The Law and Ethics of High-Frequency Trading

Minnesota Journal of Law, Science & Technology Minnesota Journal of Law, Science & Technology

The Law and Ethics of High-Frequency Trading The Law and Ethics of High-Frequency Trading

Steven R. McNamara American University of Beirut, Olayan School of Business

Article 2

Part of the Banking and Finance Law Commons

The Minnesota Journal of Law, Science & Technology is published by the University of Minnesota Libraries Publishing.

TheLawandEthicsofHigh-Frequency Trading

StevenMcNamara*

ABSTRACT MichaelLewis’srecentbookFlashBoyshasresurrectedthe controversyconcerning“high-frequencytrading”(HFT)inthe stockmarkets.WhileHFThasbeenimportantinthestock marketsforaboutadecade,andmayhavealreadypeakedin termsofitseconomicsignificance,ittouchedanervewitha publicsuspiciousoffinancialinstitutionsinthewakeofthe financialcrisisof2008-2009.Inreality,HFTisnotonething, butawidearrayofpracticesconductedbytechnologicallyadept electronictraders.Someofthesepracticesarebenign,andsome evenbringbenefitssuchasliquidityandimprovedprice discoverytofinancialmarkets.Ontheotherhand,thereare legitimategroundsforthecommonlyheardcomplaintthat “HFTisnotfair.”CertainHFTpracticessuchasco-location, flashorders,andenricheddatafeedscreateatwo-tiered financialmarketplace,whileotherpracticessuchasmomentum ignition,spoofing,andlayeringaremerelyhigh-techversionsof traditionalmarketmanipulation.Finally, thecreationof specialordertypessuchas“HideNotSlide”showsthe exchangesallowingtheirHFTclientstojumpthequeueofpricetimepriorityembeddedinRegulationNMSandstockmarket practice.Whilethecommonly-usedtechniqueofacost-benefit analysisleadstoequivocalorindeterminateresultswhen appliedtoHFTtradingactivityincomplexandoftenopaque

©2016StevenMcNamara

*AssistantProfessor,BusinessLaw,OlayanSchoolofBusiness,the AmericanUniversityofBeirut;BA,St.John’sCollege;Ph.D.,BostonCollege; J.D.,ColumbiaUniversitySchoolofLaw.Theauthormaybecontactedat [email protected],participants atthe2015meetingoftheCanadianLawandEconomicsAssociation,andhis AUBcolleaguesIbrahimJamali,GeorgeNamur,andNeilYorke-Smithfor discussionandcommentsduringthecourseofpreparationoftheArticle. ThanksalsotoLamiaHarizforresearchassistance.Allerrorsarehisown.

markets,amorebasicethicoffairnessgroundedincommonly acceptedrulesoffinancialmarketbehaviorillustratesthat certainHFTpracticesareindeedunfair.ThisArticledrawson thelegal,financial,andbusinessethicsliteraturestoillustrate exactlyhowcertainformsofHFTareunfair,andproposesfour coreprinciplestoguideHFTactivityanditsregulation.

I.

II.

TheEcosystemoftheEquityMarkets.............................77 A.ThePre-RegulationNMSLandscape........................78 B.RegulationNMS.........................................................84 C.OverviewoftheCurrentEquityMarkets.................92 ProblematicVarietiesofHighFrequencyTrading.........98 A.TheOriginsofHFT:AlgorithmicTrading................99 B.Co-location................................................................101 C.FlashOrders............................................................105 D.DirectandEnrichedDataFeeds.............................110 E.Spoofing,Layering,QuoteStuffing,etc.:HFTas

MarketManipulation..............................................114

F.SpecialOrderTypes:“HideNotSlide,”ISOs,and

JumpingtheQueue.................................................117 III.AnEthicsofAlgorithmicTrading..................................121 A.CostsandBenefitsofHFT.......................................127 B.TheUltimatumGame,Reciprocity,andthe

DutchBook...............................................................134 C.FreedomfromMisrepresentation............................141 D.ALevelPlayingField...............................................143 InstitutionalIntegrity.............................................145 E. Conclusion.................................................................................150

INTRODUCTION

DuemainlytothepublicationofMichaelLewis’sFlash Boys,1theAmericanpublicisonceagainconcernedwiththe questionofthefairnessofthefinancialmarketplace.Lewis’s specificchargeisthatthestockmarketisa“rigged”2game, withhigh-frequencytraderspreyinguponmoretraditional “MainStreet”investorsandtheinstitutionalinvestorsthat servethem.Theallegationthatthemarketsareriggedledtoa

1.MICHAELLEWIS,FLASHBOYS:AWALLSTREETREVOLT(2014). 2.See,e.g.,id.at40(whenprotagonistBradKatsuyamarealizeshe cannotbuyorsellstocksatthepricesontheconsolidatedtickerfeed,he “realizedthatthemarketsarerigged”).

swiftdenialbyMaryJoWhite,theSecuritiesandExchange Commission(SEC)Chairconcernedwithpolicingthem.3Butin thewakeofthefinancialcrisisof2008andtheslowrecovery fromthe“GreatRecession,”theaccusationthatWallStreet insiderswereprofitingatthegeneralpublic’sexpensequickly gainedtraction.4

SECChairWhiteandotherregulatorshaveanimportant roletoplayinfosteringtrustinthefinancialsystem.However, acloselookatanumberofpracticesofthehighfrequency tradersandtheexchangesthatcatertothemsuggeststhe marketsare,ifnot“rigged,”subjecttounfairanddeceptive practicescarriedoutbysomeofthenewhigh-frequencytrading (HFT)marketmakers.WhilethetotalprofitsextractedbyHFT arenotgreatbyWallStreetstandards5andhavebeenin

Exchange

3.MaryJoWhite,Chairwoman,Sec.&Exch.Comm’n,EnhancingOur EquityMarketStructure,AddressBeforeSandlerO’Neill&Partners,L.P. Global 2014), http://www.sec.gov/News/Speech/Detail/Speech/1370542004312#.U7ta8_ldWS q(statingthat“[a]llofthesemarketqualitymetricsshowthatthecurrent marketstructureisnotfundamentallybroken,letalonerigged”afterciting dataonthecostofexecutinglargeorders,marketvolatility,andbid-ask spreads).

Conference

Brokerage

4.See,e.g.,HarryBruinius,

‘FlashBoys’ReignitesDebate:IsHighFrequencyTradingaDigitalAgeMenace?,CHRISTIANSCIENCEMONITOR(Apr. 2,2014),http://www.csmonitor.com/Business/2014/0402/Flash-Boys-reignitesdebate-Is-high-frequency-trading-a-digital-age-menace;StephenGandel,Isthe SECProtectingHigh-FrequencyTraders?,FORTUNE(Mar.31,2014), http://fortune.com/2014/03/31/is-the-sec-protecting-high-frequency-traders/; JohnLanchester,Scalpers,Inc.,36LONDONREV.BOOKS7–9(June5,2014), http://www.lrb.co.uk/v36/n11/john-lanchester/scalpers-inc;MatthewO’Brien, EverythingYouNeedtoKnowAboutHigh-FrequencyTrading,ATLANTIC(Apr. 11,2014),http://www.theatlantic.com/business/archive/2014/04/everythingyou-need-to-know-about-high-frequency-trading/360411/. critical appraisalofLewis’sbook,seeFelixSalmon,TheLewisEffect,SLATE:BOOK REV., http://www.slate.com/articles/business /books/2014/04/michael_lewis_s_flash_boys_about_high_frequency_trading_re viewed.html (arguingthatthestockmarkethasneverbeenfair,and transactioncostsarefarsmallerthaninthepast).Thecontroversyprompted SenatorCarlLevin,ChairoftheSenatePermanentSubcommitteeon Investigations,toholdhearingsonHFT.SeeConflictsofInterest,InvestorLoss ofConfidence,andHighSpeedTradinginU.S.StockMarkets:HearingBefore thePermanentSubcomm.onInvestigationsoftheS.Comm.onHomelandSec. &GovernmentalAffairs,113thCong.(2014).

(Apr.

For

7,

a

5.TheTabbGroupestimates2014profits

forHFTfirmsas approximately$1.3billion,downfromanestimated$7.2billionin2009.See LarryTabb,No,MichaelLewis,theU.S.EquitiesMarketIsNotRigged,TABB F. (Mar.31,2014),http://tabbforum.com/opinions/no-michael-lewis-the-usequities-market-is-not-rigged.

declineinrecentyears,6anumberofcharacteristicsofHFT suggesttheyconstitute,tosomedegreeatleast,arent extractedfromlong-terminvestors.7ThemostsuccessfulHFT firmsexperiencelittleornotradinglossforyearsonend,8and thecentralplankusedtojustifytheiractivities,thatthey provideliquiditytothecurrentcomputerizedmarketplace,is subjecttoimportantqualifications.9Thatsaid,theargument thatHFTisdoingsomethingwrongrequiresanethicsthat rejects“caveatemptor”asaguidingprincipleforfinancial marketstransactions.Sincethemostimportantstructural elementsthatenableHFTactivity—specializedordertypes, enricheddatafeeds,and“co-location”10—are,atleastin principle,openforusebyall,theexchangesanddefendersof HFTcanarguethereisnothingintrinsicallywrongwiththeir conduct.Theclashofperspectivesresultsfromthefactthat, and practically

specialized

speaking,

highly

only

Fall,

6.SeeMatthewPhilips,HowtheRobotsLost:High-FrequencyTrading’s 2013),

Rise http://www.bloomberg.com/bw/articles/2013-06-06/how-the-robots-lost-highfrequency-tradings-rise-and-fall(HFTpeakedinsharevolumein2008-2011); seealsoDonaldMacKenzie,ASociologyofAlgorithms:High-Frequency TradingandtheShapingofMarkets29–30(June2014) (unpublished manuscript), http://www.sps.ed.ac.uk/__data/assets/pdf_file/0004/156298/Algorithms25.pdf (notingthatprofitsofHFTfirmsappeartohaveshrunktoapproximately 1/20thofacentpertrade,althoughtherearereasonstobeskepticaloflow estimatesofprofitability).ForanearlydiscussionofthelimitationsofHFT, seeRichardBookstaber,TheArmsRaceinHighFrequencyTrading,RICK BOOKSTABERBLOG(Apr.21,2009),http://rick.bookstaber.com/2009/04/armsrace-in-high-frequency-trading.html.

7.Seeinfranote219andaccompanyingtext. 8.DaveCummings,ownerofKansasCity-basedTradebot,reportedly toldagroupofbusinessschoolstudentsthathisfirmhadnotlostmoneyona singledayoverafour-yearperiod.SeeStephenGandel,IsKCFirmtheNew KingofWallStreet?,TIME(May18,2010),http://business.time.com/2010/05 /18/is-kc-firm-new-king-of-wall-street/.Similarly,VirtuFinancialdisclosed thatithadexperiencedonlyonedayoftradinglossesoverafive-yearperiod. SeeSamMamudi&LesliePicker,VirtuFilingShinesLightonHigh2014), Frequency http://www.bloomberg.com/news/articles/2014-03-11/virtu-filing-shines-lighton-business-of-high-frequency-trading;cf.MacKenzie,supranote6,at30 (statingthatoneintervieweeestimatedthathisfirmprofitedonfifty-three percentofitstrades,typicallyforaonecentprofit).

Trading,

11,

9.Liquidityprovisionandpricediscoveryarethetwomostfrequently citedbenefitsHFTbringstothefinancialmarkets,andHFTfirmsappearto providesignificantliquiditytothestockmarkets.Foradiscussionofvarious costsandbenefitsHFTbringstofinancialmarkets,seeinfraSectionIII.A.

10.SeeinfraPartIIforadetailedexplanationofthevarietiesandspecific

structuralelementsofHFTactivity.

firms

cantakeadvantageof

technologicallyadept the opportunitiesofferedbytheexchanges,andthetrading strategiestheyenableallowfor,atleastforsomeHFTfirms, thenearlyrisk-freeextractionofprofitsattheexpenseof slowertraders.Furthermore,anumberofthespecializedorder typesusedbytradersappeartohavebeenspecificallycreated bytheexchangesandHFTshopstoallowtechnologicallyadept playerstopreyuponslower,unawaretraders.11

ThisArticleaimstoprovideacloserlookatthefeaturesof HFTthathavecausedcriticstocomplainthatitisunfair.12In particular,itcomestothefollowingconclusions:

11.TheallegationthatHFTfirmsandtheexchangescolludedtocreate techniquestoenableHFTtoprofitattheexpenseofslowertradersinthe marketisoneofthemostdisturbingchargesleveledagainstHFT.SeeSCOTT PATTERSON,DARKPOOLS:THERISEOFTHEMACHINETRADERSANDTHE RIGGINGOFTHEU.S.STOCKMARKET204–05(2013) (“High-speedfirms workedhandinhandwiththetradingnetworkstocreateexoticordertypes thatwouldbehaveinveryspecificways.”);accordLEWIS,supranote1,at163 (“BygivingHFTwhatitwanted(speed,inrelationtotherestofthemarket; complexityonlyHFTunderstood;andpaymenttobrokersfortheircustomers’ orders,sothatHFThadsomethingtotradeagainst),thenewstockexchanges hadstolenmarketsharefromtheoldstockexchanges.”);id.at171(explaining thatthepurposeofthespecialordertypeswas“tocreateanedgeforHFTat theexpenseofinvestors”);seealsoinfranotes263–82andaccompanyingtext.

12.ThisArticleaimstoprovideadetailedlookatthepotentially objectionablepracticesofHFTwithparticularattentionpaidtofairness issues.Indoingso,itpresentsaconsiderablymorenegativeoverallevaluation ofHFTthanearliercommentatorsinthelegalliterature.See,e.g.,Charles Korsmo,HighFrequencyTrading:ARegulatoryStrategy,48U.RICH.L.REV. 523,528,562(2014)(hedgingagainstcriticismofHFTbystatingthatithas generatedan“appearanceofunfairness”);TomC.W.Lin,TheNewFinancial (“Whileactualandpotential Industry,65ALA.L.REV.567,595(2014) challengespresentedbycy-fi[cyberfinance]aremany,serious,andreal,so areitsactualandpotentialbenefits.”(internalfootnoteomitted ;MiHyun Yoon,TradinginaFlash:ImplicationofHigh-FrequencyTradingfor SecuritiesRegulatorsWorldwide,24EMORYINT’LL.REV.913,943(2010) (arguingthattheSECshouldregulateonlycertainHFTstrategiesthatcreate systemicrisk);MichaelJ.McGowan,Note,TheRiseofComputerizedHigh FrequencyTrading:UseandControversy,2010DUKEL.&TECH.REV.,¶31 (2010)(“Despitethesereservations,thereisstillevidencetosuggestthatHFT benefitsthemarketsandcurbingitsusewouldbetothedetrimentofall investors.”).Morerecentarticlesinthelegalliteraturehavecastamore criticaleyeonHFT.SeegenerallyMerrittB.Fox,LaurenceR.Glosten& GabrielV.Rauterberg,TheNewStockMarket:SenseandNonsense,65DUKE L.J.191,196(2015)(proposingatripartiteframeworkofadverseselection, principal-agentproblems,andmultiplevenuetradingtoanalyzeHFT);Frank Pasquale,Law’sAccelerationofFinance:RedefiningtheProblemofHighFrequencyTrading,36CARDOZOL.REV.2085,2087(2015)(drawingonthe “LegalTheoryofFinance”toarguethatbenefitsfromever-fastertradingmay beillusory);GregoryScopino,The(Questionable)LegalityofHigh-Speed

HFT,likemanyotherfinancialmarketactivities,is complex,andthetermencompassesanarrayof specificpractices.SomeHFTactivityismorally objectionable,whileothercommonHFTactivities aremorallyneutralorevenbeneficial.

WhilethedollarsatstakeinHFTmaybedeclining, isasymbolicallyimportantproblemthat it ordinary resonateswithequity Americans,particularlythosewhoinvestinthe financialmarketsthrough401(k)andothertypesof retirementplans.Regulatorsandfinancialmarkets participantsthereforeoughttotreatitcarefully.

concerns

Anaccurateevaluationof

but

alone,

thefairnessofa marketplaceactivitycannotbeunderstoodfrom quantitativemeasures requires understandingthestructuralinteractionofmarket participantsintermsoftheirexpectationsand beliefswhenenteringintoatransaction.This perspectivehighlightsarolefordeontologicalmoral theoryinfinancialregulationbecauseapurelycosttheinternal benefitanalysiswilloftenmiss perspective inthe system.

stakeholders

important

Theinvasion,andnowdominance,ofthefinancial the marketsbytheinformationtechnologyof digitalageappearstounderminethegrounding assumptionsofRegulationNationalMarketSystem (RegulationNMSorReg.NMS).RegulationNMS maycometobeseenasanunstableandultimately unworkableregimebetweentheworldofthe1934 Actandwhatcomesnextasinformationtravelsat thespeedoflight.Inhindsight,warningsoverthe dangeroffragmentationduringthedebatesonthe implementationofa“nationalmarketsystem” appearprescient.

“Pinging”and“Front-Running”intheFuturesMarkets,47CONN.L.REV.607, 616–17(2015) (arguingthathigh-speedpingingshouldbeviewedasa deceptivetradingpracticeundertheCEA);YeshaYadav,HowAlgorithmic TradingUnderminesEfficiencyinCapitalMarkets,68VAND.L.REV.1607, 1644–57(2015)(arguingthatHFT’simpactonpricesharmstheallocative efficiencyofthestockmarket).

I.THEECOSYSTEMOFTHEEQUITYMARKETS

Aswithmostotherfeaturesofcontemporarylife,stock market transactionshavebeenradicallychangedbythe informationtechnologyrevolutionoverthecourseofthepast twentyorsoyears.Lessthantenyearsago,significanttrading tookplaceontheflooroftheNewYorkStockExchange (NYSE),buttheimplementationofRegulationNMSin2007 quicklytransformedthefloorintolittlemorethanabackdrop fortelevisionstoriesrelatingtotheeconomy.13Thechanges forcedbyregulatorswereintendedtobenefittheaverage investor,andthenewrulessucceededinimportantways.14At thesametime,however,theycreatedasignificantlynew financialecosysteminwhichabusiveactivityhasthrived.The interactionofRegulationNMSwithatechnologicalarmsrace hasgivenrisetoequitymarketsthatruntheriskofbecominga two-tieredsystem.

served

theirmember-owners

Overthecourseofthepasttwentyyears,thestock exchangeshavebeentransformedfrominstitutionsthat primarily for-profit corporationsservingtheinterestsofshareholders.15Atthe sametimethecorporateformtheyoperatedunderchanged,the economiceffectsofthenewregulatorystructuretransformed theircustomerbase.16Manybroker-dealerswentoutof business,andonlysix“DesignatedMarketMakers”17(DMMs)

into

13.SeeJerryW.Markham&DanielJ.Harty,ForWhomtheBellTolls: TheDemiseofExchangeTradingFloorsandtheGrowthofECNs,33J.CORP. L.865,866(2008).

Eliminating

14.SeeSteveWunsch,HFTandtheComputerizationofWallStreet,Part 2013), 3: http://tabbforum.com/opinions/hft-and-the-computerization-of-wall-street-part3-eliminating-the-middleman.

theMiddleman,

(July

31,

F.

15.SeeAndreasM.Fleckner,StockExchangesattheCrossroads,74 FORDHAML.REV.2541,2593(2006);JonathanR.Macey&MaureenO’Hara, FromMarketstoVenues:SecuritiesRegulationinanEvolvingWorld,58STAN. L.REV.563,585(2005)(“Asexchangesconverttobecomeprofit-seeking firms...thegainsofthefewturnintoprofitsofthecorporation.”).See generallyJOHNC.COFFEE,JR.&HILLARYA.SALE,SECURITIESREGULATION 572–76(13thed.2015).

16.SeeMacey&O’Hara,supranote15,at581,593. 17.TheseareBarclaysCapitalInc.;BrendanE.Cryan&Co.,LLC;IMC FinancialMarkets;J.Streicher&Co.,LLC;KCG;andVirtuFinancialCapital EXCH., Markets, www.nyse.com/market-model/overview(lastvisitedOct.13,2015).Thereare alsosevenfirmsdesignatedas“SupplementalLiquidityProviders”:Citadel SecuritiesLLC;Goldman,Sachs&Company;HRTFinancialLLC;KCG

LLC.MarketModel

Overview,N.Y.

servetheNYSEintheroleformerlyexercisedbyspecialistson theexchange.18Insteadofservingtheinterestsofarelatively largegroupofmembersholdingseatsontheexchange,who werepreviouslycompensatedforbylargespreads,19the businessmodelnowfollowedbytheexchangesispredicatedon chargingrelativelysmallfeesfornumeroustransactionsand sellingtheinformationaboutsuchtransactions.20Inthenew environmentwhereexchangescompeteforfees,anymarket participantbringingalargevolumeoftradestoanexchange willbeavaluableandfavoredclient.

A.THEPRE-REGULATIONNMSLANDSCAPE

ThehistoryofthestockmarketsintheUnitedStates stretchesbackto1792,whenstocktradingwassaidtohave beenorganizedunderabuttonwoodtreeonWallStreet.21The “ButtonwoodAgreement”wassignedbytwenty-fourbrokers.22 Itspecifiedthatbrokerswouldtradeonlyamongstthemselves, withaminimumcommissionrateof0.25%.23Thesetwo featuresoftradingonwhatbecametheNewYorkStock Exchangeformedthecornerstoneofthetradingsystemthat persisteduntilthetwenty-firstcentury.24TheNewYorkStock Exchangegrewinthetwentiethcenturyintothemost

AmericasLLC;LatourTrading,LLC;TradebotSystems,Inc.;andVirtu FinancialBDLLC.Id.

(May

18.SeeRyanC.Fuhrmann,DeclineoftheIndependentBroker-Dealer, http://www.investopedia.com 2013),

INVESTOPEDIA /articles/professionals/050613/decline-independent-brokerdealer.asp (discussinghowtheestimated“numberofbroker-dealersregisteredwith FINRAfellbynearly10%between2005and2010,atrendthathascontinued overthepastcoupleyears”).

19.SeegenerallyMarkham&Harty,supranote13(describingthedemise

oftraditionalexchangetradingfloors). 20.SeeBruinius,supranote4. 21.SeeCHARLESR.GEISST,WALLSTREET:AHISTORY13(1997);Stuart Banner,TheOriginoftheNewYorkStockExchange,1791-1860,27J.LEGAL STUD.113,115(1998).SeegenerallyHansR.Stoll,ElectronicTradinginStock Markets,20J.ECON.PERSPS.153,153(2006).Forahistoryofthestock seeJOELSELIGMAN,THE marketsandtheSECsincethe1930s, TRANSFORMATIONOFWALLSTREET:AHISTORYOFTHESECURITIESAND EXCHANGECOMMISSIONANDMODERNCORPORATEFINANCE(3ded.2003).

22.SeeBanner,supranote21,at115n.3(citingPeterEisenstadt,How theButtonwoodTreeGrew:TheMakingofaNewYorkStockExchange Legend,19PROSPECTS75,75(1994 .

23.SeeEisenstadt,supranote22,at77. 24.Cf.Banner,supranote21,at114(chroniclingtheNYSE’sgrowth

fromitsButtonwoodbeginningstothecenteroftheAmericaneconomy).

importantvenueforstocktradingintheUnitedStates.25While theexchangewasopentooftenjustifiedcriticismthatit benefitteditsmembersattheexpenseofthepublic,itdid presentapredictablesystemforbuyingandsellingequityin public theSecurities ExchangeActof1934(theExchangeAct)weretoalarge degreeeffectiveinsuppressingtheworstabuses.26

companies,andregulationsunder

KeyaspectsoftradingstocksonanexchangeinthepreRegulationNMSlandscapeincludedboththefactthattrading inastocktendedtogravitatetoonevenueandtherelatively highcostoftrading.Tradingwaslimitedtobroker-dealers holdingseatsontheexchange,whotradedinthecontextofthe specialistsystem.27Inadditiontothememberswhowere allowedtotradeonbehalfoftheirclients,specialistssupported tradinginNYSE-listedstocks,with“482individualspecialists attheNewYorkStockExchange”28coveringanaverageof threetosixstockseachdayintheearly2000s.29Aspecialistin acompany’sstockhadtheobligationtosupporttradingwhen

25.SeeEisenstadt,supranote22,at75(“[B]etween1865and1920... theexchange[NYSE]emergedasthedominantcapitalmarketintheworld.”). 26.See,e.g.,FrankB.Cross&RobertA.Prentice,TheEconomicValueof SecuritiesRegulation,28CARDOZOL.REV.333,337(2006)(“[E]mpirical evidenceplainlydemonstratesgreateconomicvaluefromourrelatively stringentsystemofgovernmentalsecuritiesregulation.”);StevenA.Ramirez, TheVirtuesofPrivateSecuritiesLitigation:AnHistoricandMacroeconomic Perspective,45LOY.U.CHI.L.J.669,684(2014)(“[The]originalconceptionof thefederalsecuritieslawsusheredinanunprecedentederaoffinancial stability.”).Fortheviewthatitisenforcementoflawthathasbeneficial effectsforamarketeconomy,seeJohnC.Coffee,Jr.,LawandtheMarket:The ImpactofEnforcement,156U.PA.L.REV.229,230(2007) (“[H]igher enforcementintensitygivestheU.S.economyalowercostofcapitaland highersecuritiesvaluations.”).SeealsoHowellE.Jackson&MarkJ.Roe, PublicandPrivateEnforcementofSecuritiesLaws:Resource-BasedEvidence, 93J.FIN.ECON.207,227–30(2009)(arguingthatpublicenforcementof financialmarketslawispositivelyassociatedwithdeepercapitalmarkets thanprivateenforcement).Foraskepticalcounterargument,seeGeorgeJ. Benston,RequiredDisclosureandtheStockMarket:AnEvaluationofthe SecuritiesExchangeActof1934,63AM.ECON.REV.132,153(1973).

27.SeegenerallyLARRYHARRIS,TRADING&EXCHANGES:MARKET MICROSTRUCTUREFORPRACTITIONERS494–513(2003);SELIGMAN,supranote 21,at335–44(describingthefunctionandhistoryofspecialists);GideonSaar, SpecialistMarkets,inTHEENCYCLOPEDIAOFQUANTITATIVEFINANCE(Rama Conted.2010);NicholasWolfson&ThomasA.Russo,TheStockExchange Specialist:AnEconomicandLegalAnalysis,1970DUKEL.J.707,707–08 (1970)(discussingspecialists’rolesandhistory).

28.HARRIS,supranote27,at496. 29.

liquiditydriedup;30abouttenpercentoftradesweresupported byspecialistseitherpurchasingorsellingstock.31Andwhile specialistscouldseetheordersinthe“orderbook,”andthus hadknowledgeofwherethemarketwasheaded,32“frontrunning”instockswasforbidden.33Inthe“openoutcry”auction system,tradingoccurredatthepostonthefloorofthe exchangewhereaspecialistwaslocated,withbrokersshouting bidsorofferstooneanother.34Intheclassictwentiethcentury formoftheexchange,tradinggravitatedtotheexchange becausebrokerswantedaccesstothegreatest trading volume.35Inthisway,stockexchangesheldontothevastbulk ofthetradinginstockslistedontheirexchanges.36

Theothercentraleconomicfeaturewasthecostoftrading. Sincestockswerelistedin1/8thofadollarincrements,spreads

30.Seegenerally17C.F.R.§240.11b-1(a)(2)(ii)(2015)(outliningtherole ofspecialistsinthestockmarket).Specialists,however,hadacertainamount ofdiscretioninmeetingthisobligation.SeeWolfson&Russo,supranote27,at 711–12.

31.SeeNanS.Ellis,LisaM.Fairchild&HaroldD.Fletcher,TheNYSE theFailureofSelf-

ResponsetoSpecialistMisconduct:AnExampleof Regulation,7BERKELEYBUS.L.J.102,107n.20(2010).

32.SeeGeraldT.Nowak,Note,AFailureofCommunication:An ArgumentfortheClosingoftheNYSEFloor,26U.MICH.J.L.REFORM485, 501–02(1993)(presentingtheeconomicbenefitsaccordedtospecialists).

IndustryRegulatoryAuthority(FINRA)

33.Front-runningbyspecialistswasacommonpracticeintheyears beforetheExchangeAct.SeeGEISST,supranote21,at213(“[S]pecialistswere inaprivilegedpositiontoseepricesbeforeexecutingforthepublicandwould oftenactforthemselvesbeforefillinganorderfromthepublicbeingexecuted throughafloorbroker.”).Althoughfront-runningisnotexplicitlyprohibited bythefederalsecuritieslaws,therulesofthevariousexchangesandthe is Financial prosecutedasaviolationofSection10(b)oftheExchangeAct.SeeFoxetal., supranote12,at227n.87;DavidM.Bovi,Rule10b-5LiabilityforFront Running:AddingaNewDimensiontothe“MoneyGame,”7ST.THOMASL. REV.103,111–14(1994).Fordiscussionoffront-runningandothermisconduct byspecialists,seeEllisetal.,supranote31,at115–17(describingfrontrunningaswellasotherformsofspecialistmisconduct).Forananalysisofthe specialistsystem,seeJonathanMacey&HidekiKanda,TheStockExchange asaFirm:TheEmergenceofCloseSubstitutesfortheNewYorkandTokyo StockExchanges,75CORNELLL.REV.1007,1025–34(1990).

forbiditandit

34.SeeLeoMelamed,TheMechanicsofaCommodityFuturesExchange: ACritiqueofAutomationoftheTransactionProcess,6HOFSTRAL.REV.149 (1977)(presentinganoverviewanddefenseofopenoutcrytrading).

35.SeeCraigPirrong,CompetitionandVerticalIntegrationinFinancial Exchanges,7COMPETITIONPOL’YINT’L90,93(2011)[hereinafterPirrong, CompetitionandVerticalIntegration](explainingtheliquiditynetworkeffect forexchanges).

