manufact 報告 管理 点 見 日 業 企 の時 計 へ

The Japanese Association of Management Accounting

NII-Electronic Library Service

The 　Japanese 　Assoclatlon 　of　Management 　Accountlng

H本管理会 計 学 会 誌

管 理 会1言t学　2r，14_{年 　第}22_巻@ 第｝　

論

文

報 告

利 ^益

管 理

の 観

点

か ら

見

る

日

^本

^企 業 ^{の 時}

^{価 会}

^計 へ

対 応

　健 民 〈

文 要 旨^〉 本 研 究 で^は、^{時 価}会^{計 の 導入}に 対し て日本 企業 が ど の よ うな報告^利 益^{管 理 を}

っ た のか に っ い^て検証 した。本 研 究^で は 、 日 本^企業が時価 ^会計の 導入に^対し 、 損^失回 の 報告^利 益管 理 を 行ったことを確 認 し た 。 特 に 、 成^長企 業^{や 非}製^{造 業}に 属 す^{る 企}業が

質 報^告 利益 管 理 を行 い^、 損失 回 避行動 をとった

と^が確_認で_きた

〈^キー ワード 〉 発 生 項 目 報 告 利^益 管理^、実^質報^告利益管

、特 別 項目^、時^{価 会 計} How　Japanese 　Firms　 Respond 　 to 　Mark − to− Market 　A

ounting^？AnEarnings 　Management

PerspectiveChine ⁻

n　Kevin　

nAbstractln this　pape^臥linvestigate　how」apanese 　firms　respond 　to　mark−to−market　accounting fr

　 an　eamingsmanagement ^{perspec 百ve．　I}hypothesize ，　and　find　evidence　to　show^{．　that　Japanese　firms 　of}

er 　expensesoccurring 　from　income−decreaslngextraordinary 　items　Ihrough　bo1h 　accrualsmanage

nt　 and　realeamjngs 　management 　I　also 　present 　evidence 　that　firms _with 　growth　potentialtend　to　ma ge　^{earningsby 　using 　real 　eamings 　management}．^{亘also 　obtain 　evidence 　that ，　unlike}

manufact

ing 　firrns， nonmanufacturing　firms 　tend to_manage 　eamings　upward 　byusing　ov

production ^， Keywords ^：Accruals　Management ，　Real Eamings 　Management ，　Extraordinary ^{Items ，}M

k−to−marke

Accounting 2013 年4月8日　受付　　　　Subrnitte

@8　April 　2013 2014 ^{年 玉}月20日^受理　^Accepted

　January 　2014 台湾 国 立中 ^{正丿く学会計及び情報技術学科　Assistant 　Prefesser ，}　 Department 　ofAccounting d 　Information 助 理 教 授　　Teclinology_．National　Chung　ChengL≡］liN

er

ty ．　_{Taiwan49N 工工 一}

erx ft:･tvas²²igrg1e

4SooO =- ------r

4000D

350DO +---- .n-..-MM..rt..T-.,

tf I fEe

t

-:slgl

13oooa -------.....

2SOOOt---･----

safx..

1i''

..,.... [

ileoooi--- ,-

/.di, .fit /

Ai` Yisagl

^{,,,, n}

lsooe"---N.'..S.M-Tm."m..ttttttttt w.--.u...--za.-...,v.fi./-･;･ge---･--x^zsi'.,lwt1lk

ftf'

f,esWg

v

1000D

hllipuIJItu"s/A-"^{"- '" '}

5000 .---･･----..-･---.--- ... .-.. ..."...

ktS,

-----･･---･--/･di-.Xes.7 -'-

[

i1

.ifgF"'llsSilpt.rl""y"lts-tlttL.orsiYofSk?E"ist14SstSg ¢N'IYIIeN"}ipeYiif.,ipts*"%ts&..,ptS*i,'ej Figure1-Nikkei_{StockMarketAverage} (Monthly^{C]oslngPrices)}

