Financial exibility and the choice between dividends and ... · total payouts and the division of these payouts between dividends and stock repurchases. In Section 3, we provide estimates

Journal of Financial Economics 57 (2000) 355}384

Financial #exibility and the choice betweendividends and stock repurchasesq

Murali Jagannathan!, Cli!ord P. Stephens!,*,Michael S. Weisbach"

!University of Missouri, Columbia, MO 65211, USA"University of Illinois, Champaign, IL 61802-6980, USA

Received 19 January 1999; received in revised form 9 November 1999

Abstract

This paper measures the growth in open market stock repurchases and the manner inwhich stock repurchases and dividends are used by U.S. corporations. Stock repurchasesand dividends are used at di!erent times from one another, by di!erent kinds of "rms.Stock repurchases are very pro-cyclical, while dividends increase steadily over time.Dividends are paid by "rms with higher `permanenta operating cash #ows, whilerepurchases are used by "rms with higher `temporarya, non-operating cash #ows.Repurchasing "rms also have much more volatile cash #ows and distributions. Finally,"rms repurchase stock following poor stock market performance and increase dividendsfollowing good performance. These results are consistent with the view that the #exibilityinherent in repurchase programs is one reason why they are sometimes used instead ofdividends. ( 2000 Elsevier Science S.A. All rights reserved.

JEL classixcation: G32; G35

Keywords: Payout policy; Stock repurchases; Dividends

qWe would like to thank Jim Brickley, Harry DeAngelo, Stuart Gillan, Kevin Hallock, JarradHarford, Ron Howren, David Ikenberry, Ed Kane, Laurie Krigman, Harold Mulherin, RonOaxaca, Neil Pearson and Bill Schwert (the editor), as well as seminar participants at Arizona,Illinois, Maryland, Missouri, New Mexico, Oklahoma, UC-Riverside, and Washington Universityfor their helpful comments. The NSF (Grant SBR-9616675) provided "nancial support.

*Corresponding author. Tel.: #1-573-884-0945; fax: #1-573-884-6296.

E-mail address: [email protected] (C.P. Stephens).

0304-405X/00/$ - see front matter ( 2000 Elsevier Science S.A. All rights reserved.PII: S 0 3 0 4 - 4 0 5 X ( 0 0 ) 0 0 0 6 1 - 1

1This agrument does not say anything about the reason for these expectations. Rather, at leastsince 1956 when Lintner published his famous paper, "rms have been making such implicitcommitments with their dividends and market expectations have grown very strong. As an illustra-tion of these strong expectations, FPL Group in 1994 decreased their dividend by 32% andsimultaneously announced their intention to pay out the cash as a repurchase instead. FPL's stockfell by 13.7% on the announcement of this change. (See Esty and Schreiber (1995) for more detail.)

1. Introduction

One of the most signi"cant trends in corporate "nance during the 1990s is theincreasing popularity of open market stock repurchase programs. Between 1985and 1996, the number of open market repurchase program announcements byU.S. industrial "rms has increased 650% from 115 to 755, and their announcedvalue has increased 750% from $15.4 billion to $113 billion. Correspondingly,dividends have only risen by a factor of just over two during the same period;aggregate dividends for all industrial "rms listed on Compustat have risen from$67.6 billion to $141.7 billion. Repurchases are clearly an increasingly importantmethod of paying out cash to shareholders.

In this paper, we examine "rms' decisions to distribute cash #ows and theirchoices between paying out cash #ows in the form of dividends or stockrepurchases. Our goal is to assess the increasing importance of repurchases inpayout decisions and to isolate factors that a!ect the choice between repur-chases and dividends. Out primary hypothesis is that dividends represent anongoing commitment and are used to distribute permanent cash #ows, whilerepurchases are used to pay out cash #ows that are potentially temporary.Repurchases thus preserve "nancial #exibility relative to dividends because theydo not implicitly commit the "rm to future payouts.

At least since Lintner (1956), "rms have been reluctant to cut dividends andhave been greeted by a signi"cantly negative stock market reaction when theydo; Ghosh and Woolridge (1988) and Denis et al. (1994) report an average stockprice decline of about 6% on the three days surrounding the announcement ofa dividend cut. Repurchases, on the other hand, involve no such commitment orrisk. Firms sometimes announce programs but fail to repurchase any shares.Even if a "rm completes a program, it is under no explicit or even implicitobligation to begin another new repurchase program. Given these marketexpectations, stock repurchases would be a sensible way for "rms to pay outcash #ows that have a high likelihood of not being sustainable.1

In Section 2 of the paper, we discuss the construction of a database ofrepurchase announcements and actual share repurchases for all U.S. public"rms for the period 1985 to 1996. We begin with an initial sample of allrepurchases programs announced from 1985 through 1996, as reported bySecurities Data Company (SDC). Since it is not always possible to determinewith publicly available information exactly how many shares a particular "rm

356 M. Jagannathan et al. / Journal of Financial Economics 57 (2000) 355}384

purchased in a given year, we construct estimates of the upper and lower boundsof the number of shares repurchased by each of these "rms that are listed onboth Standard and Poor's Compustat database and the Center for Research inSecurities Prices (CRSP) tapes. We use these estimates to calculate estimates oftotal payouts and the division of these payouts between dividends and stockrepurchases.

In Section 3, we provide estimates of aggregate repurchases, as well as theaggregate value of other forms of payouts. We do so both to test the aggregateform of the #exibility hypothesis that repurchases are more pro-cyclical thandividends and because our estimates of repurchases are improvements on thosein the literature. At an economy-wide level, we "nd that actual share repurchaseshave increased over our sample period and represent an economically importantsource of payouts. Over the period 1985 to 1996, aggregate actual sharerepurchases by industrial "rms total between $249 billion and $339 billion. Thiscorresponds to 53 to 72% of announced repurchase levels and 20 to 27% ofaggregate dividends.

In contrast to dividends, which grow smoothly, aggregate share repurchasesare volatile and vary considerably with the business cycle. Firms increase theirrepurchases disproportionately relative to dividends during boom times andreduce them more during recessions. Even though dividends continue to makeup the majority of total payouts, repurchases are responsible for much of theyear-to-year variation. Overall, repurchases have not replaced dividends as theprimary payout vehicle. Even in 1996, which was the largest year for repurchasesin this sample, dividends amounted to more than double the total actual sharerepurchases and 126% of the total announced share repurchases. Firms are stillgenerally increasing dividends every year; the fact that they are doing so and notincreasing repurchases even faster suggests that there still is a dividend puzzle.

Section 4 of the paper considers the cross-sectional determinants of payoutpolicy. We "rst reexamine Lintner's (1956) famous arguments. Lintner's premiseis that managers prefer to increase dividends regularly and avoid decreasingdividends if possible. These arguments predict that dividend increases will bemade by "rms with higher and more stable cash #ows, that dividend increaseswill be related to permanent but not necessarily to temporary components ofcash #ow, and that dividend decreases will be less frequent than increases andaccompanied by very poor performance. We test these predictions empiricallyand "nd them supported by the data. Our evidence is also consistent with therecent empirical work on the relation between dividend changes and futureearnings (see Bernartzi et al., 1997; DeAngelo et al., 1996). Although dividendsappear to be paid out of permanent earnings, we "nd little evidence of sub-sequent earnings improvements following dividend increases.

We then test the "nancial #exibility hypothesis cross-sectionally. In particu-lar, we expect to observe "rms choosing repurchases rather than dividends whenthe value of "nancial #exibility is highest; i.e. when there is a high likelihood that

M. Jagannathan et al. / Journal of Financial Economics 57 (2000) 355}384 357

2Guay and Harford (2000) report similar conclusions using a di!erent empirical approach.

3See, for example, Hulbert (1997). Putting their money where their mouths are, Forbes, April 21,1997.

4Although "rms are not required to announce share repurchases, the announcement of a "rm'sintent to repurchase shares on the open market (or through privately negotiated transactions) is one ofthe safe harbor provisions under the stock price manipulation provisions of the Securities andExchange Act. Additionally, a "rm must satisfy the four criteria detailed in S.E.C. rule 10b-18 of theSecurities and Exchange Commission's antimanipulation guidelines. These four criteria are: (1) on anyone day, "rms may not purchase more than 25% of the average daily volume of their own sharesduring the prior four weeks, block trades and privately negotiated transactions are exempt from thisguideline; (2) "rms may not purchase their own shares at the open, or during the last one-half hour oftrading; (3) "rms may not purchase their own shares at a price higher than the last independent bid, orthe last reported sale price; and (4) all purchases on a single day must be executed through the samebrokerage "rm. This rule was adopted in 1982 and caused an increase in the number of open marketrepurchase programs adopted due to the resolution of the legal ambiguity (see Ikenberry et al., 1995).

5 In terms of observable costs, dividend increases may also appear to be essentially costless;however, a dividend increase provides an explicit commitment to increase current payouts and animplicit commitment to keep future payouts at this increased level. An open market repurchaseannouncement provides no such commitments and there is no evidence of a reputational penalty forthose "rms failing to follow through with their announced repurchase program.

the cash #ows being paid out are temporary. We construct empirical measures ofwhether cash #ow is more or less likely to be temporary, and use these measures topredict whether a payout-increasing "rm is more or less likely to increasedividends, repurchases, or both in any given year. Such temporary cash #ows arelikely to occur when a "rm's cash #ows are made up of a higher proportion ofnon-operating income relative to operating income, when a "rm's earningsvolatility prior to the repurchase is high, and when the "rm's future cash #ows areexpected to decrease. In our data, each of these measures increases the likelihoodof a "rm using repurchases rather than dividends as a means of distributing cash#ows. These "ndings suggest that managers tend to use dividends to pay outpermanent cash #ows and repurchases to pay out temporary cash #ows.2

2. Measuring share repurchases

Although open market stock repurchases are increasingly common and haverecently received much publicity, they are surprisingly di$cult to measure.3A "rm can legally repurchase its own stock whenever it chooses withoutannouncing its intention to do so; however, by announcing a repurchaseprogram the "rm protects itself from liability under the stock price anti-manipulation provision of the Securities and Exchange Act.4 Since announcinga repurchase program is essentially costless to the "rm, making such an an-nouncement would appear to be the dominant strategy for "rms that areplanning to repurchase stock.5 There is little evidence of "rms repurchasing


6 In recent years it has become common for "rms to disclose the number of shares repurchased ontheir 10-Ks and 10-Qs. In 1995 approximately 75% of a random sample of "rms announcing openmarket repurchase programs made such disclosures. However, this appears to be a relatively recentphenomena; prior to 1992 very few "rms disclosed any details of their repurchasing activities andalmost none of the "rms reported the number of shares repurchased on their 10-Ks or 10-Qs.

