a strong long‐run reversal pattern over the 571‐month sample period.Specifically, on a relative return basis, the return for losers rebounds by 59 basis points,while the winner return falls by 31 basis points. Both values are economically andstatistically significant.

12 The Journal of Financial Research

Page 11: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

The FM regression results in Panel B (with January excluded) support theprominent role that January plays in long‐run reversals as the reversal for losersdisappears when January is omitted from the analysis. With January excluded, thecoefficient on the loser dummy drops from 0.59 to 0.15, and it becomes insignificant.With respect to winners, excluding January reduces the prominence of the winnerreversal; however, the reversal remains highly statistically significant. These resultscorrespond closely with George and Hwang’s (2007) findings.

The Monetary Environment and Long‐Run Price Reversals

We next investigate the relation between monetary conditions and long‐run pricereversals. In spite of the strong theoretical association between the two variables, theirassociation has not previously been examined. Panel A of Table 3 reports the returns to aportfolio that is long the stocks contained in the loser quintile and short the stocks in thewinner quintile. This zero‐cost portfolio is referenced as the LMW (loser minus winner)portfolio hereafter. Panel A reports LMW returns for expansive and restrictive monetaryenvironments, and Panel B presents data on monetary and reserve aggregates across thetwo monetary environments.

The LMW return during expansive periods, 1.85%, is substantial. In starkcontrast, the LMW return during restrictive periods is less than one‐third the size.8

TABLE 2. Long‐Run Reversals and January Returns.

b0 Loser Winner

Panel A. Full Sample

Coefficient 1.27 0.59 �0.31t‐statistic 5.34��� 3.34��� �3.70���

Panel B. January Excluded

Coefficient 1.01 0.15 �0.23t‐statistic 4.26��� 0.90 �2.69���

Note: The table shows the results of monthly Fama–MacBeth (1973) regressions of the form:

Rit ¼ b0t þ b1t Loseri þ b2t Winneri þ eit ;

whereRit is the return to stock i in month t. Loser (Winner) is a dummy variable that takes the value of 1 if stock i is inthe lowest (highest) quintile based on continuously compounded returns over month –60 to month –13. Thet‐statistics are based on time‐series standard errors (Newey–West adjusted with four lags). Returns are adjusted fordelisting by using the delisting return from CRSP (however, if the delisting is for performance‐related reasons, thedelisting return is –55% if trading on NASDAQ or –30% if on NYSE/AMEX). In the table, Loser and Winnerrepresent the corresponding (time‐series average) coefficient estimates, in%. The sample portfolio‐formation periodis from June 1963 through December 2010 (571 months).���Significant at the 1% level.

8Following Jensen and Moorman (2010), monetary conditions are determined based on a one‐month lagrelative to measured returns. This represents a conservative approach to avoid having an announcement‐periodreaction drive our results. The approach is also consistent with measuring stringency differences via changes in theaverage monthly federal funds rate.

The Monetary Environment 13

Page 12: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

Parametric (t‐tests) and nonparametric (Wilcoxon rank sum) tests confirm the significantdifference between LMW returns in expansive versus restrictive environments.

Panel B of Table 3 reports changes in two reserve measures and one measure ofmonetary aggregates. The three aggregates are advocated by Thornton (1998). In testingfor differences in the aggregates, we employ a nonparametric test in addition to the t‐testdue to the high incidence of extreme outliers in the monetary and reserve aggregate data.

The large differences between changes in the aggregates across the twoenvironments support the efficacy of the monetary‐conditions classification approach indifferentiating the monetary environment. The data indicate that fund availability differssubstantially between the two environments, with fund availability increasing to a muchgreater degree during expansive periods.9

The LMW returns reported in Panel A provide strong support for the contentionthat the long‐run reversal phenomenon is strongly influenced by monetary conditions.When conditions are expansive, there is a strong long‐run reversal in stock prices, whilethe reversal is greatly diminished when the monetary environment is restrictive. Theprominence of the portfolio return during expansive environments is an unprecedented

TABLE 3. LMW Returns, Reserve/Monetary Aggregates, and Monetary Environments.

