Order anticipation around predictable trades

• Date: 01 March 2020
• Author: Mehmet Sağlam
• DOI: http://doi.org/10.1111/fima.12255
• Published date: 01 March 2020

DOI: 10.1111/fima.12255

ORIGINAL ARTICLE

Order anticipation around predictable trades

Mehmet Sa˘

glam

TheCarl H. Lindner College of Business,

Universityof Cincinnati, Cincinnati, OH

Correspondence

MehmetSa ˘

glam,The Carl H. Lindner College of

Business,University of Cincinnati, Cincinnati, OH.

Email:mehmet.saglam@uc.edu

Abstract

I study the presence of order anticipation strategies by examining

predictable patterns in large order trades. I construct three sim-

ple signals based on child-order execution patterns and find empir-

ical evidence that stronger signals are correlated with higher execu-

tion costs. I use the SEC's (Securities and Exchange Commission's)

ban on unfiltered access and increase in noise trading as shocks

to order anticipatory activities of algorithmic traders and find that

the price impact of predictability is smaller when order anticipation

becomes difficult. The empirical findings are mostly consistent with

the back-running theory that predicts delayedprice impact as strate-

gic traders learn about large orders gradually.

1INTRODUCTION

The recent advances in trading technology combined with machine learning theory have provided tools to sophisti-

cated investors to extractvaluable signals about asset prices and make dynamic trading decisions at a high frequency.

Usuallyreferred to as high frequency traders (HFTs), they follow various trading strategiesranging from market making

to arbitrage trading between multiple assets or venues.In response to their rise in trading activity in multiple markets,

therehave been several academic studies that attempt to decipher the costs and benefits associated with this new form

of trading.1The empirical evidence has been mostly positive pointing to smaller bid-ask spreads and faster price dis-

covery.However, the implications are still controversial, especially in the public's perception.2The keyquestion in this

this debate is: Can strategic algorithmic traders use order anticipation strategies to sniff out large orders and exploit

this information to profit at the expense of other investors?

Schedule-based algorithms seeking to match time-weighted average prices (TWAP) or volume-weighted average

prices (VWAP) in the marketare believed to be the main source of information leakage in large order executions. In a

recent survey done byITG, a financial technology company implementing algorithmic execution services, roughly 50%

of buy side investors reported that their biggest source of information leakage occurs in schedule-based algorithms.3

In some cases, predictable patterns emerging from these algorithms can even be discerned by human traders.On July

19, 2012, four large cap stocks, Coca-Cola, IBM, McDonald's, and Apple, displayed identical price patterns potentially

due to a schedule-based algorithm.4In Figure 1, I plot the trade prices for Coca-Cola. In odd half-hour intervals, the

price is roughly decreasing and in half-hour intervals, the price is increasing. Interestingly, the peaksoccur roughly at

c

2018 Financial Management Association International

Financial Management. 2020;49:33–67. wileyonlinelibrary.com/journal/fima 33

34 SA ˘

GLAM

FIGURE 1 Schedule-Based TradingAlgorithms. This figure illustrates the trade prices for Coca-Cola on July 19,

2012 from the TAQdatabase. There is a strong sawtooth trading pattern potentially caused by schedule-based

algorithms

the half-hour mark, while the lows appear at the start of the hour. These price dynamics can be exploitedby strategic

traders. For example, in anticipation of the continuing pattern, the low price corresponding to t=4.5 occurs a few

minutes earlier leading to a trade priced at $76.75. This is significantly lower than its past 30-minute average price.

Overall, this real world example providesevidence that some strategic traders may take advantage of these patterns

with further help from machine-learning techniques.

Investors and policymakers are concerned about these order anticipation strategies due to the potential negative

effects on price discovery and liquidity.If a trader can infer the presence of a large buy order submitted by a long-term

investor for liquidity reasons, they can trade along with the investor during the initial period of the execution to

overshoot the price and then sell back to the investor their accumulated position at an elevated price to achieve

roughly riskless profits. In return, this means larger execution costs for the long-term investors (Brunnermeier &

Pedersen, 2005). If the long-term investor is tradingdue to private information that was generated with costly effort,

then a strategic trader can extract this information using order anticipation strategies and be part of the resulting

profit at the expense of the investor.This implies that the incentive of the investor to invest in information acquisition

would decrease in the presence of order anticipation. In this case, as argued by Stiglitz (2014) and Weller (2018), the

markets can be less informative if algorithmic traders can share the information rents of the fundamental investors

who spend resources to obtain information about the real economy. In order to differentiate these two different

motivations, I will use “predatory trading” and “back-running” (as introduced by Yang& Zhu, 2015) to refer to the first

and second types of anticipatory trading, respectively.In predatory trading, the resulting price impact is expected to

be temporary as the investoris not informed, while in the presence of back-running, the price impact is permanent due

to information-based trading.