36.Seeid.at94(reportingthattheNYSE’smarketshareplungedfrom

85%toapproximately30%aftertheintroductionofReg.NMSin2007).

onstocktradeswereaminimumof12.5cents,with25cents beingmostcommonand50centsormorealsopossible.37The fixedpricesin1/8thincrementspreservedfatprofitsformarket makers,astheircompensationformakingatradewasthe spreadbetweenthebidandtheask,orthepricesabroker wouldpayandsellforastock,respectively.38Whilethesehigh costsrepresentedasignificantrentextractedfromthebrokers’ clients,theywerealsotransparent,39indistinctiontothe stockbrokers presentmarkets. traditionallychargedhighfixedcommissions toclients, althoughfixedcommissionsweredealtadeathblowwiththe amendmentsmadetotheExchangeActin1975.40Despitethis transparency,thecostoftradingwasacentralmotiveforthe implementationofa“NationalMarketSystem”thatbeganwith theenablinglegislationin1975andfinallyledtothe promulgationofRegulationNMSin2005.41

Inadditionto

spreads,

Thegravitationoftradingtoexchangesonwhichstocks werelistedandthehightransactioncostschargedtocustomers weretwomainfeaturesofthepre-RegulationNMStrading environment.Inhindsight,itappearsthatforinvestorsthe

37.SeeStoll,supranote21,at160,164. 38.SeeDIV.OFMKT.REG.,U.S.SEC.&EXCH.COMM’N,MARKET2000:AN EXAMINATIONOFCURRENTEQUITYMARKETDEVELOPMENTSIV-9(1994) [hereinafterMARKET2000](“[M]arketmakersbenefitbecausetheycanaffect executionsatwiderspreadsthanifasmallervariationwasused.”);seealso infranotes54–55andaccompanyingtext.

39.Seeid.at17,IV-1. 40.SeeStanislavDolgopolov,InsiderTrading,ChineseWalls,and BrokerageCommissions:TheOriginsofModernRegulationofInformation FlowsinSecuritiesMarkets,4J.L.ECON.&POL’Y311,324n.62(2008).The SecuritiesActsAmendmentsof1975prohibitedfixedcommissionsonthe stockexchanges.SecuritiesActsAmendmentsof1975,Pub.L.No.94-29,§4, 89Stat.97,107–09;seealsoSELIGMAN,supranote21,at473–86(“The practicaleffectoftheCommission[’s]decision...wastoaddtothealready considerablepressurestounfixcommissionratestotally.”).

41.SeeJonathanR.Macey&DavidD.Haddock,ShirkingattheSEC: TheFailureoftheNationalMarketSystem,1985U.ILL.L.REV.315,322 (1985)(notingtheobjectivesbehindthenationalmarketsystemlegislation were“toassureeconomicallyefficientsecuritiestransactions;faircompetition betweenbrokersanddealers,andbetweenexchangesandmarkets;the availabilityofinformation;theexecutionofinvestors’ordersinthebest market;andanopportunitytoexecuteorderswithouttheparticipationof dealers,”which“reflectExchangeAct§11A”);seealsoRegulationNMS,70 Fed.Reg.37,496,37,497(June29,2005)(codifiedat17C.F.R.pt.242)(“[T]he NMSisdesignedtoachievetheobjectivesofefficient,competitive,fair,and orderlymarketsthatareinthepublicinterestandprotectinvestors.”); SELIGMAN,supranote21,at511–14.

formeraspectwasonthewholebeneficial42whilethelatter waslargelydetrimental.43(Ofcourse,otherpartieswithinthe system,suchasstockbrokers,wouldhavedifferentviews.44) Whilemanymarketparticipantsandscholarswereconcerned aboutthedangersoffragmentationasaresultoftheeffortto introducecompetitioninthemarketsystem,whetherornot suchfragmentationwouldoccur,anditseffectifitdid,wasthe subjectofintensedebate.45Inretrospect,thepushto“reform” themarketsystemthatbeganinthe1960sand1970s,in conjunctionwiththeriseofcomputertechnology,ledtothe current,highlyfragmentedstateofthemarketswithother, newdysfunctions.46

Inthepre-RegulationNMSperiod,stockmarketswerefor themostpartauctionmarketsrunonan“openoutcry system.”47Inthisformat,transactionsweredrivenbyorders conveyedtostockbrokerstoeitherbuyorsell.48TheNational

42.

Increasedliquidityandpricecompetitionarethemainbenefitsthat economictheorywouldpredictfollowfromtheliquiditynetworkeffect.See Pirrong,CompetitionandVerticalIntegration,supranote35,at93;Stephen CraigPirrong,TheSelf-RegulationofCommodityExchanges:TheCaseof [hereinafter MarketManipulation,38J.L.&ECON.141,154–55(1995) Pirrong,TheSelf-RegulationofCommodityExchanges](discussingnatural monopolyamongcommoditiesexchanges);seealsoMarcoPagano,Trading VolumeandAssetLiquidity,104Q.J.ECON.255,255(1989).Seegenerally AnatR.Admati&PaulPfleiderer,ATheoryofIntradayPatterns:Volumeand PriceVariability,1REV.FIN.STUD.3(1988)(examiningforcesthatleadto concentrationoftradesatcertainperiodsoftheday). 43.Seesupranote40andaccompanyingtext. 44.Differentplayersinthemarketwillhavedifferentideasconcerning theidealmarketstructure.SeeCOFFEE&SALE,supranote15,at577;seealso RegulationNMS,70Fed.Reg.at37,500(notingthatatpointstheinterestsof short-termtradersandlong-terminvestorsmaydiverge).Seegenerally ConceptReleaseonEquityMarketStructure,ExchangeActReleaseNo.3461358,75Fed.Reg.3594,3597(Jan.14,2010)(notingthatitcanbedifficult toreconcilethefiveobjectivessetforthbyCongressinExchangeAct§11A).

45.See,e.g.,MARKET2000,supranote38,atIII-2(arguingthatU.S. equitymarketsasof1994werenot“fragmentedtothepointthatprice discoveryandliquidityhavebeenaffectedadversely”);RobertaS.Karmel, TurningSeatsintoShares:CausesandImplicationsofDemutualizationof StockandFuturesExchanges,53HASTINGSL.J.367,392–95(2002);Mark Klock,TheSEC’sNewRegulationATS:PlacingtheMythofMarket FragmentationAheadofEconomicTheoryandEvidence,51FLA.L.REV.753, 760–64(1999);ArthurLevitt,DynamicMarkets,TimelessPrinciples,2000 COLUM.BUS.L.REV.1,9–11(2000);Macey&Haddock,supranote41,at341– 42(characterizingfearsoverfragmentationas“unwarranted”).

46.Cf.Karmel,supranote45,at404–05(examiningdiscussions,

proposals,andactualchangesmadethroughthe1960sand1970s).

47.SeeMelamed,supranote34,at150. 48.Seeid.

AssociationofSecuritiesDealersAutomatedQuotations (NASDAQ)systemstandsasanexception,whichthroughits influenceonequitytradinggaverisetothecurrentstock marketecosystem.NASDAQrepresentedafundamentallynew modeofequitytradingbothinitsuseoftechnologyandits format.49TheNationalAssociationofSecuritiesDealersformed NASDAQtoprovideitsmemberswithawaytotradestocks thatwouldbedifficulttolistontheNYSEorothermajor exchanges.50Thesestocksweretraditionallyissuedbysmaller start-ups,oftentechnologyventures,thatwouldhavedifficulty meetingtheNYSElistingrequirements.51TheNASDAQ tradingsystemdifferedfromtheNYSEandotherstock marketsthatusedtradingfloors.Itusedaquote-driven system,wherebrokersputupquotestobuyorsellstocksonan electronicbulletinboard,andwaitedforcounterpartiesto them.52Italsoincorporatedtheuseofcomputer accept technology,andsowastheprecursortotoday’selectronic tradingsystems.53

FormanyyearsNASDAQwassimilartotheNYSEinthat itpreservedtheabilityofdealerstoearnsubstantial commissionsthroughlargespreads,butaregulatorychangein 1996representedabreakinthebulwarkprotectingdealers’ easyprofits.In1994,financeprofessorsWilliamChristieand PaulSchultzpublishedWhyDoNASDAQMarketMakers AvoidOddEighthQuotes?54Aftersurveyingpricesinheavily

49.SeeKorsmo,supranote12,at533(notingthatNASDAQwasfirstto usecomputerstotradein1971);Markham&Harty,supranote13,at899; PeterGomber,BjörnArndt,MarcoLutat&TimUhle,HighFrequency Trading manuscript), http://ssrn.com/abstract=1858626.

2011)

50.SeeMarkham&Harty,supranote13,at877. 51.SeeMichaelJ.Simon&RobertL.D.Colby,TheNationalMarket SystemforOver-the-CounterStocks,55GEO.WASH.L.REV.17,35(1986) (characterizingtheoriginsoftheNASDAQsystemas“realtimepinksheets”). SeegenerallyOverviewofNYSEQuantitativeInitialListingStandards, NYSE, https://www.nyse.com/publicdocs/nyse/listing/NYSE_Initial_Listing_Standard s_Summary.pdf requirementsfortheNYSE).

(lastvisitedNov.18,2015)

(presentinggeneral

listing

52.SeeStoll,supranote21,at159–60. 53. 54.WilliamG.Christie&PaulH.Schultz,WhyDoNASDAQMarket MakersAvoidOdd-EighthQuotes?,49J.FIN.1813(1994) [hereinafter Christie&Schultz,WhyDoNASDAQMarketMakers];seealsoWilliamG. Christie&PaulH.Schultz,DidNasdaqMarketMakersImplicitlyCollude?,9 J.ECON.PERSPS.199(1995);Stoll,supranote21,at159–60.

tradedNASDAQstocks,theysuggestedthatbrokerscolluded tokeepspreadswidethroughavoidingquotesinoddeighths, favoringeveneighthquotesinstead.55Theirpapergaveriseto significantcontroversy,56andasaresult,thenSECChair ArthurLevittchampionedorderhandlingrules,whichrequired dealerstoincludeintheirownquotationscustomerlimitorders improvingontheirownquotes.57Whilepriceswerestillquoted in1/8thincrements,nowacustomer’sbidoroffercould investor-friendly improve regulationwasastepinthedirectionoftakingcontrolaway fromtheintermediariesinthesystem,thestockbrokers,and givingittotheend-users,theinvestors.

onitsbroker’squote.58This

B.REGULATIONNMS

RegulationNationalMarketSystem,whichbecame effectivein2007,hasitsgenesisinSection11Aofthe ExchangeAct,entitled“Nationalmarketsystemforsecurities; securitiesinformationprocessors.”59Foreseeingtheproblemsof fragmentedsecuritiesmarketsandtheabilityoftechnologyto helpinovercomingthem,CongressauthorizedtheSEC“touse itsauthorityunderthischaptertofacilitatetheestablishment ofanationalmarketsystemforsecurities.”60Giventhe acceleratingpaceofchangeinthenationalsecuritiesmarkets, particularlyinthe1990sandfirstdecadeofthe2000s,this regulatoryprojectrepresentedadifficultchallengeforthe SEC.61Andgiventhedevelopmentssince2007thatcameabout

55.Christie&Schultz,WhyDoNASDAQMarketMakers,supranote54,

at1838–39.

56.See,e.g.,FloydNorris,MarketPlace;U.S.InvestigatingNasdaq’s PricingPractices,N.Y.TIMES,Oct.20,1994,atD1;WilliamPower,Academics FanFlamesofNasdaqFairnessDebate,WALLST.J.,Nov.7,1995,atC1;The OddCoupleWins,ECONOMIST,Jan.13,1996,at70.

57.SeeOrderExecutionObligations,62Fed.Reg.40,732(July30,1997) (codifiedat17C.F.R.pt.240)(promulgatingthe“orderhandlingrules”); RobertL.D.Colby&ErikR.Sirri,ConsolidationandCompetitionintheUS EquityMarkets,5CAP.MKT.L.J.169,183–85(2010);seealsoCollusioninthe Stockmarket,ECONOMIST,Jan.17,1998,at89.

58.SeeOrderExecutionObligations,62Fed.Reg.40,732(July30,1997) (codifiedat17C.F.R.pt.240)(requiring“marketmakersandexchange specialiststopubliclydisplaycertaincustomerlimitorders”).

Id.§78k-1(a)(2).

59.15U.S.C.§78k-1(2012). 60. 61.RegulationNMS,70Fed.Reg.37,496(June29,2005)(codifiedat17 C.F.R.pt.242),representstheculminationofthisprocess.Thefinalrelease wasprecededbytwoproposedreleases,RegulationNMS,69Fed.Reg.77,424 (Dec.27,2004)andRegulationNMS,69Fed.Reg.11,126(Mar.9,2004).Fora

largelyinresponsetothenewregulatorylandscape, itis temptingtoseeRegulationNMSasafailure.Whilethenew systemcertainlycreatedopportunitiesforprivateactorsto gameit,theseproblems,withwhichthisArticleislargely concerned,shouldbeunderstoodasunintendedconsequences ofanattempttofostercompetitionamongdisparatesecurities marketsacrosstheUnitedStates.62

RegulationNMSconsistsofExchangeActRules600–612.63 Thekeyprovisionswerepromulgatedin2005.TheyareRule 611(theOrderProtectionRule),64Rule610(theAccessRule),65 Rule612(theSub-PennyRule),66andamendmentstoRules 601and603(theMarketDataRules).67Theserulesknit togetherthemanyexchangesandtradingvenuesintheUnited Statesthroughthemechanismofasingleprice,theNational BestBidandOffer68(NBBO).Theyalsosetupacomplicated systemofrulesmandatingtheprovisionofquotestothe marketatlarge,andhowbrokersandotherscaninteractin thisecosystem.69Whilelargelysuccessful initsaimsof reducing amongst exchanges,thenewregimealsoledtothecreationofahighly fragmentedsystem—contrarytoitsintentions—inwhicha subsetoftraderscanpreyuponslowertraders.70

competition

fostering

spreads

criticalhistoryofthedevelopmentofRegulationNMS,seeDaleA.Oesterle, RegulationNMS:HastheSECExceededItsCongressionalMandateto Facilitatea“NationalMarketSystem”inSecuritiesTrading?,1N.Y.U.J.L.& BUS.613(2005).SeealsoSELIGMAN,supranote21,at495–534.

62.SeeRegulationNMS,70Fed.Reg.at37,498(“TheNMSispremised onpromotingfaircompetitionamongindividualmarkets,whileatthesame timeassuringthatallofthesemarketsarelinkedtogether,throughfacilities andrules,inaunifiedsystemthatpromotesinteractionamongtheordersof buyersandsellersinaparticularNMSstock.”).

63.17C.F.R.§§242.600–.612(2014);seeRegulationNMS,70Fed.Reg.at

37,496.

Id.at37,502. Id.at37,503. Id.

64.RegulationNMS,70Fed.Reg.at37,501. 65. 66. 67. 68.17C.F.R.§242.600(b)(42)(“Nationalbestbidandnationalbestoffer means,withrespecttoquotationsforanNMSsecurity,thebestbidandbest offerforsuchsecuritythatarecalculatedanddisseminatedonacurrentand continuingbasisbyaplanprocessorpursuanttoaneffectivenationalmarket systemplan....”).

69.Seeid.§242.600(b)(11),(26)–(29)forNMSRegulationdefinitions

relatingtoquotes.

70.SeeCraigPirrong,PickYourPoison—FragmentationorMarket Power?AnAnalysisofRegNMS,HighFrequencyTrading,andSecurities MarketStructure,26J.APPLIEDCORP.FIN.8,8(2014)[hereinafterPirrong,

ThecornerstoneofRegulationNMSistheOrderProtection Rule.71Thisruleprevents“trade-throughs,”72ortradesonone marketatapriceinferiortooneavailableonanothermarket.73 Therulerequiresthateachtradingcenterestablish“policies andproceduresthatarereasonablydesignedtopreventtradethroughsonthattradingcenterofprotectedquotationsinNMS stocks,”otherthaninthecaseofaspecifiedexception.74Inthis way,ordersareprotectedfrombeingexecutedataprice inferiortotheNBBO.WhiletherewasconcernthattheOrder ProtectionRulecouldconflictwiththedutyof“bestexecution” bybrokers,aslarge“blockorders”couldfacedifficultyinbeing executedquickly,75executionatthebestpricewasgiven priorityinconstructingthenationalmarketsystem.76The OrderProtectionRule,therefore,iscommensuratewiththe generalprincipleofprice-timepriorityprevailinginthe markets.77Priceistheprimaryfactorindeterminingwhethera tradewillbemade.Whenthereismorethanonequoteata particularprice,thetimeofsubmissionisgenerallyusedasa secondaryfactor.78The“IntermarketSweepOrder”(ISO) exceptiontotheprohibitionagainsttrade-throughs79allows brokerstoexecutelargetradesquicklyinthemarketsatthe requestofclients,wherethedesireforspeedinexecutingthe transactionoutweighsthedesiretosearchforthebestpossible priceamongdifferentmarketcenters.80Asexploredfurther

(characterizingReg.NMSasleadingtothecurrent

PickYourPoison] fragmentedmarketsystem). 71.17C.F.R.§242.611. 72.

Id.§242.600(b)(77)(“Trade-throughmeansthepurchaseorsaleofan NMSstockduringregulartradinghours,eitherasprincipleoragent,ata pricethatislowerthanaprotectedbidorhigherthanaprotectedoffer.”).

73.Seeid.§242.611. 74. 75.SeeRegulationNMS,70Fed.Reg.37,496,37,523(June29,2005)

Id.§242.611(a).

(codifiedat17C.F.R.pt.242). Id.at37,497. Id.SeegenerallyHARRIS,supranote27,at116–20(“Rule-Based

76. 77.

Order-MatchingSystems”).

78.HARRIS,supranote27,at117(“Allorder-matchingmarketsuseprice priorityastheirprimaryorderprecedencerule. ...Marketsusevarious secondaryprecedencerulestorankordersbasedontheirtimeofsubmission, ontheirdisplaystatus,andontheirsize.”).

79.17C.F.R.§242.611(b)(5).The“IntermarketSweepOrder”isdefined

at17C.F.R.§242.600(b)(30).

80.SeeRegulationNMS,70Fed.Reg.at37,523(“TheCommission... providedexceptionsforintermarketsweepordersthatrespond[ed]totheneed

below,theISObecameanessentialiteminthetoolkitofhighfrequencytraders,butitsoriginlayinthedesireoftheSECto allowbroker-dealers “best execution”obligations.81

inmeeting

flexibility

their

TheAccessRule82providesforuniformaccesstoquoteson thevarioustradingvenuesbyrestrictingthepricechargedfor aquoteto$0.003pershare83andbyforbiddingmarketcenters fromdiscriminatinginaccesstoquotes.84Theseruleswere necessarybecausewithoutequalaccesstoquotes,requiring brokerstoexecuteattheNBBOwouldbe“futileifbrokerdealersandtradingcenterswereunabletoaccessthoseprices fairlyandefficiently.”85TheAccessRuleisimportanttoHFT fortwoverydifferentreasons.First,itmayplayasupporting roletootherregulationsthatareimplicatedinthequestionof whethertheexchangeshaveinfactactedina“discriminatory” fashioninproviding“enriched”datafeedsand“flashorders.”86 Whileofferingenricheddatafeedsandflashordersdoesnot appeartoviolatetheprovisionsoftheAccessRule,theymaybe atoddswiththegeneralintentionbehindtherule,whichwas to“furtherthegoaloffairandefficientaccesstoquotations fromunfairly primarilybyprohibitingtradingcenters discriminatingagainstnon-membersornon-subscribers”who accessquotationsthroughanexchangememberorsubscriber.87

ofmarketparticipantstoaccessmultiplepricelevelssimultaneouslyat differenttradingcenters.”).

81.SeeinfraSectionII.F. 82.17C.F.R.§242.610(“Accesstoquotations”). 83.

Id.§242.610(c)(1).Wherethe“priceofaprotectedquotationorother quotationislessthan$1.00,thefeeorfeescannotexceedoraccumulateto morethan0.3%ofthequotationpricepershare.”Id.§242.610(c)(2).

Id.§242.610(a).

84. 85.RegulationNMS,70Fed.Reg.37,502. 86.Theanti-discriminationprincipleofRule610(a)maybeasupplement toRule603(a)(2),whichmandatesthatanexchange“thatdistributes informationwithrespecttoquotationsforortransactionsinanNMSstockto asecuritiesinformationprocessor...shalldosoontermsthatarenot unreasonablydiscriminatory.”17C.F.R.§242.603(a)(2).Ontheotherhand, “theunfairdiscriminationstandardofRule610(a)willapplyonlytoaccessto quotations,nottothefullpanoplyofservicesthatmarketsgenerallyprovide onlytotheirmembers.”RegulationNMS,70Fed.Reg.at37,540.For discussionoftheapplicationofRegulationNMStoenricheddatafeedsand otherHFTpractices,seeJohnC.Coffee,Jr.,HighFrequencyTradingReform: TheShortTermandtheLongerTerm,COLUM.L.SCH.BLUESKYBLOG(July 21, clsbluesky.law.columbia.edu/2014/07/21/high-frequencytrading-reform-the-short-term-and-the-longer-term/.Seealsoinfranotes240– 41andaccompanyingtext.

2014),http://

87.RegulationNMS,70Fed.Reg.at37,539.

Second,theAccessRule’sprohibitiononlockedandcrossed quotationscreatesanenvironmentinwhichanumberof abusivepracticesofHFTthrive.88Ironically,theprohibition onlyarosebecauseofthedisorderlyconsequencesoftraders attemptingtocaptureliquidityrebatesinthe“maker/taker” paymentmodelusedbytheearlyelectronicexchanges.89A “locked”marketoccurswhenbidsandoffersarepostedatthe sameprice,whereasa“crossed”marketiswhenabidisposted atahigherpricethanthecurrentoffer(oranofferatalower pricethanthecurrentbid).90Lockedandcrossedmarkets representfundamentallyirrationalconditions,andmarkets reactbyseizingup.91Asaresult,RegulationNMSbans displayingquotesthatlockorcrosspreviouslydisplayedquotes intheinterestofpromoting“fairandorderlymarkets”and “marketefficiency.”92

Whydolocksandcrossesoccur?Themotiveforposting lockingorcrossingquotationsina“maker/taker”systemisto avoidbeinga“taker”ofliquidity,andtoinsteadbea“maker”of liquidity.Thisisbecauseliquiditytakersarechargedanaccess feeforhittingapre-existingorder,whileliquiditymakersare paidarebate,typicallytwo-thirdsofthefeepaidbythe liquiditytaker.93Unfortunately,banninglockingandcrossing quotesonlyleadstoafurtherproblem:thedevelopmentof abusiveordertypessuchas“HideNotSlide”whichprovide traderswithameanstoovercomethebanonlockingand crossingquoteswhilesimultaneouslyjumpingthequeueof price-timepriorityembeddedinmarketpractice.94

Next,theSub-PennyRuleprohibitsquotesinpricesless than$0.01fornationalmarketsecuritiesvaluedatmorethan $1.00,andinpriceslessthan$0.0001forstocksvaluedatless than$1.00.95Themainpurposeofthisrulewastoprotectthe

88.Seegenerallyid.at37,546–48(discussinglockingandcrossing

quotations).

89.Foradiscussionof“maker/taker”paymentmodel,seeinfranotes148– 51andaccompanyingtext.Foradiscussionoftheproblemoflockedand crossedmarketspriortothepassageofRegulationNMS,seeIvySchmerken, Nasdaq’sBattleoverLockedCrossedMarkets,WALLST.&TECH.,May2003, at12,12–18.

90.SeeSchmerken,supranote89,at14. 91. 92.RegulationNMS,70Fed.Reg.at37,547. 93.SeeSchmerken,supranote89,at16–18. 94.SeeinfraSectionII.F,whichdiscussesthe“HideNotSlide”orderin

detail.

95.17C.F.R.§242.612(2014);RegulationNMS,70Fed.Reg.at37,503.

quotes

decimals,

from1/8ths

useoflimitordersinthesecuritiesmarkets.96Alimitorderisa standingordertopurchaseorsellacertainamountofagiven securityatagivenprice;97thelistofstandinglimitorders formsthecentrallimitorderbookofamarketandconstitutes theavailableliquidityforastock.98Inthewakeofthe conversion or “decimalization,”whichoccurredin2000,therewaspressure ontheexchangestoallowquotesinfractionsofapenny.99This inevitablyledtotheuseofminuteincrementstojumpaheadof standinglimitordersinagivenmarket;atradercouldjump aheadofabidforagivenstockat$20.55,forexample,by bidding$20.551.100Allowingsuchminute“improvements”on priceswouldjeopardizetheuseoflimitorders,threateningthe mainmechanismforbuildingdeepliquidityinthemarkets.101 TheSub-PennyRuletherefore“wasdesignedtolimitthe abilityofamarketparticipanttogainexecutionpriorityovera competinglimitorderbysteppingaheadbyaneconomically insignificantamount.”102

Finally,RegulationNMSinstitutedMarketDataRulesin Rule601103andRule603104thatformtheregulatorybackdrop fortheinformationsystemsthatdisseminatetheNBBOand otherinformationtomarketparticipants.105(Rules602and 604–606wereleftmostlyunmodifiedbythefinalimplementing release.106)Theserulesprovideforthecontemporaryversionof the“consolidatedtape”107andservethecrucialfunctionof

96.SeeRegulationNMS,70Fed.Reg.at37,503. 97.SeeHARRIS,supranote27,at73(“Alimitorderisaninstructionto tradeatthebestpriceavailable,butonlyifitisnoworsethanthelimitprice specifiedbythetrader.”).

98.Seeid.at75–78(discussingtheliquiditybenefitsofstandinglimit

orders).

99.SeeRegulationNMS,70Fed.Reg.at37,550–51.

100.Seegenerallyid.at37,551–52(listing“avarietyofadditional

problemscausedbysub-pennyquoting”).

101. 102.Seeid.at37,551. 103.17C.F.R.§242.601(2014)

(governingthe“[d]isseminationof transactionreportsandlastsaledatawithrespecttotransactionsinNMS stocks”). 104.

Id.§242.603(governingthe“[d]istribution,consolidation,anddisplay informationwithrespecttoquotationsforandtransactionsinNMS

of stocks”).

105.RegulationNMS,70Fed.Reg.at37,503. Id.at37,570. 106. 107.Consolidated

Tape,

COMM’N, https://www.sec.gov/answers/consolt.htm(lastupdatedOct.15,2012)(“The

SEC.

&

asRule

“[a]nynational

disseminatingasingleNBBOforallexchange-listedstocks. Theyarealsocrucialforanyquestionsoffairnesssurrounding informationconcerning disseminationofvarioustypesof “exclusive specifies quotes, any processor[108]...distribut[ing] informationwithrespectto quotations...inanNMSstocktoasecuritiesinformation processor[(SIP)109]shalldosoontermsthatarefairand reasonable.”110Furthermore, securities exchange,nationalsecuritiesassociation,broker,ordealerthat distributesinformationwithrespect toquotationsforor transactionsinanNMSstock...shalldosoontermsthatare notunreasonablydiscriminatory.”111Theinvocationofequity intheserulesisespeciallyimportanttodebatesconcerningthe provisionofinformationbytheexchangesandothervenuesto HFTfirmsoutsidethemechanismofaSIP.112 currently

information four disseminationsystemsinheritingtheroleoftheconsolidated tape:TheConsolidatedQuotationSystem,theConsolidated TapeSystem,theNASDAQSystem,andtheOPRASystem.113 ThesefourentitiesqualifyasSIPsunderRegulationNMS,and eachmustdisseminatedatatothemarketinaccordancewith itsTransactionReportingPlan.114TheSIPsarecrucialnodes intheNationalMarketSystembecausetheyaremeanttobe theentitiesthatdisseminatetheNBBOtomarketparticipants atlarge.115TheSIPsareresponsibleforprovidingeachNMS

consolidated

There

are

‘consolidatedtape’isahigh-speed,electronicsystemthatreportsthelatest priceandvolumedataonsalesofexchange-listedstocks.”).

108.15U.S.C.§78c(a)(22)(B)(2014)(defining“exclusiveprocessor”). 109. Id.§78c(a)(22)(A)(defining“securitiesinformationprocessor”). 110.17C.F.R.§242.603(a)(1)(2015). 111. 112.See,e.g.,Coffee,supranote86;Foxetal.,supranote12,at270–71;

Id.§242.603(a)(2).

seealsoinfranotes238–40andaccompanyingtext.

113.RegulationNMS,70Fed.Reg.37,496,at37,503n.40,37,571(June29, 2005)(codifiedat17C.F.R.pt.242);seeOesterle,supranote61,at621,630; seealsoCollectionandDisseminationofTransactionReportsandLastSale Data,45Fed.Reg.12,377(Feb.26,1980).

114.RegulationNMS,70Fed.Reg.at37,571;seealso17C.F.R.§ 242.600(82)(2015)(“Transactionreportingplanmeansanyplanforcollecting, processing,makingavailableordisseminatingtransactionreportswith respecttotransactionsinsecuritiesfiledwiththeCommissionpursuantto, andmeetingtherequirementsof,§242.601.”).

