[Senate Hearing 112-748]
[From the U.S. Government Publishing Office]
S. Hrg. 112-748
COMPUTERIZED TRADING: WHAT SHOULD THE RULES OF THE ROAD BE?
=======================================================================
HEARING
before the
SUBCOMMITTEE ON
SECURITIES, INSURANCE, AND INVESTMENT
of the
COMMITTEE ON
BANKING,HOUSING,AND URBAN AFFAIRS
UNITED STATES SENATE
ONE HUNDRED TWELFTH CONGRESS
SECOND SESSION
ON
EXAMINING COMPUTERIZED TRADING
__________
SEPTEMBER 20, 2012
__________
Printed for the use of the Committee on Banking, Housing, and Urban
Affairs
Available at: http: //www.fdsys.gov /
U.S. GOVERNMENT PRINTING OFFICE
80-168 WASHINGTON : 2009
-----------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Printing
Office Internet: bookstore.gpo.gov Phone: toll free (866) 512-1800; DC
area (202) 512-1800 Fax: (202) 512-2104 Mail: Stop IDCC, Washington, DC
20402-0001
COMMITTEE ON BANKING, HOUSING, AND URBAN AFFAIRS
TIM JOHNSON, South Dakota, Chairman
JACK REED, Rhode Island RICHARD C. SHELBY, Alabama
CHARLES E. SCHUMER, New York MIKE CRAPO, Idaho
ROBERT MENENDEZ, New Jersey BOB CORKER, Tennessee
DANIEL K. AKAKA, Hawaii JIM DeMINT, South Carolina
SHERROD BROWN, Ohio DAVID VITTER, Louisiana
JON TESTER, Montana MIKE JOHANNS, Nebraska
HERB KOHL, Wisconsin PATRICK J. TOOMEY, Pennsylvania
MARK R. WARNER, Virginia MARK KIRK, Illinois
JEFF MERKLEY, Oregon JERRY MORAN, Kansas
MICHAEL F. BENNET, Colorado ROGER F. WICKER, Mississippi
KAY HAGAN, North Carolina
Dwight Fettig, Staff Director
William D. Duhnke, Republican Staff Director
Dawn Ratliff, Chief Clerk
Riker Vermilye, Hearing Clerk
Shelvin Simmons, IT Director
Jim Crowell, Editor
______
Subcommittee on Securities, Insurance, and Investment
JACK REED, Rhode Island, Chairman
MIKE CRAPO, Idaho, Ranking Republican Member
CHARLES E. SCHUMER, New York PATRICK J. TOOMEY, Pennsylvania
ROBERT MENENDEZ, New Jersey MARK KIRK, Illinois
DANIEL K. AKAKA, Hawaii BOB CORKER, Tennessee
HERB KOHL, Wisconsin JIM DeMINT, South Carolina
MARK R. WARNER, Virginia DAVID VITTER, Louisiana
JEFF MERKLEY, Oregon JERRY MORAN, Kansas
MICHAEL F. BENNET, Colorado ROGER F. WICKER, Mississippi
KAY HAGAN, North Carolina
TIM JOHNSON, South Dakota
Kara Stein, Subcommittee Staff Director
Gregg Richard, Republican Subcommittee Staff Director
(ii)
C O N T E N T S
----------
THURSDAY, SEPTEMBER 20, 2012
Page
Opening statement of Chairman Reed............................... 1
Opening statements, comments, or prepared statements of:
Senator Crapo................................................ 2
WITNESSES
David Lauer, Market Structure and High-Frequency Trading
Consultant, Better Markets..................................... 4
Prepared statement........................................... 28
Responses to written questions of:
Senator Reed............................................. 72
Senator Hagan............................................ 73
Andrew Brooks, Head of U.S. Equity Trading, T. Rowe Price........ 5
Prepared statement........................................... 40
Chris Concannon, Partner and Executive Vice President, Virtu
Financial, LLC................................................. 8
Prepared statement........................................... 46
Responses to written questions of:
Senator Reed............................................. 76
Senator Hagan............................................ 77
Larry Tabb, Founder and Chief Executive Officer, TABB Group...... 10
Prepared statement........................................... 49
Additional Material Supplied for the Record
Statement of Senator Carl Levin.................................. 80
Statement of Micah Hauptman, Financial Campaign Coordinator,
Public Citizen's Congress Watch Division....................... 83
Statement of Cameron Smith, President, Quantlab Financial, and
Richard Gorelick, CEO, RGM Advisors............................ 94
Summary and Excerpts from ``High Frequency Trading and Price
Discovery'' by Terry Hendershott and Ryan Riordan.............. 108
(iii)
COMPUTERIZED TRADING: WHAT SHOULD THE RULES OF THE ROAD BE?
----------
THURSDAY, SEPTEMBER 20, 2012
U.S. Senate,
Subcommittee on Securities, Insurance, and Investment,
Committee on Banking, Housing, and Urban Affairs,
Washington, DC.
The Subcommittee convened at 10:03 a.m., in room 538,
Dirksen Senate Office Building, Hon. Jack Reed, Chairman of the
Subcommittee, presiding.
OPENING STATEMENT OF CHAIRMAN JACK REED
Chairman Reed. Let me call the hearing to order. I want to
thank all of you for joining us today, especially our
witnesses, and thank my colleague, Senator Crapo.
As everyone in the room knows, computer trading has changed
markets in fundamental ways, not the least of which is the
speed at which trading now occurs. The benefits of automated
trading for retail investors include access to real-time market
data and being able to execute orders within a fraction of a
second. The explicit costs of trading have also decreased. In
one study that focused on pre- and post-decimal trading in the
New York Stock Exchange listed stocks, average effective
spreads decreased significantly for both small and large
trades. The market is also more efficient, processing large
volumes of data and more accurately setting market prices based
on sometimes minute changes and market additions.
However, since the May 6, 2010, Flash Crash, there have
been a series of high-profile computer trading errors that
highlight some of the dangers and costs of fast-moving
computer-driven trading. In March, a computer glitch scuttled
the initial public offering of one of the Nation's largest
electronic exchanges, the BATS Global Markets, Inc. In May,
computer problems at the NASDAQ Stock Market plagued the
initial public offering of Facebook stock. And last month, the
Knight Capital Group, a brokerage firm at the center of the
Nation's stock market for almost a decade, nearly collapsed
after it ran up more than $400 million of losses in minutes
because of errant technology.
Taken together, these failures in electronic trading appear
to be affecting investor confidence in the U.S. market
structure. Most of us consider American capital markets to be
the best in the world. Our markets are known for their strength
and resiliency, their openness and transparency, and their
fairness to all market participants. But our marketplace has
been evolving very quickly and it is not clear that our rules
have kept up.
In particular, we need to focus on whether markets have the
ability to avoid systematic failure triggered by a computer
problem. Are our markets still fair? Is everyone playing by the
same set of rules? And perhaps most importantly, are our
markets still focused on long-term capital formation and the
creation of jobs?
Following the Knight Capital incident in early August, the
SEC announced it would convene a roundtable of trading
technologists in an effort to determine if brokers and
exchanges are in control of their trading systems. SEC
Chairwoman Mary Shapiro said she had asked staff to, quote,
``accelerate ongoing efforts to propose a rule that would
require trading venues to maintain the capacity and integrity
of their systems,'' and this roundtable is scheduled for
October 2, 2012.
Clearly, recent market events caused by technology-related
issues have sharpened the debate on market infrastructure and
the need to limit the broader impact of computer trading
errors. However, in light of the enormous growth of high-
frequency and algorithmic trading, there is a growing consensus
that the entire regulatory scheme surrounding high-frequency
trading firms and their algorithms should be assessed.
Do current regulations reflect the impact of high-frequency
algorithms on trading? Or can computer algorithms be programmed
to operate properly in stressed market conditions? What
challenges face firms when they are testing and implementing
new systems?
In addition, just last week, on September 14, the SEC
settled first of its kind charges against the New York Stock
Exchange for compliance failures that gave certain customers an
improper head start on trading information. Computer issues,
which include both disparities in the design of the New York
Stock Exchange hardware systems and software problems, resulted
in some customers receiving stock pricing trading data several
seconds ahead of the general public. The SEC order marks the
first ever SEC financial penalty against an exchange.
The order and fine involving the New York Stock Exchange
highlight another set of issues I hope we can discuss at
today's hearing. In effect, should the exchanges have control
over the collection, aggregation, and distribution of market
data? Should data be available to everyone at the same time, or
should enriched data be provided more quickly to those who are
willing to pay for it?
We look forward to hearing your testimony on all of these
topics. The capital markets are a public good, much like a
highway. We need to have clear rules about the speed limits,
who can use the HOV lanes. With our rapidly evolving capital
market structure, we need to make it clear what the rules of
the road are.
Now, let me turn to my colleague, Senator Crapo, for his
comments. Senator.
STATEMENT OF SENATOR MIKE CRAPO
Senator Crapo. Thank you, Mr. Chairman.
More than half of all U.S. households, 57 million,
according to one study, participate in our markets through
either stocks or mutual funds, and the health and
competitiveness of these markets have an immediate and direct
effect on our broader economy as well as on the wealth and
prosperity of the American people.
Over the last several decades, new forms of competition,
technology, global growth in trading, and broader investor
participation have integrated and interconnected the world's
capital markets and the financial services industry as never
before. According to a 2010 study by James Angle of Georgetown,
Larry Harris of the University of Southern California, and
Chester Spatt of Carnegie Mellon, technological innovations
have led to major improvements in market quality. Their study
found that execution speeds have increased, making it much
easier for retail investors to monitor execution. In addition,
retail commissions have significantly dropped and bid/ask
spreads have narrowed.
However, as Chairman Reed outlined, there have been too
many technological failures in our market infrastructures since
the Flash Crash of 2010.
I thank Chairman Reed for holding this important hearing on
a topic that has clearly captured the attention of academics,
practitioners, and regulators.
Yesterday, Georgetown University hosted a conference on
financial market quality and the question of what is the
empirical evidence on the role of alternative trading systems,
algorithmic trading, high-frequency trading, dark pools, and
new trading technology on market quality. In October, as has
been indicated, the SEC will be holding a technological
roundtable to discuss how to minimize trading errors and market
malfunctions as well as how to respond to them in real time
when they do occur.
As we look at ways to help fortify our markets, especially
during times of market stress, it is important that we examine
all of the relevant empirical evidence in order to make
informed policy decisions. I hope the witnesses at today's
hearing can provide some of that much needed evidence and I
look forward to hearing their testimony.
Thank you, Mr. Chairman.
Chairman Reed. Thank you very much, Senator Crapo.
Let me now introduce our panel. I want to thank the panel
for joining us today. We are extraordinarily lucky to have such
qualified and insightful witnesses.
Our first witness is Mr. David Lauer. Mr. Lauer is a Market
Structure and High-Frequency Trading Consultant at Better
Markets. He has helped financial service firms deploy high-
performance trading infrastructure. He has also worked as a
quantitative analyst and high-frequency trader himself. Thank
you, Mr. Lauer.
Mr. Larry Tabb is the founder and CEO of TABB Group. The
financial markets research and strategic advisory firm focuses
exclusively on capital markets. Founded in 2003, TABB Group
analyzes financial market issues. Thank you, Mr. Tabb, for
being here.
Mr. Chris Concannon is currently a partner in Virtu
Financial, where he serves as an Executive Vice President and
is the Chief Compliance Officer at Virtu Financial BD. Prior to
joining Virtu in 2009, Mr. Concannon spent 6 years as Executive
Vice President of the NASDAQ OMX Group. Thank you, Chris.
Mr. Andrew Brooks is Vice President of T. Rowe Price Group
and T. Rowe Price Associates. He has served as head of U.S.
equity trading for the firm since 1992. Thank you very much,
Andy, for joining us today.
We will begin with Mr. Lauer and then just go right down
the panel. Please try to limit your testimony to 5 minutes.
Your complete written testimony will be made part of the
record, so feel free--in fact, I encourage you to summarize.
Mr. Lauer.
STATEMENT OF DAVID LAUER, MARKET STRUCTURE AND HIGH-FREQUENCY
TRADING CONSULTANT, BETTER MARKETS
Mr. Lauer. Good morning, Chairman Reed, Ranking Member
Crapo, and Members of the Subcommittee. Thank you for the
invitation to Better Markets to testify today, and thank you
very much for holding this hearing on such a critical issue to
our financial markets and our economy.
Better Markets is a nonprofit, nonpartisan organization
that promotes the public interest in financial markets. I will
not take the time to go over all that we do today, but you can
find it at bettermarkets.com. My name is David Lauer. As you
indicated, I am a Market Structure and High-Frequency Trading
Consultant to Better Markets. I also consult for IEX Group, a
new private company that is developing an investor-focused
equity market.
Prior to consulting, I worked as a high-frequency
researcher and trader at two of the largest HFT firms in the
industry. And before that, I worked at a technology vendor that
provided high-performance hardware to many high-speed trading
desks on Wall Street.
I would like to start with a story and tell you about my
experience during the Flash Crash in May of 2010. The Flash
Crash, as you know, caused the market to drop by a trillion
dollars, nearly a trillion dollars, and then bounce back
inexplicably within 20 minutes. I was on the trading floor that
day as the market crashed and I witnessed something, quite
frankly, unthinkable. The market simply disappeared. It was
gone. We had no idea what was happening and we had no idea what
would happen next.
Our firm, like other HFT firms, immediately withdrew our
liquidity from the market because we had no idea what was
happening. We did not trust the information from our data
feeds. And we had no obligation to remain in the market. More
than half of the liquidity in the stock market was pulled in a
matter of seconds and that dramatically worsened an already
unstable situation. Anybody who seeks to minimize the role that
HFT played in the Flash Crash either was not on a trading floor
that day or is incentivized to maintain the status quo.
Remarkably, since the Flash Crash, there have been no
substantial changes in market structure. The U.S. equity
markets are in dire straits. We are truly in a crisis right
now. The past decade of technology revolution on Wall Street
has been marked by two primary trends, extreme marketplace
fragmentation and the rapid growth of HFT as the primary
supplier of liquidity. Complexity is the hallmark of this
market, whether from the fragmentation of 13 exchanges and over
50 dark pools or the interaction between algorithms listening
to high-frequency market data. While complex systems can often
provide elegant solutions to intractable real world problems,
they can also spin out of control at any moment and in
completely unexpected ways.
There is no doubt that electronic trading has tremendous
value to offer, enhancing the smooth functioning of the stock
market and increasing competition, thus driving down the cost
of trading. What we must be concerned with is whether the
pendulum has swung too far and whether the nearly unregulated
activity of 50 to 70 percent of the stock market should be
allowed to continue unabated.
A quick look at today's markets reveal that stock market
volatility has increased. Spreads have increased. Catastrophic
event frequency has increased. IPOs have dramatically dropped.
And retail investors are fleeing the stock market.
As my written testimony demonstrates, the micro structure
of the equity market has been altered by HFT and extreme market
fragmentation, resulting in an excessively fragile market. This
fragility is apparent in the impact that a single firm or even
a single computer can have on the market at large, whether
accidental or nefarious and predatory.
I believe the market is so fragile because of structural
inefficiencies, including the proliferation of the maker-taker
business model and pay for order flow deals, the disappearance
of affirmative market making obligations, the fragmentation of
equity markets into lit and unlit venues with little regulation
of the unlit venues, and the rubber stamp approval of exotic
order types without proper study or justification.
So what are we to do? Here are my suggestions, just a few
of them, for rules of the road. I believe we need to unify
trading rules regardless of lit or unlit venue, and the bar
must be raised on off-exchange execution standards. In order to
receive a rebate from any venue, market makers must be
registered and subject to affirmative market making
obligations. A unique identifier should be assigned to every
supervisory individual and attached to every quote. This would
provide a contact person in emergencies and remove the cloak of
anonymity that has allowed for manipulative behavior. We must
eliminate pay for order flow practices. The SEC should pilot a
50-millisecond minimum quote time. And strong market technology
standards are needed and regular audits to ensure they are
being followed.
In addition, I would like regulators to try a novel
approach to surveillance and enforcement inspired by the
Internet and open access and standards. Regulators could open
up access to market data and incentivize independent
programmers with prizes similar to how the X Prize functions,
or percentage of fines like their whistleblower program, to
help design better surveillance and enforcement tools for
regulators. This can be part of advancing the SEC's technology
capabilities with a substantial investment in technology and
personnel to bring the SEC into the 21st century.
Thank you very much, and I am happy to answer any
questions.
Chairman Reed. Thank you very much.
Mr. Brooks, please.
STATEMENT OF ANDREW BROOKS, HEAD OF U.S. EQUITY TRADING, T.
ROWE PRICE
Mr. Brooks. Good morning, Chairman Reed, Ranking Member
Crapo, and distinguished Members of the Senate Subcommittee on
Securities, Insurance, and Investment. Thank you for the
opportunity to testify today on behalf of T. Rowe Price
regarding the effects of recent significant changes in trading
technology and practices on market stability.
My name is Andy Brooks. I am a Vice President and Head of
U.S. Equity Trading at T. Rowe Price Associates, Incorporated.
I joined the firm in 1980 as an equity trader. I assumed my
current role in 1992. This is my 33rd year on the T. Rowe Price
trading desk.
T. Rowe Price is celebrating its 75th year anniversary of
advising clients. We are a Baltimore-based global advisor with
over $540 billion in assets under management as of June 30,
2012, and more than three million client accounts. We serve
both institutional and individual investors. We welcome the
opportunity for discussion regarding the industry and market
practices.
Our firm is particularly focused on the interests of long-
term investors. We appreciate the role other types of investors
can have in creating a dynamic marketplace. However, as we talk
with our clients, there is a growing distrust of the casino-
like environment that the marketplace has developed over the
past decade. We worry that the erosion of investor confidence
can undermine our capital markets, which are so important to
the economy, job growth, and global competitiveness.
Reaffirming a strongly rooted commitment to fairness and
stability in the market's infrastructure is critically
important.
Over the past two decades, the markets have benefited from
innovation and from new technology and competition. Generally,
markets open on time, close on time, and trades settle.
However, there are problems below the surface. Here are some of
the things we find concerning.
Order routing practices. We question the nature of various
order routing practices. The maker-taker model, payment for
order flow, and internalization of orders all seem to present a
challenge to order routing protocols. Are order routing
practices and incentives--are they an impediment to the over-
reaching requirement to seek best execution on all trades?
Colocation and market data arbitrage. We believe that the
widespread use of colocation creates an uneven playing field
that favors those who can and will pay for it. We question
whether this has produced a market that values speed over fair
access. In no other regulated industry is one party allowed a
head start in exchange for payment. Our understanding is the
current colocation practices allow for a market data arbitrage
where some investors get quotations and trade data faster than
others. This advantage is traded upon, causing some
participants to believe they are victims of front running, or
at the least, disadvantaged.
Speed and impact on market integrity. Our sense is the
almost myopic quest for speed has threatened the very market
itself. It also seems to many that high-frequency trading
strategies are designed to initiate an order to simply gauge
the market's reaction, then quickly react and transact faster
than other investors can. This seems inherently wrong. Our
understanding is that the continued push for speed is not
producing any marginal benefit to investors and, in fact, may
be detrimental. This pursuit of speed as a priority is in
direct conflict with the pursuit of market integrity as a
priority.
