market janurary

7/25/2019 market janurary

1/18

Please cite this article in press as: Easterday, K. E., & Sen, P.K. Is the January effect rational? Insights from the accounting valuation

model. TheQuarterly Reviewof Economics and Finance (2015), http://dx.doi.org/10.1016/j.qref.2015.05.001

ARTICLE IN PRESSG Model

QUAECO-852; No.of Pages18

The Quarterly Review of Economics and Finance xxx (2015) xxxxxx

Contents lists available at ScienceDirect

The Quarterly Review ofEconomics and Finance

journa l homepage: www.elsevier .com/ locate /qref

Is theJanuary effect rational? Insights from the accounting valuation

model

Kathryn E. Easterday a,, Pradyot K. Sen b

a Wright StateUniversity,UnitedStatesb University of Washington-Bothell, UnitedStates

a r t i c l e i n f o

Article history:

Received 30 May 2014Received in revised form 23 April 2015

Accepted 18 May 2015

Available online xxx

Keywords:

January effect

Permanent earnings

Tax-loss selling

a b s t r a c t

Employing a permanent earnings valuation model and a novel sample partition, we find evidence that the

January effect anomaly is consistent with rational economic market behavior. Investors in firms which

experienceJanuary effect return premiums appear to discount first quarter earnings performance but

reward permanent earnings and expectations offuture improvements. Our evidence also supports a tax-

loss selling explanation for theJanuary effect. We find that theJanuary effect is experienced by relatively

few firms in the sample overall, but a substantial percentage ofJanuary effect firms are identified as

potential tax-loss sellers. Our results complement prior research suggesting that the January effect is

neither a result of irrational noise traders nor consistent with systemic risk factor explanations. Our

study reconciles the assumption ofarbitrage inherent in trading studies with a fundamental accounting

valuation approach and offers some further insights into the nature ofthis market phenomenon.

2015 The Board ofTrustees ofthe University ofIllinois. Published by Elsevier B.V. All rights reserved.

1. Introduction

This paper finds that the January effect anomaly is associated

with accounting earnings and expectations about future earn-

ings, in a manner both economically rational and consistent with

accountingtheory. This workextends thatofHenkerand Debapriya

(2012), who argue against an irrational noise trader explana-

tion for the January effect. It complements that of Klein and

Rosenfeld (1991), who present evidence that the January effect can

be explained at least in part by new information in January about

upcoming earnings announcements. Finally, our accounting valua-

tion approach complements Mashruwala and Mashruwala (2011),

who argue that return seasonality is incompatible with systemic

risk explanations.Fama (1998) and Gerlach (2007, 2010) both argue that many

so-called market anomalies are tenuous in the sense that they are

sensitive to the methodologies used to detect or measure them.

Far from being tenuous, the January effect a capital markets

Corresponding author at: Wright State University, Raj Soin College of Business,

Department of Accountancy, 298 Rike Hall, 3640 Col. Glenn Highway, Dayton, OH

45435-0001, United States. Tel.: +1 937 775 3304.

E-mail addresses: [email protected](K.E. Easterday),

[email protected] (P.K. Sen).

phenomenon in which return premiums are on average higher

in January than in other months of the year1 persists in defi-

ance of economic theory which says it should be arbitraged away.

Although some studies suggest that the January effect is disappear-

ing (Gu, 2003; He & He, 2011; Hensel & Ziemba, 2000; Mehdian &

Perry, 2002), numerous others provide evidence that the January

effectcontinues toappear in modern US capital markets (Anderson,

Gerlach, & DiTraglia, 2007; Brown & Luo, 2006; Ciccone, 2011;

Dzhabarov & Ziemba, 2010; Easterday, Sen, & Stephan, 2009; Haug

& Hirschey, 2006; Mashruwala& Mashruwala, 2011; Ziemba, 2011)

although it does not occur every year (Easterday et al., 2009;

Hulbert, 2008).

Tax management is the most common rationalization for the

January effect: investors take advantage of capital losses at yearend for tax purposes, resulting in temporary downward mispric-

ings that create large January returns when prices rebound after

the turn of the year (Branch, 1977; Brown, Ferguson, & Sherry,

2010; Chen & Singal, 2004; Dalton, 1993; Givoly & Ovadia, 1983;

Griffiths & White, 1993; Grinblatt & Keloharju, 2004; Jones, Lee, &

Apenbrink, 1991; Koogler & Maberly, 1994; Ma, Rao, & Weinraub,

1988; Phua, Chan, Faff, & Hudson, 2010; Sikes, 2014; Starks, Yong,

1 Some researchers call it theturn of theyear effect. Both termsare widelyused

throughout the literature, often interchangeably.

http://dx.doi.org/10.1016/j.qref.2015.05.001

1062-9769/ 2015 The Board of Trustees of theUniversity of Illinois. Published by Elsevier B.V. All rightsreserved.
http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001http://www.sciencedirect.com/science/journal/10629769http://www.elsevier.com/locate/qrefmailto:[email protected]:[email protected]://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001mailto:[email protected]:[email protected]://www.elsevier.com/locate/qrefhttp://www.sciencedirect.com/science/journal/10629769http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001


2/18





2 K.E. Easterday, P.K. Sen / TheQuarterly Reviewof Economics andFinance xxx (2015) xxxxxx

& Zheng, 2006).2 However, there is evidence that tax minimizing

behavior by itself is not enough to drive the January effect (Brown,

Keim, Kleidon, & Marsh, 1983; Fountas & Segredakis, 2002; Jones &

Wilson, 1989; Pettingill, 1989; Reinganum, 1983; Ritter, 1988; Sias

& Starks, 1997; van den Bergh & Wessels, 1985).

Rather than attempting to explain the January effect,

Mashruwala and Mashruwala (2011) exploit this seasonal

increase in stock prices to examine whether accruals quality

measures proxy for information risk. Their findings suggest thatsuch measures proxy more for firm attributes associated with

tax-loss selling than for information risk. Studies by Brauer and

Chang (1990), Peterson (1990), and Reinganum and Gangopadhyay

(1991) also provide evidence that information risk is not related to

the January effect.

Extant research into the January effect itself appears primarily

in the finance literature where it is often explained as a temporary

mispricing anomaly resulting from various market inefficiencies

and risks resulting in arbitrage opportunities. However, some stud-

ies (Loughran, 1997; Mashruwala & Mashruwala, 2011; Roll, 1983;

Seyhun, 1993; Tinic & West, 1984) argue that systemic risk factor

explanations are not compatible with seasonal market behaviors.

A recent study by Henker and Debapriya (2012) provides evidence

that the January effect is not driven by irrational noise traders.

CAPMmodelsneither predict norexplain risk,especially (or only)in

January (Best, Hodges, & Yoder, 2006; Corhay, Hawawini, & Michel,

1987; Gultekin & Gultekin, 1987; Kryzanowski & Zhang, 1992;

Ritter & Chopra, 1989; Thaler, 1987).

Insights from the fundamental valuation theory of accounting

suggest that under a no arbitrage condition, returns in January

or any time period should be positively associated with

contemporaneous accounting earnings and information affecting

expectations about future accounting performance (Feltham &

Ohlson, 1995; Ohlson, 1995, 2001). With the exception ofKlein

and Rosenfeld (1991) there is little research considering how or

whether the January effect anomaly is associated with account-

ing earnings information in a market-efficient, rational economic

manner.3 Their evidence shows that low-PEstocks with lowannual

earnings forecasts in December outperform in January relativeto other low-PE stocks and they argue that the prices of these

stocks rise in January because it becomes apparent to investors

then that actual earnings for the just-completed year will be bet-

ter than was forecasted in December. Their analyses employ a

trailing-earnings-to-price ratio and focus on earnings forecasts and

investors expectations for the earnings announcement for the year

immediately past.

We extend Klein and Rosenfeld (1991) by employing a form

of the permanent earnings model developed in Easterday, Sen,

and Stephan (2011)4 to examine the association between January

returns and earnings in the first quarter. The ESS model expresses

returns as a function of contemporaneous earnings level, earn-

ings growth, and a term representing the sustainability of earnings

growth, and the model derives directly from Ohlsons (1995, 2001)valuation framework. Employing an accounting valuation model

rather than an ad hoc trading model enables us to forgo an assump-

tion of arbitrage and demonstrate that, consistent with economic

and accounting theory, earnings information plays an important

role in the economic intuition of the January effect phenomenon.

