High Frequency Trading and US Stock Market Microstructure: A Study of Interactions between Complexities, Risks and Strategies Residing in U.S. Equity Market Microstructure

• DOI: http://doi.org/10.1111/fmii.12068
• Published date: 01 May 2016
• Date: 01 May 2016

High Frequency Trading and US Stock Market

Microstructure: A Study of Interactions between

Complexities, Risks and Strategies Residing in U.S.

Equity Market Microstructure

BYSAMIR ABROL,BENJAMIN CHESIR,NIKHIL MEHTA,AND RON ZIEGLER

We examine the conditions, complexitiesand risks of a fragmented market microstructure

to contextualize the role of algorithmic and high frequencytrading in the US equity markets.

The establishment of a national market system and Regulation NMS was meant to promote

competition, recognizing the evolutionand changing dynamics introduced by technological

innovation. This evolutionand governing rule set has had many positive effects in terms of

competition, fee compression, tighter spread potential and volumes. Our paper identifies

certain unintended consequences and complexities of the national market system including

fragmentation, sub second quoting and trading, complex order types, data asymmetry,

technological innovation, unique strategies and the algorithms that power them. When

acting in concert, these complexities give rise to opportunities as well as emerging risks.

This high-speed system can be unstable and susceptible to inherent conflicts of interest,

market abuse and price shocks. These shocks can be amplified by positive feedback loops

accelerating single stock declines and also posing systemic risks in time scales beyond

real-time physical human comprehension and reaction times. Furthermore they can produce

contagion, which we refer to as ‘Flash Splashes’ caused by rapid withdrawals and injections

of liquidity in increasingly linked asset classes, indices, sectors and global liquidity pools.

High frequencytrading strategies can be both passive and aggressive and usually display risk

averse and low inventory characteristics. These strategies leverage fragmentation as they

create or capture informational asymmetries. They interact directly with sell side algorithms

that can hide intentions, hunt liquidity and sweep the order book. These interactions create

market dynamics that can benefit and challenge anyone exposed to US equity markets.

Every market participant has a risk profile unique to their strategy and objectiveand while

regulations will be enriched or revisedand certain unfair practices eliminated great attention

should be paid to understanding modern high speed trading risks and both the positive and

negative impacts on all stakeholders. Wehave examined the regulations, complexities and

risks to bring clarity and understanding to the current trading ecosystem for its users.

I. INTRODUCTION

The first stock exchange in the United States was the Philadelphia Stock Exchange

formed in 1790 followed by the New York Stock Exchange (NYSE) in 1792,

which was famously formed under the spreading boughs of a buttonwood tree by

its brokers. Over two centuries later in 2010 the NYSE would plant buttonwood

trees alongside its Mahwah data center where most of its trading now takes place.

It was not until 1971 when technological advancements and serious competition

emerged from The NASDAQ Stock Market (NASDAQ).

NASDAQ was unique because it did not need a physical location for traders

to congregate. Its business model was the enablement of fast and transparent

[Correction added on 20 April 2021, after first online publication: author Ron Ziegler has been added

in this version.]

Corresponding author: Samir Abrol, 142 Garth Road Apt 5J, Scarsdale NY 10853, USA. Email: Samir.Abrol@

me.com.

C2016New York University Salomon Center and Wiley Periodicals, Inc.

108 Samir Abrol et al.

computer based trading, a major shift from the entirely manual legacy NYSE

specialist model. The NASDAQ model was emulated by virtually every exchange

the world over and its technological leap is the psychological underpinning to

explain why US high tech companies chose to list its shares there. The birth

of NASDAQ put pressure on the bid/ask spread made by specialists and market

makers and forced the NYSE itself to change and adapt over the years to come.

Instinet was originally created in 1967 as ‘Institutional Network’ with a mission

to provide institutional traders a place to match large orders in NYSE stocks

without being subject to human specialists handling trades. Eventually Instinet

became a large electronic trading network for over the counter stocks away from

the public markets and reported prices after trades were completed. Withelectronic

signal processing and computer assisted trading taking hold in the mid 1990’s

attempts were made to open Instinet access to the broader trading community.

Demand for a market without human intervention was growing particularly in the

years leading up to the dot com bubble and explosive stock exchange volumes.

