Distributed execution in illiquid times: an alternative explanation of trading in stock markets

Abstract

Sociologists and other social scientists typically characterize financial markets as ‘liquid’, where capital can be readily purchased or sold in global electronic networks, so that trades can be completed nearly instantaneously. In contrast, drawing on ethnographic interviews in an emerging market, Malaysia, and on cross-national statistics, secondary survey data and econometric studies, this article argues that world stock markets are normally illiquid with temporal spikes of liquidity. As a consequence, professional investors in equities markets are large and frequently slow-moving behemoths, trading turgidly. To overcome illiquidity, professional investors conduct a range of practices I term distributed execution. Illiquidity and distributed execution have a number of implications for sociological debates regarding the relationship between investment firms and society.

Keywords:

• finance
• globalization
• investment
• liquidity
• Malaysia
• stock markets

Acknowledgements

This paper would not have been possible without the generosity of time from my Malaysian interviewees. All of the names of interviewees as well as their firms are pseudonyms to protect anonymity. Pseudonyms do not reflect race, religion or gender. I gratefully acknowledge funding from a Fulbright (IIE) Research Grant, an Illinois State University Pre-Tenure Faculty Initiative Grant and a College of Arts and Sciences Travel Grant. For valuable comments on previous drafts of this paper, I wish to thank Thomas Burr, Joan Brehm, Gary Paul Green, Scott M. Elliott, Dan Hirschman, Karin Knorr Cetina, Corrin Pitluck, Simone Polillo and anonymous referees from the Academy of Management Annual Meeting and this journal. Thanks also to my undergraduate research assistants, Mourad Bouajaja, Jason Carter, Ellie Wickes and Charlee Zingraf. All errors are my own.

Notes

1. A fascinating recent development in US stock markets that postdates my data is the rapid growth of ‘high-frequency’ trading strategies that use supercomputers to make trades in one-millionth of a second (Patterson & Rogow, 2009). At present (August 2009), the precise nature of this high-velocity trading, and even its legality, remains in dispute. My reading of the journalistic accounts suggests that high-frequency trading is exceptional and atypical of asset management firms’ behaviour. High-frequency trading is intra-day trading limited to already liquid stock (Patterson, 2009) and is conducted at the trading desks of only a few firms (Dughigg, 2009; Patterson & Rogow, 2009).

2. The 20 per cent figure was spontaneously used in several of my Malaysian interviews. Schwartz & Steil (2002, p. 46) argue that a threshold of 20 per cent is the implicit standard for a popular performance benchmark in the United States, VWAP (volume-weighted average price).

3. Chan and Lakonishok's (1995, p. 1149) sample comprises every New York Stock Exchange and American Stock Exchange trade attempted by thirty-seven institutional asset management firms between July 1986 through the end of 1988. These data consist of approximately 1.2 million trades, representing roughly 5 per cent of the total value of trading on the two exchanges during this time period. A second, smaller study discussed later in the main text, by Keim and Madhavan (1995), examines attempted trades by twenty-one US institutions in sub periods between 1991 and 1993. Their data comprised 62,333 orders with a market value of approximately $83 billion.

4. An additional source of downward bias which tends to exaggerate the speed at which orders are completed is an assumption the authors made in creating their sample. ‘Orders that took longer than 21 calendar days to execute’ (Keim & Madhavan, 1995, p. 377) were removed from their sample because these trades may represent ‘potential errors or unrepresentative trades’ (1997, p. 270).

5. The concentration of order flow is even higher in the United States where 73.6 per cent of all order flow passes through the five largest brokerage firms (Cetorelli, Hirtle, Morgan, Peristiani & Santos, 2007, table 4).

6. In Malaysia, a ‘sell programme based on one-third market volume’ means that there is a fund manager in the market who plans to sell as quickly as possible, for an unspecified length of time an unspecified amount of stock, as long as such selling never exceeds one-third of the day's trading volume.

7. This article is sceptical of arguments to slow down normal cross-border capital flows because equity markets are normally illiquid. This should not be interpreted as an argument against all forms of capital controls. Intuition, history and econometric studies find that abnormal surges of cross-border capital flows, often associated with inadequate regulation, increase the risk of a financial sector crisis (Reinhart & Rogoff, 2009, pp. 155–8).