Executive Summary:
Issues in the Design of Electronic Stock Exchanges
As the NYSE seeks new directions in its two-centuries old methods of trading, research by SMUCox finance professor Kumar Venkataraman and Hendrik Bessembinder offers insight and evidence into new possibilities and difficult issues facing the Big Board. In “Does an electronic stock exchange need an upstairs market?” in July’s Journal of Financial Economics the authors find that a large proportion of institutional trading in electronic exchanges is executed away from the centralized book in the informal upstairs market because of automation, thus presenting new challenges. Recommendations for the needed features of the next generation of electronic stock exchanges beyond Euronext’s first generation are discussed.
Despite the efficiencies of computerized markets, virtually every stock market is accompanied by a parallel upstairs market, where larger traders employ the services of brokerage firms to locate counterparties and negotiate trade terms. Upstairs markets are based on relationships. As Venkataraman explained, “An electronic trading system lowers the fixed costs of trading. However, it does not allow for the informal exchange of information that is important for certain types of traders. Europe is ahead of the US, with its sophisticated automated exchanges such as Euronext and Germany’s Deutsche Bourse.” The research is a study of these markets of the future, highlighting trading problems that institutional investors may find when US exchanges go fully electronic. “They don’t get a sense of who they’re trading with (in electronic markets), how much more the other party is trading, and that is very important while trading blocks,” Venkataraman continued. “The institutions therefore seek other trading venues such as the upstairs market to lower the risk of exposing their orders.” Approximately 70% of block-size trade transactions are executed in the upstairs market in Paris.
Key Findings
The Paris Bourse provides an excellent setting to test the implications of upstairs intermediation models, because its electronic limit order market closely resembles the downstairs markets envisioned by theorists. Strong support for the following predictions were found:
(1) Upstairs brokers lower the risk of adverse selection by certifying block orders as uninformed.
(2) Upstairs brokers are able to tap into pools of hidden or unexpressed liquidity.
(3) Traders strategically choose across the upstairs and downstairs markets to minimize expected execution costs.
(4) Trades are more likely to be routed upstairs if they are large or are in stocks with less overall trading activity.
The second result is the most novel and arguably the most important. The upstairs broker completes transactions by searching for institutional investors who may be interested in the stock, but who have not as yet formally expressed their trading intentions. It is documented that executions costs of transactions completed by the upstairs broker average only 35% of what they would have paid if completed against limit orders in the centralized electronic exchange, suggesting that trading relationship and the informal exchange of information between upstairs brokers and institutional traders helps lower execution costs. One major challenge facing electronic markets is the lack of a comparable mechanism of certification and information exchange.
The Euronext market allows large transactions in some stocks to be executed outside the quotes. Such outside-the-quote transactions are not permitted in U.S. markets. For eligible stocks in Paris, market participants agree to outside-the-quote execution mainly for more difficult trades and at times when downstairs liquidity is lacking. These likely represent trades that probably could not have been otherwise completed, suggesting that market quality can be enhanced by allowing participants more flexibility to execute blocks at prices outside the quotes. These findings are particularly relevant to U.S. markets because quoted spreads and depths have decreased substantially in the wake of decimalization.
The upstairs market in Paris completes two-thirds of block trading volume, compared with 20% on the New York Stock Exchange (NYSE). A likely explanation is that the NYSE floor allows large traders to execute customized strategies through a floor broker, while avoiding the risks of order exposure. If orders submitted to electronic markets do not allow block initiators to limit order exposure and trade strategically, then order flow is likely to migrate to alternative trading venues such as the upstairs market. “The floor gives you a better sense of who you are trading with,” stated Venkataraman. “ When you’re a liquidity trader, you don’t want the system to be anonymous. If you’re an informed trader you like anonymity because you can hide in the order flow.”
To compete with broker-intermediated markets, the next generation of electronic trading systems need to include features that better meet the needs of large traders, particularly the lack of anonymity. To allow large investors to manage order exposure in an electronic exchange, a wider range of order types that include state contingent exposure and execution algorithms need to be made available. The NYSE’s recently introduced “Conversion and Parity” orders which are intended to be “smart” orders for large lots of stocks that are executed gradually through the day, contingent on market conditions, are a step in this direction.
The Role of the Specialist
The Paris Bourse, with its advanced electronic trading system, offers lessons for the NYSE which, in the wake of recent trading scandals, reconsiders the future role for a specialist with information privileges. Venkataraman and colleagues present the idea of the specialist with no information privileges in “The Value of the Non-Monopolist Specialist”, which was recently presented to the National Bureau of Economic Research (NBER) market microstructure group in August.
The study examines the improvements in liquidity when a specialist with no information advantage is introduced for French stocks. They find that the specialist posts liquidity when required and increases the frequency with which markets clear. Around the announcement of the introduction of a specialist, stocks experience an increase in firm value of nearly five percent. Overall, results suggest that the specialist can improve the terms of trade even in the absence of any information advantage by merely maintaining a regular market presence.
The specialist trades in circumstances when others do not or will not, and therefore takes on a risk which warrants compensation. The current debate centers around the model of compensation. The specialist at the Paris Bourse is compensated in cash and with investment banking business. In contrast, the NYSE specialist is compensated in the form of privileged information on order flow. In recent months, several U.S. institutions alleged that the NYSE trading abuses is an outcome of this compensation structure. The Paris model overcomes this criticism and presents an alternative for the NYSE to consider. “We’re not saying in the research that this is the optimal way to compensate the specialist,” Venkataraman explained. “Our results show however that there continues to be a role for the specialist in electronic markets. Investors value the presence of a specialist because they can get in and out of the stock with ease.”
Conclusion
Findings of the two papers suggest valuable lessons and evidence to consider for policymakers of electronic stock markets. The dynamics of automation with its results on trader behavior and market liquidity as well as the important role of the specialist both present significant gains in understanding stock markets of today and the future. |