Inaccessible quotes and high cancellation rates. The growth
of HFT has led to increased volume. However, whether the
corresponding volume is good or bad deserves analysis. Volume
does not necessarily mean liquidity for large institutional
investors. When you combine high HFT volumes and even higher
cancellation rates, these forces can combine to undermine
market integrity and cause deterioration in the quality and
depth of the order book. We feel that this volume is transitory
and misleading.
Challenges to the national market system. We believe the
original construct of Regulation NMS was laudable and designed
to encourage competition. However, we do not believe this
regulation contemplated today's highly fragmented marketplace,
where we have 13 different exchanges and over 50 unregulated
dark pools. In such a fragmented market, can one really be
confident that achieving best execution, given the explosion of
market data traffic? We question the market's ability to
process the overload of market data.
Conflicts of interest. We question whether the functional
roles of an exchange and a broker-dealer have become blurred
over the years and could warrant regulatory guidance regarding
the inherent conflicts of interest. It seems clear that since
exchanges have migrated to for-profit entities, a conflict has
arisen between seeking volume to grow revenues and their
obligation to assure an orderly marketplace for all investors.
HFT trading strategies. Professional proprietary traders
often have divergent interests from those investors concerned
about the long term. Whether the average holding period for
such traders is measured in seconds as opposed to months or
years, we have destabilized the market. Given recent market
volatility, more study is warranted to assess the impact of the
exponential growth of short-term trading strategies. Most rules
and regulations seem to further enable those with short-term
profit incentives, as evidenced by the proliferation of new
order types suggested by exchanges and approved by regulators.
Suggestions. We believe it is time to step back and examine
the market structure and how it impacts all investors. A good
first step might be to experiment with a number of pilot
programs to examine different structural and rule
modifications. We suggest a look at the appropriateness of
colocation as a general practice and enhanced oversight of
high-frequency trading and other strategies that might be
unduly burdensome to overall market functionality. We would
like to see a pilot program where all payments for order flow,
maker-taker fees, and other inducements for order flow routing
are eliminated. We envision a pilot where there are wider
minimum spreads and mandated times for quotes to be displayed
to render them truly accessible. These programs can include a
spectrum of stocks across market caps and average trading
volumes, among other factors. We also suggest a pilot program
of imposing cancellation fees for unacceptable trade-to-
cancellation ratios.
A key question is, should we foster consolidation in this
fragmented market? At a minimum, should we raise the barrier
for becoming an exchange? In our opinion, requiring a more
robust testing for new software would seem to make sense.
In conclusion, T. Rowe Price Associates appreciates all the
efforts of the SEC and Congress as we strive to make the
markets better and fairer for all participants. The
consolidated audit trail, large trader ID, limit up-down
initiatives are all improvements. We suggest any regulatory
proposals be aligned with the goal of making the markets
simpler, more transparent, and less focused on speed. We
applaud the Committee's interest in making sure the right
questions are asked. There are currently over 1,000 order types
to express your buy and sell interest and we suggest that a
simplified model may be more efficient for all investors.
The issues we face are enormously complex and we certainly
do not have all the answers. We believe it is time to revisit
the historic responsibility to provide a fair and orderly
market. Thank you.
Chairman Reed. Thank you very much, Mr. Brooks.
Mr. Concannon, please.
STATEMENT OF CHRIS CONCANNON, PARTNER AND EXECUTIVE VICE
PRESIDENT, VIRTU FINANCIAL, LLC
Mr. Concannon. Chairman Reed, Ranking Member Crapo, Members
of the Subcommittee, I want to thank you for the opportunity to
appear before you today. My name is Chris Concannon and I am
the Executive Vice President of Virtu Financial.
Virtu Financial is a global electronic market maker. We are
a market maker on over 100 markets around the globe. We make
markets from our offices in New York, L.A., London, Dublin,
Singapore, and Sydney. The company's market making activity
spans across multiple asset classes, including cash equities,
fixed income, currencies, futures, options, energy products,
metals, and other commodities.
Virtu operates as a registered broker-dealer in the U.S.,
as a registered investment firm in Europe and in Asia. We are
also a market maker on the NYSE, the NASDAQ Stock Market, the
BATS Exchange, NYSE Arca, and NYSE Markets. And in Europe, we
are a market maker on the London Stock Exchange, the Swiss
Stock Exchange, Euronext, and Deutsche Bourse.
In discussing the state of the U.S. equity market, I start
from the premise that our equity markets are the most dynamic
and the most efficient markets in the world. My firm trades
across all major financial markets and no market can compare to
the U.S. equity market in terms of pricing efficiency and
liquidity. Over the last 4 years, I have witnessed an
unprecedented number of claims that our markets are horribly
broken, unfair, and dangerous. These claims tend to be short on
facts and evidence, but long on press coverage and book deals.
Let me be clear. Our markets are not perfect. It has flaws
and unnecessary complexity. The U.S. equity markets is overly
fragmented and likely over-engineered. Stocks in the U.S. trade
electronically on 13 national securities exchanges and over 50
dark pools. If we had a blank canvas and we were able to redraw
our entire market structure, it would never look like the model
that we use today.
I would like to focus on three areas that I believe deserve
further review: First, our choice of a single market structure
for all listed securities; second, our market's failure to
enhance market maker obligations; and finally, the industry's
current risk management standards.
First, our market is currently designed as a one-size-fits-
all. What I mean by that is that most of our major market
structure rules do not distinguish between the size, the market
capitalization of the listed company, or the trading
characteristics of their stock. Our markets are designed to
execute all stocks, regardless of shape or size, using the same
market mechanism. A stock that trades once per day is traded in
the same market structure as a stock that trades one million
times per day. It is like we are putting a bicycle on the fast
lane on a highway and wishing it luck. I believe we should
revisit our current market structure in order to create a
better pricing mechanism for all stocks of different shapes and
sizes.
My second area of focus is our market's failure to enhance
market maker obligations. While my firm is a market maker and
it is easier for me to call for enhanced market maker
obligations, I fundamentally believe that we need to increase
obligated liquidity in our markets. Flash crashes, miniflash
crashes, and other market disruptions demonstrate the need for
additional obligated liquidity in our market. However, I
believe enhanced market maker obligations should be targeted
where they are most needed, and that is in the less liquid
stocks.
My final area of focus is the industry's current risk
management standards. In light of recent events, I believe that
the industry should explore ways to improve its risk management
standards. First, pretrade risk management limits are already
required by the Securities and Exchange Commission under the
market access rule. Under that rule, which has been in effect
for over a year, firms are required to establish pretrade
credit limits for every customer account, and more importantly,
for the firm's own proprietary account. These credit limits are
a firm's primary defense against unwanted trading activity by
the firm or its clients.
Second, the industry is currently exploring specialized
kill switches that would be administered by the exchanges.
These kill switches would be a systematic shut-off of a firm if
it exceeded prescribed or preset trading limits. Kill switches
would not be a primary defense but rather a secondary defense
to backstop the failure of other risk management measures
operated by the firm. Kill switches have operated effectively
in the futures exchanges for many years. Such a kill switch
would have severely limited the damage done on August 1.
The third component to enhanced risk management is one of
the most important, I believe. A simple feature referred to in
the industry as drop copies should be required as a risk
management tool. Drop copies are separate and distinct
connections offered by exchanges and other markets. Drop
copies, which are widely used by the industry, provide a real
time echo or copy of a firm's trading activity on a given
exchange.
A good example of how drop copies work is if you are on
Amazon or iTunes and you click that ``purchase'' button, you
get a confirm that pops up in your window. At the same time, an
email is sent to you. That is exactly how a drop copy works in
our industry, and these drop copy emails are actually very
helpful. I recently had my 12-year-old daughter--she actually
became a rogue trader on iTunes. So luckily, the drop copy
emails informed me that she was a rogue trader and I put in a
kill switch.
[Laughter.]
Mr. Concannon. While I believe firms should have a robust
process for developing and testing new software, the industry
must have advanced risk management systems to limit the risk of
unintended trading activity by a firm or by its clients. We
know with certainty that software has bugs, hardware crashes,
and networks go down. The industry must build risk protections
that assume the worst while a robust development and testing
process avoids the worst. Pretrade risk checks, kill switches,
and real time drop copies protect us from the worst events.
Thank you again for the opportunity to be here. I would be
happy to take questions.
Chairman Reed. Thank you very much, Mr. Concannon.
Mr. Tabb, please.
STATEMENT OF LARRY TABB, FOUNDER AND CHIEF EXECUTIVE OFFICER,
TABB GROUP
Mr. Tabb. Good morning, and thank you, Chairman Reed,
Ranking Member Crapo, distinguished Senators. I am Larry Tabb,
founder and CEO of the TABB Group, a financial markets research
firm. We provide research and advisory services to financial
markets firms regarding how the markets operate and how
investors perceive the markets and the brokers who serve them.
I am also a member of the CFTC High-Frequency Trading
Committee. I would like to thank the Committee for this
opportunity to present my views on computerized trading and
equity market structure.
Unfortunately, the structure of the U.S. equity markets has
come under scrutiny, from high-frequency trading, to the Flash
Crash, to Facebook and BATS IPO challenges, to the latest
Knight Capital Group technology issue. Market professionals as
well as the general public are concerned that the U.S. equities
markets are not functioning effectively.
I have submitted a 20-page paper answering the six
questions that the Committee has defined to that end. I will
only summarize my answer to question six. What, if any, policy
changes should be considered by regulators, Congress, in order
to better protect investors, maintain fair, orderly, and
efficient markets, and facilitate capital formation?
My first statement is do no harm. The U.S. markets are the
deepest and most liquid on the globe. The markets are also
complex and interrelated. Small changes can cause significant
impact. So, first, do nothing radical. A radical shift of
market structure will unquestionably hurt investors. Radical
changes provide incentives to traders to analyze rule changes
and profit off of them. This will only cease once investors
pressure brokers and brokers develop better countermeasures. It
can take years to restore this equilibrium.
To that extent, what I would do, first, defragment the
market. While we do not want to limit competition too much, 13
equities exchanges, 50 or so ATSs, and who knows how many
internalizing brokers is too many. Stop granting new exchanges
and ATS licenses immediately. Determine the optimal number of
exchanges, ATSs, and internalizing brokers, and as these
entities go bankrupt, merge or consolidate exiting firms and
reduce the number of licenses.
Second, better manage broker ATS solicitations. An order
for 50,000 shares can easily be executed in 200 trades across
various venues. Understanding what happens to this information
is very, very difficult. While institutions typically receive
execution information, it is more difficult to tell where these
orders were routed and not executed. Where these orders were
not executed leaks information into the market. Between brokers
soliciting the other side, ATSs routing to each other, and
exchanges routing to ATSs, virtually all professionals that
have wanted to trade against this order have seen it before it
is displayed to the public.
Third, better manage MPVs, minimum price variations.
Currently, we have a minimum 1-cent MPV for all stocks over a
dollar. I would follow the direction of the JOBS Act in
implementing a test program to widen spreads for less liquid
and smaller capitalized stocks.
Fourth, greater transparency of order types and routing
mechanisms. Currently, most exchanges post their order types.
However, these descriptions are not intuitive. Exchanges, ATSs,
ECNs, internalizers, and even brokers need to begin to provide
greater transparency, descriptions, and concrete examples how
each order type works, how fees and rebates are generated,
where in the order book orders show up, how and when orders are
routed, and how these orders change under various market
conditions.
Five, develop a market-wide consolidated audit trail for
equities, futures, and options. Develop incentives that will
facilitate the cooperation of the SEC, CFTC, and various SROs
to ensure harmonious oversight. Develop clear rules on what
manipulative behavior is in electronic markets. Provide
regulators the tools and people who can understand the market,
find the people and/or machines that are driving manipulative
behavior, and give the regulators the power to stop, fine, and
jail manipulators. If we had confidence that our regulators
were able to effectively police the market, it would give the
public more confidence that pernicious behavior was being
flagged, challenged, and resolved. It would provide investors
with the assurance that our markets are safe again for trading,
investing, and raising capital.
I would like to thank the Committee for allowing me to
present TABB Group research and my personal thoughts on how to
fix the U.S. equities market structure. If there are questions,
I would enjoy answering them at this time.
Chairman Reed. Thank you all very much for excellent
testimony. I think what you have made very clear is this is a
complex issue, that you have raised extraordinarily important
questions. It is going to take us a while, I think, to come to
the answers, but this is the first step. In fact, in
collaboration with my colleague, I would assume we would
entertain other hearings in this regard because this issue is
not a one-stop and quick fix and move on.
But in that spirit, one of the issues--I was particularly
struck by Mr. Brooks' comments about the issue of speed in
conflict with market integrity. And this issue has been raised
in many different venues, even by people back home in Rhode
Island. Are these markets just too fast now? Are we losing
something bigger? And let me pose this to all the panelists,
your thoughts briefly on that. We will do a 7-minute first
round and then we will come back for a second round and more,
and we have the luxury of asking some questions.
So this whole issue of speed. Are things too fast, and if
that is the case, how do you slow it down? Mr. Lauer, do you
want to comment first, and then Mr. Brooks, and Mr. Concannon.
Mr. Lauer. I would love to. Thank you, Senator. I think
that the issue of speed is a critical one, absolutely. And the
idea of slowing things down is difficult. These are, as you
said, complex issues. I think that speed for speed's sake is
now being well proven to create very little social welfare and
utility.
There is an interesting study that came out on the
externalities of high-frequency trading and they examined two
consecutive technology shocks which increased--or decreased
latency from microseconds to nanoseconds. And while they found
that it dramatically increased trading speed and the
cancellation ratio, there was no impact on trading volume,
spread, depth, or pricing efficiency. So I think we have found
a limit to the benefits of speed, and this is well supported
empirically.
There are some ideas around. One of those is a minimum
quote life, as I mentioned in my oral testimony and my written
testimony, as Mr. Brooks has, as well. I think that that is an
interesting first step and I think it is worth a pilot program.
I completely agree with Mr. Brooks that we need to be trying
new ideas out. We need to be going through some pilot programs
and finding out what the impact of these things would be,
because we can simply surmise on what the impact of a 50-
millisecond quote life would be, but we do not really have a
perfect sense.
That same study found that between 30 to 40 percent of
market data was--are quotes that were put into the market and
canceled within 50 milliseconds without being traded on, and
the cancellation rates on quotes that are out there for less
than 50 milliseconds is, I believe, over 96 percent. So I do
believe that is one area that we can look at.
Chairman Reed. Mr. Brooks, and then I am going to make sure
everyone has a chance to answer this question.
Mr. Brooks. Thank you. I think our sense and what we hear
from our clients is really that the marginal return for
investors is perhaps negative with the increasing speed. So it
begs the question, why do we need to be faster? Who benefits
from that? And I think our question and our concern is, who is
the speed for? We do not think it is for investors. So that
means someone else is interested in speed and maybe their
interest is not as genuine, if you will, when you think about
who are the markets supposed to serve.
So it is a very complex issue. I appreciate the Chairman
acknowledging that as we work together to find solutions. But
there is a concept that speed kills on the highway. You
referenced the highway analogy. So I think speed causes us
great anxiety.
Chairman Reed. Mr. Concannon, please.
Mr. Concannon. Thank you. I think speed is--and I agree
with the panelists that it is one of the more difficult issues
to address, because if you do conclude that we need to regulate
speed, we then are left with the question, at what speed should
we all move at? And we then may move to the slowest common
denominator in terms of speed. Speed and the ability to post a
quote and cancel a quote, they all, in turn, are frictions in
the market. And we have reduced frictions to the most extreme
levels to create what is the most efficient market around the
planet.
We see today--we do not need a pilot--we see markets that
we trade in that have minimum quote life, and the impact is
spread. If you are posting a quote in the market as a market
maker, you are taking risk. You are posting actually an option
to the market, a free option for them to execute against you
when the intrinsic value of that asset changes. That free
option will be adjusted based on the life of that option, which
ultimately is the minimum quote life. And so there is an actual
impact to our market structure when we slow down the market.
Do I think we have reached an equilibrium in speed? Yes. I
do not think moving mics--one or two mics further--is adding
any value to the investing public. But of more concern is
slowing down the market to some common denominator.
Chairman Reed. Thank you.
Mr. Tabb, please.
Mr. Tabb. My issue, my concern, is that--I agree with the
panel that we have gotten faster than we can probably handle.
The challenge is, how do you actually stop it, and what happens
if you stop it?
So the first issue would be, is, OK, so I have a 50
millisecond or even microsecond minimum quote life. What
happens and who manages that? So if I am a trader and I have a
minimum quote life, am I constantly pinging the exchange to
take me out, take me out, take me out, and they say, no, no,
no, no, no, and I say, take me out, take me out, take me out,
and they say, no, no, no, no, and finally 50 microseconds
happens and then all of a sudden we have created all this
extraneous traffic that bogs down the rest of the market, just
all these messages to get me out and how is that managed.
The second issue becomes, is we constantly have folks like,
you know, Intel and IBM and Cisco and networking companies
speeding up the clocks like with chips. Do we just say, OK, you
cannot use the best computers anymore, and how do you manage
that and who governs? Do we go to, like, a stock car race where
everybody needs to trade out of the same computer, and how does
that work and operate?
And then once you get two or three levels behind, you know,
behind actually operating specifically on the exchange, do I
actually know what my customers are doing or their customers
are doing and how do I manage that, as well?
So there are a lot--yes, we have kind of gotten to the
point where the market is too fast, but to a certain extent,
maybe the answer is that the folks like Andrew need to actually
complain to their brokers to have them do a better job of
managing the execution of their orders. And if they do not like
the execution, fire their broker.
Chairman Reed. I am going to yield to Senator Crapo, but I
have another series of questions, because I think there are
some major issues here that the whole panel ought to address
and I would like the comments. Just to sort of preview, one is
the issue of fragmentation, the issue of market maker
obligations, and there are probably a couple of others. And
Mike, if you want to get into them, be my guest, but Senator
Crapo, please.
Senator Crapo. Thank you very much, Mr. Chairman.
I have a question, first, for Mr. Concannon. What kind of a
system does your company have in place to avoid erroneous
events?
Mr. Concannon. Thank you. Well, we have a number of systems
in place, what we call filters, and filters will trigger a
lockdown to our trading system. So if we place an order that
the filter determines is out of range from the price of the
market, it will stop that order from entering the market and it
will shut down the actual trading strategy, and that works all
day long, and it has been in existence for many years. It is
actually a component of the SEC's market access rule, so it is
now technically required to have in place. Every time you send
an order, whether it is for your own account or your client's
account, you have to filter for an erroneous order entering the
market.
Senator Crapo. So this is basically an industry standard at
this point?
Mr. Concannon. Absolutely. It is an industry standard.
Senator Crapo. My next question is--I would like you to
answer, Mr. Concannon, but also, I would like to have the
entire panel weigh in on this, if you would. Yesterday, there
was a Wall Street Journal article about order types called
slide and hide. The article suggested this order type allows
high-speed firms to trade ahead of less sophisticated
investors, potentially disadvantaging them and violating
regulatory rules. What is your perspective on this?
Mr. Concannon. That is a great question. First of all, our
firm does not generally allow exchanges to slide our orders. I
do not know if you have noticed, but----
Senator Crapo. How does this work?
Mr. Concannon. Sure. The way it would work would be you
would place an order that locks the market, and what that means
is it is an order that would otherwise execute against a quote
in the market and that order would be restricted from being
posted and locked. So the order type, the slide order type,
would slide your order back to what is a compliant price and
post it at that price.
