October 2003
What Do Institutional Investors Want in a Securities Trading System?
On Wednesday, October 21, AEI hosted the third in a series of conferences on the structure of the securities markets. Participants in the conference discussed investors' preferences over various trading systems. One of the key issues was the controversy of specialists in the New York Stock Exchange (NYSE). John Colon of Greenwich Associates presented a study that polled several investors about the benefits and drawbacks of various trading systems. Kenneth Lehn of the University of Pittsburgh and Steve Sachs of Rydex Global Investors discussed the study and offered their own perspectives on the state of current trading systems.
John Colon
Greenwich Associates
In order to understand investors' priorities in conducting market transactions, we interviewed senior traders from 103 institutions earlier this year. Collectively, the traders managed over $2.5 trillion in assets, with an average of around $25 billion per institution. About half the institutions had assets in excess of $10 billion, and the rest owned between $1 and $10 billion worth of assets. The traders engaged in multiple investment styles, managing many portfolios, some of which contained large cap stocks and others small and mid cap stocks. Two-thirds of trades within these institutions occurred on exchange traded stocks. As a general tendency, those surveyed were active traders, unlike index fund managers.
Study participants agreed that specialists are more of a hindrance than a help in terms of the trading process. The top-line findings of the study are that investors consider low market impact, anonymity, price improvement, and certainty of execution in trading to be major requirements when executing orders. In the comparison across trading systems, there was a widespread belief that Electronic Computer Networks (ECNs) hold a substantial advantage in the ability to deliver anonymity and low market impact over exchange or upstairs brokers. Exchanges and brokers hold a modest advantage in certainty of execution of transactions in a timely fashion. Over three-quarters of traders considered proprietary trading by a specialist to be a conflict of interest. Two-thirds of the traders believe specialist NASDAQ market makers do not add value in trading large liquid stocks. Two central questions surrounding specialists are whether specialists are necessary for large, liquid stocks and whether specialists have the capital and willpower to maintain order in plummeting markets.
Nearly half of the respondents stated that they would prefer to trade more volume on the exchange floor, the main reason being a distrust of specialists. Those who stated a preference for not leaving volume off the exchange floor cited centralization and human touch as the benefits of the specialist system.
To compare trading systems, we asked traders to rank different venues. There is a slight leaning towards ECNs as being the most effective venue for listed stocks, but it was not dramatically preferred to upstairs brokers. Exchanges were slightly less preferred than these two venues. Those surveyed agreed by a three-to-one margin that ECNs were more likely to provide low market impact and quick execution than exchanges and upstairs brokers. Overall, ECNs delivered on more important objectives than either of the alternative venues. Traders did not, however, dismiss upstairs brokers and exchanges entirely.
Predominantly, traders cited human touch as the biggest advantage of a centralized trading system. It is faster and easier for buyers and sellers to meet on the floor and to obtain better prices for securities. Others cited less market fragmentation as a big benefit of centralized exchanges. These benefits did not seem to outweigh the inefficiencies of the system. One of the greatest findings of the survey was that 47 percent of traders stated that they would like to move more volume away from the floor-a powerful indictment against the current system. These respondents cited three main reasons: specialists are well-funded competitors in the centralized exchange, there is a lack of visible liquidity, and there are too many intermediaries without institutional investors' best interests at heart.
When asked where they were executing their volume in exchange-listed stocks, traders said about 80 percent on the exchange, 12-13 percent off the exchange, and about 9 percent through ECNs or alternative trading systems. Roughly half the people who wanted to move orders away from the floor nonetheless execute their orders on the exchange floor. Why? Most of the reasons pertain to the issue of centralization. A central location makes it easier for buyers and sellers to meet. Interaction among traders leads to greater trading activity. The human touch is an integral component of trading.
When asked about the overall quality of execution between NYSE and NASDAQ, 39 percent said executions were superior in NYSE; 31 percent said executions were significantly inferior; and 30 percent said it was approximately the same. The changes traders would like to see in the securities market include having an electronic market with sufficient liquidity compete with the NYSE, implementing rules that prevent specialists from competing with customer orders, and creating a system that simultaneously exposed orders on the floor into electronic venues.
The results of this study do not necessarily suggest a definitive answer. It appears that the best execution is in the eyes of the beholder. In general, however, most traders are looking for greater informational flows in the securities market. They are looking to gain an advantage on a fair and evenhanded basis.