7SDC now includes actual share repurchases in their database back through 1994, but this is notthe "gure generally reported in the popular press.

their own stock without having an announced program in place. Announce-ments are generally made to the various wire services, which are then collectedin the Securities Data Company (SDC) database and are often reported in TheWall Street Journal and other business publications. SDC releases the aggregatedollar value of the repurchase announcements in its database to non-sub-scribers, and this number is often quoted as a measure of total stock repurchases(see Hulbert (1997); Hylton (1995); Power 1995)).6

In fact, there are a number of reasons why the total released by SDC does notaccurately measure actual stock repurchases.7 First and foremost, the an-nouncements are simply a statement of the "rm's intention to repurchase itsstock; the "rm is in no way obligated to do so. While the majority of "rms followthrough with their announced open market repurchase programs, a signi"cantnumber of "rms repurchase few or no shares. Stephens and Weisbach (1998)document that while most "rms repurchase at least the number of sharesoriginally announced over the subsequent three years and frequently repurchasemore shares than originally announced, a signi"cant number of "rms repurchasevery few or no shares. Generally, those "rms repurchasing more shares thanoriginally announced initiate subsequent repurchase programs or announceexpansions of their existing programs, so it is reasonable to think of theannounced value as an upper bound on the amount the "rm will repurchase ina particular program.

Second, the aggregate dollar value of announced repurchase programsreported by SDC overstates the total value of the announced repurchaseprograms due to the way that the SDC constructs this "gure. SDC includesannouncements from a variety of sources, including wire services and The WallStreet Journal, so repurchase programs that are announced in more than one ofthe sources on di!erent days are included multiple times by SDC. In addition,SDC includes announcements of withdrawn programs and privately negotiatedrepurchases into its "gure. The privately negotiated repurchases are generallyannounced after the transaction has taken place and do not usually re#ect anyintention to repurchase additional shares.

Since share repurchases in the open market are neither observable at the timeof the transaction nor are they always directly measurable subsequently, it isimpossible to know with publicly available information exactly how many


8 If the "rm holds the shares as Treasury stock or retires the shares, then the value reported by theCompustat aggregation is an accurate re#ection of the cost of acquiring the shares. However, if theshares are subsequently distributed to employee bene"t plans, used for the exercise of stock optionsor reissued then this aggregation will also capture any change in market value since the time of theinitial repurchase.

shares were repurchased for all "rms. Since 1984, "rms have been required toreport the value of their repurchases on their Statement of Cash Flows and thisitem is included in the Compustat database as `Purchases of Common andPreferred Stocka (data item d115). However, this variable is an aggregation ofmany other types of transactions and overstates actual share repurchases,sometimes substantially. This aggregation includes conversions of other classesof stock into common stock, purchases of Treasury stock, retirements of commonor preferred stock, and redemptions of redeemable preferred stock. The purchasesof Treasury stock also include privately negotiated repurchases and self-tendero!ers in addition to open market repurchases. The privately negotiated transac-tions are often considerable; during our sample period there are $53.2 billion inprivately negotiated repurchases. Additionally, this aggregation will in someinstances misrepresent the cost of a repurchase.8 The combined overstatement ofrepurchases as reported to Compustat is potentially signi"cant.

Stephens and Weisbach (1998) suggest an alternative method of measuringshare repurchases, using the monthly decreases in shares outstanding asreported by CRSP adjusted for non-repurchase activity a!ecting sharesoutstanding such as stock splits and dividend reinvestment plans. Monthlydecreases in shares outstanding are not o!set with subsequent increases inshares outstanding since it is possible, even when a "rm is actively repurchasingshares, for shares outstanding to increase as the result of exercise of executivestock options, distribution of Treasury shares to employee bene"t plans or evencontemporaneous stock issues. We improve on the Stephens and Weisbach(1998) CRSP measure of share repurchases by adjusting for new stock issues;22 of the "rms in our sample issue new seasoned equity while having anactive open market repurchase program in place. It is not always possible,however, to determine how many shares were reissued to employee bene"t plansor through the exercise of executive stock options. Hence, this measure willunderestimate of share repurchases by the amount of shares contemporaneouslyreissued.

3. Estimates of aggregate repurchases and payouts

3.1. Repurchase announcements

Table 1 presents estimates of the aggregate values of the various types of sharerepurchase announcements. The "rst column contains the number and value of


Tab

le1

SDC

repu

rcha

sepro

gram

anno

unc

emen

ts

This

table

pre

sents

the

num

beran

dva

lue

ofa

llre

purc

hase

progr

amsan

nounc

edby

indu

strial"rm

san

dre

por

ted

by

the

Sec

uritie

sD

ata

Com

pany

(SD

C)

forth

eper

iod

1985

thro

ugh

1996

.Theth

reem

ostco

mm

on

met

hods

use

dby"rm

sto

repurc

has

eth

eirst

ock

(Dutc

hau

ctio

ns,s

ingl

e-price

tend

ero!er

san

dop

enm

arke

tre

purc

has

epro

gram

s)ar

epre

sente

d.The"rs

tfo

ur

colu

mns

pre

sent

the

num

ber

and

valu

eofal

lop

enm

arket

,priva

tely

nego

tiat

ed,D

utc

hau

ctio

nan

dte

nde

ro!

erre

purc

has

epro

gram

anno

unce

men

ts,re

spec

tive

ly.The

last"ve

colu

mns

pre

sent

alte

rnat

ive

mea

sure

sof

the

actu

alsh

ares

repur

chas

edsu

bse

quen

tto

the

repu

rchas

epr

ogra

man

nounc

emen

t.C

olum

n5

isth

eag

greg

ate

valu

eofs

har

esre

pur

chas

edon

the

ope

nm

arket

using

the

mea

sure

pro

pose

dby

Ste

phen

san

dW

eisb

ach

(199

8).T

heva

lue

ofsh

are

repu

rchas

esus

ing

this

mea

sure

for

onl

y"rm

sal

solis

ted

inth

eC

om

pust

atdat

abas

eis

report

edin

colu

mn

6.C

olu

mn

7pre

sents

the

aggr

egat

eva

lue

of`P

urc

has

esofC

om

mon

and

Pre

ferr

edSto

cka

(dat

aitem

d11

5)from

the

Com

pust

atdat

abas

efo

r"rm

sw

ith

announ

ced

repu

rchas

epro

gram

s;co

lum

n8

repo

rts

this

mea

sure

less

the

valu

eofD

utch

auct

ion

and

tende

ro!er

repur

chas

es.C

olum

n9

repo

rts

this

mea

sure

for

only"rm

sth

atan

nou

nced

anope

nm

arke

tre

purc

hase

pro

gram

.A

llva

lues

are

inm

illio

ns

ofdo

llars

.

Ope

nm

arket

repu

rchas

esM

easu

res

ofac

tual

open

mar

ket

shar

ere

pur

chas

es

12

34

56

78

9

Distinct

SD

Can

nounc

emen

tsP

riva

tely

nego

tiat

ed

Dut

chau

ctio

nre

pur

chas

epro

gram

s

Singl

e-price

tende

ro!

erre

purc

has

esC

RSP

mea

sure

ofsh

are

repurc

has

es

CR

SPre

purc

has

es}"rm

slis

ted

on

Com

pust

at

Val

ueof

stoc

kre

purc

has

esre

port

edby

Com

pust

at

Com

pust

atre

purc

has

es}

less

Dut

chau

ctio

n,se

lfte

nder

&priva

tely

nego

tiat

ed

Com

pust

atre

purc

has

es}"rm

sw

ith

announ

ced

open

mar

ket

repur

chas

epro

gram

sY

ear

dV

alue

dV

alue

dV

alue

dV

alue

1985

115

15,4

1655

4,14

65

1,09

269

16,6

395,

100

4,83

950

,418

28,5

418,

886

1986

161

19,8

5449

5,51

79

2,22

858

7,93

312

,924

12,2

3947

,253

31,5

7614

,485

1987

606

41,0

1934

2,79

18

1,49

457

5,74

111

,120

10,5

6857

,298

47,2

7228

,888

1988

183

28,5

1758

3,62

419

7,51

652

12,0

9438

,578

36,3

2054

,784

31,5

5028

,302

1989

356

52,5

7491

3,11

619

4,98

581

3,32

224

,970

24,5

9759

,995

48,5

7233

,539

1990

567

24,9

8898

2,81

09

1,93

166

6,42

926

,281

25,7

0643

,342

32,1

7225

,234

1991

211

13,8

7293

4,00

74

739

664,

290

9,07

69,

005

28,0

4819

,012

16,3

6119

9235

427

,835

671,

676

61,

597

471,

091

10,1

269,

940

34,0

5329

,689

19,1

6219

9332

428

,831

702,

210

560

443

553

16,6

6716

,111

37,1

6833

,801

24,6

3519

9454

753

,238

851,

065

1195

143

3,98

120

,532

19,7

0443

,102

37,1

0429

,502

1995

574

52,3

4385

12,9

518

1,44

434

1,55

335

,690

35,1

3278

,427

62,4

7946

,486

1996

755

112,

792

739,

334

172,

339

443,

576

47,3

1844

,533

91,9

8276

,733

63,3

21

Tota

l4,

753

471,

279

858

53,2

4712

026

,920

660

67,2

0225

8,38

224

8,69

462

5,87

047

8,50

133

8,80

1


9 In addition, programs are included into this category multiple times if the program is reported inmore than one source. There are a total of 93 repeat announcements and 34 withdrawn programsincluded in the SDC database, which are excluded from our statistical analysis. SDC's policy is torelease only the total number and value from this category in their database to nonsubscribers,which has resulted in somewhat larger values reported in the press than are reported here.

10SDC started collecting repurchase announcements in 1985, so there is not a comprehensivelisting of program announcements prior to 1985. However, it does appear that both the number andvalue of programs was substantially smaller prior to this point. For example, Vermaelen (1981) "ndsthat only 198 NYSE "rms initiated open market repurchase programs between 1970 and 1978.Grinblatt and Titman (1998) indicate that the dollar value of repurchases reported by Compustat issubstantially smaller prior to 1985 as well (p. 521).

distinct open market repurchase announcements as listed by SDC. SDC groupsall announcements relating to open market programs into one category. Inaddition to announcements of open market programs, SDC includes announce-ments of program withdraws and privately negotiated stock purchases.9 Wereport the number and value of privately negotiated repurchases, as well andDutch auction and single-price tender o!er repurchases in Columns 2, 3, and4 of Table 1.

Table 1 indicates that share repurchases have large #uctuations determinedby economy-wide factors. In 1985, there were only 115 announcements ofprograms, compared with 755 in 1996.10 In 1987 there was a sharp increase inprogram announcements, with 606 announcements. Many repurchase programswere initiated subsequent to the October market crash; 507 of the 606 repur-chase programs announced in 1987 occurred after October 19 with 400 of theseoccurring between October 19 and October 31. The late 1980s contained a largenumber of announcements of open market programs, with a high of 567 in 1990.There was a decrease during the recession years of the early 1990s, with 211announcements in 1991, 354 announcements in 1992 and 324 in 1993. However,since 1993 repurchase announcements have increased rapidly.

The relative popularity of other types of programs has also shifted over time.In the late 1980s, self-tender o!ers and Dutch auctions were relatively popularsince many were used as takeover defenses. Another reason for the prevalence ofself-tender o!ers was the leveraged recapitalizations common during this period.The largest years for self-tender o!ers, in value terms, were 1985 and 1988. Twocases led directly to such a large "gure for 1985, Unocal's $4.2 billion repurchaseand Union Carbide's $3.3 billion repurchase; Allegis' repurchase $2.8 billion wasthe largest repurchase in 1988. In the 1990s, there have been more privatelynegotiated transactions, with their values rising substantially in 1995 and 1996.

