Reading 41: Trading Cost and Electronic Markets

Page No. 83Date//Reading 4 : 0 strading Cost Andec 10 nic MarketsImplicit Trading Cost47 Indirect costs caused by market impact of trading (price concessionstraders make to Complete their trades is Known as marketimpact costs)Implicit trading costs are harder to measures and indude the bid-askspread , market or price impact costs , opportunity costs & delay(or slippage) costsThey must be est by measuring the results of a trade relativeto a reference point (i.e a bendmark)Implicit Cost ComponentsBid - ask (or bid-offer) Spread :Highest potential cost for a small trade.Trading at the posted bid or ask prices are quickly execulatedPatient traders may choose to trade bet" the posted bid &ask prices but run the risk of not filling their ordersLarger, liquidity demanding trades tend to move down thequote and are filled at a higher cost. Page No. 84Date//Market impact (or price impact)Cost of demanding liquidity in the marketsA larger order can be broken into smaller trades to go withMore for larger trades.the flow of the market & min the price impact costHowever it may increase the delay & opportunity losts.Delay Cost (or slippage)Cost of an adverse price movement during the lag Inexecuting a large trade.- Risk when breaking larger orders.Opportunity CostFrom unfilled ordeds or failded trading opportunity-Adverse price movement during the lag makes thetrade unattractive.Liquidity Impact of a TradeMarket orders (at posted bid /ask) demand liquidityStanding limit orders with posted bid I offer supply liquiditySpreadsQuoted Spread = offer price - bid priceMidquote = (bid price + offers price)2Inside spread = best (lowest) offer - best (highest) bid Page No. 85Date//17Effective Spread :Bid + 2 Ask ,Effective spread per share = (side ) X (trade price - mid quote)where : side = H for buy orders and -1 for sell orders.Effective Spread = 2 X (effective Spread per share)Effective Spread = qouted spread for transactions at quotedbid laskprices0 When buying int > selling int 7 dealers raise ask prices to discouragebuyers and raise bid prices to encourage seller0When buying int < Selling int / dealers lower ask price to encouragebuyer&laisebid price encouragesellerExamplesBidsAskDealerPriceSizeDeaherPriceSizeP32.253000R32.292000Q32.242500P32.313000R3500Q32.2332.353500a The spread qouted by dealer Qb] The inside spreadI The midquote for dealer501a Spread quoted by dealer Q = 32.35 - 32.24 = $0.11bBest ask price = $ 32.29 & best bid price = $32.25Inside spread = 32.29-32.35 = $0.04 Page No. 86Date//(Midquote price for dealerR = (32.23 + 32.29) = $ 32.2622] Suppose that a purchase transaction occured at $32.29 Usingdealers P quoutes only, calculate the effective spread & thequoted spread(Midquote price for dealer P = (32.25 + 32.31) = $ 32.282Quoted Spread = 32-21-32-35 = $0.06Effective spread = 2XCH) X (32.29) - 32.28) = $0.026Because effective spread < quoted spread + there was a priceimprovement on that tradeLimitations of Effective SpreadPoor indicator when a large order is split (does not accantfor the price impact cost)Does not alc for slippage or delay costs when part of theorder does not get filled at desired prices.of the order is cancelled due to adverse price movementsDoes not account for opportunity cost when an unfilled part with a folio buy overdayPage No. 87asellordertrade executed during Date//allthe time interva2VolumeweightedAvgPrice(VWAP)$ evolure of each tradeVWAPtransactionCost(side)tradesize(tradeVWAP-benchmarkVWAP)Whereside=+1forbuyorders&-lforsellordersExampleThefollowingtradesareobservedforBNCincstockduringtherelevanttimeintervalAt10:00 am100sharetradeat$12.11At11:00 am300sharetradeat12At11:30 am600Sharetradeat$11.75Supposurethatthesubjecttradeforthepurchaseof300Shares,whichwasexecutedasfollows100Shareat12.11200 Shareat$ 12ICalculatethebenchmarkVWAP&tradeVWAPforBNL2CalculatetheVWAPtransactioncostSharestradedduringtherelevanttimeinterval=1000BenchmarkVWAP100$12.11t300=X$ 12 + ( 600 - 100 X $11.7510001000= $11.86 Page No. 88Date//Trade VWAP=100x1211t200x12= $12.043003002VWAP transaction cast = 300 x (12.04 -11.86) = $54Limitations of VWAPNot useful if the trade is a significant part of the tradingvolume (benmark VWAP & the trade VWAP will beclose to each other)Does not Capture the price impact cost3] ImplementationMarket impact coslShortfallDelay costsOpportunity costsImplementation shotfall = Value of actual folio - Value of paper folioThe paper folio is valued at the prevailing price lie thedecision / arrival / strick price) when the order was placedCaguie all their implicit CostFactors Driving Electronic MarketsLower cost than human tradersAccuracy in following trading rulesAudit tlails from recording