2 Additional studies focus on the January effect and its relation to tax rules for

individual and institutional investors (Lynch, Puckett, & Yan, 2014; Poterba and

Weisbenner, 2001; Slemrod, 1982).3 Lakonishok, Shleifer, and Vishny (1994) includecurrent P/Eratioas oneindica-

tor of possiblemispricing but their study does not examine fundamental valuation

implications of accountingearnings.4

Hereafter, ESS.

Our model is consistent with the notions that (1) earnings change

not earnings level captures the permanent componentof earnings

(Ali & Zarowin, 1992; Ohlson & Shroff, 1992); and (2) information

about future earnings is essential to valuationbecause it adjusts for

transitorycomponents of current earnings. Valuation depends crit-

ically on permanent earnings (Pan,2007), as well as theexpectation

that permanent earnings will be sustained into the future. In addi-

tion,eschewing a CAPMapproachavoids the uncertainties inherent

in estimating required rates of return, a feature of valuation based

on asset pricing models.5

In order to examine the association of these anomalous returns

with accounting earnings information we introduce an innovative

sample partition, forming an ex postcategorization of firms that

experience the January effect (JE firms) versus those that do not

(NJEfirms).6 Thus,we specifically identifyfirms thatexhibitJanuary

effect return premiums rather than relying on some firm charac-

teristic(s) presumed to be associated with the January effect. NJE

firms act as a kind of comparison group; but under our model and

approach, evidence of rational economic behavior in one group

does not negate or preclude rational economic behavior in the

other.

We find that JE firms represent approximately nine percent of

all firms in our sample andrange across all market caps, suggesting

that theJanuaryeffectis drivenby relatively fewfirmsoverall andis

frequentlybut notexclusivelya smallfirm phenomenon. Our JE/NJE

partitiondelivers someintriguing results whenimplemented in our

valuation model.

For JE firms the coefficient on first quarter contemporaneous

earnings level is significantly negative, while the coefficients for

contemporaneous earnings growth and expectations for future

earnings growth remain significantly positive. Although a nega-

tive earnings level coefficient may seem at first irrational, it may

indeed be consistent with rational behavior. First and most impor-

tantly, our valuation model is more comprehensive in that it does

not rely only on current or past earnings information, but includes

all other information as captured by the construction of the ana-

lysts forecast variable. The inclusion of the term for information

about expected future earnings captures the reality that marketdecisions are based in large part on expectations for the future.

Second, it is well established that price leads earnings (Ball &

Brown, 1968; Beaver, Lambert, & Morse, 1980; Beaver, Lambert,

& Ryan, 1987; Collins, Kothari, Shanken, & Sloan, 1994; DeBondt

& Thaler, 1985, 1987; Kothari, 2001). We contend that poor year

end returns followed by superior January returns foreshadow poor

first quarter earnings followed by an earnings improvement. Our

examination of earnings for the quarters immediately preceding

and following the first quarter, as well as a correlation analysis of

sequential quarterly earnings, both support this contention. These

results are also compatible with those ofBeaver et al. (1980) and

DeBondt and Thaler (1987).7

Third, we arguethat permanentearningsand theirsustainability

should be quite relevant to higher return premiums in January,a proposal in keeping with the tax-loss trading explanation for

the January effect advanced in so many prior studies. Firms are

5 Penman (2004, p. 96) reminds us that a capital asset pricing model (CAPM)

generates a requiredrateof return,not asset value.Further,valuationmodelsrelying

on estimated rates of return canbe highly sensitive to theunderlying assumptions

used in theCAPM.6 If a firms January return premium is the highest of all 12 months of the year

then it is classified as a JE firm forthat year. Otherwise, thefirm is categorized as

NJE.7 Beaver et al. (1980) demonstrate that returns are positively associated with

earnings of the following period. DeBondt and Thaler (1987) present evidence that

earnings improve in subsequent periods for loser firms. They also observe that

January and December return premiums are negatively associated.
http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001


3/18





K.E. Easterday, P.K. Sen / TheQuarterly Review of Economics andFinance xxx (2015) xxxxxx 3

tax-loss sellers when their prices have fallen, i.e., they are losers,

and investors sell them off at year end to capture capital losses for

income tax purposes. If a low December stock price drives the high

January return and also foreshadows poor first quarter earnings,

then it is natural that (increased) January return and first quarter

earnings, however transient, should be negatively correlated. The

evidence in our added correlation analysis also indicates that the

unannounced earnings in January related to the year t1 are pos-

itively correlated with the first quarter earnings for JE firms, but

not for others. Thus the bad news embedded in these two earnings

numbers seems to be captured in the lowered December return of

the previous year. The idea that JE firms possibly tax-loss sell-

ing losers in the prior year offer investors optimistic expectations

for future sustainability of permanent earnings is not unreason-

able if we admit the possibility that bargain hunting investors in

loser firms focus more on critical value generation capabilities in

the long run and less on currentearnings numbers that may reflect

transitory components. Given their weak (strong) near term (long

term) prospects these firms become candidates for tax-loss trans-

actions in December. Then their repurchase in January increases

their price hence, the high January returns. Evidence for such a

scenario can be found in Beyer, Garcia-Feijoo, and Jensen (2013),

who show that a trading strategy targeting small, outof favor firms

achieves superior return performance in January.

It is true that without the connections between tax-loss sell-

ing and repurchasing, and the notion that poor December returns

foreshadow short term earnings troubles, the negative correlation

between January returns and first quarter earnings may appear

at first glance to be irrational. However, our scenario proposed

above provides a reasonable explanation for why we see evidence

of tax-loss selling intertwined with the (apparently) counterintu-

itiveresult of negative correlation between high January return and

poor first quarter earnings performance.

Our valuation model, which anchors on both contemporane-

ous permanent earnings and expectations for their sustainability,

coupled with our sample partition that isolates JE firms from NJE

firms, offers the opportunity to examine whether an accounting

earnings valuation approach that does not admit arbitrage canprovide someinsight intoboth the observed return premiumschar-

acteristic of the January effect and the tax-loss selling hypothesis

for them.

An effective JE/NJE partition should result in a considerable

proportion of JE firms also being identifiable as probable tax-loss

sellers. And if January effect returns are rationally associated with

information about accounting earnings, then JE firms that are tax-

loss sellers could be expected to have more emphasis placed on

permanent earnings and expectations for the sustainability of per-

manent earnings in the future, and less emphasis on current (i.e.,

first quarter) earnings that are likely to be poor.

Following Dalton (1993) we identify firms in our sample whose

previous endof yearspriceperformancemakesthem likelytax loss

sellers, andthen implementour model usingpartitions for bothtax-loss selling candidacy and occurrence of the January effect. About

45% of our samples JE firms were tax-loss selling candidates at the

endof thepreviousyear,suggesting that although tax-loss selling is

an important market dynamic in understanding the January effect,

other factors likely play in as well.

Our partitions once more deliver interesting results. The sig-

nificantly negative coefficient on earnings level appears only for

JE tax-loss sellers; the permanent earnings coefficients are signifi-

cantly positive and approximately four times larger in magnitude

than the coefficient values for either NJE tax-loss sellers or any

non-tax-loss sellers. The explanatory value of our earnings model

increases by a factor of approximately 10 when we implement the

JE/NJE partition on our tax-loss selling firms. We interpret these

results as additional evidence that the January effect anomaly is

linked to economically rational market behavior: firms that are ex

postidentified as poor price performers at the end of the year (i.e.,

tax-loss selling candidates), but whose future earnings outlook is

expected to improve, are rewarded by investors.

Robustness testing for other quarters of the year provides sim-

ilar results, but they are much more pronounced for January than

for other first months of quarter (April, July, or October). Over-

all, we interpret this as support for the concepts that the January

effect anomaly appears to be a rational economic response to value

relevant accounting earnings information; that earnings levels do

not capture permanent earnings; and that the market focuses on

andrewards valuation implications of permanentearnings that are

expected to be sustainable into the future.

This study contributes to the literature in four ways. First, we

extend both Klein andRosenfeld (1991) and Henker and Debapriya

(2012)by using a valuationmodel ratherthana trading approach.In

doing so we findevidence that the January effectis linkedto funda-

mental valuation information represented in permanent earnings

and expectations for earnings growth. Second, we complement the

workofbothMashruwalaand Mashruwala(2011)and DeBondt and

Thaler (1987) by investigating tax-loss selling firms and demon-

strating that the January effectis related to information captured in

accrual accounting. Our results offer additional support for the tax

management story in a December year end tax environment such

as the US, yet are also consistent with the argument that there is

more to the January effect than just tax-loss selling (Bley & Saad,

2010; Brown et al., 1983; Corhay et al., 1987; Fountas & Segredakis,

2002; Heston & Sadka, 2010; Su, Dutta, Xu, & Ma, 2011). Third, we

extend Penman (1987) by testing the association between returns

and quarterly earnings performance without relying on an ex-post

categorization of earnings as good news or bad news.8 Finally,

partitioning our sample between firms that experience a January

effect and those that do not is an innovation that offers a more

precise inquiry into the nature of this market phenomenon.