1.Introduction

Inthis_paper,Iinvestigatehow Japanesefirmsrespond tomark-to-market accounting

from ^{an earnings management} perspective,^Ido so by taking advantage ^of Japan's

unique institutional^setting. Beginninginthe late1990s,Japaninitiated^{a series of}

accounting standard refbrms inan effbrttocatch up with global^trendsinthewake of the internationalharmonizationof accounting standards.i Mark-to-market accounting

firstcame intoeffect inJapanin200e.Unexpectedly,itforcedJapanesefirmsto operationally shift the holdingposition^{oftheirsecurities}holdings.As shown inFigure 1,stock prices^{started todeclinein2000.Figure2}illustratestime-series trendsin^mean

cash holdings,marketable securities holdings,^and investing^securities holdings^of Japanesefimisinthe period1990-2008.2Itshows thatalthough thecash holdings position^{was sustained,} in2001,therewas a drasticshiftintheholdingspositienfor

marketable securitiesand investingsecurities.3

In ^thispaper,Idocumentthereporting of 6,582

(out

^{of 8,205)firm-yearsinwhich}

there^were income-decreasingextraordinary items in^the period^{2000-2004. However,} lossreporting increasedfrom 899firm-yearsat the ordinary incomelevel

(earnings

bqforeextraordinary items)to 1,843firm-years^at the pretaxincome level

(earnings ofer

extraordinary items).This evidence raises several questions.First,why did JapaResefirmschange their securities holdingpositionsfbllowing^the introductionof mark-te-market accounting? Iffirmsare not motivated to undertake earnings management

(regardless

^{of balancemanagement or income statement} management),

thereis^{no motivation} forthem to aggressively change theirsecuritiesholdingpositions fo11owingthe introduction^of mark-to-market accounting. Second,fbllowingthe response of securities holding_positions to the introduction^of mark-to-market

The Japanese Association of Management Accounting

NII-Electronic Library Service

The JapaneseAssociation of Management Accounting

W"itlkgeopth,#.h"6n6Hptdi*aveetuft:･tAhdistms

O.IS

O.IS

O.14 t-"-..m."....---. .-m-.-m..mm.."..-"t..-"t.ttt-tttttH/ttt-tt-t"t-/.t"".tt.t""t-tttttttttt-tttttttttttttttt.t/ttt-.

O.11

dash

tt t

O.1tttttttttttuLttttttttttutttttttuttt-ttttt-t-tttttmtttttttttttttHtttHtttt-tt-tttttttttttttttttLLLtttttttttttttttttttVt

- _}nt･･

ttt

,?F;:ii<,

-t-tt-fiH-tthtmfihattttt-ttt

o,es'rmm'T..-....t t tt''i"''""../ lt'NlnvestingSec"ritlas

o,os

o.e4 '

' MerkotibleSeEurities

o.o] '

o1990'1"11"1lggs-m.dach t t

19g4199S19961997199e1999

wwMarketaNeSecurlties(CurrentAssets)2ooolool2ool2oo!2ooa2ooS nvlnvestihgSecurlties(Non(llrrtntAsse ±s)2oe62oo72eos

E

i'lj'

lil

Figure2-_{Mean Cash,Marketable}Securities,and InvcstingSecuritiesHoldingsefjapanese Firms (based^{on 21,474}observations on lbkyoStockExchangelistedfirms}

accounting, why didprofit-reportingfirms^outnumber loss-reponingfirms?Beeausethe

introductionofmark-to-market accounting was anticipated, why didfirmswant toavoid reporting Iossesifthemarket ^was able to ascertain the effect of mark-to-market accounting?

The drasticchange insecurities holdingpositions^stems fromthe corresponding classifications inincomestatements formarketable securitiesand investing^seeurities inheritedfromtheJapaneseGenerally^Accepted AccourrtingPrinciples(GAAP).^Before

the"all-inclusive" incornestatement replaced the "curTent

operating perfomiance"

income ^statement in 1974,4 "ordinary income" (current^{income) had beenthe}

bottom-line^earnings figureintheincome^statements _prepared byJapanesefirms.Even

aftertherevision of the JapaneseGAAP in1974,^which placedtheclassification of extraordinary itemsbelow ordinary income,business_people inJapanstillconsidered ordinary income as theindexof a firm'sperfbrmance.Therefbre,^an overemphasis on ordinary income^{creates a strong} incentiveforJapanese firmstoavoid reporting losses

attheordinary incomelevel.

