The specialist trading system used by most securities exchanges in the United States recently has been under heavy attack and does not appear to be fielding a very strong defense. Little attention appears to be given to the fact that the specialist system would not have survived so long unless it provided a useful service. The specialist trading system is of particular importance to exchange-traded funds (ETFs), and this article will explain the special role that the specialist plays in the functioning of these securities.
While ETFs trade on security exchanges and electro n i c trading systems in a manner that causes them to appear to be the same as most listed securities, the underlying legal structure governing them is the Securities and Exchange Commission Act of 1940. This is the same legal structure underlying mutual funds, and most of the regulations applying to mutual funds also apply to ETFs.
Fortunately, many of the problems mutual fund managers currently face-such as having accepted order exe c utions after the close of the market, permitting rapid in-andout trading or selective disclosure of the fund's holdings- are not matters of concern for ETFs. The dual-level structure of the ETF separates trading in the market fro m the creation and redemption of the securities, making such actions either difficult or without benefit.
However, regulation of trading in ETFs under the '40 Act means that the usual method of introducing new securities to listing on an exchange, preceded by an underwriting and p re-listed marketing, is not available to them. In the design of the first ETF, the Standard and Poor's Depository Receipt ('SPDR'), the need to launch it without an underwriting and marketing program was a matter of concern. Fortunately this was not the first time that the New Product Development Group at the AMEX had to deal with this type of problem. A large part of the work then being done by the g roup was the design and launching of new stock index options. The AMEX utilizes a specialist trading system to support such options, and a basic criterion in deciding which specialist group will be assigned a new index option is the amount of market-making support and marketing effort the group agrees to provide. Given this experience, it was a simple step to decide to base the selection of a specialist for the SPDR on the same criteria.
Index option start-ups differed from that of the SPDR in that an inventory in support of start up of trading had to be provided for the SPDR. An option starts in a manner similar to that of a futures contract, in that a sale automatically creates a new security. This is not true for an ETF, which must be created before it can be traded. Technically, initial trading in the SPDR could have started without an inventory. The product design is such that authorized participants ('APs'), those organizations exclusively authorized to create new ETFs, could make sales and cover them by purchasing the stocks re q u i red to create the sold ETFs, making delivery to the fund on the same day the sales were executed.
This is possible because the ETF creation system includes the special clearing program, which the NSCC designed and built especially for the SPDR. It has been used there after for the creation of all subsequent ETFs in the U.S. The system enables an ETF to receive confirmation that the portfolio of stocks required to be delivered to the fund for the creation of a portfolio of ETFs have in fact been purchased by or is already held in the AP's account at the Depository Clearing Corporation ('DTC'), ready to be transferred to the fund.
The procedure provides that at midnight of T+1 (day after the trade) NSCC becomes the guarantor of the delivery of the required stocks to the fund. This enables the fund to order transfer of the newly created ETFs to the AP's D TC account on T+3 (three days after the trade), when final settlement of any ETFs sold occurs.
While this procedure could have been used to make the sales of ETFs without creating an initial inventory, neither the potential buyers nor the SEC were comfortable with initial orders being executed for fund shares in the absence of an inventory. Accordingly, it became necessary to have an inventory in place at DTC prior to the beginning of trading in any newly-listed ETF. In most cases it is the specialist firm that provides this inventory, and the amount that specialist groups commit to provide is an important consideration in the determination of which specialist group is assigned the security.
For the SPDR, since it was the first ETF to be launched, the quality of the market-making and the amount of marketing effort specialist groups were willing to provide was a very important consideration. Spear, Leeds and Kellogg ('SLK') received the appointment and was sufficiently successful in marketing, in conjunction with the efforts of the AMEX staff, that assets under management ('AUM') of the SPDR reached the $350 million level in the first year of trading. This was important to the AMEX since it was the amount needed to meet the minimum earnings guarantee required by State Street Bank before they would agree to take on the job of fund manager. To their credit, it must be said that of all the management groups we approached requesting management proposals for the SPDR, State Street Bank was the only organization willing to undertake the management. Undoubtedly, the very different type of management re q u i red for the SPDR in comparison with standard mutual funds was the reason for the lack of response from others.
Those interested in doing away with the specialist trading system may point to the success of ETFs in foreign markets without that system. In aggregate, certainly there is a substantial amount of trading and AUM in ETFs outside the U.S. Morgan Stanley reports that as of December 31, 2003, the total AUM worldwide was $211.1 billion with the U.S. amount at $150.7 billion. The next largest market was Japan with $27.6 billion of AUM, followed by Europe with a total of $19.5 billion. This is somewhat surprising, since t h e re are currently more ETF-type instruments listed and trading in Europe than in the U.S. (158 vs. 117). Only those ETFs based on the CAC, the DAX and the Euro Stoxx 50 have reached AUMs in the $1.5 billion range, while few of the sector-based ETFs have any significant AUM.