3.2. Values of repurchases

We present estimates of the dollar values of actual repurchases in Columns 5through 9 of Table 1. As discussed above, it is not always possible to know from


11This "nding contradicts the conclusions of Netter and Mitchell (1989). One potential explana-tion for the di!erence in "ndings is that Netter and Mitchell use just the change in the number ofshares during the year as their measure of repurchases, while we use the month-to-month change.This approach understates repurchases if there is any redistributions of shares during the year, whileours understates it only if shares are redistributed in the same month (or quarter, if the "rm onlyreports its outstanding shares quarterly).

publicly available data exactly how many shares a "rm repurchased in a givenyear. We present the CRSP repurchase measure suggested by Stephens andWeisbach (1998) with an adjustment for secondary equity o!erings occurringduring the program in Column 5. This measure provides a lower bound on thevalue of the repurchases. The total of $258 billion of repurchases is approxim-ately 55% of the total announced value of $471 billion. This value understatesthe ultimate quantity of repurchases from these programs because it does notinclude all of the repurchases from programs begun at the end of the sampleperiod. Since the two largest years for repurchase announcements were 1995 and1996, together making up more than 35% of the value of all announcements, thetotal completion rate is likely to be similar to the 70 to 80% completion ratesfound by Stephens and Weisbach (1998).

The estimates of aggregate actual share repurchases in Column 5 of Table1 suggest that repurchases vary substantially over time. The "rst few years of thesample appear to contain relatively few repurchases. However, we do not havedata on program announcements prior to 1985, so the estimate for 1985 is likelyto be particularly low since it only re#ects repurchases from programs begunduring that year. The post-crash announcements of 1987 led to a large numberof repurchases in 1988, 1989, and 1990. 11 Repurchases slowed noticeably in theearly 1990s and increased substantially in the mid-1990s. The numbers inColumn 5 measure aggregate repurchases for all programs from "rms on theCRSP tape. We are interested in relating these "gures to other variables takenfrom Compustat; therefore, we present the CRSP measure for the subset of "rmsthat are also on Compustat in Column 6 of Table 1. These numbers are similar,although slightly smaller, than those for the entire CRSP sample.

While the CRSP measure understates repurchases, the "gure reported onCompustat overstates them because it includes a number of other items inaddition to repurchases. We present Compustat's aggregate number inColumn 7 of Table 1. This "gure is considerably larger than the CRSP measurepresented in Column 6. We adjust the Compustat number in two ways tomeasure actual repurchases more accurately. First, since the Compustat "gureincludes Dutch auctions, privately negotiated deals, and self-tender o!ers, wesubtract them from the total and present the results in Column 8 of Table 1.Second, we restrict our calculations to include only "rms that we know areactively repurchasing stock. Stephens and Weisbach (1998) "nd that most "rmsrepurchase all of the stock they ultimately will during the two years subsequent


to the initiation of a program. Consequently, we de"ne a "rm-year as having anactive program as one in which a program is announced during that calendaryear or one of the two previous years. We present the aggregate repurchasemeasure for "rms with active programs de"ned this way in Column 9 of Table 1.

The estimates in Column 9, representing the Compustat repurchase measurefor "rms with active programs, are larger but of the same order of magnitude asthe CRSP measure for the same group of "rms presented in Column 6. This"nding is not surprising given that the Compustat measure overstates repur-chases while the CRSP measure understates it. A potentially troubling issue isthat the aggregate Compustat repurchase measure is so much larger than themeasure restricted to "rms with active programs; the total di!erence betweenthe two columns is around $240 billion even after adjusting for Dutch auctionsand self-tender o!ers. This di!erence could simply be due to the fact that theCompustat variable includes other components in addition to the repurchasevariable. Alternatively, it could occur because a substantial number of "rmsrepurchase stock without an announced repurchase program or repurchaseprogram announcements are missing in the SDC database. We do not wish todismiss this possibility, especially given that the incentive to announce a pro-gram for repurchasing comes from the safe harbor provisions of SEC Rule10b-18 and that "rms with repurchase programs appear to violate other provis-ions of this act (see Cook et al., 1997a).

Either of these possibilities implies that "rms are repurchasing stock withouta repurchase program listed by SDC. If so, the CRSP measures of actual sharerepurchases that only considers "rms with such programs will understateaggregate repurchases. To gauge the extent of such understatement, we calculatethe CRSP measure of repurchases based on drops in the number of sharesoutstanding for all "rm-years without active programs, but with positive repur-chases reported by Compustat. The total aggregated over all years, whichshould re#ect the total value of privately negotiated repurchases, self-tendero!ers, Dutch auction repurchases, and unannounced open market repurchasesduring our sample period, is $82 billion. This "gure, which admittedly is anunderestimate, is nonetheless substantially smaller than the sum total of theprivately negotiated repurchases, self-tender o!ers, and Dutch auctions sugges-ting that the number of "rms repurchasing stock without a program listed onSDC is relatively small.

Another issue is which of the two measures, the CRSP measure from Column6 or the Compustat-based measure from Column 9, more accurately re#ectsactual repurchases. To compare the accuracy of these two measures we rely onsurvey data from Cook et al. (1997a,b), who requested data on actual repur-chases from all "rms starting programs in 1993 and received such data from 64"rms. Of these 64 "rms, 35 had data on both CRSP and Compustat. Wecompare our measures of repurchases to each of the 64 "rm-years reportingpositive repurchases in the Cook et al. data set. The median CRSP measure is


12We use medians rather than means because there appear to be a number of outliers that wouldmake interpretation of averages di$cult. For example, there were seven cases in which "rmsreported positive repurchases to Cook et al. (1997a,b) but reported zero on their "nancial statementfor the variable that includes repurchases. There were two cases where the Stephens and Weisbach(1998) measure, which should understate repurchases, is greater than 500% of repurchases reportedin the survey, which potentially means that the corporations only reported a fraction of their actualrepurchases in their response to the request for data.

68% of the median dollar value reported in the Cook et al. data, while themedian Compustat measure is 113% of that value. 12 We interpret this "ndingas supportive of our argument that the CRSP measure underestimates repur-chases while Compustat overstates repurchases, and suggests that the Compus-tat measure adjusted for program announcements is likely to be more accurate.

Table 2 documents the magnitude of several alternative payout methods. The"rst three columns contain the estimates of aggregate repurchases: the CRSPmeasure for all "rms for which it can be computed, the CRSP measure for "rmslisted on Compustat, and the Compustat measure for all "rms with active openmarket repurchase programs. The table also contains aggregate dividends for"rms listed on Compustat and privately negotiated repurchases, self-tendero!ers, and Dutch auctions from SDC. Total payouts are the sum of aggregatedividends for "rms listed on Compustat and the dollar value of open marketrepurchases, privately negotiated repurchases, self-tender o!ers, and Dutchauctions from SDC.

It is clear from Table 2 that repurchases are growing rapidly, but thatdividends remain the predominate payout device. In 1996, dividends stillaccount for 65% of total payouts, compared to 69% in 1985. In 1985, dividendspaid by industrial "rms listed on Compustat were between 8 and 13 times thevalue of actual share repurchases and 450% of the announced value of all sharerepurchase programs. At the end of the sample period dividends for industrial"rms remain between 2 and 3 times as large as actual repurchases, regardless ofwhether one uses the underestimate or the overestimate to calculate repur-chases, and 125% of the dollar amount of announced share repurchases. Whilerepurchases have not nearly replaced dividends, the value of share repurchaseshas increased much more dramatically than the value of dividends. Even thoughrepurchases are growing rapidly, dividends are as well; dividends for Compustatindustrial "rms grew almost 25% between 1992 and 1996 and have more thandoubled over our entire sample period.

As has been recognized since at least Lintner (1956), dividends have histori-cally been characterized by relatively steady growth. This description appears tocharacterize our sample period as well. Aggregate dividends paid by industrial"rms listed on Compustat increased fairly steadily during our sample. Totalpayouts, however, were much more volatile than dividends. They dropped ina number of years, and, in fact, did not surpass 1988 levels until 1995. The sourceof this volatility is the other components of payouts, especially open market


Tab

le2

Pay

outm

easu

res

and

payo

ut

tech

niq

ues

This

table

pro

vides

com

par

isons

ofth

eag

greg

ate

valu

eof

alte

rnat

ive

pay

out

mea

sure

san

dte

chniq

ues

.C

olu

mn

1is

the

aggr

egat

eva

lue

of

shar

esre

pur

chas

edon

theopen

mar

ket

using

them

easu

repro

pose

dby

Ste

phen

san

dW

eisb

ach

(199

8).T

heva

lueofs

hare

repurc

has

esus

ing

this

mea

sure

foronl

y"rm

sal

solis

ted

inth

eC

om

pust

atda

tabas

eis

repor

ted

inco

lum

n2.

Colu

mn

3pre

sents

the

aggr

egat

eva

lue

of`P

urc

has

esofC

omm

on

and

Pre

ferr

edSto

cka(

dat

aitem

d11

5)from

the

Com

pust

atdat

abas

efo

r"rm

sw

ith

annou

nce

dre

purc

hase

pro

gram

s.The

tota

lagg

rega

team

oun

tofd

ivid

endspai

dfo

ral

l"rm

slis

ted

on

the

Com

pust

atIn

dust

rial

,Full

Cov

erag

ean

dR

esea

rch"le

sar

epre

sente

din

colu

mn

4.C

olum

ns

5th

rough

7ar

eth

eto

talva

lues

ofpriva

tely

nego

tiat

edst

ock

repurc

has

es,D

utc

hau

ctio

nre

purc

hase

san

dse

lf-te

nder

o!er

repur

chas

es,r

espec

tive

ly,a

sre

port

edby

Sec

urities

Dat

aC

ompan

y(S

DC

).The

tota

lpay

outs

pre

sente

din

colu

mn

8ar

eth

esu

mof

allopen

mar

ket

repur

chas

es,d

ivid

ends,

priva

tely

neg

otiat

edre

purc

has

es,D

utc

hau

ctio

nre

pur

chas

esan

dse

lf-ten

der

o!er

repu

rcha

ses.

All

valu

esar

ein

mill

ions

ofdolla

rs.

12

34

56

78

Yea

r

Ope

nm

arket

shar

ere

purc

hase

s}

CR

SPm

easu

re

CR

SPm

easu

reof

shar

ere

purc

hase

s}

"rm

slis

ted

onC

ompu

stat

Com

pust

atre

purc

hase

s}

"rm

sw

ith

anno

unce

dre

pur

chas

epr

ogra

ms

Agg

rega

tediv

iden

dsfo

ral

l"rm

sre

port

edon

Com

pust

at

Priva

tely

nego

tiat

edre

purc

has

es

Sel

f-te

nde

ro!er

repu

rchas

esD

utch

auct

ion

repu

rchas

es

Tot

alpay

outs"

colu

mns

3#4#

5#6#

7

1985

5,10

04,

839

8,88

667

,602

4,14

616

,639

1,09

298

,365

1986

12,9

2412

,239

14,4

8577

,574

5,51

77,

933

2,22

810

7,73

719

8711

,120

10,5

6828

,888

86,5

432,

791

5,74

11,

494

125,

457

1988

38,5

7836

,320

28,3

0210

3,99

63,

624

12,0

947,

516

155,

532

1989

24,9

7024

,597

33,5

3910

2,44

73,

116

3,32

24,

985

147,

409

1990

26,2

8125

,706

25,2

3410

7,71

82,

810

6,42

91,

931

144,

122

1991

9,07

69,

005

16,3

6110

7,77

54,

007

4,29

073

913

3,17

219

9210

,126

9,94

019

,162

107,

089

1,67

61,

091

1,59

713

0,61

519

9316

,667

16,1

1124

,635

109,

811

2,21

055

360

413

7,81

619

9420

,532

19,7

0429

,502

117,

426

1,06

53,

981

951

152,

925

1995

35,6

9035

,132

46,4

8614

2,11

812

,951

1,55

31,

444

204,

552

1996

47,3

1844

,533

63,3

2114

1,68

79,

334

3,57

62,

339

220,

257

Tota

l25

8,38

224

8,69

433

8,80

11,

271,

786

53,2

4767

,202

26,9

201,

757,

984


repurchases. At least at the aggregate level, repurchases appear to make upa disproportionately large share of the short-term #uctuations in payouts, whiledividends represent a more permanent component.

