2013-22185
Federal Register, Volume 78 Issue 177 (Thursday, September 12, 2013)[Federal Register Volume 78, Number 177 (Thursday, September 12, 2013)]
[Proposed Rules]
[Pages 56541-56574]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: 2013-22185]
[[Page 56541]]
Vol. 78
Thursday,
No. 177
September 12, 2013
Part IV
Commodity Futures Trading Commission
-----------------------------------------------------------------------
17 CFR Chapter I
Concept Release on Risk Controls and System Safeguards for Automated
Trading Environments; Proposed Rule
Federal Register / Vol. 78 , No. 177 / Thursday, September 12, 2013 /
Proposed Rules
[[Page 56542]]
-----------------------------------------------------------------------
COMMODITY FUTURES TRADING COMMISSION
17 CFR Chapter I
RIN 3038-AD52
Concept Release on Risk Controls and System Safeguards for
Automated Trading Environments
AGENCY: Commodity Futures Trading Commission.
ACTION: Concept release; request for comments.
-----------------------------------------------------------------------
SUMMARY: U.S. derivatives markets have experienced a fundamental
transition from human-centered trading venues to highly automated and
interconnected trading environments. The operational centers of modern
markets now reside in a combination of automated trading systems
(``ATSs'') and electronic trading platforms that can execute repetitive
tasks at speeds orders of magnitude greater than any human equivalent.
Traditional risk controls and safeguards that relied on human judgment
and speeds, and which were appropriate to manual and/or floor-based
trading environments, must be reevaluated in light of new market
structures. Further, the Commission and market participants must ensure
that regulatory standards and internal controls are calibrated to match
both current and foreseeable market technologies and risks. This
Concept Release on Risk Controls and System Safeguards for Automated
Trading Environments (``Concept Release'') reflects the Commission's
continuing commitment to the safety and soundness of U.S. derivatives
markets in a time of rapid technological change. The Concept Release
serves as a platform for cataloguing existing industry practices,
determining their efficacy and implementation to date, and evaluating
the need for additional measures, if any. The Commission welcomes all
public comments.
DATES: Comments must be received on or before December 11, 2013.
ADDRESSES: You may submit comments, identified by RIN 3038-AD52, by any
of the following methods:
CFTC Web site, via Comments Online: http://comments.cftc.gov. Follow the instructions for submitting comments
through the Web site.
Mail: Melissa D. Jurgens, Secretary of the Commission,
Commodity Futures Trading Commission, Three Lafayette Centre, 1155 21st
Street NW., Washington, DC 20581.
Hand Delivery/Courier: Same as ``mail,'' above.
Federal eRulemaking Portal: http://www.regulations.gov.
Follow the instructions for submitting comments.
Please submit comments by only one method. All comments should be
submitted in English or accompanied by an English translation. Comments
will be posted as received to http://www.cftc.gov. You should submit
only information that you wish to make available publicly. If you wish
the Commission to consider information that may be exempt from
disclosure under the Freedom of Information Act (``FOIA''), a petition
for confidential treatment of the exempt information may be submitted
according to the procedures established in 17 CFR 145.9. The Commission
reserves the right, but shall have no obligation, to review, prescreen,
filter, redact, refuse, or remove any or all of your submission from
http://www.cftc.gov that it may deem to be inappropriate for
publication, such as obscene language. All submissions that have been
redacted or removed that contain comments on the merits of the
rulemaking will be retained in the public comment file and will be
considered as required under the Administrative Procedure Act and other
applicable laws, and may be accessible under FOIA.
FOR FURTHER INFORMATION CONTACT: Sebastian Pujol Schott, Associate
Director, Division of Market Oversight, [email protected] or 202-418-5641;
Marilee Dahlman, Attorney-Advisor, Division of Market Oversight,
[email protected] or 202-418-5264; Camden Nunery, Economist, Office of
the Chief Economist, [email protected] or 202-418-5723; or Sayee
Srinivasan, Research Analyst, Office of the Chief Economist,
[email protected] or 202-418-5309.
SUPPLEMENTARY INFORMATION:
I. Introduction
A. Design of Concept Release and Request for Comments
II. Background
A. Characteristics of Automated Trading Environments
1. Automated Order Generation and Execution
2. Advances in High-Speed Communication Networks and Reductions
in Latency
3. Rise of Interconnected Automated Markets
4. Manual Risk Controls and System Safeguards in Automated
Trading Environments
B. The Commission's Regulatory Response to Date
C. Recent Disruptive Events in Automated Trading Environments
III. Potential Pre-Trade Risk Controls, Post-Trade Reports, System
Safeguards, and Other Protections
A. Overview of Existing Industry Practices
1. Existing DCM Risk Controls
2. Existing Trading and Clearing Firm Risk Controls
B. Overview of Risk Controls Addressed in This Concept Release
C. Pre-Trade Risk Controls
1. Message and Execution Throttles
2. Volatility Awareness Alerts
3. Self-Trade Controls
4. Price Collars
5. Maximum Order Sizes
6. Trading Pauses
7. Credit Risk Limits
D. Post-Trade Reports and Other Post-Trade Measures
1. Order, Trade, and Position Drop Copy
2. Trade Cancellation or Adjustment Policies
E. System Safeguards
1. Controls Related to Order Placement
2. Policies and Procedures for the Design, Testing and
Supervision of ATSs; Exchange Considerations
3. Self-Certifications and Notifications
4. ATS or Algorithm Identification
5. Data Reasonability Checks
F. Other Protections
1. Registration of Firms Operating ATSs
2. Market Quality Data
3. Market Quality Incentives
4. Policies and Procedures To Identify ``Related Contracts''
5. Standardize and Simplify Order Types
G. General Questions Regarding All Risk Controls Discussed Above
IV. List of All Questions in the Concept Release
V. Appendices (Specific Measures in Bold Font)
A. Pre-Trade Risk Controls
B. Post-Trade Reports and Other Post-Trade Measures
C. System Safeguards
D. Other Protections
I. Introduction
U.S. derivatives markets have experienced a fundamental evolution
from human-centered trading venues to highly automated and
interconnected trading environments. Traditionally, traders and market
participants directly initiated, communicated and executed orders,
while other personnel provided a range of order, trade processing and
back office services. In contrast, automated trading environments are
characterized precisely by their high degree of automation, and by the
wide array of algorithmic and information technology systems that
generate, risk manage, transmit and match orders and trades, as well as
systems used to confirm transactions, communicate market data and link
related systems through high-speed communication networks. Automated
trading environments have conferred a number of benefits upon market
participants, including an expanded range of potential trading
strategies, and a surge in the speed, precision and tools
[[Page 56543]]
available to execute such strategies. In addition to these benefits,
however, automated trading environments have also presented challenges
unique to their speed, interconnectedness and reliance on algorithmic
systems.
In recent years, a number of high-profile system events associated
with automated trading have raised public, Commission and industry
awareness. For example, on May 6, 2010, major equity indices in both
the futures and securities markets lost more than 5% of their value in
a matter of minutes when an automated order led to extreme downward
price movement and a liquidity crisis in the Chicago Mercantile
Exchange's (``CME'') E-mini futures contract.\1\ In August 2012, a
trading firm in the securities markets--Knight Capital Group--submitted
a significant number of errant proprietary orders to the New York Stock
Exchange (``NYSE''), causing price swings in nearly 150 securities and
costing the firm approximately $440 million in the process.\2\ Most
recently, in August 2013, trading on the Nasdaq stock market was
disrupted for three hours due to malfunctions in quote dissemination
systems and potential connectivity issues between it and another
trading platform's systems. These and other recent events in automated
trading environments are discussed in greater detail in section II.C.,
below.
---------------------------------------------------------------------------
\1\ See ``Findings Regarding the Market Events of May 6, 2010,
Report of the Staffs of the CFTC and SEC to the Joint Advisory
Committee on Emerging Regulatory Issues,'' September 30, 2010
[hereinafter, the ``CFTC and SEC Joint Report on the Market Events
of May 6, 2010''], available at http://www.cftc.gov/ucm/groups/public/@otherif/documents/ifdocs/staff-findings050610.pdf.
\2\ See Jenny Strasburg & Jacob Bunge, ``Loss Swamps Trading
Firm,'' Wall St. J. (Aug. 2, 2012), available at http://online.wsj.com/article/SB10000872396390443866404577564772083961412.html.
On October 2, 2012, the Securities and Exchange Commission
(``SEC'') conducted a roundtable entitled ``Technology and Trading:
Promoting Stability in Today's Markets'' (``SEC Roundtable''). See
SEC, Notice of Roundtable Discussion: Technology and Trading
Roundtable, 77 FR 56697 (Sept. 13, 2012). A transcript of the SEC
Roundtable [hereinafter, the ``SEC Roundtable Transcript''] is
available at http://www.sec.gov/news/otherwebcasts/2012/ttr100212.shtml. At the SEC Roundtable, then-SEC Chairman Schapiro
raised the Knight Capital incident and noted that ``[e]vents like
these demonstrate the core infrastructure and technology issues that
can be problematic in any market structure.'' See SEC Roundtable
Transcript at 11.
---------------------------------------------------------------------------
The Commission has taken steps to address the transition to
automated trading and require appropriate risk controls for designated
contract markets (``DCMs''), swap execution facilities (``SEFs''),
futures commission merchants (``FCMs''), swap dealers (``SDs''), major
swap participants (``MSPs'') and others. In April 2012, it adopted
final rules requiring FCMs, SDs and MSPs that are clearing members to
establish risk-based limits based on position size, order size, margin
requirements, or similar factors, and requiring those entities to use
automated means to screen orders for compliance with the risk limits
when such orders are subject to automated execution. Further, in June
2012, the Commission adopted final rules with respect to DCMs,
including requirements that DCMs establish and maintain risk control
mechanisms to prevent and reduce the potential for price distortions
and market disruptions. Relevant controls cited in the rule include
trading pauses and halts under conditions prescribed by the DCM. The
Commission adopted similar requirements in its final rules for SEFs in
2013. Finally, the DCM final rules also require risk control
requirements for exchanges that provide direct market access (``DMA'')
to clients.
The Commission has also adopted rules related to trading practices,
including trading in automated environments. In July 2011, the
Commission adopted final rules codified in 17 CFR Part 180 that, among
other things, (i) broadly prohibit manipulative and deceptive devices,
i.e., fraud and fraud-based manipulative devices and contrivances
employed intentionally or recklessly, regardless of whether the conduct
in question was intended to create or did create an artificial price;
and (ii) codify the Commission's long-standing authority to prohibit
price manipulation by making it unlawful for any person, directly or
indirectly, to manipulate or attempt to manipulate the price of any
swap, or of any commodity in interstate commerce, or for future
delivery on or subject to the rules of a registered entity. Further,
section 747 of the Dodd-Frank Wall Street Reform and Consumer
Protection Act (the ``Dodd-Frank Act'') \3\ amended the Commodity
Exchange Act (``CEA'' or ``Act'') to make it unlawful for any person to
engage in disruptive trading practices, and the Commission has provided
guidance on the scope and application of the new statutory
prohibitions. The Commission's measures to date are summarized in
greater detail in section II.B., below. With respect to these measures
and others discussed in this Concept Release, the Commission requests
public comment regarding any additional steps, guidance or rulemaking
that it should undertake.
---------------------------------------------------------------------------
\3\ See Dodd-Frank Wall Street Reform and Consumer Protection
Act of 2010, Public Law 111-203, 124 Stat. 1376 (2010).
---------------------------------------------------------------------------
Derivatives market participants, including DCMs, FCMs, clearing
members and others, have themselves taken a number of steps to manage
risks associated with automated trading. The Commission acknowledges
these efforts, and, through this Concept Release, seeks public comment
on the extent to which measures already in place may be sufficient to
safeguard markets in automated trading environments. In particular,
section III below summarizes relevant risk controls implemented by one
or more market participants; requests comment regarding the extent of
their implementation to date; and seeks input regarding whether
existing controls would benefit from additional granularity or
regulatory standardization.
A. Design of Concept Release and Request for Comments
This Concept Release provides an overview of the automated trading
environment, including its principal actors, potential risks, and
preventative measures designed to promote safe and orderly markets.\4\
The Concept Release was informed by controls already in use today by
one or more market participants or exchanges, and best practices,
recommendations and concepts developed by the CFTC's Technology
Advisory Committee (``TAC''); the Futures Industry Association's
(``FIA'') Principal Traders Group and Market Access Working Group; the
International Organization of Securities Commissions (``IOSCO''); the
European Securities and Markets Authority (``ESMA''); and by existing
CFTC regulatory requirements. It begins with an overview of automated
trading, including the development of automated order generation and
execution systems; advances in high-speed communication networks; the
growth of interconnected automated markets; the changed role of humans
in modern markets; and a discussion of
[[Page 56544]]
recent disruptive events in automated trading environments. The Concept
Release then addresses these developments through a series of (1) pre-
trade risk controls; (2) post-trade reports and other post-trade
measures; (3) system safeguards; and (4) additional protections
(collectively, ``risk controls'') that could be implemented by one or
more categories of Commission registrants or other market participants.
---------------------------------------------------------------------------
\4\ Many of these concepts are in harmony with evolving views of
groups responsible for setting standards and developing regulations
for other markets around the world. See, e.g., IOSCO Technical
Committee, ``Regulatory Issues Raised by the Impact of Technological
Changes on Market Integrity and Efficiency: Consultation Report''
(July 2011) [hereinafter ``IOSCO Report on Regulatory Issues Raised
by Technological Changes''], available at http://www.iosco.org/library/pubdocs/pdf/IOSCOPD354.pdf.
See also ESMA, ``Final Report: Guidelines on Systems and
Controls in an Automated Trading Environment for Trading Platforms,
Investment Firms and Competent Authorities'' (December 2011)
[hereinafter, ``ESMA Guidelines on Systems and Controls''],
available at http://www.esma.europa.eu/system/files/2011-456_0.pdf.
---------------------------------------------------------------------------
The Commission seeks extensive public comment regarding each risk
control contemplated herein. Commenters should address the
effectiveness of each measure, and the degree to which it may already
be in use by industry participants. Each commenter should identify the
specific risk controls that it already employs. For all measures
discussed in this Concept Release, commenters should also address
whether there is a need for regulatory action to provide more uniform
risk mitigation across CFTC-regulated derivatives markets.\5\ Comments
that address this question with respect to each proposed risk control
and system safeguard individually would be particularly helpful. In all
cases, commenters should discuss, and quantify wherever possible, the
costs and benefits of the pre-trade risk controls, post-trade reports
and other post-trade measures, system safeguards, and other protections
discussed in this Concept Release.
---------------------------------------------------------------------------
\5\ In this regard, the Commission emphasized in the preamble to
its final rules for part 38 that the efficacy of risk controls
depends in part on the proper functioning of electronic systems, and
that ``the Commission may address electronic system testing,
controls, and supervision-related issues in a subsequent
proceeding.'' See Commission, Final Rule: Core Principles and Other
Requirements for Designated Contract Markets, 77 FR 36612, 36638
n.298, 36648, n.389 (Jun. 19, 2012) [hereinafter, the ``DCM Final
Rules''].
Similarly, the system safeguards contemplated herein for ATSs
are an outgrowth of the basic requirement in Sec. 23.600(d)(9) that
SDs and MSPs conduct testing and supervision of trading systems.
There again, the Commission indicated that further measures would be
forthcoming by stating that it ``anticipate[d] addressing the
related issues of testing and supervision of electronic trading
systems and mitigation of the risks posed by high frequency
trading.'' See Commission, Final Rule: Swap Dealer and Major Swap
Participant Recordkeeping, Reporting, and Duties Rules; Futures
Commission Merchant and Introducing Broker Conflicts of Interest
Rules; and Chief Compliance Officer Rules for Swap Dealers, Major
Swap Participants, and Futures Commission Merchants, 77 FR 20128,
20141 (Apr. 3, 2012).