Panel A. LMW Mean Monthly Portfolio Returns

Expansive (n¼ 173) Restrictive (n¼ 398) t‐statistic z‐score

LMW portfolio return 1.85% 0.48% 2.87��� 2.66���

Panel B. Mean Monthly Changes in Reserve and Monetary Aggregates

Monthly Change in: Expansive (n¼ 173) Restrictive (n¼ 398) t‐statistic z‐score

Total reserves 3.47% 0.11% 1.72� 4.80���

Nonborrowed reserves 3.98% 0.21% 1.05 4.72���

Adjusted monetary base 1.14% 0.49% 2.08�� 4.78���

Note: Panel A shows average monthly returns for the loser minus winner (LMW) portfolio in expansive andrestrictive monetary periods, while Panel B shows average change in three prominent measures of fund availability.Following Fama and French (1996), loser and winner portfolios are determined based on continuously compoundedreturns between month –60 and month –13 relative to the portfolio‐formation month t. The sample portfolio‐formation period is from June 1963 through December 2010. Returns are adjusted for delisting by using the delistingreturn fromCRSP (however, if the delisting is for performance‐related reasons, the delisting return is –55% if tradingon NASDAQ or –30% if on NYSE/AMEX). Panel B shows the percentage change in funding aggregates acrossdifferent monetary environments. Changes in aggregates are taken from monthly observations in total reserves,nonborrowed reserves, and the adjusted monetary base. Changes in aggregates are measured inmonth tþ 1 based onmonetary conditions determined in month t. The total sample consists of 571 months. In the right‐most columns,t‐statistics and Wilcoxon z‐scores are reported for a test of the difference in value across the two monetaryenvironments.���Significant at the 1% level.��Significant at 5% level.�Significant at 10% level.

9Using the monetary‐conditions measure has two distinct advantages over relying on actual changes inmonetary/reserve aggregates. First, it incorporates the influence that Fed policy shifts have on investor expectationsregarding future fund availability. Second, the measure is derived from policy indicators that are made readilyavailable to the investing public on a timely basis.

14 The Journal of Financial Research

Page 13: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

finding, and furthermore, the finding is inconsistent with previous explanations for thelong‐run reversal. The fact that the expansive environment represents 173 months (30%)of the 571‐month sample period supports the claim that there is a systematic relationbetween monetary conditions and long‐run reversals. Furthermore, the similarity in theLMW return pattern with the monetary and reserve aggregate patterns reported in Panel Bsuggests that differences in the monetary environment offer an economic explanation forthe long‐run reversal pattern.

Table 4 extends Table 2 by separating the FM regression results into the twomonetary environments: expansive and restrictive. Panel A reports the results for theentire 571 months of the sample period. The findings document the strong influence thatmonetary conditions have on the reversal pattern. There is a very pronounced loserreversal of 1.51% when the monetary environment is expansive; however, duringrestrictive conditions the loser reversal is very small, 0.18%, and insignificant. In contrast,for winners there is a significant reversal, of comparable size, in both the expansiveenvironment (–0.37) and the restrictive environment (–0.30).

Panel B of Table 4 confirms that losers reverse even after excluding Januaryreturns, but only during expansive monetary conditions. Furthermore, the reversal is ofcomparable size to the aggregate loser reversal identified in Table 2. In contrast, duringrestrictive periods, on average, losers continue to underperform by a small andinsignificant amount. With respect to winners, when the monetary environment is

TABLE 4. Long‐Run Reversals and Monetary Environments.

Expansive Conditions Restrictive Conditions

b0 Loser Winner b0 Loser Winner

Panel A. Full Sample

Coefficient 2.24 1.51 �0.37 0.85 0.18 �0.30t‐statistic 4.27��� 4.11��� �2.23�� 3.39��� 1.01 �2.87���

Panel B. January Excluded

Coefficient 1.83 0.76 �0.19 0.68 �0.10 �0.25t‐statistic 3.81��� 2.39�� �1.18 2.60��� �0.59 �2.50��

Note: The table shows the results of monthly Fama–MacBeth (1973) regressions of the form:

Rit ¼ b0t þ b1t Loseri þ b2t Winneri þ eit ;

where Rit is the monthly return to stock i in month t. Loser (Winner) is a dummy variable that takes the value of 1 ifthe stock is in the lowest (highest) quintile based on continuously compounded returns over month –60 to month–13. The t‐statistics are based on time‐series standard errors (Newey–West adjusted with four lags). The monetaryenvironment is determined based on monetary policy variables that are lagged one month relative to the returns.Returns are adjusted for delisting by using the delisting return from CRSP (however, if the delisting is forperformance‐related reasons, the delisting return is –55% if trading on NASDAQ or –30% if on NYSE/AMEX). Inthe table, Loser and Winner represent the corresponding (time‐series average) coefficient estimates, in %. Thesample portfolio‐formation period is from June 1963 through December 2010.���Significant at the 1% level.��Significant at the 5% level.

The Monetary Environment 15

Page 14: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

expansive, there is not a significant winner reversal; however, the winner reversal ishighly significant when the environment is restrictive.

Firm Characteristics, Monetary Conditions, and Long‐Run Reversals

Our findings in Table 1 confirm that the three‐factor model explains the reversal return;however, our results in Tables 3 and 4 indicate that monetary conditions also play a veryprominent role in the reversal pattern. We next investigate the relation among reversalreturns, the firm characteristics reflected in the three‐factor model, and the monetaryenvironment. Our objective in evaluating the interdependent relation among these variablesis to provide additional elucidation regarding the reasons that long‐run reversals occur.