115.SeeRegulationNMS,70Fed.Reg.at37,503–04.Foradiscussionof Congress’sintentinprovidingforoneormoreSIPsinthe1975Amendments, seeOesterle,supranote61,at624–26.

securitywith“coredata”116consistingof“(1)lastsalereports ...;(2)thecurrenthighestbidandlowestofferforthesecurity ...;and(3)the...NBBO.”117

Rule603(a)(1)requiresthat“[a]nyexclusiveprocessor,or anybrokerordealerwithrespecttoinformationforwhichitis theexclusivesource,thatdistributesinformationwithrespect toquotationsforortransactionsinanNMSstocktoa securitiesinformationprocessorshalldosoontermsthatare fairandreasonable.”118Thisrequirementismeanttoallowa SIPtogathertherelevantdataitrequires.Rule603(a)(2), however,allowsforexchangesandotherstodisseminate informationconcerningquotesandtransactionstoother marketparticipants“ontermsthatarenotunreasonably discriminatory.”119Thisimpliesalessstringentstandardthan theoneapplicabletotherequirementin603(a)(1)thatan “exclusiveprocessor”shallprovidesuchinformationtoaSIP.120 Thedisseminationofdatabytheexchangesthroughenriched datafeedsandotherarrangementsoutsideofdisseminationto theSIPfallsunder603(a)(2),notunder603(a)(1).121The questionofthetimingofthereleaseofsuchdatacameup duringthepromulgationofRegulationNMS.122TheSEC determinedthatanexchangecouldreleasedatatomarket participantssimultaneouslywithitsreleaseofdatatotheSIP, andwasnotrequiredtodosoinsuchawayas“tosynchronize thedeliveryofitsdatatoend-userswithdeliveryofdatabya Networkprocessortoend-users.”123Thisstancehasobviously

116.Whiletheterm“coredata”isnotdefinedassuchinReg.NMS,itis thedatarequiredbyReg.NMStobesubmittedbyanationalsecurities exchangeoranationalsecuritiesassociationtoaSIP.SeeNetCoalitionv. SEC,615F.3d525,529(D.C.Cir.2010);RegulationNMS,70Fed.Reg.at 37,567.ItisacrucialconceptforthefunctioningofReg.NMS.NetCoalition concernstheSEC’sapprovaloffeesfornon-coredatachargedtosecurities tradersbyNYSEArca.NetCoalition,615F.3dat525–29.TheReg.NMS Releaseallowspriceschargedfornon-coredatatobedeterminedbymarket forces,unliketheReg.NMSregimethatgovernsdistributionofrevenuesfrom marketdata.SeeRegulationNMS,70Fed.Reg.at37,569.

Id.§242.603(a)(2).

117.SeeNetCoalition,615F.3dat529. 118.17C.F.R.§242.603(a)(1)(2015). 119. 120.SeeRegulationNMS,70Fed.Reg.at37,566–67. 121.See17C.F.R.§242.603(a). 122.SeeRegulationNMS,70Fed.Reg.at37,566–67. 123.

Id.at37,567.

acquiredfarmorerelevancewiththeadventofHFTandthe marketingofprivateandenricheddatafeedstoHFTfirms.124

ithasbeenlesssuccessful

RegulationNMSisanextensivesetofverydetailed regulationsmeanttoknittogetherthevarioustradingvenues intheUnitedStates.Whileithasaccomplisheditsgoalof forcingtraderstocompeteinanationalmarketthroughthe mechanismoftheNBBOandtheprohibitionagainsttradeincounteringthe throughs, centrifugalforcesatworkamongthegrowingfieldoftrading venuessinceitsimplementationin2007.Whilealltheblame forthecurrentproblemsintheequitytradingenvironment cannotbelaidatthedoorofRegulationNMS,itisatthecenter oftheconstellationoffactorsleadingtothecurrentsituation, whichischaracterizedaboveallbyaraceforspeedbetween multipletradingvenues.

C.OVERVIEWOFTHECURRENTEQUITYMARKETS

pace

change

rapidly

technological

InthenineyearssinceRegulationNMScameintoforce, the advanced, transformingthemarkets.125Whilesevenyearsearlier,atthe turnofthecentury,themarketslookedinmanyrespects similartothemarketsofthe1960sor1970s,justnineyears latertheyareradicallydifferent.Almostnostockistradedona tradingflooranylonger,andanestimatedfortytosixtypercent ofthevolumeontheexchangesisduetoHFT.126Algorithmicassistedtradingmakesupmuchoftherest.127Inadditionto theITrevolution,theexchangeshaveconvertedfrommutuallyownedassociationsintofor-profitcompanies,transforming theirpolicy-makingandabilitytomakequickchanges.128The businessmodeloftheexchangeshaslikewisechanged,with theirroleasinformationproviderssupplementingtheirroleas

124.SeegenerallyinfraSectionII.D. 125.SeeConceptReleaseonEquityMarketStructure,ExchangeAct ReleaseNo.34-61358,75Fed.Reg.3594,3594–96(Jan.14,2010);see generallyAndreiA.Kirilenko&AndrewW.Lo,Moore’sLawVersusMurphy’s Law:AlgorithmicTradingandItsDiscontents,27J.ECON.PERSPS.51,51–53 (2013).

126.Kirilenko&Lo,supranote125,at59;seealsoGARYSHORTER&RENA S.MILLER,CONG.RESEARCHSERV.,R43608,HIGHFREQUENCYTRADING: BACKGROUND,CONCERNS,ANDREGULATORYDEVELOPMENTS13–14(2014) (describingmarketshareheldbyHFTinrecentyears). 127.Cf.SHORTER&MILLER,supranote126,at5. 128.Seeid.at15.

venuesfortrading.129Andtradingvenueshaveproliferated, particularlyamongthe“unlit”foraofdarkpools,withthirteen exchangesandoverfortydarkpoolsandotheralternative tradingsystems(ATS).130Alongwiththefragmentationof tradingvenueshascomeariseintechnologically-inducedflash crashesandothermarketdisruptions.Finally,thegrowing complexityofthemarketplace,combinedwithquestionsabout thequalityofliquidityandavailabilityofmanyquotesposted, causedobserverstoaskwhethertheoldexchangeshavemerely beenreplacedbyelectronicmarketsofquestionablereliability andnumeroushiddencosts.131

in2000.133Themost

Thedemutualizationoftheexchangesbeganin1993with theStockholmStockExchange.132Afterthatpoint,awaveof demutualizationsoccurred,includingthatoftheLondonStock importantAmerican Exchange demutualization,however,didnottakeplaceuntil2006when theNYSEconvertedfromamember-ownedcorporationtoa corporation.134Intheconversion,“ownersofthe1,366seatson theNYSEreceived80,177sharesofNYSEGroupstockplus $300,000incashandanother$70,571individends....[T]he dealvaluedeachseatatapproximately$5.5million.”135 Proponentsofdemutualizationarguedthatacorporate structurewouldprovefarmoreeffectiveindealingwiththe

129.Seeid.(notingfor-profittradingcentersrespondingtomarketdemand

byoffering“directconnectionstotheirtradedatatransmissions”).

130.Seeid.at8. 131.See,e.g.,SALARNUK&JOSEPHSALUZZI,BROKENMARKETS:HOW HIGHFREQUENCYTRADINGANDPREDATORYPRACTICESONWALLSTREETARE DESTROYINGINVESTORCONFIDENCEANDYOURPORTFOLIO47(2012);LEWIS, supranote1,at23–55(“Brad’sProblem”);PATTERSON,supranote11,at281– 96(“ARiggedGame”);MacKenzie,supranote6,at6–7;AliceK.Rossetal., RobotWars:HowHighFrequencyTradingChangedGlobalMarkets,BUREAU 2012), OF http://www.thebureauinvestigates.com/2012/09/16/robot-wars-how-highfrequency-trading-changed-global-markets.

16,

132.SeeReenaAggarwal,DemutualizationandCorporateGovernanceof

StockExchanges,15J.APPLIEDCORP.FIN.105(2002).

133.Seeid.at106tbl.1. 134.TheNYSEwasamember-owned,non-profitcorporationfrom1971 until2006.SeeRobertaS.Karmel,TheOnceandFutureNewYorkStock Exchange:TheRegulationofGlobalExchanges,1BROOK.J.CORP.FIN.& COM.L.355,385(2007).

135.AssociatedPress,NYSEStockSurgesinFirstDayofTrading,NBC NEWS(Mar.8,2006),http://www.nbcnews.com/id/11725095/38189133#.VkyTN-rR0s.SeegenerallyAaronLucchetti,ABigExchange:NYSEMembers ApproveDeal;VoteOverwhelminglyEndorsesArchipelagoPlan,Public Listing;ClosingInon‘EndofanEra,’WALLST.J.,Dec.7,2005,atC1.

cumbersomemember-ownedcooperative.136

challengesofrunningacomplexfinancialservicesbusiness In thana particular,achiefexecutiveofficerwithaboardofdirectors presentedafarmoreefficientdecision-makingbodythana member-ownedcooperative,wheregroupsofmembershad effectivevetopoweroverproposedchanges.137

Intheyearsbeforeandafterdemutualization,therewas significantacademiccontroversysurroundingthechangesin thelegalstructureoftheexchanges.Whilesome,suchas ProfessorPaulMahoney,arguedthatas self-regulatory organizations couldeffectivelyservethe theexchanges investingpublic’sinterests,138others,suchasProfessorsMacey andO’Hara,pointedoutthatunderaCoasiananalysisforprofitcorporateexchangeswereanuneasyfitwiththe prevailingregulatorystructuregoverningtheexchanges.139 MaceyandO’Hara’sargumentthatexchangesascorporations inthecurrentregulatorycontextwouldlikelyfailtoselfregulate,asregulationwasapublicgoodsubjecttosignificant

136.SeeOnnigH.Dombalagian,DemythologizingtheStockExchange: ReconcilingSelf-RegulationandtheNationalMarketSystem,39U.RICH.L. REV.1069,1132(2005);Fleckner,supranote15,at2576–77;BennSteil, ChangesintheOwnershipandGovernanceofSecuritiesExchanges:Causes andConsequences,2002BROOKINGS-WHARTONPAPERSONFIN.SERVS.61,68– 71;cf.RubenLee,CorporatisationofExchangesandCentralCounterparties: inCOMPAQHANDBOOKOFWORLDSTOCK, TheDarkSideWillCome, DERIVATIVE&COMMODITYEXCHANGESxxxix,xl (arguingthat “[e]xchangemanagementsare...likelytotakeexcessiverisks”).

(2002)

137.SeegenerallySteil,supranote136,at69,81(presentingvarious

optionsfortheorganizationalstructureofexchanges).

138.PaulG.Mahoney,TheExchangeAsRegulator,83VA.L.REV.1453, 1453–56(1997);seealsoJakeKeaveny,InDefenseofMarketSelf-Regulation, 70BROOK.L.REV.1419,1420(2005);Karmel,supranote45,at409(“Current GovernanceStructuresandtheFutureofSelf-Regulation.”).Intellectual supportforMahoney’sself-regulationargumentisfoundinStephenJ.Choi& AndrewT.Guzman,NationalLaws,InternationalMoney:Regulationina GlobalCapitalMarket,65FORDHAML.REV.1855(1997)andDanielR. Fischel,OrganizedExchangesandtheRegulationofDualClassCommon Stock,54U.CHI.L.REV.119(1987).

139.Macey&O’Hara,supranote15,at563–66.Macey&O’Hara’s argumentreliesonCoase’sdistinctionbetweentransactioncostsandagency coststoarguethatthereductionintransactioncostsoftradingsecurities increasesthelevelofcompetitionbetweentradingvenues,whichinturn reducestheefficacyofself-regulationonthepartoftheexchanges.Seeid.at 565;seealsoCoffee,supranote86(“Withtheprivatizationofsecurities markets,marketsarenolongerselfregulatorsinanymeaningfulsense,but insteadareaggressiveprofitmaximizers.”).Foraneconomicanalysisof problemswithself-regulationatthecommodityexchanges,seePirrong,The Self-RegulationofCommodityExchanges,supranote42.

rider

problems,was

140Nevertheless, free demutualizationproceededandwascompleteamongthe world’smajorstocktradingvenuesby2006.141

prescient.

InadditiontotheintroductionofRegulationNMSand demutualization, technologicaldevelopmentsalsohada transformationaleffectontheexchanges.ScottPatterson’s DarkPools:TheRiseoftheMachineTradersandtheRiggingof theU.S.StockMarkettellsthestoryoftheriseofHFT.142 Computerizedtradingwaspioneeredinthe1990sbyJosh LevineatDatekSecurities,143whowrotecomputerprograms thatwouldtradeonNASDAQquotespostedonthesmall-order entrysystem(SOES)byNASDAQmarketmakers.144These quoteswerepostedontheelectronicbulletinboardoperatedby NASDAQandleftovernight,orsimplybecamestaleinlightof informationemergingfromothervenues.145Levine’sprograms tradedonthediscrepancybetweenSOESquotesandother prices,enablingthe“SOESbandits”atDatekSecuritiesto profithandsomely.146Datek’soperationsgaverisetotheIsland tradingnetwork,anelectroniccommunicationnetwork(ECN) thatcompetedwithInstinetformarketshareinNASDAQ stocks.147Islandalsorevolutionizedhowbothtradersand exchangesprofitedfromtheiractivities.Inordertoattract

140.SeeMacey&O’Hara,supranote15,at576. 141.SeeAggarwal,supranote132,at106. 142.PATTERSON,supranote11. 143.Seegenerallyid.at67–230. 144.WhiletheNASDAQwasformedtoprovideforelectronic,screen-based disseminationofquotes,tradeswerestillconductedoverthephone.SeeMeeta Kothare&PaulA.Laux,TradingCostsandtheTradingSystemsforNasdaq Stocks,51FIN.ANALYSTSJ.42,43(1995).NASDAQintroducedtheSOESin 1984inordertoallowforautomaticexecutionofsmallordersofupto500 sharesofcertainNASDAQ-listedstocks.Id.Inresponsetocriticismthat NASDAQdealersfailedtoexecutetradesatpostedpricesduringthestock marketcrashofOctober19,1987,NASDAQexpandedthesharelimitfor smallordersto1,000shares.Id.at44.TheSOESwassubjecttovociferous criticismbydealersbecauseitexposedthemtomorenimbletraderswhocould pickoffits“stalequotes.”Seeid.Thesenimbletradersbecameknownasthe “SOESbandits.”SeeChuckEpstein,SOESStorefronts,WALLST.&TECH., Sept.1997,at48,48–49.

145.SeegenerallyKothare&Laux,supranote144,at45–46. 146.SeePATTERSON,supranote11,at89–99;JeffreyH.Harris&PaulH. Schultz,TheTradingProfitsofSOESBandits,50J.FIN.ECON.39,39–40 (1997).

147.PATTERSON,supranote11,at116–24;RebeccaBuckman,IslandECN RaisesCapitaltoBecomeaStockExchange,WALLST.J.,May11,1999,atC1; seealsoMacKenzie,supranote6,at35–37(discussingIslandandits relationshipwiththeSOESbandits).

orderflow,Islandwasthefirsttoinstitutewhatbecameknown asthe“maker/taker”paymentsystem.148Insteadofearning commissionsthroughspreads,whichhadcollapsedinthewake ofdecimalization,149Islandpaidtraderswhosuppliedaquote thatwasactedupon(i.e.,the“makers”)andchargedthosewho tooksuchaposition(i.e.,the“takers”).150Islanditselfearned moneyfromthisarrangementbykeepingone-thirdofthefees collectedfromthetakers,whilepayingouttwo-thirdsofthe feesasrebatestothemakers.151

Thecurrentfragmentedandunsettledstateofthemarkets issignaledbyahandfulofrecenteventsconnectedwiththe riseofcomputerizedtrading.InadditiontoMichaelLewis’s FlashBoys,the“flashcrash”ofMay6,2010andthefailed IPOsofBATSGlobalMarketsandVirtuFinancialInc.brought thecurrentproblemsofthemarketstowidespreadpublic attention.152Theflashcrashinvolvedthelossofover$1trillion inmarketcapitalizationinlessthananhourandthenits subsequentrecoverybytheendingofthetradingday.153

148.SeePATTERSON,supranote11,at157–60;MaureenO’Hara,High FrequencyMarketMicrostructure,116J.FIN.ECON.257,262(2015)(arguing thatthemaker/takerfeestructureisparticularlyattractivetoHFT).Because anyone“hitting”aquotehastopayafee,andanyonepostingaquotethatis hitreceivesarebate,themaker/takersystemcreatesamismatchbetweenthe quotedbidandaskspreadandtheactual,economicbid/askspread.See LawrenceHarris&EthanNamvar,TheEconomicsofFlashOrdersand Trading manuscript), (unpublished 2011) 15, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1953524.

149.

“Decimalization”istheconversionofpricesonanexchangefrom 1/8ths(typically)todollarsandcents.Theexchangesvoluntarilyintroduced decimalizationinApril2001afterCongressthreatenedtopasslegislation requiringpricesindollarsandcents.SeeHARRIS,supranote27,at115.While decimalizationgreatlyreducedspreadsonstocks,andthereforereducedthe transactioncostsforinvestors,italsoreducedtheabilityofretailstockbrokers toearnprofits,andiswidelysuspectedofcontributingtothereductionin IPOsandthelackofliquidityforsmallandmid-capstocks.Staff,U.S.Sec.& Exch.Comm’n,ReporttoCongressonDecimalization(July2012).Foran overviewoftheregulatoryhistoryofthemovetopricingindecimalsaswellas currentconcerns,seeid.

150.SeePATTERSON,supranote11,at157–60. 151.Seeid. 152.SeegenerallyKirilenko&Lo,supranote125,at60–67(detailingfive incidentsrelatedto“newvulnerabilitiescreatedorfacilitatedbyalgorithmic trading”).

153.SeeCOMMODITYFUTURESTRADINGCOMM’N&SEC.&EXCH.COMM’N, FINDINGSREGARDINGTHEMARKETEVENTSOF6MAY2010(Sept.30,2010) [hereinafterCFTC&SEC];EdgarOrtegaBarrales,LessonsfromtheFlash CrashfortheRegulationofHigh-FrequencyTraders,17FORDHAMJ.CORP.& FIN.L.1195,1195(2012);AnanthMadhavan,Exchange-TradedFunds,

Duringthisepisode,believedtohavebeentriggeredbythesale of75,000E-miniS&P500June2010stockindexfutures(worth $4.1billion),thealgorithmictradingprogramsofanumberof firmswenthaywireandstartedacascadeofsellordersfor firmsincludingAccentureand3M,aswellascertainstock indexexchange-tradedfunds.154Whileultimatelynonconsequentialtothemajorityofmarketparticipants,some retailandinstitutionalinvestorsdidsufferseriouslosses.155 Moregenerally,thisepisodealarmedobserversasitpresented thespecterofmarketscareeningoutofcontrolduetocomputer malfunctions.156Theseandotherincidents,suchasthe“miniflashcrashes”ofSeptember27,2010,157October15,2014,158 andmanyothers,gaverisetocallsforimprovedcircuit breakers a malfunction.159Otherincidentsthathighlightedthedangersof automatedtradingincludethedelayedIPOofFacebookand thefailedIPOofBATSGlobalExchange;160thelatter, ironically,amajorplayerintheHFTworld.161TheFacebook IPOstumbledduetodelaysintheprocessingoftrades,162and

algorithmic

inthe

trading

event

MarketStructureandtheFlashCrash,FIN.ANALYSTSJ.,July–Aug.2012,at 20,20;AndreiA.Kirilenko,AlbertS.Kyle,MehrdadSamadi&Tugkan Tuzun,TheFlashCrash:TheImpactofHighFrequencyTradingonan ElectronicMarket (unpublishedmanuscript), (Nov. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1686004.

7,

154.SeeCFTC&SEC,supranote153,at83–86(describingtheimpactof theFlashCrashonAccentureand3Mstock);Kirilenkoetal.,supranote153, at5.

155.SeeKirilenkoetal.,supranote153,at1(“Broadstockmarketindices –theS&P500,theNasdaq100,andtheRussell2000,collapsedand reboundedwithextraordinaryvelocity.”);Madhavan,supranote153,at20 (“Despiteits theflashcrashaffectedmanymarket participants.”).

shortduration,

156.SeeKirilenkoetal.,id. 157.SeeGrahamBowley,TheFlashCrash,inMiniature,N.Y.TIMES,Nov.

9,2010,atB1.

158.MattEgan,Market‘FreakOut’:StocksReboundfromScaryPlunge, http://money.cnn.com/2014/10/15

CNNMONEY /investing/stocks-markets-wall-street-correction/.

15,

Proposals,

159.SeeJacobBunge&BrendanConway,RegulatorsHoneCircuit2011, ST. Breaker http://www.wsj.com/articles/SB100014240529702048313045765970234376848 18;DemosTelis&MichaelMackenzie,USExchangesEyeHarmonizedPause Rules,FIN.TIMES,May6,2011,at2.Forthediscussionofthecircuitbreaker programsimplementedaftertheMay6,2010crash,seeCFTC&SEC,supra note153,at7.

Sept.

28,

J.,

160.SeeKirilenko&Lo,supranote125,at63–65. 161. 162.

Id.at64. Id.at63–64.

in

prices

further

spurred

theBATSIPOfailedonaccountofreactionsoftraderstoinitial declines downward momentum.163AndonAugust1,2012,thetradingfirmKnight Capitalnearlywentbankruptduetonewsoftwarethat malfunctioneduponitsintroduction,leadingtolossestotaling $457millionandKnight’sabsorptionintoGETCO,aleading Chicago-basedHFTfirm.164Theseincidentshighlightedthe potentiallydangerousstateofthecurrentequitymarkets, wherehumanjudgmenthasbeenpushedasidebylightningfastalgorithmictradingprogramsengagedina“battleofthe bots”againsttheircompetitors.

II.PROBLEMATICVARIETIESOFHIGHFREQUENCY TRADING

HFTisbestdefinedascomputer-assistedtradingthat exploitsincrediblysmalltimedifferencestoyieldprofitsat minimalrisktothoseemployingit.165Whilethisdefinition capturesmostactivitiesreferredtoashighfrequencytrading, itshouldberememberedthatHFTisnotonethingandit

Id.at64–65. Id.at65.

163. 164. 165.Anumberofcommentatorshavenotedthedifficultyincomingtoa precisedefinitionof“high-frequencytrading.”See,e.g.,O’Hara,supranote 148,at258(“Highfrequencytradingisamisnomer,aseeminglypreciseterm usedtodescribealargeanddiversesetofactivitiesandbehaviors.”);Gomber etal.,supranote49(Section3,“HighFrequencyTradingDefinitionsand RelatedConcepts,”andAppendixIII,“AcademicandRegulatoryDefinitionsof High-FrequencyTrading”).TheSECoffersacharacteristic-baseddescription ofHFTstatingthat“highfrequencytrading...isrelativelynewandisnot yetclearlydefined.”ConceptReleaseonEquityMarketStructure,75Fed. Reg.3594,3606(Jan.14,2010)(codifiedat17C.F.Rpt.242).TheSECthen goesontospecifyfivecharacteristicsoffirmsengagedinHFT:

theuseofextraordinarilyhigh-speedandsophisticated (1) computerprogramsforgenerating,routing,andexecutingorders; (2)useofco-locationservicesandindividualdatafeedsofferedby exchangesandotherstominimizenetworkandothertypesof latencies; forestablishingand liquidatingpositions;(4)thesubmissionofnumerousordersthat arecancelledshortlyaftersubmission;and(5)endingthetrading dayinasclosetoaflatpositionaspossible(thatis,notcarrying significant,unhedgedpositionsover-night).

(3)veryshort

time-frames

Id.;seealsoDavidEasley,MarcosM.LópezdePrado&MaureenO’Hara,The VolumeClock:InsightsintotheHigh-FrequencyParadigm,39J.PORTFOLIO MGMT.19(2012)(“Today’shighfrequencymarketsarenottheoldlow frequencymarketsonsteroids.”).

existsonacontinuumwithitsparent,algorithmictrading.166 HFTreferstoawidevarietyoftradingstrategies,mostof whichexploitminutetimingdifferencesoruseinformation technologytoincorporatevastquantitiesofdataintotheir determinationsfarmorequicklythananyhumancould.167 Whileanumberofthemost-publicizedtypesofHFTpresent seriousfairnessproblems,HFTisnomoreintrinsicallymorally problematicthantakingataxiinsteadofwalking.Aswillbe showninPartIII,moralconcernsariseifothermarket participantsarekeptinthedarkastotheactivitiesofhigh frequencytraders,ifexchangescatertohighfrequencytraders attheexpenseofotherinvestors,orifHFTcreatesan intrinsicallyunevenplayingfieldwhenotherparticipantshave alegitimateexpectationofparityofposition.PartIIpresents anoverviewofHFTthatmovesfromitsoriginsinalgorithmic tradingtoitsprogressivelymoreproblematicaspects.

A.THEORIGINSOFHFT:ALGORITHMICTRADING

HFTdevelopedoutofalgorithmictradingtechniquesused toimplementsophisticatedinvestmentstrategies.168The earliestalgorithmicstrategieswerefirst-generationindex funds,whichboughtandsoldstockstooptimizeaninvestor’s portfolioinaccordancewithMarkowitz’sportfoliooptimization theory.169KirilenkoandLostatethatwidespreaduseof algorithmictradingbeganin1975whenfinanceprofessorBarr RosenbergfoundedBarrRosenbergAssociates,aninvestment advisoryfirmthatcalculatedcovariancematricesforU.S. equitiesandprovidedclientswithsoftwareenablingthemto tradeonthisinformationtoachieveoptimalportfolios.170Using algorithmsintradingeliminatedtheneedforhumanjudgment, whichatanyratewouldbeimpossibleinquantitative strategieswiththousandsofvariablesorinputs.171Another earlyinstancewastheequal-weightedindexfund,which requiredcontinualupdatingofitsportfoliotomaintainequal

166.SeegenerallyKirilenko&Lo,supranote125,at52(describingthe

historicalrelationshipbetweenalgorithmictradingandHFT).

167.Seeid.at58–59. 168.Seeid.at51–60foranoverviewofthisdevelopment. 169.Seegenerallyid.at55–56. 170. 171.Seeid.at67;Stoll,supranote21,at153.

Id.at55.

dollaramountsofequities.172Intheseearlyformsof algorithmictrading,humanjudgmentsconcerningindividual stocktransactionsweresetasideinfavorofquantitative strategiestakinghundredsandeventhousandsofvariables intoaccount.173Algorithmictradingthereforerepresentsa techniquetoovercomethecomputationallimitationsofthe humanmindandourlimitedhumanrationality.174

theyoftenprovidedlucrative

Asbothtechnologyandfinancialeconomicsprogressed, moresophisticatedstrategiesemployingalgorithmscameinto use.ArbitragestrategiesbasedontheBlack-Scholes/Merton capitalassetpricingmodel(CAPM)andMerton’sconceptof “dynamicspanning”enabledhedgefundsandothertradersto exploitdifferencesinpayoutsofvarioustypesofsecuritiesand derivativefinancialinstruments.175Whilesuchstrategieswere notriskless,asthecaseofLongTermCapitalManagement showed, to sophisticatedtraders.176Otheralgorithmictradingstrategies includedsoftwareusedtoengageinmarket-makingactivities andautomatedorderexecutionprogramsdesignedtocarryout largetransactionsinthemostadvantageousmanner.177 tradingdonot ofalgorithmic

seem intrinsicallymorally trading programsareusedasasortof“mentalprosthetic”that overcomesthefailingsofhumanrationality.178Ourcognitive abilitiesarerelativelylimited.Ahumantradercanonly incorporateasmallamountofdataintoherdecision-making, andsoinvestmentstrategiesrequiringthecalculationof

problematic.Computerized

opportunities

These

types

172.SeeKirilenko&Lo,supranote125,at55–56;seealsoStoll,supranote

21,at167–68(discussingtheimportanceof“indexarbitrage”inthe1980s).

173.See,e.g.,Kirilenko&Lo,supranote125,at54–55. 174.See,e.g.,Kirilenko&Lo,supranote125,at53–60;BrunoBiais& PaulWoolley,HighFrequencyTrading§II.3(Mar.11,2011)(unpublished manuscript), idei.fr/sites/default/files/medias/doc/conf/pwri/biais_pwri_0311.pdf algorithmsasmitigatingtraders’“cognitionlimits”).

(HFT

175.SeeKirilenko&Lo,id.at55. 176.SeeStephanieYang,TheEpicStoryofHowa‘Genius’HedgeFund AlmostCausedaGlobalFinancialMeltdown,BUS.INSIDER(July10,2014), http://www.businessinsider.com/the-fall-of-long-term-capital-management2014-7.

177.SeeKirilenko&Lo,supranote125,at58–59. 178.SeeFelixSalmon&JonStokes,AlgorithmsTakeControlofWall http://www.wired.com/2010/12 2010), Street,WIRED /ff_ai_flashtrading/(describingareaswherealgorithmictradingexcelsover humanoperators);seealsoKirilenko&Lo,supranote125,at53–55.

(Dec.

27,

matriceswithhundredsofthousandsofvariablesnecessarily requirecomputertechnology.Likewise,humantradersfindit verydifficulttomonitorpricesonmorethanaverysmall financialmarketssimultaneously,socommon numberof arbitragestrategiesrequirecomputerassistancetoimplement. Innoneofthesebasictypesofalgorithmictradingdowefeel thattheuseofalgorithmsorcomputersisintrinsicallywrong anymorethanwefeelthatflyinggoodsfromonemarketto anotherdoesamoralwrongtothosetraderswhocontinueto movethembyship.Arbitrageisfundamentallyatrading activity,wherebypartiesprofitbysupplyingagoodtoamarket whereitissoldatahigherpricethanwhereitisoriginally purchased.179AndwhileHFTgrowsoutofthesestrategies,its morallyproblematicaspectsdonotstemfromthemereuseof computertechnology,orevenofthespeedatwhichthis technologyoperates.

B.CO-LOCATION

“Co-location”isthefirstandmosttangiblemanifestationof HFT.Itreferstotheexchanges’practiceofrentingspaceinthe facilitiesthathousetheircomputerserverstotraderswho believetheycanbenefitfromthisproximity.180Co-location facilitatesthepracticeof“latencyarbitrage,”tradingbetween marketsbasedonpurespeed.181Tosucceedat latency arbitrage,atraderhastobefirstinlineintheprice-timeorder queuesusedbytheexchanges.Itisazero-sumgamewherethe orderthatisfirstinlinewithabidorofferatanadvantageous pricewinsthecompetition.182

57–63(2012)(“TradersandMarketMakers”).

