A股B股的区别英文论文
以下，就是有道富投资为大家翻译的A股B股的区别英文论文
A.DataandPreliminaryStatistics
Thetradi荏鱿胫协ngprocessesforA-(local)andB-(foreign)sharesontheShanghai
StockExchange(SHSE)andtheShenzhenStockExchange(SZSE)aresimilar.
Bothexchangesrunorder-driven,automatedmarkets.Neitherexchangehas
designatedmarketmakers.Traderscanonlysubmitlimitorders,whicharrive
atanelectronicconsolidatedopenlimitorderbook(COLOB).Anincomingorder
isautomaticallymatchedagainstthebeststandinglimitorderintheCOLOB,
accordingtotheprice-timepriorityprinciple.Ifitcannotbematched,thenit
isaddedtotheCOLOB.Thereisnoblocktradingsystemthatallowsliquidity
traderstotradelargevolumesinanupstairsmarket.Off-exchangetrading
andinsidertradingarebothforbidden,butthisisnottightlymonitored.The
minimumtradesizeis100sharesforlocalsharesinbothmarkets,but1,000
sharesforShanghaiforeignsharesand100sharesforShenzhenforeignshares.
Ourdataconsistofalltime-stampedtradesandquotesfromJanuary2000
toNovember2001forallstockstradedontheSHSEandtheSZSE.7Weapply
anumberoffilterstoourdata.First,welimitthesampletofirmsthattraded
A-andB-sharesthroughoutthesampleperiod,therebyreducingthenumber
offirmsfromover1,000to84.Second,weexcludeeightfirms,becauseonly
forafewdayswefindnonzerovolumeintheB-sharemarket.Third,forthe
remaining76firms,weremovedaysforwhich,forexogenousreasons,there
wasnoorverylimitedtradingineithertheA-ortheB-sharemarket.8Fourth,
weremovestalequotes,whichareeasilyrecognizedthroughzerodepth.Fifth,
thefirstandlast15minutesofeachtradingsessionareremovedfromthe
sample.
TableIprovidestradingstatisticsfortheA-andB-sharesfromJanuaryto
December2000,aperiodduringwhichbothmarketswerefullysegmented.
Thetablepresentsthecross-sectionalmean,standarddeviation,minimum,
andmaximumbasedonthe76samplestocks.Forcomparison,thetradeand
quotedatafortheB-shares(inHongKongdollarsforSZSEandinU.S.dollars
fortheSHSE)areconvertedtoyuanusingthe(fixed)officialexchangerate
forthesampleperiod.Theaveragetradepriceinthisperiodis14.29yuan
($1.73)forA-sharesand3.10yuan($0.37)forB-shares.Thiscorrespondstoan
averageB-sharediscountof72%,whichisinlinewithpreviousevidence(see,
e.g.,Baileyetal.(1999)).
Intermsofvolume,theB-sharemarketisabouthalfthesizeoftheA-share
market—anaverageof854,000versus1,684,000sharesperday.TheA-share
marketismoreactivelytraded,asexpected.Furthermore,wefindthatwhile
thetradingfrequencyishigherintheA-sharemarket,thetransactionsize
islower.Thishighlightstheimportanceofcontrollingfortradesizeinour
spreaddecompositionanalysis.Althoughtheaveragequotedspreadis0.027
yuanforA-sharesand0.035yuanforB-shares,thisdifferenceisnotsignificant
atconventionalconfidencelevels.Theeffectivespreadsinthetwomarketsare
bothequalto0.035yuan.
B.PreliminaryAnalysisofInformationAsymmetryMeasures
Inthissubsection,weestimatethepriceimpactcoefficients(PI),adverse
selectioncomponents(AS),andtheprobabilityofinformedtrading(PIN)
measures,andwestudywhethertheseinformationasymmetrymeasuresagree,
cross-sectionally,astowhichsecuritiesexhibitthehighestinformationasymmetry.
Wethenrelatethesemeasurestotheforeignsharediscount.
InpreparingthedatafortheestimationofPIandAS,wefollowGlostenand
Harris(1988)andtruncatethetradesizeto100,000sharestoavoidgivingtoo
muchweighttolargetrades.Themediantruncationfrequencyis0.28%and
1.55%fortheA-andB-sharemarket,respectively.Thisismostlikelytheresult
ofmoreinstitutionalinvestorswhotradeinlargersizes,intheB-sharemarket.
Wealsotruncatethetradesizeto200,000and400,000sharesandfindthat
theresultsaregenerallysimilar.