The study proceeds as follows. Section 2 explains the develop-

ment of our empirical model. Analysis and results are in Section 3.

Section 4 discusses robustness testing. Section 5 concludes.

2. Theoreticalmodel andempirical application

A rigorous theoretical discussion of the link between price and

accounting earnings is provided in Feltham and Ohlson (1995) and

Ohlson (1995), beginning with the following assumptions:

I. The value of the firm is equal to the present value of future

expected dividends (PVED).

Pt=

=1

REtdt+

(1)

Pt= price atdate t;dt= netdividends paid at date t;R= 1 + r= the dis-

count rate plus one;Et[.] = the expected valueoperator, conditionedon information at date t.

IICleansurplus accounting. That is,change in book value is equal

to earnings less dividends:

bt= bt1 +xt dt. (2)

Using (2) to substitute recursively for the dividend term in

(1) yields the abnormal earnings model which expresses PVED as

current book value plus capitalized abnormal earnings, defining

8 Penman(1987) appearsto acknowledge a potentiallydistorting effect of January

returns in first quarter data and focuses his analysis on the other three quarters of

the year.


4/18






abnormal earnings as accounting earnings less a charge for use of

capital.

Pt= bt+

=1

REt

xat+, provided that Et

bt+

/R 0

as (3)

bt= book value at date t;xt= accounting earnings during the period

t;xat xt (R 1)bt1 = abnormal earnings.III. Appeal to a first order autoregressive process for abnormal

earnings and information other than current abnormal earnings.

(Ohlson, 1995 refers to this as assumption [A3].)

xat+1 = xat+ t+ 1,t+1

t+1 = t+ 2,t+1

t= other value relevant information; , = parameters known bythe market but unknown to researchers.

Combining (3) and [A3] yields an expression defining firm value

as a function of book value, abnormal earnings, and other informa-

tion notyet captured in earnings butrelevant forforecasting future

earnings:

Pt= bt+ 1xat+ 2t (4a)

which can be expressed equivalently as

Pt= (1 k)bt+ k (xt dt)+2t (4b)

1 = /(R) 0

2 = R/(R)(R ) > 0

k = r/(R) 0

= R/r

Eq. (4b) presents a challenge to empirical researchers because

other value relevant information, t, is not observable. A commonempirical approach is to assume that t is equal to zero (Easton &

Harris, 1991; Easton, Harris, & Ohlson, 1992;Penman & Sougiannis,

1998 are three well known examples). Ohlson (2001) warns that

although this assumption is analytically convenient it may be

overly simplistic. ESS exploit Ohlsons assumption that earnings

expectations are observable in analysts forecasts to show that in

returns form, t can be captured as the difference between thechange in forecasted future earnings and contemporaneous change

in earnings:

RETt=

1

Pt1

1xt+

2+

3r

xt+

3r

xt+1t xt

+ 3dt+ 2dt1

(5)

RETt= Pt Pt1 + dt

Pt1

xt=xtxt1

xt+1t =xf(t+1)t x

f(t)t1 = the forecast in period t for earnings per

share of period t+ 1, minus the forecast in period t1 for earn-ings per share of period t; Pt1 =price per share at beginning of

period t.

1 = R(1)(1 )/(R)(R )

2 = r/(R)(R )

3= Rr/(R)(R )

1 +2 +3 = 1

The problematic term tis thus transformed into terms that areall readily observable and measurable. Both the second and third

terms on the right hand side are divided by the cost of capital, con-

sistent with evidence in Ali and Zarowin (1992) and Ohlson and

Shroff (1992) that earnings changes capture permanent earnings.

Note that the term xt+1t =xf(t+1)t x

f(t)t1

does not represent a fore-

cast revision, but rather the difference between future earningsforecasts for two consecutive periods.9 Because analysts forecast

total earnings including any transitory components of earnings

the third term provides a correction for the portion of current

earnings change that may notbe permanent. Returns increase with

larger earnings changes from the prior period [second expression

in Eq. (5)]. When these changes are accompanied by an expecta-

tion that future earnings growth will be of even greater magnitude

[third expression in Eq. (5)], an additional premium is placed on

the realized earnings change. If the quantityxt+1t xt

is neg-

ative, i.e., earnings growth is not judged sustainable or a decline

is expected to accelerate, then current period realized earnings

change is discounted.

ESS show for both annual and quarterly time frames that this

expressionof the otherinformationvariablesubstantiallyimprovestheexplanatorypowerof thereturns model relative to theassump-

tion that t is equal to zero. They also provide evidence that the

dividend terms can be ignored without sacrificing explanatory

power. In addition, they show that contemporaneous measure-

ment of returns and earnings is both theoretically and empirically

appropriate and that proper specification of the other information

variable in the returns model removes the need for ad hoc control

variables. Based on their derivations and results the basic form of

our empirical model is as follows:

Rt= 0 + 1xtPt1

+ 2xt xt1

Pt1+ 3

xf(t+1)t x

f(t)t1

(xt xt1)

Pt1(6)

t= t ime period of interest; Rt= return in period t, computed

using CRSP monthly holding returns; Pt1 = price per share

at beginning of period t; xt= earnings per share for quar-

ter t; 0 = intercept; 1 = 1; 2 = 2+ 3

r ; 3 = 3

r ; xf(t+1)t =

the latest earnings forecast for period t+ 1 made during period

t;xf(t)t1= the latest earnings forecast for period t made during

period t 1.

Because ourfocus is onthe January effect, ourtests ideally would

be carried out by mapping monthly returns into earnings of the

same period. Of course this is not possible because firms report

earnings information on quarterly and annual bases, not monthly.

Thenextbestcandidateistomap monthly returns to corresponding

quarterly earnings. This modification fits well with the idea that

prices lead earnings, and to the extent that earnings of February

and March (May and June, August and September, November and

December) are also included in earnings of the first (second, third,

9 ESSdiscussthispointin detailand it is presentedpictoriallyin their Fig.1 (page

1131).


5/18






Fig. 1. Timeline of eventsand measurementpoints forquarterly returns, earningsand forecastsas modeled in Eq. (7). Rqn is holding return duringthe quarter.m1 (m2, m3)

is thefirst (second, third)monthof quarter n andxqnis earningsper share in quartern.xf(n+1)qn is thelatest analyst forecast forquartern+ 1 earningsper sharethat comes out

in quarter n (denoted by theshort solid vertical marker), and is after theannouncement of quarter n1 earnings (denoted by the short dashed vertical marker).xf(n)q(n1)

is the

latest analyst forecast for quartern earningsper sharethat comes outin quarter n1 after theannouncement of quarter n2 earnings.

fourth) quarter there is a bias against finding an effect for returns

of the first month only.

There is strong evidence that firm size is negatively correlated

with the magnitude of the January effect (Blume & Stambaugh,

1983; Easterday et al., 2009; Haug & Hirschey, 2006; Hensel &

Ziemba, 2000; Keim,1983; Lamoureux& Sanger, 1989; Reinganum,

1983), and we add firm size as a control variable.10 The general

empiricalform of ourquarterly data focused, expanded model thus

becomes

Rqn = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+3

x

f(n+1)

qn x

f(t)

q(n1)xqn xq(n1)

Pq(n1)

+ 4Size (7)

qn= quarter, n= 1, 2, 3, 4; Rqn = r eturn in quarter n, com-

puted using CRSP monthly holding returns; Pq(n1) =price

per share at the end of quarter (n1); xqn =earnings per

share in quarter n; 0 = intercept; 1 = 1; 2 = 2+ 3

r ;

3 = 3

r ; xf(n+1)qn = the latest earnings forecast for quarter n+

1 made during quarter n;xf(n)q(n1)

= the latest earnings forecast

for quarter n made during quarter n 1; Size= l ogarithm of firm assets at the close of quarter n.

Although we focus on monthly rather than quarterly returns

and our addition of the firm size control variable also deviates

slightly from the ESS model, we strictly follow their efforts to

avoid uncertainty related to availability of value relevant informa-

tion. We adopt the identical data measurement timeline plan as

ESS, as shown in Fig. 1 for our analyses of quarterly returns and

earnings.