Trading,

180.SeeJerryAdler,RagingBulls:HowWallStreetGotAddictedtoLight2012), (Aug. Speed http://www.wired.com/2012/08/ff_wallstreet_trading/; Rogow, Colocation:TheRootofAllHigh-FrequencyTradingEvil?,WALLST.J.:MKT. http://blogs.wsj.com/marketbeat BEAT 2012), /2012/09/20/collocation-the-root-of-all-high-frequency-trading-evil/; also MichaelA.Goldstein,PavitraKumar&FrankC.Graves,Computerizedand High-FrequencyTrading,49FIN.REV.177,188(2014).

8, Geoffrey

20,

181.SeeDavidGaffen,MeasuringArbitrageinMilliseconds,WALL.ST.J.: 2009),

MKT. http://blogs.wsj.com/marketbeat/2009/03/09/measuring-arbitrage-inmilliseconds/.

9,

182.See,e.g.,id.(quotingahedgefundtechnologyofficerstatingthat,with

HFT,“[t]hefastestonewins”).

Toillustratetheadvantagesaffordedbyco-location,ahigh frequencytraderlocatednexttoaserverwouldhaveitsorders transmittedtotheexchange’sservernearlyinstantaneously,183 whileanordertransmittedbyatraderthroughthefiber-optic telecommunicationnetworksfromdowntownManhattantothe NYSE’sfacilitiesinMahwah,NewJerseywouldarriveinfour tosevenmilliseconds.184MichaelLewis’sFlashBoystellsthe storyoftheconstructionofafiberopticlinefromChicagoto Mahwahintendedtoshavethreetofourmillisecondsoffthe deliverytimeitwouldtakeforamessagetobesentthrougha telephonecompany’slines.185Eachmillisecondisworthan estimated$100milliontoalargebrokeragefirm,186andthe firmconstructingthelineintendedtoprofitbychargingupto 200firmsanestimated$14millioneachforafive-yearleaseto usethenewline.187

Services,

amillisecond.

183.Round-triptransmissiontimesforco-locatedcomputersarewell EUREXEXCH., Co-location under www.eurexchange.com/exchange-en/technology/co-location-services/ (last visitedOct.9,2015).TheEurexGroupstatesthatco-locationround-trip transmissiontimesarefrom0.2-0.35milliseconds.Id.Interestingly—because ofwhatitimpliesaboutfairnessandtheperceptionoffairnessamongHFT clientsoftheexchanges—allmessagesfromco-locatedcomputerswithinan exchange’sserverfacilitytraveltheexactsamelengthoffiber-opticcableto reachtheexchange’sserver;forcomputersonracksclosertotheexchange server,thecableiscoileduptoequalthelengthofthemostdistantcomputers. SeeDavidSchneider,TheMicrosecondMarket,IEEESPECTRUM(Apr.30, 2012),http://spectrum.ieee.org/computing/networks/the-microsecond-market.

184.This4–7millisecondamount

isdeterminedbyanexperiment performedbyFlashBoysprotagonistBradKatsuyama.SeeLEWIS,supranote 1,at69–71.Inordertotesttheirexecutionprogram“Thor,”Katsuyamaand RonanRyantestwhenmessagessentfromtheirofficesatOneLibertyPlaza indowntownManhattanarriveatthevariousexchangeserversinNew Jersey.Id.Ontheotherhand,thefastestamessagecantravelfromone exchangeservertoanotherintheirtestis465microseconds,or0.465 milliseconds.Id.at71.

185.Seeid.at69–71.ForadditionalinformationontheprojectinFlash Boys,seeGregoryLaughlin,AnthonyAguirre&JosephGrundfest, InformationTransmissionBetweenFinancialMarketsinChicagoandNew York1–9(RockCtr.forCorp.Governance,WorkingPaperNo.137,2012). Laughlinetal.determinethattheSpreadNetworksfiberopticlinedetailedin FlashBoysreducedtheChicago–NewJerseylatencytoapproximately6.65 milliseconds.Id.at2.In2011suchtimeswerereducedfurtherto4.2–5.2 millisecondsthroughtheuseofmicrowavecommunicationsnetworks.Id.

186.SeeCiamacC.Moallemi&MehmetSaglam,TheCostofLatencyin (unpublishedmanuscript),

High-FrequencyTrading2(Feb.5,2013) https://moallemi.com/ciamac/papers/latency-2009.pdf.

187.SeeLEWIS,supranote1,at15.

Themajorexchangesallrentspaceontheirserverracksto outsideparties.Itisdifficulttodeterminetheexactamounts chargedforco-location,althoughindustryobserversbelieve theyhaveincreasedinrecentyears.188Nevertheless,itdoes appeartoconstituteasubstantialportionof theyearly revenuesoftheexchanges.189Whilesuchamountsmaybe affordabletoaHFTfirmpayingthem, theywouldbe prohibitiveforaretailinvestor,190evenifsuchindividualhad thetechnicalwherewithaltotakeadvantageofaco-located investorscouldpresumably position.Thoughinstitutional

188.SeeSHORTER&MILLER,supranote126,at14(“[S]ecuritiesexchanges beganchargingHFTfirmssubstantiallymorefortherighttobeconnectedto theexchange’sdataservers.”);Goldsteinetal.,supranote180,at183 (attributingthedeclineinprofitsamongHFTfirmsinparttothe“costsof rightsandhardwareforkeepingcomputersclosetothemajorexchanges”).

(Feb.

2015),

189.Forexample,themostrecent10KavailableofIntercontinental Exchange,Inc.,thepresentoperatoroftheNYSEaswellasothersecurities, futures,andoptionsexchanges,reportsthatdataservicesrevenues(which includesco-locationfees)were$631millionin2014,or14.9%oftotalrevenues of$4.221billion.SeeIntercontinentalExch.,Inc.,2014AnnualReport(Form http://ir.theice.com/financial-reports/annual10-K) reports.aspx.Thisincreasedfrom13.3%ofrevenuesin2013,10.7%in2012, and9.4%in2011.Id.(“ConsolidatedStatementofIncomeData”).Likewise, theNASDAQOMXGroup,Inc.reportsthatAccessandBrokersServices Revenues(whichincludesco-locationfees)were$257millionoutof$2.067 billiontotalrevenues(lesstransaction-basedexpenses)in2014,or12.4%; $255millionoutof$1.895billionin2013,or13%;and$257millionoutof $1.674billionin2012,or15%.SeeNASDAQOMXGroup,Inc.,2014Annual Report 2015), http://files.shareholder.com/downloads/NDAQ/957273764x0xS1120193-153/1120193/filing.pdf.Whileitisdifficulttobaseconclusionsontheseamounts, becausespecificco-locationamountsarenotbrokenoutoftheirlargerrevenue categories,co-locationisbelievedtobeasignificantrevenuestreamforthe exchanges.SeeDavidEasley,MaureenO’Hara&LiyanYang,Differential AccesstoPriceInformationinFinancialMarkets1–2(Feb.2013)(unpublished manuscript), http://papers.ssrn.com/com/sol3 /papers.cfm?abstract_id=1787029.

(Form

39–40

10-K)

(Feb.

17,

190.A2009articlequotesanNYSEexecutiveasstatingthatco-locatingat exchangescostsasmuchas$500,000permonth.KristiOloffson&Stephen Gandel,High-FrequencyTradingGrows,ShroudedinSecret,TIME(Aug.5, 2009), http://content.time.com/time/business/article/0,8599,1914724,00.html. DonaldMacKenziereportsthathewastoldthat“a‘rack’,acabinetsized spacethatcanaccommodatethirtyorfortycomputers,costsbetween$1500 and$15,000amonth,”whereasanentire“‘cage’...caneasilycostmorethan $1millionayear.”DonaldMacKenzie,BeGratefulforDrizzle,36LONDON REV.BOOKS27(2014).AndDingetal.reportthat“aco-locatedserverusing onlySIPdataisapproximately$7,000permonth,”whileonewithadirect datafeedfromtheexchangecoststhreetimesasmuch.ShenweiDing,John Hanna&TerrenceHendershott,HowSlowistheNBBO?AComparisonwith DirectExchangeFeeds,49FIN.REV.313,315n.3(2014).

affordthem, theyonlymakesenseaspartofalarger investmentinHFT,whichinvolvesconsiderableoutlaysfor technologyandscarcepersonnel.191Whileeachmillisecond savedintransmittingmessagesmaybeworthmillionsof dollarsperyearperHFTfirm,iteffectivelyrequiresamultimilliondollarinvestmentinordertobeabletocapturethese gains.

Co-locationhasgivenrisetonumerouscriticismsand complaintsthatitisunfair.192Itiseasytoseethebasisforthis complaint:iftradingincomputerizedmarketsisessentiallya racetogettotheexchangefastest,co-locatedtradersaregiven aheadstartthatcanneverbecaughtupto.Inthissense,colocationgoesdirectlyagainstthe(moral)intuitionthatatsome basiclevel,participantsinthemarketshouldbeonalevel playingfield.193Herethatfieldisuneven,wherethevast majorityofparticipantshavenochanceofevercatchingup. Theexchangeshaveaverybasicresponsetosuchcriticisms: theydonotdiscriminateamongtraders,andspaceintheir facilitiesisopentoanyonewillingtopaythespecifiedcost.194 Whilesuchargumentsareprimafacietrue,theyneglectthe factthatonlyaverysmallsubsetofstakeholdershasthe requisiteabilitytoprofitfromco-location.Economically,colocationisbetterseenasasortofcollusiononthepartofthe exchangeswithasmallgroupofcustomerstoprofitateveryone else’sexpense.195

Thereisalsothequestionofthegovernment’srolein allowingco-location.TheSEChaspermittedthepractice, reasoningthataslongasanyonewishingtopayforitcanhave

191.SeeSHORTER&MILLER,supranote126,at22(whiletheoretically opentoall,co-locationisonlylikelytomakefinancialsensefor“organized, institutional,stronglycapitalized”HFTfirms).

192.See,e.g.,SHORTER&MILLER,supranote126,at22–23;Goldsteinet

al.,supranote180,at194–95;Gomberetal.,supranote49,at34–35.

193.SeeinfraSectionIII.D. 194.See,e.g.,AmeetSachdev,CME’sTerryDuffyTalksShop,CHI.TRIB. (Sept.5,2014),http://www.chicagotribune.com/business/ct-cme-duffy-0908-biz20140905-story.html(“Weofferco-locationtoeverybody...[ande]verybody canhaveitatthesamespeediftheywantit.”);seealsoJamesJ.Angel& DouglasMcCabe,FairnessinFinancialMarkets:TheCaseofHighFrequency Trading,112J.BUS.ETHICS585,590–91(2013)(“Does‘co-location’give tradersanunfairheadstart?Tocertainextent,tradershavealwaysinvested heavilytogetclosertothesceneoftrading.”).

195.ThisisoneofthecentralpointsofArnukandSaluzzi’scriticismsof HFT.SeeARNUK&SALUZZI,supranote131,at99–101;seealsoinfranotes 441–42andaccompanyingtext.

co-locationinher

access,itisacceptable.196MaryJoWhiteraisedtheissueof fairnessof recentspeechtomarket participants,butfromhercommentsitappearsunlikelythe SECwouldattempttoputthegeniebackinthebottle.197And whileregulatorsmayhavebeenlaxinallowingtheexchanges torentspaceintheircomputerfacilities,therearerelatively straightforwardproposalstomitigatetheproblemsthat result.198

C.FLASHORDERS

FlashordersareanothermanifestationofHFTthathave receivedextensiveattentioninthemedia,Congress,andeven attheSEC.199Investorsinparticularareupsetthattheyallow HFTfirmsa“sneakpeek”atthemarket,ineffectallowing themto“front-run”theirslowerordersontheirwaytothe consolidatedtape,nowknownastheSIP.200TheSEC

196.See,e.g.,OrderApprovingaProposedRuleChangetoCodifyPrices forCo-LocationServices,75Fed.Reg.38,860,38,861(July6,2010)(“The Exchangehasalsorepresentedthatco-locationservicesaregenerally availabletoallqualifiedmarketparticipantswhodesirethem.”);Noticeof FilingandImmediateEffectivenessofProposedRuleChangeExpandingColocation,78Fed.Reg.69,907,69,908(Nov.21,2013)(“Asisthecasewithall Exchangeco-locationarrangements...(ii)useoftheco-locationservices proposedhereinwouldbecompletelyvoluntaryandavailabletoallUsersona non-discriminatorybasis....”).

197.SeeWhite,supranote3(“TheSECshouldnotrollbackthetechnology clockorprohibitalgorithmictrading,butweareassessingtheextenttowhich specificelementsofthecomputer-driventradingenvironmentmaybeworking againstinvestorsratherthanforthem.”).

198.E.g.,id.Regardingtheseproposals,SECChairWhitesaidshehas “instructedthestafftopreparerecommendationsfortheCommissionto improvefirms’riskmanagementoftradingalgorithmsandtoenhance regulatoryoversightovertheiruse.”Id.;seealsoARNUK&SALUZZI,supranote 131,at220–30(offeringproposalstoimprovethecurrentstateoftheequity markets);infranote357andaccompanyingtext(discussingproposalto institutebatchauctions).

199.See,e.g.,CharlesDuhigg,S.E.C.StartsCrackdownon‘Flash’ Trading,N.Y.TIMES,Aug.5,2009,atB1;NinaMehta,EquitiesIndustry ClashesoverFlashandStep-upOrders,TRADERSMAG.,July1,2009.TheU.S. SenateheldhearingsonHFTissuesincludingflashordersonOct.28,2009. DarkPools,FlashOrders,High-FrequencyTrading,andOtherMarket StructureIssues:HearingBeforetheSubcomm.onSec.,Ins.,&Inv.oftheS. Comm.onBanking,Hous.,&UrbanAffairs,111thCong.(2009).

200.See,e.g.,AlanChan,DoHighFrequencyTradersFront-Runthe MarketbyUsingTheirSpeedAdvantage?,FORBES(Apr.3,2014), http://www.forbes.com/sites/quora/2014/04/03/do-high-frequency-traders-frontrun-the-market-by-using-their-speed-advantage/;seealsoARNUK&SALUZZI,

acknowledgedthiscontroversywithits2009Release201 proposingtheeliminationoftheexceptiontoRegulationNMS Rule602thatpermitsflashorders,butnoactionhasbeen takensince.202

Aflashorderisanorderthatanexchangecannot immediatelyfillwhichisthensubmittedtocertainmarket participantsbeforeitistransferredtotheSIPpursuantto NMSRule602.203Rule602(a)(1)(i)(A)providesanexceptionto thegeneralrequirementthatanexchangesubmititsbestbids andofferstotheSIPinthecaseofbidsandoffersthatare “executedimmediatelyaftercommunication”or“cancelledor withdrawn after communication.”204The 602was implementedin1978205whenthemarketsoperatedalmost entirelywithout computerizedinformation technology.206Thisexceptionallowedanexchangeto“flash”its ordertofloorbrokerstoseeifitcouldbefilledbeforeitwas senttotheconsolidatedtape.207Boththepartysubmittingthe orderandthepartyfillingitwouldbenefitintheeventthe

immediately toRule

executed precursor

theuse

not

if

supranote131,at84(describingflashordersasallowinga“sneakpeek”into themarket).

201.EliminationofFlashOrderExceptionfromRule602ofRegulation NMS,74Fed.Reg.48,632(Sept.23,2009)(tobecodified17C.F.R.pt.242). FlashorderswereoriginallyallowedunderSECReleaseNo.34-54422.Order ApprovingaProposedRuleChangeandNoticeofFiling,71Fed.Reg.54,537 (Sept.15,2006).

202.ThisisprobablybecauseBATS,NASDAQ,andNASDAQOMXBX decidedtodiscontinueofferingflashordersasofSeptember1,2009asthey encounteredcontroversyandthepossibilityofgovernmentregulation.See EliminationofFlashOrderExceptionfromRule602ofRegulationNMS,74 Fed.Reg.at48,635.DirectEdgeofferedthemuntilFebruary2011.SeeJacob Bunge,DirectEdgetoStop‘Flashing’OrdersonMonday,WALLST.J.(Feb.25, 2011), http://www.wsj.com/articles /SB10001424052748703409304576166930877474292.Anestimated3.1%of totalequitytradingvolumewasexecutedbymeansofaflashorderinJuly 2009.SeeEliminationofFlashOrderExceptionfromRule602ofRegulation NMS,74Fed.Reg.at48,635.

203.17C.F.R.

§

242.602

(presenting

requirements

for

“[d]isseminationofquotationsinNMSsecurities”).

Id.§(a)(1)(i)(A).

204. 205.ForinformationontheprecursortothemodernRule602,see DisseminationofQuotationsforReportedSecurities,43Fed.Reg.4342(Feb. 1,1978)(formerlycodifiedat17C.F.R.pt.240). 206.SeeFoxetal.,supranote12,at198. 207.SeeEliminationofFlashOrderExceptionfromRule602ofRegulation

NMS,74Fed.Reg.at48,634–36.

orderwasfilled,andtheexchangebenefittedbyretainingthe transaction.208

Inthemoderncomputerizedcontext,thesameeconomic principlescanholdtrue,thoughtheeconomicsofflashorders aregreatlyaffectedbythemaker/takerpaymentsystem.209 Whenflashordersareusedinthecurrentenvironment,they are“flashed”toHFTtradersfor30to150millisecondsbefore theyaresenton,ifnotfilled,totheSIP.210Partiessubmitting anordercanbenefitbyitsimmediateexecution,oftenata betterpricethanwouldotherwisebeavailable,andresponders benefitbymakingatradeatapricetheybelieveis advantageous.211Furthermore, inamaker/takersystem, partieswanttheirordersflashedtopotentialrespondents,as theywillgarnerarebateforpostingaquoteanotherparty hits.212Thatsaid,flashordersmaybeusedinwayswidely consideredunfair.213Apartythatseesaflashorderisin essencegivenanearlypictureofwhattheNBBOforastockis likelytobe,sotheycanthengointothemarketandtradeon thebasisof theorderbeing incorporatedintotheNBBOdisseminatedbytheSIP.214This “sneakpeek”servesasgroundsfortheoftenheardcomplaint thatflashtradingenables“front-running.”215Whilethisis

informationaheadof

208.Seegenerallyid.at48,637–38(detailingthebenefitsofflashordersto

variousmarketparticipants).

209.SeeHarris&Namvar,supranote148,at11–13.HarrisandNamvar explainthatflashordersareindirectlydrivenbythemaker/takerpayment system.Id.Thisisbecause“liquidityrebatesallowMakerstoquotebetter pricesatmake-or-takeexchangesthantheywouldquoteattransactionfee exchanges,whichforcestransactionfeeexchangestorouteordersforthe benefitofcustomers,whothenreceivebetterpriceswithoutpayingtheaccess fee.”Id.at12.Flashordersarean“adhocsolution”tothepricingproblemsof exchangesthatchargetransactionfees,becausetheyallowthemtosidestep thispricingproblem.Id.

210.

Id.at4.Thelongestallowableintervalanexchangecanflashanorder is500milliseconds,orhalfasecond.SeeKeithSpence,FlashTrading,29B.U. REV.BANKING&FIN.L.98,100(2009).

211.SeeHarris&Namvar,supranote148,at5–6. 212.Seeid.at3;seealsosupranote148andaccompanyingtext. 213.SeeHarris&Namvar,supranote148,at9. 214.ElaineWah&MichaelP.Wellman,LatencyArbitrage,Market Fragmentation,andEfficiency:ATwo-MarketModel,14ASS’NFOR COMPUTING (2013), ON http://web.eecs.umich.edu/srg/wp-content/uploads/2013/02/ec38-wah.pdf.

COM.

215.OneofLewis’scentralthemesisthatagreatdealofHFTactivity amountsto“electronicfront-running.”LEWIS,supranote1,at89–127,172; accordKorsmo,supranote12,at558(characterizingfront-runningbyHFTas “parasitic”orderanticipation);Wah&Wellman,supranote214.Becauseof

(usually)nottechnicallytrue,asHFTtradersarenotgenerally thebrokersforclientssubmittingtheorder,andsodonothave adutytorefrainagainsttradingonthisinformation,216it otherwiseamountstothesamething.

Anotheraspectofthecriticismoverflashordersisthe moregeneralchargethattheyareafactorinthecreationofa “two-tiered”system,wherecertaininsideplayershaveaccessto fundamentallybetterinformationthaneveryoneelse.217There ismuchtothischarge,anditillustratesthebasicphenomena of“unfairness”thatmanyassociatewithHFT.218Inatwotieredsystem,certainplayerswillbegiveninformation—in caseofflashorders,priortoitsgeneralreleasetothepublic— thateveryoneelselacks.Sinceinformationisthebasisofall investment,thistimingadvantagewillallowforHFTtoearn risklessprofitsattheexpenseofthegeneralinvestingpublic.219

timedelaysinmessagestravellingfromexchangestotheSIP,andthento othertraders,HFTfirmshavethetechnicalabilitytouseinformationgained inonemarkettotradeinothermarketsbeforenon-HFTtraderscan;flash ordersareonlyonemechanismbywhichthiscanoccur.SeeWah&Wellman, supranote214.Thequestionforinvestors,regulators,andlegalscholarsis whetherthisisillegal.InsofarasmostsmallHFTfirmsarenottradingon behalfofclientsorcustomers,itwouldnotbeillegal,astheyarenotboundby anyrestrictionsthatapplytobroker-dealers.See,however,infranote235on thequestionofHFTandinsidertrading.Ontheotherhand,theproprietary tradingoperationsofthelargebankswouldlikelybeinviolationofthese obligations,particularlywhentheyaretradingaheadofcustomertradesin theirowndarkpools.

216.Seesupranote33(discussingfront-runningandsecuritieslaw).Itis importanttonotethatamulti-districtlawsuitagainsttheexchangesalleging, amongotherthings,liabilityunderExchangeActSection10(b)forenabling “electronicfront-running”bymeansofco-locationandproprietarydatafeeds, wasrejectedforfailuretostateaclaimunderFED.R.Civ.P.12(b)(6).SeeIn reBarclaysLiquidityCross&HighFrequencyTradingLitig.,No.14-MD-2589 (JMF),2015U.S.Dist.LEXIS113323,at*2–3(S.D.N.Y.Aug.26,2015). RelyingonCentralBankofDenverN.A.v.FirstInterstateBankofDenver N.A.,511U.S.164(1994),aswellasotherprecedent,thecourtdetermined thatmanipulativeschemeclaimspursuanttoSection10(b)mustbeprimary, notsecondary,violations,andthatmerelyprovidingservicestoHFTfirms thatpostquotesdoesnotqualifyasamanipulativeact.Seeid.at*42–47.

217.E.g.,EliminationofFlashOrderExceptionfromRule602of RegulationNMS,74Fed.Reg.48,632,48,636(Sept.23,2009)(tobecodified 17C.F.Rpt.242).

Id.at48,635–39.

218. 219.WhileprofitsofHFTappeartohavesufferedsignificantdeclinein recentyears,intheheydayofHFTthemostsuccessfulfirmsmadeconsistent andremarkablyrisk-freereturns.Baronetal.analyzetheprofitsmadeonEminiS&P500futurescontractsfromAugust2010toAugust2012,finding thatHFTfirmsachievedabnormallyhighSharperatiosintheirtrading, indicatingunusuallyhighrisk-adjustedreturns.SeeMatthewBaron,

exchange’s

toplayinone

Flashordersmayalsohaveimportantsecond-ordereffects ontheexchangesandbrokerspostingordersthere.Atbase,the motivationforanexchangetoissueflashordersmaybetoget sandboxover “HFTfirms another’s.”220Exchangescompeteintenselyfororderflow,and inamaker/takersystemtheyprovideasmallrebatefororders thatarefilled.InordertogetHFTfirmstopostquotesinits marketortofillordersthere,anexchangemayuseflashorders toenticesuchfirmstobepresentinthefirstplace.221While flashordersappeartohaveconsiderablebenefitsforboth suppliersandresponderstoquotes,theymayalsoinsome circumstancesworkagainsttheinterestsofquotesuppliers.222 Thiswouldthereforebeaninstanceofexchangesprivileging theirbestcustomers,theHFTfirms,attheexpenseoftheir lessimportantones.Finally,flashordersarepartofawebof forcespervertingtherelationshipofbrokerstotheirclients.223 Inordertogetcustomerorders,exchangesarewillingtopay brokersfortheirorders,whicharethenflashedtotheirbest

(May

JonathanBrogaard&AndreiKirilenko,RiskandReturninHighFrequency Trading manuscript), 2014) http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2433118(“[W]hileHFTs bearsomerisk,theirrisk-adjustedreturnsareunusuallyhigh.”).Inaddition tothetradingdata,reportsthatfirmssuchasTradebotandVirtuFinancial sufferedvirtuallynotradinglossesforyearsonendindicatethat,atleastfora time,certainfirmsmaderisklessprofits.SeeGandel,supranote8;Reuters, High-speedTraderVirtuFinancialEyes$2.6BillionValuationinIPO, FORTUNE(Apr.6,2015),http://fortune.com/2015/04/06/virtu-financial-ipo/. ThishaspromptedsomecommentatorstolikenthereturnsinHFTto economicrents.SeeCoffee,supranote86(“[S]omehighfrequencytraders havemadeenormousprofitsthatresembleeconomicrents.”);WallaceC. Turbeville,CracksinthePipelinePartTwo:HighFrequencyTrading,DEMOS (Mar.8,2013),http://www.demos.org/publication/cracks-pipeline-part-twohigh-frequency-trading(“ThesefindingspersuasivelysuggestthatHFTdoes notmerelyprovideliquiditytothemarket,butratherextractsvaluewell beyondthevalueofliquidityprovided.”);Foxetal.,supranote12,at241–44 (explaininghowslowmarketarbitragebyHFT“addsathirdparty,the liquiditysupplier,whoseonlysocialpurposeistofacilitatetradesbetween regulartradersbutwhoinsteadistheonlygainerfromtheso-calledarbitrage activity,”andthat“[t]heregularpracticeofHFTexploitationofdarkpool midpointordersprovidesrentstoHFTs”).

220.ARNUK&SALUZZI,supranote131,at84. 221.Seeid.at75(flashordersasafavortoHFT);seealsoHarris&

Namvar,supranote148,at5–6(benefitsofflashorders).

222.SeeHarris&Namvar,supranote148,at2(althoughinmost instances“Submitters”wouldbeexpectedtobenefitthroughflashorders,they doopenthemselvesuptobeingfront-runby“Responders”).

223.See,e.g.,ARNUK&SALUZZI,supranote131,at58–59.

clients.224Thepaymentfororderflowmeansthatbrokersmay notbeactingstrictlyintheinterestsoftheirclients.225Rather, theyareactingtosomeextentintheirowninterestsby allowingthedecisionofwheretosendacustomerordertobe influencedbypaymentfromanexchange.226Insofarasthey constituteareasonthatexchangeshaveforpayingfororder flow, flashordersseemtobeacorruptinginfluenceon brokers.227

D.DIRECTANDENRICHEDDATAFEEDS

Flashordersandthecontroversysurroundingthemare justoneinstanceofhow,intheprocessofsolvingexisting marketproblems,theexchangescreateotheronesintheform ofinformationaldisparities.Insomecases,thesolutionsare enabledbyaprovisionofRegulationNMSthatexistedbefore thecomputerizationofthemarkets,asinthecaseoftheflash orderexceptionandtheintermarketsweeporder.228The “problem”solvedbydirectdatafeedsontheotherhandisa verybasiconecreatedbytheintersectionofRegulationNMS withtheinformationtechnologyinfrastructureoftheSIP:the NBBOmandatedbyRegulationNMSwillalwaysbeafew milliseconds,orevenmicroseconds,behindthetruestateofthe marketasreflectedinthepricesatthevariouscomputer serversthatconstitutetherealnationalmarketsystem.229Asa result,traderswillwantaccesstothepricesatthesametime orevenbeforetheyaresenttotheSIPforinclusioninthe NBBO.230Directdatafeedsaretheexchanges’productthat

224.Seeid.at59(linkingflashorderstopaymentfororderflow).On paymentfororderflowintheHFTenvironment,seeScottPatterson,TD AmeritradeExecutiveSaysOrdersGotoVenuesthatPayHighestFees,WALL ST.J.,June17,2014,http://www.wsj.com/articles/td-ameritrade-executivesays-orders-go-to-venues-that-pay-highest-fees-1403043559.

225.SeePatterson,supranote224. 226.SeeARNUK&SALUZZI,supranote131,at58–59. 227. 228.ForfurtherdiscussionofISOs,seeinfraSectionII.F. 229.SeeYoon,supranote12,at925n.74(statingthatdirectdatafeeds willbefivetotenmillisecondsfasterthantheNBBOdisseminatedbythe SIP).

230.

Id.at925(“Anarbitragestrategyinvolvesidentifyingandcapturing thepricediscrepancies....”);seealsoLarryTabb,IstheIncreasedSpeedof DirectMarketDataWorththePrice?,WALLST.&TECH.(Oct.20,2006), http://www.wallstreetandtech.com/is-the-increased-speed-of-direct-marketdata-feeds-worth-the-price/d/d-id/1258127?(describingthecostsandbenefits ofinvestingindirectdatafeeds).

meetsthisdemand.231Dataisreleaseddirectlytoclientsat preciselythesametimethatitissenttotheSIP;giventhe inherentdelaysintransmissionandprocessingbytheSIP,a directdatafeedisthereforeafewmillisecondsaheadofthe SIP.232Enricheddatafeeds,ontheotherhand,providea greaterbreadthofinformationthanissenttotheSIP.233As withmanyofthetechniquesbywhichHFToperates,both directandenricheddatafeedsimplicateverybasicquestionsof fairnessintheelectronicmarketplace.234Unlikemanyofthe othertechniques,however,questionsconcerningthelegalityof thetransactionstheyenablehavealsobeenraised.235

231.SeeYoon,supranote12,at925n.74(“[T]raderscanreceivethe informationintheirindividualdatafeedsprovidedbyexchangesandATSs beforetherestofthemarket.”).

232.

Id.;seealsoDingetal.,supranote190,at318–21(contrasting differencesinlatencybetween“thepublic(SIPNBBO)andproprietary (syntheticNBBO)data”).