---------------------------------------------------------------------------
The Concept Release recognizes that orders and trades in automated
environments pass through multiple stages in their lifecycle from order
generation, to execution, to clearing and allocation in proprietary or
customer accounts, and steps in between. Accordingly, the Commission
requests comment regarding the appropriate stage at which risk controls
should be placed. Potential options include risk controls applicable
to: (i) ATSs at the time of order generation; (ii) clearing firms
during the order transmission process; (iii) trading platforms prior to
exposing orders to the market; (iv) Derivatives Clearing Organizations
(``DCOs''); and (v) other risk control focal points, including, for
example, third-party ``hubs'' through which orders or order information
could flow to uniformly mitigate risks across one or more trading
platforms. Similarly, the Commission requests public comment regarding
the appropriate focal point for system safeguards and testing and
supervision standards for ATSs.
Finally, the Commission requests comment regarding a series of
issues central to its improved understanding and surveillance of
trading in automated environments. For example, the Commission requests
comments regarding any surveillance tools that it should deploy
specifically for the surveillance of automated trading and areas for
academic research to improve its understanding of ATSs' impact on
market microstructure. Section IV lists all questions raised in this
Concept Release.
The Commission's Concept Release reflects fundamental statutory
objectives under the CEA. Such objectives include fostering a system of
effective self-regulation, deterring and preventing disruptions to
market integrity, protecting market participants and ``promot[ing]
responsible innovation and fair competition among boards of trade,
other markets and market participants.'' \6\ Notably, the Commission
must ensure that U.S. derivatives markets continue to serve as
effective centers of price discovery and risk mitigation, regardless of
the technologies employed by trading platforms, market participants,
and others. The Commission must further ensure that its regulatory
framework and industry practices are fully adapted to the automated
technologies of modern derivatives markets.
---------------------------------------------------------------------------
\6\ See CEA section 3(b); 7 U.S.C. 5(b).
---------------------------------------------------------------------------
II. Background
A. Characteristics of Automated Trading Environments
1. Automated Order Generation and Execution
Automated trading environments have developed in tandem with
automated systems for both the generation and execution of orders.
Systems related to the generation of orders (``automated trading
systems'' or ``ATSs'') \7\ operate at the beginning of the order and
trade lifecycle; they reflect a set of rules or instructions (an
algorithm) and related computer systems used to automate the execution
of a trading strategy.\8\ ATSs may operate as automated execution
programs designed to minimize the price impact of large orders; achieve
a benchmarked price (e.g., volume-weighted average price and time-
weighted average price algorithms); or otherwise execute instructions
traditionally provided by a human agent.\9\ They may be employed by a
range of market participants, with varying degrees of sophistication,
for both proprietary and customer trading. For example, buy-side firms
(such as mutual funds and pension funds) may use automated systems and
execution algorithms to ``shred'' one or more large orders (called
``parent order'') into a series of smaller trades (``child orders'') to
be executed over time. Such systems can include additional algorithms
to micro-manage the size, frequency and timing (often randomized) of
child orders. In addition to automated execution, ATSs may also operate
market-making programs; opportunistic, cross-asset and cross-market
arbitrage programs; and a number of other strategies.
---------------------------------------------------------------------------
\7\ While the Commission has no regulatory definition of ATS,
the term is generally understood to mean a computer-driven system
that automates the generation and routing of orders to one or more
markets. Other elements of an ATS may also include systems for
analyzing market data as a precursor to order generation, managing
orders for conformance with establish risk tolerances, receiving
confirmations of orders placed and trades executed, etc. Section
III.E.4. of this Concept Release seeks public input regarding
whether the Commission should formally define ATS and if so, how ATS
should be defined.
\8\ See IOSCO Report on Regulatory Issues Raised by
Technological Changes, supra note 4, at 10.
\9\ See John Bates, ``Algorithmic Trading and High Frequency
Trading Experiences from the Market and Thoughts on Regulatory
Requirements'' (July 2010), available at http://www.cftc.gov/ucm/groups/public/@newsroom/documents/file/tac_071410_binder.pdf.
---------------------------------------------------------------------------
In Commission-regulated markets, orders generated by ATSs are
ultimately transmitted to DCMs that have themselves become automated
systems for the matching and execution of orders. Broadly, these
trading platforms consist of a front-end to which market participants
connect and communicate using standardized messaging formats, a
matching engine that automatically matches orders to buy and sell, and
a back-end that automatically provides all market participants with a
market feed. Trade flows may make use of straight-through processing,
where the entire trade execution process occurs without intermediation
from humans, thereby
[[Page 56545]]
dramatically reducing the amount of time required to execute each
transaction. The evolution from manual trading in open-outcry pits to
electronic trading platforms is in many cases substantially complete.
An established body of data indicates the importance of electronic
and algorithmic trading in U.S. futures markets. In 2012, approximately
91.50% of exchange trading volume in U.S. futures markets was executed
electronically.\10\ Estimates indicate that algorithmic trading first
accounted for at least 50% of orders in 2009,\11\ and accounted for
over 40% of total trading volume in 2010.\12\ By the end of the first
quarter of 2010, ATSs accounted for over 50% of trading volume in a
number of significant product categories at CME Group, Inc.'s (``CME
Group'') DCMs.\13\ For example, ATSs accounted for approximately 51% of
trade volume in E-mini S&P 500 futures and 69% of trade volume in
EuroFX futures.\14\ Increased automation in both order generation and
matching, combined with the exponentially faster communication networks
discussed in section II.A.2., below, has in many cases reduced the
trade lifecycle to as little as a few milliseconds. As a result, high-
frequency trading (``HFT'') strategies have also become an increasingly
important component of automated trading environments.
---------------------------------------------------------------------------
\10\ This figure represents transactions executed competitively
on DCM trading platforms and not off-exchange transactions such as
block trades.
\11\ See Paul Zubulake & Sang Lee, The High Frequency Game
Changer at 84, fig. 6.3 (John Wiley & Sons, Inc. 2011) (source of
data: Aite Group).
\12\ See Barry Johnson, Algorithmic Trading & DMA: An
Introduction to Direct Access Trading Strategies at 78, fig. 3-11
(4Myeloma Press 2010) (source of data: Aite Group).
\13\ See CME Group, ``Algorithmic Trading and Market Dynamics''
(July 15, 2010) at 2, available at http://www.cmegroup.com/education/files/Algo_and_HFT_Trading_0610.pdf. At the time, the
CME Group operated four DCMs: the Chicago Mercantile Exchange, the
Chicago Board of Trade, the New York Mercantile Exchange
(``NYMEX''), and the Commodity Exchange.
\14\ See id.
---------------------------------------------------------------------------
The Commission is working diligently to understand and keep pace
with the growth of ATSs and HFT in its regulated markets. The TAC, for
example, has worked to define HFT and received a definition of HFT from
its working group panel of experts. The attributes of HFT, according to
the TAC's working group, include:
(a) Algorithms for decision making, order initiation, generation,
routing, or execution, for each individual transaction without human
direction;
(b) low-latency technology that is designed to minimize response
times, including proximity and co-location services;
(c) high speed connections to markets for order entry; and
(d) recurring high message rates (orders, quotes or cancellations)
determined using one or more objective forms of measurement, including
(i) cancel-to-fill ratios; (ii) participant-to-market message ratios;
or (iii) participant-to-market trade volume ratios.\15\
---------------------------------------------------------------------------
\15\ See TAC Subcommittee on Automated and High Frequency
Trading, Working Group 1, Presentation to the TAC (Oct. 30, 2012),
available at http://www.cftc.gov/ucm/groups/public/@newsroom/documents/file/tac103012_wg1.pdf. In addition, the TAC Subcommittee
on Automated and High Frequency Trading, Working Group 1, described
high frequency trading as a mechanism used by a variety of trading
strategies, including, but not limited to, liquidity provision and
statistical arbitrage.
---------------------------------------------------------------------------
In addition, the TAC's working group described automated trading as
``cover[ing] systems employed in the decision-making, routing and/or
execution of an investment or trading decision, which utilizes a range
of technologies including software, hardware, and network components to
facilitate efficient access to the financial markets via electronic
trading platforms.'' \16\ Effectively, HFT is a form of automated
trading, but not all automated trading is HFT.\17\
---------------------------------------------------------------------------
\16\ See id.
\17\ In March 2013, the German parliament approved legislation
on high frequency trading (the ``HFT Act''). See Hans-Edzard
Busemann, ``German upper house approves rules to clamp down on high-
frequency trading,'' Reuters (March 22, 2013), available at http://uk.reuters.com/article/2013/03/22/uk-germany-trading-idUKBRE92L0L820130322. The legislation defines high frequency
trading generally as follows: The sale or purchase of financial
instruments for own account as direct or indirect participant in a
domestic organized market or multilateral trading facility by means
of a high-frequency algorithmic trading technique which is
characterized by (i) the usage of infrastructures to minimize
latency times, (ii) the decision of the system regarding the
commencement, creation, transmission or execution of an order
without human intervention for single transactions or orders, and
(iii) a high intraday messaging volume in the form of orders, quotes
or cancellations. See BaFin (Federal Financial Supervisory
Authority), ``High-frequency trading: new rules for trading
participants'' (March 26, 2013) (including Workshop on High
Frequency Trading Act Presentations dated April 30, 2013 and
Frequently Asked Questions Relating to the High Frequency Trading
Act dated March 22, 2013) [hereinafter, the ``BaFin HFT Act
Materials''], available at http://www.bafin.de/SharedDocs/Veroeffentlichungen/EN/Meldung/2013/meldung_130322_hft-gesetz_en.html?nn=2821494.
The German HFT Act also defines algorithmic trading. The HFT
Act's definition is generally as follows: Trading with financial
instruments such that a computer algorithm determines automatically
the individual order parameters without being merely a system for
the transmission of orders to one or several trading venues or to
confirm orders. Order parameters within the meaning of the preceding
sentence are decisions whether the order is given, the timing, price
and quantity of an order or how the order will be executed with
limited or no human interference. See id. As explained in footnote
103 below, the HFT Act also introduces a licensing requirement.
---------------------------------------------------------------------------
In this regard, the Commission is aware that instability in
automated trading environments may be precipitated by ATSs regardless
of whether they employ high-frequency or other trading strategies.
Accordingly, the risk controls, system safeguards and other measures
contemplated for ATSs in this Concept Release do not distinguish on the
basis of ATSs' trading strategies. However, the Commission is
interested in better understanding HFT and whether it should receive
different regulatory attention than ATSs in general. The Commission
requests comment on the following questions regarding HFT and related
topics:
1. In any rulemaking arising from this Concept Release, should the
Commission adopt a formal definition of HFT? If so, what should that
definition be, and how should it be applied for regulatory purposes?
2. What are the strengths and weaknesses of the TAC working group
definition of HFT provided above? How should that definition be
amended, if at all?
3. The definition of HFT provided above uses ``recurring high
message rates (orders, quotes or cancellations)'' as one of the
identifying characteristics of HFT, and lists three objective measures
(i) cancel-to-fill ratios; (ii) participant-to-market message ratios;
or (iii) participant-to-market trade volume ratios) that could be used
to measure message rates. Are these criteria sufficient to reliably
distinguish between ATSs in general and ATSs using HFT strategies? What
threshold values are appropriate for each of these measures in order to
identify ``high message rates?'' Should these threshold values vary
across exchanges and assets? If so, how?
4. Should the risk controls for systems and firms that engage in
HFT be different from those that apply to ATSs in general? If so, how?
2. Advances in High-Speed Communication Networks and Reductions in
Latency
Automated trading environments are also characterized by
connectivity and infrastructure solutions that enable trading platforms
to process orders and execute trades at ever increasing speeds, and
enable market participants (including ATSs) to communicate with
platforms at ever decreasing latencies.\18\
[[Page 56546]]
Notably, however, such capabilities require equally sophisticated risk
management systems whose speeds are commensurate with those of low-
latency order generation and trade execution systems. Public data from
one exchange group, for example, indicates that roundtrip trade times
on its trading platform fell from 127 milliseconds in 2004 to 4.2
milliseconds in 2011.\19\ Another exchange group reported in 2010 that
its average blended transaction time in futures and OTC markets was
1.25 milliseconds.\20\ Advances in trading speeds are partly due to the
development of dedicated fiber-optic and microwave communications
networks that have dramatically reduced latency across large distances.
As of 2012, networks were being developed to reduce roundtrip messaging
between New York and London from 65 milliseconds to 60
milliseconds.\21\ In March 2013, CME Group Inc. and Nasdaq OMX Group
Inc. announced plans to launch a wireless network that will provide
roundtrip messaging between New York and Chicago in 8.5
milliseconds.\22\
---------------------------------------------------------------------------
\18\ Latency means ``the time it takes to learn about an event
(e.g., a change in the bid), generate a response, and have the
exchange act on the response.'' See Joel Hasbrouck & Gideon Saar,
``Low-Latency Trading'' (May 2013) at 1, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1695460.
\19\ See CME Group, ``Oversight of Automated Trading at CME
Group'' (March 29, 2012) at 4, available at http://www.cftc.gov/ucm/groups/public/@aboutcftc/documents/file/tacpresentation032912_cme.pdf.
\20\ See IntercontinentalExchange, ``2010 Annual Report,'' at
26, available at http://files.shareholder.com/downloads/ICE/1747226327x0x456112/BF6F428C-F8B3-4835-B22C-3F350FF13B89/ICE_2010AR.pdf. IntercontinentalExchange indicated that it measures
round trip performance end to end within its data center and through
its matching engine.
\21\ See Matthew Philips, ``Stock Trading is About to Get 5.2
Milliseconds Faster,'' BloombergBusinessweek (Mar. 29, 2012),
available at http://www.businessweek.com/articles/2012-03-29/trading-at-the-speed-of-light.
\22\ See Jacob Bunge, ``CME, Nasdaq Plan High-Speed Network
Venture,'' Wall St. J. (Mar. 28, 2013), available at http://online.wsj.com/article/SB10001424127887324685104578388343221575294.html.
---------------------------------------------------------------------------
Two common methods for reducing latency are co-location and
proximity hosting, defined as the placement of a firm's trading
technology in close proximity to the trading platform. They may be
offered directly by an exchange or by a third-party service provider.
Co-location denotes those connectivity solutions hosted by the exchange
itself, while proximity hosting indicates services offered by third
parties.\23\ In 2010, the Commission published in the Federal Register
a Notice of Proposed Rulemaking to require DCMs and others that offer
co-location and/or proximity hosting to offer such services on an equal
access basis, ensure that fees are uniform and non-discriminatory, and
provide information about the latency for various connectivity options
(``co-location rulemaking'').\24\ The Commission intends to finalize
the co-location rulemaking by the end of the year.
---------------------------------------------------------------------------
\23\ See FIA Market Access Working Group, ``Market Access Risk
Management Recommendations'' (April 2010) at 4 [hereinafter, ``FIA
Market Access Recommendations''], available at http://www.futuresindustry.org/downloads/Market_Access-6.pdf.
\24\ See Commission, Notice of Proposed Rulemaking: Co-Location/
Proximity Hosting Services, 75 FR 33198 (Jun. 11, 2010).
---------------------------------------------------------------------------
Another important latency-reducing advance in connectivity is DMA.
For purposes of this Concept Release, DMA is defined as a connection
method that enables a market participant to transmit orders to a
trading platform without reentry or prior review by systems belonging
to the market participant's clearing firm. DMA can be provided directly
by an exchange or through the infrastructure of a third-party provider.
In all cases, however, DMA connectivity implies that a market
participant's order flow is not routed through its clearing firm prior
to reaching the trading platform.\25\
---------------------------------------------------------------------------
\25\ The Commission has taken steps to mitigate the risk
associated with DMA. Rule 1.73, passed by the Commission in April
2012, requires FCMs that are clearing members to pre-screen orders
of DMA clients against risk limits that are established by the FCM.