However, for our analyses of monthly returns we require that

our current period latest forecast of future earnings occurs during

10 Seasonal earnings change also has been shown to capture cyclical behavior in

accounting earnings (Bathke, Lorek, & Willinger, 1989; Bernard & Thomas, 1990;

Brown & Rozeff, 1979; Foster, 1977; Warfield & Wild, 1992). ESS provide evidence

thataddingseasonalchange asa controlvariableto themodeladds littleexplanatory

value to the model, and we obtained the same results when we ran our analyses

including seasonal change in our modified version of the ESS model. As they were

notsignificant,thoseresults arenot presentedfor the sake of brevity.

the first month of the current quarter, no earlier than the day of the

previous quarters earnings announcement, as follows:

Rqn,m1 = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+3

x

f(n+1)qn,m1

xf(n)q(n1)

xqn xq(n1)

Pq(n1)

+ 4Size (8)

Rqn,m1 = holding return for the first month in quarter n;xf(n+1)qn,m1 =

the latest analyst forecast for quarter n+ 1 earnings per share that

comes out during the first month of quarter n and is no earlier than

the same day as the earnings announcement for quarter n1.

All other variable are as previously defined. Fig. 2 illustrates the

timing of earnings, earnings announcements, and earnings fore-

casts, and measurement points in this study.

3. Analysis and results

3.1. Data sample and primary analysis

Our data sample consists of domestic firms trading ordinary

common shares on NYSE, AMEX or NASDAQ from 1991 through

2011. We imposea December year endrequirementin orderto sim-

plify the alignment of calendar and fiscal quarter dates. Selection

parameters are:

1. Monthly holding return, price and outstanding share data avail-

able in CRSP.

2. Share price $1, to avoid very small price deflators creating

extreme values in regression variables.

3. In order for a firm to be included in any year t, CRSP data must

be available for all 12 months in year t, and also for December of

year t1.

4. Earnings per share excluding extraordinary items (EPSXQ),

report date of quarterly earnings, cash dividends per share,

quarterly revenues, and end of quarter assets available in the

Compustat quarterly database.

5. Inorderfor a firmto beincludedin anyfirm-quarter,earnings per

share and report date of quarterly earnings for the immediately

previous quarter must also be available.


6/18






Fig. 2. Timeline of events and measurement points for monthly returns,quarterly earningsand forecasts as modeled in Eq. (8). Rqn,m1 is holding returnduring thefirstmonth

ofquartern. m1 (m2, m3) is thefirst(second,third) month of quartern andxqn is earningsper share in quarter n.xf(n+1)qn is the latest analyst forecastfor quartern +1 earnings

per share that comes out in month 1 (denoted by the short solid vertical marker), and is after the announcement of quarter n1 earnings (denoted by the short dashed

vertical marker).xf(n)

q(n1)is the latest analyst forecast for quarter n earningsper sharethat comes outin quarter n1 after theannouncement of quarter n2 earnings.

6. At least two consecutive one-quarter-ahead earnings forecasts

available in I/B/E/S: the forecast for quarter n earnings per

share that was announced during quarter n1, and the fore-

cast announced in quarter n for earnings per share of quarter

n+ 1. Individual firm-quarter observations are eliminated if the

quarter n1 earnings report date in Compustat is later than the

announcement date from I/B/E/S of the n+ 1 earnings forecast.

Stock prices and EPS are adjusted for stock splits and dividends

using the cumulative adjustment factors in CRSP andCompustat. In

order to alleviatethe distorting effects of outlierswe removethetop

1% of returns and prices and the top and bottom 1% of earnings and

earnings forecasts. Firms having all the necessary pricing, earnings

and forecast data for at least one quarter in any year tare included

in the sample.

Matching of data obtained from all three datasets results in74,871 firm-quarter observations (23,716 firm-years) representing

3950 unique firms. The number of firms in each year ranges from

447 firms (1991) to 1590 (2010). Summary statistics forthe sample

are presented in Table 1, Panels A and B.

Our data constraints, especially the requirement for I/B/E/S

earnings forecast data, tend to instill in our sample a tendency

toward larger, more established firms. Evidence from numerous

prior studies suggests that the magnitude of the January effect

return premium is negatively associated with firm size (Blume &

Stambaugh, 1983; Easterday et al., 2009; Haug & Hirschey, 2006;

Hensel & Ziemba, 2000; Keim, 1983; Lamoureux & Sanger, 1989;

Reinganum, 1983). In order to establish that the small firm January

effect is present in our sample, we divide all firm-years in the

sample into deciles based on beginning of year market value ofequity, then compute mean value weighted return premiums for

each month and firm size decile. Fig. 3 shows that the January

effect is present as described in prior research. Mean return pre-

miums for January decrease monotonically and range from 7.8%

for the smallest firms to negative 0.5% for the largest firms in the

sample.11,12

11 Later in this study we examine returns for the first month, rather than for the

entire quarter, and our regression sample size shrinks due to more restrictive data

requirements for earningsforecast data. TheJanuary effect is present in the reduced

samplealso, rangingfrom a highof 7.6%for thesmallest firmsto0.5%for thelargest

firms.12 It is importantto note that weuse returnpremiumas a categorizationtool only,

in order to partition oursample between JE and NJE firms.Using returnpremiumas

Table 1a

Sample descriptive information. Panel A: Number of firms and firm-quarter obser-vations by year and in total.

Year Firms Firm-quarter observations

1991 447 1,230

1992 483 1,316

1993 570 1,494

1994 735 1,913

1995 798 2,208

1996 881 2,485

1997 1,025 2,839

1998 1,144 3,302

1999 1,118 3,243

2000 1,014 2,491

2001 1,135 3,397

2002 1,179 3,805

2003 1,242 4,050

2004 1,312 4,4122005 1,418 4,709

2006 1,504 5,052

2007 1,500 5,132

2008 1,516 5,270

2009 1,530 5,401

2010 1,590 5,567

2011 1,575 5,555

Total 23,716 74,871

The dataset consists of firm-quarter observations having 12 months of CRSP data

for each year tas well as for December of year t1, and share price $1. EPS and

report dateof quarterly earningsare available in Compustat forthe currentand prior

quarter.One-quarter-aheadearnings forecastand forecastannouncement dates are

availablein I/B/E/S for thecurrent andprior quarter.

Some studies assert that most market anomalies do not sur-vive after being made known to investors and that the January

effect has disappeared altogether (Gu, 2003, 2004; Gu & Simon,

2007). In a second examination, we followed a methodological

example included in Gu and Simons (2007) investigation into the

September phenomenon.13 Correspondingly,we hypothesized that

a dependent variable in our regression model would be inappropriate because the

model reliesupon raw returns, as demonstratedin ourEq. (5) and in Appendix B of

ESS.13 Gu and Simon (2007,page292) argue thefollowing: Ifall monthshad anequal

likelihood of being theworst performing month of theyearoverour sampleperiod,

Septemberwould be the worst about one-twelfth of the time, or 8.3% of the time.

Theyproceedto showthatin theirsample,Septemberwas theworstmonthbetween

10.7% and17.1% of thetime.


7/18






Table 1b

Sample descriptive information.Panel B: Descriptivestatistics for quarterly earningslevels andchanges,and firmsize in theentire sample.

N Mean Std. dev. Median

Quarter 1 EPS 16,046 0.226 0.406 0.180

Quarter 2 EPS 19,672 0.247 0.463 0.200

Quarter 3 EPS 20,622 0.241 0.491 0.200

Quarter 4 EPS 18,531 0.181 0.643 0.190

Quarter 1 change in EPS 16,046 0.036 0.496 0.000


Quarter 3 change in EPS 20,622 0.000 0.405 0.010Quarter 4 change in EPS 18,531 0.073 0.587 0.000

Total assets 23,716 5,927.7 32,927.8 645.9

Total revenues 23,716 3,202.0 12,028.1 591.8

Market value of equity 23,716 3,943.3 16,348.2 688.3

Quarter n EPSis earnings per share excluding extraordinary items(EPSXQ in the Compustat quarterlydataset), adjusted for effects of stock splits and dividends. Marketvalue

ofequity is measured at the beginning of the year. Total assets, Total revenues and Market value of equity are in $millions. Descriptive statistics are cross-sectional averages

over the years 19912011.