233.See,e.g.,StopthePress–ExchangesSeeNoUnfairness!,THEMIS TRADINGBLOG(Sept.27,2012),http://blog.themistrading.com/2012/09/stopthe-press-exchanges-see-no-unfairness/(“Weallknowafewthingsaboutthe StockExchangesenricheddatafeeds.Weallknowtheyhavemore information–likeorderIDs,cancelreplaceinformation,orderidentifiersand modifiers,andofcoursehiddenorderinformation.”);seealsoARNUK& SALUZZI,supranote131,at111–17.

234.See,e.g.,StopthePress–ExchangesSeeNoUnfairness!,supranote

233.

235.Regulators,legalacademics,andplaintiffshaveallraisedtheissueof insidertradingviolationsresultingfromdirectandenhanceddatafeedsfrom theexchangestotheirHFTclients.See,e.g.,FY2015BudgetHearing:U.S. DepartmentofJustice:HearingBeforetheSubcomm.onCom.,Just.,Sci.& RelatedAgenciesoftheH.Comm.onAppropriations,113thCong.(2014) (statementofEricHolder,Att’yGen.oftheUnitedStates);MercerE.Bullard, HFTIsn’t theProblem—InsiderTradingIs,CNBC(Apr.4,2014), http://www.cnbc.com/id/101553082#;Coffee,supranote86(“[I]fadarkpool leakedinformationtoapreferred,highfrequencytraderabouttheliquidityin itspool,thenaninsidertradingviolationmayhaveoccurred.Material informationwouldhavebeenpassedbythedarkpooltothehighfrequency trader,arguablyinreturnforhighfeespaidbythetrader.Thatwouldsatisfy evenaconservativedefinitionofinsidertrading.”);EricT.Schneiderman,N.Y. StateAtt’yGen.,KeynoteAddressattheNewYorkLawSchoolSymposium: InsiderTrading2.0–ANewInitiativetoCrackDownonPredatoryPractices (Mar.18,2014),http://www.ag.ny.gov/pdfs/HFT_and_market_structure.pdf. TheplaintiffinCityofProvidencev.BATSGlobalMarkets,Inc.,14-CV-2811 (S.D.N.Y.Apr.18,2014)included“insidertrading”violationsinitscomplaint at¶¶124–28,butthecomplaintwasdismissed.InreBarclaysLiquidityCross &HighFrequencyTradingLitig.,No.14-MD-2589(JMF),2015U.S.Dist. LEXIS113323,at*2–3(S.D.N.Y.Aug.26,2015)(dismissingconsolidated litigationincludingCityofProvidence).

Directdatafeedsinvolvetheprovisionofpriceinformation totradersforhigher,“premium”feesthanthosepaidtothe ConsolidatedTapeAssociation(orConsolidatedQuote Association)fortheNBBO.236Examplesofdirectdatafeeds includeNYSEBestQuote,NYSEOpenBook,NYSEAmexBest Quote,andNASDAQITCH.237TheSECallowsexchangesto disseminatedatatotheirtraderclientssimultaneouslywith theprovisionofsuchdatatotheSIP,238despitethefactthat thisinevitablycontributestotheproblemofthe“staleness”of theNBBO.239TheSEChasdeterminedthatthesimultaneous releaseofmarketdatatotradersandtheSIPisconsistentwith therequirementofRule603(a)thatinformationbereleasedon termsthatare“fairandreasonable”and“notunreasonably discriminatory.”240WhiletheSEC’spositionislogical,ithas struggledwithitsimplicationsintheHFTenvironment,where anysmalldifferenceintimingleadstoaninformational advantagethatcannotbeovercomebyslowertraders.241 Furthermore,atleastoneexchange,theNYSE,wasfoundtobe violatingtheexistingrulebysendingdataonitsproprietary datafeedsaheadofitstransmissiontotheSIP,242andtiming disparitiesinthereleaseoffinancialdatawereattheheartof theunsuccessful lawsuitsagainsttheexchangesforHFT practices.243

236.SeeARNUK&SALUZZI,supranote131,at101–02. 237.Seeid.at101–03;Easleyetal.,supranote189,at1n.3. 238.SeeSECConceptReleaseonEquityMarketStructure,75Fed.Reg.

3594,3611(Jan.14,2010)(codifiedat17C.F.Rpt.242).

Id.at3608.

239. 240.17C.F.R.§242.603(a)(2015);seealsoFoxetal.,supranote12,at271 (exploringaninterpretationofRule603(a)(2)inwhichsendingthedatasignal simultaneouslytoHFTfirmsandtheSIPis“unreasonablydiscriminatory... [since]itispredictablethatsome[endusers]willconsistentlyreceiveitfaster thanothers”).

241.SeeSECConceptReleaseonEquityMarketStructure,75Fed.Reg.at 3611.FrankPasqualedrawsattentiontoananalogousproblemwiththe releaseofmaterialinformationbycompaniestothemarket.SeePasquale, supranote12,at2106n.114.WhileRegulationFDrequiresthatcompanies issuingmaterialinformationreleaseitatthesametimetoeveryone,newsreleasedistributorsselldirectaccesstosuchinformationtoHFTfirms,which resultinthemreceivingthe“informationslightlybeforetherestofthe market.”Id.

242.SeeN.Y.StockExch.,LLC,Admin.Proc.FileNo.3-15023(Sept.14, 2012), Jenny http://www.sec.gov/litigation/admin/2012/34-67857.pdf; Strasburg&ScottPatterson,NYSEIsFinedinDataProbe,WALLST.J.,Sept. 15,2012,atA1.

243.Complaintat7–8,Lanierv.BATSExch.,Inc.,105F.Supp.3d353 (S.D.N.Y2015)(No.14-CV-3745).Similartothe2012NYSEprobe,seesupra

Theproprietary,“enriched”datafeedsofferedbythe exchangesalsoofferagreaterbreadthofinformationthan thoserelyingsolelyontheNBBOdisseminatedbytheSIPcan obtain.WhereastheNBBOdisseminatedbytheSIPoffersonly thebestbidandbestoffer,enricheddatafeedswillofferan exchange’sinformationonothercustomers’orders,including ordercancellations,modifications,andexecutions.244Thiseven includesinformationconcerningordertypesthatcustomers believedwerehiddenfrompublicview,so-called“hiddenorder” types.245AHFTfirmmakingasubstantialinvestmentinthe requiredsoftwarecanuseanenricheddatafeedtodiscernthe directionofthemarketandthentradeasquicklyaspossibleon thatbasis.246Similartoflashorders,enricheddatafeedsallow forwhatwouldbetermed“front-running”ifengagedinbya broker-dealer.247Accordingtocritics,enricheddatafeedsallow high-frequencytraderstoeffectively“reverseengineer”the forcesthatledtoaparticulartransaction(orthefailureto consummateatransaction),offeringavaluableinsightintothe market’sdirection.248Andinsofarastheexchangespaybrokers fortheirorderflowtoofferinformationconcerningtheseorders toHFTfirms,therelationbetweenabrokeranditsclientsis alsocompromised.

Thequestionsoffairnessraisedbyproprietarydatafeeds areessentiallysimilartothoseraisedbyco-locationandflash orders:theexchangesofferonegroupofparticipants,who possessthewherewithalandinterestinprofitingfromdata, accesstothisdatainreturnforhighfees.Whatarethe expectationsofmarketparticipantswhentheyenterthestock market?Isthecreationofatwo-tieredsystem,evenwhenall are,inprinciple,permittedtoenterintotheprivilegedcohort,

note242,theplaintiffinLanierallegedthatdataisprovidedbythedefendant exchangesaheadofitstransmissiontotheSIP.Seeid.at13–14.TheLanier complaintwasdismissedonfederalpreemptiongrounds,aswellasfailureto stateaclaimunderFED.R.CIV.PRO.12(b)(6).SeeLanier,105F.Supp.3d 353,357.

244.ARNUK&SALUZZI,supranote131,at111–17. 245. 246.

Id.at113–15. Id.at112;seealsoid.121–22(statingthatsomeexchangesareeven

disseminatingmachine-readablenews).

247.Forarecapof“front-running”anditsimplications,seesupranote215

andaccompanyingtext.

248.SeeARNUK&SALUZZI,supranote131,at102(statingthat informationinproprietarydatafeeds“isvaluablebecauseitenableshigh frequencytraderstomodelthebehaviorofinstitutionalandretailinvestor orderstopredicttheshort-termpricemovementsofstocks”).

fair?Isitinevitable?Orshouldlimitationsbeputuponthe informationthatexchangesarepermittedtodisseminate, eitherforafeeorforfree?ShouldtheSECreconsideritspolicy thatexchangesmaydisseminatetheirinformationtoHFT clientssimultaneouslywithitstransmissiontotheSIPand insteadfocusonthetimeinformationisreceivedbyvarious marketplayers?Andwhatabouttheeffectsenricheddatafeeds mayhaveontherelationshipofbrokerstotheircustomers? Directandenricheddatafeedsgiverisetomanyquestionsthat concernthefundamental fairnessofthemarketsandthe creationofatwo-tieredsystem.

E.SPOOFING,LAYERING,QUOTESTUFFING,ETC.:HFTAS MARKETMANIPULATION

Thenext

typeofHFTactivityinvolvestheuseof informationtechnologytoengageinpractices thatare essentiallyveryhigh-speedversionsofthe“pump-and-dump” schemeslongknowntoregulators.249Theseactivitiesdonot relyonanintrinsicadvantageinspeedordatagiventothem byathirdparty,suchastheexchange;rather,theyinvolvethe useofHFTtechnologytocreatecertainconditionsinthe market, trading programs,whichtheoriginalHFTfirmsthentakesadvantage of.ItisasifHFTcreatestheweatherinthefinancial ecosystemandthenprofitsfromit.250

oftenmanipulating

algorithmic

other

Themostbasictypeofthisactivityinvolvescreating momentumthroughpostingordersthatarethencancelled beforetheycanbeactedupon.Suchstrategiesareoften referredtoas“momentumignition”strategies,astheyinvolve inducingotherplayersinthemarkettoactinacertainway, drivingthepriceofasecurityupordown.251Asimpleexample ofthismightbeiftheNBBOforGeneralMotorsstockconsists ofabidat$15.10andanofferat$15.14.“Spoofing”would involvemakinganoffertobuy(abid)at$15.12,thenexecuting theoppositetransaction,sellingthesecurityatthisprice,after otherplayersinthemarkethaveraisedtheirbidsinresponse

249.SeeKirilenko&Lo,supranote125,at66–68;Korsmo,supranote12, at551–57.ForlimitationsontheabilityofHFTtoengageinpump-and-dump strategieswithintheEuroboundfuturesmarket,seeIreneAldridge,Can High-FrequencyTradersGameFutures?,7J.TRADING75(2012).

250.SeeKorsmo,supranote12,at548–49(discussingmomentumignition

strategies).

251.Seeid.at548.

tothehigheroffer.252Finally,theoriginaloffertobuyat$15.12 willbecancelledbeforeotherpartiescanactonit.253Inthis waypricesinthemarketaremanipulatedhigherorlower withoutthemanipulatingpartyhavingtoactonthequotesit hasplacedonanexchangeorothertradingvenue.Thefact thatanestimated95%ofordersenteredontomarketsare cancelledmayillustratehowfrequentlythistechnique,along withquote-stuffing,isused.254

“Layering”involvesessentiallythesametechniquebutina seriesofsuccessivesteps.Here,aseriesofsuccessivelyhigher (orlower)quotesareplacedononesideofthemarketinorder todriveapriceupordown;afterthepricehasbeen manipulated,aquoteisthenplacedontheoppositesideofthe market.255Thepurposeofthistechniqueistotrickother traders’algorithmsintodetectingapatterninthemarket,and tothentakeadvantageoftheirreaction.256Forinstance,

252.Kirilenko&Lo,supranote125,at66–67;Goldsteinetal.,supranote 180,at187.Section747(“AntidisruptivePracticesAuthority”)oftheDoddFrankActof2010,whichadded§4c(a)(5)totheCommoditiesExchangeAct (CEA),prohibitsspoofinginthecommoditiesmarkets.SeeDodd-FrankWall StreetReformandConsumerProtectionActof2010,Pub.L.111-203,§747, 124Stat.1376,1739(codifiedasamendedat7U.S.C.§6c(a)(5 ;D.Denis Aktas,Spoofing,33B.U.REV.BANKING&FIN.L.89,93–95(2013).Boththe CommodityFuturesTradingCommissionandtheSEChaveactively prosecutedenforcementactionsagainsttradersallegedtobespoofing.See HuwJones&JohnMcCrank,U.S.andU.K.FineHigh-SpeedTraderfor Manipulation,REUTERS(July22,2013),http://www.reuters.com/article /2013/07/22/us-britain-fca-hft-idUSBRE96L0G620130722(notingthata$6 millionfinewasleviedbyregulatorsagainstPantherEnergyTradingLLCfor spoofing).TheSEChaslongprosecutedspoofingcases.See,e.g.,Visionary Trading,LLC,ExchangeActReleaseNo.71871,2014WL1338258(Apr.4, 2014);LeonidShpilsky,ExchangeActReleaseNo.17221,2001WL1408740, at*2(Nov.5,2001).ThefirstcriminalconvictionunderCEA§4c(a)(5)(C),of PantherEnergyTradingLLCprincipalMichaelCoscia,occurredonNov.3, 2015inUnitedStatesv.Coscia,No.14-CR-00551(N.D.Ill.2015).SeeBradley Hope,U.S.WinsConvictioninFirstSpoofingCase,WALLST.J.,Nov.3,2015, atC3.

253.SeeKirilenko&Lo,supranote125,at66–67. 254.ARNUK&SALUZZI,supranote131,at32;Easleyetal.,supranote165, at24(“[E]stimatesby[the]TABBgroupputthis[ordercancellations]at 98%.”).Fordiscussionofthelinkednatureofsubmissionsandcancellations, seeJoelHasbrouck&GideonSaar,Low-LatencyTrading,16J.FIN.MKTS. 646,658–64(2013).Forthecontributionquote-stuffingmaymaketothe cancellationrate,seeJiadingGai,ChenYao&MaoYang,TheExternalitiesof HighFrequencyTrading6–10(Aug.7,2013) (unpublishedmanuscript), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2066839.

255.SeeSHORTER&MILLER,supranote126,at23;Korsmo,supranote12,

at555.

256.SeeKorsmo,supranote12,at555.

assumethepriceofamid-capstockwithalowtradingvolume hasabestbidof$6.25andabestofferof$6.75.Layeringwould involvepostingabidof$6.28,thenpostingabidof$6.30,and thenpostingabidof$6.32.Afterotherpartiesthenfollowthe risingprice,makingbidsat,say,$6.30,theoriginalpartywill sellthestockinthemarketatthishigherpriceititselfhas induced.Finally,itwillcancelitsearlierbidsthatarestill open.Layeringoperatestotrickotheralgorithmictradersinto actingonquotesthatarenevermeanttobefilled.257TheSEC hasprosecutedpartiesforlayeringactivities;258theyare essentiallythesameaspump-and-dumpschemes,althoughthe entiresequenceofactivitycanoccurinlessthanhalfa second.259

Finally,“quotestuffing”involvesusingrapid-firequotesto overwhelmthecapacityofexchangeserversorothertradersto processthem.260AsCharlesKorsmoexplains,thisactivityhas atwo-foldpurpose:sincethetradermakingtheburstofrapid firequotescansafelyignorethem,whilethealgorithmsof othertraderswillnotknowtodothis,theoriginaltrader engaginginquotestuffingcanslowhiscompetitorsdown, therebygaininganadvantageoverthem.261Quotestuffingcan alsobedonetoslowdowntheprocessoftransmittingprices fromanexchangetotheSIP,therebycreatingarbitrage opportunitiesbetweenmarkets.262Quotestuffinginvolvesthe

e.g.,

257.See,

id.at556(“[T]heTrilliumcaseshowshowmore sophisticatedHFTscanpotentiallyexploitothertradersentirelythroughthe useoforderstheHFTsneverintendtoexecute.”).

258.SeeKirilenko&Lo,supranote125,at66–67(describingSECactions againstHoldBrothersOn-LineInvestmentServices);seealsoPressRelease, Sec.&Exch.Comm’n,SECChargesNewYorkBasedHighFrequencyTrading FirmwithFraudulentTradingtoManipulateClosingPrices(Oct.16,2014) (announcingthatAthenaCapitaluseditsalgorithm“Gravy”tomanipulate pricesofstockduringthefinalsecondsofthetradingday);Korsmo,supranote 12,at555n.155(notingthatTrilliumBrokerageServiceswassubjectto FINRAsanctionsforengaginginrepeatedinstancesoflayering).

259.Kirilenko&Lo,supranote125,at67(providingexampleofmarket

manipulationoccurringwithin839milliseconds).

260.Korsmo,supranote12,at575–76;seealsoGaietal.,supranote254, at6(“‘[Q]uotestuffing’...involvessubmittingaprofusenumberofordersto themarkettogeneratecongestionsonpurpose.”).

261.SeeKorsmo,supranote12,at575. 262.Therapidcancellationofquotesin“quotestuffing”delaysthespeedat whichtradeinformationisreported,thusdistortingtheinformationsentto otherparticipantsandcreatingarbitragepointsforinvestorsawareofthe delay.SeeGaietal.,supranote254,at27–28.

useofHFTtechnologytoincreasetheopportunityforlatency arbitrageactivities.

F.SPECIALORDERTYPES:“HIDENOTSLIDE,”ISOS,AND JUMPINGTHEQUEUE

Theproliferationofspecialordertypesisafinalsetof phenomenathattopsoffthecomplaintsaboutmarketfairness andHFT.263Giventhemarketmicrostructureconditions dictatedbyRegulationNMS,specialordertypesareintegralto thesuiteoftechniquesHFTfirmsusetowringprofitsfromthe equitymarkets.264Whiletheseordertypesareostensiblyopen toall,theycontaincertainfeaturesthat,inconjunctionwith thehightechnicalbarrierstotheiruse,giveanunfair advantagetoHFTintheeyesofmanyobservers.265 Specifically,theyallowtheiruserstojumpthequeueincertain situationsthatenablethemtotradeorcollectaliquidity provisionrebateinarisk-freemanner.266Whileanumberof ordertypeshavecomeunderfireasethicallyquestionable,267 andeachexchangehasitsownvariant,twoofthemost

J.

19,

19,

263.Forarticlesdetailingthecomplaintsconcerningspecialordertypes, seegenerallyHerbertLash,ComplaintsRiseoverComplexU.S.StockOrders, REUTERS 2012),http://www.reuters.com/article/2012/10/19/usexchanges-ordertypes-idUSBRE89I0YU20121019;ScottPatterson&Jenny Strasburg,ForSuperfastStockTraders,aWaytoJumpAheadinLine,WALL ST. 2012), http://www.wsj.com/articles/SB100008723963904439892045775992436935616 70;TomSteinert-Threlkeld,OutofOrder;NewWaystoBuyandSellShares onExchangesAreArrivingDaily.AreControlsNeededonChoice,Complexity andCompetition?,TRADERSMAG.,Jan.2013,at20.HFTpractitioner-turnedwhistleblowerHaimBodekplayedacentralroleinpublicizingtherolespecial ordertypesplayinHFT;hisstoryiscentraltothenarrativeofPatterson’s DarkPools.PATTERSON,supranote11.Onspecialordertypesgenerally,as wellasBodek’scontributiontothisissue,seeStanislavDolgopolov,HighFrequencyTrading,OrderTypes,andtheEvolutionoftheSecuritiesMarket Structure:OneWhistleblower’sConsequencesforSecuritiesRegulation,2014 J.L.TECH.&POL’Y145,147–54(2014).Specialordertypeshaveattractedthe attentionofCongressaswellasregulators.SeeComputerizedTrading:What ShouldtheRulesoftheRoadBe?:HearingsBeforetheSubcomm.onSec.,Ins. &Inv.oftheS.Comm.onBanking,Hous.&UrbanAffairs,112thCong. (2012);White,supranote3.

264.SeeDolgopolov,supranote263,at148–51. 265.See,e.g.,HaimBodek,HFTCheckmate—TheAlphainOrderTypes, [hereinafterBodek,HFTCheckmate],

TABBF. http://tabbforum.com/opinions/hft-checkmate-the-alpha-in-an-order-type.

2013)

21,

266.SeeDolgopolov,supranote263,at157. 267.Bodek,HFTCheckmate,supranote265.

controversialare“HideNotSlide”(or“HideandLight”)orders andIntermarketSweepOrders.268

TheHideNotSlideorderisdesignedtosolveaproblem createdbyRule610(d)ofRegulationNMS.269Rule610(d) prohibitsquotationsthat“lock”or“cross”theNBBO.Locking occurswhenabuyorder(bid)ispostedatthesamepriceasan ordertosell(oroffer)elsewhereinthenationalmarketsystem, oranofferpostedatthesamepriceasabid.270Crossingoccurs whenabidispostedatahigherpricethanthebestoffer,oran offeratalowerpricethanthebestbid.271TheHideNotSlide orderallowstraderstosidesteptheprohibitionagainstlocks andcrosses.272Insodoing,italsoallowsatradertojump aheadinthequeueoforders,violatingtheprincipleofpricetimepriority.273

WhenabrokerpostsaHideNotSlideorder,ifitlocks anothermarket,theexchangewill“slide”thepricedowntothe nextavailableprice.274Forexample,a$15.50HideNotSlide bidthatlocksanothermarketwouldbeautomaticallychanged to$15.49,therebyavoidingalock.Then,whenthemarket unlocks,theoriginalbid$15.50willbeplacedfirstinline.The HideNotSlideorderaccomplishestwothings.First,itallowsa usertoavoidslidingdowntoalowerbid(orhigheroffer)when abidorofferlocksorcrossesthemarket.275Second,itallows theuser,whenthemarketunlocks,tobeatthetopofthe queue.276

Thisqueue-jumpingfeatureisthekeytotheHideNot Slideorder.277ThepartysuccessfullyusingaHideNotSlide

268.SeeSugatoChakravarty,PankajJain,JamesUpson&RobertWood, CleanSweep:InformedTradingThroughIntramarketSweepOrders,47J. FIN.&QUANTITATIVEANALYSIS415(2012);HaimBodek,LockedMarkets, PriorityandWhyHFTsHaveanAdvantage:Part2:Hide&Light,TABBF. (Oct. Markets], http://tabbforum.com/opinions/locked-markets-priority-and-why-hfts-have-anadvantage-part-2-hide-and-light(explaining“HideandLight”).

[hereinafter

Locked

Bodek,

2012)

16,

269.17C.F.R.§242.610(d)

(prohibiting“[l]ockingorcrossing

quotations”).

270.SeeSchmerken,supranote89,at14. 271. 272.Bodek,LockedMarkets,supranote268. 273.Seeid. 274.SeeScottPatterson&JennyStrasburg,How‘HideNotSlide’Orders

Work,WALLST.J.,Sept.19,2012,atA12.

275.Bodek,LockedMarkets,supranote268. 276. 277.

Id. Id.

orderisnowfirstinline,abletotradeatthenewbestpricein themarket,andtopositionitselfadvantageouslytocapture anyavailableliquidityrebatesif theexchangeusesa maker/takersystem.278Thisqueue-jumpingfeatureledto significantcontroversy,buttheexchangescouldcounterthat itsusewasinfactopentoall.279Inadditiontoitsunusual characteristics,themannerinwhichtheHideNotSlideorder wasintroducedalsoledtocriticism.HaimBodekclaimsthat thedocumentationabouttheseordertypeswasinsufficient, andthattheexchangesselectivelydisclosedtheirfeaturesto theirprivilegedHFTclients.280Theseallegationsappearto have recent AdministrativeOrderagainstEDGAExchange,Inc.andEDGX Exchange,Inc.281AlthoughtheRespondentsdonotadmitor denytheSEC’sfindings,theyagreedtopaya$14millionfine for,amongotherthings,notaccuratelydisclosingtoallmarket participantshowtheHideNotSlideordertypeworked.282

beensubstantiatedby

theSECinits

AnotherkeyorderforHFTactivityistheIntermarket SweepOrderor“ISO.”283TheISO284isspecificallylistedin

278.

ItisimportanttonotethatonlythefirstpartyusingaHideNotSlide orderwilljumpthequeue—allotherswillbebehindtheslidregularorders. ThisfeaturegivesrisetoparticularsuspicionthattheHideNotSlideorder wasspecificallyformulatedtobenefitHFTtraders,asitwouldonlyafforda benefittotheoneluckyfirmthatwasfirstinlineofallfirmsplacingHideNot Slideorders.ThissuspicionappearstohavebeenconfirmedbytheJanuary 12,2015SECAdministrativeOrderagainstEDGAExchange,Inc.andEDGX Exchange,Inc.,thesuccessorstotheDirectEdgeexchange.SeeEDGA Exchange,Inc.,ExchangeActReleaseNo.74032,2015WL137640,at*3(Jan. 12,2015);seealsoPATTERSON,supranote11,at339;O’Hara,supranote148, at262.

279.SeePatterson&Strasburg,supranote263(“Theexchanges’position isthatthese[advantages]arefullydisclosed;theycanbeusedbyanyonewith therighthardwareandtechnicalsavvy....”).

280.SeeBodek,LockedMarkets,supranote268.ButseeMacKenzie,supra

note6,at44(HFTinterviewees“deniedthattheseordertypesweresecret”).

281.EDGAExchange,Inc.,2015WL137640,at*11. 282.SeeBradleyHope,BATStoPay$14MilliontoSettleDirectEdge 2015), ST.

Order-Type http://www.wsj.com/articles/direct-edge-exchanges-to-pay-14-million-penaltyover-order-type-descriptions-1421082603.

Case,

12,

J.

283.FortheimportanceofISOstoHFT,seeMacKenzie,supranote6,at 42–45.Fortheirappealtotraders,seeChakravartyetal.,supranote268,at 416(“ThebenefitsofISOusageincludefasterexecutionspeedandtheability tocapturelargercounterpartydepthbyconcurrentlysubmittingordersto severalmarkets....”).FortheroleofISOsinthe“flashcrash”ofMay6, 2010,seeMadhavan,supranote153,at23(“[T]hese[ISO]ordersmayhave triggeredtheflashcrashbyaggressivelytakingbid-sideliquidity.”).

Rule611(b)(5)asanexceptiontoRule610’sprohibitionon trade-throughs.285Thestatedpurposeofthisexceptionisto facilitatetheexecutionofblocktradesbybroker-dealerson behalfofinstitutionalclients.286AnISOallowsforthepurchase orsaleofstockatpricesinferiortothoselistedintheNBBOin ordertoaccessquotesdeeperinthelimitorderbookofan exchange.287Forexample,assumetheNBBOforGMstockis $18.00bestbidand$18.03bestoffer.ExchangeXhas3,000 sharesofGMatthebestofferprice,andTraderBwantsto purchase20,000shares.TraderBcansubmitanISOfor20,000 GMshares,purchasingthe3,000attheNBBOprice,andthen goingdeeperintothelimitorderbooktopurchasemoreshares atinferiorprices.288

TherationaleforISOsinthepre-electronictradingworld asfacilitatingblocktradesiseasytounderstand.Eventhough anISOmaycausetheclienttopayahighertotalpricefora blockofsharesthanitwouldwerethosesharesallavailableat theNBBO,itmaywishtoacquireagreateramountofstock thanisimmediatelyavailableattheNBBO.289Inthiscase,the clientlikelyfearstheeffectofablocktradeontheprice.Inthe worldofelectronictrading,however,anISOoffersanumberof otherpotentialadvantages.First,inamarketwherespeedis

284.17C.F.R.§242.600(b)(30)

(defining“[i]ntermarketsweep

Id.§242.611(b)(5).

order”). 285. 286.RegulationNMS,70Fed.Reg.37,496,37,536(June29,2005)(codified at17C.F.R.pt.242)(“Thisexceptionparticularlywillfacilitatetheimmediate executionofblockordersbydealersonbehalfoftheirinstitutionalclients.”); seealsoChakravartyetal.,supranote268,at417–19(discussing“TheISO Mechanism”);StavrosGadinis,MarketStructureforInstitutionalInvestors: ComparingtheU.S.andtheE.U.Regimes,3VA.L.&BUS.REV.311,350–51 (2008).

287.SeeGadinis,supranote286,at350. 288.SeeFredandEthelCalledandWantedtoKnowAboutIntermarket SweepOrders(ISOs),THEMISTRADINGBLOG(Dec.6,2015)[hereinafterFred andEthel],http://blog.themistrading.com/fred-and-ethel-called-and-wanted-toknow-about-intermarket-sweep-orders-isos/.17C.F.R.§242.611(c)specifies that“[t]hetradingcenter,broker,ordealerresponsiblefortheroutingofan intermarketsweepordershalltakereasonablestepstoestablishthatsuch ordermeetstherequirementssetforthin§242.600(b)(30).”Simultaneously withanISO,§242.600(b)(30)requirestradersusinganISOtopurchasethe fulldisplayedsizeofsharesonothermarketsthatconstituteprotectedoffers orbids,asapplicable,atpricessuperiortothelimitpriceoftheISO.17C.F.R. §242.600(b)(30).

289.Gadinis,supranote286,at350–51(“Despitethesecostlyfeatures, ISOsofferthemoststraightforwardmethodofexecutinglargetradesinthe NationalMarketSystem.”).

paramount,itmayallowtraderstotradeatbetterpricesthan thoseontheSIP,whentheSIPisstale.290Second,exchanges offervariantsontheISOsuchasthe“PostOnlyISO”thatwill onlybeusedinsituationswherethetraderisattemptingto capturearebateinamaker-takerpricingmarket.291Third, someISOordertypesallowuserstojumpthequeuejustas HideNotSlideordersdo.292Inshort,ISOscanbevaluableina varietyofcircumstances,andtheyareakeyordertypeforHFT firms.

