Date post: | 14-Apr-2018 |
Category: |
Documents |
Upload: | judy-moody |
View: | 227 times |
Download: | 0 times |
of 12
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
1/12
SiraprapaWatakit5502310013
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
2/12
OverviewandContribution TheoreticalBackgroundandModel e o o ogyan mp r ca esu s Conclusion
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
3/12
Overview
Thepaperexplainsabouttheeffectsofliquidity,information TheB/Aspreadofactive/inactivestocksareusuallylargebecauseof3
components(1.inventorycost/liquidityeffect2.marketpower3. n orma on ase
Theresearchareaisimportantnotonlytoacademicbutalsotopolicymaker Whatisthebestwaytostructureatradingsystemfornact ve ytra e stoc s
Withtradedataandnewtechnique,themajorfindingisthattheinformationbased componentisthemostimportantfactorfortheinactivestock
Contribution
Newempiricaltechnique,probabilityofinformedtrading(PIN)
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
4/12
TradeProcess MarketMakerModel:Wheremarketmakersquotebid/askandare
assumedriskneutral,themodelisamixofdiscrete/continuos time Investor: InformedandUninformed
, Informationevent occurs/notoccurs: Bad/Goodnewsarrival: Uninformed/Informedtradersarrival:
Notes: thesesetupareforallsinglestockswithexpectationof1even per ay,an emar e ma er sa ayes an
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
5/12
IntreediagramIfthereisnonews,only tradersparticipateifthereisnews oodorbad,both andparticipate
MarketMaker: AsheisaBayesian,hewouldupdatehispostbeliefsasaccordingtoa
priorbeliefs,heknowstherewouldbe3possibleoutcomeineachday Nonews,Good News(heavybuy),BadNews(heavysell)
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
6/12
Givenasellorderatt,theprobabilityofbidis
Sameideaforask
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
7/12
Thevalueofastockisanaverageamongthethreepossibleoutcome
Substituteinb/aequationyields
Hence,thespreadiscalculateasfollow
Theprobabilityofinformedtrade
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
8/12
Allparameters canbedirectlymodelwithalikelihoodfunction
Onceallparametersareestimated,wecandirectlycalculatefortheproportionofinformedanduninformedtradeineachparticularstock
Data
method,andonlycommonequityareconsidered
Volumeportfolios: Allsamplearecategorizeintodeciles,butonlymosthighvolumestockrank1st ,medium5th andlow8th areanalyze
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
9/12
Active Inactive
Table1 and aremostlyparticipate
in1st andmuchlessin5th and8th Prob(Inf) issmallin1st but
th th Table2
on rms a ro n ssignificantandstatisticallydifferentacrossdecile
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
10/12
TheCDFalsoshowsthethe Prob(Inf)isstrongestonthe8th ,whichmeansthatProb(Inf)mightbeanimportantfactorintheb/aspread a e3con rmssprea sw en ngass oc ecomemore nac ve
Hence,todeterminewhichfactorsdrivethespread,werun.
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
11/12
Model
Ifthemodeliscorrect,weexpect tobepositiveand tobenegative,asaccordingtoCDFandTable3
Results
Table4confirmstheexpectation
verysignificant,
highR2
7/30/2019 Easley, Kiefer,O'Hara,Paperman, 1996
12/12
Giventhenewempiricaltechnique Wecandirectlyestimatehowmanyinformedanduninformedtraders Wecanalsoconfirmthatprobabilityofinformedtradeisanimportant
factorswhichdeterminethespread,especialininactive/lowvolume Tominimizethespreadofinactivestock,giventhesefinding,we
shouldencouragethepolicywhichpromotesgreaterransparency sc osureo n orma on