See 17 CFR 1.73(a)(2)(i). See additional discussion in section II.B.
---------------------------------------------------------------------------
Investment in high-speed communication networks and other
technologies to reduce latency reflects the premium that some market
participants place on speed relative to their competitors. Reductions
in latency may be appropriately achieved through improvements in a
range of technologies for the generation, transmission and execution of
orders or management of other data. However, there are also incentives
for market participants to reduce latency by minimizing pre-trade risk
controls and other safeguards that might otherwise introduce unwanted
delays. While latency-based incentive structures have promoted evident
technological innovation in many derivatives markets, they can also
lead to a competitive race to the bottom--a concern already expressed
by some market participants.\26\ A separate concern is that market
participants may simply engage in trading at speeds greater than the
speed of their risk management systems. In a trading environment where
a single algorithm can submit hundreds of orders per second, risk
management systems operating at slower speeds could allow an algorithm
that is operating in unexpected ways to disrupt one or more markets.
---------------------------------------------------------------------------
\26\ As noted by FIA's Market Access Working Group, for example:
``[p]re-trade risk controls have become a point of negotiation
between trading firms and clearing members because they can add
latency to a trade.'' See FIA Market Access Recommendations, supra
note 23, at 8.
Similarly, the TAC's Pre-Trade Functionality Subcommittee noted
that latency is a key area where trading firms and brokers are
competing to gain an advantage. See TAC Pre-Trade Functionality
Subcommittee, ``Recommendations on Pre-Trade Practices for Trading
Firms, Clearing Firms, and Exchanges Involved in Direct Market
Access'' (March 1, 2011) at 2 [hereinafter, ``TAC Pre-Trade
Functionality Subcommittee DMA Recommendations''], available at
http://www.cftc.gov/ucm/groups/public/@swaps/documents/dfsubmission/tacpresentation030111_ptfs2.pdf.
---------------------------------------------------------------------------
5. Discussions on latency often focus on the how quickly an
exchange processes orders, the time taken to submit orders, and how
quickly a firm can observe prices of trades transacted on the exchange.
The Commission is interested in understanding whether there are other
types of messages transmitted between exchanges, firms and vendors
wherein differences in latency could provide opportunities for
informational advantage. Recent press reports have highlighted such
advantages in the transmission of trade confirmations by a specific
exchange.\27\ Are there other exchanges and trading venues where
similar differences in latency exist? The Commission is interested in
understanding whether the extent of latency in any such message
transmission process can have an adverse impact on market quality or
fairness. Should any exchanges, vendors and firms be required to audit
their systems and process on a periodic process to identify and then
resolve such latency?
---------------------------------------------------------------------------
\27\ See Scott Patterson, Jenny Strasburg & Liam Pleven, ``High-
Speed Traders Exploit Loophole,'' Wall St. J. (May 1, 2013),
available at http://online.wsj.com/article/SB10001424127887323798104578455032466082920.html.
---------------------------------------------------------------------------
3. Rise of Interconnected Automated Markets
In addition to greater automation and decreased latency,
derivatives markets are increasingly characterized by a high degree of
interconnection. ATSs and algorithms deployed to trade particular
products often interact directly and indirectly with ATSs and
algorithms active in other markets and jurisdictions. Increased
interconnectedness is facilitated by electronic access to real-time
pricing information, automated order execution, and some
standardization in communication protocols at various
[[Page 56547]]
trading platforms.\28\ ATSs can quickly execute strategies across
multiple markets within very short periods of time. Often, cross-market
activity is driven by latent arbitrage opportunities and faster access
to multiple markets has led to a proliferation of strategies that seek
to identify and trade on the basis of these relationships.\29\
---------------------------------------------------------------------------
\28\ For example, FIX language makes it possible for ATS to be
``platform independent''--to incorporate interfaces to multiple
brokers, ECNs, or exchanges. See Irene Aldridge, High-Frequency
Trading: A Practical Guide to Algorithmic Strategies and Trading
Systems at 31 (John Wiley & Sons, Inc. 2010). See also Cliff, Brown,
& Treleaven, ``Technology Trends in the Financial Markets: A 2020
Vision,'' United Kingdom Government Office for Science--Foresight,
at 10, available at http://www.bis.gov.uk/assets/foresight/docs/computer-trading/11-1222-dr3-technology-trends-in-financial-markets.pdf.
\29\ For example, ``basis trading,'' and ``futures/equity
arbitrage'' are statistical arbitrage strategies that seek to
capitalize on deviations between prices on futures contracts and
related securities contracts after macroeconomic news announcements.
See Aldridge, supra note 28, at 197-98.
---------------------------------------------------------------------------
Increased interconnectedness encourages price efficiencies when
economically identical or related contracts are traded on multiple
exchanges. However, it also increases the speed with which a disruption
on one trading platform, or within one ATS or algorithm, can impact
related markets. For example, a trading platform may experience changes
in the prices, spreads or volatility of one or more of its products due
to errors in an ATS or algorithm active in its markets. Even if this
algorithm does not trade elsewhere, such changes are likely to quickly
impact the prices, spreads, and volatility of related products on other
platforms, as automated systems attempt to arbitrage price differences.
The potential result is a cascading series of market disruptions,
brought about by the malfunction of a single ATS or algorithm trading
on a single platform.
Transmission effects such as this are illustrated by events like
the May 6, 2010 ``Flash Crash.'' On that day, major equity indices in
both the futures and securities markets fell over 5% in minutes before
recovering almost as quickly. After investigation by both the
Commission and the SEC, it was found that a fundamental seller utilized
an automated execution algorithm to sell 75,000 E-mini contracts
(valued at approximately $4.1 billion) over an abbreviated time
interval. The algorithm placed orders based on recent trading volume
but was not programmed to take price or time into account; because of
this lapse, a feedback loop triggered continued orders from the
algorithm even as prices moved far beyond traditional daily ranges.
Like the hypothetical example provided above, these declines in the
derivatives market quickly filtered over to different, but closely
related, products on many other exchanges.\30\ Soon after the initial
moves in the E-mini contract, similar extreme volatility was
experienced by the S&P 500 SPDR exchange traded fund and by many of the
500 underlying securities which make up the index itself.
---------------------------------------------------------------------------
\30\ See CFTC and SEC Joint Report on the Market Events of May
6, 2010, supra note 1, at 1-6; ``Recommendations Regarding
Regulatory Responses to the Market Events of May 6, 2010, Summary
Report of the Joint CFTC-SEC Advisory Committee on Emerging
Regulatory Issues'' (February 18, 2011), available at http://www.cftc.gov/ucm/groups/public/@aboutcftc/documents/file/jacreport_021811.pdf.
---------------------------------------------------------------------------
In response to the May 2010 flash crash, regulatory authorities and
market participants have taken steps to address volatility in U.S.
markets, including trading pauses and halts that operate as ``circuit
breakers.'' For example, in May 2012, the SEC approved a ``limit up-
limit down'' mechanism in which a price band is set at a percentage
level above and below the average price of the stock over the
immediately preceding five-minute trading period.\31\ If the stock's
price does not naturally move back within the price bands within 15
seconds, there will be a five-minute trading pause. The limit up-limit
down mechanism began implementation in April 2013, beginning with all
stocks in the S&P 500 and Russell 1000 and select exchange traded
products.
---------------------------------------------------------------------------
\31\ See SEC, ``Investor Bulletin: New Measures to Address
Market Volatility'' (Apr. 9, 2013), available at http://www.sec.gov/investor/alerts/circuitbreakersbulletin.htm.
---------------------------------------------------------------------------
In addition, the SEC approved updates to market-wide circuit
breaker rules that, when triggered, halt trading in all exchange-listed
securities in U.S. markets. Among other things, the new rules lower the
percentage-decline thresholds for triggering a market-wide trading
halt. The thresholds (Level 1 (7%), Level 2 (13%), and Level 3 (20%))
are set at levels calculated daily based on the prior day's closing
price of the S&P 500 index.\32\ To be consistent with these circuit
breakers, the CME Group, effective April 8, 2013, reduced the price
limit levels for CME and CBOT U.S. equity index futures to 7%, 13% and
20%.\33\ When a trading halt is declared in the primary securities
market in accordance with these levels, trading in the S&P 500 index
futures contracts will be halted at the CME. When trading in the
primary securities market resumes after any such halt, trading in the
S&P index futures contracts will resume. Similar rules apply to other
equity index futures contracts listed on CME. In March 2012, ICE
Futures U.S. introduced a circuit breaker functionality called Interval
Price Limits, in which prices may not move more than a pre-determined
amount away from the current market price within a pre-determined
period.\34\
---------------------------------------------------------------------------
\32\ See id.
\33\ See CME Group, ``Changes to CME and CBOT Equity Index Price
Limits: Frequently Asked Questions,'' available at http://www.cmegroup.com/education/files/faq-eq-hours-and-limits.pdf.
\34\ See IntercontinentalExchange, Inc., ``ICE Circuit Breakers
(IPL) Price Limits'' (March 2012), available at https://www.theice.com/publicdocs/technology/IPL_Circuit_Breaker.pdf.
---------------------------------------------------------------------------
Throughout section III below, the Commission seeks public comment
on the benefits of standardizing various risk controls and system
safeguards, including through the uniform application of regulatory
standards to help ensure an integrated risk management infrastructure
in regulated derivatives markets. The Commission draws commenters'
particular attention to the joint regulatory and industry response to
the Flash Crash summarized above and seeks public input regarding the
need for similar joint efforts with respect to the pre-trade risk
controls, post-trade reports, and system safeguards contemplated in
this Concept Release.
4. Manual Risk Controls and System Safeguards in Automated Trading
Environments
Orders in automated trading environments may be initiated by ATSs
and algorithms. Multiple other automated systems perform other
processing, communicating, and other functions. The speed of such
automated processes has necessarily shifted risk management functions
to parallel automated risk management systems acting with equal speed.
Within this context, manual risk controls, and particularly systems
safeguards, remain crucial to orderly markets. In many cases, manual
risk controls have shifted ``upstream'' to system design and
``downstream'' to system management. In automated trading, humans
design and test ATSs, establish decision criteria, manage
implementation, and intervene when technology systems fail. ATS
designers must identify the range of market conditions that an ATS
could reasonably face, and determine the range of permissible responses
by the ATS to each condition. Designers must also consider the array of
information that ATS operators will need to effectively monitor their
ATSs and the markets in which their ATSs operate. ATS operators, in
turn, must be
[[Page 56548]]
prepared to intervene when market conditions are outside of an ATS's
design parameters, when an ATS's trading strategy must be modified, or
when an ATS appears to be malfunctioning and must be shut down. Rapid
decisions must be made while simultaneously digesting large quantities
of information regarding multiple, fast-moving markets. Accordingly,
this Concept Release contemplates a number of risk controls and system
safeguards that emphasize the role and interaction of manual processes
with automated trading environments, particularly ATSs.
B. The Commission's Regulatory Response to Date
The Commission has responded to the development of automated
trading environments through a number of regulatory measures that
address risk controls within both new and existing categories of
registrants, including DCMs, SEFs, FCMs, SDs, MSPs and others. In April
2012, the Commission adopted rules requiring FCMs, SDs and MSPs that
are clearing members to establish risk-based limits based on ``position
size, order size, margin requirements, or similar factors'' for all
proprietary accounts and customer accounts.\35\ The rules, codified in
Sec. Sec. 1.73 and 23.609, also require these entities to ``use
automated means to screen orders for compliance with the [risk]
limits'' when such orders are subject to automated execution (emphasis
added).\36\ Such screening must, by definition, occur pre-trade. The
Commission also adopted rules in April 2012 requiring SDs and MSPs that
are clearing members to ensure that their ``use of trading programs is
subject to policies and procedures governing the use, supervision,
maintenance, testing, and inspection of the program.'' \37\ The
specific content of those policies and procedures are left up to the
SDs and MSPs.
---------------------------------------------------------------------------
\35\ 17 CFR 1.73(a)(1) and 23.609(a)(1).
\36\ 17 CFR 1.73(a)(2)(i) and 17 CFR 23.609(a)(2)(i).
\37\ 17 CFR 23.600(d)(9).
---------------------------------------------------------------------------
The Commission has also adopted relevant rules with respect to
exchange platforms, including rules with respect to DCMs (adopted in
June 2012).\38\ Regulation 38.255, for example, requires DCMs to
``establish and maintain risk control mechanisms to prevent and reduce
the potential risk of price distortions and market disruptions,
including, but not limited to, market restrictions that pause or halt
trading in market conditions prescribed by the designated contract
market.'' \39\ In addition, the acceptable practices for DCM Core
Principle 4 identify pre-trade limits on order size, price collars or
bands, and message throttles as responsive measures that a DCM may
implement to demonstrate compliance with elements of the core
principle.\40\ The Commission has adopted trading pause and halt
requirements for SEFs similar to those for DCMs.\41\
---------------------------------------------------------------------------
\38\ See DCM Final Rules, 77 FR 36612.
\39\ 17 CFR 38.255.
\40\ Part 38, Appendix B, Core Principle 4, section (b)(5),
provides: Risk controls for trading. An acceptable program for
preventing market disruptions must demonstrate appropriate trade
risk controls, in addition to pauses and halts. Such controls must
be adapted to the unique characteristics of the markets to which
they apply and must be designed to avoid market disruptions without
unduly interfering with that market's price discovery function. The
designated contract market may choose from among controls that
include: Pre-trade limits on order size, price collars or bands
around the current price, message throttles, and daily price limits,
or design other types of controls. Within the specific array of
controls that are selected, the designated contract market also must
set the parameters for those controls, so long as the types of
controls and their specific parameters are reasonably likely to
serve the purpose of preventing market disruptions and price
distortions. If a contract is linked to, or is a substitute for,
other contracts, either listed on its market or on other trading
venues, the designated contract market must, to the extent
practicable, coordinate its risk controls with any similar controls
placed on those other contracts. If a contract is based on the price
of an equity security or the level of an equity index, such risk
controls must, to the extent practicable, be coordinated with any
similar controls placed on national security exchanges. See DCM
Final Rules, 77 FR at 36718.
\41\ 17 CFR 37.405.
---------------------------------------------------------------------------
In the DCM final rules, the Commission also adopted new risk
control requirements for exchanges that provide DMA to clients.
Regulation 38.607 requires DCMs that permit DMA to have effective
systems and controls reasonably designed to facilitate an FCM's
management of financial risk. These systems and controls include
automated pre-trade controls through which member FCMs can implement
financial risk limits.\42\ As the Commission noted in the preamble to
the DCM final rules, in DMA arrangements ``it is impossible for an FCM
to protect itself without the aid of the DCM.'' \43\ The Commission
also noted in the DCM final rules, however, that ``the responsibility
to utilize these [DCM-provided] controls and procedures remains with
the FCM. Each FCM permitting direct access must use DCM-provided
controls . . . .'' \44\ Accordingly, regulation 38.607 requires DCMs to
implement and enforce rules requiring member FCMs to use these systems
and controls.\45\
---------------------------------------------------------------------------
\42\ See 17 CFR 38.607.
\43\ See DCM Final Rules, 77 FR at 36648.
\44\ Id.
\45\ See 17 CFR 38.607.
---------------------------------------------------------------------------
In addition to the foregoing, section 753 of the Dodd-Frank Act
amended section 6(c) of the CEA to prohibit manipulation and fraud in
connection with any swap, or a contract of sale of any commodity in
interstate commerce, or for future delivery on or subject to the rules
of any registered entity. In July 2011, the Commission adopted final
rules implementing this new authority under the CEA. CFTC Regulation
180.1, among other things, broadly prohibits manipulative and deceptive
devices, i.e., fraud and fraud-based manipulative devices and
contrivances employed intentionally or recklessly, regardless of
whether the conduct in question was intended to create or did create an
artificial price.\46\ CFTC Regulation 180.2 codifies the Commission's
long-standing authority to prohibit price manipulation by making it
unlawful for any person, directly or indirectly, to manipulate or
attempt to manipulate the price of any swap, or of any commodity in
interstate commerce, or for future delivery on or subject to the rules
of a registered entity.\47\
---------------------------------------------------------------------------
\46\ See 17 CFR 180.1.