In 1996 Island was created with the mission of matching trades electronically

away from exchanges and away from market makers who had been accused

over the years of front running clients, maintaining excessively wide spreads and

selectively filling customer orders. Island, essentially a pool publishing its prices,

was capable of accepting computer driven order flow. Signals could now be

automated and orders sent directly into a market-matching engine.1In only three

months in 1996, 5.6 billion shares traded through Island at roughly the same time

NASDAQ market makers agreed to a $1 billion settlement related to charges that

they were inflating spreads for retail investors. This case created new SEC ‘order

handling rules’ which forced NASDAQ to publish competitor quotes alongside

market makers on the national system creating more competition. Market makers

were still able to leveragethe Instinet market to impact prices in a foreshadowing of

strategies that would be deployed years later in modern HFT.As spreads narrowed

and more venues or pools emerged, new participants were attracted to the markets

perceiving edge that can be realized with technology.Computer based quantitative

trading firms were now sending superfast order flow and devising programs to

jump to the head of a market, discover liquidity and leverage price discrepancies.

Sell side algorithms were born to combat aggressive computer based traders first

extending standard Volume Weighted Average Price (VWAP) and Time Weighted

Average Price (TWAP) programs to others named Cobra and Sniper for exam-

ple, which attempted to mask and randomize executions to ‘fake-out’ algorithmic

predators. Algorithmic traders in turn began to developmachine learning programs

and artificial intelligence to dynamically predict other trader intentions by their

quote and trade placement patterns. They developed automated strategiesdesigned

to hunt and capture quote discrepancies based on informational asymmetries intro-

duced by an ever increasingly fragmented market. Speed advantages were sought

and exchange co-location was born to minimize latency and route times. Explosive

1Patterson, Dark Pools,2013: “By late 1996 half of Nasdaq’s SelectNet trades came from Island”.

High Frequency Trading and US Stock Market Microstructure 109

trade volume led to everincreasing competition among new venues and the change

from utility-based exchange structures to ‘for-profit’ public companies. To chase

down new order flow and cater to high speed trading customers, exchanges, pools

and ECNs created a menu of services, order types and fee schedules in an effort

to cater to their best clients. The new market makers were now computers trading

far smaller sizes at much higher frequencies with low inventory and overnight

positions. Average trade sizes actually shrunk on NASDAQ from 1500 in the mid

90’s to about 500 in 2005.2Trading speeds were now measured in microseconds

representing one-millionth of a second.

The growing web of trade routes, destinations, strategies, speed advantages

and order types introduced new complexities that have sped ahead of regulators

ability to evaluate and react to changes in match engine functionality and service

offerings available to traders of all types. The market now consists of traditional

strategies based on asset pricing arbitrage, position inventory and risk interacting

with modern strategies based on prediction and discovery of contra trader inten-

tions having risk averse low inventory profiles. Algorithms on both side of the

trade are deployed to both obscure and exploit trader patterns and goals. The com-

plexities introduced as a byproduct of fragmentation have contributed to a high

speed market system susceptible to portfolio and systemic risks. Other issues such

as statistical front running and informational asymmetry are of concern and can be

mapped to specific complexities and regulation. Many point to the ‘flash crash’,

the ‘Knight-mare’, botched Facebook and BATS IPOs as examples of increasing

fragility. Regardless of all the sensationalism it is a factthat low volume, low risk,

high turnover HFT now accounts for over half of all trades on US exchanges.3

This paper is organized as follows: Section I contains the background and our in-

troduction to the current state of play in the US equity markets. Section II presents

the rules of the game; National Market System, Regulation ATS and Regulation

NMS that have facilitated competition, narrower spreads and created fragmenta-

tion giving rise to algorithmic and High Frequency Trading (HFT). Section III

addresses complexities in modern market microstructure and specifically analyze

data asymmetries in fragmentation contributing to the National Best Bid Offer

(NBBO). Section IV focuses on new risks including conflicts of interest, market

abuses, price shocks and what we term ‘Flash Splashes’ which can introduce sys-

temic contagion caused by asset linkages and rapid injections and withdrawals of

liquidity. We present our conclusions in Section V.

II. REGULATION

NATIONAL MARKET SYSTEM

In 1975, Congress directed the Securities and Exchange Commission (SEC)

through enactment of Section 11A of the Exchange Act, to establish a national

2Patterson, Dark Pools,2013.

3Financial Stability Oversight Council, 2012 Annual Report, July 18, 2013.