-0.020

-0.010

0.000

0.010

0.020

0.0300.040

0.050

0.060

0.070

0.080

0.090

sm-1 2 3 4 5 6 7 8 9 lg-10

meanvalue-weightedreturnpremium

decile of firm size

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Fig.3. Mean monthlyreturnpremiumsfor thesample,19912011.Regressionsam-

ple firm-years are divided into deciles based on beginning-of-yearmarket value of

equity. Return premium= CRSP monthly holding return value-weighted market

return. There are 23,716 firm-years representing 3,950 unique firms.

if all months have an equal likelihood of being a firms best per-

forming month, then JE firms should make up approximately one

twelfth (8.3%) of our sample.

We conductedchi-square tests to evaluate whether the number

of JE firms in our sample is within the expected range. Inability to

reject the null hypothesis would indicate that JE firms are no more

frequent than other month effect firms, and suggest that there is

nothing special about January.14 We examined our entire sample.

Then we separated the firms by size using market value of equity,

categorizing them as small, medium, or large, and assessed each

size category. The results of our chi-square analysis are as follows:

1. For the entire sample, we reject the null hypothesis that JE firms

are no more frequent than other month effect firms, with a

probability of


8/18






Table 2

Results of regressions of quarterlyreturnson price-deflatedearnings level, earningschange,and other value relevantinformation.The ESS model, as wellas a morerestricted

form of it that uses only contemporaneousearnings level and earningschange, are also included for comparison purposes.

Rqn = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+ 3

xf(n+1)qn x

f(t)

q(n1)

xqn xq(n1)

Pq(n1)

+ 4Size

(7)

N 0 1 2 3 4 Adj. R2

Pooled 74,871 0.034

*

0.255 0.166 0.0050.037* 0.215 2.532*** 2.470*** 0.023

0.057* 0.224 2.257*** 2.468*** 0.003 0.023

Quarter 1 16,046 0.008 0.563** 0.050 0.008

0.020 0.537*** 2.642*** 2.676*** 0.030

0.055 0.585*** 2.652*** 2.685*** 0.005 0.032

Quarter 2 19,672 0.064 0.305 0.419*** 0.007

0.060 0.241 2.498*** 2.155*** 0.021

0.082 0.226 2.492*** 2.153*** 0.003 0.021

Quarter 3 20,622 0.007 0.518 0.062 0.008

0.006 0.429 2.316*** 2.402*** 0.024

0.002 0.433 2.311*** 2.399*** 0.001 0.024

Quarter 4 18,531 0.073* 0.061 0.047 0.016

0.076* 0.198 3.067* 2.907* 0.038

0.105 0.209 3.076* 2.923* 0.004 0.038

Significant at < 0.10.* Significant at < 0.05.

**

Significant at < 0.001.*** Significant at < 0.0001.

Rqn is holding period return for quarter n, Pq(n1) is beginning of period price, xqn is earnings per share in quarter n, xf(n)

q(n1) is the latest analyst forecast that comes out in

quarter n1 forquartern earnings per share. Earnings and prices are adjusted for stock splits and dividends. Regressions utilize two-way cluster robust standard errors.

Table 3

Results of regressions of returns forthe first month of each quarter on price-deflated earningslevel, earnings change, other value relevant information,and firmsize.

Rqn,m1 = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+ 3

xf(n+1)

qn,m1 xf(n)

q(n1)

xqn xq(n1)

Pq(n1)

+ 4Size

(8)

N 0 1 2 3 4 Adj. R2

Pooled 39,601 0.017 0.141* 1.802*** 1.719*** 0.000 0.021

Quarter 1 4,744 0.067* 0.194 1.915* 1.998** 0.008* 0.027

Quarter 2 11,598 0.032 0.088 1.425*** 1.272** 0.001 0.011

Quarter 3 12,657 0.032 0.081 1.608** 1.635** 0.004 0.020Quarter 4 10,602 0.036 0.222 2.297* 2.217* 0.002 0.037

* Significant at < 0.05.** Significant at < 0.001.

*** Significant at < 0.0001.

Rqn,m1 is holding period return for the first month in quarter n, Pq(n1)is beginning of period price, xqn is earnings per share in quarter n, xf(n)

q(n1) is the latest analyst forecast

that comes outin quarter n1 forquartern earningsper share,xf(n+1)qn,m1

is thelatest analyst forecastthat comes out in thefirstmonth of quarter n for quartern+1 earnings per

share. Size is measured as the logarithm of total assets. Earnings and prices are adjusted for stock splits and dividends. Regressions utilize two-way cluster robust standard

errors.

contemporaneous returns, even when limited to the first month of

the quarter.

3.2. JE firms versus NJE firms

In order to more closely examine whether the January effect is

associated with accounting earnings information, it is important

to identify firms whose returns are representative of this market

anomaly. Rather than apply an ad hoc return premium variable to

our regression that is unsupported by the economic intuition of

the ESS model, we partition the sample into firms that experience

the January effect (JE firms) and those that do not (NJE firms); the

NJE firms act as a comparison sample. Our partition method pro-

ceeds as follows: foreach firmobservation in year t, we compute its

monthly return premiums by subtracting the value weighted mar-

ket return from the CRSP monthly holding return. If the January

return premium is the highest of all 12 months in year t then

the firm is categorized as a JE firm for that year; otherwise it is

categorized as NJE.16 Table 4, Panel A shows the distribution of JE

and NJE firms in the sample by year and overall. Over the entire

time period of study 9.1% of our sample firms are classified as JE

firms; the proportion of JE firms varies from 4.4% in 2000 to 22.2%

in 1992.17 This is consistent with previously cited evidence that the

January effect is more pronounced in some years than in others.

Table 4 Panels B and C show summary statistics for the parti-

tioned sample. Unsurprisingly, firms that experience January effect

returns are on average smaller in terms of assets, annual revenues,

and market value of equity but there is little discernible differ-

ence in price-deflated earnings or change in earnings between JE

and NJE. The mean January return premium for JE firms is 22.6%,

versus negative 1.6% for the much larger population of NJE firms.

16 If January is tied with any other month(s) for highest premium, the firm is

classified as JE for that year.17 We conducted a similar analysis, not presented here for the sake of brevity, of

the 93,112 firms that met our CRSP selection criteria prior to being matched with

Compustat or I/B/E/S data. JE firms made up 10.8% of that CRSP dataset.


9/18






Table 4a

Descriptiveinformation for thepartitioned sample. Panel A: Distribution of JE and NJE firms by year andoverall.

Year Firms JE firms NJE firms % JE firms % NJE firms

1991 447 81 366 18.1% 81.9%

1992 483 107 376 22.2% 77.8%

1993 570 67 503 11.8% 88.2%

1994 735 73 662 9.9% 90.1%

1995 798 60 738 7.5% 92.5%

1996 881 64 817 7.3% 92.7%

1997 1,025 78 947 7.6% 92.4%1998 1,144 64 1,080 5.6% 94.4%

1999 1,118 75 1,043 6.7% 93.3%

2000 1,014 45 969 4.4% 95.6%

2001 1,135 191 944 16.8% 83.2%

2002 1,179 83 1,096 7.0% 93.0%

2003 1,242 89 1,153 7.2% 92.8%

2004 1,312 155 1,157 11.8% 88.2%

2005 1,418 90 1,328 6.3% 93.7%

2006 1,504 265 1,239 17.6% 82.4%

2007 1,500 123 1,377 8.2% 91.8%

2008 1,516 71 1,445 4.7% 95.3%

2009 1,530 164 1,366 10.7% 89.3%

2010 1,590 119 1,471 7.5% 92.5%

2011 1,575 106 1,469 6.7% 93.3%

Total 23,716 2,170 21,546 9.1% 90.9%

JE firms are those for which the January return premium is higher than the return premium of any other month in the calendar year. NJE firms are those for which the return

premium of at least oneother month is higherthan thereturn premium for January in thecalendar year. Return premium= CRSP monthly holding return value weightedmarket return.

Table 4b

Descriptive information for the partitioned sample. Panel B: Descriptive statistics for earnings levels, earnings changes, and firm size (JE firms).


January return premium 1,423 0.226 0.179 0.180

Quarter 1 EPS 1,423 0.225 0.449 0.173

Quarter 2 EPS 1,777 0.232 0.541 0.178

Quarter 3 EPS 1,886 0.209 0.535 0.160

Quarter 4 EPS 1,676 0.186 0.630 0.160




Quarter 4 change in EPS 1,676 0.042 0.502 0.000Total assets 2,170 4,834.6 24,537.6 505.2

Total revenues 2,170 3,156.4 13,254.1 506.1



premium of at least oneother month is higherthan thereturn premium for January in thecalendar year. Return premium= CRSP monthly holding return value weighted

market return. Quarter n EPSis earnings per share excluding extraordinary items (EPSXQ in the Compustat quarterly dataset) deflated by beginning of quarter share price

obtained from CRSP, and adjusted for effects of stock splits and dividends. Market value of equity is measured at the beginning of the year. Total assets, Total revenues and

Marketvalue of equityare in $millions. Descriptive statistics are cross-sectional averages over the years 19912011.