Kenneth Lehn
University of Pittsburgh
The three major points of the Greenwich Associates survey are: ECNs trump the NYSE on anonymity, speed of execution and market impact; specialists add no value in liquid stocks, such as IBM; and proprietary trading by specialists does constitute a conflict of interest. The study is consistent with academic evidence that execution costs are lower for ECNs and that ECNs contribute to price discovery.
I have a couple questions regarding the Greenwich study. The first is how representative the survey sample is of all institutional investors. It certainly would not be representative of retail investors, who constitute a large portion of the market. There could be a selection bias because the study was commissioned by Instinet. The second question is whether the responses are different for a sub-sample of institutional investors. Do large institutions answer differently than small institutions? Do actively-managed funds look for different things than indexed funds? And finally, are responses different for large-cap versus small-cap funds?
On the issue of specialists, it is very questionable as to whether the role still exists. For less liquid stocks, the intellectual defense of the specialist system is much stronger. Consider four pieces of evidence that suggest the specialist problem is exaggerated: a large body of evidence illustrates that the NYSE does a good job in terms of execution costs and price discovery; the bulk of all trading for exchange listed stocks still occurs on the floor; centralized exchanges have an advantage in price discovery; and NYSE seat prices have increased with the growth of ECNs. This does not imply that the structure of the NYSE is perfect by any means.
The main challenge is that the NYSE needs to adapt more effectively to changes in technology and in customer demand. It is still a bureaucratic organization, much of it arising from tight SEC regulation. The SEC should relax regulations to allow the NYSE to become a nimble organization that can adapt effectively to changes in the securities markets.
One option for the NYSE is to convert to a for-profit status whether there are actually shares rather than seats and a market in which those shares can transact. Academic evidence suggest that kind of governance structure leads to greater monitoring by shareholders and, in most cases, a more effectively run organization. The NYSE should also have a smaller board, as John Reed advocated. There are currently twenty-seven members as compared to eight directors and a much smaller board at Instinet. Lastly, the NYSE should revisit the issue of decimalization. There is widespread belief that decimalization has done more harm than good, and I believe that should be part of the policy debate.
Steve Sachs
Rydex Global Investors
What do institutional investors want in a trading system? A recent article in the Journal of Portfolio Management stated that improving market quality is the overriding objective for a market center. The important question is how to implement this objective. There are numerous changes that need to be made to the current trading structure, but more importantly, these changes need to be implemented in a way that allows institutions to manage money successfully.
Low market impact, without a doubt, is the most important requirement when executing orders in listed stocks. With ECNs, there is less information leakage in the marketplace and hence lower market impact. However, the use of ECNs leads to market fragmentation, which then raises a host of other problems, including greater opportunity costs. On the floor, there is less fragmentation, but there are greater costs.
On the issue of specialists, I would agree with the assessment that NYSE specialists and NASDAQ market makers add value in large liquid stocks. ECNs have proven that they can maintain trading order with large liquid stocks. I would also agree with the statement that proprietary trading by the specialists constitutes a conflict of interest. The problem, however, is that is that there is no incentive for market makers and specialists to maintain an orderly market if there are no profits. Is there a need for specialists for a stock like IBM? No. For small cap stocks, however, specialists maintain low spreads and help execute fast trades.
In order to reconcile the difference between the need for specialists to be profitable and to maintain a fair and orderly market with the desire for institutions to have an anonymous electronic environment or the natural buy-and-sell matching arena, the current structure of the securities market should be somewhere between NASDAQ and the NYSE. In other words, there should be anonymous electronic access to an open order book that has equal standing at the trading post. Specialists could thus maintain markets and interfere when necessary, while the conflict of interest is minimized.
In regards to the decimal pricing system, I believe the policy should be reversed. I would be a proponent of nickel increments in stock that trade above $10-a policy that would encourage liquidity and depth of liquidity at particular price points. If we can find the right price, the price we are willing to pay for, and the depth at that price, then we can limit the impact of orders. However, the reverse in this policy would detrimental to the interests of individual investors at to the benefit of institutional investors.
The last point I want to emphasize is implementation. Over the past eighteen months, the NYSE launched both the institutional express system and the liquidity quote display system-systems that work well in theory but did not have the right conditions to prove highly successful. NYSE should strive to implement changes that address the fundamental purpose of securities markets - to provide an efficient market for corporations to raise cash.
AEI Intern Kartik Arekapudi prepared this summary.