4. Firm-level analysis of payouts

4.1. Hypothesis development

To complement this aggregate evidence, we now analyze the factors a!ectingpayouts at the "rm level. Cross-sectional work on payout policy dates at least toFama and Babiak (1968), who "nd strong empirical support for the Lintnermodel of dividends during the period from 1946 to 1964. This sectionreexamines the Fama and Babiak results on a more recent sample. In addition,we empirically examine the factors that lead "rms to choose between dividendsand repurchases, focusing on the #exibility arguments discussed above.

Lintner (1956) argued that managers pay dividends out of long-run, sustain-able earnings. A company with stable earnings would thus tend to pay outa higher dividend than an otherwise similar growth "rm. His interviews withmanagers indicated that they like to increase dividends regularly and viewcutting dividends as extremely costly. They are therefore reluctant to makea dividend increase that will subsequently have to be reversed.

The Lintner model suggests a number of testable hypotheses regardingdividend behavior. Dividend-paying "rms should be larger than non-dividendpaying "rms, and should have higher and more stable cash #ows. Dividendincreases should be related to the permanent components of cash #ow, but notnecessarily to the temporary components of cash #ows. Dividend decreasesshould be relatively rare and occur only when "rms have truly bad performance.Subsequent to dividend increases or decreases, the good or bad operatingperformance should continue.

There are a number of non-mutually exclusive factors that potentially in#u-ence "rms in their choice between dividends and stock repurchases. Two suchfactors, taxes and employee stock options, are e!ectively unobservable to us.Although the tax system treats dividends and repurchases the same at thecorporate level, stock repurchases are generally tax-advantaged at the personallevel. The magnitude of this tax advantage depends on the cost bases andmarginal tax rates of the shareholders, which are not generally public informa-tion. However, it seems unlikely that taxes explain the more recent increase inrepurchase activity, since the tax advantage of repurchases was substantiallyreduced in 1986, approximately corresponding to the beginning of the currentrepurchase wave.

Fenn and Liang (1997), Jolls (1998), and Weisbenner (1998) stress the import-ance of employee stock options in the decision to repurchase stock rather than


13An alternative would be a `specially designated dividenda, which is commonly thought of as aninstitution for paying out such cash #ows. DeAngelo et al. (2000) document that specially designateddividends historically have been much more `regulara than their name implies, suggesting that "rmshistorically used them to pay out recurring rather than unusual cash #ows. In addition, theseauthors "nd that most "rms replaced their specially designated dividends with regular dividendslong before the current repurchase wave, suggesting that the repurchases studied here are nota replacement for such specially designated dividends.

increase dividends. Employee stock options could in#uence payout decisions fortwo reasons. First, employee options create incentives for stock repurchasesrather than dividends because the value of an option declines when a stock goesex-dividend but not when a company repurchases shares. Second, Weisbenner(1998) emphasizes that managers prefer to use repurchased rather than newly-issued shares for employee options to avoid diluting earnings per share. How-ever, directly controlling for the impact of executive options in our empiricalwork is infeasible given our sample size, since data on executive stock optionsare not available in machine-readable form. Nonetheless, the combination ofthese two option-induced reasons is undoubtedly an important factor in thecurrent repurchase trend.

Perhaps the most commonly discussed motive for repurchases concernsasymmetric information. Dann (1981), Vermaelen (1981), and Ikenberry et al.(1995) suggest that stock prices rise on the announcement of a repurchaseprogram. In addition, Comment and Jarrell (1991) "nd that the abnormalreturns observed around the announcement of a repurchase program are in-versely related to recent stock price performance leading up to the repurchaseannouncement. These results are consistent with the view that asymmetricinformation is an important motive for stock repurchases.

We, like Guay and Harford (2000), focus on the impact of the cash #ow'spermanence on the choice between dividends and repurchases. We hypothesizethat dividends will be used to pay out cash #ows that are likely to be permanent,while stock repurchases will be used for cash #ows that are not likely to be ableto be sustained inde"nitely.13

This view leads to a number of testable predictions. First, when there is moreuncertainty about future cash #ows, we expect the "rm to utilize repurchases toa greater extent. Second, since operating cash #ows tend to be more permanentthan non-operating cash #ows, we expect a positive relation between operatingincome and dividends, while repurchases are more likely to be related tonon-operating income. Third, if repurchases are more likely to re#ect temporarycash #ows, we hypothesize that dividend-increasing "rms will have largersubsequent cash #ows than repurchasing "rms. Finally, if "rms repurchase stockbased on management's belief that the stock is undervalued, we would expectthat "rms selecting repurchases would have lower stock returns prior to thepayout change.


4.2. Variable construction

There are a number of variables that are likely to be related to the decision topay out cash #ows. Using Compustat, we construct proxies of these variables forthe periods before and after each potential payout increase. Our primary interestis the determinants and method of cash payouts. Consequently, our analysisfocuses on cash #ow and its components: operating income, non-operatingincome, and capital expenditures. In order to reduce noise induced by year-to-year variations in many of the variables, we use three-year averages unlessotherwise noted. Average values for years !3 through !1 relative to the payoutchange are used for variables prior to the payout change; average values for years0 through #2 relative to the payout change are used for the variables subsequentto the payout change. The sample for our cross-sectional analysis is limited to theperiod from 1985 to 1994 to allow for measurement of subsequent cash #ows.

We use the book value of total assets (data item d6) in the year prior to thepayout change (year !1) as our measure of "rm size. Operating income is theaverage ratio of operating income (data item d13) to total assets. Non-operat-ing income is the average ratio of non-operating income (data item d61) to totalassets. The standard deviation of operating income is the standard deviation ofthe ratio of operating income to total assets measured over the 5-year periodfrom year !4 through 0. Capital expenditures is the average ratio of capitalexpenditures (data item d128) to total assets. The lagged dividend payout ratiois the prior year's ratio of total dividends (data item d21) to net incomeavailable to common shareholders (data item d237). The market-to-book ratiois the average ratio of the market value of equity, given by the year-end price pershare (data item d24) multiplied by the number of shares outstanding (dataitem d25), to the book value of equity (data item d62). The debt ratio is theaverage ratio of long-term debt (data item d9) to total assets. Institutionalownership is the percentage of shares outstanding owned by institutions in theyear prior to the payout change obtained from Compact Disclosure and is onlyavailable from 1991 through 1994. The increase in dividends divided by themarket value of equity is the ratio of total dividends for the prior year minustotal dividends for the current year to the market value of equity. The increase inrepurchases divided by the market value of equity is the announced value of theopen market repurchase program obtained from SDC to the market value ofequity, given by the year end price per share (data item d24) multiplied by thenumber of shares outstanding (data item d25). Stock returns are computedfrom CRSP.

4.3. Univariate diwerences

Table 3 presents univariate statistics on the di!erences between "rms thatchoose di!erent payout methods. Column 1 provides statistics for the entire


Tab

le3

Firm

char

acte

rist

ics

acco

rdin

gto

pay

out

met

hod

This

tabl

epro

vides

mea

ns(m

edia

nsin

par

enth

eses

)ofv

ario

us"rm

char

acte

rist

icsfo

rth

eca

tego

ries

ofposs

ible

pay

outco

mbi

nations

.Col

um

n1

consist

sof

all"rm

-yea

rslis

ted

onSta

ndar

dan

dP

oor's

Com

pust

at(C

ompu

stat

)ta

pes

from

1985

thro

ugh

1994

;ut

ilities

and"nan

cial"rm

sar

eex

clud

edin

all

cate

gories

.Colu

mn

2co

ntai

ns"rm

-yea

rsw

ith

incr

ease

dto

talp

ayou

tsre

sultin

gfrom

eith

erin

crea

sing

divi

dend

s,in

itia

ting

are

purc

hase

prog

ram

,orbot

h;

the

sam

ple

inco

lum

n2

isth

euni

on

ofth

esa

mple

sin

colu

mns

3,4

and

5.C

olum

n3

cont

ains"rm

-yea

rsw

ith

initia

tion

sor

expa

nsions

ofre

pur

chas

epro

gram

s.C

olu

mn

4co

nta

ins"rm

-yea

rsw

ith

both

repu

rchas

esan

ddiv

iden

din

crea

ses.

Colu

mn

5co

ntai

ns"rm

-yea

rsw

ith

div

iden

din

crea

ses.

Col

um

n6

conta

ins"rm

-yea

rsw

ith

noch

ange

incu

rren

tpay

outle

vels;t

he"rm

sm

usthav

epositive

pay

out

s,but

div

iden

dsha

veno

tch

ange

d.C

olu

mn

7co

nta

ins

"rm

-yea

rsw

ith

nopa

youts

(nei

ther

div

iden

dsno

rst

ock

repu

rchas

es)in

eith

erth

ecu

rren

tor

prior

year

.Colu

mn

8co

nta

ins"rm

-yea

rsw

ith

dec

reas

esin

tota

lpa

youts

;th

esa

mpl

ein

colu

mn

8is

the

uni

on

ofth

esa

mple

sin

clud

edin

colu

mns

9an

d10

.C

olu

mn

9co

ntai

ns"rm

-yea

rsw

ith

no

repur

chas

ean

nounc

emen

tin

thecu

rren

tye

ar,b

utw

ith

are

pur

chas

ein

itia

tion

orex

pan

sion

inth

epriorye

ar.C

olu

mn

10co

nta

ins"rm

-yea

rsw

ith

divi

dend

dec

reas

es;

ther

ear

e30"rm

-yea

rsw

ith

both

divi

den

ddec

reas

esan

dre

pur

chas

ede

crea

ses(a

sde"ned

forco

lum

n9)

incl

uded

inth

isco

lum

n.D

ivid

end

incr

ease

san

ddec

reas

esar

ede"

ned

on

aper

shar

ebas

isad

just

edfo

rst

ock

split

san

din

clude

onl

yre

gula

rdi

vide

nds

asre

port

edby

CR

SP

.A

llst

atistics

are

com

pile

dfrom

data

obta

ined

from

Com

pus

tatunl

essoth

erw

ise

stat

ed;i

nord

erto

reduc

enoise

indu

ced

by

year

-to-y

earva

riat

ionsin

man

yofth

eva

riab

les,

we

use

thre

e-ye

arav

erag

esun

less

other

wise

stat

ed.A

vera

geva

lues

for

year

s!

3th

roug

h!