Table 4c

Descriptive information for the partitioned sample. Panel C: Descriptive statistics for earnings levels, earnings changes, and firm size (NJE firms).


January return premium 14,623 0.016 0.108 0.016Quarter 1 EPS 14,623 0.226 0.402 0.180

Quarter 2 EPS 17,895 0.249 0.454 0.204

Quarter 3 EPS 18,736 0.244 0.486 0.205

Quarter 4 EPS 16,855 0.180 0.644 0.195





Total assets 21,546 6,038.0 33,656.5 658.484

Total revenues 21,546 3,206.6 11,897.7 602.341




market return. Quarter n EPSis earnings per share excluding extraordinary items (EPSXQ in the Compustat quarterly dataset) deflated by beginning of quarter share price

obtained from CRSP, and adjusted for effects of stock splits and dividends. Market value of equity is measured at the beginning of the year. Total assets, Total revenues and

Marketvalue of equityare in $millions. Descriptive statistics are cross-sectional averages over the years 19912011.


10/18
http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.qref.2015.05.001


11/18






Table 5a

Regressions on the partitioned sample. Panel A: Results of regressing January returns on price deflated first quarter earnings level, earnings change, other value relevant

information, and firm size.

Rqn,m1 = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)

Pq(n1)+ 3

xf(n+1)

qn,m1 xf(n)

q(n1)

xqn xq(n1)

Pq(n1)

+ 4Size

(8)

N 0 1 2 3 4 Adj. R2

All firms 4,744 0.067

*

0.194 1.915

*

1.998

**

0.008

*

0.027JE firms 538 0.530*** 2.434** 3.683* 2.836 0.039*** 0.293

NJE firms 4,206 0.000 0.421* 1.339*** 1.409*** 0.003 0.026


** Significant at < 0.001.*** Significant at < 0.0001.


premium of at least one other month is higher than the return premium for January in the calendar year. Return premium =CRSP monthly holding returnvalue weighted

market return. Rqn,m1 is holding period return for the first month in quarter n, Pq(n-1) is beginning of period price,xqn is earnings per share in quarter n, xf(n)

q(n1)is the latest

analyst forecast that comes out in quarter n-1 for quarter n earnings per share,xf(n+1)

qn,m1 is the latestanalyst forecast that comes out in thefirst month of quarter n for quarter

n+1 earnings per share.Size is measured as thelogarithm of total assets. Earningsand pricesare adjusted forstocksplits anddividends. Regressionsutilize two-way cluster

robust standard errors.

Table 5b

Regressionson the partitioned sample. Panel B: Results of regressing firstquarterreturnson price-deflatedearnings level, earningschange,other value relevantinformation,

and firm size after partitioning the sample into JE and NJE firms. Results from the unpartitioned sample of first quarter observations shown in Table 2 are presented in the

third row for comparison.

Rqn = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+ 3

xf(n+1)qn x

f(t)

q(n1)

xqn xq(n1)

Pq(n1)

+ 4Size

(7)

N 0 1 2 3 4 Adj. R2

All firms 15,819 0.055 0.585*** 2.652*** 2.685*** 0.005 0.032

JE firms 1,377 0.347*** 0.299 2.529* 2.288* 0.024** 0.053

NJE firms 14,442 0.012 0.672*** 2.669*** 2.707*** 0.002 0.038




premium of at least one other month is higher than the return premium for January in the calendar year. Return premium =CRSP monthly holding returnvalue weighted

market return. Rqn,m1 is holding period return for the first month in quartern, Pq(n1) is beginning of period price,xqn is earnings per share in quarter n, x

f(n)

q(n1)is the latestanalyst forecast that comes out in quarter n-1 for quarter n earnings per share,xf(n

+1)

qn.m1 is the latestanalyst forecast that comes out in thefirst month of quarter n for quarter

n+1 earnings per share.Size is measured as thelogarithm of total assets. Earningsand pricesare adjusted forstocksplits anddividends. Regressionsutilize two-way cluster

robust standard errors.

Table 6a

Potential tax-lossselling and non-tax-lossselling firms. Panel A: Distributionof PTLS andNTLS in thefull sample and forJE firms.

Year PTLS NTLS % PTLS % NTLS %PTLS and JE %NTLS and JE

1991 52 208 20.0% 80.0% 19.2% 15.9%

1992 43 292 12.8% 87.2% 30.2% 21.9%

1993 120 187 39.1% 60.9% 12.5% 9.6%

1994 122 274 30.8% 69.2% 13.1% 8.4%

1995 177 296 37.4% 62.6% 9.0% 6.1%

1996 245 317 43.6% 56.4% 9.4% 5.7%

1997 305 332 47.9% 52.1% 9.5% 6.3%

1998 377 387 49.3% 50.7% 6.9% 3.4%

1999 303 479 38.7% 61.3% 8.6% 5.4%2000 213 374 36.3% 63.7% 3.3% 4.5%

2001 242 523 31.6% 68.4% 29.8% 9.0%

2002 226 676 25.1% 74.9% 8.4% 7.0%

2003 620 319 66.0% 34.0% 7.4% 6.0%

2004 319 642 33.2% 66.8% 15.7% 9.8%

2005 249 641 28.0% 72.0% 6.4% 6.2%

2006 437 452 49.2% 50.8% 21.1% 17.3%

2007 378 448 45.8% 54.2% 11.6% 6.0%

2008 489 420 53.8% 46.2% 7.0% 4.5%

2009 305 588 34.2% 65.8% 12.1% 11.1%

2010 136 710 16.1% 83.9% 8.8% 8.0%

2011 111 682 14.0% 86.0% 7.2% 6.5%

Total 5,469 9,247 37.2% 62.8% 11.2% 8.2%

Firms represented are those for which 12 months of CRSP holding returns are available for year t1. Firms with negative holding returns for December of year t1 are

categorized as potential tax-loss-sellers (PTLS). All other firms are classified as non-tax-loss sellers (NTLS). JE firms are those for which the January return premium is

higherthan thereturn premium of any other month in thecalendaryear.Return premium =CRSP monthly holding return valueweightedmarket return.


12/18






Table 6b

Potential tax-lossselling and non-tax-lossselling firms. Panel B: Distribution of PTLS and NTLS forboth JE and NJE firms, by year and overall.

Year JE firms % PTLS % NTLS NJE firms % PTLS % NTLS

1991 43 23.3% 76.7% 217 19.4% 80.6%

1992 77 16.9% 83.1% 258 11.6% 88.4%

1993 33 45.5% 54.5% 274 38.3% 61.7%

1994 39 41.0% 59.0% 357 29.7% 70.3%

1995 34 47.1% 52.9% 439 36.7% 63.3%

1996 41 56.1% 43.9% 521 42.6% 57.4%

1997 50 58.0% 42.0% 587 47.0% 53.0%1998 39 66.7% 33.3% 725 48.4% 51.6%

1999 52 50.0% 50.0% 730 37.9% 62.1%

2000 24 29.2% 70.8% 563 36.6% 63.4%

2001 119 60.5% 39.5% 646 26.3% 73.7%

2002 66 28.8% 71.2% 836 24.8% 75.2%

2003 65 70.8% 29.2% 874 65.7% 34.3%

2004 113 44.2% 55.8% 848 31.7% 68.3%

2005 56 28.6% 71.4% 834 27.9% 72.1%

2006 170 54.1% 45.9% 719 48.0% 52.0%

2007 71 62.0% 38.0% 755 44.2% 55.8%

2008 53 64.2% 35.8% 856 53.2% 46.8%

2009 102 36.3% 63.7% 791 33.9% 66.1%

2010 69 17.4% 82.6% 777 16.0% 84.0%

2011 52 15.4% 84.6% 741 13.9% 86.1%

Total 1,368 44.7% 55.3% 13,348 36.4% 63.6%


premium of at least oneother month is higherthan the returnpremium for January in thecalendar year. Return premium= CRSP monthly holding return value weightedmarket return. Firms represented are those for which 12 months of CRSP holding returns are available for year t1. Firms are categorized as potential tax-loss sellers

(PTLS) if they have negative holding returns for Decemberof year t1. Allother firms areclassified as not tax-losssellers(NTLS).

t. PTLS firms are also on average smaller and have lower revenues

than NTLS firms.