Previous research suggests that the monetary environment is expected to have amore prominent influence on firms that have less access to capital and firms that are morereliant on external sources of capital. The findings from several studies (e.g., Chan andChen 1991; Gertler and Gilchrist 1994; Vassalou and Xing 2004) suggest that small firmsgenerally fit this classification. In addition, firm size is a prominent component of thefinancial constraint measure advocated byHadlock and Pierce (2010) andWhited andWu(2006). Although firm size is a commonly recognized proxy for access to capital, there arereasons to expect that a firm’s book‐to‐market equity (BEME) may also reflect a firm’sease in accessing external capital markets. Several studies suggest that firms withrelatively high BEME are frequently financially distressed. Banko, Conover, and Jensen(2006) report evidence supporting this view as they find that high‐BEME firms (valuefirms) tend to have greater earnings uncertainty, higher leverage, and a greater propensityto reduce their dividend. Financially distressed firms would generally be expected to facegreater problems when attempting to secure financing. Thus, there are reasons to expect amore prominent reversal for firms with high BEME ratios. The final firm characteristicincluded in the three‐factor model is beta, which is included to control for differences inmarket risk across the portfolios.

The results of the integrated analysis are reported in Tables 5 through 8. Table 5reports the findings of FM regressions (of the same type as those in Table 4) estimatedseparately for small (low‐market‐capitalization) and big (high‐market‐capitalization)firms. Our findings reported above confirm that, relative to other months, long‐runreversals are much more prominent in January, possibly due to end‐of‐year tradingeffects. Therefore, to avoid undue influence from January returns, regression results inTables 5 through 8 exclude January returns.

Hadlock and Pierce (2010) report convincing evidence that there is a nonlinearrelation between a firm’s characteristics and the level of financial constraints the firmfaces. For example, the authors report evidence showing that financial constraints arerelatively severe for firms in the smallest size deciles relative to those in the largest sizedeciles. However, there is relatively little variation in the level of financial constraintsamong decile 3 through decile 8. By performing FM regressions separately on small (low‐ME) and large (high‐ME) firms, we avoid any potential influence on our findings causedby a nonlinear relation between firm characteristics and funding deficiencies.

When examining the relation among reversal returns, firm size (ME), andmonetary conditions, we include control variables for the other two factors from the three‐

16 The Journal of Financial Research

Page 15: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

factor model (BEME and beta). Likewise, in Tables 6 and 7, we evaluate one factor whilecontrolling for the other two factors. In addition, in Tables 6, 7, and 8, we perform separateregressions on the top and bottom quintiles of ranked firms to avoid the nonlinear issuesidentified by Hadlock and Pierce (2010).

The findings in Table 5, Panel A clearly show that the reversal phenomenon ismuch more pronounced for small firms relative to big firms. For the full sample, the loserreversal is insignificant for small firms (coefficient¼ 0.19 with t‐statistic¼ 1.20) and forbig firms (coefficient¼ –0.15 with t‐statistic¼ –0.95). The winner reversal, however, ishighly significant for small firms and is more than four times larger than the reversal forbig firms (�0.38 vs. �0.09).

TABLE 5. Fama–MacBeth Regressions by Size (ME) and Monetary Conditions.

ME b0 Loser Winner

Control Variables

Ln(BEME) Beta

Panel A. All Periods

Low (small) 0.02 0.19 �0.38 0.90 0.43t‐statistic 0.07 1.20 �2.18�� 5.76��� 1.47High (big) 1.05 �0.15 �0.09 �0.09 0.00t‐statistic 5.51��� �0.95 �0.52 �0.52 0.01

Panel B. Expansive Conditions

Low (small) 0.74 0.76 �0.09 0.83 0.80t‐statistic 1.53 2.77��� �0.31 3.52��� 1.44High (big) 1.26 0.03 �0.05 �0.19 0.36t‐statistic 4.32��� 0.08 �0.34 �0.71 0.74

Panel C. Restrictive Conditions

Low (small) �0.31 �0.07 �0.51 0.92 0.26t‐statistic �1.03 �0.40 �2.36�� 4.76��� 0.81High (big) 0.96 �0.24 �0.11 �0.24 �0.16t‐statistic 4.13��� �1.42 �1.03 �1.42 �0.54

Note: Each month, stocks are grouped into quintiles based on returns between months –60 and –13. Independently,stocks are sorted into quintiles each month based on market capitalization (ME) at the beginning of the month. Eachmonth, we run Fama–MacBeth (1973) regressions of the form:

Rit ¼ b0t þ b1t Loseri þ b2t Winneri þ Control Variablesit þ eit ;

whereRit is themonthly return to stock i inmonth t.Winner (Loser) is a dummy variable that equals 1 if the stock is inthe winner (loser) quintile based on continuously compounded returns over months –60 to –13. With regard to thecontrol variables, BEME is defined as in Fama and French (1992) and beta is computed using the CRSP value‐weighted index over the prior 60 months. We add 1 to BEME before taking the natural log. The t‐statistics are basedon the time‐series (Newey–West adjusted with four lags) standard errors. The monetary conditions variable isdetermined based onmonetary policy variables that are lagged onemonth relative to the returns. Returns are adjustedfor delisting. In the table, Loser and Winner (and Ln(BEME) and Beta) represent the corresponding (time‐seriesaverage) coefficient estimates (in %). January returns are excluded.���Significant at the 1% level.��Significant at the 5% level.