\47\ See 17 CFR 180.2.
---------------------------------------------------------------------------
Finally, section 747 of the Dodd-Frank Act amended the CEA to make
it unlawful for any person to engage in disruptive trading practices.
Under section 4c(a)(5) of the CEA, it is unlawful for any person to
engage in any trading, practice, or conduct on or subject to the rules
of a registered entity that: Violates bids or offers, demonstrates
intentional or reckless disregard for the orderly execution of
transactions during the closing period, or is, is of the character of,
or is commonly known to the trade as, ``spoofing.'' In May 2013, the
Commission provided guidance on the scope and application of these
statutory prohibitions.\48\ In July 2013, the Commission issued an
order filing and settling charges against a high-speed trading firm for
engaging in the disruptive practice of ``spoofing'' by utilizing a
computer algorithm that was designed to illegally place and cancel bids
and offers in futures contracts.\49\
---------------------------------------------------------------------------
\48\ See Commission, Interpretive Guidance and Policy Statement,
78 FR 31890 (May 28, 2013).
\49\ See Commission, Press Release No. 6649-13 (July 22, 2013),
available at http://www.cftc.gov/PressRoom/PressReleases/pr6649-13.
---------------------------------------------------------------------------
C. Recent Disruptive Events in Automated Trading Environments
Recent malfunctions in ATS and trading platform systems, in both
derivatives and securities markets, illustrate the technological and
operational vulnerabilities inherent to automated trading environments.
ATSs,
[[Page 56549]]
for example, are vulnerable to algorithm design flaws, market
conditions outside of normal operating parameters, the failure of
built-in risk controls, operational failures in the communication
networks on which ATSs depend for market data and connectivity with
trading platforms, and inadequate human supervision. Incidents
involving an automated trading firm active in Commission-regulated
markets are illustrative of these concerns. For example, in 2011 NYMEX
fined a firm $350,000 for failing to adequately supervise, test, and
have controls in place related to its ATS.\50\ NYMEX cited a 2010 event
where the firm launched an ATS after limited testing. The firm was also
fined a total of $500,000 by CME for failure to effectively supervise
its ATSs on multiple occasions.\51\ A panel of the CME Business Conduct
Committee found that the firm had experienced malfunctions with the
same ATS multiple times, causing it to submit error trades.
---------------------------------------------------------------------------
\50\ See NFA, Case Summary: Infinium Capital Management, NYME
10-7565-BC (Nov. 25, 2011), available at http://www.nfa.futures.org/basicnet/Case.aspx?entityid=0338588&case=10-7565-BC+INFINIUM+CAPITAL+MGMT&contrib=NYME.
\51\ See NFA, Case Summary: Infinium Capital Management, CME 09-
06562-BC (Nov. 25, 2011), available at http://www.nfa.futures.org/basicnet/Case.aspx?entityid=0338588&case=09-06562-BC&contrib=CME.
---------------------------------------------------------------------------
In another example, in 2012 a securities trading firm, Knight
Capital Group, launched new software on the NYSE that conflicted with
already existing code.\52\ At the time, the firm was one of the largest
participants and a market maker on the NYSE. The firm's ATS
inadvertently established larger positions than intended, resulting in
a $440 million loss for the firm. The malfunction impacted the broader
market, creating swings in the share prices of almost 150 companies,
and the high volatility linked to the algorithm designed by the firm
also triggered pauses in the trading of five stocks. In addition to the
software malfunction itself, some have reported that there was a delay
of approximately 40 minutes before humans intervened.\53\
---------------------------------------------------------------------------
\52\ See Strasburg & Bunge, supra note 2.
\53\ See SEC Roundtable Transcript, supra note 2, at 55-56.
---------------------------------------------------------------------------
A leading example of ATS malfunction that impacted both the
derivatives and securities markets in the Flash Crash of May 2010. As
described in detail in section II.A.3. above, the Flash Crash
illustrates the potential consequences of ATS design flaws as an
automated execution algorithm failed to take price or time variables
into account, and feedback loops triggered continued orders from the
algorithm even as prices moved far beyond traditional daily ranges.\54\
Finally, the Commission notes the recent systems malfunction at Goldman
Sachs Group Inc. that inadvertently flooded U.S. options markets with a
large number of unintended orders.\55\
---------------------------------------------------------------------------
\54\ See CFTC and SEC Joint Report on the Market Events of May
6, 2010, supra note 1.
\55\ See Jacob Bunge, Kaitlyn Kiernan & Justin Baer, ``Bad
Trades' Ripple Effect,'' W. St. J. (Aug. 21, 2013), available at
http://online.wsj.com/article/SB10001424127887324165204579026611410016876.html.
---------------------------------------------------------------------------
In addition to ATSs, trading platforms have also suffered
malfunctions and illustrate another area in which market disruptive
events can occur. In November 2010, for example, untested code changes
implemented by a U.S. stock exchange operator resulted in errors within
its trading platforms. As a result, the platforms overfilled orders in
over 1,000 stocks, resulting in $773 million of unwanted trading
activity.\56\ In March 2012, a software problem on BATS Global Markets,
whose software had undergone testing, led to a disruption of the
exchange's own IPO. The glitch caused opening orders for ticker symbols
beginning within a certain letter range to become inaccessible on the
platform.\57\ Once the system failed, circuit breakers were triggered
and erroneous trades were cancelled.\58\ In May 2012, Facebook's IPO
experienced significant problems as a result of technical errors on
Nasdaq OMX Group Inc.'s U.S. exchange.\59\ Many customer orders from
both institutional and retail buyers were unfilled for hours or were
never filled at all, while other customers ended up buying more shares
than they had intended. Finally, the Commission notes the recent three-
hour halt in trading on the Nasdaq, which according to reports was
caused when the exchange experienced a disruption in its stock quote
dissemination systems and a disruption in its connectivity with another
trading platform's systems.\60\
---------------------------------------------------------------------------
\56\ See Securities and Exchange Act Release No. 65556, In the
Matter of EDGX Exchange, Inc., EDGA Exchange, Inc. and Direct Edge
ECN LLC (Oct. 13, 2011), available at http://www.sec.gov/litigation/admin/2011/34-65556.pdf; see also SEC News Release, 2011-208, ``SEC
Sanctions Direct Edge Electronic Exchanges and Orders Remedial
Measures to Strengthen Systems and Controls'' (Oct. 13, 2011),
available at http://www.sec.gov/news/press/2011/2011-208.htm.
\57\ See Olivia Oran, Jonathan Spicer, Chuck Mikolajczak &
Carrick Mollenkamp, ``BATS exchange withdraws IPO after stumbles,''
Reuters (Mar. 24, 2012), available at http://uk.reuters.com/article/2012/03/24/us-bats-trading-idUKBRE82M0W020120324; Michael J. De La
Merced & Ben Protess, The N.Y. Times Dealbook (Mar. 25, 2012),
available at http://dealbook.nytimes.com/2012/03/25/little-fallout-expected-from-bats-trading-error/.
\58\ See id.
\59\ See Jenny Strasburg and Jacob Bunge, ``Social network's
debut on Nasdaq disrupted by technical glitches, trader confusion,''
Wall St. J. (May 18, 2012), available at http://online.wsj.com/article/SB10001424052702303448404577412251723815184.html?mod=googlenews_wsj; Jenny Strasburg, Andrew Ackerman & Aaron Lucchetti, ``Nasdaq
CEO Lost Touch Amid Facebook Chaos,'' Wall St. J. (June 11, 2012),
available at http://online.wsj.com/article/SB10001424052702303753904577454611252477238.html.
\60\ See Chris Dieterich & Jacob Bunge, ``Nasdaq Offers Details
on Trading Outage,'' Wall St. J. (Aug. 23, 2013), available at
http://online.wsj.com/article/SB10001424127887324165204579030681671164404.html.
---------------------------------------------------------------------------
Taken together, these events illustrate the importance of effective
testing, circuit breakers, and error trade policies as vehicles for
reducing the likelihood of disruptive events and mitigating their
impact when they occur.\61\ A number of the risk controls contemplated
in this Concept Release could help limit the extent of market
disruption caused by ATS or trading platform malfunctions similar to
those described above. For example, an order ``kill switch'' enables a
market participant to immediately cancel all working orders generated
by one or more of its ATSs, and prevents the submission of additional
orders until the appropriate natural persons allow order placement to
resume. Such a kill switch could be operated by the market participant
generating orders, the clearing firm guaranteeing its trades, or the
trading platform on which its orders would be executed. As another
example, ATS monitoring and supervision standards, as well as pre-
established crisis management protocols, could help ensure that human
supervisors intervene quickly when ATSs experience degraded
performance, and that supervision staff have the both the authority and
knowledge to intervene as required. Further, requiring exchanges to
calculate and disseminate market quality metrics could enable both
exchanges and market participants to better anticipate and mitigate
destabilizing events. In addition, the Commission believes that change
management standards that are beneficial to ATSs could also be applied
to trading platforms to help prevent operational or programming errors
in that element of the automated trading environment. In section III
below, the
[[Page 56550]]
Commission seeks public comment on these and other potential risk
controls.
---------------------------------------------------------------------------
\61\ In addition, although in some ways distinct from the events
summarized above, the Commission notes the significant impact of
Hurricane Sandy in October 2012. U.S. stock markets closed for two
days partially in response to concerns over preparedness to trade
exclusively on electronic venues while trading floors were
potentially closed, as well as the availability of technology and
other relevant personnel. See Jenny Strasburg, Jonathan Cheng &
Jacob Bunge, ``Behind Decision To Close Markets,'' Wall St. J. (Oct.
29, 2012), available at http://online.wsj.com/article/SB10001424052970204789304578087131092892180.html.
---------------------------------------------------------------------------
III. Potential Pre-Trade Risk Controls, Post-Trade Reports, System
Safeguards, and Other Protections
A. Overview of Existing Industry Practices
The transition to automated trading in derivatives markets, as
described above, has been followed by an evolution in what market
participants, regulators and others understand to be necessary risk
controls for various points in the order and trade lifecycle. Many of
the measures identified herein are consistent with recommendations made
by industry groups, other regulatory authorities, international
standard setting bodies, and others. Certain measures, or variants of
them, have been discussed within the futures industry for some time, or
may already be in operation at one or more exchanges, clearing members,
or market participants. For example, the system safeguards pertaining
to the cancellation of orders or disconnecting a market participant in
emergency situations are similar to proposals made separately by FIA's
Principal Traders Working Group and Market Access Working Group in 2010
and the TAC's Pre-Trade Functionality Subcommittee in 2011.\62\
---------------------------------------------------------------------------
\62\ See FIA Principal Traders Group, ``Recommendations for Risk
Controls for Trading Firms,'' (November 2010) at 5 [hereinafter,
``FIA Recommendations for Risk Controls''], available at http://www.futuresindustry.org/downloads/Trading_Best_Pratices.pdf; FIA
Market Access Recommendations, supra note 23, at 9; TAC Pre-Trade
Functionality Subcommittee DMA Recommendations, supra note 26, at 5.
---------------------------------------------------------------------------
The Principal Traders Group also addressed the need to properly
monitor ATSs in its 2010 recommendations by noting that ``firms must
ensure their [ATSs] are supervised at all times while operating in the
markets. Staff must have training, experience and tools that enable
them to monitor and control the trading systems and troubleshoot and
respond to operational issues in a timely and appropriate manner. Firms
should have processes and procedures to ensure trading operations staff
is trained on the expected operating parameters of any [ATS] for which
they are responsible.'' \63\ ATS design and operation was addressed by
FIA's Market Access Working Group and by ESMA, the latter requiring
that market participants ``make use of clearly delineated development
and testing methodologies'' for ATSs prior to their deployment or the
deployment of system updates.\64\ Among other considerations, ESMA
emphasized that ATS testing should address embedded compliance and risk
management controls and operation during stressed market conditions.
---------------------------------------------------------------------------
\63\ See FIA Recommendations for Risk Controls, supra note 62,
at 3.
\64\ See FIA Market Access Recommendations, supra note 23; ESMA
Guidelines on Systems and Controls, supra note 4, at 33.
---------------------------------------------------------------------------
As with the pre-trade and post-trade risk controls, certain system
safeguards would be applicable to more than one entity or would require
coordination between entities. For example, ATS design and operation
tests will require that trading platform operators provide suitable
test environments that accurately recreate the ``live'' trading
platform. Similarly, safeguards that provide for the immediate
disconnection of a market participant in the event of emergency or
breach of tolerances should be available to the market participant, its
clearing firm, and the relevant trading platform so that all parties
have the capacity to initiate a disconnect when necessary. As with
other overlapping measures contemplated in this Concept Release, the
Commission requests public comment regarding the necessity of such
overlaps and the most efficient way to administer them.
1. Existing DCM Risk Controls
Risk controls implemented by one or more exchanges broadly address
market stability. One large DCM (``DCM A'') employs price reasonability
validation controls (aimed at preventing ``fat finger'' type errors)
and position validation controls (both absolute limits and net long/
short limits). In addition, DCM A has implemented a circuit breaker
protection against price spikes. This control provides floor and
ceiling price limits within a specific timeframe and market, and
recalculates new floor and ceiling price limits based on current market
prices for each new timeframe. If the floor or ceiling price is
exceeded, the market is put in a ``hold'' state, although trading will
not be halted in the opposite direction of the hold. The length of the
hold varies depending on the market and orders submitted during the
hold state will remain in the order book but will not be matched. DCM A
has also implemented kill switches that provide it and risk managers at
trading firms with the ability to halt trading.
Similarly, another large DCM (``DCM B'') also employs a limit price
to each market order and stop order to prevent orders from being filled
at significantly aberrant price levels, and maximum order size
protection to prevent entry of erroneous orders for quantities above a
designated threshold. DCM B employs a functionality that introduces a
5-20 second market pause when triggered stops would cause the market to
trade outside of predefined values. This is designed to prevent
excessive price movements caused by cascading stop orders. DCM B also
employs a functionality that introduces a 5-20 second market pause when
a sub-second, extreme market move occurs as a result of order entry.
This functionality is designed to detect significant price moves of
futures contracts occurring within a predetermined period of time, and
triggers a pause in matching activity to provide time for additional
resting orders to populate the order book.
DCM A seeks to optimize message flow through both hard limits and
market incentives. It employs a message throttle limit which sets a
maximum message rate per second for each user session and prevents the
submission of messages in excess of the maximum rate. The second form
of message control used by DCM A is a system of fees based on Weighted
Volume Ratio (``WVR'') calculations designed to discourage inefficient
messaging among firms with high message volumes. The WVR is a ratio
between the number of messages submitted by a market participant and
the total volume of orders that it executes. The ratio of unfilled
orders is also weighted based on how far away from the best bid or
offer each unfilled order was when it was entered. Orders that are
farther away from the best bid or offer when entered are weighted more
heavily. The DCM assesses fees against market participants when they
exceed WVR limits.
DCMs A and B both employ an ``orders removed upon logout'' function
in which all orders are removed upon the user's logout or
disconnection, and that they maintain error trade policies that
incorporate a no cancellation range.
With respect to ATSs, DCMs A and B both employ a certification and
testing process for connecting entities. For example, one DCM described
this process as testing a firm's messaging ability (i.e., that firm's
ability to send and receive data). As part of the testing process, the
DCM will transmit market data to the firm and this provides the firm
with the opportunity to run its own algorithms and for that firm to
determine if its algorithms are functioning as it intended. Firms must
pass additional conformance tests when the exchange's own system
functionality changes. DCM B indicated that its testing process allows
customers to test
[[Page 56551]]
new products prior to their production launch.