We perform the regression analysis in Eq. (8) on firms hav-

ing sufficient available data, examining the association between

January returns and first quarter earnings information for all cat-

egories of tax-loss seller status and JE/NJE experience. Results are

shown in Table 7.

For both tax-loss sellers and non-tax-loss sellers overall the

model performs as expected: the coefficient on earnings level is

insignificantly different from zero, coefficients related to perma-

nent earnings performance are significantly positive, explanatory

value does not exceed 4.3%. However, the JE/NJE partition once

more delivers interesting results. The large and significantly neg-

ative coefficient on earnings level again appears, but only for JE

tax-loss sellers. For allother categories it is insignificantly different

fromzero or slightly positive.Permanentearningscoefficients for JE

PTLS are significantly positive and approximately four times larger

in magnitude than the coefficient values for either NJE PTLS or any

Table 6cPotential tax-loss selling and non-tax-loss selling firms. Panel C: Descriptive statistics for PTLS and NTLS.


PTLS

Quarter 1 EPS 1,781 0.253 0.440 0.180

Quarter 2 EPS 4,043 0.257 0.460 0.220

Quarter 3 EPS 4,356 0.233 0.533 0.207

Quarter 4 EPS 3,616 0.182 0.692 0.200





Total assets 5,469 6765.2 35,367.4 765.2

Total revenues 5,469 3890.1 14,434.3 749.7

Market value of equity 5,469 5001.0 20,252.7 827.1

NTLS

Quarter 1 EPS 2,963 0.321 0.433 0.250

Quarter 2 EPS 6,796 0.326 0.455 0.270

Quarter 3 EPS 7,400 0.321 0.466 0.270

Quarter 4 EPS 6,195 0.269 0.591 0.250





Total assets 9,247 7,414.0 37,506.6 1,020.1

Total revenues 9,247 4,233.9 14,055.1 983.6

Market value of equity 9,247 5,278.0 18,610.2 1,117.0

Firms arecategorized as potential tax-loss sellers (PTLS) if they have negative holding returns for themonth of Decemberin year t-1. All other firms areclassified as not

tax-loss sellers (NTLS).Quarter n EPSis earningsper share excluding extraordinary items(EPSXQ in the Compustat quarterly dataset)deflated by beginningof quarter share

price obtained from CRSP, and adjusted for effects of stock splits and dividends. Market value of equity is measured at the beginning of the year. Total assets, Total revenues

and Market value of equityare in $millions. Descriptive statistics are cross-sectional averages over the years 19912011.


13/18






Table 7

Results of regressing January returns on price-deflated first-quarter earnings level, earnings change, other value relevant information, and firm size, after partitioning the

sample as to tax loss-sellingstatus and JE/NJE firms, 19912011.

Rqn,m1 = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+ 3

xf(n+1)

qn,m1 xf(n)

q(n1)

xqn xq(n1)

Pq(n1)

+ 4Size

(8)

N 0 1 2 3 4 Adj. R2

All PTLS 1,781 0.113

*

0.157 2.609

*

2.786

*

0.013

*

0.043JE PTLS 233 0.547*** 4.131*** 5.753*** 3.915* 0.038*** 0.395

NJE PTLS 1,548 0.020 0.432* 1.379* 1.538* 0.005 0.028

All NTLS 2,963 0.035 0.422 1.442* 1.433* 0.005 0.020

JE NTLS 305 0.456*** 0.913 1.005 0.799 0.033*** 0.203

NJE NTLS 2,658 0.014 0.492* 1.344** 1.326* 0.002 0.024





market return. Firms are categorized as potential tax-loss sellers (PTLS) if they have negative holding returns for the month of December in year t1. All other firms are

classified as not tax-loss sellers (NTLS). Rqn,m1 is holding period return for the first month in quarter n, in this case January returns.Pq(n1)is beginning of period price,xqn

is earningsper share in quarter n,xf(n)

q(n1)is thelatest analyst forecast that comes outin quarter n1 for quarter n earnings per share,x

f(n+1)qn,m1

is thelatest analyst forecast that

comes out in thefirstmonthof quartern for quartern+ 1 earningsper share.Size is measured as thelogarithmof total assets. Earningsand prices areadjusted forstock splits

and dividends. Regressions utilize two-way cluster robust standard errors.

NTLS. The explanatory value of our earnings model increases by a

factorof about tenfor theJE firms in both thePTLSand NTLS groups;

the JE PTLS explanatory value is nearly twice that of JE NTLS. Fur-

ther, none of the earnings variables are significantly different from

zero for the JE NTLS. Together with its adjusted R2 value, this sug-

gests thatin the case of JE NTLS firms the model does a good job of

explaining the variability in January effect returns but we cannot

sort out which, if any, variable(s) matter(s), apart from firm size.

In summary, consistent with previous studies suggesting that

tax-loss selling is at least a partial explanation for January return

premiums, we find a sizable, though not complete, correspondence

between JE firms andthose we categorize as potentialtax-loss sell-

ers. This suggests that our categorization scheme is reasonable.

Further, complementing the evidence of Henker and Debapriya

(2012), our partitioning methodology teases out evidence that

January effect investors behavior is economically rational: excep-

tionally high January returns on PTLS loser firms may occur as

investors discount comparatively poor current earnings perfor-

mance and give extra weight to permanent earnings and their

sustainability in the future. Finally, the existence of JE firms that

are not tax-loss sellers and our inability to determine significant

explanatory variables for them suggest that there are still some

aspects of this market puzzle that are not well understood, and

that invite future investigation.

4. Robustness testing

4.1. Expectations and EPS performance of JE firms

As a reasonableness check on our findings, we investigate

how expectations and actual performance change over the time

period from the fourth quarter of year (t1) through the first

and second quarters of year (t). Table 8 presents mean quarterly

earnings and expectations for permanent earnings sustainabilityx

f(n+1)qn x

f(n)q(n1)

xqn xq(n1)

21 for the fourth through

second quarters for JE firms. The outlook for permanent earnings

is very negative in the fourth quarter for JE firms and then highly

21 This is the other information term described previously. It is the difference

between the change in forecasted future earnings and contemporaneous change in

earnings, and it captures expectations for the sustainability of earnings growth.

Table 8

Mean quarterly earnings per share and expectations for permanent earnings sus-

tainability (JE firms).

N=1,781

Quarter 4 EPS(year t1) 0.391***

Quarter 1 EPS 0.340***

Quarter 2 EPS 0.398***

Quarter 4 expectation for permanent earnings sustainability 0.062**

Quarter 1 expectation for permanent earningssustainability 0.089***

** Significant at


14/18






Table 9a

Correlations of quarterly earnings. Panel A: JE versus NJE firms, 19912011.

NJE firms JE firms

Quarter 4 (t1) Quarter 1 (t) Quarter 2 (t) Quarter 3 (t)

Quarter 4 (t1) 1.000 0.0457*

Quarter 1 (t) 0.0057 1.000 0.0389*

Quarter 2 (t) 0.0344*** 1.000 0.0366***

Quarter 3 (t) 0.0508*** 1.000

JE firms are those for which the January return premium is higher than the return premium of any other month in the calendar year. NJE firms are those for which thereturn premium of at least one other month is higher than the return premium for January in the calendar year. Return premium= CRSP monthly holding return value

weightedmarket return. Theyear in whichthe comparison quarter fallsis designated as t. N= 3,909for JEfirms.N= 35,645 for NJEfirms. Correlationspresented areSpearman

correlations, which do not depend upon assumptions of the datas normal distribution or heteroskedasticity.* Significant at < 0.05.


Table 9b

Correlations of quarterly earnings. Panel B: JE versus NJE firms, 19912011 (PTLS).

NJE firms JE firms


Quarter 4 (t1) 1.000 0.0650*

Quarter 1 (t) 0.0229* 1.000 0.1127***

Quarter 2 (t) 0.0326** 1.000 0.0474

Quarter 3 (t) 0.0548*** 1.000




premium of at least oneother month is higherthan the returnpremium for January in thecalendar year. Return premium= CRSP monthly holding return value weighted

market return. Firms are categorized as potential tax-loss sellers (PTLS) if they have negative holding returns for the month of December in year t1. All other firms

are classified as not tax-loss sellers (NTLS). The year in which the comparison quarter falls is designated as t. N=1,607 for JE firms. N= 12,171 for NJE firms. Correlations

presented are Spearman correlations, which do not depend upon assumptions of the datas normal distribution or heteroskedasticity.

