The Financial Economists Roundtable Statement
on
The Structure of Securities Markets
November 15, 2001
The structure of U.S. securities markets is both a topic for technical
specialists and a matter that affects in very practical ways all
investors in their pocketbooks. The effectiveness and efficiency of
securities markets in processing transactions in hundreds of millions of shares each day,
at an annual cost of billions of dollars, deserves more attention from
the financial press and the public than it customarily receives.
It should receive the attention of the new Chairman of the Securities
and Exchange Commission (SEC), Harvey Pitt, as he sets his agenda. He will
confront few issues of greater importance, but he will not be writing on
a clean slate. For the last quarter century, the SEC has been pursuing the
goal of a “nationalmarket system” (NMS), without a clear specification
of what that might be. Based on thediscussion at its annual meeting in July, 2001,
the Financial Economists Roundtable (FER) adopted the following statement
to help guide policy makers in their oversight of the nation’s securities markets.
The general legal concept of a NMS stems from the Securities Acts
Amendments of 1975, which in Section 11A directed the Commission to
facilitate the establishment of a NMS but did not undertake to define it.
That was left up to the SEC, having in mind the general goals of economically
efficient executionof orders in the best market, fair competition among brokers,
dealers and markets, and availability of information about quotations
and transactions. The SEC has been moving by tentative steps ever since.
Early actions abolished fixed commission rates, a boon to all investors,
and established a consolidated transaction reporting system. The SEC also
established a consolidated quotation system for all exchanges and market-makers,
but never resolved how “firm” a quotation was and how long it lasted. The SEC required
exchanges to be linked through the Intermarket Trading System (ITS), which
permitted market makers on one exchange to send orders to another exchange
to achieve better execution. However, “best execution” of customer orders
has proved elusive of precise definition. But the general objective of
trying to achieve greaterlinkage of equity trading markets remained.
As the SEC was seeking to link markets more tightly, new electronic
communications networks (ECNs) arose to serve particular investor
needs. Some, such as Instinet and POSIT, serve institutional investors.
Others, such as Island and Archipelago, serve individual investors.
The result has been that markets have become more fragmented, not less so.
One school of thought believes that all the problems of fragmentation would
be best solved by a single, fully integrated market, coordinated by a central
computer and mandated by the SEC. Customer market orders would be gathered
from all sources, matched and executed at the best price quoted anywhere. All
limit orders would be entered and displayed in a central limit order book. Trades (except possibly large block orders) would be executed by a computer, following rules
of strict price and time priority. The rationale for such a centralized system is that it guarantees the same prices for all investors, particularly for small retail
customers, who do not now have the same access as institutional investors
such as mutual funds or pension plans.
A second school of thought observes that fragmentation is a natural result
of competition and innovation. As a variety of markets with different
technologies and trading procedures compete for somewhat different groups
of customers with different needs, the result is competing market centers–registered
exchanges (such as NYSE and AMEX) with designated specialists; NASDAQ
with competing dealers; third market dealers in listed securities; alternative
trading systems (regulated as brokers) serving institutional investors or
providing on-line trading to individual investors. This second school of
thought views the multiplicity of markets as a sign of innovation and vibrant
competition, not as a problem that requires regulatory intervention.
Markets are,in fact, linked, albeit not completely, in various ways and
degrees – for example, by information andby private order routing systems
of brokers and markets.
How should these alternative views be evaluated from the standpoint of the
public interest? It is not an easy question to answer. The central choice is
between fully integrated markets, which will level the playing field among
investors immediately but would impede future changes, and fragmented markets,
which will permit greater competition and are likely to lead ultimately to
more efficient markets. A number of factors are relevant in analyzing
this choice. One is transparency of price and quote information. Securities
markets are in large part markets for information, serving to evaluate companies and their
management and to allocate capital to the most productive uses. A second
is the degree of customer access to each market. A third is the extent
of the broker’s duty to the customer to obtain best execution. And
whenever there are multiple goals or values, there will inevitably
be trade-offs among them.
In considering those questions, the FER believes there are some principles
and empirical constraints that should be kept in mind. To begin with,
securities trades are not homogeneous, standardized products but combinations
of a bundle of attributes. Trades differ in speed, market impact, and
commission or spread cost, as well as in the price per share paid or
received. All of these enter into “best execution.”
Different customers value those attributes differently. Informed
traders (those who believe they have an informational advantage)
value anonymity, while retail customers or index fund managers,
who are rebalancing portfolios,do not. Dealers incur less risk in transacting
with uninformed traders, and can charge lower transaction fees or
spreads. Informed traders are concerned with the market impact and
speed of execution of their orders, matters that may be of less
concern to other traders. Day traders may pay far greater attention
to speed of execution than to its cost. Mutual fund complexes can,
under SEC rules, trade among their fund portfolios, pricing off
market trades or quotes but without incurring execution costs.