Herrmann,Inoue,and Thomas

(2003)

^{and Pan and Tseji}

(2013)

^{show that Japanese} firms'management of earnings tooverstate ordinary incomestems fromtheJapanese GAAP treating unrealized holdinggains^and lossesof securities and fixedassets

differentlydependingon how theyare classifiedon balancesheets.Panand Tstlji

(2013)

furthershow thatJapanesefirmscan operationally shift theclassification ofsecuritiesto

manage earnings at the ordinary incomelevel.When ^a firrn^{sells marketable securities}

or fixed^{assets classified as current assets,} itis^required toreport theunrealized holding gains^orlossesas nonoperating gains^or losses

(items

^{above ordinary income).However,}

when the firmsells nonmarketable securities or fixedassets classified as noncurrent

51

ggefte-tgeag^22tsig^I't'

assets, thefirmmust report theunrealized holding_gainsor lossesas extraordinary gains

or losses

(items

^{belowordinary income).}

The mechanism by which a Japanesefirmadjusts itsordinary income _vpward by

selling assets thatgenerate^unrealized holding_gainsisas follows.The firmfirstshifts the classification of underlying assets from^{noncurrent assets} to current assets. When the

firm^{sells underlying assets,} it^{can report} theunrealized holdinggains^as nonoperating

gains^to increase^ordinary income.Conyersely,firrns^{can overstate} theirordinary

incomeby^{selling assets} thatgenerate^unrealized holdinglosses.The firmfirstshiftsthe

classification ofthe underlying ^assets fromcurrent assets to^noncurrent assets.Then, the

fimi can report the unrealized holdinglossesas extraordinary lossesinsteadof nonoperating losses,toincreaseordinary incorne.

Hemnann, Inoue,and Thomas

(2003)

^{show that,before}mark-to-market accounting was incorporatedintothe JapaneseGAAR Japanesefirmsmanaged theirearnings by

selling fixed^{assets or marketable securities to} generate^unrealized gains^{or losses.}

However,Panand [fsoji(2013)^{argue that}thechanges documentedbyHerrmann,Inoue,

and Thomas

(2003)

^relatetotheperiodbefbreJapan'sAccountingBigBangoccurred;

hence,such changes ^{may not} havebeen^sustained since the accounting standard ^refbrms Were initiated.S

This_paper contributes totheJiteraturebyshowing thatJapanesefirmsresponded to theintroductionof mark-to-market accounting by operationally shifting theirsecurities

holding _positions.Despitethe reclassification of securities being triggered by

anticipated external events, Japanesefirmssti11triedtomanage theirearnings upward through both^{accruals and real earnings management to offset} income-decreasing

extraordinary itemsto avoid reporting losses.The ^evidence presentedinthis paper

suggests thatJapanesefirmsare strongly motivated toavoid reporting losses,

2.Literature,

Institution

^Issuesin

Japan

^and

Hypothesis

2.IEarningsManagement in.lapan

Japan-basedstudies providesubstantial evidences thatshow theJapanesemanagers' engagement in^earnings management. Shuto

(2000),

^{and Sudaand Shuto}

(2001)

^provide

evidences on accruals management. Suda and Shuto

(2005)

present^earnings

management evidences withcombined eamings distribution^{and accruals} management.

Shuto(2007)^{linksJapanesefirms'earnings management tothe} compensation schemes ofthe executives. Yamashitaand Otogawa

(2008)

investigateiffirmsmanage earnings

inresponses to the tax reduction in late1990sinJapan.Onuma

(2004)

^analyzes

Japanesecommercial banks'use of ^valuation allowance of deferredtaxes fbr^earnings

management since its^effective in2000. Pan

(2009)

presents^{evidences that Japanese} firrns^engage in^{real earnings} managernent intheearly 2000swhile reporting income decreasingaccruals suggesting that real activities earnings management can be a

pessible^substitute fbraccruals managements.

^{Earningsmanagement} byJapanesefirmsalsQ drawsattentions from outside ^ofJapan.

Darrough,Pounialaliand Saudagaran

(1998)

^{show that Japanese firms}participatein

accruals management. Mande, File^and Kwak (2000)^{findthat Japanesemanagers}

cutting R&D budgets^tosmoeth income.Herrmann,Inoue,and Thomas

(2003)

^specify

Japanesefirms'^{earnings management} activitiesthrough thesales of fixedassets ^and

The Japanese Association of Management Accounting

NII-Electronic Library Service

The JapaneseAssociation of Management Accounting

W-if]asepae)en,ne:6n6H4fiIOeetukEtA.Ostpt

marketable securities. Higgins

(2013)

^{argues and} investigatesif Japanese

stock-for-stock ^{merger acquirers manage earnings} before^merger announcernent and

present^{evidence showing} thatacquirer firmsexhibit significant positivelong-term

abnormal accruals befbrethestock-swaps mergers.