In Panels B andC, we partitionthe samplealongthe dual dimen-

sions of tax-loss selling status and JE versus NJE firms. For JE PTLS

firmsthe earnings correlation coefficientsare of the predictedsigns,

and stronger. On the other hand, Panel C shows that the first quar-

ter (t)-fourth quarter (t1), as well as the second quarter (t)-firstquarter (t) correlation coefficients are both insignificantly different

from zero for JE NTLS firms. NJE firms in Panels B and Panel C are

positively significant for the first-second quarter earnings correla-

tions, but the correlation between the first and fourth quarters for

PTLS (NTLS) firms is significantly negative (insignificantly different

from zero). Earnings correlations between quarters two and three

are positively significant in all cases except NJE PTLS, where the

correlation is insignificantly different from zero.

Taken together, we interpret these results as evidence suppor-

ting our argument that in the case of JE firms, investors seem to

rationally discount relativelypoor earnings performance in the first

quarter and focus on increased expectations for future improved

earnings in awarding large return premiums.

4.2. The role of dividends

Hand and Landsman (2005) argue that dividends could proxy

for other value relevant information because they are included

in the linear information dynamics supporting the Ohlson valua-tion framework. Their argument is consistent with evidence that

investors may look for information about expected future perfor-

mance in firmsdividendpolicies (Brucato & Smith,1997; Jin, 2000).

ESS test the role of dividends in their returns specification and find

no evidence that current or past period dividends add explanatory

power, nor that dividends proxy for other value relevant informa-

tion in the absence of their term representing information about

expected growth in permanentearnings,

xf(t+1)t x

f(t)t1

(xtxt1)

Pt1. We

re-examine dividends as a potential proxy for value relevant infor-

mation here for two reasons. First, measuring

xf(t+1)t x

f(t)t1

(xtxt1)

Pt1requires analysts forecasts, the availability of which introduces a

Table 9c

Correlations of quarterly earnings. Panel C: JE versus NJE firms, 19912011 (NTLS).

NJE firms JE firms


Quarter 4 (t1) 1.000 0.0234

Quarter 1 (t) 0.0011 1.000 0.0114

Quarter 2 (t) 0.0354*** 1.000 0.0813***

Quarter 3 (t) 0.0408*** 1.000



premium of at least oneother month is higherthan the returnpremium for January in thecalendar year. Return premium= CRSP monthly holding return value weighted

market return. Firms are categorized as potential tax-loss sellers, PTLS if they have negative holding returns for the month of December in year t1. All other firms are

classifiedas nottax-loss sellers,NTLS.The yearin whichthe comparison quarterfalls is designated as t. N= 1,991forJE firms.N= 21,336 for NJEfirms. Correlationspresented

are Spearman correlations, which do not depend upon assumptions of the datas normal distribution or heteroskedasticity.


15/18






Table 10

Results of regressing January returns on price-deflated first quarter earnings level, earnings change, dividends, and firm size on the unpartitioned and partitioned sample.

Rqn,m1 = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+ 4Size+ 5dqn

Pq(n1)+ 6

dq(n1)Pq(n1)

(9)

N 0 1 2 4 5 6 Adj. R2

All firms 3,707 0.059 0.304 0.079 0.007* 1.062 0.145 0.010

JE firms 418 0.509*** 3.139* 1.155* 0.033*** 0.759 1.923 0.275

NJE firms 3,289 0.008 0.431* 0.038 0.002 0.742 0.027 0.008

* Significant at < 0.05.*** Significant at < 0.0001.

JE firms are those for which the return in January return premium is higher than the return premium of any other month in the calendar year. NJE firms are those for which

the returnpremium of at least oneothermonthis higher than thereturn premium forJanuary in thecalendaryear. Return premium =CRSP monthly holding return value

weighted market return. Rqn,m1 is holding period return for the first month in quarter n, in this case January returns. Pq(n1)is beginning of period price, xqn is earnings per

share in quartern, dqnis cash dividend per share in quartern. Size is measured as thelogarithm of total assets. Earningsand pricesare adjustedfor stock splitsand dividends.

Regressions utilize two-way cluster robust standard errors.

selection bias toward larger, more established firms. Second, the

main difference in specification between our empirical model in

thisstudyandthatproposedbyESSisourfocusonthefirstmonthof

the quarter, when investors may be struggling to determine expec-

tations for future earnings performance. In such a setting there

may be an informational function for dividends. We investigate therole of dividends in explaining January returns using the following

empirical form of the model, which follows from Eq. (5) and in

which current (first quarter) and the previous years fourth quarter

price deflated dividends replace

xf(t+1)t x

f(t)t1

(xtxt1)

Pt1, as follows:

Rqn,m1 = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+ 4Size+ 5dqn

Pq(n1)

+6dq(n1)Pq(n1)

(9)

dqn = cash dividends paid in quarter n, in this case quarter 1 and all

other variables are as previously defined.

Results are shown in Table 10. As we are limited to using firms

for which we have two-consecutive-quarter dividend data avail-

able our sample size decreases slightly from 538 (4,206) JE (NJE)

observations in quarter 1 with analysts forecasts to 418 (3,707).

The analysis indicatesthat substituting firstand fourth quarter divi-

dends for the other information term results in a model with little

explanatory power for either NJE firms or firms overall, as their

respective adjusted R2 values are only 0.008 and 0.010. Consistent

with the idea that dividends represent a decrease in future value-generating power, the coefficients on 5 and 6, are negative inall cases but nowhere significant. In the case of JE firms adjusted

R2 using dividends is higher than for NJE firms or for the pooled

sample. Adjusted R2 values for JE firms are slightly lower for the

dividend model than for the ESS model (0.275versus 0.293). Taken

together,we findlittle support for dividends as an explanatory vari-

able that is preferable to the other information term specified by

ESS.

4.3. April (July, October) effects

As a final check, we perform similar partitions for firms that

experience an April (July, October) effect; that is, April (July,

October) return premiums are higher than return premiums of

all other months in the year. Consistent with the methodology

Table 11

Results of regressing April (July, October) returns on price-deflated second (third, fourth) quarter earnings level, earnings change, other value relevantinformation, and firm

size, afterpartitioning thesample into AE (JULE,OE) and NAE (NJULE, NOE) firms. Results forJE and NJE firms from Table5, Panel A are shown in the top line ofeachsection

for comparison purposes.

Rqn,m1 = 0 + 1xqn

Pq(n1)+ 2

xqn xq(n1)Pq(n1)

+ 3

xf(n

+1)

qn,m1 xf(n)

q(n1)

xqn xq(n1)

Pq(n1)

+ 4Size

(8)

N 0 1 2 3 4 Adj. R2

JE firms 538 0.530*** 2.434** 3.683* 2.836 0.039*** 0.293

AE firms 1,427 0.488***

0.677

1.356***

0.672 0.029*

0.074JULE firms 1,154 0.414*** 0.541*** 3.248* 2.918* 0.029*** 0.208

OE firms 1,202 0.560*** 0.800* 2.451* 2.095* 0.040** 0.192

NJE firms 4,206 0.000 0.421* 1.339*** 1.409*** 0.003 0.026

NAE firms 10,171 0.019 0.092 1.125** 0.853* 0.004* 0.014

NJULE firms 11,503 0.066* 0.445* 1.385*** 1.513*** 0.006* 0.037

NOE firms 9,400 0.026 0.296 1.572* 1.525* 0.002 0.034


** Significant at < 0.001.*** Significant at < 0.0001.

JE (AE, JULE, OE) firms are those for which the return premium in January (April, July, October) is higher than the return premium of any other month in the calendar year.

NJE (NAE, NJULE, NOE) firms are those for which the return premium of at least one other month is higher than the return premium for January (April, July, October) in

the calendar year. Rqn,m1 is holding period return for first month in quartern, Pq(n1) is beginning of period price, xqn is earnings per share in quarter n, xf(n)

q(n1)is the latest

analyst forecast that comes out in quarter n1 forquartern earnings per share,xf(n+1)

qn,m1 is thelatest analyst forecast that comes outin thefirst month of quartern for quarter

n+1 earningsper share. Size is measured as thelogarithm of total marketvalueof equity. Earnings and prices areadjusted forstocksplits anddividends. Regressions utilize

two-way cluster robust standard errors.


16/18
http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0105http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.elsevier.com/S1062-9769(15)00057-5/sbref0100http://refhub.

Documents

market janurary