The problem with that order type is it is not permitted for
any broker to take an investor's order and slide it. So it is
not actually an investor-used order. So there is no broker that
says, I have an order that is marketable, because it will lock
the market, and I am going to have an exchange slide it back to
a different price. That is a violation of Best X and they
cannot use it that way. So that order type--they would actually
have to, because it is a marketable order, they would have to
execute it, or execute it against the quote in the market.
That order type is mainly used for professional traders. So
the issue is, does that slide order used by a professional
trader, have they read the rules that the exchange filed with
the SEC that were published and did they understand the order
type that they were using? It is not that investors, retail
investors or even Mr. Brooks' firm using that order and being
disadvantaged. His broker cannot use that order and slide his
orders when they would otherwise execute at the market price.
So it is really an issue of professional trader versus
professional trader, trying to understand how the order types
work.
Senator Crapo. Mr. Lauer.
Mr. Lauer. Yes. Thank you, Senator Crapo, if I may. I think
that hide, not slide, and many other order types are great
examples of complexity in the marketplace, and at times, not
necessarily complexity for complexity's sake but perhaps
unjustified complexity. I think that the bar for approving new
order types needs to be revisited, and I think that it would be
a great service to reinstilling some confidence in the
marketplace if it was easier for the average person to get
their head around what these order types mean.
And the fact that we need so much explanation for just this
one order type, let alone--I mean, there are, as Mr. Brooks
said, a thousand order types if you take individual ones at
each market center. There are so many order types that it
just--it is something that people who are not in the business
have a very difficult time with and there is absolutely no
reason they should not have a difficult time with that.
Normally, the bar for regulating and passing rules is very
high. It demands a lot of study and justification as to the
utility of a new rule or regulation, and I think that that same
standard should be applied to all of these exotic order types.
And there should be, at this point, a great amount of evidence
as to the utility of these order types because they have been
around for a while now. I think that that would be an
interesting exercise to go through.
Senator Crapo. Mr. Brooks or Mr. Tabb, do either of you
want to weigh in?
Mr. Brooks. Thank you. I think the only thing we might
suggest is that we might refer to the point Mr. Lauer made, and
that is it really has added complexity to the marketplace and
we are wondering why. Why is someone trying to make things more
complex? And at the Georgetown conference yesterday that you
all referenced, there was some discussion about whether it
would make sense to have a moratorium on approval of new order
types until the marketplace can really come together and get
their arms around, what are we doing here? Why do we need so
many ways to express trading interest? And is there something
else going on? So it is a question that is troubling and we are
not sure what the answers are.
Senator Crapo. Mr. Tabb.
Mr. Tabb. I do not disagree. I think that--and in my
written testimony, or my oral testimony, I presented, I think,
that we need much more transparency on these order types.
Should the SEC be looking over these in more detail? I think,
yes. I think that many of the order types that are out there
are not well known, even to professional investors. And I think
that greater transparency and reducing the number would
probably be a good thing.
Senator Crapo. Yes, Mr. Concannon.
Mr. Concannon. On the topic of order types, I mean, I do
agree that we have way too many order types, but before Reg
NMS, before that rule was installed in our marketplace, we had
very few order types. We had market orders and limit orders. It
was with the complexity of Reg NMS that interconnected all of
our markets and gave us 50 dark pools that we ended up with all
these order types.
And I do have to disagree. I heard--I read in testimony of
a rubber stamp approval at the SEC. I personally filed order
types. There was no rubber stamp approval. It would take 6
months or longer to get these order types approved, and people
like Robert Cook and David Shulman and Bob Colby are far from
rubber stamping order types and they have never rubber stamped
order types. So it is a difficult process to get these things
approved. They are analyzed. There are far too many. But those
order types all came because of Reg NMS.
Senator Crapo. Thank you.
Chairman Reed. Well, thank you, Senator Crapo, and again,
the questions are, I think, right on target and the answers are
not only interesting, but provoke further questions.
Let me take up some of the issues that were suggestive of
some additional testimony. I go back to the issue of
fragmentation, of which you all commented upon, and just give
you an opportunity to add anything further that you might want
to say with respect to the issue of the fragmentation.
One of the ironies is that the national markets were put in
place because there was an oligopoly operating, and now we
have--have we created--sort of gone too far in the other
direction. I think, Mr. Tabb, you could begin.
Mr. Tabb. I think we have gone too far. I think that while
it does not sound like a bad thing to fragment the markets and
have multiple places to trade and have competition, every time
a message pings off a market, a dark pool, an internalizer, it
releases information. And so with trying to find the other side
of the trade, information is bouncing all over the place that
investors do not really even know that is occurring. And I
think, to a certain extent, it disenfranchises the investor and
leaks information.
That said, on the other side, there are organizations who
spend a lot of time, money, and effort building up their
liquidity pools and their businesses. You cannot really just
say, OK, tomorrow, they do not exist. I am not sure eminent
domain would be a great idea.
So the question then becomes is how do you defragment them
and how do you do it in a fair way, and I do not know the
answer to that, but I think we clearly have too many places to
trade.
Chairman Reed. Mr. Concannon.
Mr. Concannon. Well, when we take a look at our market and
we see 13 exchanges, 50 dark pools, you have to understand that
in every stock, there is a bid or an offer finding its way in
one of those liquidity zones. And what that does is it thins
out our overall liquidity. So on a normal day, it works. On a
day of high volatility, like May 6, that thin layer of
liquidity is easily pierced. When it is not consolidated in one
place, you can pierce through that liquidity and you have the
volatility that we saw that day. So fragmentation has caused
increased volatility, but I agree with Larry. It is very
difficult to consolidate that market. Again, on an average day,
there are dealers that are professionals that try and
consolidate that market and it is a very liquid market despite
the fragmentation. It is really those days that we have shocks
and events.
Chairman Reed. Mr. Brooks, do you have any comments? Mr.
Lauer? Mr. Brooks.
Mr. Brooks. Senator, if I might, I do not think we know the
answer in terms of what the perfect number of exchanges or dark
pools might be, but we do wonder and hear others question
whether the system, with all these different venues, can
process the data that is being generated. I mean, there is just
an enormous amount of data, and some of that data gums up the
system and slows it down and that is troubling.
And, two, to pick up on Mr. Concannon's point, it is
awfully hard to figure out the supply demand equation in the
stock today when there are so many different places that
trading interests reside. So if you are trying to find out if
there is more to buy or more to sell, good luck. It is tough
today.
Chairman Reed. Mr. Lauer, a comment, and then I have
another question.
Mr. Lauer. Yes. I think that what we have seen with this
massive fragmentation and internalization and dark pools, as
Mr. Concannon alluded to, is what is called adverse selection.
As order flow goes from place to place, from retail
institutional investors, before it makes itself--before it gets
to the lit exchange, it is picked off at every step of the way
by internalizers, by dark pools, by high-frequency traders,
proprietary trading desks. And, therefore, the flow that
eventually gets to the lit exchanges is what we refer to in the
industry as toxic flow that nobody wants to trade with. And
what we have done is we have reduced the number of natural
buyers and sellers in the market, as Mr. Concannon said when he
was just answering. So I think that that has been one of the
dramatic effects of fragmentation, internalization and dark
pools.
There was a study out of Rutgers that demonstrated that
because of all the off-exchange executions, spreads are
widening in lit exchanges by an average of 1.28 cents on the
New York Stock Exchange, and that is a dramatic effect and it
is a direct result of this adverse selection because it becomes
harder to be a profitable market maker in such a scenario.
I think that we can take a cue from other countries in this
regard and we can raise off-exchange execution standards. We
can say that you must meaningfully price improve in order to
internalize, and meaningful price improvement does not mean a
tenth of a cent. In Canada, they have one standard that says at
least a tick size, a minimum tick size, and if the spread is
one penny, then at least half a tick size. They also have a
minimum order size. So if you are going to internalize or
execute on a dark pool, it has to be an especially large order.
That is why you have dark pools and that should be the
reasoning behind that.
I also think that eliminating pay for order flow is an
important step in order to combat this adverse selection
problem.
Chairman Reed. Thank you. Let me raise another topic, and
that is this issue of market making. Mr. Concannon, your firm
is a market maker in several different venues. And related to
this, and it might not be precisely related, is another issue
that we hear about and was referred to in the testimony is
people coming into markets, making a bid and then canceling it
within fractions of a second, sort of probing, which creates
more information but is not the traditional, what people think
what markets are, where someone really wants to buy something
and they are looking for the best price and when they get it,
they are going to close the deal.
So this whole issue of market making obligations on
everyone, which I think you raised in your testimony, can you
comment upon it? Then I will ask Mr. Tabb and then everyone
else who wants to make a comment. Please.
Mr. Concannon. Sure. I think if we go back 20 years and we
look at our market makers, they had fairly strict obligations,
people like specialists on the floor, but they had exceptional
benefits. They had exclusive rights to the market and the
incoming order, and that was worse than the benefits that they
provided as market makers. So we made a decision and we changed
our markets to be an all to all.
And I do believe in some products that all to all model,
where anyone can be a market maker or market maker obligations
are light, at best, we lose a little bit in terms of liquidity.
And so there are stocks that I do not necessarily think they
need obligations and market makers all day long, every minute
of the day, but the good majority of our market really needs to
be supported by market makers with real obligations to be at
the best bid or offer.
The most important thing for me in terms of market maker
obligations that do not exist is that not only should you be at
the best price, but you should be at the next best price and
one more price below that. You need to provide--a real market
maker makes depth of liquidity markets, not just a thin layer
at the best bidder offer.
Chairman Reed. Mr. Tabb, your comments.
Mr. Tabb. I agree. I think that we should have some sort of
market maker obligations. Now, that said, market maker
obligations is not a panacea. If we wind up having some sort of
crash, people will not step in front of a train, catch a
falling knife, whatever analogy you want to make. They are not
going to go out of business because they have an obligation.
But that said, in the general market or general interday
volatility, I believe we should have stronger market making
obligations.
Chairman Reed. Thank you. Comments, Mr. Brooks, Mr. Lauer?
Mr. Lauer.
Mr. Lauer. Yes. Thank you, Senator. I believe that--I
completely agree with Mr. Concannon and I think that market
making obligations should be the cornerstone of any new
efforts. And I believe strongly that they should be tied to
receiving a rebate on any venue in which securities are
transacted, be they an exchange, ATS, ECN, dark pool. Any place
that somebody can receive a rebate, that rebate should be tied
to an obligation to make markets.
And as I explained in my written testimony, one of the
better studied and more beneficial models is what is called the
maximum spread model. I think the way obligations have been
done in the past is normally something like a 90 percent of the
time you need to be in the market. There are other models, but
that is a standard one, and both Mr. Concannon and Mr. Tabb are
correct that that would not address the issues of a crash or a
flash crash, which is why I think we need to examine much more
stringent market making obligations. And some studies have
developed some models for that that do demonstrate they would
prevent flash crashes and that the market makers in some
instances would have to be compensated for their losses. But
the overall improvement in social welfare far outweighs the
cost to that compensation.
Chairman Reed. Mr. Brooks, you have a comment, and then I
will yield.
Mr. Brooks. I do. Thank you. I guess our thought would be
that if you are going to receive a benefit as a market maker,
you have to provide one, as well.
Chairman Reed. Thank you.
Senator Crapo.
Senator Crapo. Thank you, Mr. Chairman.
As I listen to the different issues being raised here and
the potential solutions, a question occurs to me. This is a
broad question and I would like to ask each of you to just
weigh in on it, if you would. What do you think the top two or
three issues or solutions to these issues that we are
discussing here should be, the SEC should consider in its
October roundtable? What are the top areas of focus we should
recommend to the SEC or see them consider? We will just start
over here. Mr. Lauer.
Mr. Lauer. Thank you, Senator. The roundtable is a
technology roundtable, and as such, the number one thing should
be market technology standards, policing, the kill switch that
has been referred to on the panel, these types of issues. As I
pointed out in my written testimony and my oral testimony, I
think we should also--we need to get more creative. The SEC is,
I believe as Senator Reed said, hopelessly outgunned. They need
a pretty dramatic improvement in their technology capabilities.
The market is a market based on technology. I am a
technologist. I would never argue to move backwards. I would
only argue to move forwards.
But the SEC needs to take a market-wide approach to
surveillance and enforcement. They need to build a strong
market data plan. And if they could open access to market data,
which is a controversial issue, they could really inspire some
novel approaches, some creative thinking by independent
programmers, a model that has been demonstrated time and again
to be very successful for developing really interesting
applications and inspiring new innovation, and that is
something--a change in mindset that I would be very pleased to
see.
Senator Crapo. Thank you.
Mr. Brooks.
Mr. Brooks. I think our sense would be that perhaps three
things for the technology crowd to look into. One would be
order routing practices. Why are orders going where they are?
What is driving that?
Payment mechanisms, so maker-taker payment for order flow
internalization. At some point, it is all about the money, so I
think that is an important conversation to have regarding
technology.
And finally, we are concerned and our clients are concerned
about the sense of unfair advantage that market data
aggregators are giving to others and the sense that you get a
view of the horse race's finish before anybody else, and I
think technology needs to address that.
Senator Crapo. Thank you.
Mr. Concannon.
Mr. Concannon. Well, obviously, this roundtable was
scheduled because of the summer of technology mishaps that we
experienced. I think the focus will largely be on the proper
testing that we deploy, the processes that we put around our
new code that we roll into production, both exchanges and
broker-dealers.
But I hope the focus will be more around what happens when
something goes wrong. We will always have bugs in our code. We
will always have network gear breakdown. We will have networks
flat. Those all cause disruptions that we have to build risk
management procedures around. So we will never get rid of the
bugs. We will never get rid of the mishaps. But it is how we
respond to them that is more critical.
I actually think a very simple adjustment to the SEC's rule
on market access is all that needs to be made. If we--we
already have pretrade covered, so if everyone complied with
pretrade, we are going to protect ourselves from what goes out
the door. If we install these kill switches at the exchanges
that are not that difficult to deploy, we can make the kill
switch setting part of your market access obligation. So a
broker has to justify his setting in the kill switch to the SEC
when they come in and inspect.
And the final piece that I mentioned, the drop copy piece,
that covers your post-trade. So even if you miss something,
even if an exchange creates risk for you because of an outage,
you will still know on your post-trade, your drop copy, that
there is someone trading in your name and you can check that
against your risk procedures.
And all three items would be part of the market access rule
and would be all part of your procedures that you roll out for
the SEC when they examine you.
Senator Crapo. Thank you.
Mr. Tabb.
Mr. Tabb. I agree with a lot of what Mr. Concannon said
about the kill switches and risk issues. One of the other
issues, though, is that to a certain extent, some of these
rules that the SEC has put in actually have not done--have not
been specified properly and actually do not really work. And so
I think the SEC really needs to do a better job in terms of how
they specify some of their rules, like the direct access rule
did not help Knight. It is supposed to be across asset class,
but nobody has got cross-asset class capability now.
Linked markets--it is very easy to do pretrade risk in a
single market, but once you start trading across 50 markets, it
becomes very hard to manage all that pretrade risk and you wind
up with people raising their credit limits just so that they do
not have to deal with it.
The large trader rules were specified out of the wrong part
of the data base that does not actually have trade data from
the dealers. It would be nice if they actually took the time to
actually understand what they are specifying.
So, now, getting back to my other thought of defragmenting
the market, so if you look at market data, the issue is if I
have a market and I am a market maker, even if no one trades
there, I have got to quote every name that I am going to quote.
So if I have two markets, all of the liquidity is here but I
decide to quote over here, as well, so I have got all of a
sudden twice as many quotes, which is kind of on a logarithmic
scale in terms of amount of market data. So if we start pruning
back the number of exchanges and licenses, we will wind up
actually minimizing some of the market data challenges.
And last is we really do need a consolidated audit trail
and a set of regulators that can actually look over this stuff,
because at the end of the day, if the SEC is bringing people to
task, we will have a little bit better confidence that they
have their eye on the switch, you know, their eye on the wheel.
Senator Crapo. Thank you very much.
Mr. Lauer. Senator, if I may, just a quick comment----
Senator Crapo. Yes.
Mr. Lauer. ----along the lines of the consolidated audit
trail, another important idea. The problem with the
consolidated audit trail, while it is a fantastic idea and
critically needed, it is years away. If we right away said,
from now on, when you put a quote into the market, you have to
include a unique ID number, either on a per account basis, like
the consolidated audit trail specifies, or on a supervisory
individual basis, and that ID number is propagated not publicly
but down to regulators, into their market data plan, it would
have a dramatically chilling effect on manipulative behavior
and it would allow regulators to quickly reconstruct what
happened within the markets. It would allow them to enforce the
existing rules. And it would also provide a contact point to
quickly identify what is going on in an emergency. And I
believe, technologically, that is a very easy task for current
exchanges and their customers to take.
Senator Crapo. Thank you.
Chairman Reed. Thank you very much, Senator Crapo. I have a
few other questions. My colleague has to depart for an
interview, but I want to thank him. I thought it was an
excellent question, to try to focus on the SEC panel. As you
point out, it is generally a technology focus.
We have talked about technology, but also really important
policy issues, the number of orders that are available, et
cetera, and I just have two general questions. And first, Mr.
Tabb and others that might want to comment, one of the things
that has happened over the last several years is the
proliferation of these dark pools and it raises the question of
should there be a sort of uniform set of rules or conduct or
behavior that apply to the dark pools as well as to the lit
exchanges, and just any thoughts you might have.
Mr. Tabb. A very good question. They are definitely
regulated in separate ways and the exchanges would certainly
like an even playing field. And to a certain extent, I agree
with that.
The one challenge academically, and to a certain extent--
and I think Mr. Brooks was talking a little bit about this, I
think, earlier--I think it was Mr. Brooks--is that what Reg NMS
did was infuse a single kind of market structure, a fast market
priority over slow markets, and what winds up happening is that
once you wind up having a trade-through rule and you wind up
having the same priorities across markets, in effect, what
happens is the markets become tighter, more tight, more closely
linked together, and problems can cascade through them. And so
by then even tying together the dark pools with the exchanges
in a trade-through mechanism, what you are going to--you might
even create more fragility because then everything has to work
exactly the same way, and keeping, you know, 13 exchanges and
50 dark pools lined up like that may be really difficult.
One of the answers may actually be to get rid of the trade-
through rule and let markets develop a little bit more
naturally. I think that was Mr. Brooks. Maybe it was Mr.
Concannon was talking about that earlier. And that one market
structure actually does not fit all. What you may want to
actually do is actually what Mr. Lauer said, is have some sort
of trade-at rule which basically says, you know, you cannot
have de minimis price improvement and you need to price improve
at least a half a tick or something like that to bring more
liquidity into the market centers.
But, by and large, it is a complicated--it is a complicated
issue. I do not know what the right answer is.
Chairman Reed. I concur.
[Laughter.]
Chairman Reed. Any other comments on this question, and
then I have one more general question. Mr. Concannon.
Mr. Concannon. Sure. I am actually fairly sympathetic to
the exchanges' situation. They are--they do have competition in
the market called the dark pool. They are subject to heightened
obligations as an exchange. But dark pools are registered. They
are registered ATSs. They are subject to Regulation ATS. So
they do have some of the components of the exchange regulation
on them, as well.