PanelAofTableIIcontainsestimatesofthepriceimpactcoefficientsforthe
76A-andB-shares.Themeanestimatesofγandφare9.66×10−7yuanper
shareand8.58×10−3yuanfortheA-sharemarket,and2.61×10−7and6.35
×10−3fortheB-sharemarket.Therefore,althoughA-sharevolumeishigher
thanB-sharevolume(seeTableI),A-sharedepthisactuallylowerthanB-share
depth.Thisresultisinterestingasitshowsthatoneshouldnotequateliquidity
withtradingvolume.Thus,consistentwithourmodelinSectionII,evenwithout
tradingvolumeconsideration,theinformationasymmetrybetweenA-and
B-sharemarketscouldcausetheirmarketdepthstobedifferent.Thismotivates
whyinsubsequentregressionanalyseswetesttheextenttowhichourinformation
asymmetrymeasuresexplaintheB-sharediscountaftercontrollingfor
tradingactivity.
PanelBofTableIIcontainscross-sectionalstatisticsonthefixedandvariable
adverseselectioncomponent(AS)ofthespreadandongrossprofitforthe76
A-andB-shares.TheAScomponentcoefficients,z0andz1,aresignificant,and
carrytheexpectedsignforallofthe76securitiesintheA-sharemarketand
foralmostall(66and61,respectively)ofthesecuritiesintheB-sharemarket.
Thisevidenceisconsistentwithpreviousliteratureinthatthecostofadverse
selectionisasignificantcomponentofthespreadthatincreaseswiththesize
ofthetransaction.Thefixedgrossprofitcoefficients,c0,aresignificant,and
carrytheexpectedsignforallofthesecuritiesintheA-sharemarketandfor67
ofthesecuritiesintheB-sharemarket.
Thevariablegrossprofitcoefficients,c1,aresignificantfor55ofthesecuritiesintheA-sharemarketandforonly4ofthesecuritiesintheB-sharemarket.ThelastcolumninPanelBreportsthe
AScomponentforamedian-sizedtrade.ConsistentwiththePIestimate,we
findthattheaverageAScomponentislargerintheA-sharemarketthaninthe
B-sharemarket,withvaluesof56.4×10−4and46.8×10−4yuan,respectively.
AprobleminestimatingPIandtheAScomponentforthetwomarketsis
thatthereisalowertradefrequencyforB-sharesascomparedtoA-shares.In
additiontoestimatingtheAScomponentbasedontrade-by-tradedata,wealso
estimatebasedonfixed-lengthintervals.Forinstance,inthecaseof15minutes,
pricechangefromt–1totismeasuredasthepricechangeoverthe15-minute
interval,V(t)isdefinedastheabsolutevalueofnetorderflowinthe15-minute
intervalt,andQ(t)isdefinedas1or–1dependingonwhetherthenetorder
flowispositiveornegativeinthe15-minuteintervalt.Ingeneral,wefind
thatusingalternativeestimatesofPIandtheAScomponentinoursubsequent
regressionanalysesproducesqualitativelysimilarresults.
PanelCofTableIIpresentscross-sectionalstatisticsontheparameterestimatesofthePINmodel.Again,wefindconsiderableevidenceforprivately
informedtradersintheA-sharemarket,astheaveragearrivalrateofthese
typesoftraders,μ,is0.38,whichisofthesamelevelofmagnitudeasthearrival
ratesofuninformedbuyers(0.44)andsellers(0.51).Thearrivalrateof
informedtradersintheB-sharemarketislowerat0.11,andagainisofthe
samelevelofmagnitudeasthearrivalratesofuninformedbuyers(0.06)and
sellers(0.07)inthismarket.Theprobabilityofaninformationeventonaspecific
day,α,ishigherintheA-sharemarketthanintheB-sharemarket,with
ratesof0.36versus0.31,respectively.Thisisconsistentwiththeexistenceof
moreinformationintheA-sharemarket.However,theaveragelevelofthePIN
measureishigherintheB-sharemarket,asthismarketexhibitsarelatively
lownumberofuninformedtrades.Allparametersaresignificantforthemajority
ofthesecuritiesexceptfortheparameterδ,whichistheprobabilityofthe
newsbeingbadnews.
Wecompareourmeasuresofinformationasymmetrybyverifyingwhether
theyagreecross-sectionallyastowhichsecuritiesexhibitthemostinformation
asymmetry.Figure4presentsscatterplotsoftheAScomponentofthespread
(foramedian-sizedtrade)againstPIandthePINforboththeA-andB-share
markets.Theseplotssuggestastrongerrelationshipbetweenthemeasures
intheA-sharemarket,withcorrelationsof89%and59%,respectively.The
relationshipintheB-sharemarketismuchweaker.
Finally,scatterplotanalysisrevealsthattheB-sharediscountappearstobe
explainedbytheproposedinformationasymmetrymeasures.Figure5plots
foreignsharediscountsagainstthedifferentialsoftheinformationasymmetry
measuresintheA-sharemarketrelativetotheB-sharemarket,asmeasured
byPI,theAScomponent,orthePIN.Forallmeasures,wefindthatstocks
withrelativelyhigherinformationasymmetryappeartocommandhigherBshare
discounts.Thecorrelationsbetweenthethreeinformationasymmetry
measuresandthediscountsare66%forPI,67%forAS,and28%forPIN,and
arestatisticallysignificantateitherthe1%or5%level.