The ^above evidences, talcentogether, suggest thatJapanesefirmsrnay accordingly manage earnings to serve firmsinterests.

2.2DofnitionsofExtrao,dinary^{fremsin.lbpan and Plorking}thpothesis

^{Accordingto the}L`Supplement 12 ExtraordinaryItems"^of the-Japanese GAAe6

extraordinary items^{consist oftwo main} eomponents: (1)non-recurTing items

(including

[a]gain^or loss^{on sale of} fixed^assets, [b]gain^or loss^{on sales of securities acquired} for pui[poses^other than^resale, and [c]^{casualty loss),and}

(2)

priorperiodadjustments'

(including

^{adjustment of}

[a]

^reyerses,

[b]

depreciation,

[c]

^{correction of inventory}

valuation, ^and [d]^{recovery of bad debtswritten off in}priorperiods).8"9Inaddition,

"extraordinary itemsof an immaterialamount or of a recuning nature may beclassified

as ordinary item."iO

The definitionof extraordinary itemsinJapaneseGAAP allows more room fbr irrterpretations^{compared tothe US} Gne ^or IFRS (IFRS^{doesnot havea} classification

fbrextraordinary items).Inpractice,extraordinary items^reported by Japanesefirms

usually includegainsAosses^{on sales}ofiong-term investmentsinproperties,^equipment,

real estates, and other-than-trading securities, gains!lossesfrom ^retirement benefits

(since

^{2001),impairmentexpense}

(since

^{2004),and gainsfromnegative goodwill}

(since

201o).ii

^{Herrmann,Inoue,and Thomas}

(2003)

^{show that7duringthe 1990s,Japanesefirms}

increased

(or

^{decreased)earnings through thesale of fixedassets and marketable}

securities to mitigate fbrecast^{management errors} in^{operating earnings,} Beforethe

imp]ementation^of fixed_assetsimpairmentin2004,fixed^assets inJapan^{were measured}

at historical^cost lessaccumulated depreciation.Thus,book-market^value differencesin fixedassets persist^until thefixedassets are sold. InJapan,marketable securities are valued at either cost or market value, whichever is^{the lowest.Firms have had}this

option even befbremark-to-market accounting was introduced.Hence,becauseef

book-marketvalue differencesinfixedassets and marketable securities,unrealized

holdinggains^or losses^{are created.} Thus,Japanesefirms^{can exploit} the rules byusing

fixed^{assets or marketable securities} toengineer unrealized holdinggains^or losses.A

firmthatexpects toperformpoorlyinthecurrent period^{can create unrealized} holding gains^to increaseearnings and preserve^{unrealized holding} losseson the books.By

contrast, a firmthatperfbrmsbetterthan expected can create unrealized holdinglosses todecrease^eamings and preserve^unrealized holding_gainsas reserves tooffset future losses.However, the introduction^of mark-to-market accounting isexpected toprevent firmsfrommaintaining unrealized holdinggains^and Iosses^{as reserves.} Infact,Panand ,Tsaji

(2013)

^{show thatresults} presentedⁱⁿHerrTnann,Inoue,and Thomas

(2003)

^might

not apply inthe 2000s.

In thispapeg I investigatewhy 6,582

(out

^of 8,205)firm-years^{reported net} income-decreasingextraordinary itemsduringthesample period^{under study while} the

number ^of loss-reportingfirrnsinereasedfrom899to1,843firm-years,Pan

(2009)

^and

Yamaguchi

(2009)

^{show thatJapanesefirmsmanage earnings tomeet various earnings}

benchmarks

(e.g.,

^{pretaxearnings, zero earnings, earnings from}previous^{earnings, and}

53

ifgattf\ ca22tseg1e

management earnings fbrecasts).Firmsare unlikely to change theirsecuritiesholding positionsfbrno reason. The number of loss-reportingfirrnssuggests that managers are stillreluctant to record lossesafter reporting income-decreasingextraordinary items.I

then coniecture and hypothesize that firrnsmanage earnings upward while reporting

income-decreasingextraordinary items.

formal Ib/pothesis:Firmsthat^report income-decreasingextraordinary items(i.e,,

firmsreport net lossesfromextraordinary items)^{manage earnings upward} toavoid or decreaselosses.

