Informed Trading and Market Structure

• Date: 01 January 2015
• Published date: 01 January 2015
• DOI: http://doi.org/10.1111/eufm.12003

European Financial Management, Vol. 21, No. 1, 2015, 148–177

doi: 10.1111/eufm.12003

Informed Trading and Market

Structure

Charlie X. Cai

Bradford University Schoolof Management, UK

E-mail: x.cai1@bradford.ac.uk

Jeffrey H. Harris

Syracuse University, USA

E-mail: Jhharr03@syr.edu

Robert S. Hudson

University of Newcastle, UK

E-mail: robert.hudson@newcastle.ac.uk

Kevin Keasey

International Institute of Banking and Financial Services, University of Leeds, UK

E-mail: K.Keasey@lubs.leeds.ac.uk

Abstract

We examine London Stock Exchange trading around information releases and link

market quality dimensions with market structure during periods with heightened

interaction between informed and uninformed traders. We find support for both

the hypothesis that automated electronic markets minimise trading costs for

liquid stocks and the hypothesis that adverse selection costs are minimised with

intermediated trading. We examine how news affects both dealer and electronic

systems and find that electronic markets are prone to greater stealth trading and

post-trade volatility, both consistent with the proliferation of algorithmic trading

and short-term volatility events such as the May 6, 2010 ‘flash crash.’

Keywords: market structure,informed trading,trading systems,London stock

exchange

JEL classification: G10, G15

Wewould like to thank Kirsten Anderson and seminar participants at the University of Leeds

and the University of Brunel for their helpful comments. We thank John Doukas, the editor

of this journal, and the anonymous referee for their constructive and helpful comments. All

errors are our responsibility. Correspondence: Charlie X. Cai.

© 2013 John Wiley & Sons Ltd

Informed Trading and Market Structure 149

© 2013 John Wiley & Sons Ltd

1. Introduction

Technological advances have created unprecedented opportunities for developing a

variety of market structures for trading stock. In the USA, both Nasdaq and the

New York Stock Exchange (NYSE) have adopted electronic communications network

(ECN) technology as their primary trading platforms.1Indeed, technology has improved

market quality along many dimensions and, by facilitating automated executions, has

ushered in an explosion of trading volume. However, as the 6 May 2010 ‘flash crash’

demonstrates, technological changes also present challenges to financial markets. While

electronic systems facilitate algorithmic trading and high frequency trading strategies,

the diminished role of human interaction raises questions about the stability of these

electronic systems.2

A number of research papers examine various features of trading systems to assess the

efficiency and effectiveness of different market structures.3Generally, this work finds

that automation may providedirect cost advantages, but Venkataraman (2001) showsthat

electronic, disintermediated markets may suffer from large adverseselection costs which

may outweighpotential direct cost savings. In this paper, we use London Stock Exchange

(LSE) data to further examine the interaction between market quality characteristics and

the structure of the trading system itself. Our tests recognise the fact that traders face

both direct costs and adverse selection costs when executing orders. We first examine

the direct cost minimisation hypothesis that lowercost automated electronic markets are

most effective for liquid stocks with an abundance of public information available for

making trading decisions. Under this hypothesis, intermediation adds needless costs to

the trading process and, the trading system may affect the strategies that traders employ

in the search for the most cost effective execution.

Wethen examine the second (not mutually exclusive) hypothesis that adverse selection

costs are minimised with intermediated trading so that intermediated trading benefits

less liquid stocks where market makers can mitigate more substantial adverse selection

costs. Indeed, whenrelatively little information is available about smaller stocks, concern

about adverse selection costs may outweigh the prospects of higher direct costs. Under

this hypothesis, market makers add liquidity to the market, lowering the cost of informed

trading and accommodating larger trade sizes overall.

In order to test these hypotheses about how market structure affects market quality

characteristics, we would ideally like to measure market quality for stocks trading in

both dealer and electronic markets, ceteris paribus. In the absence of this opportunity,

one approach examines stocks which switch trading venues.4While this approach holds

1Using INET and ArcaEx platforms, respectively. The BATS Exchange grew from an ECN

to execute over 10% of US matched volume in just four years.

2Hendershott and Menkveld (2010), Brogaard (2010) and Jovanovic and Menkveld (2010)

each address the costs and benefits of high frequency traders and their use of trading

algorithms.

3Thiessen (2002), Handa et al. (2004) and Jain (2005) examine floor- and screen-based

systems. Berkman and Hayes (2000), Madhavan and Panchapagesan (2000), Sof ianos and

Werner (2000), and Battalio, et al. (2007) highlight the importance of floor participants,

while Benveniste et al. (1992), Madhavan and Sofianos (1998), Anand and Weaver (2004)

and Anand et al. (2010) characterise the value of a specialist.

4See, for example, Sanger and McConnell (1986), Christie and Huang (1995), Barclay

(1997), Clyde et al. (1997) and Kalay and Portniaguina (2001).