Given these varying customer needs and preferences, different
trading systems are constantly being created to serve them.
Some systems (such as Instinet) cater to institutional investors
seeking to avoid incurring the full trading costs of brokers
and exchanges. Other systems automate procedures for handling small orders.
But these alternative trading systems depend to varying degrees
on prices derived from the primary markets, so information linkages
across markets are important and desirable. Recognizing this, the SEC
has mandated transparency–the immediate dissemination of trade prices
and quantities as well as the quoted prices and quantities at which
future transactions may take place. In a transparent market, investors
can make informed decisions about where an order should be sent.
The SEC has ventured beyond transparency in mandating an intermarket
trading system (ITS) for routing orders among the exchanges and the
NASDAQ system. The purpose of better intermarket linkages is to
enable orders to be routed to the market center where they will
receive best execution. The downside is that it has tended to
discourage new trading systems because of the difficulty of
integrating them into the ITS structure.
Institutional investors can and do monitor the execution of their orders,
and develop ways to bypass market centers that they view as not
performing satisfactorily. So obtaining best execution is a greater
concern for retail investors. Retail brokers have a legal obligation
to ascertain the best market and transact in that market to get the
customer as favorable a price as possible. The issue has been –
exactly what does “best execution” mean in operational terms?
Many market makers believe it requires only that they execute customer trades at the national-best-bid-and-offer (NBBO) price,
as shown on their computer screens, and not an obligation to
seek better offers. That in turn makes it important that the
NBBO include all limit orders, so that retail customers obtain the best prices.
(1) The multiplicity of market centers currently observable has been
criticized as “fragmented” and inefficient. Indeed it is fragmented,
but it is not inefficient. Fragmentation is another term for the
existence of competitors seeking particular customer clienteles, and
like competition in general, it promotes both innovation and better
prices for customers. In our view, such competition will produce
greater efficiency and lower transaction costs than would come from
a NMS in the sense of an SEC-mandated, single integrated market.
Furthermore, market participants have themselves developed links
among market centers.
(2) Transparency of the quotes and trades is a desirable attribute
of markets. Transparency has two important benefits. First, it enhances
competition because it allows consumers to compare prices. Second, it
helps achieve best execution because customers can monitor brokers
to determine whether they are sending orders to the best market.
Consequently, it may be desirable to display more information about
trading interest at, and outside, the NBBO.
(3) Linkages among the multiple market centers – for quote information,
order routing and settlement – are definitely needed. But the market
centers and vendors have incentives to develop them in accordance with
customer specifications, and they are evolving. The FER believes that
the precise form of linkages is best left to the market centers, in
their quest for trading volume and liquidity. Linkage should not take
the extreme form of requiring a central limit order book (CLOB). A
mandated CLOB would constrain competition and innovation.
(4) Detailed specification of the duty of best execution, spelling out
price priority or price improvement or trade-through requirements, is a
highly technical subject. The FER does not believe it is in a position
to conclude that a particular set of execution rules should be adopted,
given the different needs and priorities of different traders. In its
view, the SEC has followed the correct policy of enhancing disclosure,
most recently by new rules on disclosure of execution quality by each
market center. As the data on execution quality receive attention
from intermediary firms and academics, the issues and proper balance
may become clearer. But we would urge the SEC not to adopt at this
point a specific best execution standard.
(5) In the very broadest sense, these issues raise the question of what should be the role of government regulators in the structure of securities markets. We commend the SEC for having acted prudently in addressing such a sweeping question in a field in which technology is rapidly changing. It has avoided a rigid NMS, and has made useful moves toward enhanced transparency and linkages. But we have one note of caution: the 1975 Act placed the SEC, at its own request, in the awkward position of having to approve the rules of self-regulatory organizations (for example, the NYSE or NASD) in advance. That places on the SEC the onerous and impossible responsibility for foreseeing how trading markets should evolve. A natural reaction to such a burden, particularly for complicated and contested issues, is too often to delay and to consider everything at inordinate length. The consequence is a drag on innovation and, in a global market, the possibility that trades move offshore. It would be preferable for the SEC to exercise its oversight discretion ex post, by subsequently ordering repeal or modification of rules that prove abusive or anti-competitive.
revised on November 19, 2001 by gmoore2@luc.edu
http://www.luc.edu/orgs/finroundtable/members