In theory, order anticipation strategiescan also lead to desirable market liquidity conditions usually referred to as

“sunshine trading.” This possibility is often ignored in academic discussions on order anticipation. If a large order exe-

cution is submitted by an uninformed investor,predictable executions can actually motivate market makersto provide

additional liquidity knowing that there is no adverse selection risk. Admati and Pfleiderer (1991) formalize this theory

and find that the public announcement of an uninformed liquidation may reduce trading costs. They directly assume

thatmarket makers have perfect knowledge about the uninformed liquidation ex ante. In the contextof large order exe-

cutions, anticipatory traders can gradually learn whether the large executionis informed or not by analyzing the price

impactof past trades. Depending upon the accuracy of this exploration process, algorithmic traders may be incentivized

to provide greater liquidity during the lifetime of the execution.In the empirical literature, only the perfect information

case has been extensivelystudied. For example, Bessembinder, Carrion, Tuttle,and Venkataraman (2016) find support-

ing evidence of sunshine trading in large executionsoccurring in crude oil exchange-traded funds (ETF) rolls.

SA ˘

GLAM 35

Given the mixed implications of predictable trading both theoretically and empirically,it is important to study the

net effects arising from recurring patterns in the order flow data. In this paper, I empirically investigatewhether pre-

dictable patterns in large order executionslead to higher or lower trading costs. Using natural experiments, I construct

the potential channels between order anticipation and predictable patterns. Then, I examine the consistency of my

empirical findings with predatory trading,back-running and sunshine trading.The empirical analysis utilizes more than

20,000 parent-orders constituting more than 2.5 million child-order executions.My sample includes 15 months of data

on liquid S&P 500 stocks from January 1, 2011 to March 31, 2012. The data set consists of large orders submitted by

146 distinct investors comprised of primarily institutional portfolio managers. All of the orders in the data set are exe-

cuted to match the VWAPrealized during the lifetime of the parent-order. The average order size is roughly $1 million

and corresponds to roughly 1.8% of the volume traded during the execution.

If algorithmic traders follow anticipatory trading,their activities should be particularly easy to detect in predictable

executions. Here, I define executionpredictability as the likelihood that a strategic trader can succeed in inferring the

presence of a large order execution. I provide three simple signals that quantify this measure of predictability utiliz-

ing statistics from the child-orders of the execution. These signals are constructed based on the intuition regarding

minimizing information leakage.

The first signal is computed using the volatility of the size of the child-order trades. Forexample, if a large order is

executedin a series of equal trade sizes (e.g., 150 shares), the pattern recognition algorithms may deduce the existence

of a large order with greater likelihood. In addition, using a similar intuition, I consider the regularity of the trading

intervals as a signal that can leak order size information and propose a signal that computes the volatility of time inter-

vals between successive trades. This is motivated bythe empirical evidence in the literature that execution algorithms

exhibit robust clock-time periodicity,the tendency to make trades around full seconds or half seconds as identified by

Hasbrouck and Saar (2013). It is also possible to consider the order size and trading frequency at the same time and

propose an aggregate measure involving the volatility of the trading rate.Thus, the final signal is based on the corre-

lation between the executedquantity and the elapsed time and will be very close to one for executions with an almost

constant trading rate.

It is worthwhile to emphasize that in the optimal executionliterature, some of these measures are argued as factors

thatactually decrease execution costs. However,these results are based on the absence of any other competing traders

in the marketplace. For example,in the seminal paper on optimal liquidation of large blocks of shares, Bertsimas and

Lo (1998) find that equal partitioning policy is optimal if the price impact is permanent and linear. This suggests that

the volatility of the traded quantities of child-orders or trading intervals should be at minimum. Similar deterministic

algorithms are proposed in order to accommodate U-shaped volume profiles in the markets (e.g., VWAP). However,

in the presence of strategic traders with order anticipation skills, I argue that these deterministic strategies would

actually increase the cost of trading.

This paper contributes to the literature by providing robust signals that can quantify executionpredictability and

analyzingtheir impact on execution costs using liquid S&P 500 stocks. Since the execution strategy is known and unique

along with a single broker,the data set allows me to clearly verify the impact of predictable executions. The empirical

findings are consistent with earlier literature studying the link between HFT activity and institutional trading costs.

They point to another direct channel for the cost increase that is based on execution predictability.Using a diverse

universe of institutional investors in terms of short-term trading skill, I find evidence of back-running strategies. The

cost increase is economically significant. Analyzing uninformed executions, I do not find evidenceof sunshine trading

(i.e., predictable uninformed executionsdo not lower trading costs).

The empirical analysis provides four main contributions. First, the signals are significantly correlated with the

execution costs after controlling for the standard determinants of price impact implying the potential presence of

order anticipation. As the predictability of the execution goes up, I find that the execution cost measured by its

implementation shortfall (IS), the percentage deviation of the averageexecution price from its starting price, increases

by economically substantial amounts. The median execution cost is 2.7 bps and I find that a one-standard deviation

increase in my signals increases the execution cost by a range of 1.6 to 1.8 bps after controlling for a rich set of

executionlevel statistics including the use of marketable and passive limit orders. Further,I confirm that lagged signals