I think the one thing I am a little bit cautious about is
the innovation in our market actually came from--not from the
traditional exchanges, but from the original ATSs in our
market. And today, those exchanges are actually former ATSs
that were purchased by the exchanges. So I am concerned if we
put too much regulation in the over-the-counter market for ATSs
or dark pools that you will dampen that innovation in a market
that we have enjoyed for many years.
The other concern I have around dark pools is that it is a
place for the buy side, people like Mr. Brooks, to feel
comfortable about how their orders are handled and traded and
not being pushed into the displayed market where people can
find their orders and have an impact on the market. So there is
a protective mechanism sitting inside the dark pool that is
valued by the buy side and needs to be protected.
Chairman Reed. Mr. Brooks, you have a comment.
Mr. Brooks. I do, and I appreciate that, Mr. Concannon,
because we do need protection because our orders are larger and
a lot of the marketplace today is trying to identify our order
flow and trade against it. So we are paranoid about that, and
we should be, and I would suggest we have always been paranoid
about that. So that is not new news.
But in a perverse way, the dark pools' execution size, I
believe, is not much greater than the lit markets, 200 shares.
So what it was designed for, which might be to rest and hide
order flow but be willing to trade, not get picked off, not be
identified, it is not really accomplishing that for the most
part. So that begs the question of what is going on in dark
pools, and again, I think we all are somewhat challenged by
truly understanding what is happening there and that needs
further examination by all parties.
Chairman Reed. Thank you.
If you have a quick comment, then I have one more question,
Mr. Lauer.
Mr. Lauer. Sure. I would just agree that I do believe that
some unification of rules is necessary across venues. I think
that is very important. And I think that, as I said before, the
burden to execute on a dark pool should be much greater than
executing on a lit exchange because you are removing
information and liquidity from the public domain and lit
exchanges.
Chairman Reed. You talked about the dark pool. Let us shift
focus to the exchanges, which now are for-profit enterprises
with proprietary operations, et cetera. And there have been
some questions, concerns raised about proprietary data feeds,
customers getting advantages over other traders, and the whole
issue I think you raised, Mr. Brooks, in terms of your
testimony, just about the standard of behavior that grew up in
what was a utility more than anything else and now is a for-
profit enterprise. So any comments that you might have, and I
will begin with you, Mr. Brooks, about any things that we
should do given the nature of these exchanges as for-profit
vehicles and their issues of proprietary feeds, colocation,
access. That was brought up. Any comments.
Mr. Brooks. So, again, you have picked a very tough
question to pose to everybody in the industry. We certainly do
not have the answers here, but it is challenging when exchanges
have a for-profit motive and are interested in growing volume,
and investors are not really interested in volume. They are
interested in liquidity. And so those are two really different
things.
I am certain that exchanges can be profitable and can work
well as a for-profit enterprise, but perhaps we need to
understand really what their role should be going forward, who
they should be serving, and by the way, what is the role of a
broker-dealer today. These lines, as we stated earlier, have
really become blurred and we need some guidance here. We need
to bring together some great minds to really think through what
does it mean to have these different roles and how should the--
it is sort of a Glass-Steagall question. What is appropriate to
be separate?
We are troubled by the blurring of the lines. We are
troubled by market data aggregation and dissemination
selectively. Certainly, I can understand that data should be
gathered and it can be sold and we are OK with that, but equal
access important.
You know, when we think about things, when we talk to
clients and shareholders, we are interested--they are
interested in fairness and a sense of balance, and our
Chairman, Brian Rogers, always says, how do we get back to
fairness and balance? And I think that is a good question to
always be asking yourself. Are things as fair as they could be?
You know, we have a group of terrific, dedicated traders
every day trying to find fairness in the marketplace for our
investors and they do not feel so good about things. Things are
moving on us. Things are--we are getting identified. People are
taking advantage. There is opportunity being taken away from
investors. It is the worst I have ever seen in my career, and I
have seen a lot, not maybe as much as you have or others, but
we have seen a lot of different things and it is a troubling
time.
Chairman Reed. Other comments? Mr. Concannon.
Mr. Concannon. I agree that it is a very difficult question
around proprietary trading--proprietary feeds that come out of
the exchanges. If you look around the world, most markets do
not have a consolidated feed. It is all--Europe is all based on
the feed that comes from the exchange and the direct exchange.
And so, commercially, people consolidate those feeds.
I am encouraged when we think about in the U.S. that the
proprietary feeds are being consumed by investors. Some of the
biggest consumers of those proprietary feeds are some of the
retail online firms. They do push those proprietary feeds onto
their Web site. They do do a form of consolidation. And when
those retail orders come into the market, they are executed
based on the proprietary feed. So there is some fairness being
in terms of how the proprietary feeds make it out to the
investing public and how their orders are treated under the
proprietary feeds.
But what that leaves me with a question, is then what is
the value of our consolidated feeds, and I was more shocked
that--I think the fact that they were so slow and went
unnoticed by the industry kind of speaks volumes in terms of
their value to the industry.
Chairman Reed. Anyone else? Mr. Tabb.
Mr. Tabb. Yes. I kind of agree with Chris. I think that
most professional investors--and I am not as familiar with the
retail side, Chris would be much more familiar with it. On the
institutional side, most of the large brokers who are servicing
the large institutions use proprietary feeds. A lot of their
algos and technologies are based off that. I do not know
necessarily who uses the consolidated, the aggregated feed, the
sites, things like Yahoo!.
As we start getting issues around colocation and governance
of that speed, I am--even if we ban--let us say we ban
proprietary feeds and we only use the aggregated SIP feed. The
issue also then becomes is where is it being intercepted and
read, and with the speed of light, someone who reads it in New
Jersey is always going to get it before someone who reads it in
California and how do you manage that. And I am not sure how
you wind up delaying it so everybody gets it exactly at the
same time.
Chairman Reed. Your testimony was a good lesson in physics,
so thank you very much.
Mr. Lauer, you get a final comment, and then I will
summarize.
Mr. Lauer. Yes. I think as far as colocation goes, there is
a case to be made for eliminating colocation. The argument that
is often made against it is exactly what Mr. Tabb has said. I
think that, right now, if you can think about it, though,
colocation has a very small radius of equality. You know, it is
a very limited area in which people are equal. So if you
eliminate that, you at least expand the boundary and widen the
opportunity to many more participants. You can reduce the
costs, as well.
I do think there is a case to be made for it. As Mr. Brooks
said in his opening statement, which I thought was
illuminating, there is no other regulated industry in which
access to information is sold for an advantage by somebody who
can act before someone else. To me, colocation in some ways
does reek of nonpublic information. I have a server. It is
moving very fast. I can receive a piece of market data, analyze
it, and act on it before many people have even received that
piece of market data. And again, in my written testimony, there
are several studies that have demonstrated that this race to
zero is having no effect--no beneficial effect on market
quality.
I think as far as the number of exchanges in the for-profit
exchange model goes, I think what we need is a diversity of
business models. Right now, there is pretty much only one
business model in the marketplace. That is the maker-taker
model. And so I think that it is--the environment is ripe for
innovative ideas, and, of course, in my work with IEX Group,
that is what we are working on. But I would like to see many
more like that. So we could have different forces of
defragmentation reducing identical business models, but coming
up with innovative new ways that really appeal to the investor
community rather than the trader community, and that is what
the maker-taker model does. The incentive structure is very
skewed toward high-speed trading and volume rather than
liquidity, which is a major distinction that needs to be drawn.
Chairman Reed. I have just one final question which my
staff reminded me, and this is in the wake of the volatility we
have seen in the oil commodities markets. The inference in a
lot of what we talked about today were equity markets, et
cetera, but the interaction of electronic funds trading,
commodity markets, high-frequency trading, et cetera, raises
another issue about volatility.
And this always comes up in the context, I think,
particularly of oil and commodities. Is there a way to
manipulate these markets, and either wittingly or unwittingly,
is there the possibility, the potential, or the reality of
manipulation, particularly given the incentive of oil to
everything we do? And you might want to comment, Mr. Lauer, and
then I will ask Mr. Tabb.
Mr. Lauer. Yes. The answer is simply, yes. Absolutely,
there is a way to manipulate prices and markets, especially
with the current speed of systems right now, and what we see
sometimes, which are these what are called illiquidity
contagions or these miniflash crashes. There is well-documented
evidence of practices such as quote stuffing, which is to slow
down the channel of a direct proprietary feed in order to pick
off participants that are slower or unable to keep up with a
high volume of data.
It is, in fact, well documented in one study that we see
what is called comovement of message flow within channels,
which is a rather shocking conclusion. Channels from a
proprietary feed perspective are alphabetically distributed
and, therefore, completely random, and you should expect to see
no movement in the A through C channel. You would expect to see
comovement of stocks in the same industry or exposed to the
same macroeconomic factors. But the idea that there is going to
be movement on a stock because of the letter it begins with is
perfect evidence that there is manipulation going on. That also
leads to things like stop hunts, very manipulative behavior,
and there is no way to figure out who is perpetrating that,
which is one reason why I think this unique identifier attached
to quotes is an absolutely critical issue and something that
could be moved on very quickly.
Chairman Reed. Mr. Tabb, you have a comment, and then I
will conclude.
Mr. Tabb. The way I want to answer that is yes and no. I
think what Mr. Lauer is saying is absolutely true in the short
term, very, very short term. I think there is an ability to
manipulate stocks, especially in low liquid environments. But,
generally, that manipulation is only going to take a little bit
of--a short period of time from the electronic trading crowd,
because, generally, the length of time they are holding a
position is pretty short. So if they are buying 1 minute, they
are selling the next minute. So they are only locking in small
increments.
That said, there are people who can push markets. The
people who push markets actually are not the short-term
holders. They are longer-term holders or intermediate-term
holders because they have capital at stake. They are willing to
push a position for, you know, days or weeks, and those tend to
be more hedge fund-oriented type organizations. But if you are
thinking of the traditional short-term electronic trading guy,
I think to have an impact over a couple of minutes, yes. Impact
over a day, pretty hard. You know, if you are going to have
impact over a day, it has got to be somebody with much deeper
pockets.
Chairman Reed. Thank you.
Gentlemen, thank you for excellent testimony. I think what
you have really exposed is the complexity of these issues, and
there are not just one or two. There is a series of
interrelated issues. And I think, also, the need to begin to
act promptly to address all of these issues, to give the
investing public confidence that the system is operating
fairly. I think this notion that you said very well, Mr.
Brooks, about fairness, I mean, we have recognized benefits,
obviously, from the increased liquidity, from the decreased
spreads, that now we have to sort of step back and see at what
cost and how do we make improvements, not just simply keep
pressing along.
I must say, though, I did get some reassurance, with a 5\1/
2\-year-old daughter, that iTunes does provide a drop copy.
Thank you, Mr. Concannon.
[Laughter.]
Chairman Reed. I am making a note of that so that we have
accomplished one objective today.
But I think, again, just to say, this is an issue that will
not go away. In fact, with technology, it will get even more
complicated. I am pleased that the SEC is undertaking the
technological roundtable, but I would like to be able to, with
the concurrence of my colleague, Senator Crapo, do additional
hearings, because I think you have raised some extraordinarily
complicated issues in a very thoughtful way and we have to do a
lot more work, along with the regulators, to come up with
sensible responses.
Mr. Tabb, your point, I think, is always well made. Try to
do no harm. Try to prioritize changes that are the least
disruptive and the most effective, and that is always good
advice, so thank you for that.
But, gentlemen, thank you so much. I have one other
statement to make for the record.
First of all, I want to thank all of our witnesses for
testifying today.
[Laughter.]
Chairman Reed. We appreciate both the time and effort that
you have made. That should be obvious. Thank you so much.
Now, if you have additional written statements, please feel
free to submit them. You may receive additional questions from
my colleagues. Please respond as rapidly as possible. I will
ask my colleagues to submit their questions no later than next
Thursday, September 27, and please respond as quickly as
possible to that. If any of my colleagues want to make a
statement for the record, it will be included by unanimous
consent in the record.
In addition, I ask the statement of Public Citizen be
included in the record, and hearing no objection, so ordered.
Chairman Reed. With that, thank you very much, gentlemen.
The hearing is adjourned.
[Whereupon, at 11:28 a.m., the hearing was adjourned.]
[Prepared statements, responses to written questions, and
additional material supplied for the record follow:]
PREPARED STATEMENT OF DAVID LAUER
Market Structure and High-Frequency Trading Consultant, Better Markets
September 20, 2012
Good morning Chairman Reed, Ranking Member Crapo, and Members of
the Subcommittee. Thank you for the invitation to Better Markets to
testify today.
Better Markets is a nonprofit, nonpartisan organization that
promotes the public interest in the domestic and global capital and
commodity markets. It advocates for transparency, oversight, and
accountability with the goal of a stronger, safer financial system that
is less prone to crisis and failure, thereby, eliminating or minimizing
the need for more taxpayer funded bailouts. Better Markets has filed
almost 100 comment letters in the U.S. rulemaking process related to
implementing the financial reform law and has had dozens of meetings
with regulators. Our Web site, www.bettermarkets.com, includes
information on these and the many other activities of Better Markets.
My name is David Lauer and I am a Market Structure and High-
Frequency Trading Consultant to Better Markets. I'm also consulting for
IEX Group, Inc., a private company that is in the process of building
an investor-owned and investor-focused U.S. equity market center. Prior
to working with Better Markets and IEX Group, I worked as a senior
quantitative analyst at Allston Trading and before that at Citadel
Investment Group. Prior to my career as a researcher and trader, I
worked at Tervela where I helped to design hardware and specialized in
studying and understanding the complexities of the rapidly evolving
electronic marketplace both before and after Reg NMS was implemented. I
have a Master's degree in International Economics and Finance from
Brandeis University. I grew up in Southern New Jersey.
Introduction
On May 6, 2010, the U.S. stock market demonstrated some of the most
unpredictable and disturbing behavior in its 218 year history. Over the
course of just 20 minutes, the stock market plunged and snapped back
up, losing and then regained nearly $1 trillion in market value. Nobody
had ever seen anything like it. The crash began in the S&P E-Mini
Futures market but quickly spread to and overwhelmed the equity
markets. That day I was in the center of the storm, working on the
high-frequency trading floor of one of the largest S&P 500 E-Mini
Futures trading firms in the world, on the global equity trading desk.
As I watched the market crash, I witnessed something unthinkable:
the market simply disappeared. For what felt like an eternity, but was
more likely 30 seconds to a minute, there were no bids or offers
displayed in the market for major stocks and ETF's such as SPY (the S&P
500 Index ETF).
None of us knew what to do or what would happen next. Immediately
before the market disappeared, our firm, like other high-frequency
trading firms, withdrew our orders from the market because we did not
understand what was happening, did not trust our data feeds and had no
obligation to remain active in the market. Anybody who seeks to
minimize the role that high-frequency trading had in the Flash Crash
either was not on a trading floor that day or has an interest in
maintaining the current unregulated status quo.
When more than half of the liquidity in the stock market is able to
be pulled from it in a matter of seconds, dramatically worsening an
unstable situation, something is dreadfully wrong.
U.S. equity markets are in dire straits. We are truly in a crisis.
Over the past three decades a technological revolution has swept over
Wall Street. In many ways, this has dramatically improved the
efficiency of capital markets relative to past decades: reducing
spreads and volatility, and helping them to more effectively perform
their core functions of price discovery and capital formation.
Regardless of the arguments about this extremely volatile issue, we
must judge the evolution of capital markets around these
characteristics of spread width, price volatility, price discovery and
capital formation as well as other characteristics such as the price
impact of large institutional orders and catastrophic event frequency.
The past decade of this revolution has been marked by two primary
trends: extreme marketplace fragmentation and rapid growth of high-
frequency traders as the primary suppliers of liquidity. As of today,
there are 13 lit exchanges and more than 50 dark pools/internalization
venues. Exchanges account for between 65 percent and 75 percent of the
market, and the Dark Pools, with 18 tracked by Rosenblatt, account for
12 percent-15 percent. The balance is attributed to internalization,
including OTC block trades and wholesalers. This is a far departure
from the stock market of the 20th century that was well understood by
most Americans. If you were to ask the average retail trader or even
sophisticated institutional investor what happens today when they send
a buy or sell order to the market, few if any would be able to describe
the labyrinthine path that order takes to be filled.
Complexity has become the hallmark of the new electronic landscape,
whether it is in the form of a multitude of venues and participants or
the advanced algorithms that many of them are using to analyze incoming
market data. While complex systems can often provide elegant solutions
to intractable real-world problems, they can also spin out of control
in unexpected ways. Often the interaction of these nonlinear systems is
difficult or impossible to predict. In the U.S. equity market, we have
seen first-hand glimpses of what can happen as overly complex systems
interact in nonlinear ways. The incidents are becoming more common, and
include:
The Flash Crash in May, 2010, was set off by a single large
trade estimated at $4.1 billion in the S&P 500 E-Mini Futures
Market. The cascade led to 20 minutes of extreme volatility,
wiping out nearly $1 trillion of market cap before quickly and
inexplicably recovering. The total economic cost of this event
is unmeasured, but certainly huge. We were lucky it didn't
happen near the market close--had the U.S. market closed before
it recovered, the result could have been total economic
disaster because money would have hemorrhaged out of the stock
market overnight.
In August, 2011, the stock market swung up and down by over
4.4 percent on four consecutive days, alternating up and down
days. It was wild, unprecedented volatility--only the third
time in history that had happened, with the second time having
been 3 years prior, during the crash of 2008. While the
European crisis was becoming a more important issue at the
time, this volatility was not warranted by major economic
changes or historic macroeconomic events. This was computer-
driven volatility.
``Mini flash crashes'' occur on a near-daily basis in
individual stocks. Nanex has documented almost 2,000 instances
of individual irregularities in stocks since August 2011. \1\
Single-stock circuit breakers have failed to stem the tide of
these incidents.
---------------------------------------------------------------------------
\1\ http://www.nanex.net/aqck/aqckIndex.html
IPO's in Facebook and BATS (itself an Exchange) have gone
horribly wrong due to technological ``glitches,'' continuing to
sour the already languid market for IPO's and costing untold
numbers of jobs as companies cannot raise the capital they need
---------------------------------------------------------------------------
to expand and hire.
Few realize how lucky we were on Tuesday, July 30. An order
to sell nearly $4.1 billion in the S&P 500 E-Mini Futures
Market, the same size as what precipitated the Flash Crash, was
executed three seconds before the market closed. There simply
was not enough time for the waterfall of May 6, 2010, to repeat
itself. What happens the next time when that same order is sent
in a couple of minutes sooner? While some may point to this as
evidence that the market worked, there have not been any
changes in market structure since the Flash Crash other than
circuit breakers, which are not active in the final 25 minutes
of trading.
On Wednesday, July 31, Knight Capital Group--one of the
largest market making firms, an official Designated Market
Maker on the NYSE, had a ``software glitch'' according to their
CEO. The result? A loss for them estimated at $440 million,
untold economic losses for retail investors with stop-loss
orders in one of the almost 140 stocks that were affected and
further erosion in investor confidence.
There is no doubt that electronic trading has tremendous value to
offer, at times enhancing the smooth functioning of the stock market
and increasing competition, thus driving down the spread that the
average investor has to pay to buy or sell a stock. HFT has been so
successful that it has taken over the stock market, now accounting for
between 50 percent-70 percent of equity market volume on any given day.