1re

lative

toth

epa

yout

chan

gear

eus

edfo

rva

riab

lespr

iorto

pay

outc

hange

;ave

rage

valu

esfo

rye

ars0

thro

ugh

#2

rela

tive

toth

epay

outch

ange

areus

edfo

rth

eva

riab

lessu

bse

que

ntto

the

pay

out

chan

ge.F

irm

size

isto

talas

sets

(dat

aitem

d6)

inth

eye

arpr

ior

toth

epay

out

chan

ge(y

ear!

1).O

per

atin

gin

com

eis

the

aver

age

ratio

ofoper

atin

gin

com

e(d

ata

item

d13

)to

tota

lass

ets.

Non-o

per

atin

gin

com

eis

the

aver

age

ratio

ofn

on-

oper

atin

gin

com

e(d

ata

item

d61

)to

tota

lass

ets.

The

stan

dard

dev

iation

ofop

erat

ing

inco

me

isth

est

anda

rddev

iation

ofth

era

tio

ofop

erat

ing

inco

me

(dat

aitem

d13

)to

tota

lass

etsm

easu

red

over

the

5-ye

arpe

riod

from

year

!4

thro

ugh

0.C

apital

expe

ndi

ture

sis

the

aver

age

ratio

ofca

pital

expe

ndi

ture

s(d

ata

item

d12

8)to

tota

las

sets

.The

lagg

eddi

vide

ndpay

out

ratio

isth

eprior

year's

ratio

of

tota

ldiv

iden

ds(d

ata

item

d21

)to

net

inco

me

avai

labl

eto

com

mon

shar

ehol

ders

(dat

aitem

d23

7).The

mar

ket-

to-b

ook

ratio

isth

eav

erag

era

tio

ofth

em

arket

valu

eof

equi

ty,g

iven

by

the

year

-end

price

persh

are

(dat

aitem

d24

)multip

lied

by

the

num

berof

shar

esou

tsta

ndin

g(d

ata

item

d25

),to

the

boo

kva

lue

ofeq

uity

(dat

aitem

d62

).The

deb

tra

tio

isth

eav

erag

era

tio

ofl

ong-

term

debt(d

ata

item

d9)

toto

tala

sset

s.In

stitutiona

low

ners

hip

isth

epe

rcen

tage

ofsh

ares

outs

tand

ing

owned

by

inst

itutions

inth

eye

arpr

ior

toth

epa

youtch

ange

from

Com

pac

tD

iscl

osure

.The

prior

year

mar

ket

retu

rns

are

the

CR

SPm

onth

lyre

turn

sge

om

etrica

llycu

mula

ted

for

the

year

prior

(yea

r!

1)to

the

year

inques

tion

;cu

rren

tye

arm

arke

tre

turn

sar

eth

eC

RSP

month

lyre

turn

scu

mula

ted

forth

ecu

rren

tye

ar(y

ear0)

.The

incr

ease

indiv

iden

ds/

mar

ket

valu

eofe

qui

tyis

the

ratio

ofto

taldiv

iden

ds

for

the

prio

rye

ar(y

ear!

1)m

inus

tota

ldi

vide

nds

for

the

curr

ent

year

(yea

r0)

toth

em

arket

valu

eofeq

uity.

The

incr

ease

inre

pur

chas

es/m

arke

tva

lue

ofeq

uity

isth

ean

nou

nce

dva

lue

ofth

eope

nm

arke

tre

purc

hase

pro

gram

obta

ined

from

SDC

toth

em

arke

tva

lue

ofeq

uity

.


Firm

char

acte

rist

ics

Full

sam

ple

Firm

sin

crea

sing

tota

lpa

youts

Firm

sre

purc

has

ing

shar

es

Firm

sre

purc

has

ing

and

incr

easing

div

iden

ds

Firm

sin

crea

sing

div

iden

ds

Firm

sno

tch

angi

ngcu

rren

tpa

youts

Firm

sth

atha

vem

ade

no

payo

uts

inprior

or

curr

ent

year

s

Firm

sdec

reas

ing

tota

lpa

youts

Firm

sde

crea

sing

repurc

has

es

Firm

sdec

reas

ing

div

iden

ds

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

Obse

rvat

ions

54,0

035,

873

1,48

572

33,

665

3,75

442

,023

2,35

31,

073

1,28

0

Prior

orpa

yout

chan

ge:

Firm

size

(inm

illio

ns

ofdo

llors

)87

0.30

41,

939.

978

983.

450

3,56

8.77

71,

999.

777

1,67

1.83

858

1.39

81,

142.

695

979.

937

1,28

7.04

(40.

265)

(290

.1)

(109

.15)

(836

.23)

(345

.45)

(225

.23)

(20.

16)

(146

.64)

(119

.29)

(188

.10)

Oper

atin

gin

com

e(%

)6.

906

21.1

5819

.128

22.4

0121

.695

17.1

402.

309

15.6

2618

.248

13.3

16(1

1.93

)(1

9.79

)(1

6.99

)(2

1.03

5)(2

0.25

)(1

5.58

)(8

.70)

(13.

50)

(16.

76)

(11.

51)

Non

oper

atin

gin

com

e(%

)2.

124

1.72

52.

116

1.60

91.

616

1.69

02.

302

1.95

52.

106

1.82

9(1

.37)

(1.3

7)(1

.71)

(1.3

05)

(1.2

8)(1

.25)

(1.3

9)(1

.44)

(1.5

75)

(1.3

35)

Stan

dard

dev

iation

of12

.791

(5.6

88)

10.2

914.

025

4.31

95.

693

15.4

298.

892

10.2

047.

747

oper

atin

gin

com

e(%

)(6

.49)

(3.6

9)(6

.18)

(2.9

3)(3

.34)

(4.3

3)(8

.06)

(5.7

4)(5

.835

)(5

.59)

Cap

ital

expen

ditu

res

(%)

11.3

599.

497

11.2

198.

627

9.00

38.

955

12.1

099.

984

10.6

159.

259

(6.5

5)(7

.465

)(7

.53)

(7.4

85)

(7.4

4)(7

.13)

(6.1

8)(6

.64)

(7.1

3)(6

.30)

Lag

ged

divi

dend

pay

out

12.1

7232

.079

15.7

539

.585

37.3

0946

.279

4.98

531

.482

17.9

843

.528

ratio

(%)

(0.0

)(2

4.41

)(0

.0)

(30.

98)

(29.

025)

(28.

82)

(0.0

)(8

.905

)(0

.0)

(23.

36)

Mar

ket

-to-

book

ratio

2.24

72.

297

2.27

22.

548

2.25

71.

685

2.36

41.

764

2.04

51.

505

(1.6

9)(1

.9)

(1.7

78)

(2.1

49)

(1.9

05)

(1.4

30)

(1.7

38)

(1.3

8)(1

.613

)(1

.189

)

Deb

tra

tio

(%)

24.6

3318

.247

20.4

3116

.472

17.7

5320

.376

25.1

8722

.751

20.7

9624

.436

(18.

382)

(16.

115)

(16.

93)

(15.

27)

(16.

1)(1

8.03

5)(1

8.74

)(2

0.47

)(1

7.59

)(2

2.27

)


Tab

le3

(con

tinu

ed)

Firm

char

acte

rist

ics

Full

sam

ple

Firm

sin

crea

sing

tota

lpa

youts

Firm

sre

purc

has

ing

shar

es

Firm

sre

purc

has

ing

and

incr

easing

div

iden

ds

Firm

sin

crea

sing

div

iden

ds

Firm

sno

tch

angi

ngcu

rren

tpa

youts

Firm

sth

atha

vem

ade

no

payo

uts

inprior

or

curr

ent

year

s

Firm

sdec

reas

ing

tota

lpa

youts

Firm

sde

crea

sing

repurc

has

es

Firm

sdec

reas

ing

div

iden

ds

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

Inst

itutiona

low

ners

hip

(%)

25.0

2742

.566

35.4

8650

.338

44.8

5940

.792

17.4

1632

.788

35.4

4928

.811

(18.

875)

(44.

06)

(33.

885)

(53.

49)

(47.

08)

(40.

61)

(9.5

1)(3

1.62

)(3

4.54

)(2

6.55

)

Prior

year

stock

retu

rns

0.12

80.

185

!0.

011

0.15

80.

259

0.11

80.

120

0.08

40.

222

!0.

064

(0.0

10)

(0.1

49)

(!0.

080)

(0.1

11)

(0.2

07)

(0.0

76)

(!0.

032)

(0.0

10)

(0.1

35)

(!0.

114)

Incr

ease

indiv

iden

ds/

0.00

0.57

50.

003

0.71

90.

764

0.10

10.

04!

2.27

5!

0.01

0!

4.29

6m

arket

valu

eof

equi

ty(%

)(0

.0)

0.39

9(0

.0)

(0.4

70)

(0.5

25)

(0.0

61)

(0.0

)(!

0.38

0)(0

.0)

(!2.

383)

Incr

ease

inre

purc

has

es/

**

7.97

26.

397

**

**

!6.

746

*

mar

ket

valu

eof

equi

ty(%

)*

*(5

.704

)(4

.953

)*

**

*(!

5.64

7)*

Subse

que

ntto

payo

utch

ange

:O

per

atin

gin

com

e(p

ost

)(%

)4.

082

18.9

8514

.752

20.9

1420

.139

14.2

490.

747

10.0

3713

.082

7.48

4(1

0.03

)(1

8.27

6)(1

3.99

3)(1

9.80

4)(1

9.33

7)(1

3.25

7)(7

.865

)(9

.382

)(1

2.91

8)(7

.451

)

Non-

oper

atin

gin

com

e1.

716

1.27

21.

536

1.07

41.

205

1.17

21.

858

1.15

41.

359

0.98

3(p

ost

)(%

)(0

.977

)(0

.907

)(1

.110

)(8

.368

)(0

.853

)(0

.783

)(1

.027

)(7

.87)

(0.9

85)

(0.6

42)

Cap

ital

expen

ditu

res

(post

)(%

)10

.879

9.49

19.

690

8.69

49.

569

8.31

311

.601

7.02

68.

494

5.79

1(5

.455

)(7

.397

)(6

.371

)(7

.362

)(7

.814

)(6

.273

)(4

.994

)(4

.492

)(5

.786

)(3

.476

)

Cur

rentye

arst

ock

retu

rns

0.06

40.

211

0.10

70.

252

0.24

20.

095

0.03

3!

0.06

9!

0.08

0!

0.05

6(0

.0)

(0.1

66)

(0.0

40)

(0.2

12)

(0.1

96)

(0.0

51)

(!0.

068)

(!0.

133)

(!0.

147)

(!0.

116)


14Dividend increases and decreases are de"ned on a per share basis adjusted for stock splits andinclude only regular dividends as reported by CRSP.

sample, which consists of all "rm-years on Compustat from 1985 through 1994excluding utilities and "nancial "rms. Column 2 presents the means andmedians for "rm-years with increased total payouts. Columns 3, 4 and 5 breakdown the payout-increasing "rm-years: Column 3 includes observations withinitiations or expansions of repurchase programs, Column 4 contains cases withdividend increases in addition to an initiation or expansion of a repurchaseprogram, and Column 5 includes the "rm-years with dividend increases but nonew repurchase program.14 Column 6 presents the means and medians for"rm-years with positive payouts that remain constant from the previous year,while Column 7 includes observations for "rm-years with no payouts (neitherdividends nor stock repurchases) in either the current or prior year.Column 8 presents the means and medians for "rm-years with decreases in totalpayouts. Columns 9, and 10 break down these payout-decreasing observations:Column 9 contains observations with no repurchase announcement in thecurrent year, but with a repurchase initiation or expansion in the prior year,while Column 10 presents the means and medians for "rm-years with dividenddecreases. The 30 "rm-years in which both dividends and repurchases decreaseare included with the dividend decreases in Column 10.