In addition to their internal risk mitigation programs, DCMs also
provide risk mitigation tools to intermediaries such as FCMs, allowing
the intermediaries to set risk control parameters on controls that
reside at the trading platform level. Clearing firms, for example, are
able to set risk tolerance levels for their customers based on position
size, order activity, executions, among other variables.
2. Existing Trading and Clearing Firm Risk Controls
Risk controls at the level of individual market participant firms,
whether trading firms or clearing firms, are necessarily entity
specific. Accordingly, industry groups have collaborated to determine
best practices for risk controls. As noted previously, other entities,
including the TAC, have also developed best practices or
recommendations. One goal of this Concept Release is to determine how
consistently these, and other, recommendations are today being
implemented by market participants. As noted by FIA, ``all principal
traders have a vested interest in well-functioning markets with
effective risk controls, clear error trade policies that focus on trade
certainty, and a strong regulatory framework.'' \65\ Comments to this
Concept Release will allow the Commission to best ensure this strong
framework. Questions about the general use of automated risk controls
at the level of a firm are also informed by two reports prepared by
authors affiliated with the Federal Reserve Bank of Chicago. One report
details the current practices of nine proprietary trading firms, with
special attention to risk mitigating practices currently applied to
their automated systems.\66\ Through interviews, the authors found that
(1) all firms have maximum order sizes in place and intraday position
limits; (2) all but one firm has credit limits by account, which
monitor open positions, dollar value of positions and quantity of
working orders; \67\ (3) half of the firms have price protection points
for orders; (4) most firms had message throttles, set at order volume
per unit of time; and (5) all firms had kill buttons. The risk controls
included in this list, and others discussed within the report, are
expanded upon in the below discussion. In its questions for comment,
the Commission seeks to understand what types of risk controls are most
commonly used throughout the industry, and the degree to which those
risk controls are standardized across the industry.
---------------------------------------------------------------------------
\65\ See FIA Recommendations for Risk Controls, supra note 62,
at 2.
\66\ See Carol Clark & Rajeev Ranjan, ``How Do Proprietary
Trading Firms Control the Risks of High Speed Trading?'' (March
2012), available at http://www.chicagofed.org/digital_assets/publications/policy_discussion_papers/2012/PDP2012-1.pdf.
\67\ The final firm also sets credit limits, but only for new
traders. See id. at 7.
---------------------------------------------------------------------------
A second report \68\ summarized interviews with five Broker/Dealers
(``B-Ds'') and FCMs, again detailing their current practices in
automated risk controls. As at the trading level, some firms have
implemented pre-trade and post-trade checks, along with other credit
related controls to mitigate trading losses and resulting burdens on
the clearing firm. The report details categories of risks considered by
the B-D or FCM when signing on a new client, or updating controls as a
client enters new businesses or expands on old ones. These include:
Credit risks, market risks, counterparty risks, portfolio risks and
regulatory risks. Through these assessments, clearing firms are able to
determine appropriate risk thresholds for a given client, and apply
them as necessary at multiple points in the trading chain. Specific
controls come in forms quite similar to those outlined above in the
case of the trading firm. Pre-trade risk controls span order size
limits, intraday position limits, credit limits, and message throttles.
These can vary by asset class, exchange, and other market factors,
along with coincident market dynamics such as volatility levels and
current positions of the trading firm. The monitoring done by the
clearing firm is aided by post-trade measures such as the drop-copy of
executions, which allows for the monitoring of positions and associated
credit risks.
---------------------------------------------------------------------------
\68\ See Carol Clark & Rajeev Ranjan, ``How Do Broker-Dealers/
Futures Commission Merchants Control the Risks of High Speed
Trading?'' (June 2012), available at http://www.chicagofed.org/Webpages/publications/policy_discussion_papers/2012/pdp_3.cfm.
---------------------------------------------------------------------------
B. Overview of Risk Controls Addressed in This Concept Release
The risk controls presented below describe specific measures which
could be taken by exchanges and participants in automated trading
environments. To better understand current industry practices, the
Commission is interested in determining, for each risk control: (1)
Whether the entity commenting has implemented the control; (2) whether
the entity believes implementation of the control within the
marketplace is consistently applied; and (3) the benefits and costs of
a regulatory mandate of the control. If the Commission determines that
the types of risk controls employed across the industry vary widely,
the Commission would be aided by understanding the extent of this
variance, the reasons for it, and whether regulatory standardization
can be of benefit. By gathering this information, the Commission will
be better informed regarding beneficial future regulation surrounding
automated systems.
The Commission emphasizes that this Concept Release is intended to
serve as a high-level enunciation of potential measures intended to
reduce the likelihood of market disrupting events and mitigate their
impact when they occur. Many of the risk controls listed below are in
effect, in part or in full, across multiple entities. Others have been
included in recommendations by industry groups and standard-setting
bodies, or addressed by foreign regulatory authorities. The Commission
also notes that a number of the measures described below offer similar
risk controls at various stages in the life of an order (e.g., a
safeguard applicable to the ATS generating an order and a similar
safeguard applicable to the trading platform receiving such order).
Added security through redundancy of risk controls is a feature of
safeguard documents reviewed by the Commission in preparing this
Concept Release. The Commission seeks public comment on merits of
single versus redundant risk control models. Market participants and
members of the public are encouraged to comment on the potential risk
controls, and the Commission anticipates further refinement of the
measures described herein based on the comments received.
The discussion of risk controls below is followed by a number of
general questions on which the Commission requests comment (see section
III.G. below). These questions are applicable to all the risk controls
discussed below.
C. Pre-Trade Risk Controls
The Commission includes below a set of pre-trade risk controls
aimed at reducing market disruptions related to automated trading due
to errors, system malfunctions or other events with similar effects. In
general, pre-trade risk controls seek to protect against the
accumulation of a large volume of orders, executions, or positions over
an abbreviated period of time. Some market participants are currently
using controls which address this accumulation, including maximum order
size limits, message rate limits, and similar measures. Pre-trade risk
controls can also promote fair and orderly markets, through the use of
circuit breakers, execution throttles and self-trade
[[Page 56552]]
monitoring. Finally, the pre-trade risk controls also include pre-trade
credit limits designed to protect clearing firms, and their clients,
with respect to customer and proprietary orders.\69\ Each of these
groups is discussed below in greater detail.
---------------------------------------------------------------------------
\69\ The pre-trade risk controls contemplated herein are
consistent with general principles or specific recommendations (in
DMA context) expressed in the TAC Pre-Trade Functionality
Subcommittee DMA Recommendations, supra note 26, at 2-5; IOSCO
Technical Committee, Final Report on Principles for Direct
Electronic Access to Markets (August 2010) at 20, available at
http://www.iosco.org/library/pubdocs/pdf/IOSCOPD332.pdf; and the FIA
Recommendations for Risk Controls, supra note 62, at 4. The pre-
trade risk controls described herein are also consistent with the
principles included in the ESMA Guidelines on Systems and Controls,
supra note 4.
---------------------------------------------------------------------------
In order to fully address possible disruptions, the pre-trade risk
controls apply at one or more of three points in the execution chain:
(1) Individual firms; (2) intermediaries of many forms (including SDs,
MSPs, FCMs, Floor Traders, Commodity Pool Operators (``CPOs'') and
DCOs); and (3) exchanges (including DCMs and SEFs). In many cases, the
same or similar risk controls are implemented at more than one point in
the execution chain, such as first at the firm, then perhaps at the
clearing firm, and then finally at the DCM. The Commission believes
that this approach offers a number of advantages.\70\ First, it allows
individual entities to calibrate the relevant risk control in
accordance with their own objectives and risk tolerances. For example,
an exchange may set a per-product maximum order size to ensure orderly
trading in its markets, with the same limit applying equally to all
market participants. A clearing firm, however, may wish to address its
customers' distinct risk profiles by setting different maximum order
sizes for different customers.
---------------------------------------------------------------------------
\70\ In this regard, the Commission notes that the TAC's Pre-
Trade Functionality Subcommittee described ``three levels in the
electronic trading `supply chain' where pre-trade risk safeguards
could happen: Trading firms (as principal or agent), clearing firms
(as principal or agent), and exchanges.'' The Subcommittee's
recommendations to the TAC noted that it ``believe[s] strongly that
all three levels of the supply chain should institute pre-trade risk
management measures.'' See TAC Pre-Trade Functionality Subcommittee
DMA Recommendations, supra note 26, at 1.
---------------------------------------------------------------------------
Second, by indicating that some risk controls should reside at the
exchange level in addition to the market participant and clearing firm
levels, the Commission is responding to competitive and ``race to the
bottom'' concerns raised by several observers. FIA's Market Access
Working Group, for example, noted that ``[p]re-trade risk controls have
become a point of negotiation between trading firms and clearing
members because they can add latency to a trade. To avoid such
negotiations, the Market Access Working Group believes that certain
risk controls should reside at the exchange level and be required for
all trading to ensure a level playing field.'' \71\
---------------------------------------------------------------------------
\71\ See FIA Market Access Recommendations, supra note 23, at 8.
See also TAC Pre-Trade Functionality Subcommittee DMA
Recommendations, supra note 26, at 2. The TAC Pre-Trade
Functionality Subcommittee called for a ``realistic view'' of the
incentives under which market participants, clearing firms, and
exchanges operate. The Subcommittee identified these incentives as
follows:
``Trading firms are competing with one another to have
the smallest time delays (lowest latency) in getting their orders
into the exchange's matching engine, and are thus negotiating with
brokers to reduce latency. At the same time they are trying to
protect their capital from rogue trading, technological deficiencies
or other adverse, unintended events.
Brokers (clearing FCMs) are competing with one another
to attract the business of these high-volume, speed-seeking trading
firms, and are thus trying to reduce latency. At the same time, they
are trying to protect themselves from loss due to unauthorized
trading by their trading firm clients or other adverse, unintended
events.
Exchanges (Designated Contract Markets, or DCMs, and
Foreign Boards of Trade, or FBOTs) are competing with one another to
provide low latency execution, and will soon be competing with Swaps
Execution Facilities (SEFs), to attract the business of these
trading firms.''
The Subcommittee expressed its concern that risk controls should
ensure fairness so that one trading firm is not disadvantaged
relative to another ``because its clearing firm chose to act more
responsibly.''
---------------------------------------------------------------------------
Third, the risk controls listed below acknowledge a variety of
industry practices with respect to order generation, such as whether
the order passes through intermediaries prior to execution. The
Commission seeks to understand how increased standardization in risk
controls at the level of exchanges or exchange members could provide
strengthened protection for the markets and the public.\72\ Notably, if
the Commission were to require the placement of credit controls,
maximum order size limits, and maximum message rate limits at both
exchanges and clearing members, it could address both traditional means
of order flow (i.e., through a clearing firm) and newer DMA practices,
which require controls at the exchange set by the relevant clearing
firm. In combination, these reasons demonstrate the strength, in
certain cases, of putting into practice standardized risk controls,
with similar goals, at multiple entity types.\73\
---------------------------------------------------------------------------
\72\ For example, trading platforms provide a range of risk
controls, but there is limited standardization in the types of risk
controls available to customers from one exchange to the next. The
Commission seeks to understand whether diverse risk management tools
and policies at various exchanges complicate risk management for
intermediaries and traders.
\73\ The Commission notes that some existing regulations address
pre-trade risk controls. See supra section II.B.
---------------------------------------------------------------------------
Finally, the Commission notes the importance of risk controls
designed to protect the financial integrity of DCOs, and to address
risks posed by market participants utilizing DMA. Throughout the range
of pre-trade risk controls discussed below, and other measures
discussed later in this Concept Release, the Commission specifically
solicits public comment regarding the following questions:
6. Are there distinct pre-trade risk controls, including measures
not listed below, or measures in addition to those already adopted by
the Commission, that would be particularly helpful in protecting the
financial integrity of a DCO?
7. Are there distinct pre-trade risk controls, including measures
not listed below, or measures in addition to those already adopted by
the Commission, that should apply specifically in the case of DMA?
The following sections describe the pre-trade risk controls
inquired about in this Concept Release, and present a series of
questions to assist the Commission in determining the effectiveness,
adoption rate, and need for any additional action with respect to these
pre-trade risk controls or others that commenters may think advisable.
1. Message and Execution Throttles
The Commission seeks public comment regarding the potential
benefits and existing use of maximum message rate and execution rate
throttles (``execution throttles''). The Commission also seeks public
comments regarding the types of execution throttles that would be most
effective at alerting market participants to potential algorithm
malfunctions and limiting the extent of market disruption when there is
a malfunction.\74\
---------------------------------------------------------------------------
\74\ The Commission understands that some trading firms and
several exchanges already have limits on the number of orders that
can be sent to a trading venue during a specified period of time.
See Clark & Ranjan, ``How Do Proprietary Trading Firms Control the
Risks of High Speed Trading,'' supra note 66, at 7; Oliver Linton &
Maureen O'Hara, ``Economic impact assessments on MiFID II policy
measures related to computer trading in financial markets,'' United
Kingdom Government Office for Science--Foresight (August 2012) at
24-25, available at: http://www.futuresindustry.org/epta/downloads/Economic-Impact-assessments-on-MiFID-2-policy-measures_083012.pdf.
However, the Commission would like to understand whether requiring
some measure of standardization and the use of such tools among
exchanges, FCMs, and trading firms would provide additional
protection for the market.
---------------------------------------------------------------------------
[[Page 56553]]
Execution throttles prevent an algorithm from exceeding its
expected message rate or rate of execution, and when tripped, can alert
monitors at both the exchange and the trading firm. Such alerts can
facilitate rapid detection of malfunctioning algorithms. Depending on
the nature of the malfunction, execution throttles may also reduce the
damage and monetary losses caused by the disruptive algorithm during
the time when it is being investigated. The Commission understands that
trading firms \75\ and exchanges \76\ employ individual variants of
throttles to limit the number of orders that can be transmitted to or
processed by an exchange. The Commission requests public comment
regarding the extent to which market participants that already utilize
execution throttles apply them in a static manner (i.e., a fixed
threshold, beyond which notifications are generated), or dynamically
(i.e., dependent on the time of day or the previous activity of the
algorithm).\77\ The Commission also requests public comments regarding
the extent to which throttles are applied by trading firms on a per-
algorithm basis, calibrated to take into account the expected message
and execution rates of each algorithm for a given time period.
---------------------------------------------------------------------------
\75\ See Clark & Ranjan, ``How Do Proprietary Trading Firms
Control the Risks of High Speed Trading,'' supra note 66, at 7.
\76\ See Carol Clark & Rajeev Ranjan, ``How Do Exchanges Control
the Risks of High Speed Trading?'' (November 2011) at 3, available
at http://www.chicagofed.org/digital_assets/publications/policy_discussion_papers/2011/PDP2011-2.pdf.
\77\ The Commission notes that the Futures and Options
Association (``FOA'') expressed the opinion that throttles may
hinder price formation and market integrity if applied dynamically
during a period of market stress. However, the FOA generally
supported the use of throttles that are ``pre-defined, transparent
and certain (i.e., the member obtains connections with a specified
bandwidth in terms of maximum messages per second).'' See FOA,
``ESMA's Consultation Paper: Guidelines on Systems and Controls in a
highly automated trading environment for trading platforms,
investment firms and competent authorities: A response paper by the
Futures and Options Association'' (October 2011) at 2, available at
http://www.esma.europa.eu/system/files/11-FOA.pdf.
---------------------------------------------------------------------------
In addition, the Commission asks whether maximum message rates and
execution throttles could be used as a mechanism to prevent individual
entities from submitting messages or executing orders at speeds that
are misaligned with their risk management capabilities. Execution
throttles of this type would be unique to individual firms or accounts,
and could be set by the exchange or clearing firm after reviewing the
risk management capabilities of the entity to which the throttle will
apply. For some firms, there may be a delay before effective risk
management begins; in these cases, execution throttles may mitigate
harm to the firm or other market participants prior to the firm's
response to a malfunction. Last, message rate limits could be used to
mitigate the risk of manipulative or disruptive messaging strategies
such as ``order stuffing,'' where firms use ATSs to submit large
numbers of orders that are cancelled before execution in order to slow
down the matching engine and create arbitrage opportunities in or
across products.