3.Research Design

3.1Modeling ProxiesforRealEarningsManagement

Tb estimate real earnings management, fo11owingRoychowdhury

(2006),

^{Iuse cash}

fiowfrem operations

(CFO)

^{and use} discretionaryexpenses and production ^costs to proxy ^{real earnings management} activities.Severalauthors

(e.g.,

^{Gunny 2010;Cohen}

and Zarowin201O;^and Zang2012)have^{obtained and validated results on real earnings}

management by^using thesevariables. FollowingRoychowdhury

(2006),

^{Ifbcuson the}

abnormal levelsof CFO, discretionaryexpenses, and production ^costs, These three

variables are used toproxy thereal eamings management activities listedbelow.

1.Acceleratingsalesbyoffering increased_pricediscountsor more lenientcredit terrns.

2.0verproducingto report rower^{cost ofgoods} sold.

3.Cuttingadvertising, R&D, ^and SG&A ^expenses todecreasediscretionary^expenses.

SblesAcceleration:Salesvolumes can be temporary boosted^by offbring price discountor more generous^{credit terms.However, temporary increasedsales volumes}

are likelytodisappearonce thefirmsstop offeging such discounts.The increasedsales volumes wili increasethe^current period^eainings butnot thecuirentperiod^cash flows.

OveT:production:Fixed^{costs of a specific} preduct^{will not vary} withthe increase

volumes produced.Thus,aslong^as thedecreasedper-unitfixed^{costs are not offset} by

the increasedper-unit^{marginal costs,}overproducing will lowertotalper-unit^{costs of a} product.Assume that the sales ^volumes stay the same, the firmscan report higher

operating margin incurrent period ^with thedecreasesinreported costs of goods ^sold

(COGS).

CuttingDiscretionar:}slixpenses:Firms _can boostcurrent period^earnings by ^cutting

expenses likeadvertising, R&D ^and selling, general^and administrative ^expenses

(SG&A).

^{Theseexpenses are generallypaidincash; therefbre,cutting theseexpenses}

will also leadtoan increasein^current _period^cash flows,ⁱ²

Thesethreeactivities generate^abnorrnal levels^of CFO, discretionaryexpenses, ^and production ^costs. Iuse these variables toproxy^{real earnings management} activities.

First,I use a model developed by Dechow, Kothariand Watts (1998)^and

Roychowdhury

(2006)

^{to estimate thenorrnal levelof CFO,} discretionaryexpenses, and

production^costs. The^norrnal level^ofCFO is^{expressed as} thefbllowinglinearfunction

ofcurrent salesand itschange: ⁱ³

The Japanese Association of Management Accounting

NII-Electronic Library Service

The JapaneseAssociation ofManagement Accounting

ewRiUEljptoue,ti,,h;6n4H4fuloeetuk't7ttxsoNms

ewo,,,^{= ae} + !?ISZiles,,,⁺ 13,^{ASales,,,+ s,,,.}

(l)

1[b^obtain an estimate ^{of abnorrnal} CFO, Isubtract estimated CFO basedon

(l)

^from

actual CFO.

^{Following the existing} literature,Idefine_production^{costs as} the^{sum of} COGS ^and thecurrent change ininventory.COGS ^{ismodeled as a} linearfunction^ofcurrent sales:

COGS,,,^{= ao +}

1]iSbiesi,,

^{+ si,,･}

(2)

The change ininventoryismodeled byusing thefbllowing regresslon:^.

AllNZ･,,⁼ a, + Assales,,,⁺ AASales,,,-i^{+ s,,,.} (3)

Equations

(2)

^and (3)^{combine to}generate^estimated production^costs:

PRODi,,⁼ ao + jC3iSblesi,,+

f72asalesi.,

⁺ Aasalesi,,-]^{+ ei,,･}

(4)

Normaldiscretionaryexpenses are modeled as a linearfunctionofsales:

Dlsexi,t=ae ⁺

13iSbles,,,

⁺ /3>ASalesi,,^{+ s,,,.}

(5)

However,iffirrns^manage sales upward toincreasereported earnings inthecurrent

year,^modeling discretionaryexpenses ^{as a} functionof current sales causes theerror

term tobebiaseddownward.^{Tb address thisissue,Ilet}discretionaryexpenses be a

fimction^{ofsales inthe}_{previousyear:}

Disexi,t⁼ ao + fiiSblesij-i^{+ si,,.}

(6)

The^{variables used} intheabove regressions are definedasfollows.