The Monetary Environment 17

Page 16: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

Relative to the full‐period results, the results during expansive monetaryenvironments (Panel B) differ dramatically as the loser reversal for small firms issubstantial in both economic and statistical terms (coefficient¼ 0.76 with t‐statistic¼2.77). Because January returns are excluded, this finding contrasts sharply with the claimof George and Hwang (2007) that outside of January there is no evidence of a loserreversal. In contrast to the prominent loser reversal when the monetary environment isexpansive, there is no evidence of a significant winner reversal for small or big firmsduring expansive conditions.

Panel C reports reversal returns when monetary conditions are restrictive. Thefindings show that the reversal of winners is restricted to small firms during restrictiveperiods. Again, this result is inconsistent with the capital gains lock‐in hypothesis, whichimplies the winner reversal is invariant to the monetary environment. Furthermore, duringrestrictive conditions, losers show no evidence of a significant rebound in performancewhether the firms are small or big.

Interestingly, the results reported in Table 5 offer no evidence of a significantreversal pattern for big firms for the full sample period or either monetary environmentsubperiod. This evidence supports the contention that monetary conditions are a moreimportant consideration for small firms because of their limited access to alternativesources of capital. According to this view, firms that have experienced a prolonged periodof poor stock performance (losers) face considerable hurdles in obtaining the fundsnecessary for financing their operations. For small‐cap losers, the access to financing iseven more limited because of the firms’ relatively weak market presence (e.g., Gertler andGilchrist 1994; Thorbecke 1997). Continuing with the theory, when monetary conditionsare restrictive, lenders are likely to direct their limited capital to firms that have greatermarket recognition and more stability (large‐cap firms). For small‐cap losers, anexpansive environment represents an opportunity to gain access to much‐neededfinancing. The returns suggest that small‐cap losers are the greatest beneficiary of anexpansive monetary environment. For investors, an expansive environment reflectshigher expectations that small‐cap losers will survive and potentially recover from theirordeal, which encourages investors to bid up the stocks’ price from their depressed levels.Furthermore, the findings are consistent with the results of Jensen and Moorman (2010)that an environment characterized with greater fund availability provides investors andspeculators with an incentive to increase their allocation to equities that have lowerliquidity and higher margin requirements.

Our proposed theory implies a comparable scenario exists with respect to winnersin that small‐cap winners, relative to large‐cap winners, are likely to be more sensitive tothe potential for funding to diminish. Stocks that have experienced a prolonged period ofprice appreciation (winners) may become reliant on the availability of financing tocontinue their success. According to the theory, when the monetary environment isrestrictive, small‐cap winners, relative to large‐cap winners, face greater uncertaintyregarding the future availability of financing. Again, lenders are more likely to direct theirlimited capital to firms with more prominent positions in the industry. Investors reactwhen conditions are restrictive by reducing portfolio exposure to the small‐cap winners,which are likely to bemost reliant on a steady supply of accessible capital. Thus, the firms’stock prices are driven down from their elevated levels as investors reallocate their capital.

18 The Journal of Financial Research

Page 17: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

The evidence in Table 6 shows that high‐BEME losers reverse significantly, evenafter excluding January returns, but only during periods of expansive monetary policy.Once again, this finding is counter to the capital gains lock‐in hypothesis advocated byGeorge and Hwang (2007), which implies that losers only reverse during January. Thisevidence is, however, consistent with our contention that high‐BEME losers faceconsiderable problems in obtaining funding. According to our monetary‐based theory,lenders would be reluctant to lend to a firm that has experienced a prolonged period ofpoor stock performance and has a price multiple that suggests its prospects are weak.When monetary conditions are expansive, these financially distressed firms are likely tobe the biggest beneficiary of the more readily available financing. Recognizing this

TABLE 6. Fama–MacBeth Regressions by Book‐to‐Market and Monetary Conditions.

BEME b0 Loser Winner

Control Variables

Ln(ME) Beta

Panel A. All Periods

Low (growth) 0.01 0.14 �0.11 0.10 0.18t‐statistic 0.04 0.73 �1.39 2.28�� 0.73High (value) 1.62 0.03 �0.23 �0.17 0.29t‐statistic 5.79��� 0.22 �1.31 �3.32��� 0.91

Panel B. Expansive Conditions

Low (growth) 1.37 0.49 0.18 �0.04 0.26t‐statistic 2.08�� 1.34 1.42 �0.61 0.55High (value) 2.43 0.46 0.00 �0.33 0.78t‐statistic 4.98��� 1.94� 0.01 �3.85��� 1.26

Panel C. Restrictive Conditions

Low (growth) �0.60 �0.01 �0.24 0.17 0.15t‐statistic �1.41 �0.07 �2.57�� 3.01��� 0.55High (value) 1.25 �0.17 �0.33 �0.10 0.07t‐statistic 4.23��� �1.29 �1.70� �1.61 0.22

Note: Each month, stocks are grouped into quintiles based on continuously compounded returns between months–60 and –13. Independently, stocks are sorted into quintiles each month based on book‐to‐market equity (BEME)defined as in Fama and French (1992). Each month, we run Fama–MacBeth (1973) regressions of the form:

Rit ¼ b0t þ b1t Loseri þ b2t Winneri þ Control Variablesit þ eit ;

whereRit is themonthly return to stock i inmonth t.Winner (Loser) is a dummy variable that equals 1 if the stock is inthe winner (loser) quintile based on continuously compounded returns over months –60 to –13. Regarding thecontrol variables, ME is lagged one month relative to the return and Beta is computed using the CRSP value‐weighted index over the prior 60 months. The t‐statistics are based on the time‐series (Newey–West adjusted withfour lags) standard errors. The monetary conditions variable is determined based on monetary policy variableslagged one month relative to the returns. In the table, Loser and Winner (and Ln(ME) and Beta) represent thecorresponding (time‐series average) coefficient estimates (in %). January returns are excluded.���Significant at the 1% level.��Significant at the 5% level.�Significant at the 10% level.

The Monetary Environment 19

Page 18: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

possibility, investors bid up the price of the stocks to take advantage of a potentialturnaround. With respect to restrictive monetary environments, there is little evidence ofany significant reversal in performance for losers; however, winners with growthcharacteristics do reverse significantly. According to our theory, the restrictiveenvironment would be viewed as a detriment for growth firms to obtain the fundsnecessary to finance their recognized investment opportunities.

Finally, Table 7 reports reversal returns relative to monetary conditions and beta.The full‐period results indicate that significant return reversals are observed for both

TABLE 7. Fama–MacBeth Regressions by Beta and Monetary Conditions.

Beta b0 Loser Winner

Control Variables

Ln(ME) Ln(BEME)

Panel A. All Periods

Low 0.86 0.13 �0.05 0.03 0.22t‐statistic 3.41��� 0.82 �0.52 0.90 2.92���

High 1.94 0.35 �0.34 �0.11 0.30t‐statistic 2.78��� 2.00�� �2.92��� �1.74� 3.15���

Panel B. Expansive Conditions

Low 2.04 0.45 0.21 �0.09 0.15t‐statistic 4.56��� 1.53 1.23 �1.50 1.07High 3.35 1.05 �0.15 �0.26 0.29t‐statistic 2.67��� 3.29��� �0.79 �2.78��� 1.90�

Panel C. Restrictive Conditions

Low 0.32 �0.02 �0.17 0.09 0.26t‐statistic 1.15 �0.09 �1.38 2.12�� 2.89���

High 1.30 0.03 �0.42 �0.04 0.31t‐statistic 1.60 0.15 �2.85��� �0.48 2.70���

Note: Each month, stocks are grouped into quintiles based on continuously compounded returns between months–60 and –13. Independently, stocks are sorted into quintiles eachmonth based on beta, where beta is computed usingthe CRSP value‐weighted index over the prior 60 months. Each month, we run Fama–MacBeth (1973) regressionsof the form:

Rit ¼ b0t þ b1t Loseri þ b2t Winneri þ Control Variablesit þ ei;

whereRit is themonthly return to stock i inmonth t.Winner (Loser) is a dummy variable that equals 1 if the stock is inthe winner (loser) quintile based on continuously compounded returns over months –60 to –13. With regard to thecontrol variables, market capitalization (ME) is lagged onemonth relative to returns andBEME is defined as in Famaand French (1992). We add 1 to BEME before taking the natural log. The t‐statistics are based on the time‐series(Newey–West adjusted with four lags) standard errors. The monetary conditions variable is determined based onmonetary policy variables that are lagged one month relative to the returns. Returns are adjusted for delisting. In thetable, Loser andWinner (and Ln(ME) and Ln(BEME)) represent the corresponding (time‐series average) coefficientestimates (in %). January returns are excluded.���Significant at the 1% level.��Significant at the 5% level.�Significant at the 10% level.

20 The Journal of Financial Research

Page 19: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

losers and winners; however, the reversals are limited to those stocks that are mostsensitive to market conditions (high‐beta stocks). Therefore, even excluding Januaryreturns, we once again identify a significant loser reversal for a cross‐section of firms. Thefindings reinforce the prominent influence that monetary conditions have on long‐runreversals. High‐beta losers reverse significantly, but only during expansive conditions. Incontrast, high‐beta winners reverse only when conditions are restrictive.

There are three prominent reversal‐related observations reported in Tables 5, 6,and 7. First, the reversal phenomenon is dependent on the monetary environment; losersreverse only when monetary conditions are expansive and winners reverse only whenconditions are restrictive. Second, the reversal of losers is driven by the price moves ofsmall‐capitalization stocks, high‐BEME stocks, and stocks with high systematic risk, thatis, small, distressed, risky stocks. In contrast to the implications of the capital gains lock‐inhypothesis, even with January returns excluded, a significant loser reversal is identifiedwhen the monetary environment is expansive. Third, like the loser reversal, the winnerreversal is limited to stocks with particular features; specifically, small‐market‐cap,growth features and high systematic risk. Furthermore, winner reversals are isolatedwithin restrictive environments, which further support the crucial role that monetaryconditions play for the long‐run reversal phenomenon.

Overall, our evidence provides an explanation for Fama and French’s (1996)finding that the three‐factor model explains the long‐run reversal effect.10 Specifically,given the strong association we identify among long‐run reversals, the monetaryenvironment, and the three factors (firm size, beta, and firm BEME), one would expectthat controlling for these factors would greatly diminish the prominence of the reversalpattern. In general, according to our monetary‐based theory, the three characteristicscorrelate with a firm’s ability to obtain capital and with the level of financing available toinvestors and speculators. Our findings suggest that investors’ assessment of the value ofstocks exhibiting long‐run extreme price moves is conditional on the monetaryenvironment.

Firm Financial Constraints, Monetary Conditions, and Long‐Run Reversals

Hadlock and Pierce (2010) report evidence suggesting that firm size and firm age serve asthe most effective measures in identifying a firm’s level of financial constraints. Theauthors evaluate several alternative measures that have been advocated as indicators offinancial constraints and present evidence supporting the efficacy of firm size, firm age,and a composite measure composed of these two firm characteristics.

Based on the evidence presented by Hadlock and Pierce (2010), we evaluate therelation among long‐run reversals, monetary conditions, and Hadlock and Pierce’sadvocated measure of financial constraints (SA Index). We follow the format used tocreate Tables 5 through 7 and report the findings in Table 8. Once again, we excludeJanuary returns from the analysis and we control for the factors in the three‐factor model.

10Our findings are consistent with an explanation based on the availability of financing and are inconsistentwith the most prominent alternative explanations; however, the results do not preclude an alternative explanation.

The Monetary Environment 21

Page 20: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

Table 8 reports long‐run reversal returns relative to firm financial constraints (SAIndex) and monetary conditions. The findings indicate that long‐run reversals are limitedto firms with a high level of financial constraints. Once again, however, the findingsconfirm that the reversal pattern is conditional on the monetary environment. High‐constraint losers reverse significantly only when monetary conditions are expansive(Panel B), whereas a significant reversal of high‐constraint winners occurs only whenconditions are restrictive (Panel C).

Overall our findings suggest that investors view the fortunes of losers andwinners tobe dependent on monetary conditions, and thus, long‐run reversals are conditional on themonetary environment. Furthermore, we argue that the findings are consistent with amonetary‐conditions‐based explanation for the relation between firm characteristics and theprominence of the reversal pattern. In particular, long‐run reversals are shown to be more

TABLE 8. Fama–MacBeth Regressions by Firm Constraints (SA Index) and Monetary Conditions.

SA Index b0 Loser Winner

Control Variables

Ln(BEME) Ln(ME) Beta

Panel A. All Periods

Low 1.23 �0.30 �0.15 0.08 �0.04 0.06t‐statistic 3.20��� �1.72� �1.42 0.59 �1.04 0.23High 1.23 0.04 �0.21 0.84 �0.42 0.46t‐statistic 3.04��� 0.25 �1.64 4.20��� �4.57��� 1.70�

Panel B. Expansive Conditions

Low 2.34 �0.06 0.03 �0.13 �0.16 0.48t‐statistic 3.89��� �0.19 0.18 �0.62 �2.66��� 0.85High 2.47 0.47 0.03 0.49 �0.64 0.65t‐statistic 3.59��� 2.08��� 0.15 1.75� �3.68��� 1.28

Panel C. Restrictive Conditions

Low 0.72 �0.40 �0.23 0.18 0.02 �0.13t‐statistic 1.82� �1.81� �1.66 1.02 0.37 �0.44High 0.67 �0.16 �0.32 1.00 �0.32 0.37t‐statistic 1.46 �0.97 �2.09�� 4.12��� �3.12��� 1.25

Note: Each month, stocks are grouped into quintiles based on continuously compounded returns between months–60 and –13. Independently, stocks are sorted into quintiles based on financial constraints (SA), computed followingHadlock and Pierce (2010). Each month, we run Fama–MacBeth (1973) regressions of the form:

Rit ¼ b0t þ b1t Loseri þ b2t Winneri þ Control Variablesit þ ei;

whereRit is themonthly return to stock i inmonth t. Loser (Winner) is a dummy variable that equals 1 if the stock is inthe lowest (highest) quintile based on long‐run past returns. Control variables are defined in Table 7. The t‐statisticsare based on the time‐series (Newey–West adjusted with four lags) standard errors. The monetary conditions arelagged one month relative to returns. January returns are excluded from the regression.���Significant at the 1% level.��Significant at the 5% level.�Significant at the 10% level.

22 The Journal of Financial Research

Page 21: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

pronounced for firms that we characterize as being more reliant on fund availability, forexample, small firms, high‐risk firms, and financially constrained/distressed firms. As notedby Fama and French (1996) and others, losers tend to be firms that are financially distressedto some degree. During periods when the monetary environment is restrictive, such firms arelikely to have limited access to funds. According to our proposedmonetary‐conditions‐basedtheory, funding problems are likely to be exacerbated for firms that are small, risky, orcapitally constrained. Thus, for such firms, favorable monetary conditions offer relief fromconcerns about the availability of financing. An expansive environment offers comfort toinvestors that financing will be available to loser firms to keep their operations going. Inaddition, a favorable monetary environment provides investors with more confidence inholding low‐liquidity/high‐margin equities. Thus, when the monetary environment isexpansive, investors are more optimistic about losers, which causes their price to rebound.

Likewise, our proposed theory implies that the monetary environment has acomparable influence onwinners.Winners have experienced strong performance over thepast several years and have raised investor expectations. For such firms, a period ofmonetary restraint may be viewed as an impediment to continued success because of lessreadily available financing. Investors become less confident that winners will be able toobtain the necessary funds to continue financing the firms’ opportunities; this isparticularly problematic for firms that have limited access to alternative sources of capital.Thus, the prospect of more scarce future financing causes investors to discount the valueassigned to the firms’ opportunities, which results in a drop in their stock price.

V. Summary and Conclusions

The existence of a reversal in stock prices for firms that experienced the best, and worst,performance over the previous several years was first documented by DeBondt and Thaler(1985). Subsequently, this return pattern became known as the long‐run price reversal. Long‐run price reversals were originally attributed to irrational investor pricing decisions, but morerecently have been explained as the consequence of tax‐based trading decisions.Motivated byprevious empirical evidence and theoretical arguments, we evaluate the influence thatmonetary conditions have on long‐run price reversals.We find strong evidence supporting anexplanation for the long‐run reversal phenomenon that is based on themonetary environment.

We start by establishing the significance of the long‐run reversal pattern that hasbeen identified by previous researchers. Specifically, we confirm that stock performancereverses for firms that have experienced weak performance over the past five years(losers) and for firms showing strong performance over the previous five years (winners).We also confirm that January plays a prominent role in the long‐run reversal of losers. Infact, our initial findings support George and Hwang (2007) who conclude that, outside ofJanuary, there is no reversal of losers.

A major contribution of our research is to show that the long‐run reversal patternis conditional on the monetary environment. In particular, we show that, even outside ofJanuary, there is a strong reversal of losers when monetary conditions are expansive, andthe loser reversal is trivial when the monetary environment is restrictive. In contrast,winners are shown to only reverse when monetary conditions are restrictive. These

The Monetary Environment 23

Page 22: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

findings are inconsistent with the prevailing explanation for long‐run reversals (the capitalgains lock‐in hypothesis) and are also counter to the irrational investor behaviorexplanations; however, the findings correspond with an explanation based on themonetary environment. Furthermore, consistent with our monetary‐based explanation ofreversals, we find that the reversal pattern is more pronounced for firms that havecharacteristics that suggest they are more reliant on external financing and for firms withdiminished access to external funding sources.

According to the monetary‐conditions‐based theory that we propose, firms thathave experienced a prolonged period of poor stock performance (losers) become starvedfor financing. When monetary conditions are restrictive, market participants are reluctantto offer losers financing because of their poor performance. The situation is exacerbatedfor losers that are small or financially distressed. Lenders are likely to withhold scarcefunding from such firms in favor of the more dominant, stable firms in the industry. Inaddition, investors shun these firms in favor of their larger more stable peers because theirstocks are less liquid and have higher margin requirements. An expansive monetaryenvironment is most beneficial for these funding‐rationed losers. During expansiveenvironments the firms have the opportunity to access needed financing, which iswithheld during restrictive environments. When funding is expansive, investors areencouraged to reallocate funds into losers to take advantage of their depressed prices.

With respect to winners, our proposed theory contends that firms that haveexperienced a prolonged advance in stock price have raised investor expectations for furtherexpansion in firm operations and stock price. For such firms, a restrictive monetaryenvironment confines future opportunities for expansion. Small firms (financiallyconstrained firms) are discounted during an environment that suggests future financingmay be limited due to their limited borrowing capacity and greater reliance on externalsources of capital. These firms suffer to a greater degree when the funds required to maintaintheir superior performance are less assured. Investors, recognizing that the firms facesignificant obstacles in obtaining funding at attractive rates, allocate scarce funds elsewhere.Furthermore, investors discount the financially constrained winners’ accumulatedinvestment opportunities, which are viewed as a riskier component of firm value.

Our empirical evidence supports our proposed monetary‐based theory as anexplanation for the long‐run reversal phenomenon. The strong association that wedocument between the monetary environment and the reversal pattern is inconsistent withother theories that have been offered to explain the reversal pattern.

References

Banko, J., M. C. Conover, and G. Jensen, 2006, The relationship between the value effect and industry affiliation,Journal of Business 79, 2595–2616.

Barberis, N., A. Shleifer, and R. Vishny, 1998, A model of investor sentiment, Journal of Financial Economics 49,307–43.

Berk, J. B., R. C. Green, and V. Naik, 1999, Optimal investment, growth options, and security returns, Journal ofFinance 54, 1553–1607.

Bernanke, B. S., and A. Blinder, 1988, Credit, money, and aggregate demand, American Economic Review 78,435–39.