8. If, as contemplated above, maximum message rates and execution
throttles were used as a mechanism to prevent individual entities or
accounts from trading at speeds that are misaligned with their risk
management capabilities, how should this message rate be determined?
9. Message and execution throttles may be applied by trading firms
(FCMs and proprietary trading firms), clearing firms, and by exchanges.
The Commission requests public comment regarding the appropriate
location for message and execution throttles.
a. If throttles should be implemented at the trading firm level,
should they be applied to all ATSs, only ATSs employing HFT strategies,
or both?
b. What role should clearing firms play in the operation or
calibration of throttles on orders submitted by the trading firms whose
trades they guarantee?
10. Should the message and execution throttles be based on market
conditions, risk parameters, type of entity, or other factors?
11. What thresholds should be used for each type of market
participant in order to determine when a message or execution throttle
should be used? Should these thresholds be set by the exchange or the
market participant?
12. Are message and execution thresholds typically set by contract,
or by algorithm? What are the advantages and disadvantages to each
method?
13. Who should be charged with setting message rates for products
and when they are activated?
14. Would message and execution throttles provide additional
protection in mitigating credit risk to DCOs?
2. Volatility Awareness Alerts
Automated volatility awareness alerts implemented by trading firms
are another form of risk control contemplated in this Concept Release.
Volatility awareness alerts could be triggered when price movements in
a given product move beyond a certain threshold within a previously
specified time period. Such alerts could assist in identifying market
conditions that may exceed an algorithm's parameters, or may highlight
unintended effects of an algorithm's orders. Given an alert, human
monitors at the trading firm could then intervene either by halting the
relevant algorithms under their control, or by conveying the
information to other relevant parties. Unlike exchange trading pauses
and halts, volatility awareness alerts inform firm personnel as to
changes in market conditions that may disrupt the parameters within
which their ATSs and algorithms were programmed to operate, rather than
immediately triggering a pause in trading.
15. The Commission is aware that alarms can be disruptive or
counterproductive if ``false alarms'' outnumber accurate ones. How can
volatility alarms be calibrated in order to minimize the risk that
false alarms could interrupt trading or cause human monitors to ignore
them over time?
3. Self-Trade Controls
A trade that results from the matching of opposing orders between a
firm or a single or commonly owned account, such as a wash trade, does
not shift risk between different market participants. In addition, such
trades may inaccurately signal the level of liquidity in the market and
may result in a non-bona fide price. Risk controls that identify and
limit self-trading may result in more accurate indications of the level
of market interest on both sides of the market and help ensure arms-
length transactions that promote effective price discovery. Some
regulated exchanges have tools specifically designed to identify and
limit self-trading. The Commission is interested in better
understanding those risk controls and how widespread their use may be.
For example, the Commission understands that in June 2013, CME
Group introduced a voluntary self-match prevention functionality that
allows market participants to prevent buy and sell orders for the same
account (or for an account with common beneficial ownership) from
matching with each other.\78\ Market participants
[[Page 56554]]
that wish to opt-in to this functionality populate a new FIX tag on all
orders with a ``Self Match Prevention Identifier,'' in addition to an
executing firm number. When the exchange's matching engine detects buy
and sell orders at the same executable price level in a particular
contract and both orders have the same Self Match Prevention Identifier
and the same executing firm number, the engine will automatically
cancel the resting order(s) on one side of the market and process the
incoming order on the other side of the market.
---------------------------------------------------------------------------
\78\ See CME Group, ``CME Globex Self-Match Prevention
Functionality FAQ'' (2013), available at http://www.cmegroup.com/globex/resources/smpfaq.html. On July 9, 2013, CME Group requested
Commission approval to issue a market regulation advisory notice
intended to provide guidance with respect to the types of activity
that may constitute a violation of the exchange's wash trades rule
and to provide additional information concerning its self-match
prevention technology. This notice, which is under review by the
Commission, is available at http://www.cftc.gov/stellent/groups/public/@rulesandproducts/documents/ifdocs/rul070913cmecbotnymexcomandkc1.pdf.
---------------------------------------------------------------------------
In addition, the Commission understands that ICE Futures U.S.
(``ICE'') offers voluntary self-trade prevention functionality for
preventing inter- and intra-company orders from matching in the
exchange's matching engine. This functionality was initially designed
to prevent the matching of inter- and intra-company trades by
automatically rejecting the taking order. The Commission understands
that in May 2013, this functionality was expanded to allow for the
rejection of the resting order.
16. What specific practices or tools have been effective in
blocking self-trades, and what are the costs associated with wide-
spread adoption of such practices or tools?
17. Please indicate how widely you believe exchange-sponsored self-
trading controls are being used in the market.
18. Should self-trade controls cancel the resting order(s)? Or,
instead, should they reject the taking order that would have resulted
in a self-trade? If applicable, please explain why one mechanism is
more effective than the other.
19. Should exchanges be required to implement self-trading controls
in their matching engines? What benefits or challenges would result
from such a requirement?
20. Please explain whether regulatory standards regarding the use
of self-trading control technology would provide additional protection
to markets and market participants.
21. If you believe that self-trading controls are beneficial,
please describe the level of granularity at which such controls should
operate (e.g., should the controls limit self-trading at the executing
firm level? At the individual trader level?) What levels of granularity
are practical or achievable?
22. If you believe that self-trading controls are beneficial,
please explain whether exchanges should require such controls for
market participants and identify the categories of participants that
should be subject to such controls. For example, should exchanges
require self-trading controls for all participants, some types of
participants, participants trading in certain contracts, or
participants in market maker and/or incentive programs? What benefits
or challenges would result from imposing such controls on each category
of participant?
4. Price Collars
The Commission is also inquiring about price collars for both
orders and executions. Price collars on orders prevent orders outside
of acceptable price ranges from either entering the order book or
executing at extreme levels; in effect, collars prevent market or stop
orders (which execute as market orders) from trading at levels far
beyond that expected at order entry. Similarly, price collars for
execution prevent an order that is already in the book from being
executed by the matching engine if it is outside of the acceptable
range. Price collars can be contract specific and dynamic, responding
to changes in market prices and market volatility for each contract.
Price collars may reduce realized volatility by preventing a large,
aggressive order from sweeping the book and matching at prices outside
the range allowed by the collar, or allowing isolated market orders to
execute during periods when one-sided liquidity is extremely low.\79\
---------------------------------------------------------------------------
\79\ The Commission currently estimates that about half of the
trading firms operating ATS have limits that check orders against a
specific price range before sending them to the exchange. See Clark
& Ranjan, ``How Do Proprietary Trading Firms Control the Risks of
High Speed Trading,'' supra note 66, at 7. However, the Commission
would like to better understand whether standardizing such controls
at the level of exchanges or requiring such controls at the level of
trading firms would further promote stable and reliable markets.
---------------------------------------------------------------------------
23. The Commission is aware that some exchanges already have price
collars in place for at least a portion of the contracts traded in
their markets. Please comment on whether exchanges should utilize price
collars on all contracts they list.
24. Would price collars provide additional protection in mitigating
credit risk to DCOs?
5. Maximum Order Sizes
Maximum order sizes are intended to protect against execution of
orders for a quantity larger than a predetermined ``fat finger'' limit.
Like other controls, these limits can function at multiple levels; for
example, at the firm level, in which firms prevent the submission of
orders beyond certain limits, or at the clearing level, in which
clearing members prohibit transmission of customer orders in excess of
predetermined limits.
The Commission believes that most, if not all, exchanges currently
have the capability to set maximum order sizes, but understands that
such controls may vary among exchanges in their ability to set limits
by product, product class, customer, or clearing member.\80\ The
Commission is interested to understand the following:
---------------------------------------------------------------------------
\80\ See Carol Clark & Rajeev Ranjan, ``How Do Exchanges Control
the Risks of High Speed Trading?'' supra note 76, at 3.
---------------------------------------------------------------------------
25. Are such controls typically applied to all contracts and
customers, or on a more limited basis?
26. Do exchanges allow clearing members to use the exchange's
technology to set maximum order sizes for specific customers or
accounts?
27. Would additional standardization in the capabilities of this
technology or more uniform application of this technology to all
customers and contracts improve the effectiveness of such controls?
The Commission understands that some, but perhaps not all clearing
firms may utilize the exchange's systems, and possibly their own
systems, in order to conduct pre-trade maximum order size screens.\81\
The Commission is interested to understand the following:
---------------------------------------------------------------------------
\81\ See, e.g., Carol Clark, Rajeev Ranjan, John McPartland, &
Richard Heckinger, ``What Tools Do Vendors Provide to Control the
Risks of High Speed Trading?'' (October 2011) at 2-3, available at
http://www.chicagofed.org/digital_assets/publications/policy_discussion_papers/2011/PDP2011-1.pdf.
---------------------------------------------------------------------------
28. To what extent are clearing firms and trading firms conducting
pre-trade maximum order size screens? Please explain whether firms are
conducting such screens by utilizing: (1) Their own technology; (2) the
exchange's technology, or (3) a combination of both.
29. Would regulatory standards regarding the use of such technology
provide additional protection to the markets?
6. Trading Pauses
The Commission wants to better understand the existing
implementation of trading pauses for trading platforms, and whether any
additional types of pause mechanisms would be beneficial. A wide range
of pause methodologies are currently in effect at exchanges, such as
stop-logic functionality and interval price limits. These methodologies
include market pauses when the execution of resting stop orders would
cause excessive price movements, when prices move in excess of a
dynamic threshold over a given time period, or simply when prices have
[[Page 56555]]
moved more than a given amount during the trading day.\82\ Often, the
market will monitor the order book during the pause, and determine when
it is ``safe'' to re-open the market to further executions or re-open
after a specified interval. Trading pauses have mitigated price
movements during particularly volatile times in the past.\83\
---------------------------------------------------------------------------
\82\ The Commission understands that some triggers leading to a
market pause are not necessarily best classified as ``pre-trade''
risk controls. Some pauses, as described, may be in anticipation of
a certain set of executions, and are pre-trade, while others may be
in response to a given execution. The discussion here implicitly
includes all of the above, and the Commission requests comment on
the full range of pause types.
\83\ See CFTC and SEC Joint Report on the Market Events of May
6, 2010, supra note 1, at 6 (noting that CME's stop logic
functionality that triggered a halt in E-Mini trading shows that
pausing a market can be an effective way of providing time for
market participants to reassess their strategies, for algorithms to
reset their parameters, and for an orderly market to be re-
established).
---------------------------------------------------------------------------
The Commission is interested in better understanding the relative
costs and benefits of each type of pause functionality and whether
certain types of pause mechanisms are more effective than others with
respect to ATS trading. The Commission is also interested to understand
whether additional types of pause triggers would be advisable. These
might cover a wider array of adverse states of an automated central
limit order book, including, for example, significant depth imbalance,
a significant number of aggressive orders, or a significant number of
cancelled orders.
30. Trading pauses, as currently implemented, can be triggered for
multiple reasons. Are certain triggers more or less effective in
mitigating the effects of market disruptions?
31. Are there additional triggers for which pauses should be
implemented? If so, what are they?
32. What factors should the Commission or exchanges take into
account when considering how to specify pauses or what thresholds
should be used?
33. How should the re-opening of a market after a trading pause be
effected?
7. Credit Risk Limits
Credit risk limits are a valuable protection for limiting the
activity of malfunctioning ATSs. Risk limits are most valuable when
implemented as a pre-execution filter. Alternatively, low-latency post-
trade risk limits may also provide some risk mitigation. Credit risk
controls may be implemented by different entities, including the
trading firms that originate orders, the clearing firms that guarantee
the orders, the trading platforms matching the orders, and the DCOs
that clear the orders. The Commission acknowledges that some trading
firms and FCMs conduct post-trade credit checks with varying degrees of
latency and that pre-trade credit risk screens are already required
pursuant to Sec. Sec. 1.73 and 23.609.\84\ As noted above, however,
the Commission seeks public comments regarding any additional measures
that could help protect the financial integrity of DCOs, including
measures discussed in this Concept Release or other measures that may
be recommended by interested parties.
---------------------------------------------------------------------------
\84\ See Commission, Final Rule: Customer Clearing
Documentation, Timing of Acceptance for Clearing, and Clearing
Member Risk Management, 77 FR 21278 (Apr. 9, 2012).
---------------------------------------------------------------------------
The TAC has received proposed models for implementing certain pre-
trade risk controls for swaps, particularly those pertaining to credit
risk.\85\ Relevant solutions for implementing credit-based pre-trade
risk controls include those in which credit limits reside at the FCM,
at the trading platform (based on instruction from the clearing firm),
or, for example, at a ``hub'' which applies credit controls on a per-
order basis.\86\ The Commission is interested to understand whether the
``hub'' model, one of several proposed solutions received by the TAC,
could be usefully applied to futures markets.
---------------------------------------------------------------------------
\85\ See, e.g., ``Managing Credit Lines in a SEF/Cleared
World,'' a presentation by MarkitServ at the March 29, 2012 TAC
meeting [hereinafter, the ``MarkitServ Presentation'']. Available
at: http://www.cftc.gov/ucm/groups/public/@aboutcftc/documents/file/tacpresentation032912_markitse.pdf.
\86\ See id. The presentation also noted that post-trade checks
at the DCO is another form of risk control based on end-customer
position or credit limits. See section III(D) for additional
discussion of post-trade reports and other post-trade measures.
---------------------------------------------------------------------------
The Commission is also interested in credit risk limits as a
mechanism for limiting the disruptive activity of a malfunctioning ATS.
Therefore, the Commission requests comment on the following:
34. What positions should be included in credit risk limit
calculations in order to ensure that they are useful as a tool for
limiting the activity of a malfunctioning ATS? Is it adequate for such
a screen to include only those positions entered into by a particular
ATS or should it include all the firm's positions?
35. Should pre-trade credit screens require a full recalculation of
margin based on the effect of the order?
36. In light of your answers to the previous two questions, where
in the lifecycle of an order should the credit limits be applied and
what entity should be responsible for conducting such checks?
37. If credit checks are conducted post-trade, what should be done
when a trade causes a firm to exceed a limit?
38. Please describe any technological limitations that the
Commission should be aware of with respect to applying credit limits.
39. The Commission is particularly interested to receive public
comment on the ``hub'' model and its applicability to different types
of pre-trade risk controls. What are the strengths and weaknesses of
this approach relative to other pre-trade or post-trade approaches to
checking trades against credit limits? How would the latency between
the ``hub'' and the exchanges be managed to provide accurate limits for
high frequency ATS?
40. If you believe that post-trade credit checks would be an
effective safeguard against malfunctioning ATSs, what is the maximum
amount of latency that should be allowed for conducting such checks?
What technological or information flow challenges would have to be
addressed in order to implement post-trade checks with that degree of
latency?
41. With respect to any entity that you believe should be
responsible for applying credit risk limits, please describe the
technology necessary to implement that risk control and the cost of
such technology.
The pre-trade risk controls described above are summarized in
Appendix A.
D. Post-Trade Reports and Other Post-Trade Measures
The Commission understands that, even with the presence of the most
robust set of pre-trade risk controls, unanticipated events occur
within a complicated marketplace. For example, the emergence of
unexpected feedback loops between multiple algorithms, or
malfunctioning pre-trade risk controls can lead to unintended order
submissions that adversely impact market quality and investor
confidence. Post-trade reports have the potential to mitigate the
impact of such events, particularly if the post-trade reports are made
available and utilized on a low-latency basis, such that market
participants are quickly aware of any malfunction. Other post-trade
measures, including enhanced error trade policies, may help
counterparties to errant trades to better anticipate and address risk
associated with trade uncertainty when such events occur. The post-
trade reports and other measures are summarized below.