III.ANETHICSOFALGORITHMICTRADING

Thetechniquesusedbyhighfrequencytradersoutlined abovegeneratedintensecontroversyuponthepublicationof MichaelLewis’sFlashBoys.293Whilethecontroversywas encapsulatedintheperhapsoverstatedclaimthatthemarkets were“rigged,”acloselookatsomeoftheobjectionable techniquesofHFTsupportsthenotionthatsomeHFT practicesareindeedunfair.294ThepurposeofPartIIIisto

290.SeeChakravartyetal.,supranote268,at416(“ThebenefitsofISO

usageincludefasterexecutionspeed.”).

291.SeeFredandEthel,supranote288. 292.Seeid. 293.SeeLEWIS,supranote1,at2. 294.TheacademicliteraturetodatepresentssomediscussionofwhyHFT isoftenperceivedasunfair.See,e.g.,Angel&McCabe,supranote194,at 586–91(describingvarioustechniquesusedbyHFTs);Dolgopolov,supranote 263,at149–50(citingselectivedisclosureofspecialordertypesasespecially unfair);Goldsteinetal.,supranote180,at194–95(“Thefairnessofveryhighspeedtrading”);Kirilenko&Lo,supranote125,at60,67–68(advantages conferredbytechnologicalexpertiseasgivingrisetofairnessconcerns); Korsmo,supranote12,at528(characterizingthe“speedandtechnological sophistication”ofHFTfirmsas“generatinganappearanceofunfairness”); O’Hara,supranote148,at269(“[T]hegreatercomplexity,lowertransparency, andhigheruncertaintyofhighfrequencymarketsallcontributetoasense thatmarketscanbemorefairforsomethanforothers.”);McGowan,supra note12,¶45(discussinghowcoststoretailtradersinmaker-takerpayment systempromptfairnessconcerns);Gomberetal.,supranote49,at34–36 (explainingfairnessconcernsasdrivenbythespeedadvantagesderivedfrom co-location);MacKenzie,supranote6,at5–6(attributingthehostility surroundingHFTtoitshiddennature).TheSECalsoquestionsthefairnessof investmentsinITandhumanresourcesthateffectivelyresultinatwo-tiered marketplace.ConceptReleaseonEquityMarketStructure,ExchangeAct ReleaseNo.34-61358,75Fed.Reg.3594,3605(Jan.14,2010)(codifiedat17 C.F.R.pt.242)(“[I]situnfairformarketparticipantstoobtainacompetitive advantagebyinvestingintechnologyandhumanresourcesthatenablethem totrademoreeffectivelyandprofitablythanothers?”);cf.CharlesJones,What DoWeKnowAboutHigh-FrequencyTrading?42(Colum.Bus.Sch.,Research PaperNo.13-11,2013),http://ssrn.com/abstract=2236201(“[M]anyofthe

explainwhymanyintheinvestingpublicreactedastheydid uponlearningaboutHFT.Inotherwords,PartIIIoffersa theoreticalgroundingtosupporttheclaimthatsomeofHFT’s techniquesareunfair.

thetransformationof

ItisimportantattheoutsettoacknowledgethatHFTis simultaneouslypartof theequity marketsthathasgreatlyreducedtransactioncostsformany investors,andsohasbroughtgreatbenefits.295Manydefenders ofHFTrepeatedlypointthisout,andanumberofacademic studiesdocumentmarketbenefitsattributable,atleastinpart, toHFT.296However,ageneralrebuttaltotheunfairness argumentthatHFThasloweredcostsformanymarket participants,eveniftrue,missesthepoint.HFTmaybepartof atransformationof themarketsbenefittingtheaverage investoratthesametimethatsomeofitsmoreaggressive techniquescommitobjectionablewrongs.297Wemaybeforced to“pickourpoison,”toparaphraseProfessorPirrong,298butis itwrongtohopefornopoisonatall?Itisimportanttoexplain thelegitimatebasesforcomplaintsandtoseehowHFT implicatesanumberofissuesofwiderconcerntothefinancial system.

bedetailedenoughto

InordertoevaluatethemoralstatusofthevariousHFT strategies,aframeworkfornormativeanalysisisrequired.The frameworkmust ina determinativefashiontofinancialmarketsactivity,atthe sametimethatitshouldconnecttoatleastoneofthemajor ethical informeconomicandpublicpolicy discussions.Infact,relativelylittlehasbeenwrittenon specificallyethicalissuesinfinance.299Thisislikelybecause

schools

apply

issuesassociatedwithHFTarethesameissuesthatariseinmoremanual markets.”).Inadditiontotheacademicliterature,thenarrativesofMichel LewisandScottPattersonarelargelydrivenbyfairnessconcerns.LEWIS, supranote1;PATTERSON,supranote11.

295.Jones,supranote294,at51(“Basedonthevastmajorityofthe empiricalworktodate,HFTandautomated,competingtradingvenueshave substantiallyimprovedmarketliquidityandreducedtradingcostsforall investors.”).

296.Seeid.;seealsoinfraSectionIII.A. 297.See,e.g.,Korsmo,supranote12,at528–29,577–80(suggestingthat even

the ofHFTcan“inadvertently—or deliberately—causeextremevolatilityeventssuchastheFlashCrash”).

technological

advantages

298.SeePirrong,PickYourPoison,supranote70,at14. 299.SeeJohnR.Boatright,EthicsinFinance,inFINANCEETHICS3,3 (JohnR.Boatrighted.,2010)(“[T]heacademicstudyoffinanceethicshas receivedsurprisinglylittleattentionfromscholarsineitherfinanceor

implicitinfinancialmarketsdiscussion,andeconomicsmore generally,isarelativelywell-definedethicalposition.Thisisof courseutilitarianism.300Whileeconomicsisnotnecessarily

themostrelevant

businessethics.”).WithinFINANCEETHICS, tothis discussionofHFTareRobertW.Kolb,EthicalImplicationsofFinance,in FINANCEETHICS,supra,at23;JohnDobson,BehavioralAssumptionsof Finance,inFINANCEETHICS,supra,at45;andEugeneHeath,Fairnessin FinancialMarkets,inFINANCEETHICS,supra,at163.Inthewiderbusiness ethicsliterature,seealsoJOHNR.BOATRIGHT,ETHICSINFINANCE(3ded. 2014);BOUDEWIJNDEBRUIN,ETHICSANDTHEGLOBALFINANCIALCRISIS (2015);Angel&McCabe,supranote194,at585;JohnDobson&KenReiner, TheRationalityofHonestyinDebtMarkets,22MANAGERIALFIN.20,31–32 (1996)(characterizinghonestborrowersasnecessaryforthefunctionofdebt marketequilibrium);JamesO.Horrigan,TheEthicsoftheNewFinance,6J. BUS.ETHICS97(1987)(presentingnormativeideasimplicitinthe“New Finance”);TimothyC.Johnson,ReciprocityasaFoundationofFinancial Economics,131J.BUS.ETHICS43,43(2014)(arguingforreciprocityasa fundamentalprincipleofcontemporaryfinance);HershShefrin&Meir Statman,Ethics,FairnessandEfficiencyinFinancialMarkets,49FIN. ANALYSTSJ.21,21–22(1993);MeirStatman,RegulatingFinancialMarkets: ProtectingUsfromOurselvesandOthers,65FIN.ANALYSTSJ.22,22–24 (2009).Forarticlesexploringthelinksbetweenethics,trust,andfinancein thelegalliterature,seeDanAwreyetal.,BetweenLawandMarkets:IsThere aRoleforCultureandEthicsinFinancialRegulation?,38DEL.J.CORP.L. 191(2013);MargaretM.Blair&LynnA.Stout,Trust,Trustworthiness,and theBehavioralFoundationsofCorporateLaw,149U.PA.L.REV.1735,1737– 39(2001)(arguingfortheimportantroleoftrustasaninternalconstraintin thecorporation);TamarFrankel,RegulationandInvestors’Trustinthe SecuritiesMarkets,68BROOK.L.REV.439(2002)(examining“therelationship betweeninvestors’trustandgovernmentregulationofthemarkets”);PeterH. Huang,Trust,Guilt,andSecuritiesRegulation,151U.PA.L.REV.1059,1159 (2003);KevinT.Jackson,TheScandalBeneaththeFinancialCrisis:Gettinga ViewfromaMoral-CulturalMentalModel,33HARV.J.L.&PUB.POL’Y735 (2010);IanB.Lee,EfficiencyandEthicsintheDebateAboutShareholder Primacy,31DEL.J.CORP.L.533,535–37(2006);SusannaKimRipken,The DangersandDrawbacksof theDisclosureAntidote:TowardaMore SubstantiveApproachtoSecuritiesRegulation,58BAYLORL.REV.139,203–04 (2006);LynnA.Stout,TheInvestorConfidenceGame,68BROOK.L.REV.407, 408(2002)(investortrustasafoundationofthesecuritiesmarkets);JohnH. Walsh,ASimpleCodeofEthics:AHistoryoftheMoralPurposeInspiring FederalRegulationoftheSecuritiesIndustry,29HOFSTRAL.REV.1015,1016 (2001).Forcriticalappraisalsoftheroleofethicalconceptsinexplaining financialphenomena,seeRobertT.Miller,MoralsinaMarketBubble,35U. DAYTONL.REV.113(2009);LarryE.Ribstein,Lawv.Trust,81B.U.L.REV. 553(2001).

300.SeeKennethJ.Arrow,SomeOrdinalist-UtilitarianNotesonRawls’s TheoryofJustice,70J.PHIL.245,246(1979)(“Theimplicitethicalbasisof issomeversionofutilitarianism.”).Fora economicpolicyjudgment philosophical economics, examination utilitarianism,andpolicyscience,seeJeanneL.Schroeder,TheStumbling Block:Freedom,Rationality,andLegalScholarship,44WM.&MARYL.REV. 263(2002).

relationship

between

utilitarian,asJudgePosnerpointsout,301itisconventionally so.

Economicsbearsthestampofutilitarianismbecauseits prescriptivecounselscenteraroundchoosingthepathofaction thatresultsinthegreatestamountof“socialwelfare.”302Social welfareisusuallydefinedasthestateofaffairsresultinginthe greatestamountofutilityforagivenpopulation.303Inorderto makethisconceptmoresusceptibletomeasurement,subjective andamorphouselementssuchasfairness,justice,andmorality arecommonlyexcludedfromitsdefinition.304Theconceptof socialwelfareisimplicitlyutilitarianbecauseitdefinesthe goodasthemaximizationofapossibleamountofutility.305 SuchadefinitionhearkensbacktoJohnStuartMill’s“Greatest HappinessPrinciple,”inwhich“actionsarerightinproportion astheytendtopromotehappiness,wrongastheytendto producethereverseofhappiness.Byhappinessisintended pleasure,andtheabsenceofpain;byunhappiness,pain,and theprivationofpleasure.”306Socialwelfarethereforebridges thedividebetweenethical inquiryandthedisciplineand techniquesofmoderneconomics,andinsodoing,illuminates therootsofeconomicsinutilitarianism.307

its

appeal,

Despite

thenatural

of and utilitarianismandeconomics,utilitarianismfacesanumberof powerfulobjectionsatthetheoreticallevel,aswellasobstacles toitsapplicationinpractice.Twoareparticularlyrelevantto thediscussionofHFT.Thefirstistheproblemofdetermining theappropriateframeofreferenceforthedistributionofcosts

affinity

301.SeeRichardA.Posner,Utilitarianism,Economics,andLegalTheory,

8J.LEGALSTUD.103(1979).

302.Schroeder,supranote300,at351(“Inconventionaleconomicanalysis, normativeanalysisisnodifferentfromprescriptiveanalysis,sincethegoalof thelegalsystemistomaximize‘socialwelfare’....”)(quotingChristineJolls, CassR.Sunstein&RichardH.Thaler,ABehavioralApproachtoLawand Economics,50STAN.L.REV.1471,1474–75(1998 .

597(2004).

304.Seeid.at664–65(presentingtheconceptofsocialwelfareasanarrow

aggregateofutilities,excludingfairnessanddistributionalconcerns).

305.SeeAmartyaSen,UtilitarianismandWelfarism,76J.PHIL.463,464 (1979)(“‘[W]elfarism’...[as]theprinciplethatthegoodnessofastateof affairsdependsultimatelyonthesetofindividualutilitiesinthatstate,and— moredemandingly—canbeseenasanincreasingfunctionofthatset.”).

306.JOHNSTUARTMILL,Utilitarianism,in10COLLECTEDWORKSOF

JOHNSTUARTMILL203,210(JohnM.Robsoned.,1969)(1861).

307.SeeSHAVELL,supranote303,at608–12(discussing“Welfare

EconomicsandNotionsofMorality”).

andbenefits:astraightforwardutilitarianismonlycompares thesumtotalofpleasuresandpainsinvariouspossiblestates of theworld,neglectingtheirdistributionthroughthe populationingeneral.308Aworldwhereoneindividual experiencedintensepleasureandacompleteabsenceofpain whilefourotherslivedmiserablycouldbedeemedsuperiorto oneinwhichfiveindividualslivedverymoderatebutrelatively lives.Astraightforwardutilitarianismtherefore equal countenancesthesacrificeofcertainindividuals,ortheir interests,forthegreatergood.309ThediscussionofHFTfaces ananalogousframingproblem:dothegeneralmarketbenefits itmaybringcompensatefor,orevencancelout,thewrongsit commitsagainstspecificinvestors,orclassesofinvestors,in specificinstances?Whatistheappropriateframeofreference forevaluatingcostsandbenefits?Isittheequitytrading marketasawhole?Doesn’tthisimplythattheinterestsof someinvestors,orclassesofinvestors,maybesacrificedforthe greatergood,intheformoftheefficiencyofthelargertrading system?Whilewedonotwanttoignorethebenefitsofthe currentmarketsystem,itseemsobtusetodenythatsometimes HFToperatesinamannerthatisunfair.310

Second,theattempttopututilitarianismintopracticein manycontemporarysettingsishamperedbythecomplexityof causalrelationshipsamongcomponentsofagivensocial, technological,orecologicalsystem,andtheconcomitant difficultyofmeasuringinputsandoutputsinsuchasystem.311 Forexample,inthefieldoffinance,standardeconomicpolicy prescriptionsaresubjecttodebatebecauseoftheinherent thesystemandthedifficultyofassessing complexityof whetheragivenpolicyinfactresultsinanincreaseinsocial

308.SeeSamuelScheffler,Introduction,inCONSEQUENTIALISMANDITS CRITICS1,2–3(SamuelSchefflered.,1988)(pointingouttheobjectionthat “[p]rovidednetaggregatesatisfactionismaximized...utilitarianismis indifferentastohowsatisfactionsanddissatisfactionsaredistributed”);John inCONSEQUENTIALISMANDITSCRITICS Rawls,ClassicalUtilitarianism, supra,at14,17(“Thusthereisnoreasoninprinciplewhythegreatergainsof someshouldnotcompensateforthelesserlossesofothers....”).

309.SeeScheffler,supranote308. 310.E.g.,RISHINARANG,INSIDETHEBLACKBOX283(2ded.2013)(“In sum,it’shardtoseeanymeritintheideathatHFTisunfairorcreatesatwotieredmarketplace.”).

311.See,e.g.,JeffreyN.Gordon,TheEmptyCallforCost-BenefitAnalysis inFinancialRegulators,43J.LEGALSTUD.S351(2013)(emphasizingthe difficultyofmakingpredictionsaboutthefinancialsystem,whichisitself constructedbyregulationandsubjecttounpredictablesecond-ordereffects).

welfare.312Whileitisonlyapartofthemuchlargerfinancial system,thesesameproblemsaffecttheattempttoofferpolicy prescriptionsconcerningHFT.Weknowthatalargeamountof equitytradingoccursinthe“dark”markets—thoughexactly howmuchwedon’tknow—whileatthesametimethesystem of regulators,and regulationissufficientlycomplexthatdefinitivestatements abouttheoperationofcauseandeffectinthissystemareoften impossible.

tradingvenues,marketparticipants,

Becausethecurrentequitytradingmarketspresent considerableobstaclestothesuccessfulimplementationofthe standardeconomicframeworktakenfromutilitarianism,Part IIIpresentsanalternativeethicalframeworkforfinancial marketsactivity.Whileeconomiccostsandbenefitsaretaken intoaccountinthisframework,313atitsbaseistheprincipleof reciprocity,whichreflectsafundamentallydeontological ethicalstance.Uponthisfoundationwecanconstructthe additionalprinciplesoffreedomfrommisrepresentation,alevel playingfield,andinstitutionalintegrity.Whiletheexactextent oftheseprinciplesissubjecttodisagreement,itislikelythat

312.SeeJohnC.CoatesIV,Cost-BenefitAnalysisofFinancialRegulation: CaseStudiesandImplications,124YALEL.J.882,887(2015)(“[C]asestudies showthatquantified...[cost-benefitanalysis]amountstonomorethan ‘guesstimation,’entailing:(a)causalinferencesthatareunreliableunder standardregulatoryconditions;(b)theuseofproblematicdata;and/or(c)the samekindsofcontestable,assumption-sensitivemacroeconomicorpolitical modelingusedtomakemonetarypolicy.”);seealsoGordon,supranote311,at S351(“Insteadofweighingcostsandbenefits,financialregulationnecessarily isbasedonaseriesoftrade-offsofnormativelyderivedvalues,whichmay entailprinciplesofpragmaticdesign.”).ButseeRobertB.Ahdieh,Reanalyzing Cost-BenefitAnalysis:TowardaFrameworkofFunction(s)andForm(s),88 N.Y.U.L.REV.1983,1983–84(2013)(proposingaframeworkforcost-benefit analysisofregulations);EricPosner&E.GlenWeyl,Benefit-CostAnalysisfor FinancialRegulation,103AM.ECON.REV.393,393(2013) (offering suggestionsforhow“regulatorybenefitscanbequantified”);CassR.Sunstein, Cost-BenefitAnalysisandtheKnowledgeProblem(Mossavar-RahmaniCtr.for Bus.&Gov’t,Harv.KennedySch.,Harv.U.,WorkingPaperNo.RPP-2015-03, 2014)(presentingasolution-baseddefenseofcost-benefitanalysisdespitethe “knowledge problem”), http://papers.ssrn.ssrn.com/sol3/papers.cfm?abstract_id=2508965.

313.SeeBOATRIGHT,supranote299,at26–29(providingaframeworkfor financeethicsthatincludestheutilitarianconceptofwelfareaswellasduties, rights,andfairness);Shefrin&Statman,supranote299,at23(“Policymakers operateasiftheyhaveutilityfunctionsthatdependonbothefficiencyand fairness.”).Thisapproachimplicatesthewellknown“equity/efficiencytradeoff.”SeeARTHURM.OKUN,EQUALITYANDEFFICIENCY:THEBIGTRADEOFF (1975).

theirveryreal,yetimperfect,instantiationineconomiclife allowsfortheflourishingofourfinancialmarkets.314These principlesalsoallowustounderstandwhycriticsclaimthat HFTis“notfair”andinwhatsensetheyarecorrect.First, however, itisimportanttosurveythefinanceliterature exploringthecostsandbenefitsofHFT.

A.COSTSANDBENEFITSOFHFT

InresponsetotherapiddevelopmentofHFToverthe courseofthepastdecade,numerousacademicstudieshave attemptedtoquantifythebenefitsandcostsofHFT.The effectsofHFTonpricediscovery,liquidity,spreads,andstock marketvolatilityarethesubjectsofdetailedstudies,aswellas HFT’spotentialroleincreatingasociallywasteful“armsrace” amongtradersandheighteningsystemicriskinthefinancial system.BenefitscommonlydiscussedrelatetotheroleHFT playsinadvancingpricediscoveryandtheprovisionof liquidity;alternatively,costsrelatetoliquidity,technological armsraces,adverseselection,andcontributiontosystemic risk.

Contributiontotheprocessofpricediscoveryisfrequently claimedtobeabenefitofHFT.InHighFrequencyTradingand PriceDiscovery,JonathanBrogaard,TerrenceHendershott, andRyanRiordanexaminetransactionleveldatafrom2008to 2009for120randomlyselectedNASDAQandNYSEstocks, dividingthemintolarge,medium,andsmallcapcategories.315 DividingtradersintoHFTandnon-HFTcategories, the authorsfindthatHFTcontributestopricediscoverythrough sellinginthedirectionofpermanentpricechangesandaway fromtransitorypricingerrors.316Overall,HFTtradersbenefit

314.See,e.g.,TamarFrankel,TrustingandNon-Trusting:Comparing Benefits,CostandRisk(B.U.Sch.L.,WorkingPaperNo.99-12,1999), stable http://papers.ssrn.com/sol3/papers.cfm?abstract_id=214588 financialsystemrequiresthesupportoffiduciarylawprinciples,judicial enforcement,andmonitoringbyindependentregulatoryagencies.”);seealso LawrenceE.Mitchell,FairnessandTrustinCorporateLaw,43DUKEL.J.425, 425(1993)(citing“trustinthecorporatefiduciary’sfairness”asaprecondition ofthecorporateenterprise);Stout,supranote299,at408(“Investortrust providesthefoundationonwhichtheAmericansecuritiesmarkethasbeen built.”).Seealsoinfranote456andaccompanyingtext.

(“[A]

315.JonathanBrogaard,TerrenceHendershott&RyanRiordan,High FrequencyTradingandPriceDiscovery,27REV.FIN.STUDIES2267,2271 (2014)[hereinafterBrogaardetal.,PriceDiscovery].

316.

Id.at2280;seealsoJonathanBrogaard,AlCarrion,Thibaut,Ryan Riordan,AndriyShkilo&KonstantinSokolov,HighFrequencyTradingand

significant

frompricediscovery,317buttheinformationHFTusesislimited toathreetofoursecondspan.318Consequently,“[i]fthis informationwouldbecomepublicwithoutHFTs,thenthe potentialwelfaregainsmaybesmallornegativeifHFTs impose longer-term investors.”319Similarly, inastudyoftheforeignexchange marketsbetween2003and2007,AlainChaboud,Benjamin Chiquoine,ErikHjalmarsson,andClaraVegafindthat algorithmictradingimprovespriceefficiency.320Theyalso determinethatalgorithmictradingincreasestheefficiencyof pricesthroughliquidityprovision,notfromreactiontoposted quotes.321

selection

adverse

on

Perhapsevenmoreimportantly,HFTissaidtobenefit measuresofmarketplaceliquidity.TerrenceHendershott, CharlesJones,andAlbertMenkveldstudyagroupof943 stocksfromFebruary2001throughDecember2005inDoes ImproveLiquidity?322TheNYSE AlgorithmicTrading introducedits“Autoquote”systemoverthecourseofJanuary 29,2003toMay27,2003.323Thisprovidesanimportant opportunitytotesttheeffectsofalgorithmictradingonthe marketbecauseAutoquoteenabledalgorithmictraderstoplace quotesonamarketthatpreviouslyoperatedmanually.324The authors“usetherateofelectronicmessagetrafficasaproxyfor

Extreme http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2531122.

PriceMovements

2014)

21,

2011)

317.SeeTerrenceHendershott&RyanRiordan,AlgorithmicTradingand manuscript), Information http://faculty.haas.berkeley.edu/hender/atinformation.pdf (“AT[(algorithmic trading)]contribute[s]moretotheefficientpricebyhavingmoreefficient quotesandATdemandingliquiditysoastomovethepricestowardsthe efficientprice.”).Foratheoreticalmodelwiththissameimplication,seeBruno Biais,ThierryFoucault&SophieMoinas,EquilibriumHighFrequency Trading27(Sept.2011)(unpublishedmanuscript),http://papers.ssrn.com/sol3 /papers.cfm?abstract_id=1834344(“HFTordersimpoundinformationinto pricesfaster.”).

318.Brogaardetal.,PriceDiscovery,supranote315,at2302. 319. 320.AlainP.Chaboud,BenjaminChiquoine,ErikHjalmarsson&Clara Vega,RiseoftheMachines:AlgorithmicTradingintheForeignExchange Market,69J.FIN.2045,2045(2014).

Id.at2048.

321. 322.TerrenceHendershott,CharlesM.Jones&AlbertMenkveld,Does

AlgorithmicTradingImproveLiquidity?,66J.FIN.1,6(2011).

Id.at14.

323. 324.Seeid.at13(describingthepurposeofAutoquote).

theamountofAT[algorithmictrading]”325andthereductionin effectivespreadsasaliquiditymeasure.326Theyconcludethat “ATdoesinfactimproveliquidityforlarge-capstocks.”327 Likewise,HasbrouckandSaarlookatNASDAQdatafortwo groupsofstocksfromOctober2007andJune2008.328Focusing on“strategicruns”ofmessagesrangingfromthreetomore thanten,theyfindthat“higherlow-latencyactivityimplies lowerpostedandeffectivespreads,greaterdepth,andlower short-termvolatility.”329Interestingly,theyfindthattheeffect ofHFTactivityonmeasuresofmarketqualityincreasedinthe June2008sample,whichwasatimeofheightenedmarket stresscausedbythefirsttremorsofthefinancialcrisisof 2008.330Theysuggestthat“[i]tisreasonabletoassumethat highervolatilitycreatesmoreprofitopportunitiesforhighfrequencytraders.”331

Despitethesepapersdetailingcertainbenefits,some financialeconomistspresentresultsthatappeartosupportthe popularcriticismsofHFTregardingliquidityvoicedbySal ArnukandJosephSaluzzi,332MichaelLewis,333andScott Patterson.334AndreiKirilenkoandAndrewLonotethat“[i]n contrasttoanumberofpublicclaims,high-frequencytraders donotasaruleengageintheprovisionofliquiditylike traditionalmarketmakers.Infact,thosethatdonotprovide liquidityarethemostprofitableandtheirprofitsincreasewith thedegreeof‘aggressive,’liquidity-takingactivity.”335

Morespecifically,economistVincentvanKervelexamines tradingdataforagroupoftenrandomlyselectedFTSE100

325. 326. 327.

Id.at6. Id.at7–10. Id.at3;cf.EricBudish,PeterCramton&JohnShim,TheHigh FrequencyTradingArmsRace:FrequentBatchAuctionsasaMarketDesign Response manuscript), http://faculty.chicagobooth.edu/eric.budish/research/HFTFrequentBatchAuctions.pdf(arguingthatwhileliquidityhasimprovedin recentyears,itisprobablyfromtheincreasedroleofinformationaltechnology as“thereislittlesupportforthepropositionthatthespeedracepersehas improvedliquidity”).

4,

Id.at667. Id.at670. Id.at673.

328.Hasbrouck&Saar,supranote254,at648. 329. 330. 331. 332.ARNUK&SALUZZI,supranote131,at27–28. 333.LEWIS,supranote1,at107–09. 334.PATTERSON,supranote11,at253. 335.Kirilenko&Lo,supranote125,at60.

stockstradingontheLondonStockExchangeaswellasthe Chi-X,BATS,Turquoise,andNASDAQOMXEuropevenues.336 Hefindsthatatradeononevenueresultsinsubstantial cancellationsoflimitordersontheothervenues:“within100 millisecondsafteratradeonsomevenues,39–85%oftheorder sizeiscancelledoncompetingvenues.Afteronesecondthis numberincreasesto98–125%,whichshowsthattheimpactof atradeonliquidityisinfacttwicethetradesize.”337He thereforeconcludesthat“theinterrelationofliquidityacross tradingvenuescausessubstantialoverestimationofliquidity aggregatedoverthesetradingvenues.”338

JiadingGai,ChenYao,andMaoYelookattheveryfastest tradingenvironments,whereordersoccuratmicrosecond(and evennanosecond)339speeds.340Theycontrast theHFT environment,wheretimeisinfinitelydivisiblebutpricesare constrainedbyaminimum$0.01ticksize,withtheidealworld ofWalrasianeconomicswhere“priceisinfinitelydivisiblebut timeisnot.”341IntheHFTworld,then,thecompetitionfor speedisaconsequenceofafailedpricecompetition,andan increaseinspeeddoesnotchangethecostof liquidity— althoughitdoesdecidewhosuppliesthisliquidity.342Looking atavarietyofdatasetsincludingNASDAQTotalViewITCH, theauthorsconcludethatinthenanosecondenvironment, “[o]urempiricalresultssuggestthattheimpactofspeedon liquidityisinsignificant.”343

Tworecentpapersinthelegalliteraturealsocastdoubt upontheallegedbenefitsofHFTonshareprices.InTheNew StockMarket:SenseandNonsense,344MerrittFox,Lawrence Glosten,andGabrielRauterbergarguethat“electronicfront running”harmsinformedtraders345andso“[t]heeliminationof electronicfrontrunningwouldmakeitmoreprofitablefor

336.VincentvanKervel,Liquidity:WhatYouSeeIsWhatYouGet?3,18– (unpublishedmanuscript),http://www.rsm.nl/fileadmin

19(May2012) /home/Department_of_Finance__VG5_/LQ5/VanKervel.pdf.

Id.at29. Id.at2.

337. 338. 339.Amicrosecondisonemillionthofasecond,andananosecondone

billionth.Ananosecondistoonesecondasonesecondisto31.71years.

340.Gaietal.,supranote254,at2. 341. Id.at4. 342.Seeid.at16. 343. 344.Foxetal.,supranote12. 345.

Id.at230–32.

Id.at25.