Fortunes have been made, with estimated annual profits exceeding $21
billion \2\ at its peak, and estimates varying but still in the
billions of dollars today.
---------------------------------------------------------------------------
\2\ Rob Iati, ``The Real Story of Trading Software Espionage'',
http://advancedtrading.com/algorithms/
showArticle.jhtml?articleID=218401501, July 10, 2009.
---------------------------------------------------------------------------
What we must be concerned with is whether the pendulum has swung
too far, and whether the nearly unregulated activities of anywhere from
50 percent-70 percent of stock market volume should be permitted to
continue down this path. For the proponents of HFT to make the case
that the market is functioning well, or that only incremental reforms
are needed rather than wholesale changes, they must make the case
unequivocally that the market satisfies the aforementioned
characteristics of tightening spreads, decreasing volatility and is a
more efficient and low-cost mechanism for price discovery and capital
formation today than at any time in the past, and that proposed reforms
are not even worth trying.
Despite the large quantity of HFT-funded research touting the
benefits of HFT (lower spreads, increased liquidity, lower volatility),
there are also many research studies that prove the opposite. However,
a higher-level debate on the impact of HFT on today's market should
also consider the fact that catastrophic event frequency has increased,
IPO's have dramatically dropped and retail investors have been fleeing
the stock market in droves. The flight of the retail investor during a
period of incredible stock market returns is a sure sign that this
exodus is a result of mistrust rather than economic conditions.
Investor confidence is nonexistent, with only 15 percent of the public
expressing trust in the stock market in the latest Chicago Booth/
Kellogg School Financial Trust Index--8 percent lower than those that
trust the banks! The figure that follows is one of the most disturbing
illustrations of the flight of the retail investor:
Historically, the correlation between market performance and net
flows is undeniable. Despite the rally over the past 3 years, fund
inflows have not followed. Since the Flash Crash in May 2010, over
$283bn has flown out of the U.S. Equity markets. Over that time period,
the S&P 500 has risen by over 21 percent.
This trend, along with unpredictable and increasing volatility, has
also driven companies away from the stock market. From 1990-2000, the
average number of firms going public each year was 530. Since 2001,
that number has plummeted to 125. When companies are able to easily go
public and access large amounts of capital, they are able to grow,
expand and hire. When fragility, volatility and mistrust are the
defining characteristics of the market, companies do not see an IPO as
a viable means to raise capital, or in their long-term best interests.
That has been the reality in the market for over a decade, becoming
most acute since 2008. This is the connection from the market to the
economy and the country as a whole, and it demonstrates how changes in
market structure are affecting nonmarket participants. The void in the
IPO market has correlated strongly with the U.S. unemployment rate over
the past decade.
It's important to note that these numbers are before 2012, and the
debacles of the BATS and Facebook IPO's. Those technology ``glitches''
may well end up costing us countless jobs that cannot be realized
because companies are increasingly reluctant to use the public
marketplace to raise funding.
It does not make sense to address a problem before establishing
that one exists. While it is clear that the symptoms of the problem--
increasingly frequent market disruptions, retail investor flight and a
stagnant IPO market--are dramatically demonstrated, it is instructive
to examine the current state of U.S. equity markets. I will seek to
demonstrate that market quality has not been improved by high-frequency
trading, and in fact that the market is more fragile than ever, to the
degree that individual firms, and even individual servers, can have a
disproportionate impact on the entire U.S. equity market and by
extension the global market. I will also demonstrate that the new,
fragmented market is detrimental to long-term investors. Finally I will
propose remedies to address these problems.
Market Quality and High-Frequency Trading
To suggest that regulatory changes are necessary to address
problems in current market structure, it must be demonstrated that
current market structure is not serving the needs of long-term retail
and institutional investors, and that the current regulatory regime is
insufficient to address problems in the market. To begin with, it must
be demonstrated that there has been a material, adverse change in
market structure that must be addressed.
In an interesting paper published in 2010, \3\ Reginald Smith asked
the question ``Is high-frequency trading inducing changes in market
microstructure and dynamics?'' He went on to demonstrate that while
historically, the Hurst exponent \4\ of the stock market has been
measured at 0.5, consistent with Brownian motion and our general
understanding of how the market functions on small time scales, since
2005 and Reg NMS that value has been increasing. While this is
certainly an esoteric statistical discussion, his conclusion is
striking:
---------------------------------------------------------------------------
\3\ Smith, Reginald, ``Is High-Frequency Trading Inducing Changes
in Market Microstructure and Dynamics?'' (June 29, 2010). Available at:
http://arxiv.org/PS_cache/arxiv/pdf/1006/1006.5490v1.pdf
\4\ The Hurst exponent (H) is a measure of self-similarity, the
autocorrelation of a time series. Computed using intraday data, it
measures the tendency of a time series to form clusters or remain
random. Random price oscillation is the cornerstone of our
understanding of the stock market, and how the price discovery process
occurs.
we can clearly demonstrate that HFT is having an increasingly
large impact on the microstructure of equity trading dynamics .
. . this increase becomes most marked, especially in the NYSE
stocks, following the implementation of Reg NMS by the SEC
which led to the boom in HFT . . . volatility in trading
patterns is no longer due to just adverse events but is
becoming an increasingly intrinsic part of trading activity.
Like Internet traffic Leland, et. al. (1994), if HFT trades are
self-similar with H > 0.5, more participants in the market
---------------------------------------------------------------------------
generate more volatility, not more predictable behavior.
This is one of several shocking conclusions that we will examine in
this testimony. It also speaks to the detrimental impact that the
complexity of HFT algorithms, and their nonlinear interactions, has on
market microstructure. These algorithms are complex enough that it
takes seasoned practitioner to understand them just in isolation, and
nobody understands what happens as they interact with each other. The
so-called ``Great Quant Meltdown of 2007'' \5\ from August 7-10, 2007,
was an ominous harbinger of what was to come. Many people aren't even
aware that there was significant turmoil in financial markets in early
August, 2007 because of unexpected behavior of long-short quantitative
equity funds. Many of these funds were holding the same undervalued/
overvalued securities, and suffered great losses when they did not
revert and entered a positive feedback loop of liquidation. A full
discussion of this is outside the scope of this testimony, but it was
an indication that, while they were not HFT strategies, seemingly
simple long/short quantitative strategies were not well understood by
those who wrote them, and they certainly didn't understand the
ramifications of their interactions with each other in the marketplace.
---------------------------------------------------------------------------
\5\ More information on this can be found here: http://
select.nytimes.com/2007/08/18/business/18nocera.html.
---------------------------------------------------------------------------
Many HFT strategies have been poached from other firms, or arrived
at independently by quantitative analysts using identical techniques.
This is one of the main reasons we have seen such a massive investment
in technology. The sophistication of your trading strategy is no longer
a defining characteristic of its success, rather the number of
microseconds that it takes your software to react to a piece of market
data has become one of the most important factors of success in the HFT
industry.
The implications of this are grave. As we see on a nearly regular
basis, algorithms reacting to one another, or to manipulative behavior
by nefarious actors, can exhibit nondeterministic behavior. The ``mini
flash crashes'' that Nanex documents on a regular basis are excellent
examples of this nondeterministic behavior as algorithms enter positive
feedback loops either with themselves or in relation to other
algorithms. Many high-frequency trading strategies rely on correlation
of securities with other assets and use price movements in different
securities to inform the fair-value price of another security, through
high-, medium-, and even low-correlation relationships. On a widescale
basis, this can have unintended and nondeterministic consequences,
including positive feedback loops in which liquidity is rapidly
withdrawn from the market.
Giovanni Cespa and Thierry Foucault wrote a paper in March, 2012,
calling this type of phenomenon an ``Illiquidity Contagion.'' \6\ They
noted that the same type of positive feedback loop I have discussed
here can cause a sudden drop in market liquidity, and shift the market
to a new equilibria characterized by high illiquidity and low price
informativeness. That a crash in the futures market could ``infect''
the equity market is another symptom of the increasing
interconnectedness of global security markets--equity index futures,
commodities, stocks/ETF's and currencies are all now being actively
traded by high-frequency traders. Using the aforementioned historical
correlations has the unfortunate side effect of becoming self-
reinforcing because many HFT models observe the same historical
relationships and often trade assets in this manner. At the limit, when
a positive feedback loop is loosed onto the markets-at-large, you can
have a correlation=1 event--in other words, the Flash Crash.
---------------------------------------------------------------------------
\6\ Cespa, Giovanni, and Foucault, Thierry, ``Illiquidity
Contagion and Liquidity Crashes'' (May 8, 2012). Available at SSRN:
http://ssrn.com/abstract=1804351.
---------------------------------------------------------------------------
The increase in correlations is a well-documented characteristic of
the new electronic marketplace. Correlations between stocks are as high
as they have ever been, including during the crisis in 2008, as shown
in this chart from Goldman Sachs:
According to JPMorgan, this ``new normal'' of persistently high
correlations is driven by the macroeconomic environment, the increased
use of ETFs and index futures, and high-frequency trading. \7\ With the
stock market moving in lock-step, this means that the traditional
benefits of diversification that retail investors have relied on for
decades are no longer there to protect them from dramatic moves in
stock market indices. It also means that the market becomes a poor
indicator of company value and performance, undermining one of its core
functions. In addition, the increases in volatility and correlation
drive options prices higher, which increase the hedging costs to
businesses that pass those costs on to their customers. Furthermore,
any increase in stock market volatility has a ``negative wealth
effect.'' Retail investors are notoriously poor at timing market moves,
and have a tendency to buy when the market is high and sell when it's
low. Increasing volatility exacerbates this behavior and costs retail
investors money.
---------------------------------------------------------------------------
\7\ JPMorgan, ``Why We Have a Correlation Bubble'', http://
www.cboe.com/Institutional/JPMDerivativesThemesCorrelation.pdf, October
5, 2010.
---------------------------------------------------------------------------
It is clear that the structure of the market has changed
dramatically since the advent of high-frequency trading, but that is
not necessarily a bad thing. One could argue that the previous market
structure was inefficient; the specialist model extracted unreasonably
high rents for the service they provided, and that the benefit of
tighter spreads in this new electronic marketplace outweighs the costs
of a more highly correlated market. If indeed tighter spreads and lower
volatility characterize this new, more efficient marketplace, there is
a case to be made that the value far outweighs the cost. We are
therefore confronted with these questions:
1. Are spreads tighter, and tightening? Is volatility lower, and
continuing to drop? Is the price discovery process efficient?
Is the price impact of large trades acceptable? Are companies
comfortable accessing capital by going public? Are regulators
comfortable in their understanding of market mechanics, and
able to effectively write new rules and enforce existing ones?
2. Is this the best we can do, or is there a market structure under
which total transaction costs to investors are even lower? Is
there a market structure that is more stable, and one that will
instill confidence for institutional and retail investors?
The traditional mantra of the high-frequency trading industry is
that HFT has helped to decrease trading costs by providing tighter
spreads and lower volatility. One of the oft-cited studies in support
of this claim was authored by employees of RGM Advisors, LLC, a
prominent HFT firm. Another study was done by an employee of Credit
Suisse, a major proponent of HFT. However, an increasing number of
independent academic papers have demonstrated the opposite:
Watson, Van Ness, and Van Ness (2012) using Dash-5 data
find that the average bid-ask spread from 2001-2005 was $0.022
(2.2 cents), while the average bid-ask spread from 2006-2010
was $0.027 (2.7 cents), a dramatic increase of 22.7 percent. In
addition they document increasing volatility as measured by the
standard deviation of the price of stocks, from 0.13 for 2001-
2005 to 0.161 for 2006-2010, an increase of 23.85 percent.
Zhang (2010) found that ``high-frequency trading is
positively correlated with stock price volatility after
controlling for firm fundamental volatility and other exogenous
determinants of volatility. The positive correlation is
stronger among the top 3,000 stocks in market capitalization
and among stocks with high institutional holdings. The positive
correlation is also stronger during periods of high market
uncertainty.'' \8\ Zhang finds that estimated HFT trading
volume is 78 percent, explaining that ``78 percent is clearly
excessive if HFT is meant to provide liquidity. If HFT were to
provide all of the market's liquidity, the volume of HFT would
still be at most 50 percent.'' \9\
---------------------------------------------------------------------------
\8\ Zhang, Frank, ``High-Frequency Trading, Stock Volatility, and
Price Discovery'' (December 2010). Available at SSRN: http://ssrn.com/
abstract=1691679.
\9\ Ibid.
Cartea and Penalva (2011) \10\ examine the impact of HFT on
financial markets using a model with three types of traders:
liquidity traders, market makers, and high frequency traders.
The finding of their model is that high frequency traders
increase the price impact of liquidity trades, increasing
(decreasing) the price at which liquidity traders buy (sell).
These costs increase with the size of the trade suggesting that
large liquidity traders (i.e., large institutional traders
making sizable changes to their portfolio) will be most
affected by HFT. The authors also propose that HFT increases
price volatility and doubles volume.
---------------------------------------------------------------------------
\10\ Cartea, Alvaro, and Penalva, Jose, ``Where Is the Value in
High Frequency Trading?'' (December 21, 2011). Available at SSRN:
http://ssrn.com/abstract=1712765.
Kirilenko, Kyle, Samadi, and Tuzun (2010) \11\ examine the
behavior of high frequency traders in E-mini S&P 500 futures
contracts during the events surrounding the flash crash. HFT
patterns surrounding the flash crash are inconsistent with
traditional market making. They conclude that while high
frequency traders may not have caused the flash crash, their
response to the high selling pressure exacerbated volatility.
(This is certainly consistent with my experience on the trading
floor that day.)
---------------------------------------------------------------------------
\11\ Kirilenko, Andrei A., Kyle, Albert S., Samadi, Mehrdad, and
Tuzun, Tugkan, ``The Flash Crash: The Impact of High Frequency Trading
on an Electronic Market'' (May 26, 2011). Available at SSRN: http://
ssrn.com/abstract=1686004.
Mao Ye, Chen Yao, and Jiading Gai (2012) \12\ point out two
effects of high frequency trading: ``First, it may enable
investors to seize trading opportunities. Second, it may
increase adverse selection problem for slow traders and
generate negative externality. Our results indicate that the
second effect dominates in the sub-millisecond environment.
Also, an increase in the cancellation ratio or message flow
also creates another negative externality. Stock exchanges need
to continuously upgrade trading systems to accommodate more
message flow. These costs, finally, are covered by fees from
traders. However, the current fee structure only charges
trades, not cancellations. Therefore, cancellation actually
creates an externality for traders with true intentions to
trade, who subsidize the traders with lots of cancellations.''
---------------------------------------------------------------------------
\12\ Ye, Mao, Yao, Chen, and Gai, Jiading, ``The Externality of
High Frequency Trading'' (August 31, 2012). Available at SSRN: http://
ssrn.com/abstract=206683.
From the industry, Morgan Stanley concluded in a recent report that
institutional orders are having a much larger impact on asset prices
now than in the period before 2007. They found that the maximum
percentage of average daily volume that a Volume-Weighted Average Price
trade (a very common institutional trading strategy) can handle without
adverse price impact has declined from 10-15 percent to around 4-5
percent now. \13\ They believe the primary reason for this is that the
percent of volume attributable to natural buyers and sellers has
declined by 40 percent when comparing the volume during 2001-2006 to
that since 2008.
---------------------------------------------------------------------------
\13\ Crow, Charles, and Emrich, Simon, `` `Real' Trading Volume,
Morgan Stanley Quantitative and Derivative Strategies Group'', April
11, 2012.
---------------------------------------------------------------------------
These studies show a clear, detrimental impact to spreads, price
impact and volatility. Because there are other studies pointed to by
the industry that enable them to claim the opposite, it is instructive
to not only examine the academic literature and studies, but to look at
the ultimate result. It is clear that long-term investors do not trust
the market. The Morgan Stanley paper cited above concludes that the
volume of trading attributable to institutional traders dropped from 47
percent in the period 2001-2006 to 29 percent since 2008, a decline of
40 percent. \14\ It is clear that long-term investors are fleeing
equity markets at unprecedented rates. It should also be clear that
there are severe inefficiencies in the current market structure, and
that indeed these inefficiencies are structural--they are not going
away without structural changes. The HFT industry continues to make
billions of dollars each year by exploiting these structural
inefficiencies. While there have recently been marginal declining
returns to scale, it has not stopped the performance and technology
race, which the HFT industry realizes is the only way that they can
differentiate themselves from one another.
---------------------------------------------------------------------------
\14\ Ibid.
---------------------------------------------------------------------------
Structural Inefficiencies Have Created a Fragile Marketplace
The new electronic marketplace has several structural
inefficiencies. These are what have permitted HFT to become a
destructive force in the market, rather than a passive liquidity
providing mechanism. This should not be construed to say that all HFT
is bad, but there are 2 important points to make--the structural
inefficiencies present in the market have created a massive
misallocation of resources into technology that provides no social
benefit, and structural deficiencies in market structure have allowed
for nefarious or accidental actions to disrupt the market.
The overriding aspect in the current market that we should fear
most is the inordinate impact that a single market participant, or even
a single server, can have on the market-at-large. This is a grave
concern from any perspective. Whether we are dealing with nefarious,
predatory behavior such as quote stuffing and pinging, or simply
accidental mistakes such as the Knight Capital fiasco appears to be,
one thing should be clear: The market has become more fragile than we
should expect or accept given the tremendous advances in technology
over the past decades.
These structural inefficiencies have been created by the unintended
consequences of regulations:
The approval of the maker/taker model and Pay-For-Order-
Flow deals.
The disappearance of affirmative market-making obligations.
The fragmentation of equity markets into lit and unlit
market centers, with little regulation of the unlit centers.
The rubber-stamp approval of exotic order types without
proper study or justification.
The unbridled latency race to zero without concern over the
impact on markets, and the massive investment in technology
required to keep pace.
While proponents of the existing market structure will argue that
they must be given a free market to allow capitalism and competition to
fix any inefficiency, they ignore the fact that many of the
inefficiencies have been created by regulation and must therefore by
remedied by regulation. They must also be reminded of the consequences
of unfettered, unregulated industries with negative externalities whose
costs are not borne by the producers. In much the same way that
polluting enterprises have to be regulated so that they bear the cost
of negative externalities, so must HFT firms.
In a ground-breaking study done at the University of Illinois at
Urbana-Champaign, Mao Ye, Chen Yao, and Jiading Gai demonstrated the
following: \15\
---------------------------------------------------------------------------
\15\ Ye, Mao, Yao, Chen, and Gai, Jiading,`` The Externality of
High Frequency Trading'' (August 31, 2012). Available at SSRN: http://
ssrn.com/abstract=206683.
While aggressive investment in new technology to reduce
latency has obvious benefits for market participants making
that investment, it is unclear whether there is an associated
social benefit when negative externalities are properly
---------------------------------------------------------------------------
accounted for.
They examined two consecutive technological shocks that
decreased latency from microseconds to nanoseconds. They found
that those shocks ``drastically increase both the trading speed
and the cancellation ratio, which escalates from 26:1 to 32:1.
However, there is no impact on trading volume, spread, depth,
or price efficiency.'' \16\
---------------------------------------------------------------------------
\16\ Ibid.
They go to conclude that ``high-frequency trading may cause
an adverse selection problem for slow traders'' and that this
effect ``dominates at the sub-millisecond level.'' \17\
---------------------------------------------------------------------------
\17\ Ibid.
They also demonstrate conclusive evidence for quote
stuffing, a practice in which the infrastructure of data feeds
is manipulated in order to slow down traders with inferior
technology and to take advantage of this. They shockingly
demonstrate ``clear evidence of comovement of message flow for
stocks in the same channel through factor regression. This
result is consistent with quote stuffing, because message flow
of a stock slows down the trading of the stock in the same
channel, but does not have the same effect on stocks in a
different channel.'' \18\ This claim is reinforced by Egginton,
Van Ness and Van Ness who ``find that quote stuffing is
pervasive with several hundred events occurring each trading
day and that over 74 percent of U.S. listed equity securities
experience at least one episode during 2010.'' \19\ Why is this
such a shocking conclusion? Stock symbol distribution across
channels is alphabetic--essentially random. You would expect
comovement of message flow across industries or correlated
stocks, but not based on an alphabetic distribution. This is
clear evidence of market manipulation.
---------------------------------------------------------------------------
\18\ Ibid.
\19\ Egginton, Jared F., Van Ness, Bonnie F., and Van Ness, Robert
A., ``Quote Stuffing'' (March 15, 2012). Available at SSRN: http://
ssrn.com/abstract=1958281.
Finally they demonstrate that 50ms would be a reasonable
``speed limit'' for minimum quote life, as they find that 30-40
percent of orders are canceled within 50 milliseconds and they
have a trivial if any contribution to liquidity. The
elimination of these orders could have a $0.0000378 increase in
quoted spread, or 0.5 share decrease in depth within 10 cents
of the best bid and ask. In addition, limit order books without
these orders have the same variance ratio (a measure of price
---------------------------------------------------------------------------
discovery and efficiency) as order books with these orders.
It seems that in the latency race to zero, it is those participants
who are not engaged in this race that are paying the unintended costs.
Market data feed volume has risen exponentially in recent years, and
increasing technology investments are being borne by retail and
institutional investors in the fees that they must pay to trade. The
genesis of this problem was the maker-taker business model of the for-
profit exchange/ECN/ATS. This business model is designed to compensate
those who provide liquidity by charging those who take it. Almost every
single exchange now makes money on the spread between the rebate paid
to liquidity providers (previously market makers, now HFT firms) and
the fees charged to institutional and retail investors to take
liquidity. This has created a tremendous conflict-of-interest for
exchanges, especially now that they are publicly traded and beholden to
shareholders (at least in the case of Nasdaq, NYSE, and eventually
BATS). They make money through volume and churn, and have little
incentive to maintain fair, orderly markets.
Another consideration with the maker-taker model is the complete
lack of viable alternatives. Exchanges have had ``glitches'' and blow
ups, notably BATS and Nasdaq during high-profile IPO's. Yet their
market share has not suffered because, frankly, there are no viable
alternatives. This is not to say that there's no hope. Other efforts
are underway to provide an alternative business model, of which I am
devoting my expertise and time to one of those potential alternatives.
In our current environment, a two-pronged approach is critical:
regulators must address structural inefficiencies, transparency, level
playing field, antifraud and related matters, while allowing private
industry's innovation and developments to operate in a fair and open
marketplace.
The increasing fragmentation of the marketplace and the advent of
pay-for-order-flow deals have led to a phenomenon called adverse
selection. This means that profitable trades (from a market-making
perspective) never reach the market. Retail and institutional order
flow pass through a gauntlet of internalizers and high-frequency
trading desks, which pick off any profitable order flow before it ever
reaches the public, lit market. While these orders are filled within
the NBBO, meaning that the originator of the order is no worse off on
that particular order, market quality as a whole suffers. Natural
buyers and sellers are virtually nonexistent under this structure, and
the majority of the volume on the exchanges becomes ``toxic flow'' an
industry term for orders that nobody wants to interact with. The end-
product is consistent with the Morgan Stanley report cited above,
although for different reasons than the authors of the report would
point to. Other countries mandate significant price improvement to
internalize order flow, but the SEC has not.
Dan Weaver, a professor out of Rutgers University has designed a
research study to measure what the impact of dark pools and extreme
fragmentation has been on the lit venues. He has found that trading in
off-market venues such as dark pools directly impacts the spread in the
lit markets, and that the cost of trading is 1.28 cents higher in NYSE
lit markets because of this. \20\ Weaver goes on to say that ``as the
percentage of internalization increases, average trades will have an
increasing impact on prices. Finally, for all market segments, higher
levels of internalization are associated with higher levels of return
volatility.'' \21\ And in fact, in his literature review, Weaver
reviews two studies that demonstrate that ``the internalizing of
uninformed order flow by discriminating dealers reduces the number of
uninformed orders for the nondiscriminating dealers to spread their
informed losses over. The result of this is a widening of spread
charged by the nondiscriminating dealer.'' \22\
---------------------------------------------------------------------------
\20\ Weaver, Daniel G., ``Internalization and Market Quality in a
Fragmented Market Structure'' (May 19, 2011). Available at SSRN: http:/
/ssrn.com/abstract=1846470.
\21\ Ibid.
\22\ Ibid.
---------------------------------------------------------------------------
The current level of fragmentation and complexity is helpful to HFT
firms. They are the masters of the complexity, some of the few actors
in the market that understand the relationships and interconnectedness
and realize high returns based on arbitrage. They also use these unlit
venues as signaling mechanisms, and many of these venues cater to them,
for a price.
Adding to the complexity in the current market structure is the
proliferation of exotic order types. While the SEC deliberates
extensively on any rule change and regulation, the hurdle is far lower
for new order types. It is extremely rare to have the SEC refuse an
exchange's request for a new order type, regardless of whether the
order type does anything to further price discovery or make markets
more efficient.
Many of the newest order types appear to have been designed by the
HFT firms themselves, with little to no utility outside of their
automated strategies. It remains completely unclear what social utility
comes of hidden midpoint pegs, sliding, hide-and-slide, post-only, PNP,
PL select and the rest of the alphabet soup of order types. Even
sophisticated sell-side algorithmic trading desks rarely use anything
other than a limit order, an order for the opening/closing auction and
maybe a midpoint or peg order. HFT firms thrive on this contrived,
structural complexity. They make it their business to understand these
order types and how best to exploit them.
What Can Be Done?
The current structure of the U.S. equity markets is demonstrably
unwieldy, overly complex, and extremely fragile. It is subject to
manipulation, whether nefarious or accidental, on a daily basis.
Spreads are no longer tightening and volatility is no longer dropping.
The price impact of large, institutional orders is rising.
Technological mayhem is more frequent and likely to increase. These
events are not technology ``glitches'' and ``bugs,'' as the industry
and its allies like to dismissively refer to them as, because they
wreak havoc on the market in multiple material ways. It is simply a
matter of time before we have another catastrophe of the same magnitude
or worse than the Flash Crash. The next time it happens, we may not be
so fortunate with regard to the timing--it was only luck that the Flash
Crash didn't start in the morning, inciting markets around the world to
crash, or at 3:45 p.m. EST, with the market closing after the drop, but
before it could recover. If this were to happen, there would be an
overnight exodus from the market with disastrous consequences for the
U.S. economy.
In addition, until confidence in markets is restored, retail
investors will continue to stay away, regardless of the returns they're
missing (as has been shown over the last 2 years), and companies will
hesitate to go public costing untold jobs.
I'd like to start with a proposal for some concrete steps that
regulators can take to address much of the instability and unfairness
in capital markets:
1. Unify trading rules, regardless of venue--exchange, ATS, ECN and
dark pools should all abide by the same general rules. Require
substantial price improvement to internalize flow. This means
more than $0.001.
2. To receive a rebate from any venue on which securities are
transacted, a market participant must be a genuine registered
``market maker'' subject to affirmative market making
obligations. All such rebates or other compensation must be
disclosed.
3. Mandate a unique identifier for every supervisory individual.
This ID would have to be attached to every quote submitted to
any venue, and provide a mechanism for regulators to associate
quotes with the supervisory individual on the trading desk.
4. Eliminate pay-for-order-flow practices.
5. Establish strong, clear market technology standards and regularly
audit firms to ensure they are being followed.
6. Revoke order type approval for order types that do not have a
clearly demonstrated utility to long-term investors and market
stability.
It cannot be legitimately denied that the fragmentation of the
equity market has added unnecessary complexity and created structural
inefficiencies. A change of mentality is required from a regulatory
point-of-view: from the view that there are market centers to be
regulated to the view that there is a marketplace to be regulated,
independent of individual market centers. Reg NMS (National Market
System) was a first step down this path, but the mentality must be
embraced at every level.
Rules and regulations should apply to all actors and all venues
where security transactions are taking place. All of the ideas referred
to here are guided by this principle. In addition, much greater
coordination and visibility must be achieved across asset classes. HFT
firms trade equities, futures, FX and treasuries without blinking. It's
near impossible to regulate the industry without a cross asset-class
viewpoint.
As mentioned before, the maker-taker business model leads to skewed
incentives for exchanges and ``good times'' for liquidity providers. A
simple change could have a dramatic impact on the quality of liquidity
in the market and the level of volatility: any venue that offers any
type of rebate for liquidity provision should be required to tie that
rebate to affirmative market-making obligations. According to a review
by Charitou and Panayides (2006), many markets around the world have
embraced a form of this model, including the Toronto Stock Exchange,
the London Stock Exchange, the Deutsche Bourse, Euronext, and the main
stock markets of Sweden, Spain, Italy, Greece, Denmark, Austria,
Finland, Norway, and Switzerland. All of them designate market makers
with affirmative obligations to supply liquidity for at least some
stocks. The London Stock Exchange requires market makers to register
and to maintain quotes within a maximum spread band, and with minimum
size.
Another study by Bessembinder, Hao, and Lemmon (2011) concluded the
following about the maximum spread model: \23\
---------------------------------------------------------------------------
\23\ Bessembinder, Hendrik (Hank), Hao, Jia, and Lemmon, Michael
L.,`` Why Designate Market Makers? Affirmative Obligations and Market
Quality'' (June 2011). Available at SSRN: http://ssrn.com/
abstract=989061.
A maximum spread rule will lead to a narrowing of bid-ask
spreads, not a widening, and that will lead ``to increased
trading, which can improve allocative efficiency in the
---------------------------------------------------------------------------
presence of information-based externalities.''
This type of rule ``can improve social welfare'' as long as
the ``spread is not constrained to be less than the real
friction, i.e., the social cost of completing trades.''
Such a rule can also increase the speed of price discovery
by encouraging more trading by both informed and uninformed
traders.
Most importantly, they show that ``future flash crashes can
be potentially avoided, and economic efficiency enhanced, by
agreements calling for one or more designated market makers to
continue to provide liquidity during periods of enhanced
information asymmetries. While the DMMs would need to be
compensated for their losses suffered at such times, the social
gains from trade would exceed the costs.''
There is no doubt that many if not all of these ideas are
controversial. However, historically ideas such as these have not been
so controversial. It's not clear what mechanism can be used to
compensate DMMs for losses during severe events. It should be clear
that it is in the public interest to have DMMs provide liquidity during
high-stress events, so it is worth thinking creatively about whether
the public or individual exchanges can help to backstop liquidity
providers during these events. There is no doubt, however, that
liquidity providers are making substantial profits, but without
affirmative obligations to maintain markets they are shifting much of
the risk to the public. If the market is truly a public good then it is
incumbent on regulators to ensure that it remains stable, fair and
orderly, rather than ensuring that exchanges continue to profit from
the maker-taker model.
The SEC should also maintain a database of supervisory individuals
and assign a unique identifier to each of them. Any venue that accepts
incoming orders can mandate the inclusion of another field that would
contain the unique ID of the supervisory individual. While this
information would not be propagated publicly, it would be available to
exchange surveillance teams and to regulators. This would have a two-
pronged effect. It would allow rapid identification of individuals
responsible for aberrant order flow, and provide the ability to quickly
contact them and find out what is happening. In addition, it would
remove the cloak of anonymity that participants currently enjoy, and
thereby act as a deterrent to predatory behavior such as quote
stuffing, stop-hunts and other manipulative behavior. The technological
effort to implement this is not trivial, but it is not complex either.
As someone who has worked with these applications, protocols and
connections for nearly a decade, I can tell you that it could be
implemented in a month or less if so mandated.
The practice of selling retail order flow, what is commonly
referred to as Pay-For-Order-Flow, should be ended. It exacerbates the
problem of adverse selection discussed earlier, and removes natural
buyers and sellers from the market. It is having a negative impact on
market quality, with no benefit other than to the firms selling their
order flow and to those firms able to pick out the lucrative orders
before sending the toxic ones to market.
Finally, the bar for order type approval should be raised to be
similar to any regulatory or rule change. The SEC demands evidence that
any changes benefit the market, and should do the same with exotic
order types. As Scott Patterson demonstrates in his book, Dark Pools,
these order types are having a poorly understood impact on markets and
long-term investors. Many order type approvals should be revoked, and
the burden of proof placed on the exchange to demonstrate their social
utility.
Additionally, I believe the SEC should consider some more creative,
novel ideas for limited implementations or pilots to assess their
efficacy and actual impact on the market. These ideas include:
A 50ms minimum quote life/time-in-force
Open up access to historical and current market data, and
incentivize programmers to help design better tools for
regulators.
Greatly increase the SEC's technology capabilities. This
means a substantial investment in technology and personnel, and
creative thinking about market-wide surveillance.
The SEC should consider a 50ms minimum quote life, at least on a
pilot basis, to observe what the actual impact would be. With so much
evidence against the utility of ``fleeting orders'' it is at least
worth considering. This is the type of effort that could be rolled out
quickly on a preliminary basis, in order to examine what the impact
actually is. Not only would removing those orders help reduce the
technology burden to firms to participate in markets, but the
requirement to stand by a quote, at least for one-sixth of the time it
takes to blink an eye, could help to change some of the behavior of HFT
strategies. Claims that this would widen spreads are unsubstantiated,
and must not be accepted until proven in a pilot program. One
reasonable concern is the cross-asset nature of securities, and
therefore in a broader rollout, such a program should be considered
within a cross-asset perspective and effort.
Finally, a dramatic change in how market data and surveillance are
viewed should be considered. The Internet and Open Source efforts have
taught us that open systems are nearly always preferable to closed. In
that spirit, and under the premise that markets are a public good,
market data feeds and tick data history should be opened up. It is
critical to understand that many academic papers are skewed because
they are either funded directly by the industry, or provided access to
expensive and proprietary data by the industry. Opening up access to
this data would have a dramatic effect.
Access to the historical data of direct market data feeds should be
made available freely to the public, and a prize-based incentive
created for those who can find innovative ways of designing
surveillance systems and algorithms. While the exchanges will surely
argue vigorously against this idea as market data is a major profit
engine for them, it is in the public's interest for the regulation and
enforcement to move out of the 20th century.
The SEC should also consider implementing a market-wide
surveillance mechanism. In this new interconnected market, individual
surveillance groups at market centers are not sufficient. The SEC must
build sophisticated surveillance capabilities. They have taken a first
step, although unfortunately a very inexplicable one. They have
contracted with a prominent HFT firm to build a ticker plant, the first
step towards processing and storing market-wide tick data. This is
reminiscent of the fox guarding the hen house. This HFT firm is not in
the business of building ticker plants. They are in the business of
making money trading. They would not have taken this on if they did not
believe it was in the long-term best interests of their trading profit
center.
In addition, there are a multitude of vendors who do this all day
everyday--why one of them was not chosen for this contract is
inexplicable. The SEC should consider canceling the contract with the
HFT firm or just contract with a technology vendor, just as any other
firm would. The SEC should also hire quantitative researchers, and
compensate them as close as possible to industry rates, along with
bonuses based on a percentage of the fines that they are able to
uncover through data analysis. The SEC needs to attract top-line
talent, and to do so they must be able to compensate competitively.
This is an arms race that Wall Street is winning easily right now.
While the SEC has obvious budgetary constraints, they can get creative
about bonuses, potentially offering percent of fines, similar to their
whistleblower program.
The quantitative data analysis will be made much more robust with
the unique identifiers associated with supervisory individuals
mentioned earlier. The SEC should work very closely with individual
surveillance teams at the exchanges, leveraging best-of-breed ideas,
and helping those that are behind to catch-up. A market-wide
surveillance system could eventually be expanded into a robust set of
technology standards, tripwires/speed bumps, and other mechanisms for
quickly detecting aberrant or nefarious behavior and immediately
throttling or cutting off the offending firm's market access.
Many of the ideas referenced here are not popular within the HFT
industry. Whenever an idea is proposed that they do not agree with,
they respond that there is no way they can operate under such rules,
they will go out of business, spreads will blow out, the market will
cease to function, and generally the sky will come crashing down and
life will end as we know it. I have worked with these folks for years
now, and I must have a much higher opinion of their capabilities than
they do. These are some of the smartest people in the world, and they
will figure out how to continue to make money, and compete, albeit on a
more level playing field. Some will go out of business, as is the
nature of capitalism, and some will thrive. In the end our markets will
be much more stable, resilient and effective price discovery and
capital formation mechanisms, and confidence will be restored to the
retail and institutional investors.
______
PREPARED STATEMENT OF ANDREW BROOKS
Head of U.S. Equity Trading, T. Rowe Price
September 20, 2012
Introduction
Chairman Reed, Ranking Member Crapo, and distinguished Members of
the Senate Subcommittee on Securities, Insurance, and Investment, thank
you for the opportunity to testify today on behalf of T. Rowe Price \1\
regarding the effects of recent significant changes in trading
technology and practices on market stability. My name is Andrew (Andy)
M. Brooks. I am Vice President and Head of U.S. Equity Trading of T.
Rowe Price Associates, Inc. I joined the firm in 1980 as an equity
trader and assumed my current role in 1992. This is my 33rd year on the
T. Rowe Price trading desk.
---------------------------------------------------------------------------
\1\ T. Rowe Price Associates, Inc., a wholly owned subsidiary of
T. Rowe Price Group, Inc., together with its advisory affiliates
(collectively, ``T. Rowe Price''), had $541.7 billion of assets under
management as of June 30, 2012. T. Rowe Price has a diverse, global
client base, including institutional separate accounts;
T. Rowe Price sponsored and sub-advised mutual funds, and high net
worth individuals. The T. Rowe Price group of advisers includes T. Rowe
Price Associates, Inc., T. Rowe Price International Ltd, T. Rowe Price
Hong Kong Limited, T. Rowe Price Singapore Ltd., T. Rowe Price
(Canada), Inc., and T. Rowe Price Advisory Services, Inc.
---------------------------------------------------------------------------
T. Rowe Price is celebrating its 75th year of advising clients. We
are a Baltimore-based global adviser with over $540 billion in assets
under management as of June 30th, 2012, and more than 3 million client
accounts. We serve both institutional and individual investors.
We welcome the opportunity for discussion regarding the industry
and market practices.
Our firm is particularly focused on the interests of long-term
investors. We appreciate the role other types of investors can have in
creating a dynamic marketplace. However, as we talk with our clients,
there is a growing distrust of the casino-like environment that the
marketplace has developed over the past decade. We worry that the
erosion of investor confidence can undermine our capital markets, which
are so important to the economy, job growth, and global
competitiveness. Reaffirming a strongly rooted commitment to fairness
and stability of the market's infrastructure is critically important.
Over the past two decades the markets have benefited from
innovation from new technology and competition. Generally, markets open
on time, close on time, and trades settle. However, there are problems
below the surface.
Here are some things we find concerning:
Order Routing Practices
We question the nature of various order routing practices. The
maker-taker model, payment for order flow, and internalization of
orders all seem to present a challenge to order-routing protocols. Are
order routing practices and incentives an impediment to the overarching
requirement to seek best execution on all trades?
Colocation/Market Data Arbitrage
We believe that the widespread use of colocation creates an uneven
playing field that favors those who can and will pay for it. We
question whether this has produced a market that values speed over fair
access. In no other regulated industry is one party allowed a head
start in exchange for payment. Our understanding is that current
colocation practices allow for a market-data arbitrage where some
investors get quotations and trade data faster than others. This
advantage is traded upon, causing some participants to believe they are
victims of front-running or are at least disadvantaged.
Speed and Impact on Market Integrity
Our sense is that the almost myopic quest for speed has threatened
the very market itself. It also seems many high frequency trading (HFT)
strategies are designed to initiate an order to simply gauge the
market's reaction and then quickly react and transact faster than other
investors can. This seems inherently wrong. Our understanding is that
the continued push for speed is not producing any marginal benefit to
investors and in fact may be detrimental. This pursuit of speed as a
priority is in direct conflict with the pursuit of market integrity as
a priority.
Inaccessible Quotes and High Cancellation Rates
The growth of HFT has lead to increased volume; however, whether
the corresponding volume is ``good'' or ``bad'' deserves analysis.
Volume does not necessarily mean liquidity for large institutional
investors. When you combine high HFT volumes and even higher
cancellation rates, these forces can combine to undermine market
integrity and cause deterioration in the quality and depth of the order
book. We feel that this volume is transitory and misleading.
Challenges to the National Market System (Regulation NMS)
We believe the original construct of Regulation NMS was laudable
and designed to encourage competition. However, we do not believe this
regulation contemplated today's highly fragmented marketplace, where we
have 13 different exchanges and over 50 unregulated ``dark pools.'' In
such a fragmented market, can one really be confident in achieving best
execution given the explosion of market data traffic? We question the
markets' ability to process the overload of market data.
Conflicts of Interest
We question whether the functional roles of an exchange and a
broker-dealer have become blurred over the years and could warrant
regulatory guidance regarding the inherent conflicts of interest. It
seems clear that since the Exchanges have migrated to ``for-profit''
entities, a conflict has arisen between seeking volume to grow revenues
and their obligation to assure an orderly marketplace for all
investors.
HFT Trading Strategies
Professional and proprietary traders often have divergent interests
from those of investors concerned about the long-term. When the average
holding period for such traders is measured in seconds as opposed to
months or years, have we destabilized the market. Given recent market
volatility, more study is warranted to assess the impact of the
exponential growth of short-term trading strategies. Most rules and
regulations seem to further enable those with short term profit
incentives as evidenced by the proliferation of new order types
suggested by exchanges and approved by regulators.
Suggestions
We believe it is time is to step back and examine market structure
and how it impacts all investors. A good first step might be to
experiment with a number of pilot programs to examine different
structural and rule modifications. We suggest a look at the
appropriateness of colocation as a general practice and enhanced
oversight of high frequency trading and other strategies that might be
unduly burdensome to overall market functionality. We would like to see
a pilot program where all payments for order flow, maker-taker fees,
and other inducements for order flow routing are eliminated. We
envision a pilot where there are wider minimum spreads and mandated
time for quotes to be displayed to render them truly accessible. These
programs can include a spectrum of stocks across market caps and
average trading volumes, among other factors. We also suggest a pilot
program of imposing cancellation fees for unacceptable trade to
cancellation ratios. A key question is should we foster consolidation
in this fragmented market? At a minimum, should we raise the barrier
for becoming an exchange? In our opinion, requiring a more robust
testing for new software would seem to make sense.
Conclusion
T. Rowe Price appreciates all the efforts of the SEC and Congress
as we strive to make the markets better and fairer for all
participants. The Consolidated Audit Trail, Large Trader ID, limit up/
down initiatives are all improvements. We suggest any regulatory
proposals be aligned with a goal of making the markets simpler, more
transparent, and less focused on speed. We applaud the Committee's
interest in making sure the right questions are asked.
There are currently over 1,000 order types to express your buy and
sell interest and we suggest that a simplified model may be more
efficient for all investors. The issues we face are enormously complex.
We certainly do not have all the answers. We believe that it is time to
revisit the historical responsibility to provide a fair and orderly
market.
PREPARED STATEMENT OF CHRIS CONCANNON
Partner and Executive Vice President, Virtu Financial, LLC
September 20, 2012
Chairman Reed, Ranking Member Crapo, and Members of the
Subcommittee, I want to thank you for the opportunity to appear before
you today. My name is Chris Concannon and I am an Executive Vice
President for Virtu Financial, LLC.
Virtu Financial (``Virtu'' or the ``Company'') is a global
electronic market maker. Virtu is an active market maker on more than
100 markets around the globe. Virtu makes markets from our six offices
in New York, Los Angeles, London, Dublin, Sydney, and Singapore. The
Company's market making activity spans across multiple asset classes,
including cash equities, fixed income, currencies, futures, options,
energy products, metals and other commodities. Virtu, through its
subsidiaries, is directly registered as a broker dealer or investment
firm and operates as a registered market maker on most primary markets
around the globe. In the U.S., Virtu operates two registered broker
dealers that are also registered as market makers or designated market
makers on the NYSE, Nasdaq, BATS Exchange, NYSE Arca, and NYSE MKT. In
Europe, Virtu operates a registered investment firm that is also
registered as a market maker on the London Stock Exchange, the Swiss
Exchange, Euronext, and the Deutsche Bourse Exchange. Obviously, Virtu
believes in the benefits of market making and is committed to providing
continuous, obligated liquidity in the markets we serve.
In discussing the state of the U.S. equity market, I start from the
premise that our equity market is the most dynamic and efficient market
in the world. The U.S. equity market is a special asset that should be
celebrated. Our markets are envied by Nations and financial centers
around the globe. Our U.S. equity market is also the most liquid and
robust pricing mechanism on the planet. My firm trades across all of
the major financial markets and no market can compare to the U.S.
equity market in terms of pricing efficiency and liquidity. Companies
listed on our U.S. markets enjoy the most efficient and liquid market
which contributes to higher returns for their investors. Over the last
4 years, I have witnessed an unprecedented number of claims that our
markets are horribly broken, unfair and dangerous. These claims tend to
be short on facts and evidence, but long on press coverage and book
deals. Our market is not perfect. And it has recently experienced some
dramatic mishaps. But, despite its flaws, it is a market that has
withstood the most unprecedented volatility and repricing of equity
values in our lifetime while maintaining the same levels of pricing
efficiency.
Let me be clear, our market is not perfect. It has flaws and
unnecessary complexity. The U.S. equity market is overly fragmented
and, likely, over engineered. Stocks in the U.S. trade electronically
on 13 national securities exchanges and over 40 dark pools. The current
state of our equity market is not one that we would set out to design
if we did it all over again. The U.S. equity markets began evolving
into a fully electronic market during the 1990s. For the last decade,
our markets have been largely automated. That means every exchange and
every market in the U.S. is a fully automated, electronic destination.
Virtually every order arrives at its intended exchange in electronic
form. The automation that exists in our market today is not a new
phenomenon. Technology has been operating our markets for the last 15
years.
With fragmentation and technology comes complexity. Our market is
one of the most complex securities markets on the planet. It is not
naturally complex. It is complex because of the number of major
regulatory reengineering events that have taken place in the U.S. over
the last 15 years. For example, the list of major market structure rule
changes includes the Limit Order Display Rule, Regulation ATS,
Decimalization, T+3 Settlement Cycle, Regulation NMS, Regulation SHO,
Single Stock Circuit Breakers and, more recently, the Market Access
Rule. Each of these major regulatory reengineering events required
substantial technological enhancements to be delivered by all industry
participants and exchanges. These were not simple software programming
endeavors. These were all major technology projects completed across
the industry.
I would like to focus on three areas that I believe deserve further
review: (1) our choice of a single market structure for all listed
companies; (2) our markets failure to enhance market maker obligations;
and (3) the industry's current risk management standards.
First, our market is currently designed as a ``One-Size-Fits-All''
market. What I mean by this is that most of our major market structure
rules do not distinguish between the size or market capitalization of
the listed company, or the trading characteristic of its stock. Our
markets are designed to execute all stocks, regardless of shape or
size, using the same market mechanism. As the list of public companies
continues to grow, a more diverse number of public companies trade on
our market while subject to the same market structure. A stock that
trades once per day is traded in the same market structure as a stock
that trades one million times per day. Our market is solely designed
for Cisco, Microsoft, and Bank of America and not for a stock that
trades by appointment. I believe we should revisit our current market
structure in order to create a better pricing mechanism for all stocks
of different shapes and sizes. This One-Size-Fits-All approach is
further exacerbated by an expansion of the portfolios of our largest
investors. As institutional holdings expand further into less liquid
stocks, like Russell 2000 stocks, our largest institutions are
struggling to trade in our poorly designed market structure for those
types of stocks.
My second area of focus is on our markets' failure to enhance
market making obligations. While my firm is a market maker and it is
easier for me to call for enhanced market making obligations, I
fundamentally believe that we need to increase obligated liquidity in
our markets. Flash crashes, miniflash crashes and other market
disruptions demonstrate the need for additional obligated liquidity in
our market. However, I believe enhanced market maker obligations should
be targeted where they are most needed and that is in our less liquid
stocks. And so, my earlier point about our flawed, single market
structure should be considered with enhanced market making obligations
as a component of a new market model. New market models for less liquid
stocks should be accompanied with enhanced market maker obligations.
My final area of focus is the industry's current risk management
standards. In light of recent events, I believe that the industry
should explore ways to improve risk management standards. Industry
participants have already identified several areas of risk management
enhancements that should be implemented and could be delivered in short
order. First, pretrade risk management limits are already required by
the Securities and Exchange Commission (SEC) under SEC Rule 15c3-5
(also known as the ``Market Access Rule''). Under the Market Access
Rule, which has been in effect for over a year, firms are required to
establish pretrade credit limits for every customer account and for the
firm's own proprietary account. The credit limits required by the
Market Access Rule must be administered in real-time and at all times.
These credit limits are a firm's primary defense against unwanted
trading activity by the firm or by its client.
In addition, the industry is currently exploring specialized ``Kill
switches'' that would be administered by exchanges. These ``Kill
switches'', as currently being discussed, would provide a systematic
shut-off of a firm if it exceeded prescribed or preset trading limits.
``Kill switches'' would not be a primary defense, but rather, a
secondary defense to back stop the failure of other risk management
measures operated by a firm. Kill switches have operated effectively on
futures exchanges in the U.S. for many years. These same trading limits
could be implemented across all U.S. equity exchanges. Like the futures
exchange limits, firms would be required to establish limits on each
equity exchange. Such a kill switch would have severely limited the
damage done on August 1st of this year.
The last component to enhanced risk management is one of the most
important. We believe a simple feature referred to in the industry as
``drop copies'' should be required as a mandatory risk management tool.
``Drop copies'' are separate and distinct connections offered by
exchanges and other markets. Drop copies, which are widely used by the
industry, provide a real-time echo, or copy, of a firm's trading
activity on a given exchange. Drop copies are primarily used by the
industry to run reconciliations that compare a firm's known trading
activity against what the exchange believes was traded by the firm.
This is commonly referred to as a ``Street vs. House'' comparison. If
such a drop copy comparison is conducted in real-time by systems that
are independent from the firm's trading system, a firm will always have
an accurate assessment of its positions and trading activity, including
both intended and unintended activity (See, Exhibit I).
While I believe firms should have a robust process for developing
and testing new software, the industry must have advanced risk
management systems to limit the risk of unintended trading activity by
a firm or its client. We know with certainty that software has bugs,
hardware crashes and networks go down no matter the robustness of a
firm's development and infrastructure process. The industry must build
risk protections that assume the worst while a robust development and
testing process avoids the worst. Pretrade risk checks, ``Kill
switches'' and real-time drop copies protect us from the worst events.
Thank you again for the opportunity to be here today to speak on
this subject. I would be pleased to answer the Committee's questions.
PREPARED STATEMENT OF LARRY TABB
Founder and Chief Executive Officer, TABB Group
September 20, 2012
RESPONSES TO WRITTEN QUESTIONS OF CHAIRMAN REED
FROM DAVID LAUER
Q.1. Germany is considering legislation that would impose new
rules on high-frequency trading. It is my understanding that
this legislation would require traders to register with
Germany's Federal Financial Supervisory Authority, collect fees
from those who use high-speed trading systems excessively, and
limit the number of orders that may be placed without a
corresponding trade. The new rules would also grant the
regulator the power to compel firms to detail their trading
practices.
Similar European-wide legislation is being considered by
the Committee on Economic and Monetary Affairs of the European
Parliament. These measures include a requirement for orders to
rest on the exchange order book for a minimum of half a second
and the testing of algorithms that allow preprogrammed trading.
Canada began increasing the fees charged to firms that
flood the market with orders earlier this year, with firms
charged for all the orders they cancel, not just the trades
they execute. Additional rules are expected to cut the amount
of trades going to dark pools. Starting in October, the pools
will be allowed to take orders only if they offer a
significantly better price than the public exchange.
What is your view of these various reform proposals? Would
these measures make the U.S. markets more of less fair and
efficient? Would these measures be feasible in the U.S.
markets? Why or why not?
A.1. Adoption of several of these reforms would help make U.S.
markets more fair and efficient. All of them are feasible, but
that does not mean that they are advisable.
I think many of the reform proposals are commonsense ideas
with substantial empirical support. Requiring traders to
register is an excellent idea and one that I continue to
advocate for. This is part of the idea I presented in my
written testimony requiring firms to register trading
strategies and tag all of their orders with a strategy-level
ID. This would deter nefarious activity and help regulators
enforce existing regulations.
I presented studies in my written testimony that attempt to
quantify the negative externalities that HFT generate. Part of
that is from excessive order rates and order cancellation
rates. At the moment the cost of these externalities is not
being borne by the producers. Germany and Canada's move to
change that dynamic with cancellation fees is laudable.
Finally, the Canadian rules are exactly the type of rules I
argued for in my written testimony. Increasing the requirements
for off-exchange execution is critical to restoring natural
liquidity in the lit markets. These requirements should include
substantial price improvement and a minimum execution size.
I would not advocate for the European-wide legislation. I
do not believe a Financial Transaction Tax would have the
intended consequences. We have seen dramatic evidence of
fleeing liquidity in those markets that have adopted this tax,
although the market quality implications of that flight are
still under study and may not be as dire as the liquidity
flight would appear. A minimum resting period of a half second
also seems extreme and not supported by empirical evidence. I
have advocated for a pilot program to test a 50 millisecond
holding period based on academic studies. I now believe that
such a policy would only be advisable if it were done across
all affected asset classes, include equities, options, and
futures.
Q.2. In discussing the use of a ``hide not slide'' order type,
there was testimony that stated that such an order is mainly
used by professional traders and that it is not permitted for a
broker to take an investor's order and slide it. Can anyone
other than a professional trader use this order type? For
example, can a broker-dealer acting on behalf of institutional
investors use a ``hide not slide'' order? Why or why not?
A.2. Anyone can use this order type who has a direct connection
to an exchange, and control over the flags that they are
setting in the order that is sent to the exchange. A broker-
dealer can use this order type, although generally they only
use the simplest order types when sending out orders on behalf
of investors (or rebate-maximizing order types). Customers
would have to specifically request that this order type is
used, and ordinarily they would not do that.
------
RESPONSES TO WRITTEN QUESTIONS OF SENATOR HAGAN
FROM DAVID LAUER
Q.1. High-speed trading is an issue that has received
considerable attention from market regulators throughout the
world.
On September 26, 2012, German Chancellor Angella Merkel's
cabinet approved draft rules that would require high-speed
traders to register with securities regulators and would
mandate that automated orders be labeled as such when submitted
to regulators. Can you evaluate the strengths and weaknesses of
the approach taken by Germany?
A.1. As covered in my written testimony, I believe that
registration of High-Frequency Traders is a simple, critical
first step to regulating their activities. I also believe that
the German proposal does not go far enough, as I would like to
see registration of individual trading strategies with unique
ID's assigned to them. These ID's should be placed in every
order sent by these traders for regulatory tracking purposes,
and also to support the function of enhanced, adaptive kill
switches.
Q.2. In August, the Australian Securities & Investment
Commission (ASIC) proposed reforms to automated share trading
that are intended to prevent market disruptions. The proposal
would require high-speed traders to put in place pretrade
``filters'' and to limit or suspend automated orders that would
``interfere with the efficiency and integrity of the market.''
This mandate is coupled with a fine on traders that could not
trace orders and trading messages. Can you evaluate the
strengths and weaknesses of the approach taken by Australia?
A.2. I believe it is important to unify risk controls across
the industry, and Australia's efforts here are a good first
step. What they are lacking is a centralized means of enforcing
these regulations on a real-time basis, instead relying on
analysis and fines after-the fact. In the U.S. most individual
firms already have some risk controls in place, but regulators
can do a better job of following Australia's example and unify
these requirements. I don't believe that static kill switches
are the answer however, and I don't believe that individual
market centers should have this responsibility. Market activity
cannot be analyzed by individual market centers, and therefore
a centralized approach is necessary. The SEC can take on this
role on a real-time basis, provided these firms are registered
and tagging their orders with unique ID's.
Q.3. In July, Hong Kong's Securities and Futures Commission
(SFC) released a proposal targeting automated trading. The
SFC's proposal would require intermediaries to have
``appropriate policies, procedures and controls . . . when they
conduct electronic trading,'' example include pretrade risk
management controls and post-trade monitoring. Can you evaluate
the strengths and weaknesses of the approach taken by Hong
Kong?
A.3. I believe Hong Kong's proposal suffers from the same
weaknesses as Australia's. Individual firms can have an
inordinate impact on the market, and therefore fining firms
after-the-fact will not help to protect the market during a
stressful event.
Q.4. Given that a wide range of countries have pursued changes
that address computerized trading, how will the competitiveness
of U.S. financial markets be impacted by a failure to adopt
similar reforms?
A.4. The U.S. markets are developing a poor reputation because
of our failure to properly regulate HFT. While some of the
international reforms are knee-jerk reactions, and therefore
inadvisable, many of them are sensible and will help to restore
trust. These include registration of HFT firms, attempts to
quantify negative externalities and shift the cost of those to
the producer via cancellation fees, and increasing the hurdle
for off-exchange executions.
Q.5. In an October, the Federal Reserve Bank of Chicago
released a letter titled ``How to keep markets safe in the era
of high-speed trading.'' The letter outlines several risk
controls that could improve market structure and investor
confidence. Please discuss the costs and benefits of the
following controls:
Intraday position limits.
A.5. This is a reasonable control, and is one most firms have
already implemented.
Q.6. Limits on the number of orders that can be sent to an
exchange within specified period of time.
A.6. This has similar problems to static kill switches. While
it seems like a reasonable approach for a normally operating
market, during times of market stress this could lead to severe
problems.
Q.7. Automated trading is a logical extension of technology in
a competitive market that may provide benefits such as
increased transparency and better execution for market
participants. What would happen to price discovery,
transparency, and execution prices if high-speed trading ceased
or declined?
A.7. It is important to distinguish between technology advances
and high-frequency trading. Generally, most of the benefits to
investors of increased liquidity and tighter spreads came
before 2007 and the advent of high-frequency trading. In fact,
several studies have shown increased spreads over the last
couple of years, and no comprehensive studies have been done on
execution costs overall. Spreads can be a poor proxy for
execution costs, as so much liquidity in the market is fleeting
(up to 40 percent according to one study cited in my written
testimony). That being said, it is clear that in the near-term,
market quality, price discovery and execution prices would all
be harmed if high-speed trading were to cease, as it has become
the primary supplier of liquidity in the market. We should
instead take a measured approach to reducing nefarious
activity, reining in certain practices, and evening the playing
field. Many of the recommendations I make in my written
testimony would be incremental approaches to confronting these
issues, rather than blunt instruments such as a transaction
tax.
Q.8. In the United States the Minimum Price Variation (MPV) for
all stocks over one dollar is one penny. In Europe MPVs are
less uniform.
How do MPVs impact high-speed trading?
A.8. They have little impact on high-speed trading. Computers
are very good at adjusting to this type of variable. They do
help to reduce competition to high-speed traders, as smaller
MPV's in low-priced or less liquid stock make it less
profitable for regular market makers to provide liquidity. The
reason is that when a market maker offers stock in a less
liquid name, with an MPV of $0.01, a high-speed trader can
simply step in front of them. While this may mean the
appearance of a tighter spread, as market makers are driven out
of business, spreads in these names widen again, and market
depth suffers.
Q.9. What factors should be considered when determining the
appropriate MPV?
A.9. An interesting proposal is to simply allow a firm to
choose its own MPV, under the advice of an investment bank or
other fiduciary. Barring this free market approach, market
capitalization and liquidity should be the main factors.
Q.10. Should there be different MPV's for stocks of varying
price or trading volume?
A.10. I agree that a one-size-fits-all approach to MPV has
harmed smaller capitalization and less liquid companies.
Alternatives should be explored and a pilot program would be an
excellent first step.
Q.11. Several witnesses mentioned the proliferation of order
types that are designed for high-speed traders.
How do the order types available to trades in the United
States compare to those available in markets abroad?
A.11. Foreign markets have a much smaller set of order types.
Many order types are a result of Reg NMS requirements, and
therefore other markets do not require such an extravagant set.
Q.12. How do exotic order types advantage or disadvantage
retail investors, institutional investors, exchanges, and high-
speed traders?
A.12. High-frequency traders thrive on complexity. They make it
their business to understand the nuances of every order type,
and the exchanges often cater to them by providing access to
the developers of the logic, or consulting with them when
creating a new order type. HFT firms use exotic order types to
segment order flow so that they don't interact with certain
types of traders, or to play tricks to jump the queue. In
addition, a lack of sophistication by institutional investors
means that they can end up using order types that disadvantage
them in terms of cost or priority.
Q.13. Please list those order types that you believe could be
eliminated without doing harm to markets.
A.13. Post-only, PNP, PL Select, Hide-Not-Slide and the retail
price improvement orders that have recently been approved.
Q.14. The increased fragmentation of exchanges and dark pools
complicates the task of developing a market wide approach to
data aggregation and analysis by regulators.
What type of technology investments could advance the
ability of the Securities and Exchange Commission and the
Commodity Futures Trading Commission to better monitor markets
and develop a market wide approach to data?
A.14. I have developed and presented a full proposal in my
written testimony for a real-time, cross-asset monitoring
system. This strikes me as the bare minimum that the regulators
need to understand and stay on top of the current market. The
Consolidated Audit Trail is important, but will fall far short
of what is needed on a real-time basis. Markets are moving too
quickly for regulators to always be a day or more behind.
Regulators can take a balanced approach, including some
investments in technology and leveraging the existing
investments exchanges have already made. This real-time,
market-wide surveillance system combined with a firm-level or
strategy-level registration system would allow regulators to
finally do their job, and enforce existing regulations
properly.
------
RESPONSES TO WRITTEN QUESTIONS OF CHAIRMAN REED
FROM CHRIS CONCANNON
Q.1. Germany is considering legislation that would impose new
rules on high-frequency trading. It is my understanding that
this legislation would require traders to register with
Germany's Federal Financial Supervisory Authority, collect fees
from those who use high speed trading systems excessively and
limit the number of orders that may be placed without a
corresponding trade. The new rules would also grant the
regulator the power to compel firms to detail their trading
practices.
Similar European-wide legislation is being considered by
the Committee on Economic and Monetary Affairs of the European
Parliament. These measures include a requirement for orders to
rest on the exchange order book for a minimum of half a second
and the testing of algorithms that allow preprogrammed trading.
Canada began increasing the fees charged to firms that
flood the market with orders earlier this year, with firms
charged for all the orders they cancel, not just the trades
they execute. Additional rules are expected to cut the amount
of trades going to dark pools. Starting in October, the pools
will be allowed to take orders only if they offer a
significantly better price than the public exchange.
What is your view of these various reform proposals? Would
these measures make the U.S. markets more or less fair and
efficient? Would these measures be feasible in the U.S.
markets? Why or why not?
A.1. Virtu Financial, LLC, (Virtu) is a global electronic
market maker and active in more than 150 markets around the
world. Virtu actively is actively engaged in dialogues with
regulators around the Globe. Virtu actively supports global
market reforms that include:
1. Registration requirements for active trading firms;
2. The ban on ``Naked-sponsored'' access;
3. Increased capital standards for market makers;
4. Implementation of market access and risk controls;
5. The establishment of real-time credit limits;
6. Limit-up/Limit-down Trading Controls;
7. The establishment of exchange kill switches;
8. Modernization of the U.S. surveillance systems and
mechanism; and
9. Enhanced market making obligations.
However, rules or restrictions which limit technology or
alter micromarket structure should be studied and have a
measurable benefit. Our current market complexity was the
result of micromarket structure changes implemented over the
last 15 years. Those changes were not adequately studied and
were driven by other initiatives.
------
RESPONSES TO WRITTEN QUESTIONS OF SENATOR HAGAN
FROM CHRIS CONCANNON
Q.1. Automated trading is a logical extension of technology in
a competitive market that may provide benefits such as
increased transparency and better execution for market
participants. What would happen to price discovery,
transparency, and execution prices if high-speed trading ceased
or declined?
A.1. Our markets converted to automated markets 15 years ago.
Since that conversion was made, trading costs of declined
materially for all investors large and small. During this
evolutionary change that benefited investors, our market
intermediaries had to change as well. Our markets' market
makers are automated out of the need to survive in an automated
world. The elimination or decline of our electronic liquidity
providers would directly impact price discovery and execution
quality. Investors would ultimately be harmed by introducing
unnecessary frictions into our markets.
Q.2. Several witnesses mentioned the proliferation of order
types that are designed for high-speed traders. How do the
order types available to traders in the United States compare
to those available in markets abroad?
A.2. The implementation of regulatory changes over the last 15
years, such as Regulation ATS and Regulation NMS, has resulting
in greater market fragmentation and increased market
complexity. It is this unique fragmentation and market
complexity that has resulted in the proliferation of order
types in the United States.
Q.3. In the United States the Minimum Price Variation (MPV) for
all stocks over one dollar is one penny. In Europe MPVs are
less uniform.
How do MPVs impact high-speed trading?
A.3. MPVs impact all trading participants. MPVs set the minimum
spread for quoting and trading of an instrument. Therefore,
MPVs determine the price that all investors pay, both large and
small, when entering or exiting the market. An MPV set too high
will introduce substantial costs to investors. An MPV set too
low can lower displayed liquidity and cause the appearance of
increased price volatility.
Q.4. What factors should be considered when determining the
appropriate MPV?
A.4. As I stated in my written testimony, our market is
currently designed as a ``one size fits all'' market. More
regulatory flexibility around MPV and other market structure
features should be considered in order to accommodate the wide
spectrum of publicly traded stocks in our market.
Q.5. Should there be different MPV's for stocks of varying
price or trading volume?
A.5. Yes. ``One size fits all'' market structures are not
appropriate.
Q.6. Several witnesses mentioned the proliferation of order
types that are designed for high-speed traders.
How do the order types available to trades in the United
States compare to those available in markets abroad?
A.6. U.S. markets are highly complex and fragmented as a result
of regulatory changes. Market participants have requested a
variety of order types to address their unique needs to deal
with this highly complex and fragmented market. These order
types are filed with the Securities and Exchange Commission and
approved by the SEC.
Q.7. How do exotic order types advantage or disadvantage retail
investors, institutional investors, exchanges, and high-speed
traders?
A.7. A variety of order types have been designed and made
available to investors, both retail, institutional, and
professional. These order types, which range from simple to
complex, have been designed, delivered, and accessible to meet
the unique needs of these investors.
Q.8. Please list those order types that you believe could be
eliminated without doing harm to markets.
A.8. Because many order types have been designed to meet the
unique needs of investors both large and small, such an
analysis would require the solicitation of views from a very
broad audience to ensure that certain order types are not
mistakenly eliminated.
Q.9. The increased fragmentation of exchanges and dark pools
complicates the task of developing a market wide approach to
data aggregation and analysis by regulators.
What type of technology investments could advance the
ability of the Securities Exchange Commission and the Commodity
Futures Trading Commission to better monitor markets and
develop a market wide approach to data?
A.9. The two agencies seem to recognize that market structure
and technology advancement issues present challenges to
monitoring our modern markets. Consequently, the agencies have
been openly discussing these market surveillance issues with
interested parties including Virtu. We believe the design and
implementation of a multimarket surveillance system is very
achievable and will deliver the long term benefits needed to
protect our markets.
Additional Material Supplied for the Record
STATEMENT OF SENATOR CARL LEVIN
I would like to thank Chairman Jack Reed, Ranking Member Crapo, and
all my colleagues on the Securities, Insurance, and Investment
Subcommittee for holding this hearing.
Millions of American workers and their pension funds have long put
their savings into the U.S. capital markets in hopes of a better
retirement and better future. In recent years, though, a number of
headline-grabbing events have raised serious questions about our
markets in the minds of ordinary Americans, sophisticated investors,
policy makers, and regulators.
Are our markets too fragile and vulnerable to collapse?
Are our markets fair for all investors?
And are investors, who we need to fund the growth and
development of our companies and our economy, fearful about
subjecting their savings to these markets?
Nearly 2 years ago, some of you joined with members of the
Permanent Subcommittee on Investigations, which I chair, in a joint
hearing to examine the efficiency, stability, and integrity of the U.S.
capital markets. At that hearing, held in the wake of the so-called
Flash Crash, experts raised a number of important issues, and we
highlighted a number of potential regulatory improvements.
I am pleased that the SEC took some steps to limit dangerous holes
in our market regulation, such as unlimited, unsupervised direct market
access. And just recently, the SEC finally issued a rule that will
someday create a consolidated audit trail. These and other recent steps
are a good start, but they are little more than Band-Aids. As anyone
who has watched or participated in the markets in recent years knows,
more must be done.
It is time for a broad reexamination of the actual workings of our
markets. When the horse and buggy gave way to the automobile, our
national infrastructure needed an upgrade. During these tumultuous
times, while we as a country learned how to take advantage of this
great new innovation, there were innumerable horrific accidents: cars
running into older horse-drawn vehicles; cars running off the edges of
unpaved roads; and cars simply crashing into one another.
What our country learned is that everything needed an upgrade. The
roads themselves needed upgrades. They needed to be wider and more
stable to bear the heavier traffic. The cars themselves needed to be
safer. And the drivers needed an upgrade. They needed to adopt and
follow ``rules of the road,'' like when to make a left hand turn and
what to do when an emergency vehicle drives by. These upgrades needed
to be enforced. Roads need to be inspected, as did the cars driving on
them. Drivers needed to be tested, and police officers needed to ensure
that drivers followed the rules.
As the cars got faster over the next several decades, the benefits
and needs for upgrades grew as well. We needed seatbelts and airbags.
Roads needed to be wider, straighter, and bridges sturdier. Drivers
needed to be smarter too. Those driving unique vehicles, like tractor
trailers and busses, needed further training. And enforcement also
needed upgrades. Because police officers can't be everywhere at once,
speed cameras and red-light cameras have started to pop up.
The goals of all of these innovations were simple: (1) minimize the
number of crashes, and (2) make the crashes that do occur less harmful.
In recent years, our capital markets have undergone a
transformation that is no less stunning. The old New York Stock
Exchange floor has been hollowed out. The screaming that once typified
the floor is now replaced with the whir of computers. Traditional
market makers have given way to computer trading firms with servers
located right next to the computers of the trading venues.
Computerized traders use automated trading systems to very quickly
place and cancel orders--more than 90 percent are canceled--to make
markets and arbitrage very small price differences between markets. In
this very competitive industry, milliseconds can be the difference
between millions in profits and significant losses. Because of that,
computerized trading firms now pay thousands of dollars a year for the
right to special proprietary data feeds from exchanges and other
execution venues. Many rent space, for another hefty sum, to collocate
their servers at execution venues. Doing this reduces the response time
by a few milliseconds, which is worth millions of dollars to these
firms. Orders to buy or sell stock are diced by computers into small
bits, routed in milliseconds to any of several dozen venues, and
executed in fractions of a second. All while other computerized trading
programs are looking to sniff out those orders, and sneak in front of
them for a profit.
Investors who hold onto stocks for weeks, months, and years
comprise less of the market than ever, giving way to computer trading
programs holding onto stocks for fractions of a second.
It is not surprising that this rapid development has resulted in
more than our share of crashes, the most notable of which was the Flash
Crash. And there are plenty of lessons to be learned from just that
event.
But, since then, numerous other crashes and technological errors
have roiled the markets. Many investors were shocked to see the failure
of the IPO of BATS Global Markets. BATS is itself a major exchange on
which investors trade more than 500 million shares a day. Yet, due to a
computer glitch on BATS's own systems, its IPO went haywire and
ultimately had to be shelved.
Just a few months later, another high profile IPO faced a different
set of technical glitches. This time, trading glitches at Nasdaq left
investors who tried to buy shares of Facebook, one of the most well-
known and valuable IPOs in recent memory, with no idea whether their
trade had been completed for hours, and, reportedly in some cases,
days. Now, Facebook stock is worth barely half of its initial value,
and lawsuits continue to be filed as Nasdaq, the underwriters, and
investors fight over the damage.
Still another computerized trading error occurred last month. This
time, Knight Capital, one of the largest market makers, nearly
collapsed when a rogue trading program went unchecked for just about 40
minutes. Despite new rules intended to stop wild price swings, millions
of unintended trades flung prices and volumes in about 150 stocks all
out of whack. Ultimately, a seemingly small programming glitch cost
Knight about $440 million and forced them to sell off a 73 percent
stake in the company, nearly wiping out existing shareholders.
These accidental trades were one of the first significant tests of
the rules that the SEC put in place after the Flash Crash. Most of
those protections weren't triggered. While this is certainly a warning
to other firms about the dangers of a single software error, it should
also be a warning to regulators about the strength and security of our
markets.
And these incidents don't include the dozens of ``mini-crashes''
that have occurred in single issuers. For a company and their
employees, it must be more than a bit discomforting to watch your stock
price spike and plummet over minutes, often for no discernible reason.
It's past time for some more upgrades.
Regulators don't yet have effective systems to stop catastrophic
collapses, and they lack the ability to see all of the activity that's
occurring in all of the venues in a way that lets them understand
what's going on. The consolidated audit trail was proposed years ago.
It doesn't cover nearly enough activity, and is still years away from
being a reality.
Regulators don't seem to know what the traders are really doing.
While so-called ``naked access,'' wherein traders are given unfettered
access to the trading markets without real oversight, is banned, there
are still no driving tests for these traders. There are still no
licenses for these traders, unless they opt into them.
System failures are not the only concerns with our markets. There
are very real questions about whether our U.S. capital markets are
fair. Since the joint PSI/Securities Subcommittee hearing in 2010,
there have been a number of high-profile instances where some market
participants have been given preferences over ordinary investors, and
where our opaque market structure has allowed trading abuses that have
victimized unwitting investors to go otherwise undetected.
Recently, the Wall Street Journal reported how some exchanges have
created special order types that allow sophisticated algorithmic
traders to jump the line and take advantage of ordinary investors. In
some cases, exchanges are creating order types for sophisticated
traders to use that are designed not to trade. Some are hidden from
few. These orders aren't about maintaining fair and transparent
markets. Just the opposite.
Also recently, the NYSE settled charges by the SEC that it sent its
data feeds to certain customers faster than the data it provided to
other market participants. Collectively, these reports show that some
traders seem to be given a leg up on the rest of us. They get
information faster. They process it faster. And they are given tools to
take advantage of that information to the detriment of the rest of the
market.
One of the more offensive examples of this type of favoritism is in
so-called ``flash orders,'' where a venue ``flashes'' an order to a
select group of favored customers before showing that order to the rest
of its members. What's to prevent one of those favored customers from
taking advantage of that information to the detriment of the customer
whose order was flashed? Not much. Senator Schumer and I, and others,
have called for the banning of this practice. But, years later, that
still hasn't happened.
There are other fairness risks. Given the complexity of order
routing, and the incentives of the current ``maker/taker'' model,
brokers are often faced with a conflict of interest. They can send an
order to a venue with the best price, or they can send it to the venue
which gives them the most profit. If these decisions were made by
humans, we could ask them why they make particular decisions. But they
aren't. These decisions are buried deep within the codes of smart order
routers. Institutional traders and ordinary investors alike should
wonder whether the smart order router is really working for them or
their broker. To be sure, many brokers are fulfilling their duties to
their customers and getting the best executions, but how is the
customer to know? How is a regulator going to know?
Another significant challenge is the lack of transparency in the
off-exchange trading venues. In a case last year, the SEC fined
Pipeline Trading Systems, a dark pool operator, for trading in front of
customers, taking advantage of their role in the center of the
marketplace. While Pipeline was supposed to be matching up buyers and
sellers, it was secretly trading against its customers, often for a
profit. Is this type of activity happening at other dark pools? Most of
them disclose that the firm itself may take the other side of a trade.
So should it be ok if the customer knows that he's likely to be taken
advantage of? What's his alternative? Is it to post the order on the
New York Stock Exchange and be taken advantage of there?
I look forward to today's hearing, and we should all look forward
to pressing our regulators to address these questions. For more than a
century, American markets have been the envy of investors across the
world. We must continue to examine the regulations and structure that
has been put in place to ensure that today's equity market is fair for
all and has the right protections in place to prevent technical
glitches from causing a collapse.
Once again, I wish to thank Chairman Reed and Ranking Member Crapo
for holding this important hearing today. I hope this hearing will lead
to a strengthening of our markets to ensure that they remain the best
in the world in coming years.
STATEMENT OF MICAH HAUPTMAN, FINANCIAL CAMPAIGN COORDINATOR, PUBLIC
CITIZEN'S CONGRESS WATCH DIVISION
STATEMENT OF CAMERON SMITH, PRESIDENT, QUANTLAB FINANCIAL, AND RICHARD
GORELICK, CEO, RGM ADVISORS
SUMMARY AND EXCERPTS FROM ``HIGH FREQUENCY TRADING AND PRICE
DISCOVERY'' BY TERRY HENDERSHOTT AND RYAN RIORDAN