4.3.1. Dividend changesTable 3 allows us to examine the empirical predictions of the Lintner model

discussed above. The results are consistent with the predictions of this model.Firms making payouts are substantially larger than "rms that have not madepayouts in the current or prior year. Dividend-increasing "rms have higheroperating incomes and similar non-operating incomes than "rms that do notchange payouts. Given that operating cash #ows are relatively permanent whilenon-operating incomes are more temporary, this "nding suggests that dividendincreases are funded out of permanent cash #ows. The standard deviation ofoperating income, a proxy for the stability of cash #ows, is lower for thedividend-increasing "rms than for the "rms making no change to existing (butpositive) payouts. The standard deviation of operating income for the "rmskeeping current payouts constant is lower than the standard deviation ofoperating income for "rms that have not historically made any payouts. Sub-sequent to the payout change, dividend-increasing "rms continue to havesubstantially higher operating income than "rms that do not change payouts,which is in turn higher than subsequent operating income for "rms that have nothistorically paid dividends. These di!erences are signi"cant at the 1% levelusing two-tailed test of means and a Wilcoxon nonparametric test. Allthese results are consistent with the predictions of the Lintner model discussedabove.


15This di!erence is potentially due to taxes: If low-tax investors purchase stocks of dividend-paying companies and high-tax investors purchase stocks of repurchasing companies, then stock-holders of dividend-paying "rms will prefer "rms to increase dividends and stockholders ofnon-dividend paying "rms prefer to increase repurchases. Many thanks to a referee for suggestingthis possibility.

Finally, the Lintner view suggests that "rms will avoid dividend decreases if atall possible, so that dividend decreases will be less frequent than increases andassociated with genuinely poor performance. The data in Table 3 are consistentwith this view. There were 1,280 cases in which the nominal dividend per sharedecreases; in comparison, there are 4,388 cases of dividend increases (723 ofthese observations also repurchase stock in the same year). However, thedividend decreases were much larger than the increases (4.3% of equity valuecompared to 0.76% of equity value). In these cases, the "rms are indeed in"nancial di$culty. During the year prior to dividend decreases, the average "rmpaid out 43.53% of net income as dividends. The average operating income toassets was substantially lower for these "rms than for "rms keeping payoutsconstant or increasing them. Subsequent to the payout decrease, the averageoperating income over assets for these "rms decreases to 7.48%, just over a thirdof the average for dividend-increasing "rms.

4.3.2. The choice between stock repurchases and dividendsTable 3 also documents systematic di!erences between the dividend-increas-

ing and repurchasing "rms. Dividend-increasing "rms are generally larger thanrepurchasing "rms. Prior to the payout increase, dividend-increasing "rms havehigher operating cash #ows. However, consistent with the "nancial #exibilityhypothesis, non-operating cash #ows are higher for repurchasing-increasing"rms than for dividend-increasing "rms. The standard deviation of the operat-ing income for the repurchasing "rms is about twice as high as for the dividend-increasing "rms, suggesting that cash #ows for repurchasing "rms are substan-tially more uncertain than they are for dividend-increasing "rms.

Perhaps the most striking number in Table 3 is the dividend payout ratioprior to the decision. The average dividend payout ratio for "rms that increasedboth dividends and repurchases is 39.6% and is 37.3% for "rms that increaseonly dividends. In comparison, the average lagged dividend payout for repur-chasing "rms is only 15.8% and the median is zero.15 In fact, the median "rmincreasing repurchases previously did not pay out any dividends at all. Thislarge di!erence suggests that dividend-increasing "rms are following a historicalpolicy of paying out cash #ows, while repurchases are less frequent eventsallowing "rms to pay out a cash surpluses that are likely to be temporary.

Subsequent to the payout increase, operating income is substantially andstatistically signi"cantly higher for the dividend-increasing "rms than for therepurchase increasing "rms. Non-operating income continues to be higher for


repurchasing "rms than for either of the other payout-increasing types, withboth di!erences statistically signi"cant at the 1% level. These results are consis-tent with the view that repurchases are used to pay out temporary cash #owswhile dividends are used to pay out permanent ones.

Another di!erence between repurchases and dividends concerns the size of thepayouts. Firms increasing repurchases announce programs with announcedtargets equal to about 8% of equity value. Given that they typically buy about75% of this target over the subsequent two or three years (see Stephens andWeisbach, 1998), this value implies an increase in annual payouts of about 2%to 3% of equity value. In contrast, dividend increases are much more commonbut average only 0.76% of equity value. Repurchases allow "rms to distributelarge quantities of cash to shareholders relatively quickly. Thus, the quantity ofcash a "rm wants to distribute is likely to be an important factor in choosingbetween dividends and repurchases.

One of the common explanations of stock repurchases is that they occur whenmanagers believe their stock is undervalued (see Dann, 1981; Vermaelen, 1981).Consistent with this view is evidence that "rms tend to announce programsfollowing poor stock market performance (see Comment and Jarrell, 1991;Stephens and Weisbach, 1998). This result holds in our sample as well; theaverage stock return for "rms that announce repurchase programs but do notincrease dividends during the year prior to the announcement is !1.1% andthe median is !0.8%. In contrast, the average return for "rms announcingdividend increases but not a repurchase program is 25.9% and the median is20.7%. This large di!erence is statistically signi"cant at the 1% level.

Table 3 also presents the mean and median level of institutional ownership foreach category. These observations are consistent with the view that institutionsprefer established companies that have performed well recently. Firms that areincreasing payouts have substantially higher institutional ownership than "rmskeeping payouts constant or decreasing them. Among the payout-increasing"rms, the ones increasing repurchases but not dividends have the lowest levels ofinstitutional ownership. One explanation of this "nding is that many institu-tions are tax-exempt, so that they do not share in the tax bene"ts of repurchases.Given this interpretation, this result is consistent with the arguments of Allenet al. (1998), who suggest that one motive of dividend payments is to attractinstitutional investors that will subsequently provide monitoring bene"ts to the"rm.

Overall, the univariate comparisons in Table 3 suggest that dividend-increas-ing and repurchasing "rms are noticeably di!erent. Repurchasing "rms havemore uncertain cash #ows and have not historically had high payout ratios.They have lower operating incomes but higher non-operating incomes. Firmstend to increase repurchases following poor stock-market performance whilethey tend to increase dividends following good performance. Finally, institu-tions tend to favor dividend-paying over repurchasing "rms.


16Weisbenner (1998) uses multinomial logit similarly to estimate the impact of stock optionprograms on the choice between repurchases and dividends.

4.4. Multivariate diwerences

The univariate comparisons suggest that there are di!erences between "rms'payouts depending on the "rms' characteristics. Since, these characteristics arecorrelated with one another we examine these di!erences in a multivariatecontext.

To do so, we estimate a model in which "rm characteristics predict payoutpolicy. One complicating factor is that there are many potential choices ofpayout methods, roughly corresponding to the columns of Table 3. Given thatthese choices do not have a clear ordering, a natural approach to this problem isa multinomial logit model.16

The multinomial logit model assumes that the probability of an outcome> isgiven by:

Prob(>"j)"eb@

j xi

1#+Jk/-

eb@kxi

for j"1, 2,2, J, (1)

and

Prob(>"0)"1

1#+Jk/-

eb@k xi

. (2)

Each bjis a vector of dimension equal to the number of independent variables,

which can be estimated by maximum likelihood. These estimates are conve-niently expressed in terms of the log odds of any two outcomes, which equal:

lnCP

ijPikD"x@

i(b

j!b

k). (3)

The coe$cient on each independent variable in this equation equals the di!er-ence between the bs for two di!erent outcomes. The p-value on such a coe$cientprovides a test of the hypothesis that the independent variable a!ects theprobability of each outcome in the same manner.

Tables 4 and 5 provide estimates of a multinomial logit models predictingpayout method choices. The di!erence between the two tables is that Table5 includes institutional ownership as an independent variable, which limits thenumber of observations since our institutional ownership data is not availablefrom Compact Disclosure before 1991. Each model reported here groups payoutchoices into four categories: increasing only repurchases, increasing only divi-dends, increasing both repurchases and dividends, or not increasing payouts.We limit the categories because we are primarily interested in the choicebetween dividends and repurchases. Adding more outcome types complicates


both the estimation and the reporting of our results, since each additional choicerequires one extra parameter per choice per independent variable. To ensurethat limiting the model in this fashion does not a!ect the inferences we drawfrom our results, we have also estimated similar models allowing for payoutdecreases to be a separate choice and models allowing for repurchase decreasesand dividend decreases to be separate choices. The results reported below arerobust to these speci"cations.

The coe$cients reported in Tables 4 and 5 are estimates of the log-oddsbetween each pair of categories. The multivariate results tell a similar story tothe multivariate results. The category of "rms that do not increase payouts aresmaller and have worse operating income both before and after the potentialpayout increase than any of the payout-increasing categories. Higher standarddeviations of operating income also predict a higher probability of not payingout cash. Within the categories of "rms that do increase payouts, higheroperating incomes increase the probability of a dividend increase, whilehigher standard deviations of operating incomes increase the probability ofrepurchases. One univariate result that does not hold here concerns non-operating income prior to the potential payout, which, in a multivariate context,is not signi"cantly related to the choice between dividends and repurchases.However, the non-operating income subsequent to the payout change is signi"-cantly positively related to the likelihood of repurchases. Finally, dividend-increasing "rms and "rms that both increase dividends and repurchases havesigni"cantly higher market returns than repurchasing "rms for both prior andcurrent years.

Finally, the results reported in Table 5 incorporate the impact of institu-tional ownership, similar to DeAngelo et al. (2000). Institutional ownership iscorrelated with both the "rm's decision to increase payouts and the choice ofpayout methods. Similar to the univariate results, "rms that increase payoutshave signi"cantly higher institutional ownership than "rms that do not increasepayouts. In addition, dividend-increasing "rms have higher institutional owner-ship than repurchasing "rms.

Overall, our multivariate analysis is consistent with the univariate compari-sons discussed earlier and both generally support the hypothesis that the"nancial #exibility inherent in open market repurchase programs is an impor-tant consideration in the choice of payout methods.

5. Conclusions

In this paper we analyze the recent rise in open market stock repurchases. Westart with a complete listing of program announcements supplied by SDC. Fromthese announcements, we construct a database consisting of both an underesti-mate and an overestimate of actual repurchases for every Compustat "rm


Tab

le4

Oper

atin

gan

dst

ock

perfor

man

cech

arac

terist

ics

a!ec

ting

a"rm's

choi

ceofpay

out

met

hod

This

tabl

epre

sent

sth

ere

sults

from

estim

atio

nofa

multin

omia

llo

git

mode

lofva

rious

oper

atin

gch

arac

terist

ics

and

stoc

kre

turn

sth

ough

tto

pred

ict

a"rm's

choic

eof

alte

rnat

ivem

etho

dsof

distr

ibut

ing

cash#ow

sin

agi

ven

year

.Eac

hco

lum

npr

ovid

esth

epar

amet

eres

tim

ates

obt

ained

from

thelo

g-od

dsra

tios

.The"rs

tco

lum

nco

mpar

es"rm

-yea

rsw

ith

no

incr

ease

inpa

yout

sto"rm

-yea

rsin

whi

chpay

outs

incr

ease

thro

ugh

the

useofr

epurc

has

es,i

ncre

ased

div

iden

dsorbo

th.T

he

seco

ndco

lum

nco

mpar

es"rm

-yea

rsw

ith

repurc

hase

prog

ram

initia

tionsorex

pan

sionsto

the

other

pos

sible

outc

om

es.T

he

third

colu

mn

com

par

es"rm

-yea

rsw

ith

both

stock

repur

chas

esan

ddiv

iden

din

crea

sesto

theoth

eral

tern

ativ

esan

dth

efo

urt

hco

lum

nco

mpa

res"rm

-yea

rsw

ith

div

iden

din

crea

ses

toth

eot

her

alte

rnat

ives

.All

indep

enden

tva

riab

les

are

the

sam

eas

de"ned

inT

able

3. Firm

-yea

rsw

ith

Firm

-yea

rsw

ith

repur

chas

eF

irm

-yea

rsw

ith

no

repur

chas

ean

nounce

men

tsan

dF

irm

-yea

rsw

ith

incr

ease

inpay

out

san

nou

nce

men

tsdiv

iden

din

crea

ses

div

iden

din

crea

ses

Coe$

cien

tp-

valu

eC

oe$

cien

tp-

valu

eC

oe$

cien

tp-

valu

eC

oe$

cien

tp-

valu

e

Firm

-yea

rsw

ith

no

incr

ease

inpa

youts

Inte

rcep

t3.

329

0.00

04.

024

0.00

02.

353

0.00

0Firm

size

!14

.021

0.01

0!

31.1

720.

000

!17

.918

0.00

0O

per

atin

gin

com

e(p

rior)

!2.

971

0.00

0!

7.35

90.

000

!6.

836

0.00

0N

on-

ope

rating

inco

me

(prior)

!3.

159

0.01

3!

2.65

70.

324

!1.

297

0.37

6Pay

out

ratio

(prior)

0.02

80.

618

!0.

126

0.00

0!

0.12

10.

000

Std.d

evia

tion

ope

rating

inco

me

2.56

70.

000

30.5

740.

000

27.6

990.

000

Oper

atin

gin

com

e(p

ost

)!

2.61

90.

000

!7.

052

0.00

0!

6.19

20.

000

Non-

ope

rating

inco

me

(post

)!

2.86

10.

037

4.33

10.

059

2.61

10.

066

Prior

year

stoc

kre

turn

0.59

30.

000

0.32

40.

007

!0.

046

0.03

9C

urr

entye

arst

ock

retu

rn!

0.05

40.

470

!0.

588

0.00

00.

545

0.00

0

Firm

-yea

rsw

ith

repu

rchas

ean

noun

cem

ents

Inte

rcep

t!

3.32

90.

000

0.69

40.

000

!0.

976

0.00

0Firm

size

14.0

210.

010

!17

.150

0.00

3!

3.89

60.

481

Oper

atin

gin

com

e(p

rior)

2.97

10.

000

!4.

389

0.00

0!

3.86

40.

000

Non-

ope

rating

inco

me

(prior)

3.15

90.

013

0.50

20.

863

1.86

10.

317


Pay

out

ratio

(prior)

!0.

028

0.61

8!

0.15

40.

015

!0.

149

0.31

3Std

.de

viat

ion

ope

rating

inco

me

!2.

567

0.00

028

.008

0.00

025

.133

0.00

0O

per

atin

gin

com

e(p

ost

)2.

619

0.00

0!

4.43

10.

000

!3.

571

0.00

0N

on-

ope

rating

inco

me

(post

)2.

861

0.03

77.

192

0.00

55.

472

0.00

3Prior

year

stoc

kre

turn

!0.

593

0.00

0!

0.26

70.

064

!0.

638

0.00

0C

urre

ntye

arst

ock

retu

rn0.

054

0.47

0!

0.53

40.

000

!0.

491

0.00

0

Firm

-yea

rsw

ith

repu

rchas

ean

noun

cem

ents

and

div

iden

din

crea

ses

Inte

rcep

t!

4.02

40.

000

!0.

694

0.00

0!

1.67

00.

000

Firm

size

31.1

720.

000

17.1

500.

003

13.2

540.

001

Oper

atin

gin

com

e(p

rior)

7.35

90.

000

4.38

90.

000

0.52

30.

417

Non-

ope

rating

inco

me

(prior

)2.

657

0.32

4!

0.50

20.

863

1.35

90.

631

Pay

out

ratio

(prior)

0.12

60.

000

0.15

40.

015

0.04

30.

657

Std

.de

viat

ion

ope

rating

inco

me

!30

.574

0.00

0!

28.0

080.

000

!2.

874

0.11

0O

per

atin

gin

com

e(p

ost

)7.

052

0.00

04.

431

0.00

00.

860

0.20

3N

on-o

pera

ting

inco

me

(post

)!

4.33

10.

059

!7.

192

0.00

5!

1.72

10.

484

Prior

year

stoc

kre

turn

!0.

324

0.00

70.

267

0.06

4!

0.37

00.

002

Cur

rentye

arst

ock

retu

rn0.

588

0.00

00.

534

0.00

00.

042

0.66

7

Firm

-yea

rsw

ith

div

iden

din

crea

ses

Inte

rcep

t!

2.35

30.

000

0.97

60.

000

1.67

00.

000

Firm

size

17.9

180.

000

3.89

60.

481

!13

.254

0.00

1O

per

atin

gin

com

e(p

rior)

6.83

60.

000

3.86

40.

000

!0.

523

0.41

7N

on-

ope

rating

inco

me

(prior

)1.

297

0.37

6!

1.86

10.

317

!1.

359

0.63

1Pay

out

ratio

(prior)

0.12

10.

000

0.14

90.

013

!0.

043

0.65

7Std

.de

viat

ion

ope

rating

inco

me

!27

.699

0.00

0!

25.1

330.

000

2.87

40.

110

Oper

atin

gin

com

e(p

ost

)6.

192

0.00

03.

571

0.00

0!

0.86

00.

203

Non-o

pera

ting

inco

me

(post

)!

2.61

10.

066

!5.

472

0.00

31.

721

0.48

4Prior

year

stoc

kre

turn

0.04

60.

039

0.63

80.

000

0.37

00.

002

Cur

rentye

arst

ock

retu

rn0.

545

0.00

00.

491

0.00

0!

0.04

20.

667

Obse

rvat

ions

16,0

47Log-

likel

ihood

!9,

157


Tab

le5

Oper

atin

gan

dst

ock

perfor

man

cech

arac

terist

ics

a!ec

ting

a"rm's

choi

ceofpay

out

met

hod

This

tabl

epre

sent

sth

ere

sults

from

estim

atio

nofa

multin

omia

llo

git

mode

lofva

rious

oper

atin

gch

arac

terist

ics

and

stoc

kre

turn

sth

ough

tto

pred

ict

a"rm's

choic

eof

alte

rnat

ivem

etho

dsof

distr

ibut

ing

cash#ow

sin

agi

ven

year

.Eac

hco

lum

npr

ovid

esth

epar

amet

eres

tim

ates

obt

ained

from

thelo

g-od

dsra

tios

.The"rs

tco

lum

nco

mpar

es"rm

-yea

rsw

ith

no

incr

ease

inpa

yout

sto"rm

-yea

rsin

whi

chpay

outs

incr

ease

thro

ugh

the

useofr

epurc

has

es,i

ncre

ased

div

iden

dsorbo

th.T

he

seco

ndco

lum

nco

mpar

es"rm

-yea

rsw

ith

repurc

hase

prog

ram

initia

tionsorex

pan

sionsto

the

other

pos

sible

outc

om

es.T

he

third

colu

mn

com

par

es"rm

-yea

rsw

ith

both

stock

repur

chas

esan

ddiv

iden

din

crea

sesto

theoth

eral

tern

ativ

esan

dth

efo

urt

hco

lum

nco

mpa

res"rm

-yea

rsw

ith

div

iden

din

crea

ses

toth

eot

her

alte

rnat

ives

.All

indep

enden

tva

riab

les

are

the

sam

eas

de"ned

inT

able

3. Firm

-yea

rsw

ith

Firm

-yea

rsw

ith

repur

chas

eF

irm

-yea

rsw

ith

no

repur

chas

ean

nounce

men

tsan

dF

irm

-yea

rsw

ith

incr

ease

inpay

out

san

nou

nce

men

tsdiv

iden

din

crea

ses

div

iden

din

crea

ses

Coe$

cien

tp-

valu

eC

oe$

cien

tp-

valu

eC

oe$

cien

tp-

valu

eC

oe$

cien

tp-

valu

e

Firm

-yea

rsw

ith

no

incr

ease

inpa

youts

Inte

rcep

t3.

642

0.00

05.

326

0.00

03.

389

0.00

0Firm

size

16.6

890.

311

!16

.708

0.01

5!

15.7

680.

000

Oper

atin

gin

com

e(p

rior)

!2.

011

0.00

0!

6.91

50.

000

!6.

599

0.00

0N

on-

ope

rating

inco

me

(prior)

!2.

842

0.23

8!

1.39

10.

817

6.20

30.

059

Pay

out

ratio

(prior)

0.17

90.

108

!0.

052

0.90

9!

0.00

80.

719

Std.d

evia

tion

ope

rating

inco

me

1.10

90.

065

27.4

570.

000

25.4

230.

000

Oper

atin

gin

com

e(p

ost

)!

2.49

50.

000

!5.

288

0.00

0!

4.90

60.

000

Non-

ope

rating

inco

me

(post

)!

5.59

40.

007

!2.

243

0.75

6!

5.31

90.

121

Inst

itutiona

low

ner

ship

!0.

021

0.00

0!

0.03

60.

000

!0.

026

0.00

0Prior

year

stoc

kre

turn

0.69

70.

000

0.58

10.

011

0.02

40.

706

Curr

entye

arst

ock

retu

rn!

0.18

50.

104

!0.

889

0.00

0!

0.62

70.

000

Firm

-yea

rsw

ith

repu

rchas

ean

noun

cem

ents

Inte

rcep

t!

3.64

20.

000

1.68

40.

000

!0.

253

0.21

7Firm

size

!16

.689

0.31

1!

33.3

970.

056

!32

.457

0.05

2O

per

atin

gin

com

e(p

rior)

2.01

10.

000

!4.

904

0.00

0!

4.58

80.

000

Non-

ope

rating

inco

me

(prior)

2.84

20.

238

1.45

10.

819

9.04

50.

021

Pay

outra

tio

(prior)

!0.

179

0.10

8!

0.18

50.

123

!0.

189

0.09

7


Std

.dev

iation

ope

rating

inco

me

!1.

109

0.06

526

.347

0.00

024

.314

0.00

0O

per

atin

gin

com

e(p

ost

)2.

495

0.00

0!

2.79

30.

063

!2.

411

0.01

0N

on-

ope

rating

inco

me

(post

)5.

594

0.00

73.