generated / maintained by electronic systeFraud prevention Keep hidden orders hidden ( ho leakage)Continuous markets - not affected by weather etcEffectivenessarround that clockConvinionsAscessibility Page No. 89Date//AdvantageDisadvantageEasy to interpertInterpreting VWAP transaction costAllews trader to determineest are problem atic when tradeswheathertrade is betterare a significant portion of VWAPworsethan avg tradingbeachmark or take place a thepricesame rate as of other benchmark =trades & VWAP benchmarkTraded VWAP (trades are not at costly )Electronicbondorder - matching exist investors but theyprimarily serve dollours¬publicMarket fragmentationsMarket fiag mentation results from the same securitytradingmultiple platformsduetoproliferationofonelectronicmarketAutomated trading Strategies such as liquiditity book agyreg ationalgorithms create a super order displayingliquidity across all marketsSmart order reuting elegoriths send order s to themarket with the best prices 4 Size@ Trading the same instrument in multiple venue - Marketfregmentation increases potential for price & liquidity diffacross venuesFor example, trading of exchange - listed equities is now dividedamong several exch alternative trading systems & 00 numerous dealears Page No. 90Date//0 Trading respond to market fragmention by searching for liquidityacross multiple venues and time to control market impactof large trades*Electronic Trades IMPElectronic news readers analyze high speed news feed and generateorders before market price adj.Electronic dealer :- post standing orders and profit from spreadElectronic arbitrageurs :- Use round - trip trades in multiple marketsto Capture princing discrepanciesElectronic front Tunn ers - use patterns / events to trades ahead ofanticipated ordersBuy - side traders use electionic brokers to execute trade forportfolio managersElectronic quote matches use the valuable option provided byStanding limit orders Eg a posted bid provides a floorat which (un get outEffects on trades transactioncostGrowth in electronic trading has the following effects on transactionCosts:-Grouth in electronic Numerous studies prove the decline inSubstantial decline in bid- ask spreads which lowerstransaction costs due to electronic trading due to cost efficient.transaction Costs for retail trades and So instritionstrading small orders Page No. 91Date//Transaction costs have also decreased for larger ordersOne study shows that the implementation shortfall costcost of filling thousand of equity orders for usstocks has dropped.*Types of traders (electronic)Proprietary tradersBuy-side tradersElectronic brokersNotes:-3> Proprietary trades are profit - motivated trades and includedealers arbitrageurs & various types of front runners.Biy- side traders trade to fill orders for invest & riskmanagers who use the markets to establish positions forhomeself and derive various utilitarian & profit - motivatedbenefitsHow electronic traders send orders to marketsProprietory traders registers as brokers / dealers directly sendtheir orders to exch.Proprietary tradess who are not brokers / dealers must sendthier orders to brokers who forward them to exch who givethem sponsared access to fast electronic processing systemAElectronic proprietary traders includes high-fieg trade (HFT)and low-latency tiaders Both must trade very quicklyin response to new info Page No. 92Date//CHigh freq TradersComplete trades very quicklyMay trade in & out of sec or contracts thausands of timebut usually only in small sizeHold positions for periods less than a dayLow- Latency AdvantagesLatency is the time lag bet occurrence of an eventandSubmission of a traded on that eventLatency is Composed of :-Time gap bet" publication of news and receipt of that Dewby the trader (communication speed )Time gap bet recepit of news and determination ofappropriate order (commputing speed)1 Time gap bet" determination of orders & recepit ofthe order by the exch ( Communication speed )Include news traders & parasitic traders- News traders feed on electronic news feed- Parasitic traders are speculators who base their pridictionabout future prices on info they obtain about orderswhich other traders will fill or intend to fillNeed to send or cancel orders very quickly when openingorclosing positions.May hold positions for as long as a day of evenlonger in confrast to HFT. Page No. 93Date//* Need for SpeedCommunication Speed - enhanced by using high speed fibernetwork & reduction of distance travelledCollocate - Servers at the exch if such a service is atavailableComputing speed Overclocking the latest processors, fast memoryno frills as , assembler languagein Taking Making CancellingCharacteristics and uses of electionic trading systemsBuy -side traders use electronic brokers for advanceofadvanced orders, trading tactics & alogrithms providedbytheirelectromic brokers to search for liquidity*Advanced OrdersUse dynamic limit prices + when an exch supports itThe limits prices change automatically based on somerule (old order gets deleted & replaced by a new order)Example:- A trader may submit a bid price of 2 ticks belowthe best ask As prices changes , the trader's quote alsoautomatically changesTrading TacticsSubmission of multiple orders : Used to uncover hiddenorders or discretionaly orders (orders that have a limit+ - discretion amt specified). These are Stiu ctured to be Page No.94Date//C immediate or cancel (10C) orders .Often used while trading at dark pools (venues that bideliquidity and restrict trading to a select clientele)Also to execute midspread orders. (i.e limit orders peggedto the midpoint of the quoted bid-ask prices)Electronic Trading StrategiesHidden Orders : Limit orders hidden from the orders books(so as the not provide a valuable option to trade)Leapfrog : To beat the best bid / offers price Efficiencynarlow inside tradeFlickering quotes : 10C orders designed to not given valuabletrading option that standing limit orders provideExposue limit order & cancel imidatlyMachine Learning : Techniques that allow the model to evolveQuate matchingbased on new data Holy grail : money machineElectronic ArbitrageTake liquidity on both sides : Market orders to trade atposted bid/ask prices simultaneously in diff marketsOffer liquidity on one side : One leg is market order andthe other is a limit order Risk of not filling thelimit orders. Page No. asDate//Offer liquidity on both sides : Limit orders for bothlegs at prices worse than the best bid lasklowers trading cost but more riskyEx_ : BenKhuslowobservesthefollowingquotesforstock22MarketBest BidBid sizeBest askAsk sizeAlpha10.0620,00010.1218,000Bravo10.0215,00010.05(1000!Khuslow takes liquidity on both sides of the market,calculate the profit per share that Khuslow Can expectto carn of the quotes hold up for both sides of thetrade)Sop The best - ask across the two markets is $10.06 - $ 10.05Marketpurchase order at $10.05 simultaneously a market sellorder at $10.06 ( both for 11,000 share )4the prices doesnot move bet" the time Khuslow getsthe quotes, submits the order and the orders are receivedby the exch both segment would / be fillied t resultingthe $0.01 per share arbitrage profit (or $110 on11000 shares ) Page No.9611th Wed 2022Date / /Example 2Ben Khuslow observes the following goutes for stock 22:Market Best Bid Bid size Best ask Ask sizeAlfa10.0620,00010.1218000Bravo10.0215,00010.0511000How can Khuslow increase the expected return per share by eitherb] offering liquidity on one side of the market , ora]offering liquidity on & both sides of the market ?What are the risks of each of these Strategics ?Khuslow can submit a limit purchase order with a bid of$10.04 in market baravo If trades gets filled a marketsell order in market Alfa at posted bid of $10.08would now result in a profit per share of $ 10.06 - $10.04-$0.02bA limit bid of $10.04 in market baravo & limit ask of $ 10.07inaffa would result in profit per shale of $0.03Limit order have higher failure - to - fill riskElectronic Trading RisksHFT arms race : An unfair advantage for some traders Page No. 97Date//,Systemic Risk :pacties other than the party responsible for the tradeThe Cost of bad trades being borne byRunway alogorithms : Eg flash crash of May 2010TFat finger errors : If large enough can spike volatilityOvercharge orders : Q Demand more liquidity than avail .Malevolent orders : Designed to manipulate marketsMitigation of Systemic RiskProper checks & balance-Authorized order input / software updates-Software should be adequately back tested- Brokers should inspect orders before submitting them-Markets should have limits to price changes & automatedtrade halft when vdatility spikes 6ver a thresh old lul7Abusive Trading PracticesDemand Supply1for Trade which "Demand liquity" Use qudte matching when tradeFront running :- Trading ahead of Known orders Algorithmssniff out large orders that are being split &Sent to an exchMarket ManipulationTrading for market impact :- Sometimes at a large cost.Rumormangering :- The dissemination of false infoWash trading To create false impressions about liquidity Page No. 98Date / /Spoofing / layering :- Fake limitation orders to create fakeoptimism / pessimismBluffing :- To prey on momentum traders Leg pump & dumpshort and distroits)Gunning the market :- Pressure others into a bad & tradeSqueezing & Cornering :- Obtaining control of settle thenwithdrawing resources needed to contracts(eg short squeeze, Cornering in a commodity)