1. Order, Trade, and Position Drop Copy
The Commission is inquiring about the potential advantages of
increased
[[Page 56556]]
standardization of real-time order, trade, and position reports for use
by clearing firms and market participants. Real-time information is
critical to market participants managing the risk of their own, and
their customers' trades. The Commission is inquiring as to the
advisability of requiring all exchanges and DCOs to provide real-time
order and trade reports to each market participant, and the clearing
firm serving that client for that particular trade. This information
would give clearing firms real-time updates of their customers' order
and trading activities.
These reports could improve the effectiveness of automated credit
risk limits, which require current order and trade information in order
to calculate current positions and monitor credit risk effectively. In
some cases order information may be available to a trading platform
before it is available to the relevant clearing member (e.g., in the
case of DMA-enabled participants), and trade information is always
available first to the trading platform. Therefore, there is a strong
interdependency between exchanges, DCOs and clearing firms as the
latter seek to manage their credit risk.
Any time lag in the clearing firm's ability to construct a
retrospective view of their customers' positions could diminish a
clearing firm's ability to assess its customer's risk profile before
such customer enters additional orders or establishes additional
positions and accumulates greater risk.
More generally, widespread use of order and trade reports may be
beneficial in both DMA and non-DMA situations to help market
participants to track all order and trade activity quickly and
efficiently. The Commission notes that some or all DCOs already provide
post-trade information to clearing members, and that some DCOs charge
for that information and others do not.\87\ However, the Commission
believes that the content of the data vary among DCOs and that not all
market participants choose to purchase data when it is available. As
described above, the Commission preliminarily believes that more
standardized access to real-time data from exchanges and DCOs could be
valuable to clearing firms, and possibly to trading firms, as they
manage their risks. The Commission encourages interested parties to
comment, again, on the current use of real-time reports, the
consistency of this use, and the potential benefits and nature of
additional order and trade reports.
---------------------------------------------------------------------------
\87\ See Carol Clark & John McPartland, ``How Do Clearing
Organizations Control the Risks of High Speed Trading?'' (May 2012)
at 6-7, available at http://www.chicagofed.org/Web pages/
publications/policy_discussion_papers/2012/pdp_2.cfm.
---------------------------------------------------------------------------
42. What order and trade reports are currently offered by DCMs and
DCOs? What aspects of those reports are most valuable or necessary for
implementing risk safeguards? Please also indicate whether the report
is included as part of the exchange or clearing service, or whether an
extra fee must be paid.
43. If each order and trade report described above were to be
standardized, please provide a detailed list of the appropriate content
of the report, and how long after order receipt, order execution, or
clearing the report should be delivered from the trading platform to
the clearing member or other market participant.
2. Trade Cancellation or Adjustment Policies
The Commission is interested to know whether it would be beneficial
for exchanges to develop more uniform and objective trade cancellation
or adjustment policies. These policies should apply to cancellation or
adjustment of individual trades, as well as to cancellation or
adjustment of a large quantity of trades in response to a disruptive
market event at the direction of a regulatory body or in accordance
with the exchange's own determination that such cancellation or
adjustment of a large quantity of trades is necessary. The policies
could include (1) Clear principles on when trades will be cancelled or
adjusted; (2) a requirement that traders notify the exchange of error
trades within a specified number of minutes; and (3) a requirement that
the exchange notify market participants of possible adjusted or busted
trades immediately. Requiring traders to notify the exchange quickly
and requiring the exchange to communicate the situation to market
participants immediately helps to ensure that any market participants
potentially affected by impending adjustment or cancellation actions
are made aware of the additional risk they bear and can take steps to
mitigate that risk.
It may be advisable to base cancellation and adjustment policies on
pre-defined, objective criteria in order to minimize the time for
identification and notification. Such criteria may include the minimum
trade size for which cancellation will be considered, the minimum and
maximum range in which a trade will be adjusted, the time a market
participant has to request the cancellation or adjustment, the specific
circumstances under which trades will be adjusted or canceled (e.g., an
exchange system error, specific types of human errors) and factors to
be taken into account (e.g., market conditions, whether other market
participants have relied on the price). Last, the Commission is
inquiring as to the advisability of policies to favor trade adjustment
over trade cancellation in order to help ensure that market
participants are able to keep the positions they have entered into,
even if the prices are adjusted. The Commission is interested in
receiving comments on whether additional standardization in error trade
policies would be beneficial, and whether this prioritization scheme is
appropriate.\88\
---------------------------------------------------------------------------
\88\ The Commission notes that error trade policies may vary for
different exchanges and for different products at each exchange. See
id. at 7.
---------------------------------------------------------------------------
44. Is a measure that would obligate exchanges to make error trade
decisions (i.e., decisions to cancel a trade or to adjust its price)
within a specified amount of time after an error trade is reported
feasible? If so, what amount of time would be sufficient for exchanges,
but would be sufficiently limited to help reduce risk for
counterparties to error trades?
45. Should exchanges develop detailed, pre-determined criteria
regarding when they can adjust or cancel a trade, or should exchanges
be able to exercise discretion regarding when they can adjust or cancel
a trade? What circumstances make pre-determined criteria more effective
or necessary than the ability to exercise discretion, and vice versa?
46. Do error trade policies that favor price adjustment over trade
cancellation effectively mitigate risk for market participants that are
counterparties to error trades? Are there certain situations where
canceling trades would mitigate counterparty risk more effectively? If
so, what are they and how could such situations be identified reliably
by the exchange in a short period of time?
47. Should error trade policies be consistent across exchanges,
either in whole or in part? If so, how would harmonization of error
trade policies mitigate risks for market participants, or contribute to
more orderly trading?
E. System Safeguards
In this Concept Release, the Commission inquires about a range of
system safeguards for trading platforms,\89\ clearing firms, and market
participants (including ATSs). Those system safeguards are intended to
address a number of operational, market
[[Page 56557]]
abuse and transmission risks, and may protect against potential
disruptions and abuses that are unique to electronic trading. The
potential system safeguards are broadly grouped into those that address
(1) Controls related to order placement; (2) policies and procedures
for the design, testing and supervision of ATSs; (3) self-
certifications and notifications; (4) ATS or algorithm identification;
and (5) data reasonability checks. Each system safeguard is summarized
below.
---------------------------------------------------------------------------
\89\ The Commission notes that the system safeguards
contemplated herein for DCMs address trading-related risks, and are
therefore distinct from the requirements of DCM Core Principle 20
and SEF Core Principle 14, which address business continuity and
disaster recovery capabilities.
---------------------------------------------------------------------------
1. Controls Related to Order Placement
a. Order Cancellation Capabilities
The Commission is inquiring about various standards related to
order cancellation capabilities. Auto-cancel on disconnect requirements
would ensure that working orders do not remain in the limit order book
when a firm loses connectivity with the exchange, ensuring that
unwanted trades avoid execution even if the firm is unable to cancel
them. The speed of disconnect notification and the cancellation of
orders on disconnect can be helped by the exchange of ``heartbeat''
messages between exchange and user which continuously monitor the
response ability of a given algorithm. In addition, by requiring
exchanges to develop and maintain the capacity to selectively cancel
working orders at the level of individual algorithms, individual
accounts, or individual firms, as deemed necessary in an emergency, the
trading platform would be able to mitigate the risk or quantity of
error trades due to a malfunction.\90\
---------------------------------------------------------------------------
\90\ In addition to order cancellation capabilities, the
Commission is inquiring about various related measures that concern
connectivity testing, including that trading platforms and all
entities connected to a trading platform for purposes of
transmitting orders together must test that the systems of all such
entities are properly connected to and communicating with the
trading platform, and that trading platforms must provide, and
market participants operating ATSs must utilize, heartbeats that
indicate proper connectivity between the trading platform and an
ATS.
---------------------------------------------------------------------------
The Commission is also inquiring as to the advisability of
requiring market participants operating ATSs, clearing members, and
exchanges to develop and maintain ``kill switch'' capabilities. A
market participant's kill switch could immediately cancel all working
orders from that firm to the exchange and could prevent them from
submitting further orders until natural persons with the proper
authority at both the firm and the exchange allow the firm to resume
trading. A kill switch at clearing members could cancel all working
orders attributable to the clearing member, including both proprietary
orders and orders placed on behalf of their clients, and prevent the
clearing member from transmitting additional orders until natural
persons at both the clearing firm and the exchange allow the clearing
member to resume trading. An exchange's kill switch could cancel all
working orders from an individual market participant or clearing firm
and could prevent additional orders from the same market participant or
clearing firm from being accepted at the exchange until authorized
natural persons at both the exchange and affected market participant or
clearing firm allow trading to resume.
48. The Commission's discussion of kill switches assumes that
certain benefits accrue to their use across exchanges, trading and
clearing firms, and DCOs. Please comment on whether such redundant use
of kill switches is necessary for effective risk control.
49. What processes, policies, and procedures should exchanges use
to govern their use of kill switches? Are there any different or
additional processes, policies and procedures that should govern the
use of kill switches that would specifically apply in the case of DMA?
50. What processes, policies, and procedures should clearing firms
use to govern their use of kill switches when using such a safeguard to
cancel and prevent orders on behalf of one or more clients?
51. What objective criteria regarding kill switch triggers, if any,
should entities incorporate into their policies and procedures?
52. What benefits or problems could result from standardizing
processes, policies, and procedures related to kill switches across
exchanges and/or clearing firms?
53. Please explain how kill switches should be designed to prevent
them from canceling or preventing the submission of orders that are
actually risk reducing or that offset positions that have been entered
by a malfunctioning ATS.
54. The Commission requests comment regarding whether kill switches
used by clearing firms already have or should have the following
capabilities: (a) Distinguish client orders from proprietary orders;
(b) distinguish among orders from individual clients; and (c) cancel
working orders and prevent additional orders from one or more of the
clearing firm's clients, or for all the clearing firm's proprietary
accounts, without cancelling and preventing all orders from the
clearing firm.
55. The Commission is aware of proposals that would enable FCMs to
establish credit limits for customers that are stored at a central
``credit hub'' for the purpose of pre-trade credit checks.\91\ If such
a model were implemented, is it possible that it could also be enabled
with kill switches that cancel existing working orders and prevent
additional orders from being submitted by one or more market
participants? Should such an approach be designed to complement kill
switches that are controlled by exchanges, clearing members, and
trading firms, or to replace these kill switches? What benefits and
drawbacks would result from each approach?
---------------------------------------------------------------------------
\91\ See MarkitServ Presentation, supra note 85.
---------------------------------------------------------------------------
b. Repeated Automated Execution Throttle
A further potential risk control of interest to the Commission is a
``Repeated Automated Execution Throttle.'' This risk control was
highlighted in FIA's Principal Traders Group recommendations regarding
risk controls.\92\ For this control, ATSs would be required to monitor
the number of times a strategy is filled and then re-enters the market
without human intervention. After a configurable number of repeated
executions the system should be disabled until a human re-enables it.
The Commission would like to better understand the value of this
safeguard. The Commission understands that it would disable automated
systems which have experienced activity levels far beyond that
anticipated by its designers, and then notify monitors regarding this
activity. Through this, human review would independently verify the
operation of an ATS at regular intervals, and in doing so, could help
to ensure that an algorithm's strategy is currently acting as
anticipated and that it is appropriately responding to current market
conditions. The Commission requests comments as to whether there could
be adverse effects of automatically disabling an ATS after a given
number of order executions, and also requests comment regarding the
potential value, proper use, and limitations of this safeguard.
---------------------------------------------------------------------------
\92\ See FIA Recommendations for Risk Controls, supra note 62,
at 4.
---------------------------------------------------------------------------
2. Policies and Procedures for the Design, Testing and Supervision of
ATSs; Exchange Considerations
Taken as a whole, the ATS monitoring and supervision standards, ATS
design and testing standards, ATS crisis management procedures
standards, and ATS monitoring staff training standards inquired about
in this Concept Release constitute a set of standards related to
[[Page 56558]]
policies and procedures for firms operating ATSs. Existing rules
require SDs and MSPs to ensure that their ``use of trading programs is
subject to policies and procedures governing the use, supervision,
maintenance, testing, and inspection of the program,'' \93\ but there
is no corresponding rule for FCMs or other market participants
operating ATSs. Moreover, even when applied to SDs and MSPs, section
23.600(d)(9) does not have any prescriptive requirements related to
supervision and testing and does not require formal review or approval
of each firm's policies and procedures by an informed, independent
party other than at the time of registration.\94\ As a consequence,
there is no minimum amount of testing that SDs and MSPs or other market
participants operating ATSs are required by the Commission to perform
before deploying an algorithm or before re-deploying an algorithm that
has been altered. Nor are there any minimum standards for training or
sophistication in the areas of supervision, maintenance, and inspection
of the ATS.\95\ Because of this, the Commission is interested in better
understanding whether more standardized requirements, or clearer
minimum standards, related to policies and procedures for firms
operating ATSs would benefit the markets and the public. The policies
and procedures relating to the design, testing and supervision of ATSs
are summarized below, and addressed in greater detail in Section V,
Appendix C.
---------------------------------------------------------------------------
\93\ See 17 CFR 23.600(d)(9).
\94\ 17 CFR 23.600(b)(4) requires SDs and MSPs to ``furnish a
copy of its written risk management policies and procedures to the
Commission, or to a futures association registered under section 17
of the Act, if directed by the Commission, upon application for
registration and thereafter upon request.''
\95\ It is also possible that SDs and MSPs could fail to
incorporate emerging industry best practices for managing
operational risk of ATSs into their policies and procedures as
effective risk management technology and practices are introduced to
the market.
---------------------------------------------------------------------------
a. ATS Development, Change Management, and Testing; Development, Change
Management, and Testing of Exchange Systems
The Commission requests public comment regarding the necessity for
ATS development, change management and testing standards in CFTC-
regulated markets. Potential benefits to such standards include
ensuring that ATSs are designed and modified in an environment where
there is no risk that the ATS could interfere with activity in or
related to the live market and ensuring that appropriate personnel have
approved changes and verified proper testing before a system is moved
to the production environment. Standards concerning the retention and
control of access to current and historical versions of source code may
help to ensure that changes are only made by appropriate personnel and
reviewable when necessary. Finally, audit trail material may assist
regulators when investigating problems.
With respect to testing, a firm's ATS testing standards could
require it to test an ATS on the trading platform(s) where it will
trade, prior to deploying such ATS into the live environment. Such
testing standards may reduce the incidence of technical errors at the
level of individual algorithms and firms. In addition, a firm's ATS
testing standards may require it to test an ATS on the trading
platform(s) after modifying the underlying algorithms or other system
components to a degree subject to further definition. ATS testing could
include tests against historical data, especially periods for which the
relevant algorithm would likely have been stressed, or would have been
active during periods with unanticipated market activity. In addition,
exchanges could also be required to provide a test environment to
simulate production trading so that market participants can conduct
exchange-based conformance testing, which would include tests of
compatibility with the matching engine (including initiation and
cessation of the ATS connection) and verification of risk controls
required by the trading platform.
The Commission is particularly interested to understand when it is
most beneficial for firms to test an ATS after it has been modified.
Some have asserted that the amount of testing should be calibrated to
the significance of the change and the risk it poses to the proper
function of the ATS.\96\ The Commission would like to better understand
how market participants estimate the significance of a change and the
risk that a given change might pose to the proper function of an ATS.
Also, the Commission would like to understand what current best
practices are for testing ATSs and how those practices are tailored to
the extent of the modification.
---------------------------------------------------------------------------
\96\ See SEC Roundtable Transcript, supra note 2, at 49-51.
---------------------------------------------------------------------------
56. Please describe the necessary elements of an effective ATS
testing regime, in connection with both the initial deployment and the
modification of an ATS.