351

0.65

20.

274

0.94

3In

stitutiona

low

ner

ship

0.02

10.

000

!0.

016

0.00

2!

0.00

50.

121

Prior

year

stoc

kre

turn

!0.

697

0.00

0!

0.17

90.

441

!0.

722

0.00

0C

urr

entye

arst

ock

retu

rn0.

185

0.10

4!

0.70

40.

001

!0.

441

0.00

4

Firm

-yea

rsw

ith

repu

rchas

ean

noun

cem

ents

and

divi

dend

incr

ease

s

Inte

rcep

t!

5.32

60.

000

!1.

684

0.00

0!

1.93

70.

000

Firm

size

16.7

080.

015

33.3

970.

056

0.94

00.

891

Oper

atin

gin

com

e(p

rior)

6.91

50.

000

4.90

40.

000

0.31

60.

816

Non-

ope

rating

inco

me

(prior)

1.39

10.

817

!1.

451

0.81

97.

594

0.23

8Pay

out

ratio

(prior)

0.05

20.

909

0.18

50.

123

!0.

004

0.94

0Std

.dev

iation

ope

rating

inco

me

!27

.457

0.00

0!

26.3

470.

000

!2.

033

0.53

9O

per

atin

gin

com

e(p

ost

)5.

288

0.00

02.

793

0.06

30.

381

0.78

9N

on-

ope

rating

inco

me

(post

)2.

243

0.75

6!

3.35

10.

652

!3.

076

0.68

4In

stitutiona

low

ners

hip

0.03

60.

000

0.01

60.

002

0.01

10.

022

Prior

year

stoc

kre

turn

!0.

518

0.01

10.

179

0.44

1!

0.54

20.

009

Curr

entye

arst

ock

retu

rn0.

889

0.00

00.

704

0.00

10.

262

0.20

8

Firm

-yea

rsw

ith

div

iden

din

crea

ses

Inte

rcep

t!

3.38

90.

000

0.25

30.

217

1.93

70.

000

Firm

size

15.7

680.

000

32.4

570.

052

!0.

940

0.89

1O

per

atin

gin

com

e(p

rior)

6.59

90.

000

4.58

80.

000

!0.

316

0.81

6N

on-

ope

rating

inco

me

(prior)

!6.

203

0.05

9!

9.04

50.

021

!7.

594

0.23

8Pay

out

ratio

(prior)

0.00

80.

719

0.18

90.

097

0.00

40.

940

Std

.dev

iation

ope

rating

inco

me

!25

.423

0.00

0!

24.3

140.

000

2.03

30.

539

Oper

atin

gin

com

e(p

ost

)4.

906

0.00

02.

411

0.01

0!

0.38

10.

789

Non-

ope

rating

inco

me

(post

)5.

319

0.12

1!

0.27

40.

943

3.07

60.

684

Inst

itutiona

low

ners

hip

0.02

60.

000

0.00

50.

121

!0.

011

0.02

2Prior

year

stoc

kre

turn

0.02

40.

706

0.72

20.

000

0.54

20.

009

Curr

entye

arst

ock

retu

rn0.

627

0.00

00.

441

0.00

4!

0.26

20.

208

Obse

rvat

ions

5,12

4Log-

like

lihoo

d!

2,85

1


between 1985 and 1996. The data indicate that repurchases have grown over thisperiod. In 1996, actual repurchases amounted to between $44.3 and $63.3billion, suggesting that they are an economically important source of payouts.However, repurchases are still considerably smaller than the $141.7 billion individends paid that year.

Repurchases are noticeably more volatile than dividends. They appear to varyprocyclically, they were high during the rising markets of the late 1980s, droppedin the recession of the early 1990s and increased during the boom of themid-1990s. Repurchases are responsible for a disproportionately large fractionof the variation in total payouts. The smoothness of the dividend series com-bined with the volatility and procyclicality of the repurchase series are consis-tent with the view that dividends are paid out of sustainable cash #ows whilerepurchases are paid out of temporary cash #ows. Repurchases do not appear tobe replacing dividends; rather they seem to serve the complementary role ofpaying out short-term cash #ows.

A cross-sectional analysis of "rms' decisions to increase dividends or repur-chases is consistent with this view. Firms with higher operating cash #ows aremore likely to increase dividends, while "rms with higher non-operating cash#ows are more likely to increase repurchases. Firms with a higher standarddeviation of cash #ows are more likely to use repurchases. Subsequent to thepayout increase, cash #ows of repurchasing "rms continue to be lower thanthose of dividend-increasing "rms.

Textbook discussion of payout policy generally suggests that dividends andstock repurchases are more or less equivalent ways of paying out cash #ows (seeBrealey and Myers, 1996 or Grinblatt and Titman, 1988). Generally, the dis-cussion of the choice between the two revolves around Black's (1976) focuson repurchases' tax advantage relative to dividends. Our empirical work sug-gests that much more than taxes are necessary to explain di!erences in howdividends and repurchases are used in practice. Dividends and repurchases areused at di!erent places in the business cycle by di!erent types of "rms. Ourinterpretation of these results emphasizes the "nancial #exibility inherent inrepurchases. While this is likely to be part of the explanation, the completeanswer is undoubtedly more involved. For example, the striking di!erence instock market performance prior to dividend increases and repurchase programssuggests that valuation is also an important factor in determining payoutmethods.

Even though repurchases have not replaced dividends, they have become animportant source of payouts. Many branches of "nance research have focusedon dividends, and their yields in many di!erent contexts. For example, muchresearch has examined the extent to which expected returns can be predictedusing dividend yields (see Campbell and Shiller, 1988; Fama and French, 1988;Kandel and Stambaugh, 1996; Kothari and Shanken, 1997; or Ponti! andSchall, 1998). Another related literature has focused on excess volatility and the


issue of whether the variation in stock prices can be explained by movements individends (see Shiller, 1981; Leroy and Porter, 1981; Marsh and Merton, 1986;Kothari and Shanken, 1992). These two strands of the literature could beextended productively by considering total payouts rather than only dividendsin their econometric work.

References

Allen, F., Bernardo, A., Welch, I., 1998. A theory of dividends based on tax clienteles. Unpublishedworking paper, UCLA.

Bernartzi, S., Michaely, R., Thaler, R., 1997. Do changes in dividends signal the future or the past?The Journal of Finance 52, 1007}1034.

Black, F., 1976. The dividend puzzle. Journal of Portfolio Management 2, 5}8.Brealey, R., Myers, S., 1996. Principles of Corporate Finance., 5th Edition. McGraw-Hill,

New York.Campbell, J., Shiller, R., 1988. The dividend-price ratio and expectations of future dividends and

discount rates. Review of Financial Studies 1, 195}228.Comment, R., Jarrell, G., 1991. The relative signaling power of Dutch-auction and "xed-price

self-tender o!ers and open market share repurchases. Journal of Finance 46, 1243}1271.Cook, D., Krigman, L., Leach, J., 1997a. Safe harbor or smoke screen? compliance and disclosure

under SEC rule 10b-18. Unpublished working paper, University of Colorado.Cook, D., Krigman, L., Leach, J., 1997b. Corporate repurchase programs: execution strategy and the

competing market-maker e!ect. Unpublished working paper, University of Colorado.Dann, L., 1981. The e!ects of common stock repurchase on security holder's returns. Journal of

Financial Economics 9, 101}138.DeAngelo, H., DeAngelo, L., Skinner, D., 1996. Reversal of fortune: dividend signaling and the

disappearance of sustained earnings growth. Journal of Financial Economics 40, 341}372.DeAngelo, H., DeAngelo, L., Skinner D., 2000. Special dividends and the evolution of dividend

signaling. Journal of Financial Economics 57, 309}354.Denis, D., Denis, D., Sarin, A., 1994. The information content of dividend changes: cash #ow

signaling, overinvestment, and dividend clienteles. Journal of Financial and Quantitative Analy-sis 29, 567}587.

Esty, B., Schreiber, C., 1995. Dividend policy at FPL Group (B), Harvard Business School Case9-295-106, Harvard Business School Press, Boston, MA.

Fama, E., Babiak, H., 1968. Dividend policy: an empirical analysis. Journal of the AmericanStatistical Association 63, 123}1161.

Fama, E., French, K., 1988. Dividend yields and expected stock returns. Journal of FinancialEconomics 22, 3}26.

Fenn, G., Liang, N., 1997. Good news and bad news about share repurchases. Unpublished workingpaper, Milken Institute and Board of Governors of the Federal Reserve System.

Ghosh, C., Woolridge, J., 1988. An analysis of shareholder reaction to dividend cuts and omissions.The Journal of Financial Research 11, 281}294.

Grinblatt, M., Titman, S., 1998. Financial Markets and Corporate Strategy. Irwin/McGraw Hill,Boston, MA.

Guay, W., Harford, J., 2000. The cash #ow permanence and information content of dividendincreases vs. repurchases. Journal of Financial Economics 57, 385}415.

Hulbert, M., 1997. Putting money where their mouths are. Forbes April 21, 1997.Hylton, R., 1995. Stock buybacks are hot } here's how you can cash in. Fortune 132, 133}134.


Ikenberry, D., Lakonishok, J., Vermaelen, T., 1995. Market underreaction to open market sharerepurchases. Journal of Financial Economics 39, 181}208.

Jolls, C., 1998. Stock repurchases and executive compensation. Unpublished working paper, Har-vard Law School.

Kandel, S., Stambaugh, R., 1996. On the predictability of stock returns: an asset allocationperspective. Journal of Finance 51, 385}424.

Kothari, S., Shanken, J., 1992. Stock return variation and expected dividends: a time series andcross-sectional analysis. Journal of Financial Economics 31, 177}210.

Kothari, S., Shanken, J., 1997. Book-to-market, dividend yields and expected market returns: a timeseries analysis. Journal of Financial Economics 44, 169}203.

Leroy, S., Porter, R., 1981. The present value relation: tests based on implied variance bounds.Econometrica 49, 555}574.

Lintner, J., 1956. Distribution of incomes of corporations among dividends, retained earnings, andtaxes. American Economic Review 46, 97}113.

Marsh, T., Merton, R., 1986. Dividend variability and variance bound tests for the rationality ofstock market prices. American Economic Review 76, 483}498.

Netter, J., Mitchell, M., 1989. Stock repurchase announcements and insider transactions after theOctober 1987 stock market crash. Financial Management 18, 84}96.

Ponti!, J., Schall, L., 1998. Book-to-market ratios as predictors of market returns. Journal ofFinancial Economics 49, 141}160.

Power, W., 1995. Heard on the street: most buybacks are stated not completed. The Wall StreetJournal, March 7, 1995, c1}c2.

Shiller, R., 1981. Do stock prices move too much to be justi"ed by subsequent changes in dividends?American Economic Review 71, 421}436.

Stephens, C., Weisbach, M., 1998. Actual share reacquisitions in open market repurchase programs.Journal of Finance 53, 313}334.

Vermaelen, T., 1981. Common stock repurchases and market signaling. Journal of FinancialEconomics 9, 139}183.

Weisbenner, S., 1998. Corporate share repurchases in the mid 1990s: what role do stock optionsplay? Unpublished working paper, MIT.


Documents

Financial exibility and the choice between dividends and ... · total payouts and the division of these payouts between dividends and stock repurchases. In Section 3, we provide estimates