347Likewise,

thesetraderstoengageintheiractivity....Asaresult,prices willbemoreaccurate.”346Thiswouldbenefittheeconomyat largebyimprovingtheallocativeefficiencyofthecapital marketsandthesignalsgiventomanagement,boards,and inHowAlgorithmicTrading investors. UnderminesEfficiencyinCapitalMarkets,YeshaYadav focusesonharmtotheaccuracyofpricesandtheconsequences forthefinancialregulatorysystematlarge,whichhasreliedin recentdecadesontheassumptionthatstockmarketprices accurately informationconcerning available securities.348Whilethesepapersdonotpresenteconometric analysesofHFTdata,theydopresenttheoreticalarguments thatcallintoquestionthebeliefthattheeffectsofHFTon priceswillbeonnetsociallybeneficial.

reflect

TheideathatHFTmaybeatbaseasociallywastefularms raceisbehindoneofthemostcelebratedrecentpapersonHFT, TheHigh-FrequencyTradingArmsRace:FrequentBatch AuctionsasaMarketDesignResponse.349EricBudish,Peter Cramton,andJohnShimlookatmarketdatafromtheChicago MercantileExchangeandtheNYSErelatingtotworelated securities,theSPDRS&P500exchange-tradedfund(SPY)and theS&P500E-minifuturescontract(ES),fromtheperiod 2005–2011.350TheyfindthatwhilethepricesoftheSPYand theESarealmostperfectlycorrelatedinnormalhumantime, atthemillisecondlevelatwhichHFToccursthecorrelation breaksdowncompletely.351ThisbreakdownallowsfortheHFT practiceof“sniping,”pickingoffstalequotesasanarbitrage activity.352Mostimportantly,whilethetimehorizonover whichthearbitrageopportunitiesexistshrinksdrasticallyover theperiodstudied,thevalueofthearbitragedoesnot.353 SimilartoGai,Yao,andYe,Budish,Cramton,andShim believethatthecurrentmarketdesignforcestradersintoa sociallywastefularmsracetoavoidbeingpickedoff.354HFT firmsfaceaprisoner’sdilemma:theywouldallbebetteroffnot

Id.at234. Id.

346. 347. 348.Yadav,supranote12,at1607–08. 349.Budishetal.,supranote327. 350. 351.

Id.at2. Id.at2–3,12(Figure1.1,“ESandSPYTimeSeriesatHuman-Scale

andHigh-FrequencyTimeHorizons”).

352. 353. 354.

Id.at5. Id.at2. Id.at1.

engaginginsuchacompetition,butavoidingitindividually woulddoomafirmtofailure.355Budishandothers,arecareful topointoutthattheirmodeldoes“notimplythatonnetHFT hasbeennegativeforliquidityorsocialwelfare,”butsimply “thatsnipingisnegativeforliquidityandthatthespeedraceis sociallywasteful.”356Theirpaperproposestheintroductionof batchauctionsevery100millisecondsasamechanismthat couldundercutthearmsraceandforcetraderstocompeteon price,notspeed.357

WhilethecostsofHFTarmsracesarebornedirectlyby thecompetitorsinthetournamentforspeed,HFTimposes substantialadverseselectioncostsonallthosewhochoosenot toenterthiscompetition.358Theseslowertradersaretypically institutionalinvestorsandtheirbrokers.359Theymayattempt toavoidtheworstofthecostsbyusingsmartorderroutersto sendlargeorders,orbytradingindarkpools,buttheypay highercostsbecausetheycannotcompeteintheraceforspeed. Avarietyofeconomistshaveexploredadverseselectioncostsin theHFTcontext,360thoughcomingtoareliableestimateofthe totalamountwouldseemtobeimpossible.

Finally,HFTmaycontributetosystemicriskinthe financialsystem,althoughitsexactroleasacauseisfarfrom clear.Itappearstohavebeenacontributingfactorinthe “FlashCrash”ofMay6,2010,thoughnotafundamental cause.361HFTisalsosuspectedofcontributingtothenumerous

Id.at35. Id.at37. Id.at39–42.

355. 356. 357. 358.SeeBrogaardetal.,PriceDiscovery,supranote315,at2268;seealso

Budishetal.,supranote327,at35.

359.SeeFoxetal.,supranote12,at231. 360.See,e.g.,Brogaardetal.,PriceDiscovery,supranote315,at2268(“We showthatHFTsimposeadverseselectioncostsonotherinvestors.”);Easleyet al.,supranote189,13–14(includingan“adverseselectioneffect”intheir modelwhichultimatelymakesthemarketlessliquid);LinTong,ABlessingor aCurse?TheImpactofHighFrequencyTradingonInstitutionalInvestors31 http://papers.ssrn.com/sol3 (Oct. /papers.cfm?abstract_id=2330053(“Ifindstrongevidencethatanincreasein HFTisassociatedwithanincreaseinthetradingcostsofinstitutional investors.”).

361.Kirilenkoetal.,supranote153,at2;seealsoKirilenko&Lo,supra note125,at62–63.Forafarmoreskepticalreading,seeNanex,Reexamining 2014), HFT’sRole www.nanex.net/aqck2/4650.html.

intheFlashCrash,NANEXRES.

12,

“miniflashcrashes”thathaveoccurred.362HFTcancontribute tosystematicriskbyspeedinguptheresponsetofinancial informationandremovingtheabilityofhumanstosenserisk inthetradingenvironment,ineffectputtingtradingsystems onanextremelyfastautopilot.

Inthefinalanalysis,itishardtoknowwhetherHFT providesanetbenefittothefinancialmarketsbecauseofthe complexityofthetradingenvironmentandtheincredible difficultyofaccuratelyestimatingcostsandbenefits.363It seemssafetoconcludethatHFTispartoftherevolutionthat greatlyloweredtransactioncostsforinvestors,although decimalizationandthegeneralapplicationof information technologytothefinancialmarketswouldalsoappearto accountformuchofthebenefit.Ontheotherhand,itseems clearthatHFTimposessignificantadverseselectioncostson slowertraders,andtheliquiditybenefitsitprovidesaresubject itsbenefitstotheprice toimportantcaveats.Moreover, discoveryprocessmaybequestionableif itonlyimparts informationtothemarketsafewsecondssoonerthanitwould otherwisearise.364Itappearsimpossibleinpractice,ifnot theory,toanswerthequestion“IsHFTgoodorbad?”usinga conventionaleconomicanalysisthatreliesonenumerationof costsandbenefits.

362.SeeBowley,supranote157;NeilJohnson,ZhaoGuannan,Eric Hunsader,JingMeng,AmithRavinder,SpencerCarran&BrianTivnan, FinancialBlackSwansDrivenbyUltrafastMachineEcology1(Feb.7,2012) (unpublishedmanuscript),http://arxiv.org/abs/1202.1448.

363.CraigPirrong’sconclusionontherelationofinformedtouniformed

traderscapturesthispoint:

Icannotseeanywayofevaluatingthewelfareeffectsoffinancial trading,andinparticularinformedtrading.Thesocialbenefits (howdomoreinformativepricesimprovetheallocationofreal resources)areimpossibletoquantify:theyareoftendifficulteven toidentify,exceptinthemostgeneralway(‘capitalallocationis improved’).Unlikethetradeformostgoodsandservices,thereis noreasontobelievethatsocialandprivatebenefitsalign.

CraigPirrong,Pinging:WhoIsthePredator,andWhoIsthePrey?, 2014), STREETWISE http://streetwiseprofessor.com/?p=8340.

(Apr.

364.SeegenerallyJosephE.Stiglitz,TappingtheBrakes:AreLessActive MarketsSaferandBetterfortheEconomy?2–3(Fed.Res.BankofAtlanta, https://www.frbatlanta.org Apr. /news/conferences/2014/140415-fmc/agenda.aspx of (surveying excessivelyactivemarkets).

issues

2014)

15,

B.THEULTIMATUMGAME,RECIPROCITY,ANDTHEDUTCH BOOK

Modernstockmarketspresentahighlyunusualcontextfor ethicalanalysis.Onascalerepresentingthedegreeof interpersonalinteractiontheyareatanextreme,evenwhen comparedtootherarm’s-lengthbargainingcontextssuchas corporatetransactions, commercialsales,oremployment negotiations.365Stockmarketspresentanenvironmentdevoid offace-to-facecontactbetweencounterparties,animportant contextualfactorinmoralanalysisthathasalsohadimportant legalsignificance.366Stockmarketsdominatedbyalgorithmic

365.See,e.g.,MacKenzie,supranote6,at58(“Algorithmicmarketsin whichmostactorsarethemselvesalgorithmsarethemostdepersonalizedof currentmarket studyMacKenzie distinguishesbetweenthreetypesofmarkets:

thebeginningofhis

forms.”).At

1.Marketactorsareallhumanbeings,and‘themarket’involves directinteractionamonghumanbeings.2.Themarketisan algorithm,buttheactorsremainmostlyhumanbeings;they interactwiththemarketviacomputerscreen,keyboardand mouse.3.Themarketisanalgorithm,andmostactorsinitare alsoalgorithms.

Id.at3.Intheeconomicsociologyliterature,MarkGranovetteremphasizes “theroleofconcretepersonalrelationsandstructures(or‘networks’)ofsuch relations ingeneratingtrustanddiscouragingmalfeasance.”Mark Granovetter,EconomicActionandSocialStructure:TheProblemof Embeddedness,91AM.J.SOC.481,490(1985).Hearguesfurtherthat“the anonymousmarketofneoclassicalmodelsisvirtuallynonexistentineconomic lifeandthattransactionsofallkindsarerifewiththesocialconnections described.”Id.at495;seealsoWayneE.Baker,TheSocialStructureofa NationalSecuritiesMarket,89AM.J.SOC.775,807(1984)(explainingthat computerizedtradingmay“affectthesocialstructureofthemarket,butit wouldneverescapethefactthatmarketsaresociallystructured”).Whilethis Articlearguesthatthecontemporary,computerizedequitymarketsrepresent anextremetypeofmarketthatinsomewaysapproachestheanonymousand abstractmarketofneoclassicaltheory,thefactthatHFTiscontroversialatall demonstratesthatwestillbringourmoralbeliefstoquestionssurrounding thefunctioningofthealgorithmsandcomputerserversthatconstitutethis market.

366.Whetheratransactionwasconductedface-to-faceoronananonymous exchangewasoftendeterminativeundertheearlytwentiethcenturystatelaw governinginsidertrading.SeeA.C.Pritchard,UnitedStatesv.O’Hagan: AgencyLawandJusticePowell’sLegacyfortheLawofInsiderTrading,78 B.U.L.REV.13,22–26(1998);StephenM.Bainbridge,TheLawand EconomicsofInsiderTrading:AComprehensivePrimer7–8(Feb.2001) manuscript), (unpublished http://papers.ssrn.com/sol3/papers.cfm?abstract_id=261277; seealsoElinor Ostrom,CollectiveActionandtheEvolutionofSocialNorms,14J.ECON. PERSPS.137,140(2000) face-to-face communicationinpublicgoodsgames).

(highlightingtheimportanceof

tradingtakethisaspecttoitsextreme,where“traders”inthe conventionalsensereallyjustrepresentalgorithms.367Yetas theoutcryoverHFTshows,eveninthishighlyattenuated feelmoralprinciplesdo,or contextmanyindividualsstill should,apply.

Becauseofthisunusualsocialcontext,anyethicsofthe financialmarketswillinevitablybespare.Weunderstandthat ourpotentialtradingpartnersonanexchangearewholly motivatedbyself-interestintransactingwithus,andtherewill benoongoingrelationshipbeyondasinglesale.368Unlikeother businessrelationships,thereisnorelianceonanotherparty concerningfuturebehavior.Nevertheless,thereisasetof trust, minimalexpectations,andthusalimitedformof surroundingaone-timetransactionintheanonymous, mediatedtradingenvironmentofthemodernstockexchange.

TheresultsoftheUltimatumGameaspresentedby behavioraleconomistRichardThalerareanappropriate startingpoint.369IntheUltimatumGameanindividual,the Allocator,isgivenasumofmoneyandtoldtodivideitbetween himselfandtheRecipient.370TheRecipientthendecides whethertoacceptthisallocation.371Ifhedoes,bothparties receivetheamountsasdeterminedbytheAllocator;ifthe Recipient theproposedallocation,bothreceive nothing.372Accordingtotheviewofrationalityespousedby conventionaleconomictheory,Recipientsshouldacceptany offersgreaterthanzero,andtheoffersproposedbyAllocators shouldapproachzero.373

rejects

Theresultsofexperimentaltestingdivergeconsiderably fromthisview.BothAllocatorsandRecipientsdemonstratea preferencetowardsdivisionsapproachinga50/50split,which Thalerinterpretsasdemonstratingthat“notionsoffairness canplayasignificantroleindeterminingtheoutcomesof negotiations.”374InanUltimatumGameconductedbyGerman

367.MacKenzie,supranote6,at58. 368.SeegenerallyJoelSobel,InterdependentPreferencesandReciprocity,

43J.ECON.LITERATURE392,411–12(2005)(discussing“RepeatedGames”).

369.RichardH.Thaler,Anomalies:TheUltimatumGame,2J.ECON.

PERSPS.195(1988).

370. 371. 372. 373.

Id.at195. Id.at195–96. Id. Id.at196(citingArielRubinstein,PerfectEquilibriumina

BargainingModel,50ECONOMETRICA97(1982 .

374.

Id.at205.

economistsWernerGüth,RolfSchmittberger,andBernd Schwarze,21pairsdividedasum“c,”between4and10 Deutschmarks,withameanofferof.37c.375In7outof21cases theofferwasa50/50split.376Inafurtherexperiment conductedbyKahneman,Knetsch,andThaler,Allocatorswere presentedwiththechoiceofdividing$20withRecipientsin eitheran18/2ora10/10split;76%ofthesubjectsdividedthe $20evenly.377

TheUltimatumGameindicatesthathumansgaugewhat isappropriateinatransactiononsomebasisotherthana narrowlydefinedrationalityalone.Ifrationalityisdefinedas “themaximizationofutilityunderabudgetconstraint,”378a rationalRecipientwouldacceptanyoffer,nomatterhowsmall, madebyanAllocator.Likewise,Allocatorswouldofferamounts assmallaspossibleinordertoretainthemaximumamountfor themselves.Yetsuchextremeself-interestdoesnotcomport withtheresultsoftheUltimatumGame.Humanbehaviorlies somewherebetweentheextremesofcompletegenerosityand completegreed,asThalerobserves.379Mostimportantly,our inherentgreedinessistemperedbythenotionoffairnessboth asAllocatorsandRecipients.380

375.

Id.at196–97(citingWernerGuth,RolfSchmittberger&Bern Schwarze,AnExperimentalAnalysisofUltimatumBargaining,3J.ECON. BEHAV.&ORG.367(1982 . Id.at197. Id.at198.

376. 377. 378.SeeKolb,supranote299,at24(citingKennethJ.Arrow,Rationality ofSelfandOthersinanEconomicSystem,59J.BUS.S385,S388(1986 (“Key AssumptionsinFinance:RationalityandSelf-Interest”);seealsoDobson, supranote299,at47–49(“TheAxiomsofRationality”);Nien-hêHsieh, EfficiencyandRationality,inFINANCEETHICS,supranote299,at63.

379.SeeThaler,supranote369,at205(“[I]tissafetosaythatmostpeople arenotwelldescribedbyeitherextremeview.Rather,mostpeopleprefer moremoneytoless,liketobetreatedfairly,andliketotreatothersfairly.”).

380.Evolutionarypsychologistspositthatthetendencytoincorporatea conceptoffairnessintostrategicnegotiationbehavior,whichbothdeparts fromstrictrationalityandisgroundedinreciprocity,promotedsurvivalinthe environmentofevolutionaryadaptation.SeeJohnTooby,LedaCosmides& MichaelE.Price,CognitiveAdaptations forn-PersonExchange:The EvolutionaryRootsofOrganizationalBehavior,27MANAGERIAL&DECISION ECON.103,118(2006).Reciprocitywouldthereforebeasuperiorsurvival strategytogreedinenvironmentscharacterizedbysignificantuncertainty, wherebondsoftrustbetweenmembersofagroupcouldmeanthedifference betweensurvivalandextinction.Forthepresentationoftheseideasinthe languageofeconomics,seeSobel,supranote368,at397–412.Cf.ArmenA. Alchian,Uncertainty,Evolution,andEconomicTheory,58J.POL.ECON.211, 211(1950)(discussingtheinsufficiencyofthetypicalrationaleconomictheory

beliefs

principle

ourmoral

TheUltimatumGamepointstowardsreciprocityasa grounding concerning exchange.381Thedeviationfromrationalityobservedinthe UltimatumGameresultsinAllocatorsgivingmorethanwould seemnecessary,orrational,andRecipientsdemandingmore thanaminimalamount.WhilenotallAllocatorssplittheir sumsevenly,asurprisingportionofthemdo.Themovement towardsanequaldivisionimpliesthatweconceivefairnesson thebasisofputtingourselvesinacounterparty’sshoes—we ask,whatwouldwehopeforifwewereintheotherperson’s position?Aneven50/50splitwouldbetheidealreflectionofthe other’sinterestsinthiscase.

Theprincipleofreciprocityisreflectedinmanyofthe world’smajorethicalandreligioustraditions.Aristotle’s treatmentofexchangeasproportionateequalitycaptures this,382asdoestheGoldenRuleoftheNewTestament,“Doto toyou.”383The othersasyouwouldhave intersubjectivityimplicit inKant’sCategoricalImperative capturesthenotionofreciprocityaswell.Incommandingthat oneoughtneveractexceptinsuchawaythatI“willalsothat [my]maximshouldbecomeauniversallaw,”theCategorical Imperativeplacestheselfonthesameplaneasitsfellow humans.384Indeed,Kant’snotionoftheKingdomofEndstakes

themdo

of“predictable,individualbehavior,”andtheneedtoincorporateideasof uncertaintyand“environmentaladoption”ineconomicmodels).

381.SeegenerallyJohnson,supranote299,at43.WhileJohnsongrounds hisclaimsinwhatheterms“pragmaticphilosophy,”whichisinturnbasedon Aristotle’sethics,whatfollowsinPartIIIisbasedonKantiandeontology.Id. WhileJohnsondiscussesandrejectsthephilosophyofRawlsasapossible basisforhistheorythatreciprocityisthefoundationoffinancialeconomics,on thegroundsthatRawlsbelievedthathisconceptionofjusticewas“political notmetaphysical,”hedoesnotexploreKantiandeontologyasabasisfor reciprocity.Id.at45.

382.SeeARISTOTLE,NICHOMACHEANETHICSbk.V,at88–91(AlbertKeith Whitakered.,JoeSachstrans.,FocusPublishing2002)(c.384B.C.E.). JohnsonbaseshisdiscussionofreciprocityonAristotle’sconceptionof“justice inassociationsforexchange.”SeeJohnson,supranote299,at48.Aristotle believedthatafairexchangewascharacterizedbyproportionateequality,so thattherewasnogivinginexchange.SeeLindsayJudson,AristotleonFair in15OXFORDSTUDIESINANCIENTPHILOSOPHY147,147–51 Exchange, (C.C.W.Taylored.,1997).

383.LUKE6:31(NewInternationalVersion);seealsoMATTHEW7:12(New InternationalVersion)(“Soineverything,dotootherswhatyouwouldhave themdotoyou,forthissumsuptheLawandtheProphets.”).

384.SeeIMMANUELKANT,GROUNDWORKFORTHEMETAPHYSICSOF MORALS19(AllenW.Wooded.&trans.,YaleUniv.Press2002)(1785).The firstpresentationoftheCategoricalImperativeinGroundworkiscommonly

thistoitsconclusion,aswearetotreatallpersonsasendsin themselves,notmeans.385WhileKant the CategoricalImperativeisnotsynonymouswiththeGolden Rule,386bothaskustoplaceourselvesinthepositionofothers whenexercisingmoraljudgment.

isclearthat

Reciprocityasafundamentalprincipleofethicshas importantimplicationsforfinance.387First,itmeansthatwe shouldnotactinawaythatwewouldwishothersnotact towardsus.388Whilethismightappearatfirstglancetorule outtheoverwhelmingmajorityofmarkettransactions,which areofcoursebasedonself-interest,thisisnotreallythe case.389AsAristotle’sdiscussionofproportionateequalityin exchangemakesclear,afairexchangeoccurswhenaparty exchangessomethingofequalvalueforsomethingelse.390This allowsforpartiestotradewhentheyhavedifferingneeds.For example,ifIsellaloafofbreadfortwodollars,Imaydoso becauseIhavethreeloavesalreadyandcannoteatthemall.It alsocruciallyallowsforindividualstotradeonthebasisof differingbeliefsaboutfutureevents.391IfIbelievethecountry

referredtoasthe“FormulaoftheUniversalLawofNature”byphilosophers. See,e.g.,RobertJohnson,Kant’sMoralPhilosophy,STAN.ENCYCLOPEDIA PHIL.,http://plato.stanford.edu/entries/kant-moral/ (lastmodifiedApr.6, 2008).

385.KANT,supranote384,at45–46. 386.

Id.at48.Kantherestatesthattheprinciplequodtibinonvisfieri[do

notdotootherswhatyoudonotwantdonetoyourself]

isonlyderivedfromthatprinciple...itcannotbeauniversallaw, foritdoesnotcontainthegroundofdutiestowardoneself,northat ofthedutiesof lovetowardothers(formanywouldgladly acquiescethatothersshouldnotbebene!centtohim,ifonlyhe mightberelievedfromshowingbene!cencetothem),or!nallyof oweddutiestooneanother,forthecriminalwouldargueonthis groundagainstthejudgewhopunisheshim,etc.

Id.ForadetailedexpositionoftherelationbetweenKant’sCategorical ImperativeandtheGoldenRule,seeDanielBerthold,TheGoldenRulein KantandUtilitarianism,inTHEGOLDENRULE:ANALYTICALPERSPECTIVES 83(JacobNeusner&BruceChiltoneds.,2009).

387.See,e.g.,Johnson,supranote299,at43. 388.Seeid.at55. 389.SeeBOATRIGHT,supranote299,at1(“[A]moment’sreflectionreveals

thatfinancewouldbeimpossiblewithoutethics.”). 390.SeeARISTOTLE,supranote382,at88–91. 391.SeegenerallyARISTOTLE,POLITICS,bk.I,ch.11(TrevorJ.Saunders trans.,OxfordUniv.Press1995)(c.350B.C.E.).Afterhewasreproachedfor hispoverty,ThalesputadepositdownonalltheolivepressesontheGreek islandsofChiosandMiletus,knowingthatthecomingyearwouldbringa largeolivecrop.Id.at1258.Whenthecropcamein,hewasabletorentout theolivepressesatahighprice,earningasubstantialamount.Id.at1259.

willenteraperiodofdrought,Imaybewillingtopayten dollarsforaportionofgrain,andifmycounterpartybelieves rainfallwillbegoodthisyearhemaywanttoselltomeatthis price.Reciprocitydoesnotmeanthatonewon’ttradeonthe basisofcomparativeadvantageordifferingbeliefsaboutthe future,orevendifferinglevelsofinformation,butmerelythat onecanenvisionmakingthetradeonthegiventermsifone wereontheoppositesideofthetransaction.

Reciprocitydoes,however,conflictwithtradingpractices generallyconsideredabusive.Forexample, insidertrading violatestheprincipleofreciprocity,asweinstinctivelyfeelwe wouldnotwanttotradewithsomeonewhohadinformation aboutasecuritythatwas,inprinciple,onlyavailabletoa few.392Likewisewefeelthatfront-runningonthepartofour brokerisunfair,asitnegativelyaffectsthepricewereceivefor astock,ifonlybyasmallamount.393Bothoftheseactivities strikemostofusasfundamentallyunfair.Still,itmustbe rememberedthatreciprocityisamoralideal:wehopeitwillbe respectedwhenweactinthemarketplace,eventhoughwe knowitsometimesisnot,andwesuccumbonoccasiontothe temptationtotreatothersaswewouldnotwanttobetreated.

Reciprocityfundamentallyrepresentsanequalitybetween thetwosidesofanexchange.Oneinterestingviolationof the“DutchBook.”394This iswhenthe reciprocityis probabilitiesinabettinggameadduptomoreorlessthan one.395Iftheprobabilitiesarelessthanone,agamblercan makeacertainprofitagainstthehouse;ifmorethanone,the

illustratinghowa WhileAristotlerelatesthisstoryforthepurposeof monopolycanbeprofitable,thereisnohintthatwhatThalesdidwasimmoral orshameful,eventhoughAristotlebelievesthatthepursuitofwealthis inferiortophilosophy.Id.

392.SeeStevenMcNamara,InsiderTradingandEvolutionaryPsychology: StrongReciprocity,CheaterDetection,andtheExpandingBoundariesofthe Law,22VA.J.SOC.POL’Y&L.241,244(2015)(notingthat“weintuitivelyfeel that[insidertrading]isunfair”andexaminingtherelationshipbetweenthis socialunderstandingandthelaw).

393.Seeid.(“[W]estillbringourmoralintuitionstothecontemporary

environmentofcomplextechnologicalandfinancialsystems.”).

394.SeeJohnson,supranote299,at55.The“DutchBook”argumentwas presentedbythemathematicianBrunodeFinettiin1931.SeeKlausHeilig, CarnapanddeFinettionBetsandtheProbabilityofSingularEvents:The DutchBookArgumentReconsidered,29BRIT.J.PHIL.SCI.325,326–27(1978); seealsoWilliamM.Briggs,DutchBook:Or,HowtoGamble,W.M.BRIGGS BLOG(Feb.21,2012),wmbriggs.com/?s=dutch+book(providingexamplesof theDutchBook).

395.SeeBriggs,supranote394.

housewillalwayscomeoutahead.396Intheconventional situationwhereacasino,lotteryagency,pari-mutuelbetting operation,orMafianumbersracketmakesaDutchBook,the houseineffectpaysitselfthroughthelopsidedodds.397Suchan amountisknownasthe“skim,”“juice,”“vigorish,”or“vig.”398It maybethatthisamountofguaranteedprofittothehouseis justthecostofgambling,andthatgamblersconsenttopaying itbecauseitisunderstoodthatinthelongrunthehouse alwayswins.399

Query,however,whethersuchanunderstandingobtains fortransactinginthestockmarket.Dowebelievethatinthe longrun,ourinvestmentswillalwaysloseor,moreprecisely, thattheywillalwaysbecompromisedbysomesmallamount? Equallyimportant,dowefeelthatitisaviolationofthe principleofreciprocityiftherearesuchsmallamountsby whichweinvariablylose,asasortoftaxorrentonour transactions,andwedonotknowbeforehandofeithertheir existenceorextent?OneofthefundamentalcriticismsofHFT isthatitexactsasmallandhiddencostonallothertradersin themarkets.400Ifso,HFTresemblesaDutchBook.The analogymaybeimperfect,butacasinoandasuccessfulHFT firmbothgenerate(relatively)risklessprofits.401

396.

theNon-event).Agambler

Id.ConsiderthefollowingexamplegivenbyWilliamBriggs:Abookie givesevenodds,or1:1,fortheEventoccurring,andoddsof3:1fortheNoneventoccurring.Heretheprobabilitiessumto0.75(0.50fortheEventplus canmakeacertainprofitby 0.25for simultaneouslywagering$20ontheEventand$10ontheNon-event.Should theEventcometopass,thegamblerwillreceiveapayoutof$40,comprisedof theoriginal$20wageredplusanother$20.Inthealternative,theNon-Event comestopassandthegamblerreceivesapayoutof$40,comprisedofthe original$10plusanother$30.Thegamblerwagersatotalof$30andineither casereceives$40becausetheprobabilitiessumtolessthan1.Wherethe probabilitiessumtogreaterthan1,thehousealwayswins.Seeid.

397.See,e.g.,id.(explaininghowAmericanRouletteisaDutchBookfor

thecasino).

398.Seeid. 399.Seeid.(statingthattheamountskimmedmayalsobereferredtoasa

“transactioncost”).

400.See,e.g.,ARNUK&SALUZZI,supranote131,at47(“Explicitlyyour transactioncostsmayhavecomedown....Implicitly,youpaymoreforthe stocksyoubuyoryoureceivelessfromthoseyousell.”).

401.SeeBriggs,supranote394(describingtherisklessprofitsgenerated byaDutchBook);seealsosourcescitedsupranote8(describinghowsome HFTsexperiencevirtuallynolosses).

C.FREEDOMFROMMISREPRESENTATION

Theprincipleofreciprocityservesasafoundationforthree moreparticularethicalprinciplesthatwillapplyinthespecific environmentofthefinancialmarkets.Firstamongtheseis freedomfrommisrepresentation.402Wedonotwanttobeliedto abouttheparticularsofaspecificsecurityoraboutaparticular stateofaffairsinthemarkets.

Freedomfrommisrepresentationisacornerstoneof marketregulationandformsthebasisofAmericansecurities law.403Atbase,marketparticipantswantsomeassurancethat theycanrelyontheinformationcommunicatedtothem. Examplesofthisincludethedisclosurerequiredbysecurities regulation,aswellasmoreshort-livedaspectsrelatingto prices.404Canweassumethataquotepostedisactually availabletobeactedupon?SinceHFThastransformedthe basicunitofmarketinformationfromthetradetotheorder,405 quotesthemselvesarenowtheprimaryitemsofinformationin thehighfrequencymarkets,andthereissubstantialconcern abouttheirreliability.406Anestimated96%–98%ofquotesare nowcancelledbeforetheycanbefilled.407Theenvironmentof flickering,constantlyrevisedquotationscallsintoquestion someofthebasicassumptionsofsecuritiesmarkets,butitdoes

402.Freedomfrommisrepresentationisafocusofthebusinessethics literaturedealingwiththefinancialmarkets.See,e.g.,BOATRIGHT,supranote 299,at175–77(“Fraudandmanipulation”);Shefrin&Statman,supranote 299,at25.

403.SeeJamesM.Landis,TheLegislativeHistoryoftheSecuritiesActof 1933,28GEO.WASH.L.REV.29,34–35(1959);ElaineA.Welle,Freedomof ContractandtheSecuritiesLaws:OptingOutofSecuritiesRegulationby PrivateAgreement,56WASH.&LEEL.REV.519,534–35(1999).

404.Shefrin&Statman,supranote299,at24. 405.O’Hara,supranote148,at263(“Algorithmictradingmeansthat tradesarenotthebasicunitofmarketinformation–theunderlyingorders are.”).

406.TheinabilityofHaimBodektoexecutetradesatthepricesoffered formedtheimpetusforhisquesttounderstandthenewtradingenvironment. SeePATTERSON,supranote11,at54–57.MaureenO’Harareportsthat,by someestimates,upwardsof98%ofquotesarecancelled,“creatinguncertainty astotheactuallevelofcurrentprices.”O’Hara,supranote148,at259,267; seealsoJoelHasbrouck,HighFrequencyQuoting:Short-TermVolatilityin BidsandOffers4(Jan.20,2015)(unpublishedmanuscript),http://ssrn.com /abstract=2237499(examiningtherelationshipbetweenshort-termvolatility inbidsandhighfrequencyquoting).