In Japan the intense competition among the brokerage houses authorized to introduce ETFs has resulted in the brokers themselves creating substantial amounts of AUM. Even with this, only those ETFs based on Japanese major market stock indexes, the Topix and the Nikkei 225 have attracted substantial AUM. Among the Japanese sector ETFs the largest, banking, only has approximately $30 million dollars of AUM.
While the Tokyo Stock Exchange does not utilize the specialist system, the government's policy of limiting the number of brokerage houses permitted to introduce and manage ETFs has provided a similar result to that of the U.S. specialist system in terms of providing support for the securities. This is also true in Korea and Taiwan, where only one or two organizations have been given permission to introduce ETFs. These restrictions have had the effect of creating an environment similar to that of a specialist system where specific firms have the responsibility to provide liquidity to the funds.
In Hong Kong, the most successful ETF thus far is the Hong Kong Tracker, utilized by the Hong Kong government to aid in returning the government's inventory of stocks to the market. To accomplish this the government had to offer special discounts to buyers below the intrinsic value of the ETF stock portfolio. In its present form, the tracker is not operating as a true ETF but its design is such that it will operate as such once the government is no longer the sole c reator of AUM.
In the smaller foreign markets securities exchanges have been eager to introduce ETFs, having noted the exchange trading volume in the U.S. and other parts of the world. In some cases, their operational trading system will not support an ETF-type of security. The Shanghai stock exchange, for example, is interested in introducing ETF-type securities. While it has high-speed computers, which permit it to settle trades on T+1, it has no equivalent of the U.S. NSCC clearing and settlement system. The exchange solves this problem by requiring that a security be held in a seller's account before it can be sold. This solves the fail problem but makes it impossible to execute many of the different types of transactions, such as arbitrage/creation combinations or short sales, which have played a large part in the ETF's success.
In summary, the successes in introducing ETFs to markets where trading is purely electronic has been largely limited had the effect of creating an environment similar to that of a specialist system where specific firms have the responsibility limited to those based on major market indexes with supporting futures and options markets. This is also true in the U.S. where there is a great deal of computerized ETF trading, but largely limited to the those well-established ETFs based on the major market indexes.
My initial purpose in designing the first ETF, the SPDR, was to provide the AMEX with badly-needed trading volume. However after launching the second ETF, the Mid Cap SPDR (MDY), things changed and I consulted with Morg a n Stanley Capital International (MSCI) in the design of the foreign-based ETFs they called World Equity Benchmark S h a res (WEBS), now known as iShares. Thereafter I acted as consultant for Barclays Global Investors in the development and launching of the iShare program, still the only effective marketing organization introducing ETFs to retail investors. In addition, I have been involved in the introduction and growth of these securities in varying degrees in many other parts of the world.
Based on this experience I believe that while ETFs could exist without the specialist trading system, the restrictions imposed by the SEC '40 Act would make it much more difficult to launch new ETFs in the U.S. without that trading system. In foreign markets, I believe that basing trading only on electronic trading markets has largely been the reason that the only successful ETFs are those based on the wellestablished, broad-market indexes.
Those making a case against the specialist system in the U.S. base it primarily on their claim that there is no need for specialists in providing markets in the larger listed securities. This is true a majority of the time, although specialist organizations are often needed in putting together large trades even in the highly-traded securities. However, this overlooks the fact that U.S. specialist organizations are required to support trading in smaller listed securities as well as in the large established ones. Many of the smaller stocks would be difficult to buy or sell at reasonable prices without the support of the assigned specialist's market making. A look at the markets in the smaller listed stocks on Nasdaq would demonstrate the type of markets that would then exist.
Further, in judging the specialist system, we must re cognize that specialist trading organizations are businesses intended to make profits. If they were restricted to only market making in the smaller listed securities it would be difficult for them to justify maintaining the staff, equipment and capital required to run such operations.
During the nineteen years that I was at the AMEX, I worked closely with the specialist groups on the trading floor. They were essential to the work that my new pro ducts group was doing in designing and launching new securities. If the specialist trading system now suffers real damage, or in the extreme is done away with in its entirety, I believe that the U.S. securities markets, and particularly ETFs, will suffer for it.