CFOi,,Salesi,tASblesi.t

AINVI･,, PROD,,,

Disexi,t Skelesi,F]

ASZilesi,i-i

= cash flowsfromoperations inperiod^t

= salesgeneratedin_periodt

= changes in^sales inperiod^t

= changes ininventoryinperiod^t

== production^costs,definedas^cost of goods^{sold + change} ininventory

= discretionaryexpenses, definedasSG&Ai4

= sales generatedin_periodFl

= changes insalesinthepriorperiod^t-1.

^Ifurtherdefinethe abnormal leyels^of CFO, _produ^¢tioncosts, and discretionary

expenses as thedifferencesbetweentheiractual^{values and} theirestimated values based

on

(1),

(4),^and

^(6),

respectively. Theseestimated abnormal levelsofvariables are used to proxy^{real earnings management} activities.Iexpect thatfirmsthatconduct ^real eamings management toboostreported earnings ^upward will exhibit thefo11owing:

(i)

an unusually low level^of CFO; (ii)^{an unusually low levelof} discretionary^expenses;

and!or (iii)^{an unusually highlevel}ofproduction costs.

55

gmpa:-t\^as22tsee1e

ProxiesforAccruals-BasedEarningsManagement

I^use standard procedures^toproxyaccruals-based earnings management, Following Jones(1991),^{Imodel normal accruals as fo11ows:}

Zxl,,,⁼ a, + iBIASales,,, + JBI,PPEI., + s,,,_,

(7)

where

7]4i.,ASalesi,t pPEi,,

= earnings befbreextraordinary items^minus operating cash fiow

= change insalesfrompreviousyear

= gross^value ofpreperty, plant,^{and equipment.}

3.2ResearchDesignforFirmsRqportingineom'e-DecreasingElrtraordinarlyfrems ^[Ibexamine ifJapanese firmsmanage earnings upward tooffset expenses arising

fromincome-decreasingextraordinary items,^{Iestimate the fbllowing}regression, which

isbased on thatused by Roychowdhury (2006).

X ^{= ao + ll,Size,+ 13,MTB,+ /3kEBEL+}

17,Ei\TIL4,

^{+ s,}

(8)

^[Eb determine^whether firmsthat report income-decreasingextraordinary items

engage in^{real eamings management} activities,Iestimate theexcess levelsof accruals,

CFO, _production^{costs, and} discretionary^{expenses. The} dependentvariable }1

represents theexcess levelof each ofthese ^variables inperiod^{t.These excess} levelsare estimated fromregressiens

(l), (4), (6),

^and

(7),

^{Foreach variable, theexcess amount is}

rneasured as thediffbrencebetweenitsactual value and its

(regression-based)

predicted

value fbreach industry-year.The indicator^variable EX77RA is^equal to one fbrfirms

reponing net extraordinary losses,^{and zero otherwise.}

The ^regression includesthreecontrol variables: the market-to-book ratio ef equity

(MIZB),

^firmsize

(Size),

^{and earnings befi)re}extraordinary items(ordinary^income)

(EBEI),

^{which isan earnings measure. Allthreevariables are measured as deviations} from^their industry-yearmeans. IincludeSize

(measured

^{by market} capitalization) to

control fbrsystematic differencesineach excess variable. Throughout,Iuse market capitalization as a scaling factor.Hence, I take the natural logarithm^of market capitalization. MZB is^used to control fbrdifferencesinfirrns'growthopportunities.

EBIIIisincludedto control fbrfirmperfbrmance.^iS

4.

Sample

^and DescriptiveStatistics 4.1Dataand Sample

Financialand stock pricedata^{are obtained} from theNikkeiEconomic Electronic DatabankSystem-Financial

QUEST

^{online databaseservice,Allfirmsinthesample are}

listed^on thefirst^and second sections of the rlbkyo

SecuritiesExchange.Allfirms includedinthesarnple report ^results inthe fiscal_yearending inMarch.The datacover

theperiodfromfiscal_year^2000tofiscal_year^{2004.i6 Tb ensure 'data}consistency, firms

thatchanged theend dateoftheir fiscal_yearduringthe^sample period^{were removed. In}