D.ChangesAftertheB-ShareMarketWasOpenedUptoLocalInvestors
InMarch2001,regulatorsopenedtheB-sharemarkettodomesticinvestors.
Weusethisregulatorychangeeventtofurthertestourinformationasymmetry
hypothesisintwoways.First,weexpectB-sharediscountstoshrinkorvanish
and,moreimportant,ourinformationasymmetrymeasurestoincreaseforthe
B-sharemarketafterthisevent,becausebetter-informeddomesticinvestorsare
nowallowedtoparticipateinthismarket.Second,werepeatourcross-sectional
regressionsforthenewsampleperiodasarobustnesstest.Weanalyzethesame
76stocksforthesampleperiodofApriltoNovember2001.
Wefindthat,consistentwithourhypothesis,thediscountlevelsdecrease
fromanaverageof72%to43%,andthelevelofinformedtradingintheBshare
marketincreasescomparedwiththemodelestimatesforthepreevent
period.Themainreasonforthisgradual,ratherthansudden,declineindiscounts
isthelackofforeigncurrencyamongdomesticinvestors.PanelAin
TableVpresentspre-andpostentryestimatesofPI,andshowsthatPIalmost
doubles(+81%)fortheB-sharemarketfrom2.6preentryto4.7postentry.This
increaseislargerthanthe21%increaseintheA-sharemarketandistherefore
consistentwiththearrivalofbetter-informeddomesticinvestorsinthe
B-sharemarket.PanelBconfirmstheseresultsbasedontheAScomponentof
thespread.Forboththefixed(z0)andthevariable(z1)AScomponent,wefind
considerableincreasesintheB-sharemarketof52%and101%,respectively,
afterdomesticinvestorswereallowedtoenterintothismarket.
Wefindthat
theAScomponentformedian-sizedtradesincreasesby44%from46.8centsto
67.4cents.ThisincreaseisagainlargerthantheincreaseintheA-sharemarket
andthusconsistentwiththePIfindings.PanelCofTableVrevealsthat
thearrivalrateofinformedinvestorsintheB-sharemarketincreasesby164%.
Thisrateactuallydecreasesby29%intheA-sharemarket.Furtherevidence
ofincreasedinformationintheB-sharemarketisthattheprobabilityofan
informationeventincreasesfrom0.31to0.42,whereasintheA-sharemarket,
theprobabilityincreasesonlyfrom0.36to0.38.Nevertheless,thelevelofthe
PINmeasuredoesnotchangemuchineithermarket,becausethechangesinthearrivalrateofuninformedtradersroughlymatchthoseofinformed
traders.
Theregressionsoftheforeignsharediscountontheinformationasymmetry
measuresandcontrolsintheposteventperiodareconsistentwithearlierfindings.
TableVIpresentstheresultsforPI,AS,andPINincludingandexcluding
thecontrolvariables.Inalloftheregressions,theinformationasymmetrymeasures
aresignificantlypositive,withthePIandAScomponentmeasuresexplaining
61%and71%ofthevariationinforeignsharediscounts,respectively.
E.DiscussionandInterpretationoftheResults
Overall,ourresultsprovidestrongevidencethattheinformationasymmetry
measures,especiallytheadverseselectioncomponentofthebid-askspread,are
farmoreimportantthananyofthecontrolvariablesinexplainingthecrosssectional
variationinB-sharediscounts.Ourinterpretationisthattheinformation
asymmetrymeasuresreflecttheextentofprivateinformationavailable
todomesticinvestors.Whenthereisahigherdegreeofinformationasymmetry,
asmeasuredbyahigherpriceimpactcoefficientoradverseselectioncomponent
intheA-sharemarketthanintheB-sharemarket,domesticinvestorsare
morewillingtopayahigherpricethanforeigninvestors,resultinginB-share
discounts.However,severalissuesareworthyofdiscussion.
Thefirstissueconcernsthetypeofprivateinformationtowhichwerefer.It
iswidelybelievedthatChineseinvestorstradeonrumorsratherthanfundamentals.
Furthermore,sharemanipulationiswidespread,pushingpricesaway
fromtheintrinsicvalueforarelativelylongtimeperiod.Meietal.(2003)argue
thatA-sharepricesareareflectionofspeculativebubbles.Thus,itisnotclear
whetherourinformationasymmetrymeasuresreflectfundamentalnews.Although
weagreewiththisviewofthespeculativebehaviorofinvestors,wedo
notthinkthatitshoulddisqualifyourpriceimpactcoefficientoradverseselection
componentfrombeingeffectivemeasuresofinformationasymmetry.