24 The Journal of Financial Research

Page 23: THE MONETARY ENVIRONMENT AND LONG-RUN REVERSALS IN STOCK RETURNS

Bernanke, B. S., and A. Blinder, 1992, The federal funds rate and the channels of monetary transmission, AmericanEconomic Review 82, 901–21.

Bernanke, B. S., V. R. Reinhart, and B. P. Sack, 2004. Monetary policy alternatives at the zero bound: An empiricalassessment, Finance and Economics Discussion Series, Federal Reserve Board, Washington, DC.

Booth, J. R., and L. C. Booth, 1997, Economic factors, monetary policy, and expected returns on stocks and bonds,Economic Review: Federal Reserve Bank of San Francisco 32–42.

Brav, A., and J. B. Heaton, 2002, Competing theories of financial anomalies, Review of Financial Studies 15, 575–606.

Brunnermeier, M. K., and L. H. Pedersen, 2009, Market liquidity and funding liquidity, Review of Financial Studies22, 2201–38.

Chan, L. K. C., and N. Chen, 1991, Structural and return characteristics of small and large firms, Journal of Finance46, 1467–84.

Chopra, N., J. Lakonishok, and J. R. Ritter, 1992, Measuring abnormal performance: Do stocks overreact? Journalof Financial Economics 31, 235–68.

Chordia, T., A. Sarkar, and A. Subrahmanyam, 2005, An empirical analysis of stock and bond market liquidity,Review of Financial Studies 18, 85–129.

Cooper, M. J., R. C. Gutierrez, and A. Hameed, 2004, Market states and momentum, Journal of Finance59, 1345–66.

Daniel, K., D. Hirshleifer, and A. Subrahmanyam, 1998, Investor psychology and investor security market under‐and overreactions, Journal of Finance 53, 1839–86.

DeBondt, W. F. M., and R. Thaler, 1985, Does the stock market overreact? Journal of Finance 40, 793–808.DeBondt, W. F. M., and R. Thaler, 1987, Further evidence on investor overreaction and stock market seasonality,

Journal of Finance 42, 557–81.Fama, E. F., and K. R. French, 1992, The cross‐section of expected stock returns, Journal of Finance 47, 427–65.Fama, E. F., andK. R. French, 1996,Multifactor explanations of asset pricing anomalies, Journal of Finance 51, 55–

84.Fama, E. F., and J. D. MacBeth, 1973, Risk, return, and equilibrium: Empirical tests, Journal of Political Economy

81, 607–36.Fujimoto, A. 2004. Macroeconomic sources of systematic liquidity, University of Alberta Working Paper.George, T. J., and C. Hwang, 2004, The 52‐week high and momentum investing, Journal of Finance 59, 2145–76.George, T. J., and C. Hwang, 2007, Long‐term return reversals: Overreaction or taxes? Journal of Finance 62,

2865–96.Gertler, M., and S. Gilchrist, 1994,Monetary policy, business cycles, and the behavior of small manufacturing firms,

Quarterly Journal of Economics 109, 309–40.Grinblatt, M., and T. J. Moskowitz, 2004, Predicting stock price movements from past returns: The role of

consistency and tax‐loss selling, Journal of Financial Economics 71, 541–79.Hadlock, C. J., and J. R. Pierce, 2010, New evidence on measuring financial constraints: Moving beyond the KZ

index, Review of Financial Studies 23, 1909–40.Hong, H., and J. Stein, 1999, A unified theory of underreaction, momentum trading and overreaction in asset

markets, Journal of Finance 54, 2143–84.Jensen, G. R., J. M. Mercer, and R. R. Johnson, 1996, Business conditions, monetary policy and expected security

returns, Journal of Financial Economics 40, 213–37.Jensen, G. R., and T. Moorman, 2010, Inter‐temporal variation in the illiquidity premium, Journal of Financial

Economics 98, 338–58.Klein, P., 1999, The capital gain lock‐in effect and equilibrium returns, Journal of Public Economics 71,

355–78.Lakonishok, J., A. Shleifer, and R. Vishny, 1994, Contrarian investment, extrapolation, and risk, Journal of Finance

49, 1541–78.Lewellen, J., and J. Shanken, 2002, Learning, asset‐pricing tests, and market efficiency, Journal of Finance 57,

1113–45.Patelis, A., 1997, Stock return predictability: The role of monetary policy, Journal of Finance 52, 1951–72.Perez‐Quiros, G., and A. Timmermann, 2000, Firm size and cyclical variations in stock returns, Journal of Finance

55, 1229–62.Thorbecke, W., 1997, On stock market returns and monetary policy, Journal of Finance 52, 635–54.Thornton, D. L., 1998, The information content of discount rate announcements: What is behind the announcement

effect? Journal of Banking and Finance 22, 83–108.Vassalou, M., and Y. Xing, 2004, Default risk in equity returns, Journal of Finance 59, 831–68.Whited, T. M., and G. Wu, 2006, Financial constraints risk, Review of Financial Studies 19, 531–59.

The Monetary Environment 25