407.SeeO’Hara,supranote148,at259(“[I]tisnowcommonforupwardof 98%ofallorderstobecanceledinsteadofbeingexecutedastrades.”);Gaiet al.,supranote254,at6(“[A]nincreaseintradingspeedsleadstoadramatic increaseinthecancellation/executionratiofrom26:1to32:1....”).

notnecessarilyindicateaviolationoftheprincipleoffreedom frommisrepresentation.Thereseemstobenothingwrongin principlewithcancellinganorder,evenifitleadstoavast majorityofordersbeingcancelled.Aconsequenceofthe markedincreaseintheorder-to-cancellationratio,however,is thattraderscannolongerrelyontheassumptionthataquote representssharesactuallyavailabletobeboughtorsold.408

Ontheotherhand,therearenumerousabusesinthehigh frequencyarenathatdoamounttoaviolationoftheprinciple of the freedomfrommisrepresentation.Asseenabove, momentumignitionstrategiesofspoofingandlayeringarein essencecontemporaryformsofstockmarketmanipulation.409 thecatandmousegameofdueling Thesearepartof algorithms,buttheyalsoinvolvethemisuseofaquotethat amountstoanimplicitmisrepresentation.Ifwepostaquote thatindicatesthatwearewillingtobuysomethingwiththe intentofchangingtheprice,andthenturnaroundandsellit, theoriginalquotemisrepresentsourintent.410Ofcourse,we couldsaythesameaboutpostingaquoteandthencancelling it,butmomentumignitionstrategiestakethisonestepfurther. Inthecontextofspoofingandlayering,aquoteispostedwith theintentofcausingittochangethepricesprevailinginthe markets,andthenanotherquoteisactedontothedetrimentof othersrespondingtothechange.Notsurprisingly, these activitieshavebeentargetedwiththelawsproscribingmarket manipulation.411

Thebeliefinpublicprotectionfrommisrepresentationisa cornerstoneofsecuritieslaw,andbusinesslawingeneral,yet itsapplicationpresentsseriousdifficultiesintheHFTworld.412 Thisisbecausewewanttocontinuetorelyonsignalsthatwe

408.SeeO’Hara,supranote148,at267;Gaietal.,supranote254,at5–6. supranotes249–59andaccompanyingtext 409.See

(discussing

momentumignitionandspoofing).

410.SeePasquale,

supranote12,at2106–12(“Responding

supranote

411.SeeAktas,

MisrepresentationofMarketConditions”). (discussingmarket 252, manipulation);Scopino,supranote12,at660–63(“SpoofingasViolating CommodityExchangeActSections4c(a)(2)(B)and9(a)(2)”);see,e.g.,Kirilenko &Lo,supranote125,at66(SECissuedacease-and-desistorderagainstan investmentfirmusing“spoofing”and“layering”techniquestomanipulate prices);Korsmo,supranote12,at554–55(TrilliumBrokerageServs.,LLC wasfined$2millionforutilizingamomentumignitionstrategy).

93–94

at

412.SeegenerallyTaraE.Levens,Note,TooFast,TooFrequent?HighFrequencyTradingandSecuritiesClassActions,82U.CHI.L.REV.1511 (2015).

havealwaysassumedwerereliable—forexample,thatquotes postedareinfactopentobefilled—butforreasonsof technologicaladvancemaynolongerbeso.413Partofthe consternationtraderssuchasHaimBodekhavefeltstemsfrom animplicitfeelingofdeceptioninthemarkets,thatquotes postedarenotinfactavailabletobefilledbutaremerelythere toinduceotherbehaviorfromothertraders.414Freedomfrom fraudandmisrepresentationarejustasmuchpreconditionsof computerizedmarketsastheywereofearlierones.Increased enforcement lawconcerningmarket manipulationagainstdeceptiveuseofquotationsisrequiredto policesuchbehavior.

existing

D.ALEVELPLAYINGFIELD

OneofthemostfrequentcomplaintsofHFTisthatit createsatwo-tieredsysteminwhichasmallamountofplayers havepreferentialtreatmentorareotherwiseadvantaged.415 Thiscriticismassumesthatsomehowthesecuritiesmarkets are,orshouldbe,alevelplayingfield.

Thenotionofalevelplayingfieldinfinanciallawmeans thatlikemarketparticipantswillbetreatedalike.416In securitiesmarketsthismeansthatwiththenecessary prerequisites—therequiredinvestmentsoftimeandmaterial resources—anyoneshouldbeabletocompetesuccessfullyin

413.See,e.g.,id.at1511–12. 414.Bodek,HFTCheckmate,supranote265. 415.ConceptReleaseonEquityMarketStructure,ExchangeActRelease No.34-61358,75Fed.Reg.3594,3605(Jan.14,2010)(codifiedat17C.F.R.pt. 242).

416.SeeJesseM.Fried,InsiderAbstention,113YALEL.J.455,460–62 (2003)(characterizingthe“discloseorabstain”ruleofinsidertradinglawas anattemptto“leveltheplayingfield”);ZacharyJ.Gubler,Reconsideringthe InstitutionalDesignofFederalSecuritiesRegulation,56WM.&MARYL.REV. 409,424–26(2014)(characterizingthedisclosureregimeoftheExchangeAct asanattemptto“leveltheplayingfield”forexchange-listedfirms);seealso GuhanSubramanian,TheDrivers inRevlon Transactions,28J.CORP.L.691,699–700(2003)(notingthecreationofa “levelplayingfield”underDelawarecorporatelawforbiddersincorporate auctions).Fordiscussionsofalevelplayingfieldinthelawitself,seeSec. Exch.Comm’nv.TexasGulfSulphurCo.,401F.2d833,852(2dCir.1968) (reiteratingcongressionalintentthat“allmembersoftheinvestingpublic shouldbesubjecttoidenticalmarketrisks”);SelectiveDisclosureandInsider Trading,65Fed.Reg.51,716,51,716(Aug.24,2000)(codifiedat17C.F.R.pts. 240)(“[I]nvestorsloseconfidenceinthefairnessofthemarketswhenthey knowthat informational advantages’....”).

otherparticipantsmayexploit

ofMarketEfficiency

‘unerodable

thesecuritiesmarkets,andthattheseprerequisitesshouldbe availabletoall.417Ofcourse,inrealitynotallmarketplayers willmaketheseinvestments,andmorefundamentally,these prerequisitesmaynotbeavailabletoall.418Alevelplayingfield existsmoreasagoaloridealthanasaconcretereality.419An exampleofthiscomesfromthelawofinsidertrading.Oneof themotivationsofinsidertradinglawistopreventindividuals fromtradingonthebasisofinformationthatisinprinciplenot availabletoall,becausethisisthoughttobeunfair.420Noone presumesthattradersinfacttradeonequalinformation,but thelawdoes—tosomedegreeatleast—aimtorestricttrading oninformationthatisnotaccessibletoall.421Equalaccessto informationiskeytowhatalevelplayingfieldmeansinU.S. securitieslawgenerally.422IntheHFTarenaalevelplaying fieldalsoimplicatesissuesofspeedandprocessingpower,and equaltreatmentbythelawandkeymarketinstitutions.423

Theraceforspeedandco-locationcallintoquestionthe notionofalevelplayingfieldbecausethesmalladvantages thataccruetoHFTcanneverbemadeupbyothertraders.In thezero-sumgameoftrading,traderswithspeedadvantages willalwayswinthecompetitiontocomeinfirst,asotherscan nevermakeupforanyinitialdelay.424Sincetradinginthe currentmarketsoftenreducestoamerecompetitionforspeed,

417.SeeShefrin&Statman,supranote299,at28(“Theideaofthelevel playingfieldisassociatedwiththeviewthatinvestmentactivityisagameof skill.”).

418.Seesupranotes188–90(describingthelargeinvestmentsnecessaryto

competeinHFT).

419.SeeUnitedStatesv.O’Hagan,521U.S.642,658(1997)(“Although informationaldisparityisinevitableinthesecuritiesmarkets,investorslikely wouldhesitatetoventuretheircapitalinamarketwheretradingbasedon misappropriatednonpublicinformationisuncheckedbylaw.”).

420.Of

courseinAmericansecurities

tradingonnonpublic law, informationisnot, inprinciple,enoughtoresultinaninsidertrading violation;accesstosuchinformationmustalsoinvolvetheviolationofa fiduciaryduty.O’Hagan,521U.S.at647.Forperceptionsofunfairnesson tradinginawidevarietyofscenarios,seeStuartP.Green&MatthewB. Kugler,WhenIsItWrongtoTradeStocksontheBasisofNon-Public Information?PublicViewsoftheMoralityofInsiderTrading,39FORDHAM URB.L.J.445(2011).

421.SeeO’Hagan,521U.S.at658–67(detailingthepurposeandprevious

courtinterpretationsofthemisappropriationtheory).

422.SeeGubler,supranote416,at424–26(arguingthatoneofthe principlesthatmotivatedCongresstoenactsecuritieslawsisaconcernfor fairnessandalevelplayingfield).

423.Levens,supranote412,at1512. 424.Seesupranote182andaccompanyingtext.

slowertradersfeelthattheycannothopetocompetewithcolocatedtradersforthebesttrades,leadingtothebeliefthat thereisatwo-tieredmarket.425

Alevelplayingfieldisaconceptthatrepresentsanideal, notareal,condition.426Evenwherethelawisrelatively successfulincreatingalevelplayingfield,inrealitythereis informationinmarkets.427Much, significantasymmetric thoughnotall,ofthisistheresultofvaryinginvestmentsin humancapital,technology,research,andanalysis.428And behindthesedifferencessitourdifferenthumanendowments, whichresultinvaryinglevelsofambition,intelligence,and intereststhatinturncontributetovaryinglevelsofmaterial success.Thatsaid,theconceptoffairnessinthefinancial marketsrequiresthatregulatorsstrivetokeepaccessto informationandtheinherentabilitytoprofitfromitopento all.429Whileintheoryallmayhavetheopportunitytousethis information,inpracticeHFTcallsintoquestionthenotionofa levelplayingfieldbyrestrictingtheabilitytoprofitfromitto onlytheverymostsophisticatedactors.430

E.INSTITUTIONALINTEGRITY

Thefinalbasicprincipleofafinancialmarketsethicsfor HFTisthepresumptionofinstitutionalintegrity.Anumberof thepracticesoutlinedaboveimplicatequestionsoftheintegrity oftheinstitutionscentraltothefinancialmarkets.Primary amongthesearethestockexchanges,butthepracticesofdark poolsandbroker-dealershavealsocomeunderfireasunfairto certainoftheircustomers.Beforeweexaminethesecharges, however,itisusefultodefineinstitutionalintegrity.431

425.Alevelplayingfieldalsoimplicatesthe(un)equaltreatmentof differentclassesof institutions,primarilythe exchangesandothertradingvenues.Sincethisrelatestobusinessdecisions madebytheseinstitutions,theyaretreatedininfraSectionIII.E.

tradersbykeymarket

426.SeeO’Hagan,521U.S.at658. 427.SeeSubramanian,supranote416,at694(notingthat“information

willnotbewidelydistributed”because“informationcostsarehigh”).

428.Seeid.;Levens,supranote412,at1511. 429.SeeCoffee,supranote86(“Tothispoint,wehavebeenduckingthe largerissueandfocusingonlyonwhetherRegulationNMSwasviolated.But shouldco-locationandprivatedatafeedsfromexchangestopreferred shareholdersbepermissible?”).

430.Bodek,HFTCheckmate,supranote265. 431.

Institutional

integrityhasreceivedsubstantialdiscussioninthe businessethicsliterature.See,e.g.,ThomasMaak,UndividedCorporate Responsibility:TowardsaTheoryofCorporateIntegrity,82J.BUS.ETHICS353

institution’s

Atthecoreofinstitutionalintegrityistheideathatan institutiondisplaysintegritywhenitactsinthefurtheranceof itspriorcommitmentstostakeholders.432Whilehonestyis importanttointegrity,thetwoconceptsarenotsynonymous.433 Integritydenotesthewilltohonorobligationsinawaythat demonstratesacoherenceoftheactionsandcommitmentsof theinstitutioninquestionboththroughtimeandacrossthe various range stakeholders.434Atbaseinstitutionalintegritymeanslivingup toaninstitution’scommitments.435Animportantsub-question askshowweknowaninstitutionhasmadeacommitmenttoa particular thesecommitmentsor obligationsarisenaturallyinthecourseofbusinessandare implicitratherthanexplicit.436Whiletheremaybesubstantial roomfor particular as commitments,itisimportanttorecognizethattheycanarise implicitlyinthecourseofbusiness.Suchimplicitcommitments formalegitimatebasisformanyofthecomplaintsfoundinthe HFTdebateaswellasinwiderquestionsofbusinessethics.

stakeholder.Manyof

disagreement

obligations

identifying

its

Problemsatthestockexchangesarenotnew,butthe conjunctionofRegulationNMSandrapidtechnologicalchange transformedtheexchanges.Theywentfrombeingrelatively transparententitiesthatprivilegedmembersattheexpenseof retailclients,toopaqueonesprivilegingonesetofclientsatthe expenseofothers.437ThetemptationofcateringtoHFTinthe

(2008);LynneMcFall,Integrity,98ETHICS5(1987);LynnSharpPaine, ManagingforOrganizationalIntegrity,HARV.BUS.REV.,Mar.–Apr.1994,at 106.

432.SeeMaak,supranote431,at358. 433.SeeMcFall,supranote431,at7–8. 434.SeeMaak,supranote431,at364;McFall,supranote431,at7–9. 435.SeeMaak,supranote431,at362. 436.SeeRonaldJeurissen,TheSocialFunctionofBusinessEthics,10BUS. ETHICSQ.821,823(2000)(characterizingthe“socialcontractmodel”of businessethicsasan“implicit‘socialcontract’”);cf.AnitaJose&MaryS. Thibodeaux,TheInstitutionalizationofEthics:ThePerspectivesofManagers, 22J.BUS.ETHICS133,135(1999)(discussingimplicitformsofethicsin organizations).Forananalogousargumentinthelegalliterature,seeLon.L. Fuller,HumanInteractionandtheLaw,14AM.J.JURIS.1(1969)(arguing customarylaworiginatesinreciprocalexpectations);GeraldPostema,Implicit Law,13L.&PHIL.361,362(1994)(“ImplicitRulesandSocialInteraction”).

437.ThiscontrastisthefulcrumpointofCraigPirrong’sPickYourPoison, whichemphasizestheabilityoftheexchangestoprofitfromnaturaltrading monopoliespriortotheintroductionofReg.NMS’sOrderHandlingRule, which“effectivelysocializedorderflow.”Pirrong,PickYourPoison,supranote 70,at11.

newandhighlycompetitiveenvironment—whererevenue comesfromthespreadbetweenmakerandtakerfees— compromisedtheinstitutionalintegrityoftheexchanges.438In theracetocaptureaviableflowoforders,exchangeswere incentivizedtocatertotheirbestcustomers,theHFTshops.439 Whilethemaker/takerpricingsystemisalargefactorinthe competitionfororderflow,perhapsequallyimportanttoHFT clientsisthenatureoftheorderflowonanexchangewith manyinstitutionalandretailclients:itisgenerallynotwell informed,i.e.,not“toxic.”Thisuninformedorderflowholds greatappealforHFT,andtheexchangeshaveofferedHFT clientsanumberoftoolstoprofitfromtradingwithit.

“cut”

ThevarioustechniquesreviewedinPartIIcanallbeseen asmechanismsthatfacilitateHFT’smissiontoprofitfrom slowerandlessinformedtraders.440Oneofthemosttrenchant criticismsofHFTisthattheexchangesengineeredthese techniquesspecificallytoallowHFTfirmstoprofitatthe expenseoftheirothercustomers.441Thefeesgarneredfor facilitatingHFTactivitycanthereforebeseenasthe compromised exchanges’ institutionalintegrityoftheexchangesconsistsofofferingup onegroupofcustomersaspreytoanother,morelucrative group.Acommitmenttoinstitutionalintegritywouldentail resistingthetemptationtoprofitattheirslowercustomers’ expensebyrefusingtofacilitatetheabusiveformsofHFT. WhiledefendersofHFTandtheexchangesmayarguewhether suchastrategywouldevenbefeasibleinanenvironment wherethebulkoftradingisaccomplishedalgorithmicallyand themaker/takersystempredominates,atleastonetrading venueiscurrentlytestingthispropositioninthemarket.443

profits.442The

438.Seesupranotes148–50andaccompanyingtext(describingthe

maker/takerpaymentsystem).

439.Cf.supranotes189–90andaccompanyingtext(exchangesasderiving

substantialrevenuefromthesaleofdatatoHTFs).

440.Afewofthetechniquesdiscussedaboveareco-location,flashorders,

spoofing,layering,quote-stuffing,andISOs.SeesupraPartII.

441.See,e.g.,Bodek,HFTCheckmate,supranote265;Hope,supranote 282(notingtheBATSexchangewasfined$14milliontosettleclaimsthatit, amongotherthings,“gavesomehigh-speedtradersanadvantageoverothers bynotprovidingdetailsaboutcertainordertypes”).

442.Cf.Foxetal.,supranote12,at233(“Ifelectronicfrontrunningwere eliminatedtomorrow,HFTco-locationfacilitieswouldbeworthlesstothe HFTs....Thismightreducetherentscollectedbytheexchanges.”).

443.SeeKarlAntle,IEX&theRe-MutualizationofMarketInfrastructure, (May28,2014),http://www.wallstreetandtech.com

IEXstandsasanimportanttestofwhetherthetrading environmentwillsupportatradingvenuethatmarketsitselfto traders,inlargepart,onethicalgrounds.444Whileithasonly beeninoperationsinceOctober2013,445ithasenjoyeda successfullaunch.446

Inadditiontotheexchanges,atleastonedarkpoolhas allegedlybeentemptedtomisrepresentitselftoclientsinabid toprofitfromHFT.447Darkpoolsareoftenseenasmore favorabletoinstitutionalandretailtradersthanlitexchanges becausenoquotesarepostedbeforetradesaremadeandthe HFTfirmsarethereforedeprivedofakeyelementofthe informationneededtotradeaheadofslowerinvestors.448 Conversely,HFTpractitionersfinddarkpoolsalucrative environmentduetotheslownessandrelativelyuninformed natureoftheaverageinstitutionseekingtotradethere.449 AccordingtoNewYorkAttorneyGeneralEricSchneiderman, BarclaysclaimeditpreventedHFTfirmsfromtradinginits darkpool,BarclaysLX,whilesimultaneouslyandcovertly

/infrastructure/iex-and-the-re-mutualization-of-market-infrastructure/a/did/1252879.

444.Seegenerallyid.(describingthestructureandsupportofIEX). 445.

IEXAnnounces:ProductionLaunchFriday,October25th,2013asa 2013),

Non-Displayed ATS, http://www.iextrading.com/trading/alerts/2013/001/.

IEX

11,

446.

IEXnowaccountsfor0.8–0.9%ofthedailystocktradingvolume,and israisingfundstobecomearegisteredexchange.SeeMichaelJ.DelaMerced, Lower-FrequencyPlatformRaisesFundingtoBecomeFullStockMarket,N.Y. TIMES:DEALBOOK(Sept.2,2014),http://dealbook.nytimes.com/2014/09/02/iexupstart-trading-platform-raises-75-million-in-new-financing/.

447.SeeKeriGeiger,SamMamudi&ChrisDolmetsch,DarkPoolGreed DroveBarclaystoLietoClients,N.Y.Says,BLOOMBERG:BUSINESS(June25, 2014), http://www.bloomberg.com/news/articles/2014-06-25/barclays-darkpools-said-to-face-suit-from-n-y-attorney-general;ScottPatterson&Bradley Hope,BarclaysDarkPoolDrewEarlyAlarms,WALLST.J.(July20,2014, 7:14PM),http://www.wsj.com/articles/barclays-dark-pool-drew-early-alarmson-fast-trading-1405898592.AlsorelevantistheSEC’s$2millionfineofdark poolLiquidnetfordisclosinginformationconcerningtraderstoissuerswhose stockwastradedonitsdarkpool;suchdisclosureviolatesRegulationATS Rule301(b)(10).PressRelease,U.S.Sec.&Exch.Comm’n,SECChargesNew York-BasedDarkPoolOperatorwithFailingtoSafeguardConfidential http://www.sec.gov Trading /News/PressRelease/Detail/PressRelease/1370542011574; seealsoWilliam Alden,Liquidnet,aDarkPoolOperator,IsFined$2MillionoverCustomer 2014), Data, http://dealbook.nytimes.com/2014/06/06/liquidnet-a-dark-pool-operator-isfined-2-million-over-customer-data/?_r=0.

Information

N.Y.

448.SeeSHORTER&MILLER,supranote126,at5. 449.SeeGeigeretal.,supranote447.

toHFTfirms.450Themisrepresentation marketingitself allegedlyconsistedofprovidingfalsenumbersfortheestimated amountof tradingdonebyHFTfirmsunderBarclays’ “liquidityprofiling”metric.451IftheallegationsinAttorney GeneralSchneiderman’s true,Barclays compriseditsinstitutionalintegrityinanattempttoprofitin thecompetitiveworldofdarkpools.

complaint

are

Broker-dealersarethefinalgroupofmarketactorswho mayhavesuccumbedtothetemptationtoprofitattheexpense oftheirclients.Themaker/takerpaymentsystemappearsto incentivizebrokerstoroutetheirclients’limitorderstothe tradingvenuethatoffersthehighestliquidityrebate,rather thantheoneofferingthebestlikelihoodofexecution.452Since mostbrokerschargeclientsfixedcommissions,whilethe exchangesoffervariouslevelsofrebatesinthemaker/taker pricingsystem,“itmaybeprofitmaximizingforbrokersto considerliquidityrebatesratherthantheprobabilityoflimit orderexecutionwhenmakingroutingdecisions.”453Asaresult, thecurrentmarketstructureappearstoincentivizebrokerdealerstoviolatetheirdutyofbestexecutionforclients.454 Alongwiththeexchangesanddarkpools,thecurrenttrading

450.SeeComplaintat2–4,Peoplev.BarclaysCapitalInc.,1N.Y.S.3d910 (N.Y.Sup.Ct.2015) (IndexNo.451391/2014).Unliketheconsolidated litigationsInreBarclaysLiquidityCross&HighFrequencyTradingLitig., No.14-MD-2589(JMF),2015U.S.Dist.LEXIS113323,at*81(S.D.N.Y.Aug. 26,2015),andLanierv.BATSExch.,Inc.,105F.Supp.3d353,367(S.D.N.Y. 2015),AttorneyGeneralSchneiderman’ssuitsurvivedamotiontodismiss. Peoplev.BarclaysCapitalInc.,1N.Y.S.3d910,917(N.Y.Sup.Ct.2015).

451.SeeSamMamudi&DoniBloomfield,CracksOpeninDarkPool DefensewithBarclaysLawsuit,BLOOMBERG:BUSINESS(June26,2014), http://www.bloomberg.com/news/articles/2014-06-26/cracks-open-in-dark-pooldefense-with-schneiderman-barclays-suit.

452.SeeRobertH.Battalio,ShaneA.Corwin&RobertJennings,Can BrokersHaveitAll?OntheRelationBetweenMake-TakeFeesandLimit (unpublishedmanuscript), OrderExecutionQuality1–5(Oct.20,2015) http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2367462; also BLACKROCK,USEQUITYMARKETSTRUCTURE:ANINVESTORPERSPECTIVE7 http://www.blackrock.com/corporate/en-us/literature/whitepaper (2014), /viewpoint-us-equity-market-structure-april-2014.pdf.

453.ConflictsofInterest,InvestorLossofConfidence,andHighSpeed TradinginU.S.StockMarkets:HearingBeforethePermanentSubcomm.on InvestigationsoftheS.Comm.onHomelandSec.&GovernmentalAffairs, 113thCong.(2014)(statementofRobertH.Battalio,ProfessorofFinance, MendozaCollegeofBusiness,UniversityofNotreDame).

454.SeeBLACKROCK,supranote452,at2.

environmentappearstoofferincentivestemptingbrokerdealerstocompromisetheirinstitutionalintegrity.455 Liketheotherprinciplesoutlinedabove,

institutional integritydependsontheexpectationsofpartiesenteringthe financialmarkets.Whenweenterthefinancialmarketplace, wedonotexpectthatthevenueonwhichwetradeiscomplicit inorderingusupaspreytoanotherclassoftraders.Similarly, lyingwhentheymake wepresumeinstitutionsarenot representationstoparticipantsinthemarketsandthat,at somebasiclevel, thereisa“levelplayingfield.”These principlesarethe“conditionsofthepossibilityof”456financial marketsingeneral,andthatiswhycriticsofHFTultimately believeithasaseriouseffectonmarketconfidence.While marketactorscanviolatetheseprinciplestosomedegree,in orderforfinancialmarketstoflourishtothegreatestextent possible,theseprinciples,whichareessentiallymoral,mustbe respected.

CONCLUSION

Thefourprinciplesoutlinedaboveareabstractand essentiallymoral,buttheyalsoreflectbedrockprinciplesof Americansecurities Inmandatingextensive disclosureandpolicingthemarketsforfraud,insidertrading, andotherabuses,Americansecuritieslawgoessignificantly beyonda“caveatemptor”approach.457Sinceitisdifficultto drawusefulestimatesofcostsandbenefitsinasystemas complicatedasthecurrentequitymarkets,theprinciplesof

regulation.

455.Seeid. 456.ThismodeofargumentismodeledonKant’sphilosophicalmethod. Kantbelievedthatthe“categoriesofhumanunderstanding”functionedasthe “aprioriconditionsofapossibleexperienceingeneral”whichinturnallowus toanswerthequestions:“Howispuremathematicspossible?Howispure scienceofnaturepossible?”IMMANUELKANT,CRITIQUEOFPUREREASON56, 138(N.Kemp-Smithtrans.,St.Martin’sPress1965)(c.1781).Likewise,a modicumoftrustintheethicalprinciplesoutlinedaboveseemsanecessary preconditionforthepossibilityoffinancialmarkets.Forarelatedargument basedontheintersectionofKant’smoraltheorywitheconomics,seeJeanJacquesLaffont,MacroeconomicConstraints,EconomicEfficiencyandEthics: AnIntroductiontoKantianEconomics,42ECONOMICA430(1975).This Kantianmodeofargumentisalsofoundinthelawandeconomicsliterature. SeePasquale,supranote12,at2117(“Lawisconstitutiveofso-calledfinancial markets,notsomemeresideconstraintonthem.”).Pasquale’sargumentis basedonPistor’s“LegalTheoryofFinance.”Id.at2087–89(citingKatharina Pistor,ALegalTheoryofFinance,41J.COMP.ECON.315(2013 .

457.SeeGubler,supranote416,at415–23.

freedomfrommisrepresentation,alevelplayingfield,and institutionalintegrityshouldplayastrongroleinbuildinga regulatorystrategyforHFTthatbestservesthegoalofa stable,healthy,andjusteconomy.Theseprinciplesultimately reflectabeliefthatfinancialmarketstransactionsshould qualifyasreciprocalexchange.

financial

Whileequityconcernsareoftenpushedtothebackground regulation,wherefundamental inquestionsof uncertaintysurroundstheoperationofcausesandeffectsina systemascomplexasthecurrentfinancialmarkets,theyare particularlyusefulasexanteregulatoryguides.Asetofethical principlescanalsofunctionasafloororminimumwherea cost-benefitanalysisfailstoadequatelysafeguardtheinterests ofindividualsorgroupsofmarketactors,asisarguablythe casewithHFT.DespiteHFT’sroleinthemarketrevolution thathasgreatlyloweredcostsformostinvestors,thedetailed lookattheobjectionableHFTpracticespresentedinPartII illustratesthattheinteractionofvariousgroupsoffinancial marketstakeholderswithHFTprovidesgroundsforthe complaintthatHFTisnotfair.

AcloselookatHFTisalsoilluminatingbecauseitsuggests thatthegreatestchallengepresentedbyHFTmaybethatit underminessomeoftheimplicitassumptionsofRegulation NMS.AsDonaldMacKenzieobserves,RegulationNMSis predicateduponaNewtonianworldwhileHFTushersinan Einsteinianone.458RegulationNMSmakessenseifpricescan betransmittedinstantaneouslytotheSIP,butthespeedat whichHFToperatesrevealsgapsbetweendifferentmarkets, i.e.,theserversusedbytheexchanges,whichHFTexploits. RegulationNMSwasalsoformulatedwithlittleattentionto theradicalshiftunderwayinthebusinessmodelofthe exchanges,anditseemstopresumethatexchangescould continuetoearnmoneyintheirtraditionalmannerasthey competedwithoneanotherinanationalmarketsystem.In realitytheadoptionofmaker/takerpricingsystemsforces exchangestocompetewithoneanotherunderaradicallynew businessmodel.Iftheexchangesareforcedtocompromise theirinstitutionalintegrityforthesakeoftheirverysurvival inthenewtradingenvironment,thisposesthequestionof whetherthereisaneedformultipletradingvenuesinaworld whereallexchangesarecloselyknittogetherintoadigitized

458.SeeMacKenzie,supranote6,at41–42.

nationalmarketsystem.Attheveryleast,thecurrenttravails showthatthedisruptionsbroughtaboutbytheinformation technologyrevolutionrequireanewvisionofanequitable tradingenvironment.

--- FOOTNOTES ---

BUSINESS

BLOOMBERG:

BLOOMBERG:

BUSINESS

TABB

STOCK

EXCH.

INVESTIGATIVE

JOURNALISM

WALL

179.SeegenerallyROBERTJ.SHILLER,FINANCEANDTHEGOODSOCIETY

WIRED

BEAT

MACH.

CONF.

ELEC.

WALL

303.SeeSTEVENSHAVELL,FOUNDATIONSOFECONOMICANALYSISOFLAW

PROFESSOR

WALLST.&TECH.

DEALBOOK

TIMES: