the e!ect of international institutional factors on properties of accounting

qWe gratefully acknowledge the helpful comments of David Alexander, Eli Bartov, Sudipta Basu,Gary Biddle, Sir Bryan Carsberg, Dan Collins, Peter Easton, Bob Holthausen (the editor), ScottKeating, Christian Leuz, Gerhard Mueller, Christopher Nobes, the late Dieter Ordelheide, PeterPope, Abbie Smith, Peter Taylor, Ross Watts, Greg Waymire, Steve Ze!, Jerry Zimmerman, thereferee, and seminar participants at: Carnegie Mellon University, University of California atBerkeley, University of California at Los Angeles, CUNY Baruch College, EIASM Workshop inEuropean Accounting in Krakow, 1999 Financial Accounting and Auditing Conference of theInstitute of Chartered Accountants in England & Wales, UniversitaK t Frankfurt am Main, HongKong University of Science and Technology 1998 Summer Symposium, IAAER/CIERA 1998Conference, Inquire Europe Autumn 1998 Seminar, University of Iowa, Harvard Business School,KPMG-AAA International 1999 Accounting Conference, London Business School, London Schoolof Economics, Massachusetts Institute of Technology, Melbourne Business School, New YorkUniversity, Ohio State University, University of New South Wales, University of Technology inSydney, and Washington University in St Louis. The paper has received a Vernon K. ZimmermanAward and an Inquire Europe Prize. Ball and Kothari received "nancial assistance from the JohnM. Olin Foundation and the Bradley Policy Research Center, and Kothari acknowledges "nancialassistance from the New Economy Value Research Lab at the MIT Sloan School of Management.

*Corresponding author. Tel.: #(617) 253-0994; fax: #(617) 253-0603.E-mail address: [email protected] (S.P. Kothari).

Journal of Accounting and Economics 29 (2000) 1}51

The e!ect of international institutional factorson properties of accounting earningsq

Ray Ball!, S.P. Kothari",*, Ashok Robin#

!Graduate School of Business, University of Chicago, Chicago, IL 60637, USA"Sloan School of Management, Massachusetts Institute of Technology, Cambridge, MA 02142, USA

#College of Business, Rochester Institute of Technology, Rochester, NY 14623, USA

Received 3 August 1998; received in revised form 9 June 2000

Abstract

International di!erences in the demand for accounting income predictably a!ect theway it incorporates economic income (change in market value) over time. We character-ize the &shareholder' and &stakeholder' corporate governance models of common andcode law countries respectively as resolving information asymmetry by public disclosureand private communication. Also, code law directly links accounting income to current

0165-4101/00/$ - see front matter ( 2000 Elsevier Science B.V. All rights reserved.PII: S 0 1 6 5 - 4 1 0 1 ( 0 0 ) 0 0 0 1 2 - 4

payouts (to employees, managers, shareholders and governments). Consequently, codelaw accounting income is less timely, particularly in incorporating economic losses.Regulation, taxation and litigation cause variation among common law countries. Theresults have implications for security analysts, standard-setters, regulators, and corporategovernance. ( 2000 Elsevier Science B.V. All rights reserved.

JEL classixcation: F00; F30; G15; M41

Keywords: International accounting; Standard setting; Regulation; Conservatism

1. Introduction

We show that di!erences in the demand for accounting income in di!erentinstitutional contexts cause its properties to vary internationally. The propertiesof accounting income we study are timeliness and conservatism. Timeliness isde"ned as the extent to which current-period accounting income incorporatescurrent-period economic income, our proxy for which is change in market valueof stockholders' equity. Conservatism is de"ned in the Basu (1997) sense as theextent to which current-period accounting income asymmetrically incorporateseconomic losses, relative to economic gains.

A central result is that accounting income in common-law countries issigni"cantly more timely than in code-law countries, due entirely to quickerincorporation of economic losses (income conservatism). Conversely, informa-tion asymmetry more likely is resolved in code-law countries by institutionalfeatures other than timely and conservative public "nancial statements, notablyby closer relations with major stakeholders. In contrast with Roe (1994), weconclude that enhanced common-law disclosure standards reduce the agencycosts of monitoring managers, thus countering the advantage of closer share-holder}manager contact in code-law countries.

We believe that timeliness and conservatism together capture much of thecommonly used concept of "nancial statement &transparency.' In comparisonwith a system that allows economic losses to be re#ected in accounting incomegradually over time, timely incorporation of economic losses in accountingincome incents managers to stem the losses more quickly. Because accountingincome #ows into balance sheet accounts, conservatism as we de"ne it alsomakes leverage and dividend restrictions binding more quickly. It makesoptimistic non-accounting information released by managers less credible touninformed users. Conservative accounting thus facilitates monitoring ofmanagers and of debt and other contracts, and is an important featureof corporate governance.

The principal institutional variable we study is the extent of political in#uenceon accounting. Our simplest proxy for political in#uence is a dichotomous

2 R. Ball et al. / Journal of Accounting and Economics 29 (2000) 1}51

classi"cation of countries into code law systems with high political in#uenceversus common law systems in which accounting practices are determinedprimarily in the private sector. We hypothesize that politicization of accountingstandard setting and enforcement weakens the demand for timely and conserva-tive accounting income, and conversely increases the demand for an incomevariable with low volatility. In our sample, Australia, Canada, UK and USA areclassi"ed as common-law countries (they comprise a group known as G4#1,exclusive of New Zealand) and France, Germany and Japan are classi"ed ascode-law.

In code-law countries, the comparatively strong political in#uence on ac-counting occurs at national and "rm levels. Governments establish and enforcenational accounting standards, typically with representation from major politi-cal groups such as labor unions, banks and business associations. At the "rmlevel, politicization typically leads to a &stakeholder' governance model, involv-ing agents for major groups contracting with the "rm. Current-period ac-counting income then tends to be viewed as the pie to be divided among groups,as dividends to shareholders, taxes to governments, and bonuses to managersand perhaps also employees. Compared to common-law countries, the demandfor accounting income under code law is in#uenced more by the payoutpreferences of agents for labor, capital and government, and less by the demandfor public disclosure. Conversely, because these groups' agents are representedin corporate governance, insider communication solves the information asym-metry between managers and stakeholders. We hypothesize that their prefer-ences penalize volatility in payouts and thus in income. Thus, code-lawaccounting standards give greater discretion to managers in deciding wheneconomic gains and losses are incorporated in accounting income. Managersreduce income volatility by varying the application of accounting standards orby in#uencing operating, "nancing and investment decisions (for example, bydeferring discretionary expenditures such as R&D in bad earnings years).

Under the &shareholder' governance model that is typical of common-lawcountries, shareholders alone elect members of the governing board, payouts areless closely linked to current-period accounting income, and public disclosure isa more likely solution for the information asymmetry problem. In comparisonwith the more political process in code-law countries, the desirable properties ofaccounting income in common law countries are determined primarily in thedisclosure market. We hypothesize those properties include timeliness in incor-porating negative economic income (i.e., asymmetric conservatism).

We caution that the code/common classes are by no means homogeneous,with "nancial reporting in no country being determined in a purely market orplanning system. Notable historical examples of overlapping include the codi"-cation imposed on a predominantly common-law reporting system by theCompanies Acts in the UK and by the Securities and Exchange Acts in the US,and the enactment of French and German legislation to permit consolidated

R. Ball et al. / Journal of Accounting and Economics 29 (2000) 1}51 3

1Asset impairment standards such as SFAS 121 in the US are an important example ofaccounting standards linking accounting and economic incomes. Information asymmetry impliesthat implementation of impairment standards depends on the incentive of "nancial statementpreparers (managers and auditors) to disclose information about economic losses, which variesinternationally (compare US and Japan).

"nancial statements prepared under common-law accounting standards. Des-pite these limitations, our results indicate the code/common classi"cation isa valid proxy for the extent of political relative to market determination of"nancial reporting. Nevertheless, we develop "ner hypotheses based on tax andregulatory di!erences across individual countries.

We also caution that institutional determinants of "nancial reporting varyover time. As a coarse test, we divide the sample into two sub-periods, andobserve an increase in asymmetric conservatism of accounting income in mostcountries. One interpretation is that timely incorporation of economic losses inaccounting income is an e$cient corporate governance mechanism, providingbetter incentives to attend to losses and hence maximize value, which increasedinternational product market competition has created incentives for even code-law corporations to adopt.

The sample studied is more than 40,000 "rm-year accounting incomes re-ported during 1985}95, under the accounting rules of seven countries. Code-lawincome in this sample is substantially less timely and less conservative onaverage than common-law income. It does not even exhibit more timeliness thandividends. Within the common-law group, there is less asymmetric conservatismin accounting income in the United Kingdom, a country we characterize interms of lower political involvement in accounting, lower litigation costs andless issuance of public debt. In addition to a detailed analysis of seven countries,we also study properties of accounting income in a sample of eighteen othercountries. The results are consistent with our general thesis, that importantproperties of accounting income (conservatism in particular) around the worldare a function of the varying demands that accounting income satis"es underdi!erent institutional arrangements.

Our research design addresses the incorporation of economic income inaccounting income over time, under di!erent international institutions. This hasseveral advantages over simply studying international variation in accountingstandards. First, much accounting practice is not determined by accountingstandards, for reasons that include: practice is more detailed than standards;standards lag innovations in practice; and companies do not invariably imple-ment standards.1 Second, the extent to which accounting practice is determinedby formal standards varies internationally, and the incentive to follow ac-counting standards depends on penalties under di!erent enforcement institu-tions, so studying accounting standards per se is incomplete and potentially


misleading in an international context. Third, reported income is in#uenced bymanagers' operating, "nancing and investment decisions, as well as by ac-counting standards. For example, managers can reduce volatility in accountingincome by deferring discretionary expenditures (such as R&D) in bad years.Because the use of accounting income in corporate governance varies interna-tionally, we expect managers' operating, "nancing and investment decisions toa!ect accounting income di!erentially across countries, and report evidenceconsistent with that expectation. For both these reasons, we study internationalvariation in properties of the actually reported income numbers, inferred from theway they incorporate economic income over time.

Our research design's validity depends on two measures. First, we study the#ow of market-valued economic income into book-valued accounting income,using the "scal-year change in market value of equity (adjusted for dividendsand capital transactions) as a proxy for economic income. A major concern isthat the accuracy of this proxy is correlated with the institutional independentvariables in the study, and in particular that code-law countries have endogen-ously lower market liquidity and public disclosure standards. Second, theresearch design requires us to infer independent variables, such as the degree ofpolitical versus market determination of reported income, from our character-ization of salient institutional facts. While our characterization is based onsurveying a wide range of sources, it undoubtedly is subject to error. In theconcluding section, we argue that both types of measurement error create a biasagainst our hypotheses.

We contribute to a growing literature on the e!ects of international ac-counting di!erences, including Jacobsen and Aaker (1993), Alford et al. (1993),Amir et al. (1993), Bandyopadhyay et al. (1994), Harris et al. (1994), Joos andLang (1994), Barth and Clinch (1996), and Pope and Walker (1999). We alsocontribute to the literature on international corporate governance, includingBaums et al. (1994) and La Porta et al. (1997).

The following section outlines the model used to test the timeliness andconservatism of accounting income. The third section describes the data. Sectionfour surveys the salient institutional facts used to develop and then test hypo-theses on properties of income internationally. Section "ve extends these tests toa comparison of income with dividends and cash #ows. The sixth section reportsspeci"cation tests and the concluding section discusses the research design andthe implications of the results.

2. A model of incorporation of economic income in accounting income

The research design infers timeliness and conservatism from the way "rms'accounting incomes incorporate their economic incomes over time. We there-fore specify accounting income as the dependent variable. We measure "rms'


2We use clean surplus accounting as a concept to motivate the research design. We assume thedegree of violation of clean surplus accounting is not systematically related to the internationalinstitutional factors we investigate. Research on the return-earnings relation typically exacerbatesclean-surplus violation by excluding extraordinary items.

economic incomes as "scal-year changes in market values of equity, adjusted fordividends and capital contributions (Hicks, 1946).&Clean surplus' accounting (Ohlson, 1988) implies two relevant identities for

all "rms. First, accounting income equals "scal-year change in book value ofequity, adjusted for dividends and capital contributions. Second, a "rm's ac-counting and economic incomes summed over its lifetime are identical.2 Weinvestigate the temporal process of the incorporation of economic income inaccounting income, i.e., the accounting model of income determination, andhow it is a!ected by international institutional factors. Our research designallows for three fundamental features of the accounting model of incomedetermination: accounting &recognition' principles that generally reduce thetimeliness of accounting income by smoothing its incorporation of economicincome over time; the e!ectiveness of accounting accruals in ameliorating serialcorrelation in operating cash #ows; and accounting income-statement conserva-tism.

The most fundamental feature of accounting determining the incorporation ofeconomic income in accounting income over time is the accounting &recognition'principles (FASB, 1985, Paras 78}89), including the Revenue Realization andExpense Matching principles. Whereas economic income immediately incorpor-ates changes in expectations of the present values of future cash #ows, therecognition principles incorporate such changes in accounting income graduallyover time, generally at points close to when the actual cash #ow realizationsoccur. Hence, accounting income systematically lags economic income (Ball andBrown, 1968) and the lag extends over multiple periods (Beaver et al., 1980;Easton et al., 1992; Kothari and Sloan, 1992). The recognition principles there-fore cause economic income to be incorporated in accounting income ina lagged and &smoothed' fashion over time.

This feature of accounting income arises because there is demand for anincome variable with properties additional to timeliness. While timeliness per seis desirable, information asymmetry between managers and users creates a de-mand for an income variable that is observable independently of managers.Accounting income thus incorporates only the subset of available value-relevantinformation that is independently observable, whereas economic income incor-porates information that is not independent of managers, such as plans andforecasts (our proxy for economic income incorporates the sharemarket reactionto managers' forward-looking statements). In other words, accounting incomedoes not attempt to anticipate future cash #ows to the same extent as economic


income. The "rst-order e!ect of the recognition principles thus is to makeaccounting income a complex moving average of past economic incomes.

The second fundamental feature of the accounting income-determinationmodel is that accounting accruals imperfectly ameliorate serial correlation inoperating cash #ows. The accounting model provides for some anticipation offuture cash #ows through accrual accounting. For example, if managers pay anaccount for inventory early, then there is a decrease in current-period operatingcash #ow and, ceteris paribus, an o!setting increase in subsequent periods.Accrual accounting rules attempt to insulate income from the e!ect of the earlypayment, by expensing an amount in both periods that is based on inventoryusage, not payments. In general, short term variation in "rms' operational"nancing and investment decisions (such as changes in inventories, accountspayable and accounts receivable) causes negative serial correlation in operatingcash #ows, which accrual accounting attempts to remove from accountingincome (Dechow et al., 1998). Hence, operating cash #ow can be viewed asa noisier and less timely version of accounting income. However, accrualaccounting is imperfect, because it is costly, so anticipation of future cash #owsvia accruals does not completely remove the noise in the cash #ow time series.

The "rst two features of the accounting model of income determinationtogether imply:

>it"f

j(*<

it, *<

it~1, *<

it~2, *<

it~3,2, <

it) (1)

where > and *<, respectively, denote accounting and economic income, and< denotes noise due to imperfect accounting accruals. Economic income, *<, is"scal-year change in the market capitalization of equity plus dividends andminus capital contributions during the year (Hicks, 1946). We hypothesize thatthe accounting model is applied di!erently across countries, and assume themodel's parameters hold for all "rms i that report under the accounting systemsof country j. Assuming that *< is independent over time, this simpli"es to

>it"g

j(*<

it, g

it) (2)

The disturbance git

incorporates lagged changes in market values (*<it~1

,*<

it~2, *<

it~3, 2) as well as noise due to the residual serial correlation in cash

#ows not removed by accounting accruals (<it). Both components of the

disturbance term a!ect the R2 of regression (2), which is used as a proxy for thetimeliness property of accounting income. After scaling by opening marketvalue, <

it~1, the dependent and independent variables are annual rate of return

(Rit,*<

it/<

it~1) and earnings yield (NI

it,>

it/<

it~1), and a linear speci"ca-

tion gives

NIit"a

0j#a

1jR

it#m

it(3)

The third fundamental feature of the accounting income model we study isconservatism. A longstanding example of income conservatism is the &lower of


3This 1995 accounting standard formalized what already had become common practice. Elliottand Hanna (1996) report an increase in negative &one time' charges against income for US "rmsaround 1970, rising to 20% of "rms annually by the early 1990s. Collins et al. (1997) report a similarincrease, and that by the early 1980s 25}30% of US "rms reported negative incomes before one-timecharges. These data are consistent with our view that formalized common-law accounting standardsprimarily arise endogenously from common practice in the market for accounting.

4Asset revaluations occur for acquisitions accounted under the purchase method, but do not #owthrough income.

cost or market' inventory rule, which incorporates inventory losses more quick-ly in income than gains. A topical example is new information about future cash#ows from long-term assets. The recognition principles normally incorporatethis information in accounting income at or near the point when the actual cash#ow realizations occur. However, a variety of accounting rules and practicescause immediate write-o!s against income when expected future cash #owsdecrease, without waiting for the cash #ow decreases to be realized. In the US,SFAS 121 recently formalized longstanding write-o! practices for long-livedassets in the form of asset impairment rules.3 Upward revaluation is compara-tively rare in the US: it has not been practiced since the Securities and ExchangeCommission (SEC) was established in 1934, though it is practiced in somecountries. Consequently, unrealized increases in asset values generally do not#ow into income until approximately when the underlying cash #ow increasesoccur, but unrealized decreases are more likely to be incorporated quickly.4

Following Basu (1997), we incorporate conservative asymmetry in accountingincome timeliness by modifying (3) for asymmetric incorporation of negativeeconomic income:

NIit"b

0j#b

1jRD

it#b

2jR

it#b

3jR

itRD

it#e

it(4)

The dummy variable RDtassumes its value based on the sign of stock return,

not earnings: one if return Rtis negative, and zero otherwise. b

2jand (b

2j#b

3j)

capture the incorporation in current-year accounting income of positive andnegative economic income respectively, in country j.

This speci"cation has several attractive features. One advantage of specifyingaccounting income as the dependent variable is avoiding the need for a noisyearnings expectations model. Here, the independent variable (annual stockreturn) is relatively free of short-term microstructure, liquidity or mispricinge!ects. An additional advantage of the speci"cation is that it incorporates thefundamental tenets of accounting income recognition. In particular, it incorpor-ates lags that arise from the demand for an independent income measure, andpiecewise linearity allows us to study international di!erences in asymmetrictimeliness, or conservatism.

Initially, we estimate separate individual-country relations for each country j,pooling all "rms i reporting under the country's accounting standards and allyears t. International di!erences in income timeliness, for positive and negative


5German companies do not deduct minority interest from consolidated net income, but GlobalVantage alters their numbers to comply with US practice. We therefore de"ne German Y as IC data32# data 27, which reconciles to the numbers actually reported. The adjustment has only a triviale!ect on the results.

6 In correspondence, Christian Leuz and the late Dieter Ordelheide note that some French andGerman "rms recently have issued consolidated "nancial statements prepared under InternationalAccounting Standards. This practice largely post-dates our sample. If Global Vantage classi"esthese observations as &German,' the errors create a potential bias against our hypotheses.

7The rationale is to eliminate observations potentially with errors or with extreme values due toscaling. The disadvantage is that potentially informative observations are deleted and there is thedanger of an incorrect inference.

economic income combined, are re#ected in the R2's of individual-countryregressions (4).

3. Data

Accounting income, cash #ow, and dividends over 1985}95 are from theGlobal Vantage Industrial/Commercial (IC) "le. Accounting income NI

tis net

income before extraordinary items (IC data 32).5 Dividends (DIV) is dividendspaid (IC data 36). Operating cash#ow (OCF) is net income before extraordinaryitems (IC data 32) plus depreciation (IC data 11), minus the change in non-cashcurrent assets (IC data 75 minus data 60), plus the change in current liabilitiesother than the current portion of long-term debt (IC data 104 minus data 94). Allvariables are scaled by market value of equity, calculated from the GlobalVantage Issue "le as price times number of outstanding shares, adjusted forstock splits and dividends using the Global Vantage adjustment factor. Changein accounting income *NI

tis NI

t!NI

t~1. Stock return R is the holding-

period return, including dividends, over the "rm's "scal accounting year. Each"rm/year observation is assigned to a country based on Periodic DescriptorArray 13 on the IC "le, indicating the accounting standards used in preparing its"nancial statements that year (normally the country of the "rm's home ex-change).6

We exclude the two extreme percentiles of each variable (NI, *NI, DIV, OCFand R).7 Next, we exclude each "rm/year with a missing value for any variable,giving the same observation set for the various variables and models estimated.Finally, we exclude countries with less than 1000 "rm/year observations overthe thirteen years. This leaves us with a "nal sample of 40,359 "rm/yearobservations in eleven years from seven countries: Australia, Canada, UK andUSA (common law countries) and France, Germany and Japan (code lawcountries). We also summarize results for a secondary sample of 18 countrieswith at least 100 "rm-year observations.

Table 1 contains sample descriptive statistics. Individual-country samples arepooled "rm-years, ranging in size from 1054 (France) to 21,225 (US). In Panel A,


Tab

le1

Sam

ple

char

acte

rist

ics!

Pan

elA

:de

scri

ptiv

est

atis

tics

NR

NI

DIV

OC

F

kM

edp

kM

edp

kM

edp

kM

edp

Aus

tral

ia1,

321

17.3

9.9

52.6

2.6

6.7

19.0

3.5

3.6

3.3

11.6

10.9

33.3

Can

ada

2,90

112

.16.

046

.83.

25.

315

.42.

01.

42.

415

.112

.228

.9U

S21

,225

12.7

8.1

42.8

3.1

6.2

14.4

1.9

1.0

2.4

11.7

10.3

21.4

UK

5,75

813

.59.

838

.06.

67.

28.

83.

43.

32.

012

.710

.418

.8Fra

nce

1,05

414

.96.

643

.06.

16.

610

.82.

32.

11.

922

.115

.736

.2G

erm

any

1,24

58.

94.

231

.53.

74.

29.

02.

02.

01.

718

.414

.326

.8Ja

pan

6,85

53.

7!

2.8

33.4

1.7

1.8

2.1

0.7

0.7

0.4

4.7

4.2

7.8

Pan

elB

:O

bser

vati

ons

byx

scal

year

end

(mon

th)

12

34

56

78

910

1112

Tota

l

Aus

tral

ia5

032

430

964

322

450

620

11,

321

Can

ada

113

2613

151

2110

547

144

155

7231

2,00

52,

901

US

964

405

872

401

446

2,09

456

250

61,

666

700

425

12,1

8421

,225

UK

297

119

1,22

233

311

934

412

114

359

016

466

2,24

05,

758

Fra

nce

02

230

03

012

245

098

51,

054

Ger

man

y0

194

00

759

018

37

394

51,

245

Japa

n94

363

5,34

139

6829

914

150

6714

453

76,

855

Tota

l1,

473

934

7,62

582

868

43,

614

780

821

2,81

31,

015

675

19,0

9740

,359


Pan

elC

:O

bser

vati

ons

byye

ar

8586

8788

8990

9192

9394

95Tota

l

Aus

tral

ia11

2988

9111

616

516

016

418

117

813

81,

321

Can

ada

189

221

277

276

306

325

306

311

317

325

482,

901

US

1,58

61,

661

2,23

42,

240

2,15

92,

098

2,00

62,

111

2,15

32,

189

788

21,2

25U

K98

314

438

508

621

703

722

751

756

733

114

5,75

8Fra

nce

966

7472

9814

715

515

113

114

65

1,05

4G

erm

any

2074

8376

9014

518

018

917

019

226

1,24

5Ja

pan

332

866

873

984

188

185

085

787

481

40

6,85

5Tota

l1,

916

2,69

33,

862

4,00

24,

231

4,46

44,

379

4,53

44,

582

4,57

71,

119

40,3

59

!Sam

ple

consist

sof40

,359"rm

-yea

robse

rvat

ions

sele

cted

from

the

Glo

bal

Van

tage

Indust

rial

/Com

mer

cial

and

Issu

e"le

sove

r19

85}95

,using

the

follo

win

gpro

cedu

re.F

irst

,fo

rea

chva

riab

le(see

bel

ow)w

eel

imin

ate

the

two

extr

eme

per

cent

iles

of"rm

-yea

robse

rvat

ions

.Sec

ond,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sw

ith

missing

valu

esfo

rone

orm

ore

variab

les,

tofa

cilit

ate

com

para

bilit

yw

ith

resu

lts

inpre

viou

sta

bles

.Thi

rd,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sfrom

count

ries

with

less

than

1,00

0ob

serv

atio

ns,l

eavi

ngse

ven

count

ries

repre

sent

ed.A

ustr

alia

,Can

ada,

Uni

ted

Stat

es,a

ndth

eU

nited

Kin

gdom

are

the

com

mon

-law

count

ries

,th

ere

star

eco

de-

law

count

ries

.R"

buy-

and-h

old

secu

rity

retu

rnin

clusive

ofdiv

iden

dsove

rth

e"sc

alye

ar;

NI"

annu

alea

rnin

gspe

rsh

are

befo

reex

trao

rdin

ary

item

sde#

ated

bybe

ginnin

gofper

iod

price

;D

IV"

annua

ldiv

iden

ds

per

shar

ede#

ated

bybeg

innin

gof

period

price

;O

CF"

annu

aloper

atin

gca

sh#ow

per

shar

ede#

ated

bybe

ginnin

gofper

iod

price

;N"

the

num

ber

of"rm

/yea

robs

erva

tions

.


mean "scal-year returns range from 3.7% (Japan) to 17.3% (Australia). MeanNI ranges from 1.7% (Japan) to 6.6% (UK). All countries' mean andmedian OCF (from which capital investments are not deducted) exceed NI(from which depreciation, a weighted average of past capital investments, isdeducted).

Consistent with asymmetric conservatism, accounting income is negativelyskewed (all medians exceed means), which contrasts with the positive skew ofstock returns (all means exceed medians). That is, conservative accounting tendsto incorporate economic losses as larger but transitory, capitalized amounts,and to incorporate economic gains as smaller but persistent #ows over time,thus generating the negative skew of accounting income. The di!erential skewof earnings relative to returns calls into question the traditional linearearnings-returns speci"cation, and supports the Basu (1997) piecewise linearversion.

Also consistent with our model, in all seven countries the rank orderof volatilities across variables is (highest "rst): R, OCF, NI and DIV. Ourinterpretation is: (i) accounting income is a lagged function of present andpast years' returns (a type of moving average), and hence has lowervolatility than individual-year stock returns; (ii) cash #ow from operations isnoisier than income, and hence is more volatile; and (iii) dividends is afurther lagged function of accounting income, and hence is the least volatilevariable. We comment on the relative magnitudes of the R and NI volatilitiesbelow.

Panel B reports the distribution of ending months for companies' "scal years.December is the norm in Canada, France, Germany and US. March is the normin Japan and June is the norm in Australia. Code law countries exhibit moreconformity with their norm, presumably due to the greater in#uence of regula-tion and tax accounting in those countries, discussed below. UK, which wecategorize below as the least regulated country in the sample, exhibits thegreatest dispersion in "scal year-ends. Panel C reports the sample distributionby calendar year.

4. Hypotheses and tests: international timeliness and conservatism of accountingincome

We develop hypotheses concerning the in#uence of institutional variables onthe two properties of accounting income captured by our model: timelinessand conservatism (or asymmetric timeliness). There have been numerousattempts to classify nations' accounting systems, based on a variety of institu-tional variables, but little empirical research has been directed at determining


8American Accounting Association (1977) classi"es accounting systems using eight variables:political system (traditional oligarchy, totalitarian oligarchy, modernizing oligarchy, tutelarydemocracy, political democracy), economic system, stage of political development, "nancial report-ing objectives, source of accounting standards, education and licensing of accountants, mechanismto enforce standards, and accounting client (public or private). Nobes (1992) and Nobes and Parker(1995, Chapter 4) survey classi"cation schemes. How these variables a!ect important properties ofaccounting information internationally is largely untested. International accounting texts typicallylist variables to justify classi"cations, without correlating them with the national accountingstandards listed in subsequent chapters, let alone with properties of the "nancial statements actuallyprepared under those standards.

9David and Brierley (1985) provide a survey.

which variables explain di!erences in important properties of accountingincome.8

4.1. Demand for timely accounting income in code and common law governance

Perhaps the most fundamental institutional variable causing accountingincome to di!er internationally is the extent of political in#uence on bothstandard setting and enforcement. An admittedly imperfect proxy for politicalin#uence } but a proxy our results legitimate } is whether standard setting andenforcement occur under codi"ed law (a governmental process) or common law(a market process).9 We hypothesize that the demand for timely incorporationof economic income in accounting income is lower under the code-law &stake-holder' model of corporate governance than under the common-law &share-holder'model. Below we describe the origins of and salient features of code- andcommon-law institutional environments. The discussion highlights the di!er-ences between the two systems and their di!erential implications for propertiesof accounting income. However, there is overlap between the code- and com-mon-law institutions and accounting standard setting, which likely weakensempirical support for the hypotheses we develop.

Common law arises from individual action in the private sector. It emphasizesfollowing legal procedure over rules (David and Brierley, 1985, p. 24; Posner,1996). Common laws } including accounting standards } evolve by becomingcommonly accepted in practice. While it might be e$cient for private-sectorbodies to codify generally accepted accounting rules and make them binding ontheir members, such standards arise in an accounting market, not in govern-ment. Common law enforcement is a private matter, involving civil litigation.Common law originated in England and it is now found in UK and manyformer British colonies. The common-law countries in our sample are Australia,Canada, UK and US.

Common law historically has evolved to meet the demands of contractingin markets. The &shareholder' model of corporate governance, in which


10Romans implanted code-law (also referred to as civil law, La Porta et al., 1997) in manycontinental European countries. The most detailed accounting code is the French Plan comptablege&ne& ral, adapted under occupation from Germany's 1937 standardization of accounting for warplanning (Standish, 1997, p. 60). Japan has a code-law system, derived from the German legal andFrench accounting systems during the Meiji Era (1868}1910). Scandinavian law is another code-lawcategory with origins in Roman law (David and Brierley, 1985 and La Porta et al., 1997).

11See Nobes and Parker (1995, Chapter 12), Roe (1994) and Miwa (1996).

shareholders alone elect the governing board, predominates in common-lawcountries. Alchian and Demsetz (1972) argue this is e$cient due to the addi-tional incentive of residual claimants to e!ectively monitor managers. Com-pared to code-law governance, board members are less likely to hold largeblocks, there is more monitoring of managers by external debt and equitymarkets (including analysts), and lenders and employees seldom have boardrepresentation. We hypothesize that, because parties contracting with the "rmoperate at greater &arm's length' from managers, information asymmetry incommon-law countries is more likely to be resolved by timely public disclosure.

Code-law originates from collective planning in the public sector. Govern-ments or quasi-governmental bodies, such as France's Conseil National de laComptabilite& or Japan's Business Accounting Deliberation Council (which ad-vises the Ministry of Finance) establish code-law accounting standards.10 Thecode prescribes regulations ranging from abstract principles (e.g. &prudence') todetailed procedures (e.g. the format of "nancial statements). Code-law enforce-ment is a governmental function, involving administrative bodies undertakingcriminal prosecution for code violation. Code-law countries in our sample areFrance, Germany and Japan.

The code- and common-law common classes overlap in practice, in that"nancial reporting in no country is determined in a purely market or planningsystem. The UK Companies Acts imposed codi"cation on a predominantlycommon-law system during the 19th century. In the US, the Securities andExchange Acts played a similar role in the 1930s, among other things creatingthe SEC as a US government agency with responsibility for regulation ofaccounting standards. We acknowledge that the code- and common-law separ-ation is not watertight and that there is some overlap in the nature of standardsetting in the two categories of countries. Nevertheless, we believe the distinctionis informative because, after reviewing the institutional details, we conclude itcaptures di!erences in the extent of political in#uence on accounting standard-setting (versus private contracting in the markets).

An important di!erence between common-law and code-law countries is themanner of resolving information asymmetry between managers and potentialusers of accounting income, including debt and equity investors, employees,suppliers and customers. Code-law corporate governance tends to be conductedby elected or appointed agents for these parties.11 These agents tend to be


12La Porta et al. (1997) predict that, in addition to di!erences in legal systems, di!erences in theextent of law enforcement a!ects the development of capital markets across countries. It is possiblethat variation in law enforcement has predictive power with respect to the demand and supply ofaccounting information that we ignore.

13Daimler-Benz reported 1993 German-rule income of DM615 million, but subsequent USGAAP disclosure revealed that a loss of DM1839 million had been hidden by various accountingadjustments (Ball, 1998).

informed by private &inside' access to information. Thus, employees and stock-holders each elect 50% of the supervisory board of German Aktiengesellschaft(stock corporations). Banks typically dominate stock voting due to their large-block holdings and due to the German practice of banks voting individuals'stocks as agents (KoK ndgen, 1994). The supervisory board appoints and monitorsthe managerial board, and approves the "nancial statements. For the system tobe tractable, the number of contracting parties must be small, so managers haveclose relations with intermediaries: notably, banks, other "nancial institutions,labor unions, governments, major customers and suppliers. There is no pre-sumption that parties operate at a distance. Consequently, the demand fortimely public disclosure in code-law countries is not as great as in common-lawcountries.12 We propose that this reduces the demand for timely incorporationof economic income in code-law accounting income.

Conversely, we propose that code-law accounting income meets other de-mands. The stakeholder model views accounting income as a common &pie'divided among stakeholders, as dividends to shareholders, taxes to govern-ments, and bonuses to managers and perhaps employees. The portfolio weightsof managers and employees typically are skewed toward their employer "rms, sotheir incentive is to reduce volatility in payouts. Regulation of bank leverageratios penalizes volatility in bank income and thus in the accounting incomeand/or dividends on their equity investments. Code-law banking systems typi-cally are hierarchical, so bank representatives are well aware of governments'incentives to reduce volatility of tax receipts. Employee representatives typicallyare re-elected annually. While incentives to reduce volatility in accountingincome exist in common-law countries (Healy, 1985), we hypothesize thatcode-law governance ampli"es them.

Volatility can be reduced, at the expense of timeliness, through accountingmethods that &smooth' accounting income over time, incorporating economicincome gradually over several periods. The Recognition Rules inherentlysmooth accounting income in all countries (though we also argue they fre-quently are overridden in the case of negative economic income). Nevertheless,code-law accounting gives managers considerably more latitude in timing in-come recognition. In good years, income can be reduced by asset write-downs(e.g. excessive allowances for bad debts), by provisions (e.g. excessive provisionsfor future losses or future expenses) and by transfers to reserves. In bad years,accounting income can be increased by reversing these adjustments.13 We


14The code-law ratio range is 0.06}0.29; common-law is 0.33}0.36 (excluding 0.23 for UK,discussed below).

predict that code-law accounting income incorporates a lower proportion ofcurrent-period economic income, &smoothing' its incorporation to a greaterextent over time.

H1: Code-law countries+ accounting incomes are more &smoothed+ and less timely

in incorporating current-period changes in market value than common-lawcountries'.

The summary statistics reported in panel A of Table 1 are broadly consistentwith this hypothesis, in that code-law countries exhibit a lower ratio of NIvolatility to R volatility.14 We formally test the hypothesis using a variety ofBasu regressions (4): individual-country regressions with observations pooledacross time; annual Fama/MacBeth cross-sectional regressions for individualcountries; a secondary sample of eighteen other countries; and (in Section 4.3)a pooled-countries regression that allows formal tests of di!erences amongcountries.

4.1.1. Evidence from country regressions with observations pooled across timeInitially, we test Hypothesis 1 by comparing the adjusted R2's of the indi-

vidual country, each estimated from pooled time series and cross-section data.While we make a distinction between common- and code-law countries, thecategories are by no means homogeneous. For example, in Section 4.4 we arguethat the UK is di!erent from the remaining three common-law countries due todi!erences in the extent of accounting regulation, litigation environment, andthe existence of public versus private debt. The empirical analysis thereforereports results separately for the common-law countries excluding the UK, forthe UK, and for the code-law countries.

The left-hand side of panel A of Table 2 reports the results and Fig. 1 graphsthem. With the possible exception of France, there is a clear di!erence betweenindividual code-law country R2s (ranging from 4.2% to 12.6%) and common-law country R2's (from 9.1% to 17.0%), consistent with the hypothesis. Whencountries are grouped, the R2 for the pooled code-law sample is 5.2%, approx-imately one-third the common-law equivalent of 14.4% (excluding UK). Theinternational di!erences in the degree of income-return association we docu-ment are similar to those in Alford et al. (1993), despite slightly di!erent timeperiods and models.

To assess whether the di!erences in R2's are statistically signi"cant, weestimate the standard deviation of estimated R2's, which Cramer (1987) shows isa function of sample size, the number of independent variables and the true R2.For four independent variables including intercepts, a true R2 of 5%, andsample size 1,000 (5,000, 20,000), the standard deviation of the estimate is 1.3%


Fig. 1. International di!erences in earnings timeliness R-squares from individual country regres-sions of earnings on (a) annual return and (b) annual return times negative return dummy.

(0.6%, 0.3%). For a true R2 of 15%, it is 2.0% (0.8%, 0.5%). Assumingindependence across countries, di!erences in the order of 5% between oursample countries are signi"cant, and the pooled common-law sample R2 of14.4% signi"cantly exceeds the code-law R2 of 5.2%. Since independence likelyis violated, we report alternative tests below.

4.1.2. Annual cross-sectional regressionsAn alternative test, due to Fama and MacBeth (1973), estimates separate

annual cross-sectional Basu regressions (4), using observations for each countryas well as for each of three pooled country groups (code-law, common-law andUK). We exclude the 6 of 77 country-years with fewer than 20 "rm observations.For each country and each country-group, the right-hand side of panel A ofTable 2 reports the time-series average of the estimated annual slope coe$cientsand the average of the annual regression R2's and their t-statistics. The t-statisticis the ratio of the sample mean to the standard deviation of the time-seriesdistribution of the estimated coe$cients or R2's, divided by the square root ofthe number of annual cross-sections ("11 or 10 depending on the availabilityof 20 or more observations). From this test, the mean R2 for the common-lawgroup is 16.7% (t-statistic "12.85) compared to 7.0% (t-statistic "6.07) for thecode-law group and 16.8% (t-statistic "8.81) for the UK. The common-lawcountries' average R2 is signi"cantly greater than that for the code-law countriesat the 0.01 level.


Tab

le2

Con

tem

por

aneo

us

asso

ciat

ion

bet

wee

nea

rnin

gsan

dre

turn

s.Sta

tist

ics

bas

edon

pool

edcr

oss

-sec

tion

and

tim

e-se

ries

regr

essions

and

annual

cross

-se

ctio

nal

regr

essions

using"rm

-yea

rob

serv

atio

ns

for

each

countr

y.In

terc

epts

are

not

repor

ted!

Pan

elA

:N

I"b 0

#b 1

RD#

b 2R#

b 3RHR

D#

e

Pool

edre

gres

sions

Annua

lcr

oss

-sec

tion

alre

gres

sion

s

b 2t(b 2

)b 3

t(b 3

)A

dj.R

2

(%)

NA

vg.b

2t(b 2

)A

vg.b

3t(b 3

)A

dj.R

2

(%)

t(R

2)

Aust

ralia

!0.

01!

0.53

0.37

8.63

9.1

1,32

10.

020.

990.

336.

1211

.15

3.68

Can

ada

0.00

0.12

0.40

17.2

117

.02,

901

0.01

0.51

0.39

6.82

18.1

95.

59U

SA

0.03

8.57

0.29

34.0

214

.721

,225

0.03

3.43

0.33

14.3

417

.15

13.8

2U

K0.

0410

.14

0.15

13.3

213

.85,

758

0.05

4.23

0.14

5.41

16.7

68.

81Fra

nce

0.08

7.30

0.07

2.30

12.6

1,05

40.

062.

950.

142.

7117

.59

4.52

Ger

man

y0.

054.

280.

103.

275.

41,

245

0.04

2.01

!0.

01!

0.06

7.47

4.85

Japa

n0.

015.

950.

012.

584.

26,

855

0.00

0.55

0.02

3.26

6.88

4.07

Com

mon

0.02

7.07

0.31

39.1

014

.425

,447

0.02

2.36

0.34

15.3

116

.74

12.8

5U

K0.

0410

.14

0.15

13.3

213

.85,

758

0.05

4.23

0.14

5.41

16.7

68.

81C

ode

0.04

13.2

70.

012.

195.

29,

154

0.04

4.55

0.04

1.91

7.00

6.07


Tab

le2

(cont

inued

)

Pan

elB:

R"

b 0#

b 1N

I#e

R"

b 0#

b 1N

I#b 2

*N

I#e

b 1t(b 1

)A

dj.R

2(%

)b 1

t(b 1

)b 2

t(b 2

)A

dj.R

2(%

)N

Aust

ralia

0.47

6.30

2.9

0.44

5.75

0.12

2.84

3.4

1,32

1C

anad

a0.

7914

.49

6.7

0.72

13.0

60.

266.

448.

02,

901

USA

0.90

46.0

89.

10.

7636

.52

0.37

18.5

810

.521

,225

UK

1.44

26.6

711

.01.

2120

.40

0.46

9.07

12.2

5,75

8Fra

nce

1.40

12.2

212

.31.

2810

.86

0.46

4.23

13.7

1,05

4G

erm

any

0.76

7.84

4.6

0.71

6.78

0.13

1.30

4.7

1,24

5Ja

pan

3.16

16.8

24.

02.

2311

.04

2.85

11.8

35.

96,

855

Com

mon

0.85

47.5

28.

20.

7439

.71

0.30

18.9

29.

425

,447

UK

1.44

26.6

711

.01.

2120

.40

0.46

9.07

12.2

5,75

8C

ode

1.44

22.4

45.

21.

2919

.10

0.47

7.04

5.7

9,15

4

!Sam

ple

cons

ists

of40

,359"rm

-yea

robs

erva

tions

sele

cted

from

the

Glo

bal

Van

tage

indust

rial

/Com

mer

cial

and

Issu

e"le

sove

r19

85}95

,us

ing

the

follo

win

gpro

cedu

re.F

irst

,fo

rea

chva

riab

le(see

bel

ow)w

eel

imin

ate

the

two

extr

eme

per

cent

iles

of"rm

-yea

robse

rvat

ions

.Sec

ond,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sw

ith

missing

valu

esfo

rone

orm

ore

variab

les,

tofa

cilit

ate

com

para

bilit

yw

ith

resu

lts

inpre

viou

sta

bles

.Thi

rd,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sfrom

count

ries

with

less

than

1,00

0ob

serv

atio

ns,l

eavi

ngse

ven

count

ries

repre

sent

ed.A

ustr

alia

,Can

ada,

Uni

ted

Stat

es,a

ndth

eU

nited

Kin

gdom

are

the

com

mon

-law

count

ries

,th

ere

star

eco

de-

law

count

ries

.R"

buy-

and-h

old

secu

rity

retu

rnin

clusive

ofdiv

iden

dsove

rth

e"sc

alye

ar;

RD"

equa

lson

eif

retu

rnR

isneg

ativ

ean

dze

rooth

erw

ise.

NI"

annu

alea

rnin

gper

shar

ebef

ore

extr

aord

inar

yitem

de#ec

ted

by

beg

innin

gofpe

riod

pric

e;*N

I"ch

ange

inN

Ipe

rsh

are,

adju

sted

for

stock

split

san

dst

ock

dev

iden

ds,

de#at

edby

beg

innin

gofper

iod

pric

e;N"

the

num

ber

of"rm

/yea

rob

serv

atio

ns.

Inpan

elA

,the

repo

rted

t-st

atistics

forth

eav

erag

eoft

heslope

coe$

cien

tsfrom

annual

cross

-sec

tion

alre

gres

sionsfo

rea

chco

unt

ryar

eth

era

tiosoft

he

mea

nes

tim

ated

coe$

cien

tsto

the

stan

dar

ddev

iation

ofth

edistr

ibution

of11

annual

estim

ated

slope

coe$

cien

ts,d

ivid

edby

the

squa

rero

otof11

(see

Fam

aan

dM

acBet

h,19

73).


15We describe asymmetric timeliness in incorporating economic losses as &income conservatism.'The concept is related to, but di!erent from, balance sheet conservatism (reporting low book value ofequity by under-stating assets and/or over-stating liabilities). Income conservatism implies balancesheet conservatism, but not vice versa: while code-law companies typically report conservative bookvalues, they also are more likely to boost income in bad years. This reduces the asymmetrictimeliness of accounting income and is di$cult to describe as &conservative' in its e!ect on income.Asymmetric timeliness is di!erent from, but related to, Gray's (1980) concept of conservatism.

4.1.3. Secondary sample of eighteen countriesA secondary sample of eighteen countries with between 100 and 1000 "rm-

year observations provides similar results. We use the Mueller et al. (1997, pp.11}12, Exhibits 1}4) classi"cations of countries as following a British}Americanor a Continental accounting model to proxy for our private-sector common-lawand public-sector code law categories. We thereby classify eight additionalcountries as common-law (Hong Kong, India, Ireland, Malaysia, Netherlands,New Zealand, Singapore, and South Africa), nine as code-law (Austria, Belgium,Denmark, Finland, Italy, Norway, Spain, Sweden, and Switzerland), and addThailand as a code-law country based on our own assessment. While theMueller et al. classi"cations agree with ours for each of the seven countries inour primary sample, we are less con"dent of the eighteen additional countryclassi"cations. Speci"cally, we suspect that many countries followingBritish}American accounting rules nevertheless lack common-law litigationenforcement and exhibit reduced demand for timely accounting income.Classi"cation errors are expected to reduce the signi"cance of the secondarysample results.

For each of the eight common-law and ten code-law countries, we estimatea pooled Basu regression (4). Annual cross-sectional regressions for each coun-try are not feasible due to insu$cient observations. In the interest of parsimony,we summarize the main results. Detailed results in tabulated form are availableto interested readers upon request. The average R2 for the code-law countries is6.5%, the median is 7.0%, the minimum is 1%, and the maximum is 14.6%. Incomparison, the average R2 for the common-law countries is 15.3%, the medianis 12.4%, the minimum is 9.8%, and the maximum is 22.5%. The di!erence inmean R2's of the common and code-law countries is statistically signi"cant. Theresults are consistent with those reported for the seven countries in Table 2 andprovide an independent con"rmation of H

1.

4.2. Universal demand for conservative accounting income

In this section we argue that conservatism, de"ned as asymmetric timelinessin incorporating economic gains and losses (Basu, 1997), is a general property ofaccounting income.15 This section serves as a lead-in to Section 4.3, which


16For the separate positive and negative return regressions, we examined plots of the residualsagainst returns, and found no evidence of non-linearity (which was a concern for the negative-returnsample in particular).

focuses on the e!ect of code- and common-law institutional di!erences on thedegree of accounting conservatism.

Accountants contract to supply users with asymmetrically conservative in-come (i.e., to incorporate economic losses in a more timely fashion than gains)due to three properties of the accounting information market. First, managerspossess speci"c information (Alchian, 1984) that is costly for external users toproduce themselves (is not independently veri"able). Because managers haveasymmetric incentives to disclose positive and negative speci"c information,information of negative innovations in expected future cash #ows (economiclosses) is more credible than positive innovations, and accountants are morelikely to incorporate it in income. Second, lenders are important users ofaccounting information, including income and book value (a function of in-come), and they are asymmetrically a!ected by economic gains and losses(Watts and Zimmerman, 1986). Third, we propose that timely disclosure ofeconomic losses is an important corporate governance mechanism. We assumethat reversing bad investment decisions and strategies is personally more costlyto managers than continuing good ones, and that informed monitoring byboards, analysts, investors and lenders is a mechanism to force them to under-take the cost. For these reasons, accountants supply income and book valuesthat incorporate economic losses in a more timely fashion than economic gains.

The above properties of the market for accounting information are universal(though we argue below that they vary in degree internationally). We thereforeexpect accounting income to be asymmetrically conservative in all countries.Empirically, accounting income should exhibit higher R2's for bad news (i.e.negative "scal-year return) observations than for positive. Table 3 reports thatnegative-return R2's exceed their positive-return counterparts in all common-law countries and in Germany.16 The exceptions are the code-law countriesFrance and Japan, discussed below.

Table 3 and Fig. 2 show that in all seven countries the coe$cient on negativereturns exceeds its counterpart on positive returns. Accounting income in theUS is approximately ten times as sensitive to negative as to positive returns(estimated slopes of 0.32 and 0.03). The median country in terms of relativesensitivity to negative versus positive returns is the UK, with a ratio of approx-imately "ve (0.19: 0.04). A formal test of the asymmetry is provided in Table 2,panel A. The incremental slope b

3on negative "scal-year returns (i.e. return

interacted with the return dummy variable, RHRD) is signi"cantly positive for allseven countries. A Fama}MacBeth test, with the b

3coe$cients estimated from


Table 3Contemporaneous association between earnings and returns separately in good and bad news years

Statistics from pooled cross-section and time-series regressions using "rm-year observations foreach country. Intercepts are not reported!

Good news:NI"b

0#b

1R#e(R50)

Bad news:NI"b

0#b

1R#e(R(0)

b1

Adj. R2 (%) N b1

Adj. R2 (%) N

Australia !0.01 !0.1 813 0.36 10.1 508Canada 0.00 !0.1 1,688 0.40 17.9 1,213USA 0.03 0.8 12,721 0.32 11.7 8,504UK 0.04 3.2 3,612 0.19 11.6 2,146France 0.08 9.0 611 0.16 4.7 443Germany 0.05 2.4 712 0.16 4.9 533Japan 0.01 1.1 3,141 0.02 0.8 3,714

Common 0.02 0.4 15,222 0.33 12.2 10,225UK 0.04 3.2 3,612 0.19 11.6 2,146Code 0.04 3.3 4,464 0.05 1.7 4,690

!Sample consists of 40,359 "rm-year observations selected from the Global Vantage indus-trial/Commercial and Issue "les over 1985}95, using the following procedure. First, for each variable(see below) we eliminate the two extreme percentiles of "rm-year observations. Second, we eliminateall "rm-year observations with missing values for one or more variables, to facilitate comparabilitywith results in previous tables. Third, we eliminate all "rm-year observations from countries withless than 1,000 observations, leaving seven countries represented. Australia, Canada, United States,and the United Kingdom are the common-law countries, the rest are code-law countries.

R"buy-and-hold security return inclusive of dividends over the "scal year;NI"annual earnings per share before extraordinary items de#ated by beginning of period price;N"the number of "rm/year observations.

separate annual cross-sectional country regressions, also is reported in Table 2,panel A. The average incremental slope is signi"cant for all countries exceptGermany.

Unreported results for the secondary sample of 10 code-law and 8 common-law countries also are consistent with asymmetric timeliness being a universalproperty of accounting income. The incremental slope on negative returns ispositive in 12 of the 18 countries.

4.3. Demand for greater income conservatism under common-law governance

We propose that common-law accounting income is more asymmetricallyconservative than code-law, due to greater demand for timely disclosure ofeconomic losses. Each of the three properties of the market for accountinginformation described above is qualitatively weaker in code-law governance.First, the closer code-law relation between managers and agents for contracting


Fig. 2. International di!erences in asymmetry in earnings response to good and bad news; sensitiv-ity to positive and negative returns from pooled regression of earnings on (a) annual return and (b)annual return times negative return dummy.

groups (banks, labor unions) reduces information asymmetry. Second, banksand other "nancial institutions tend to supply both debt and equity capital,making their loss function more symmetric. Third, there is less reliance onexternal monitoring of managers. Consistent with lower demand for timelyincorporation of economic losses, the expected litigation cost of untimelyincorporation is lower. Conversely, in code-law countries there is a demand fora low-volatility income variable, due to the more direct relation betweencurrent-year accounting income and short-term payouts such as dividends andbonuses. Reluctance to cut dividends (Lintner, 1956) leads to a preference forgradually incorporating an economic loss in income over time, for example bywaiting for the reduced underlying cash #ow realizations to occur, as distinctfrom recognizing it as a large, capitalized, transitory amount.

H2: Common-law accounting income is more asymmetrically conservative than

code-law.


17Unreported incremental negative return slopes in the additional 18-country sample are incon-sistent with our hypothesis. While the R2's are consistent with greater conservatism in common-lawcountries and the slopes are consistent with universal conservatism, the slopes do not suggest greaterconservatism in common-law countries. However, due to small sample sizes many of the positivecode-law country incremental slopes are not signi"cant.

18Nobes and Parker (1995, pp. 298}300).

The separate positive and negative return samples in Table 3 reveal that thegreater overall timeliness of common-law accounting income reported above isdue entirely to years with economic losses. For the common-law group, the12.2% negative-return year R2 signi"cantly exceeds the 0.4% in positive-returnyears. In all common-law countries, R2's in negative-return years exceed those inpositive-return years. Accounting income in Australia, Canada and the US overthe sample period essentially ignores current-year economic gains. Common-law accounting income seems directed primarily toward incorporation of eco-nomic losses. With the exception of Germany, the same cannot be said ofcode-law accounting. For the code-law sample, the negative-return year R2 of1.7% actually is lower than 3.3% for the positive-return years. These results areconsistent with the hypothesis that, relative to code-law countries, the common-law demand for accounting income originates more in arm's length corporategovernance, debt contracts, and investor litigation, and less in determiningshort-term payouts.

As another test of a di!erential degree of conservatism in code- and common-law countries, we focus on the slope coe$cients from the Basu regression (4)results in Table 2 and Fig. 2. The b

2jslopes 2 show that accounting income in

every common-law country exhibits less sensitivity to positive returns than inFrance and Germany (accounting income in Japan exhibits the least timelinessfor both return samples). In contrast, the incremental negative-return slopesb3j

for common-law countries range from 0.15 to 0.40, considerably larger thanthe code-law range of 0.01}0.10. The pooled incremental slopes are 0.31 for thecommon-law group as a whole (excluding UK) and 0.01 for the code-law group.The di!erence is statistically signi"cant at the 0.01 level.17

The incremental b3j

slope of 0.01 for Japan is the lowest in the sample (due torelative sample sizes it dominates the code-law group slope). Low incomeconservatism is consistent with the stylized institutional facts for Japan, in whichone-time accounting write-o!s are rare, economic losses reputedly are dribbledslowly into reported income over time (the banking sector being a notoriouscase), there is no provision for under-funded pension liabilities or other post-retirement liabilities, and the &lower of cost or market' rule is not even used forinventories.18


4.3.1. Pooled sampleTo formally test for international di!erences in the degree of conservatism, we

estimate (4) using data pooled across countries:

NI"b0#+

j

b0j

CDj#b

1RD#+

j

b1j

RD CDj#b

2R

#+j

b2j

R CDj#b

3R RD#+

j

b3j

R RD CDj. (5)

Six dummy variables identify the country of the accounting standards used foreach "rm/year, with CD

j"1 for "rm/years under country j and "0 otherwise.

US is the &base country,' with zero values for all country dummy variables. Thecoe$cient b

2on return measures the incorporation of current economic income

into US "rms' accounting incomes. The coe$cient b2j

on the product of returnand the country j dummy variable measures the country's incremental incorpo-ration, relative to the US. The coe$cients b

3and b

3jmeasure the asymmetric

conservatism of accounting income under US standards and the incrementalconservatism under other countries' standards. Thus, b

2#b

2j#b

3#b

3jmeasures the incorporation of negative economic income in country j. Table 4reports results of the pooled country-dummy regression Eq. (5). Accountingincome in all sample code-law countries (France, Germany and Japan) exhibitssigni"cantly less incremental sensitivity to negative economic income thanunder US standards.

4.4. Regulation, litigation and debt diwerences among common-law countries

The distinction between common-law and code-law countries provides usefulinsights, but as we observe above the categories are by no means homogeneous.We consider two important institutional di!erences within the class of com-mon-law countries: the method and extent of their regulation of accounting; andthe extent to which their securities litigation rules favor plainti!s.

4.4.1. RegulationWe propose that, among common-law countries, income conservatism in-

creases with regulation of accounting standard setting and enforcement. Build-ing on Peltzman (1976), Watts and Zimmerman (1986, pp. 229}231) argue thatthe political process and the SEC as its agent have an incentive to avoidperceived responsibility for investor losses. We argue that responsibility isattributed more to managers, and less to the political process, if losses aredisclosed in "nancial statements in a more timely fashion. Consequently, regula-tion adds criminal penalties to the common-law civil remedy of damages foruntimely disclosure of material bad news. Accounting is regulated to varyingdegrees in all common-law countries. For reasons summarized below, we


Table 4Comparative asymmetry in the contemporaneous returns-earnings relation: pooled regressions withindividual-country e!ects!

Reported statistics are for the following model using the pooled cross-section and time-series of"rm/year observations for all countries:

NI"b0#+

j

b0j

CDj#b

1RD#+

j

b1j

RD CDj#b

2R#+

j

b2j

R CDj#b

3R RD

#+j

b3j

R RD CDj#e

Results are not reported for the intercept, the negative-return intercept, and their respective countrydummies. The country category models use the common law countries of Australia, Canada and theUS as the base category; dummies are used for (1) the UK and (2) the code law countries of France,Germany and Japan.

Coe!. t-Stat.Panel A: country dummies model

Earnings &good news+ sensitivity

b2

(Return) 0.03 9.88b2j

(Return*Country Dummies):Australia !0.03 !3.95Canada !0.03 !3.80UK 0.02 2.40France 0.05 4.16Germany 0.03 1.57Japan !0.02 !2.64

F-stat. for country return dummies 11.35p(0.01

Incremental *bad news+ sensitivity

b3

(Return Dummy*Return) 0.29 39.21b3j

(Return Dummy*Return*Country Dummies)Australia 0.07 2.64Canada 0.11 5.28UK !0.14 !8.26France !0.22 !5.81Germany !0.19 !4.41Japan !0.29 !15.74

F-stat.: Country Negative Ret Dummies 67.44p(0.01

Regression

N 40,359Adj. R2 15.4%F 272.98


Table 4 (continued)Panel B: country category dummies model

Earnings *good news+ sensitivity Coe!. t-stat.

b2

(Return) 0.02 8.36b2j

(Return*Country Category Dummy):UK 0.02 3.52Code 0.01 2.37

F-stat. for country category return dummies 7.94p(0.01

Incremental &bad news+ sensitivity

b3

(Return Dummy*Return) 0.31 46.23b3j

(Return Dummy*Return*Country Category Dummy):UK !0.16 !9.50Code !0.30 !19.00

F-stat.: Country Category Negative Return 200.92Dummy p(0.01

Regression

N 40,359Adj. R2 14.9%F 642.07

!Sample consists of 40,359 "rm-year observations selected from the Global Vantage indus-trial/Commericial and Issue "les over 1985}95, using the following procedure. First, for eachvariable (see below) we eliminate the two extreme percentiles of "rm-year observations. Second, weeliminate all "rm-year observations with missing values for one or more variables, to facilitatecomparability with results in previous tables. Third, we eliminate all "rm-year observations fromcountries with less than 1,000 observations, leaving seven countries represented. Australia, Canada,United States, and the United Kingdom are the common-law countries, the rest are code-lawcountries.

R"buy-and-hold security return inclusive of dividends over the "scal year;NI"annual earnings per share before extraordinary items de#ated by beginning of period price;RD"the proxy for bad news"1 if R(0 and "0 otherwise;CD

j"country identi"er"1 for "rm/years in country j and "0 otherwise. USA is the &base

country' with CDj"0 ∀j.

characterize UK as the least regulated accounting market among our samplecommon-law countries.

The Securities Exchange Act of 1934 created the SEC as a US Governmentagency authorized to mandate and administer accounting standards. The SECby-and-large has accepted the standards of the accounting profession, but


nevertheless has retained a close supervisory role. It has intervened in standardsetting on several occasions, as has Congress on occasions (e.g., in the debate onmark-to-market accounting, and by examining some "rms' in-process R&D write-o!s). We conclude that the US operates a closely regulated common-law system.

The e!ect is to compound the civil and criminal penalties for non-disclosureof material bad news. It has become a violation of both the common-lawobligation to disclose (the civil law penalty for which is remedial damagesawarded in private litigation) and a similar and sometimes stronger statutoryobligation (the criminal law penalty for which is a "ne, incarceration and/orprohibition from practice). We conjecture that this dual system of penaltiescreates additional incentives to recognize economic losses in regulated com-mon-law countries.

Australia and Canada are widely viewed as having evolved from the looselyregulated UK model (described below) to the US regulatory model. Bothcommenced with a largely self-regulating profession, initially adopted provincial(rather than federal) regulation, and then moved to a system in which govern-mental or semi-governmental bodies set national accounting standards. Austra-lia created a regulatory body (now constituted as the Australian SecuritiesCommission) &obviously modeled on the SEC' (Nobes and Parker, 1995, p. 106).It also removed standard setting from the private sector, assigning it in 1984 toa government-appointed authority (now the Australian Accounting StandardsBoard), whose accounting standards can be overturned only by the AustralianParliament. These changes predate our sample period, so we characterizeAustralian accounting as highly regulated (see Choi and Mueller, 1992, p. 86).

Canada has evolved to a regulated federal model in a similar fashion. Ac-counting was increasingly regulated with the establishment of the provincialsecurities commissions, and the enactment of provincial and federal securitieslaws. Various revisions of the Ontario Companies Act have been particularlyin#uential. In 1975, federal regulations required "nancial statements to complywith the accounting standards promulgated in the Canadian Institute of Char-tered Accountants' CICA Handbook, thereby giving them statutory status. TheSEC has indirectly in#uenced Canadian accounting, due to large Canadiancorporations listing in New York.

Among the common-law countries in our sample, we classify the UK ashaving the least regulated accounting market over our sample period. There isno UK regulatory body comparable to the SEC in the US. UK "nancialmarkets (the City of London) have historically been viewed as primarily &self-regulating,' and the UK parliament has seldom intervened in accounting mat-ters (Choi and Mueller, 1992, Chapter 3; Nobes and Parker, 1995, Chapter 6;Radebaugh and Gray, 1997, Chapter 5). In 1990, UK established the AccountingStandards Board (ASB), modeled on the US FASB and accountable to anewly created Financial Reporting Council. The ASB was authorized to issueFinancial Reporting Standards (FRSs) that are backed by law (Radebaugh and


19The English rule is applied in UK, Australia, and most of Western Europe (Posner, 1986, p.537), though not uniformly. It a!ects expected litigation costs via the frequency of litigation and thelegal costs of the defendant, but not the size of awards, with unclear net e!ect. Katz (1987) argues:&While it is conceivable that the English rule would lower the total number of cases brought to trial,it would likely increase the average expenditure per case. 2Even if the number of cases were to fallby as much as 30 percent, total expenditure could rise by more than 50 percent under the Englishrule.'Katz' analysis does not take account of one institutional detail, however: English courts awardonly reasonable costs of successful defendants, thus reducing the incentive to spend. See also Shavell(1982), Posner (1986, pp. 534}540), Hughes and Snyder (1995), Posner (1996) and Miceli (1997).

Gray, 1997, pp. 91}92; Choi and Mueller, 1992, pp. 116}118). Nevertheless, thelaw allows a company to deviate from FRSs if it discloses the e!ect on itsaccounts. Even these changes did not take place until near the end of our sampleperiod, 1985}95. We therefore treat UK accounting as less regulated than othersample common-law countries (Australia, Canada and US).

4.4.2. LitigationAmong common-law countries, we also propose that income conservatism

increases in the expected costs to accounting "rms and their clients fromsecurities litigation. Expected litigation costs a!ect managers' and auditors'disclosure decisions (Kothari et al., 1988). The expected costs are a function oflawsuit probability, award size and legal fees. Securities lawsuits induce a de-mand for conservatism because the payo! function is asymmetric: they almostinvariably allege investor losses arising from insu$ciently conservative dis-closures. We expect countries with higher expected litigation cost of non-disclosure are more likely to demand accounting income that incorporateseconomic losses in a timely fashion.

Our assumption is that expected litigation costs are lower in the UK than inAustralia, Canada and US. Relevant UK institutional facts include: punitivedamages are more di$cult to obtain; juries are seldom used in civil litigation;absence of class action suits; and the so-called &English rule,' under which losingplainti!s pay part of defendants' costs.19 In code-law countries, civil litigation iscomparatively rare and the size of awards is comparatively small.

4.4.3. Private debtUK corporate debt is predominantly private. We conjecture there is less

information asymmetry between managers and private lenders than in the caseof public debt, thus reducing the demand for timely incorporation of economiclosses in UK accounting income.

4.4.4. Regulation, litigation and private debt considered jointlyOverall, we classify UK as having lower regulatory and litigation costs and

predominantly private debt, so we predict less conservatism in UK accountingincome than in other common-law countries. Nevertheless, relative to code-lawcountries, UK is expected to be more income conservative.


20 In response to a prior draft of this paper, Pope and Walker (1999) claim that the UK incomebefore extraordinary items appears less conservative than US because UK "rms enjoy greaterdiscretion in reporting economic losses as extraordinary items. They use data from Datastreamwhereas we use the Compustat Global Vantage database. Reconciliation of the di!erences betweenthe two studies is beyond the scope of our research.

H3: U.K. accounting income is less conservative than in other common-law coun-

tries and more conservative than in code-law countries.

The data weakly support this hypothesis. The UK negative-return R2 of11.6% in Table 3 falls between that of the common-law group (12.2%) and thecode-law group (1.7%). The incremental negative-return slope b

3of 0.15 in

panel A of Table 2 is signi"cantly smaller than those for the other common-lawcountries (b

3from 0.29 to 0.40), and it is marginally signi"cantly greater than

those for the code-law countries (b3

from 0.01 to 0.10).20

4.5. Advantage of the Basu (1997) piecewise linear model

The misspeci"cation in linear models for the common-law countries is seenfrom their generally lower R2's in panel B of Table 2, compared with the Basupiecewise linear regressions in panel A, which is graphically depicted in Fig. 3.Code-law countries exhibit considerably less asymmetry, and hence their R2'sdi!er little between the speci"cations. In contrast, the asymmetry is su$cientlystrong in Australia, Canada and US to cause R2's from the Basu model to beapproximately 1.6 to 3.2 times their linear model equivalents. Thus, linearearnings-returns models are potentially misleading in a common-law contextand in international comparisons, due to substantial international di!erences intiming of economic loss incorporation.

5. Hypotheses and tests: accounting income versus dividends and cash 6ows

In this section, we compare accounting income with dividends and cash #owsinternationally. We argue that code-law institutional links between current-yearincome and dividends imply they incorporate similar information. Consistentwith this hypothesis, we report that accounting income is timelier than divi-dends in common-law countries but not in code-law countries. Income is lesstimely than dividends in Germany and Japan. In contrast, accounting income istimelier than cash #ows from operations in all countries.

5.1. Laws and practices linking current-period income and dividends

Accounting income is in#uenced, in varying degrees across countries, by "rms'current dividend decisions. One in#uence results from laws on the taxation


Fig. 3. International di!erences in the contemporaneous relation between earnings and returns;R-squares from (1) the pooled reverse regressions of earnings against (a) annual return and (b)Annual return times negative return dummy and (2) annual return against NI and *NI.

21Except as indicated, taxation facts are from annual editions of Coopers & Lybrand Interna-tional Tax Summaries.

of corporate income and of dividend distributions.21 We address these linkagesin code and common-law countries separately.

5.1.1. Code-law links between dividends and accounting incomeIn the code-law &stakeholder' corporate governance model, governments

frequently are viewed as stakeholders, and tax payments are viewed as govern-ment's share of the same income &pie' from which dividends and bonuses arepaid. Politics requires reported and taxable incomes to converge, particularlysince the government is responsible for both tax and accounting codes. Tax rulesin the three sample code-law countries allow a deduction against taxable incomeonly if also taken against reported accounting income (Choi and Mueller, 1992,p. 104). In code-law countries generally, it thus is considered imprudent toreport income in excess of that required to justify dividends and bonuses, tominimize corporate tax. The consequence is that accounting income is in-#uenced by short-term dividend policy.


Table 5Contemporaneous association between dividends and returns!

Statistics are from regressions using the pooled cross-section and time-series of "rm/year observa-tions for each country. Intercepts are not reported.

Panel A: DI<"b0#b

1RD#b

2R#b

3RHRD#e

b2

t(b2) b

3t(b

3) Adj.R2(%) N Ratio

Australia 0.00 !1.85 0.06 8.67 11.1 1,321 0.82Canada !0.01 !9.06 0.04 11.26 7.3 2,901 2.34USA !0.02 !26.54 0.05 34.24 9.2 21,225 1.60UK 0.01 8.69 0.02 9.25 9.7 5,758 1.42France 0.01 4.22 0.02 3.90 8.8 1,054 1.43Germany 0.02 6.66 0.01 0.99 9.6 1,245 0.56Japan 0.00 2.28 0.01 9.72 5.5 6,855 0.78

Common !0.01 !25.23 0.05 35.67 8.4 25,447 1.70UK 0.01 8.69 0.02 9.25 9.7 5,758 1.42Code 0.01 10.08 0.01 3.78 5.5 9,154 0.96

The in#uence is particularly strong in Germany, where (Nobes and Parker1995, pp. 269}272) the Handelsgesetzbuch (&commercial code') includes the Mas-sgeblichkeitsprinzip (&authoritative principle') that tax accounting be based on the"rm's Handelsbilanz (&commercial balance sheet'). Choi and Mueller (1992, p. 96)conclude of Germany: &The dominance of tax accounting rules means that there isliterally no di!erence between "nancial statements prepared for tax purposes and"nancial statements published in "nancial reports. 2Financial reports re#ect taxlaws } not primarily the information needs of investors and other "nancial marketparticipants.' The relation between income and dividends is tightened in Germanyby two additional institutional factors. First, Federal law forbids managementfrom paying dividends less than 50% of income without stockholders votingapproval. Second, undistributed pro"ts are taxed at a higher rate than distributedpro"ts (currently 45% versus 30%, excluding the &solidarity surcharge').

These code-law institutional factors impose an additional role on the (alreadyreduced) public-disclosure role of accounting income: corporate policy on cur-rent payouts to &stakeholders,' including governments (via taxation), share-holders (via dividends), and managers and employees (via bonuses). If thesedistributions per se are uninformative, then the testable implication is thatcode-law accounting income is noisier and less oriented to incorporating currenteconomic income, relative to dividends, than common-law income.

H4: Code-law accounting income is less timely relative to dividends than common-

law.

Table 5 reports the contemporaneous association between returns and divi-dends, for the same sample of "rm-years as for accounting income in Table 2.


Tab

le5

(cont

inued

)

Pan

elB

:D

I<"

b 0#

b 1R#

e(R5

0)D

I<"

b 0#

b 1R#

e(R(

0)

b 1A

dj.R

2(%

)N

Rat

ioV

uong

b 1A

dj.R

2(%

)N

Rat

ioV

uong

Aust

ralia

0.00

0.2

813

!0.

360.

650.

0618

.350

80.

552.

41C

anad

a!

0.01

3.8

1,68

8!

0.02

5.22

0.03

7.9

1,21

32.

26!

3.69

USA

!0.

024.

412

,721

0.18

9.67

0.04

10.4

8,50

41.

12!

1.48

UK

0.01

1.9

3,61

21.

73!

1.74

0.03

9.3

2,14

61.

24!

1.48

Fra

nce

0.01

2.2

611

4.12

!2.

460.

039.

544

30.

491.

88G

erm

any

0.02

4.8

712

0.50

1.43

0.02

3.8

533

1.31

!0.

6Ja

pan

0.00

0.1

3,14

111

.00

!2.

740.

014.

53,

714

0.19

6.06

Com

mon

!0.

013.

415

,222

0.13

8.71

0.04

10.3

10,2

251.

19!

2.45

UK

0.01

1.9

3,61

21.

73!

1.74

0.03

9.3

2,14

61.

24!

1.48

Cod

e0.

011.

54,

464

2.17

!2.

870.

012.

74,

690

0.62

1.89

!Sam

ple

cons

ists

of40

,359"rm

-yea

robs

erva

tions

sele

cted

from

the

Glo

bal

Van

tage

indust

rial

/Com

mer

cial

and

Issu

e"le

sove

r19

85}95

,us

ing

the

follo

win

gpro

cedu

re.F

irst

,fo

rea

chva

riab

le(see

bel

ow)w

eel

imin

ate

the

two

extr

eme

per

cent

iles

of"rm

-yea

robse

rvat

ions

.Sec

ond,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sw

ith

missing

valu

esfo

rone

orm

ore

variab

les,

tofa

cilit

ate

com

para

bilit

yw

ith

resu

lts

inpre

viou

sta

bles

.Thi

rd,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sfrom

count

ries

with

less

than

1,00

0ob

serv

atio

ns,l

eavi

ngse

ven

count

ries

repre

sent

ed.A

ustr

alia

,Can

ada,

Uni

ted

Stat

es,a

ndth

eU

nited

Kin

gdom

are

the

com

mon

-law

coun

trie

s,th

ere

star

eco

de-

law

coun

trie

s.H

ow

ever

,the

com

mon

law

coun

try

cate

gory

inth

eth

ird

row

from

the

bott

om

excl

udes

the

UK

.R"

buy-

and-h

old

secu

rity

retu

rnin

clusive

ofdiv

iden

dsove

rth

e"sc

alye

ar;

RD"

1if

R"

0an

dR

D"

0if

R'

0;D

IV"

annua

ldiv

iden

ds

per

shar

ede#

ated

bybeg

innin

gof

period

price

;N"

the

num

ber

of"rm

/yea

robs

erva

tions

.R

atio"

inpan

elA

,the

ratio

ofth

eR

2'soft

heeq

uiva

lent

earn

ings

(Tab

le2)

and

div

iden

ds(this

tabl

e)re

gres

sions

,in

pan

elB

,the

ratio

oft

heR

2'soft

he

equi

vale

ntea

rnin

gs(T

able

3)an

ddiv

iden

ds

(thi

sta

ble

)re

gres

sions.

Vuo

ng

stat

isticis

alik

elih

ood

ratio

stat

istic

that

com

par

esth

e"tof

there

turn

/div

iden

dm

ode

lin

this

table

inpan

elB

agai

nstth

ere

turn

/ear

nin

gsm

ode

lin

Tab

le3,

bot

hes

tim

ated

sepa

rate

lyusing

the

good

new

san

dba

dnew

sre

turn

obse

rvat

ions

.T

heV

uong

stat

istic

isdistr

ibut

edunit-N

orm

al,i.e

.,a

Z-st

atistic.


Fig. 4. Earnings versus dividends; R-squares from the individual country reverse regressions of(1) NI and (2) DIV against (a) annual return and (b) annual return times negative return dummy.

Fig. 4 depicts comparative R2's for income and dividends. Results generallysupport the hypothesis that common-law accounting income is more timelyrelative to dividends than code-law. In the last column of panel A, the ratio ofthe returns/income R2 to the returns/dividends R2 exceeds unity for three of thefour common-law countries. For the pooled common-law countries, the ratio is170%. In contrast, dividends are more timely than income in two of the threecode-law countries (Germany and Japan). In Germany, where there are parti-cularly binding institutional links between them, accounting income capturesonly 56% as much current value relevant information as dividends. The equiva-lent ratio for Japan is 78%. For the pooled code-law group, the ratio is 96%,meaning current-year income and dividends capture approximately the same


22The Australian imputation system is described in Hamson and Ziegler (1990). The Canadiansystem is dividend credits, not full imputation; and the marginal Canadian investor could pay UStax and thus not receive credits.

23We expect this to be reinforced in US by stock repurchases to distribute cash to investors.Repurchases contain information about future income (Dann et al., 1991), likely reducing informa-tion in dividends.

amount of current-year economic income. These results are consistent with thetimeliness of code-law accounting income being constrained by its direct role indetermining current-year dividend payouts.

Panel B of Table 5 reports separate dividend/return relations in positive andnegative return years, comparable to Table 3 for accounting income. Since theseare univariate regressions, we can report Vuong's (1989) likelihood ratio statisticfor selection among non-nested models in which the dependent variable is thesame, but the independent variables di!er. We calculate the Vuong statistic witha common dependent variable (returns) but competing independent variables(accounting income and dividends). The results from the Vuong test in panelB of Table 5 are weakly consistent with Hypothesis 4. For example, for the badnews or negative returns sample, the Vuong test statistic (distributed as unit-Normal, i.e., a Z-statistic) shows that common-law accounting income explainsreturns signi"cantly more than dividends, but that the opposite is true forcode-law countries. We also estimate, but do not report, a linear relation modelwith dividends levels and changes, comparable to panel B of Table 2. Theseresults also support Hypothesis 4.

5.1.2. Common-law tax imputation link between dividends and accounting incomeAmong common-law countries, there is considerable variation in the taxation

of dividend distributions. We focus on &imputation,' which penalizes corpora-tions for reporting taxable income in excess of distributed dividends. There is anincentive to structure transactions and use accounting standards that maketaxable income conform to dividend policy, which a!ects accounting incomebecause it is correlated with taxable income. Australia, Canada and the UKoperate imputation systems (Coopers & Lybrand, 1982}95), but not the US.22Provided dividends per se are uninformative, the testable implication is thataccounting income in the US exhibits greater timeliness relative to dividendsthan it does in Australia and Canada.23

H5: The diwerential timeliness of accounting income relative to dividends is greater

in the US than in common-law countries with dividend imputation (Australia,Canada, UK).

There are mixed results for this hypothesis as seen from the &ratio' reported inthe last column of panel A of Table 5. The results for Australia and UK reported


Table 6Contemporaneous association between cash#ows and returns!

Statistics are from regressions using the pooled cross-section and time-series of "rm/year observa-tions for each country. Intercepts are not reported.

Panel A: OCF"b0#b

1RD#b

2R#b

3RHRD#e

b2

t(b2) b

3t(b

3) Adj. R2(%) N Ratio

Australia 0.07 2.74 0.16 2.11 2.4 1,321 3.80Canada 0.02 1.07 0.25 5.28 3.6 2,901 4.75USA 0.04 7.81 0.10 7.39 3.1 21,225 4.78UK 0.11 11.40 !0.02 !0.87 3.7 5,758 3.69France 0.19 4.78 !0.16 !1.39 3.3 1,054 3.79Germany 0.16 4.28 !0.04 !0.37 2.6 1,245 2.10Japan 0.01 1.17 0.08 7.18 1.7 6,855 2.49

Common 0.04 8.04 0.12 9.19 3.0 25,447 4.79UK 0.11 11.40 !0.02 !0.87 3.7 5,758 3.69Code 0.08 8.92 0.00 !0.04 2.2 9,154 2.34

are consistent with the hypothesis (the ratios are less than that for USA), butthose for Canada are not. One conjecture is that the marginal investor inCanada is a US resident and does not receive imputation credits, so that Canadae!ectively is in the same tax category as the US.

5.2. Accounting income and cash yows from operations

We propose that cash #ows are noisier than accounting income in re#ectingcontemporaneous value-relevant information (Dechow, 1994) in all countries.Provided there is no new information in managers' current "nancing andinvestment decisions, accruals are expected to make accounting income incor-porate economic income in a more timely basis than cash #ows. This logic holdsunder the accrual-accounting systems of all countries.

Table 6 and Fig. 5 report results for the contemporaneous association be-tween returns and operating cash #ow, for the same sample of "rm/years as forincome and dividends in Tables 2, 3 and 5. In all countries, the income R2's arehigher than those of operating cash #ow, the ratio ranging from 2.10 (Germany)to 4.78 (US). This result would be obtained in less than 1% of cases by chance,assuming independence across the seven countries. The Vuong statistic and theratio of explanatory powers of the income/return and cash #ow/return relationsreported in panel B indicate that the greater timeliness of accounting incomethan operating cash #ow is due largely to bad news years. These "ndingsgeneralize well-known US results on timeliness of accounting income comparedto cash #ow (Ball and Brown, 1968; Dechow, 1994; Basu, 1997), and contrastsharply with results reported above for dividends.


Tab

le6

(cont

inued

)

Pan

elB

:O

CF"

b 0#

b 1R#

e(R5

0)O

CF"

b 0#

b 1R#

e(R(

0)

b 1A

dj.R

2(%

)N

Rat

ioV

uong

b 1A

dj.R

2(%

)N

Rat

ioV

uong

Aus

tral

ia0.

070.

981

30.

250.

920.

231.

450

812

.88

!3.

38C

anad

a0.

020.

01,

688

384.

000.

420.

272.

81,

213

2.82

!6.

23U

SA

0.04

0.5

12,7

219.

34!

1.51

0.14

1.4

8,50

47.

31!

12.8

8U

K0.

113.

43,

612

0.55

0.18

0.09

0.7

2,14

613

.85

!6.

93Fra

nce

0.19

2.9

611

0.76

!2.

120.

03!

0.2

443

!47

.40

!1.

67G

erm

any

0.16

2.2

712

2.22

!0.

160.

130.

353

313

.96

!2.

18Ja

pan

0.01

0.0

3,14

110

.00

!2.

790.

092.

23,

714

2.07

2.61

Com

mon

0.04

0.4

15,2

228.

12!

0.13

0.16

1.6

10,2

256.

45!

14.6

2U

K0.

113.

43,

612

0.55

0.18

0.09

0.7

2,14

613

.85

!6.

93C

ode

0.08

1.3

4,46

41.

20!

3.10

0.08

0.5

4,69

05.

49!

2.18

!Sam

ple

cons

ists

of40

,359"rm

-yea

robs

erva

tions

sele

cted

from

the

Glo

bal

Van

tage

indust

rial

/Com

mer

cial

and

Issu

e"le

sove

r19

85}95

,us

ing

the

follo

win

gpro

cedu

re.F

irst

,fo

rea

chva

riab

le(see

bel

ow)w

eel

imin

ate

the

two

extr

eme

per

cent

iles

of"rm

-yea

robse

rvat

ions

.Sec

ond,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sw

ith

missing

valu

esfo

rone

orm

ore

variab

les,

tofa

cilit

ate

com

para

bilit

yw

ith

resu

lts

inpre

viou

sta

bles

.Thi

rd,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sfrom

count

ries

with

less

than

1,00

0ob

serv

atio

ns,l

eavi

ngse

ven

count

ries

repre

sent

ed.A

ustr

alia

,Can

ada,

Uni

ted

Stat

es,a

ndth

eU

nited

Kin

gdom

are

the

com

mon

-law

coun

trie

s,th

ere

star

eco

de-

law

coun

trie

s.H

ow

ever

,the

com

mon

law

coun

try

cate

gory

inth

eth

ird

row

from

the

bott

om

excl

udes

the

UK

.R"

buy-

and-h

old

secu

rity

retu

rnin

clusive

ofdiv

iden

dsove

rth

e"sc

alye

ar;

RD"

1if

R4

0an

dR

D"

0if

R'

0;O

CF"

annual

ope

rating

cash#ow

per

shar

e,w

her

eoper

atin

gca

sh#ow

isde"

ned

asea

rnin

gsplu

sdec

reas

ein

non-

cash

curr

entas

sets

plu

sin

crea

sein

non-d

ebtcu

rren

tlia

bilit

ies

plus

depr

ecia

tion

;N"

the

num

ber

of"rm

/yea

robs

erva

tions

.R

atio"

inpa

nel

A,t

hera

tio

oft

heR

2'soft

he

equi

vale

ntea

rnin

gs(T

able

2)an

dop

erat

ing

cash#ow

(thi

sta

ble

)reg

ress

ions,

inpan

elB

,the

ratio

oft

he

R2's

ofth

eeq

uiva

lent

earn

ings

(Tab

le3)

and

ope

rating

cash#ow

(thi

sta

ble)

regr

essions.

Vuo

ng

stat

istic

isa

like

lihood

ratio

stat

istic

com

paring

the"tof

the

retu

rn/o

per

atin

gca

sh#ow

model

inth

ista

ble

with

the

retu

rn/e

arni

ngs

mode

lin

Tab

le3,

both

estim

ated

sepa

rate

lyfrom

the

good

new

san

dbad

new

sobs

erva

tion

sin

pane

lB

.


Fig. 5. Earnings versus cash#ows R-squares from the individual country reverse regressions of (1)NI and (2) OCF against (a) annual return and (b) annual return times negative return dummy.

The incremental timeliness of accounting income (relative to cash #ows)measures the extent to which accruals under a particular country's accountingrules are oriented toward timely incorporation of economic income. We hy-pothesize that common-law accruals re#ect a greater demand for timely ac-counting income, and hence predict that common-law accounting income hasmore incremental timeliness (relative to cash #ows) than code-law accountingincome.

H6: Common-law accounting income is more timely relative to operating cashyows.

Results generally support Hypothesis 6. In panel A of Table 6, three of thetop four ratios of return/income R2 to return/cash #ow R2 are common-lawcountries. Accounting income in Germany and Japan incorporates the leastamount of current-period economic income, relative to cash #ows. Thus, thegreater timeliness of common-law income is not due entirely to underlyingtransactions.


6. Speci5cation tests: income and returns

The relation between accounting income and stock returns is robust withrespect to a variety of speci"cation tests, reported in Tables 7 and 8.

6.1. Variation in expected returns

When economic income is the independent variable, a possible misspeci"ca-tion arises from variation in expected returns (across time, countries and "rms).Assume the true model for accounting income incorporates only the componentof *<

tdue to information about future cash #ows a, and not due to variation in

expected returns r. Assume these components are independent, both of eachother and over time. Our model (2) then can be simpli"ed as

NIt"h(a

t, f

t) (6)

where the disturbance term ft

now incorporates expected return e!ects, inaddition to lagged information about cash #ows and accounting noise. Using*<

tas a proxy for a

tthen would introduce measurement error in the indepen-

dent variable, due to variation in expected returns.Table 7 reports pooled regressions with two alternative controls for market-

wide expected return e!ects. In column (2) the independent variable, annualreturn, is de"ned relative to the mean return for the "rm's country/year. Incolumn (3) the dependent variable, accounting income, is scaled by the coun-try/year long-term interest rate. In both cases, results are similar to those incolumn (1), which repeats the results from Table 4 for comparison. We concludemarket-wide e!ects do not substantially in#uence our results.

6.2. Consolidated versus parent income and control for special items

In countries where parent companies are not required to &equity account' theirshare of a$liated-company income, their accounting income omits a componentthat accrues to their stockholders. This could be viewed as introducing measure-ment error that likely is positively correlated with parent-company income,thereby biasing the regression slopes downward. A counter-argument is that weare interested in properties of accounting income as reported under di!erentcountries' accounting rules, including any omission of a$liate income. Never-theless, we replicate column (1) results for the 33,441 total "rm/years for whichPeriodic Descriptor Array No. 13 on Global Vantage speci"cally labels incomeas &fully consolidated,' thus including parent companies' proportional shares ofa$liates' incomes. The sample size reduction is due almost entirely to Japan,


Tab

le7

Com

par

ativ

eas

ymm

etry

inth

eco

nte

mpor

aneo

usre

turn

s-ea

rnin

gsre

lation

:al

tern

ativ

esp

eci"

cations

!

Rep

orte

dst

atistics

are

for

the

follo

win

gm

odel

using

the

poo

led

cros

s-se

ctio

nan

dtim

e-se

ries

of"rm

/yea

rob

serv

atio

ns

for

allco

untr

ies:

NI"

b 0#

+ j

b 0jC

Dj#

b 1R

D#

+ j

b 1jR

DC

Dj#

b 2R#

+ j

b 2jR

CD

j#b 3

RR

D#

+ j

b 3jR

RD

CD

j

Res

ults

are

not

repo

rted

for

the

inte

rcep

t,th

ene

gative

retu

rnin

terc

ept,

and

thei

rre

spec

tive

count

rydum

mie

s.The

countr

yca

tego

rym

odel

sus

eth

eco

mm

on

law

count

ries

ofA

ustr

alia

,Can

ada

and

the

US

asth

ebas

eca

tego

ry;d

um

mie

sar

euse

dfo

r(1

)th

eU

Kan

d(2

)the

code

law

count

ries

ofF

rance

,G

erm

any

and

Japan

.

Mode

l(1

)(2

)(3

)(4

)(5

)>

NI

NI

NI/

RF

NI2

yrN

IC

ontr

ol

Rjt

SIC

Coe!.

t-Sta

t.C

oe!

.t-Sta

t.C

oe!

.t-Sta

t.C

oe!

.t-Sta

t.C

oe!

.t-Sta

t.

Pan

elA

:co

untr

ydu

mm

ies

mod

el

Ear

ning

s&g

ood

new

s+se

nsit

ivit

y

b 2(R

etur

n)0.

039.

880.

013.

960.

369.

88!

0.05

!84

.24

0.04

10.2

5b 2j

(Ret

urn *

Cnt

ryD

ums):

Aust

ralia

!0.

03!

3.95

!0.

04!

3.17

!0.

48!

4.24

!0.

02!

3.02

!0.

04!

4.15

Can

ada

!0.

03!

3.80

!0.

01!

1.63

!0.

32!

3.54

0.06

10.0

3!

0.02

!3.

28U

K0.

022.

400.

033.

740.

141.

570.

092.

920.

011.

57Fra

nce

0.05

4.16

0.06

3.48

0.65

3.86

0.55

21.2

80.

053.

69G

erm

any

0.03

1.57

0.04

1.81

0.55

2.54

0.21

19.3

20.

021.

42Ja

pan

!0.

02!

2.64

0.00

!0.

51!

0.17

!1.

690.

076.

82!

0.02

!3.

12

F-s

tat

for

Cou

ntry

11.3

57.

7310

.28

164.

139.

75R

etur

nD

umm

ies

p(0.

01p(

0.01

p(0.

01p(

0.01

p(0.

01


Incr

emen

tal&b

adne

ws+

sens

itiv

ity

b 3(R

dum

*Ret

urn)

0.29

39.2

10.

2947

.25

3.66

38.1

10.

4326

.13

0.29

31.0

4b 3j

(Rdu

m*R

et*C

ntry

Dum

s):

Aust

ralia

0.07

2.64

0.01

0.37

0.19

0.50

0.12

1.85

0.08

2.89

Can

ada

0.11

5.28

0.05

2.92

0.45

1.70

0.12

2.43

0.11

5.25

UK

!0.

14!

8.26

!0.

17!

11.1

1!

2.05

!9.

10!

0.24

!2.

47!

0.14

!7.

96Fra

nce

!0.

22!

5.81

!0.

14!

4.40

!2.

87!

5.99

!0.

68!

7.40

!0.

21!

5.54

Ger

man

y!

0.19

!4.

41!

0.24

!5.

74!

2.49

!4.

55!

0.31

!7.

48!

0.20

!4.

73Ja

pan

!0.

29!

15.7

4!

0.29

!16

.80

!3.

43!

14.9

0!

0.42

!10

.70

!0.

28!

15.1

6

F-s

tat:

Cou

ntry

67.4

472

.74

54.5

537

.81

63.4

3N

egat

ive

Ret

Dum

sp(

0.01

p(0.

01p(

0.01

p(0.

01p(

0.01

Reg

ress

ion

N40

,359

40,3

5939

,240

33,0

8240

,359

Adj

.R2

15.4

%16

.8%

13.8

%22

.5%

16.2

%F

272.

9830

3.32

234.

1941

0.56

117.

16

Pan

elB

:co

untr

yca

tego

rydu

mm

ies

mod

el

Ear

ning

s&g

ood

new

s+se

nsit

ivit

y

b 2(R

etur

n)0.

028.

360.

012.

820.

268.

22!

0.05

!83

.70

0.03

8.99

b 2j(R

etur

n *C

ntry

Dum

s):

UK

0.02

3.52

0.04

4.35

0.23

2.75

0.20

19.1

20.

022.

58C

ode

0.01

2.37

0.02

2.87

0.24

2.99

0.15

15.2

60.

011.

53

F-s

tat

for

Cnt

ryC

at7.

9412

.24

7.20

297.

973.

98R

etur

nD

ums

p(0.

01p(

0.01

p(0.

01p(

0.01

p(0.

01


Tab

le7

(cont

inued

)

Incr

emen

tal&b

adne

ws+

sens

itiv

ity

Mode

l(1

)(2

)(3

)(4

)(5

)>

NI

NI

NI/

RF

NI2

yrN

IC

ontr

ol

Rjt

SIC

Coe!.

t-Sta

t.C

oe!

.t-Sta

t.C

oe!

.t-Sta

t.C

oe!

.t-Sta

t.C

oe!

.t-Sta

t.

b 3(R

Dum

*Ret

urn)

0.31

46.2

30.

3053

.64

3.76

43.2

80.

4630

.27

0.31

36.6

8b 3j

(RD

um*R

et*

Cnt

ryC

atD

ums):

UK

!0.

16!

9.50

!0.

17!

11.7

2!

2.14

!9.

69!

0.34

!8.

14!

0.16

!9.

25C

ode

!0.

30!

19.0

0!

0.28

!18

.82

!3.

47!

17.4

1!

0.48

!13

.65

!0.

29!

18.3

1

F-s

tat:

Cnt

ryC

at20

0.92

218.

6617

4.82

111.

5618

7.17

Neg

ativ

eR

etD

ums

p(0.

01p(

0.01

p(0.

01p(

0.01

p(0.

01

Reg

ress

ion

N40

,359

40,3

5939

,240

33,0

8240

,359

Adj

.R2

14.9

%16

.2%

13.5

%21

.3%

15.7

%F

642.

0771

1.90

557.

8493

8.83

148.

07

!Sam

ple

cons

ists

of40

,359"rm

-yea

robs

erva

tions

sele

cted

from

the

Glo

bal

Van

tage

indust

rial

/Com

mer

cial

and

Issu

e"le

sove

r19

85}95

,us

ing

the

follo

win

gpro

cedu

re.F

irst

,fo

rea

chva

riab

le(see

bel

ow)w

eel

imin

ate

the

two

extr

eme

per

cent

iles

of"rm

-yea

robse

rvat

ions

.Sec

ond,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sw

ith

missing

valu

esfo

rone

orm

ore

variab

les,

tofa

cilit

ate

com

para

bilit

yw

ith

resu

lts

inpre

viou

sta

bles

.Thi

rd,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sfrom

count

ries

with

less

than

1,00

0ob

serv

atio

ns,l

eavi

ngse

ven

count

ries

repre

sent

ed.A

ustr

alia

,Can

ada,

Uni

ted

Stat

es,a

ndth

eU

nited

Kin

gdom

are

the

com

mon-

law

countr

ies,

there

star

eco

de-law

coun

trie

s.The

sam

ple

size

sfo

rm

odel

(3)t

hat

uses

NI/

RF

asth

edep

enden

tva

riab

lean

dm

odel

(4)th

atuse

sonly

the

two-

year

NI

valu

esar

e39

,240

and

33,0

82re

spec

tive

ly.

R"

buy-

and-h

old

secu

rity

retu

rnin

clusive

ofdiv

iden

ds

ove

rth

e"sc

alye

ar;

NI"

annua

lea

rnin

gsper

shar

ebe

fore

extr

aord

inar

yitem

sde#

ated

by

beg

innin

gof

period

price

;N

I/R

F"

earn

ings

de#

ated

by

the

long-

term

inte

rest

rate

;N

I2yr

"th

esu

mof

net

inco

me

over

the

two

year

stan

dt#

1;R

D"

the

pro

xyfo

rbad

new

s"1

ifR(

0an

d"

0ot

her

wise;

CD

j"co

unt

ryid

enti"er"

1fo

r"rm

/yea

rsin

coun

try

jan

d"

0oth

erw

ise.

USA

isth

e&b

ase

count

ry'w

ith

CD

j"0

∀j.

Inm

ode

l(2)

the

regr

ession

isco

ntr

olle

dfo

rm

ean

annu

alco

untr

yre

turn

by

scal

ing

NIan

dby

rede"

ning

RD

;in

mod

el(5

)the

regr

ession

isco

ntro

lled

for

the

10m

ostpre

vale

nt

2-di

git

SIC

code

by

crea

ting

the

appro

priat

edum

my

variab

les

bas

edon

thes

eSI

Cco

des

.In

tere

stra

tes:

RF

ism

easu

red

asth

elo

nge

stte

rmbond

yiel

dav

aila

ble

inth

eIM

F's

Inte

rnat

iona

lF

inan

cial

Stat

istics

for

the

cale

ndar

year

most

coin

ciden

tw

ith

the"rm's"sc

alye

ar[!

11,0

].R

ates

for

the

indiv

idual

countr

ies

are

for:

trea

sury

bonds

2ye

ars

(Aus

tral

ia);

govt

.bon

dyi

eld'

10yr

(Can

ada)

;gov

t.bond

yiel

d(G

erm

any)

;gov

t.bond

yiel

d,m

oyen

s(F

ranc

e);l

ong

term

govt

.bond

yiel

d(U

K);

govt

.bon

dyi

eld

(Jap

an);

govt

.bond

yiel

d10

yr(U

SA).


24Excluding these observations is not clearly necessary because Japanese rules require equityaccounting in the parent's books (Nobes and Parker, 1995, Chapter 13). Few observations areexcluded for Germany where public companies have issued consolidated income since the EuropeanUnion's Seventh Directive.

25The exception is the positive-return slopes for the US and Australia. The sample size decreases,due mainly to losing one year. Results are for overlapping samples.

which loses most of its observations.24 Results (unreported) are almost identicalexcept for Japan (due to the small sample size).

The reported results in the paper are based on using accounting incomebefore extraordinary items. Special items, which are included in calculatingaccounting income, have characteristics similar to extraordinary items, in thatthey tend to be negative and transitory (e.g., Collins et al., 1997). We repeat theanalysis using accounting income net of special items and obtain qualitativelysimilar results.

6.3. Extending the lag in accounting income

Column (4) of Table 7 reports results with the dependent variableNI2

yr"NI

t#NI

t`1, allowing an additional year to incorporate economic

income in accounting income. As expected, the coe$cients generally increase:the median two-year coe$cient is 1.6 times its one-year equivalent.25 Ac-counting income thus incorporates economic income over time (it is &smoothed,'and has &momentum' or &persistence').

6.4. Control for industry composition

Another concern is correlated omitted variables. For example, if the propor-tion of growth options relative to assets in place varies across countries, thencommon application of the revenue realization rule will cause accounting incometo vary internationally in timeliness, due to &real' rather than &accounting' e!ects.This concern is reduced by our result that international di!erences in timelinessare due in part to accounting accruals and not entirely to di!erences intimeliness of cash #ows. A related concern is that our sample contains only listedcorporations. Public corporations are less prevalent under code-law systemsprobably because the latter do not facilitate public disclosure and public capitalmarkets (see La Porta et al., 1997) to the same extent as common-law systems.For example, UK has approximately four times as many listed corporations asGermany. To alleviate concern that our results are due to di!erent samplecomposition across countries, column (5) reports results after controlling forindustry e!ects, in the form of interactive dummy variables for the ten most


Tab

le8

Cont

empor

aneo

us

asso

ciat

ion

bet

wee

nea

rnin

gs,div

iden

ds,

cash#ow

san

dre

turn

ssu

b-p

erio

dan

alys

is!

Ana

lysi

s:Sta

tist

ics

are

from

regr

essions

using

the

pool

edcr

oss

-sec

tion

and

tim

e-se

ries

of"rm

/yea

robse

rvat

ions

for

each

countr

y.In

terc

epts

are

not

repor

ted.

1985}90

1991}95

b 2t(b 2

)b 3

t(b 3

)A

dj.R

2(%

)N

b 2t(b 2

)b 3

t(b 3

)A

dj.R

2(%

)N

Pan

elA

:N

I"b 0

#b 1

RD#

b 2R#

b 3R

*RD#

e

Aust

ralia

0.02

1.04

0.26

5.31

13.0

500

!0.

02!

0.99

0.48

7.55

9.4

1,82

1C

anad

a0.

00!

0.23

0.38

12.9

419

.81,

594

0.01

1.30

0.44

11.6

315

.81,

307

USA

0.03

6.14

0.28

26.7

917

.311

,978

0.03

6.54

0.33

22.1

412

.99,

247

UK

0.04

7.49

0.09

6.92

13.7

2,68

20.

069.

020.

1911

.64

17.4

3,07

6Fra

nce

0.09

5.96

0.04

1.09

19.1

466

0.06

3.15

0.22

4.24

10.5

588

Ger

man

y0.

053.

860.

040.

976.

148

80.

041.

950.

194.

085.

375

7Ja

pan

0.01

5.21

0.02

4.96

7.0

3,46

00.

000.

160.

011.

561.

03,

395

Com

mon

0.02

5.64

0.29

30.1

217

.314

,071

0.02

5.32

0.36

26.2

212

.711

,374

UK

0.04

7.49

0.09

6.92

13.7

2,68

20.

069.

020.

1911

.64

17.4

3,07

6C

ode

0.03

10.9

70.

000.

425.

34,

413

0.05

8.05

0.01

1.38

4.4

4,73

9

Pan

elB

:D

I<"

b 0#

b 1R

D#

b 2R#

b 3R

*RD#

e

Aust

ralia

0.00

!0.

360.

065.

0914

.450

0!

0.01

!1.

890.

077.

3010

.382

1C

anad

a!

0.01

!5.

370.

048.

127.

41,

594

!0.

01!

6.60

0.05

7.69

7.0

1,30

7U

SA

!0.

02!

18.0

20.

0525

.51

9.8

11,9

78!

0.01

!19

.45

0.05

23.0

98.

99,

247

UK

0.00

2.40

0.03

7.80

8.1

2,68

20.

018.

640.

025.

9911

.33,

076

Fra

nce

0.01

4.03

0.02

2.02

13.7

466

0.01

1.42

0.04

4.24

6.7

588

Ger

man

y0.

025.

400.

011.

0010

.748

80.

013.

690.

011.

348.

275

7Ja

pan

0.00

6.61

0.00

5.16

9.2

3,46

00.

00!

0.61

0.01

7.81

7.2

3,39

5C

omm

on

!0.

01!

17.2

50.

0526

.29

9.2

14,0

71!

0.01

!18

.40

0.05

24.3

38.

011

,374

UK

0.00

2.40

0.03

7.80

8.1

2,68

20.

018.

640.

025.

9911

.33,

076

Cod

e0.

019.

310.

00!

0.51

4.2

4,41

30.

019.

990.

001.

889.

74,

739


Pan

elC

OC

F"

b 0#

b 1R

D#

b 2R#

b 3R

*RD#

e

Aust

ralia

0.01

0.21

0.14

1.31

1.0

500

0.09

3.02

0.23

2.07

3.3

821

Can

ada

!0.

02!

0.73

0.29

4.62

3.4

1,59

40.

042.

070.

223.

033.

91,

307

USA

0.03

3.82

0.09

5.13

2.5

11,9

780.

057.

150.

135.

993.

89,

247

UK

0.11

7.24

!0.

02!

0.55

3.1

2,68

20.

129.

05!

0.03

!0.

784.

53,

076

Fra

nce

0.20

3.42

!0.

15!

0.96

2.6

466

0.23

3.95

!0.

17!

1.00

4.3

588

Ger

man

y0.

122.

400.

120.

872.

448

80.

254.

08!

0.19

!1.

403.

075

7Ja

pan

0.03

4.99

0.03

1.54

1.9

3,46

0!

0.01

!0.

750.

126.

723.

83,

395

Com

mon

0.02

2.87

0.12

6.88

2.5

14,0

710.

058.

130.

156.

983.

711

,374

UK

0.11

7.24

!0.

02!

0.55

3.1

2,68

20.

129.

05!

0.03

!0.

784.

53,

076

Cod

e0.

087.

30!

0.07

!2.

341.

64,

413

0.21

11.4

9!

0.08

!2.

396.

34,

739

!Sam

ple

cons

ists

of40

,359"rm

-yea

robs

erva

tions

sele

cted

from

the

Glo

bal

Van

tage

indust

rial

/Com

mer

cial

and

Issu

e"le

sove

r19

85}95

,us

ing

the

follo

win

gpro

cedu

re.F

irst

,fo

rea

chva

riab

le(see

bel

ow)w

eel

imin

ate

the

two

extr

eme

per

cent

iles

of"rm

-yea

robse

rvat

ions

.Sec

ond,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sw

ith

missing

valu

esfo

rone

orm

ore

variab

les,

tofa

cilit

ate

com

para

bilit

yw

ith

resu

lts

inpre

viou

sta

bles

.Thi

rd,w

eel

imin

ate

all

"rm

-yea

robs

erva

tion

sfrom

count

ries

with

less

than

1,00

0ob

serv

atio

ns,l

eavi

ngse

ven

count

ries

repre

sent

ed.A

ustr

alia

,Can

ada,

Uni

ted

Stat

es,a

ndth

eU

nited

Kin

gdom

are

the

com

mon

-law

coun

trie

s,th

ere

star

eco

de-

law

coun

trie

s.H

ow

ever

,the

com

mon

law

coun

try

cate

gory

inth

eth

ird

row

from

the

bott

omin

each

panel

excl

udes

the

UK

.R"

buy-

and-h

old

secu

rity

retu

rnin

clusive

ofdiv

iden

dsove

rth

e"sc

alye

ar;

RD"

1if

R4

0an

dR

D"

0if

R'

0;N

I"an

nual

earn

ings

per

shar

ebe

fore

extr

aord

inar

yitem

sde#

ated

bybe

ginnin

gofper

iod

price

;D

IV"

annua

ldiv

iden

ds

per

shar

ede#

ated

bybeg

innin

gof

period

price

;O

CF"

annu

alop

erat

ing

cash#ow

per

shar

ede#

ated

by

beg

innin

gofpe

riod

price

,whe

reoper

atin

gca

sh#ow

isde"

ned

asea

rnin

gsplu

sdec

reas

ein

non-c

ash

curr

ent

asse

tspl

us

incr

ease

inno

n-deb

tcu

rren

tlia

bilit

ies

plus

dep

reci

atio

n;N"

the

num

ber

of"rm

/yea

robs

erva

tions

.


prevalent 2-digit SIC codes. Little change is apparent. The regression adjustedR2 rises slightly and the F-statistic falls. The F-statistics for the countries'dummy slopes fall slightly. All remain signi"cant at the 0.01 level. There is littlechange in the coe$cients for both individual countries and country groups.

6.5. Subperiod results

Splitting the sample into two subperiods, 1985}90 and 1991}95, reveals threeinteresting results, reported in Table 8. First, the incremental coe$cients onnegative economic income increase in the second subperiod for six of the sevenindividual-country regressions (4), the exception being Japan. Several countriesexhibit statistically signi"cant increases, including Australia (0.48 versus 0.26),UK (0.19 versus 0.09), France (0.22 versus 0.04) and Germany (0.19 versus 0.05).The coe$cient on negative economic income for US increases, but not signi"-cantly (0.33 versus 0.28). These results suggest Basu (1997) might have erred inattributing increased asymmetry over time in US accounting income to changesin US litigation rules, since similar increases have occurred in France andGermany. An intriguing possibility is that timely incorporation of economiclosses has become a more important corporate governance mechanism overtime worldwide (if not in Japan), and that pressure to adopt it has come fromincreased international product market competition (if not from changes in thepolicies of standard-setters around the world). Under this explanation, timelyincorporation of economic losses in accounting income is an e$cient gover-nance mechanism that reduces managers' incentives to continue with loss-making investments and strategies, thereby reducing agency-related negativeNPVs. We believe this phenomenon is worthy of further study.

Second, despite increased sensitivity to economic losses, the R2's in six ofseven country regressions (4) decreased between subperiods, generalizing the USresult (Ramesh and Thiagarajan, 1995; Lev, 1997). Third, the results for cash#ow from operations, reported in panel C, contrast unexpectedly with those foraccounting income. Between subperiods, the point estimates of the slopes onpositive returns increased for six of seven countries, the incremental slopes onnegative returns show no systematic change, and the R2's in all of the indi-vidual-country cash #ow regressions increase. We have no explanation for theseapparently systematic cash #ow changes.

7. Conclusions, implications, and limitations

The worldwide trend toward &internationalization' of markets, especiallycapital markets, in which accounting information is used has rekindled aca-demic and professional interest in di!erent national accounting models. Theproperties of accounting information prepared under common-law accounting


standards are of particular contemporary interest because the InternationalAccounting Standards Committee (IASC) recently completed a set of &interna-tional' accounting standards widely viewed as re#ecting a largely common-lawapproach of &transparent,' timely disclosure.

We show that common-law accounting income does indeed exhibit signi"-cantly greater timeliness than code-law accounting income, but that this is dueentirely to greater sensitivity to economic losses (income conservatism). Thisresult has important implications for corporate governance. We conjecture thatearly incorporation of economic losses, as distinct from gradual incorporationover time, increases managers' incentives to attend to the sources of losses morequickly. It brings more and quicker pressure from security analysts, makesleverage and dividend restrictions binding more quickly, and a!ects currentmanagers' and employees' bonuses. It also makes excessively optimistic state-ments by managers less credible. Conservative accounting thus facilitatesmonitoring of managers and is an important feature of common-law corporategovernance. In contrast with Roe (1994), we conclude that enhanced common-law disclosure standards reduce the agency costs of monitoring managers, thuscountering the advantages of closer shareholder-manager contact in code-lawcountries.

Our results suggest an explanation for the emergence of a largely common-law model in international transacting, and in particular for the IASC adoptionof a more &transparent' common-law approach to disclosure. The cost ofcross-border transacting by parties who are geographically, culturally andlinguistically separated from the "rm's management presumably is greater ina system that assumes they are informed insiders than in one with more timelypublic disclosure. Timely disclosure of economic losses is particularly importantto cross-border lenders. Whether this model will prevail in international trans-acting depends on whether high disclosure-quality "rms (i.e., "rms knowna priori as likely to recognize economic losses in a timely fashion) can signaltheir quality e!ectively to users, or equivalently on whether low-quality "rmscan be excluded from false signaling. In the absence of common-law penalties tofalse signaling, it is di$cult to see how high-quality "rms in code-law countrieswill be able to reduce contracting costs through improved "nancial reporting,other than by listing in a common-law jurisdiction and exposing themselves tocommon-law penalties for low-quality disclosure.

Greater common-law income conservatism should be no surprise, consideringthe use of accounting income in common-law arm's-length debt and equitymarkets, and especially considering common-law litigation. Nevertheless, Ger-man accounting in particular is widely presumed to be more conservative,because German managers have unusual discretion to reduce reported incomeduring good years. However, they also have unusual discretion to delay recogni-tion of economic losses, and thus to increase reported income in bad years. Thiswas a common UK practice at the beginning of the twentieth century (Yamey,


26 Rex v. Lord Kylsant 1932 1 KB 442. This case-law was codi"ed in the 1948 UK Companies Act,which required companies to distinguish reserves from provisions, &making the creation of secretreserves more di$cult.' (Nobes and Parker, 1995, p. 103). The case can lay claim to be the legalgenesis of the Rule 10b-5 in the US. US German practice is exempli"ed by the notorious Daimler-Benz case (Ball, 1998).

1962), but e!ectively was extinguished in common-law countries by the RoyalMail case.26 Similar observations can be made about accounting in Japan.

Our research design is subject to several limitations. The most obviousconcern is the validity of stock returns as a proxy for economic income,particularly in code-law countries, which have endogenously lower liquidity andpublic disclosure standards. &Noise' in annual stock returns as a measure ofmarket-valued income could be a correlated omitted variable. We counter withthree observations. First, poor public disclosure does not necessarily impede the#ow of information into stock prices, since the information #ow can occurinstead via the trading of informed insiders. In the absence of e!ectively enforcedinsider-trading laws, which in many ways are fundamentally incompatible withcode-law governance, corporate insiders' incentives are to trade on informationand thereby incorporate it into prices. This, Ramseyer (1993) conjectures,explains the practice of Japanese banks trading in their clients' stock. To somedegree, equating poor public disclosure with uninformed stock prices involvesprojecting common-law precepts onto code-law institutions, which are morelikely to solve information asymmetry by private rather than public commun-ication. Second, our results seem inconsistent with the hypothesis that stockreturns re#ect poor information #ows in code law countries, which would implyless anticipation of accounting income, a stronger income &surprise' e!ect, anda stronger return-income association. We observe the opposite. Third, we studystock returns over intervals of one and two years, not over short &event win-dows,' which is long enough for international di!erences in liquidity, marketmicrostructure and disclosure timing e!ects to have minimal impact.

Correlated omitted variables are another obvious concern. For example, if theproportion of growth options relative to assets in place varies across countries,then common application of the revenue realization rule will cause income tovary internationally in timeliness, due to &real' rather than &accounting' e!ects.Similarly, our results hold only for listed corporations. The number of listedcode-law corporations is endogenously small compared with common-lawcountries, in part because code-law systems are not designed for the demands ofpublic disclosure, so private corporations are comparatively more e$cient andconsequently more prevalent. Thus, our sample is less representative of code-lawaccounting in general. This concern is addressed to a degree in the Table7 control for SIC codes. Another concern is whether the sample period isrepresentative. Japan, for example, went through boom and bust during


1985}95. Speci"cation tests in Table 7 address this concern, where the resultschange little when annual "rm return is de"ned relative to the mean return forthat country/year, and when income is scaled by the country/year interest rate.

Another concern is our characterization of the relevant institutional featuresof the seven countries in our primary sample. In particular, the code/commonlaw categorization is a proxy for an underlying economic construct, the extent towhich accounting is determined by market supply and demand relative topolitical forces. The categories overlap, as is obvious from the roles of theCompanies Acts in the UK and the SEC in the US. To some degree, this concernis alleviated by the consistent evidence we report from our primary sample. Inaddition, we report consistent evidence from a secondary sample of eighteencountries, suggesting that our results are generalizable.

A "nal concern is that institutional determinants of "nancial reporting varyover time. For example, shareholder lawsuits are reported as rising in Japan(Wall Street Journal, January 7, 2000, p. A13), which could signal a change ingovernance or incentives of "nancial statement preparers. While we have mech-anically divided the 1985}95 period into approximately equal sub-periods andhave thereby observed a systematic increase in conservatism, a "ner partitioningbased on changes in institutional determinants of accounting might obtainclearer results.

The research design also has its strengths. It studies the incorporation ofeconomic income into accounting income over time, under di!erent interna-tional institutions, using a model of accounting income determination that isbased on fundamental properties of accounting. We believe the results to be ofinterest to accountants, analysts, standard-setters, regulators and students ofcorporate governance.

References

Alchian, A., 1984. Speci"city, specialization, and coalitions. Journal of Institutional and TheoreticalEconomics 140, 34}49.

Alchian, A., Demsetz, H., 1972. Production, information costs and economic organization. AmericanEconomic Review 62, 777}795.

Alford, A., Jones, J., Leftwich, R., Zmijewski, M., 1993. The relative informativeness of accountingdisclosures in di!erent countries. Journal of Accounting Research 31 (Suppl.), 183}223.

American Accounting Association, 1977. Report of the Committee on International Accounting.Supplement to Accounting Review 52.

Amir, E., Harris, T., Venuti, E., 1993. A comparison of the value-relevance of U.S. versus non-U.S.GAAP accounting measures using form 20-F reconciliations. Journal of Accounting Research 31(Suppl.), 230}264.

Ball, R., 1998. Daimler-Benz AG: evolution of corporate governance from a code-law &stakeholder'to a common-law &shareholder value' system. Unpublished manuscript, University of Rochester.

Ball, R., Brown, P., 1968. An empirical evaluation of accounting income numbers. Journal ofAccounting Research 6, 159}178.


Bandyopadhyay, S., Hanna, D., Richardson, G., 1994. Capital market e!ects of U.S.-Canada GAAPdi!erences. Journal of Accounting Research 32, 262}277.

Barth, M., Clinch, G., 1996. International accounting di!erences and their relation to share prices:evidence from U.K., Australian, and Canadian "rms. Contemporary Accounting Research 13,135}170.

Basu, S., 1997. The conservatism principle and the asymmetric timeliness of earnings. Journal ofAccounting and Economics 24, 3}37.

Baums, T., Buxbaum, R., Hopt, K. (Eds.), 1994. Institutional Investors and Corporate GovernanceWalter de Gruyter, Berlin.

Beaver, W., Lambert, R., Morse, D., 1980. The information content of security prices. Journal ofAccounting and Economics 2, 3}28.

Choi, F., Mueller, G., 1992. International Accounting. Prentice-Hall, Englewood Cli!s, NJ.Coopers & Lybrand, 1982}95. International Tax Summaries, Annual Editions. Wiley, New York.Collins, D., Maydew, E., Weiss, I., 1997. Changes in the value-relevance of earnings and book values

over the past forty years. Journal of Accounting and Economics 24, 39}67.Cramer, J., 1987. Mean and variance of R2 in small and moderate samples. Journal of Econometrics

35, 253}166.Dann, L., Masulis, R., Mayers, D., 1991. Repurchase tender o!ers and earnings information. Journal

of Accounting and Economics 14, 217}251.David, R., Brierley, J., 1985. Major Legal Systems in the World Today. Stevens & Sons, London.Dechow, P., 1994. Accounting earnings and cash #ows as measures of "rm performance: the role of

accounting accruals. Journal of Accounting and Economics 18, 3}42.Dechow, P., Kothari, S., Watts, R., 1998. The relation between earnings and cash #ows. Journal of

Accounting and Economics 25, 133}168.Easton, P., Harris, T., Ohlson, J., 1992. Aggregate accounting earnings can explain most of security

returns. Journal of Accounting and Economics 15, 119}142.Elliott, J., Hanna, J., 1996. Repeated write-o!s and the information content of earnings. Journal of

Accounting Research 34, 135}155.Fama, E., MacBeth, J., 1973. Risk, return, and equilibrium: empirical tests. Journal of Political

Economy 81, 607}636.Financial Accounting Standards Board (FASB), 1985. Statement of "nancial accounting concepts

no. 6. FASB, Stamford, Connecticut.Gray, S., 1980. International accounting di!erences from a security analysis perspective. Journal of

Accounting Research 18, 64}76.Hamson, D., Ziegler, P., 1990. The impact of dividend imputation on "rms' "nancial decisions.

Accounting and Finance 30, 29}53.Harris, T., Lang, M., MoK ller, H., 1994. The value relevance of German accounting measures: an

empirical analysis. Journal of Accounting Research 32, 187}209.Healy, P., 1985. The e!ect of bonus schemes on accounting decisions. Journal of Accounting and

Economics 7, 85}107.Hicks, J.R., 1946. Value and Capital 2nd edition. Clarendon Press, Oxford.Hughes, J., Snyder, E., 1995. Litigation and settlement under the English and American rules: theory

and evidence. Journal of Law and Economics 38, 225}250.Jacobsen, R., Aaker, D., 1993. Myopic management behavior with e$cient, but imperfect, "nancial

markets. Journal of Accounting and Economics 16, 383}405.Joos, P., Lang, M., 1994. The e!ects of accounting diversity: evidence from the European Union.

Journal of Accounting Research 32 (Suppl.), 141}168.Katz, A., 1987. Measuring the demand for litigation. Journal of Law, Economics, & Organization 3,

143}176.KoK ndgen, J., 1994. Duties of banks in voting their clients' stock. In: Baums, Buxbaum, Hopt (Eds.),

(Chapter 18). Institutional Investors and Corporate Governance Walter de Gruyter, Berlin.


Kothari, S., Lys, T., Smith, C., Watts, R., 1988. Auditor liability and information disclosure. Journalof Accounting, Auditing and Finance 3, 307}339.

Kothari, S., Sloan, R., 1992. Information in earnings about future earnings: implications for earningsresponse coe$cients. Journal of Accounting and Economics 15, 143}171.

La Porta, L., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 1997. Legal determinants of external"nance. Journal of Finance 52, 1131}1150.

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

Miceli, T., 1997. Economics of the Law. Oxford University Press, Oxford.Miwa, Y., 1996. Firms and Industrial Organization in Japan. New York University Press, New

York.Mueller, G., Gernon, H., Meek, G., 1997. Accounting: An International Perspective. Richard D.

Irwin, Chicago, IL.Nobes, C., 1992. International Classi"cation of Financial Reporting. Routledge, London.Nobes, C., Parker, R., 1995. Comparative International Accounting 4th Edition. Prentice-Hall,

Englewood Cli!s, NJ.Ohlson, J., 1988. Accounting earnings, book value and dividends: the theory of the clean surplus

equation. Unpublished manuscript, Columbia University.Peltzman, S., 1976. Toward a more general theory of regulation. Journal of Law and Economics 19,

211}240.Pope, P., Walker, M., 1999. International di!erences in the timeliness, conservatism, and classi"ca-

tion of earnings. Journal of Accounting Research 37 (Suppl.), 53}87.Posner, R. A., 1986. Economic Analysis of Law, 3rd Edition. Little, Brown, Boston, MA.Posner, R.A., 1996. Law and Legal Theory in the UK and USA. Clarendon Press, Oxford.Radebaugh, L., Gray, S., 1997. International Accounting and Multinational Enterprises, 4th Edi-

tion. Wiley, New York.Ramesh, K., Thiagarajan, R., 1995. Inter-temporal decline in earnings response coe$cients. Unpub-

lished manuscript, Northwestern University.Ramseyer, J., 1993. Columbian cartel launches bid for Japanese "rms. Yale Law Journal 102,

2005}2020.Roe, M., 1994. Some di!erences in corporate governance in Germany, Japan and America. In:

Baums, Buxbaum, Hopt (Eds.), (Chapter 2). Institutional Investors and Corporate GovernanceWalter de Gruyter, Berlin.

Shavell, S., 1982. Suit, settlement and trial: a theoretical analysis under alternative methods for theallocation of legal costs. Journal of Legal Studies 11, 55}81.

Standish, P., 1997. The French Plan Comptable. Expert Comptable MeH dia, Paris.Vuong, Q., 1989. Likelihood ratio tests for model selection and non-nested hypotheses. Econo-

metrica 57, 307}333.Watts, R., Zimmerman, J., 1986. Positive Accounting Theory. Prentice-Hall, Englewood Cli!s,

NJ.Yamey, B., 1962. Some topics in the history of "nancial accounting in England, 1500-1900. In:

Baxter, W.T., Davidson, S. (Eds.), Studies in Accounting Theory. Sweet & Maxwell, London.


the e!ect of international institutional factors on properties of accounting

Documents

university of chicago

accounting conference

university of iowa

washington university

financial accounting

journal of accounting

hongkong university

codelaw accounting income