57. With respect to testing of modifications, how should the
Commission and market participants distinguish between major
modifications and minor modifications? What are the objective criteria
that can be used to make such distinctions? Should any testing regime
applicable to ATS modifications distinguish between major and minor
modifications, and if so, how?
58. What challenges or benefits may result from exchanges
implementing standardized procedures regarding the development, change
management, and testing of exchange systems? Please describe, if any,
the types of standardized procedures that would be most effective.
b. ATS Monitoring and Supervision
The Commission is aware that many exchanges and software design
firms offer extensive testing platforms to validate algorithm
functionality before deployment in a live trading environment. The
Commission wants to better understand the extent to which testing is
utilized and would like to better understand the methodology supporting
these test environments. Further, the Commission believes that many, if
not all, firms operating ATSs have human monitors supervising ATSs when
they are operating. However, the Commission is uncertain to what degree
such monitors have been sufficiently trained in how to respond to
unexpected problems, and been given the requisite authority to
intervene at these times.\97\ A firm's ATS training standards could
require that relevant staff members be able to understand how to
identify malfunctions, evaluate the risk resulting from those
malfunctions, and respond constructively to those malfunctions,
including elevating the problem to the attention of more senior
personnel. The Commission would like to better understand whether
regulatory measures or new standards in this area would promote more
effective ATS monitoring and supervision.
---------------------------------------------------------------------------
\97\ The Commission would like to better understand what sorts
of training and policies market participants use in order to ensure
that human monitors have the capability to respond to operational
issues in a timely way. In particular, the Commission is interested
in better understanding what training monitors receive in the
rationale for the trading patterns executed by the ATS, the scope of
intervention authority given to human monitors, and the procedures
firms use to escalate questions or decisions from such human
monitors to more senior personnel during a crisis.
---------------------------------------------------------------------------
c. Crisis Management Procedures
Well-designed crisis management procedures may help to ensure that
[[Page 56559]]
firms are prepared to conduct rapid triage in the event of a problem,
including the ability to escalate decisions quickly to the proper
individuals or provide notification to their clearing firms, exchanges,
or the Commission.\98\ Such procedures may promote common expectations
among monitoring staff, firm leadership, and exchange leadership about
basic procedures in the event of market destabilizing events,
facilitating more rapid intervention and mitigating the effects of an
individual disruption.
---------------------------------------------------------------------------
\98\ See SEC Roundtable Transcript, supra note 2, at 133-34.
---------------------------------------------------------------------------
59. Should basic crisis management procedures be standardized
across market participants? If so, what elements should be addressed in
an industry-wide standard?
60. Are there specific, core requirements that should be included
in any crisis management procedures? Similarly, are there specific
types of crisis events that should be addressed in any crisis
management procedures? If so, please identify such requirements and/or
crisis events and the level of granularity or specificity that the
procedures should have with respect to each.
3. Self-Certifications and Notifications
a. Self-Certification and Clearing Firm Certification
To ensure that market participants employ the pre-trade risk
controls, post-trade reports and other measures, and system safeguards
described herein, the Commission is inquiring whether it would be
appropriate to require a periodic self-certification program for all
market participants operating ATSs and for clearing firms providing
services to those market participants. These certifications could refer
to the extent of implementation of those risk control mechanisms
discussed in the other sections of this Concept Release. With respect
to ATSs, an acceptable certification might attest that: (1) The ATS
contains structural safeguards to provide reasonable assurance that the
trading system will not be disruptive to fair and equitable trading;
(2) the market participant's ATSs have been designed to avoid
violations of the CEA, Commission regulations, or exchange rules
related to fraud, disruptive trading practices, manipulation and trade
practice violations; and (3) such systems have been sufficiently tested
and documented in a manner that is appropriate to the intended design
and use of that system. Additionally, the Commission asks whether the
chief executive officer, chief compliance officer, or similar ranking
official of each market participant should attest to the certification.
The Commission is interested in receiving comment on the costs and
benefits of a certification program, what elements should be included
in the program, and whether that program should be self-executed, or,
if not, overseen by what authority.
61. How often should a market participant certify that their pre-
trade risk controls, post-trade reports and other measures, and system
safeguards meet the necessary standards?
62. Which representative of the market participant should be
required to attest that the certification standards have been met?
Should it be the market participant's chief executive officer, chief
compliance officer, or similar high-ranking corporate official, or some
other individual?
63. Which entity(ies) should receive certifications from market
participants? For example, should it be the market participant's
clearing firm, its designated self-regulatory organization (if
applicable), one or more trading platforms, a registered futures
association, the Commission, or other entity?
64. Should DCMs, SEFs or clearing member firms be required to audit
market participant certifications? What would be covered in an audit
and how often should these audits occur? Should the same entity that
receives the certification be required to perform the audit?
b. Risk Event Notification Requirements
The Commission also seeks information as to whether it would be
beneficial for market participants operating ATSs to notify one or more
of trading platforms, their clearing firms, the Commission, or others
of risk events.\99\ Entities receiving notifications could, when they
deem it appropriate based on the magnitude of a single event or a
pattern of smaller related events, review further with the market
participant to remedy the underlying cause(s) of the risk event. Such
reviews would allow market participants, clearing firms, trading
platforms, and the Commission to respond and proactively reduce risk in
automated trading environments.
---------------------------------------------------------------------------
\99\ The SEC is presently considering a set of rules that would
require self-regulatory organizations, significant alternative
trading systems, certain disseminators of market data, and exempt
clearing agencies to notify SEC staff of events including systems
disruptions, compliance issues, or intrusions. See SEC, Notice of
Proposed Rulemaking: Regulation Systems Compliance and Integrity, 78
FR 18084 (Mar. 25, 2013). Under the proposed rules, these entities
would be required to notify and provide the SEC with detailed
information when such systems issues occur as well as when there are
material changes in its systems. Id. The Commission notes that it
may consider distinctive aspects of the SEC's proposed rules, and
public comments with respect to it, when developing any future
proposals arising from this Concept Release. Commenters with respect
to this Concept Release are encouraged to indicate in their comments
any elements of the SEC's proposed rules that they believe are
relevant.
---------------------------------------------------------------------------
The Commission seeks comment on the types of risk events that
should be reported. For example, reportable risk events generally could
include any instances where design parameters of an ATS are violated
and where risk control processes or technologies do not function as
anticipated, regardless of whether these events lead to error trades or
market destabilization. Violated design parameters and unanticipated
lapse of risk management processes and technology create conditions
that may presage future malfunctions, even absent a current disruption.
65. Do commenters believe that risk event notifications would help
to better understand and ultimately reduce sources of risk in automated
trading environments? What information should be contained in a risk
event notification to maximize its value?
66. What types of risk events should trigger reporting
requirements, and what entities should receive risk event notifications
from market participants operating ATSs?
67. Which entities should receive risk event notifications?
4. ATS or Algorithm Identification
The Commission is considering measures to improve the
identification of ATS or their underlying algorithms in messages
generated by ATSs. The Commission believes that identification of ATSs
or underlying algorithms could help both firms and trading platforms to
more quickly identify malfunctioning systems that could disrupt
markets. Fuller identification of automated systems may also improve
oversight by the Commission, including the ex post analysis of
disruptive events aimed at preventing or mitigating similar
recurrences.
The Commission is aware of the inherent complexity in any ATS or
algorithm identification system and seeks public comment on this
potential measure. Specific questions of interest to the Commission
include:
68. Should the Commission define ATS or algorithm for purposes of
any ATS identification system that may arise from this Concept Release?
If so, how should ATS or algorithm be defined? Should a separate
designation be reserved for high frequency trading algorithms and if
so, what is the threshold difference?
[[Page 56560]]
69. What are the existing practices within trading firms for
internally identifying ATSs or algorithms and for tracking their
performance, including profit and loss? What elements of existing
practices could be leveraged in any ATS or algorithm identification
system proposed by the Commission in the future?
70. The Commission understands that an ATS may consist of numerous
algorithms, each of which contributes to a trading decision. If an
algorithm-based identification system is proposed, which of the
potentially multiple algorithms that constitute an ATS should carry the
ID? In addition, what degree of change to an algorithm should
necessitate the use of a new ID, and how often does this change
typically occur? What is the appropriate definition of ``algorithm''
for purposes of an algorithm identification system?
71. If the identification system resides at the ATS level, how
should such IDs be structured to ensure that they are nonetheless
sufficiently granular to identify components that may be leading or
have led to unstable market conditions?
72. What message traffic between an ATS and a trading platform
should include the ATS or algorithm ID (all messages, orders only,
etc.)?
73. What relationship should this ATS ID have to the legal entity
identifier (LEI)?
5. Data Reasonability Checks
The Commission is interested in the range of information sources
used by ATSs to inform their trading decisions, and in how market
participants form reasonable beliefs as to the accuracy of such data.
For example, following recent media reports regarding the adverse
market impact of false information distributed through unauthorized use
of a social media outlet used by the Associated Press, the Commission
is asking questions to broaden its understanding of the extent to which
ATSs in derivatives markets use social media to inform their trading
decisions, and the extent to which information derived from social
media is verified by the ATS prior to its use. One potential risk
control of interest to the Commission is the ``market data
reasonability check,'' which was included in FIA's Principal Traders
Group recommendations regarding risk controls.\100\ In those
recommendations, the FIA recommended that trading firms' systems have
``reasonability checks'' on incoming market data.
---------------------------------------------------------------------------
\100\ See FIA Recommendations for Risk Controls, supra note 62,
at 4.
---------------------------------------------------------------------------
74. Please describe existing practices in the industry concerning
how and the extent to which ATSs use (1) market data; and (2) news and
information providers, including social media, to inform trading
decisions.
75. The Commission requests comment regarding any risk controls,
including reasonability checks, currently being used by market
participants operating ATSs to review market data and news and
information providers, including social media. Please describe the risk
control, including the purpose of the control, the extent of its use
among derivatives market participants, and any other aspects of the
risk control that you believe would be helpful for the Commission to
understand.
In addition, the data analyzed by trading algorithms can include
government economic reports (e.g., GDP, unemployment, and inflation
data), as well as economic reports from non-governmental organizations
such as universities, trade groups, and other sources. While government
reports are released pursuant to a lock-up process that is intended to
ensure that no entity receives them ahead of others, it has been
reported that early access to some non-government economic reports is
available for a fee. For example, according to recent reports, the
University of Michigan's consumer report was available to certain
investors two seconds ahead of the rest of the market.\101\
---------------------------------------------------------------------------
\101\ See Brody Mullins, Michael Rothfeld, Tom McGinty & Jenny
Strasburg, ``Traders Pay for Early Peek at Key Data,'' Wall St. J.
(June 12, 2013), available at http://online.wsj.com/article/SB10001424127887324682204578515963191421602.html.
---------------------------------------------------------------------------
76. The Commission requests public comment concerning the lock-up
process for government economic reports, and any additional measures
that might be taken to protect against inappropriate disclosure.
77. Please describe the extent to which potentially market-moving
data from non-governmental economic reports can be obtained prior to
its public release for a fee. Are there specific reports or types of
reports for which early disclosure should not be permitted? What
process should be used for identifying non-governmental economic
reports whose early release should not be permitted? Should the data
release process for such reports be similar to the data lock-up process
implemented for the release of government economic data?
The system safeguards described above are also listed in Appendix
C.
F. Other Protections
1. Registration of Firms Operating ATSs
Although the Commission can currently take several actions to seek
information from firms, such as the issuance of subpoenas to
investigate a firm's trading activities on a registered exchange or to
compel a firm to provide books and records, some have suggested that a
registration requirement for firms operating ATSs and not otherwise
registered with the Commission would enhance the Commission's oversight
capabilities. Additionally, a registration requirement may allow for
wider implementation of some or all of the pre-trade controls and risk
management tools discussed in this Concept Release and currently
deployed in various degrees in the market today.
In considering the registration of specific entities using ATSs and
not otherwise registered with the Commission, the ``floor broker''
definition in CEA 1a(23), in pertinent part, states that, in general,
the term ``floor trader'' means any person who, in or surrounding any
pit, ring, post or other place provided by a contract market for the
meeting of person similarly engaged, purchases, or sells solely for
such person's own account.\102\
---------------------------------------------------------------------------
\102\ See CEA section 1a(23), as amended by section 721 of the
Dodd-Frank Act; 7 U.S.C. 1a(23) (emphasis added).
---------------------------------------------------------------------------
In addition to seeking input on whether it would be beneficial to
require registration, the Commission also requests specific public
comments in response to the following questions:\103\
---------------------------------------------------------------------------
\103\ In March 2013, the German parliament approved the HFT Act,
which requires any firm using HFT strategies to become licensed as a
financial services institution subject to the supervision of BaFin
(Germany's banking regulator) or to passport an existing license
granted by another member state of the European Economic Area. The
licensing requirement includes ``indirect'' trading, meaning that it
applies to foreign firms that are trading through a direct exchange
member on a German-regulated market or a German multilateral trading
facility. As a result of becoming licensed, HFT firms become subject
to a general regulatory framework applicable to investment firms
under German statutes, and specific organizational requirements
applicable to HFT firms imposed by the HFT Act. See BaFin HFT Act
Materials, supra note 17.
---------------------------------------------------------------------------
78. Should firms operating ATSs in CFTC-regulated markets, but not
otherwise registered with the Commission, be required to register with
the CFTC? If so, please explain.
79. Please identify the firm characteristics, trading practices, or
technologies that could be used to trigger a registration requirement.
80. Should all firms deploying ATS be required to register, and
should there be different standards for firms deploying
[[Page 56561]]
HFT strategies? What are the appropriate thresholds levels below which
registration would not be required?
81. Since the floor trader distinction only addresses proprietary
traders, please explain whether there is any other category of market
participant, such as those deploying ATS or HFT strategies and trading
on behalf of clients (aside from market participants already subject to
Commission jurisdiction, such as Introducing Brokers and FCMs) that the
Commission should consider with respect to potential registration
requirements.
82. Should software firms providing algorithms be required to
register, and under what authority? What standards should apply to such
firms?
83. Please identify the functionalities discussed in this Concept
Release that could be applied to floor brokers that operate ATSs. Are
there any other controls not mentioned in this Concept Release that
should be under consideration?
84. Please supply any information or data that would help the
Commission in deciding whether firms may or may not meet the definition
of ``floor trader'' in Sec. 1a(23) of the Act.
85. Do you believe that the registration of such firms as ``floor
traders'' would effectuate the purposes of the CEA to deter and detect
price manipulation or any other disruptions to market integrity?
86. Considering the broad deployment of automated trading systems
across both equities and derivatives markets, the Commission seeks to
understand the appropriate level of coordination between itself and the
SEC in defining and applying possible standards to the ATS and HFT
trading space. How closely should the CFTC and SEC coordinate on
possible rules and requirements for trading firms? The Commission also
seeks public comment on the appropriate level of coordinated oversight
between itself and relevant Self-Regulatory Organizations such as
National Futures Association and FINRA.
87. Using the Flash Crash as an example, is it important to have
identical definitions and remedies in the case of ATS and HFT
registration requirements or do the existing market controls, such as
circuit breakers, provide the necessary market protections in both the
equities and derivatives markets? If the rules are not coordinated,
what impact would this have on market interaction and oversight?
88. If trading venues apply mandatory functionalities to access
derivatives markets, what benefit would a registration requirement
provide to the Commission?
2. Market Quality Data
The Commission is inquiring as to the advisability of requiring
each trading platform to provide market quality indicators for each
product traded on its platform at a regular frequency. Some metrics of
the type below are currently calculated by exchanges, often at an
account level, and provided to market participants. Some metrics are
currently used in aid of various exchange programs (such as order
efficiency programs). Other metrics are not currently used but may,
nonetheless, provide the Commission and the public potentially useful
information.
The Commission envisions that increased transparency through the
regular disclosure of market quality indicators will allow the
Commission and market participants to better understand, among other
things (1) The stability and efficiency of each market, (2) the degree
of informed versus uninformed order flow, and (3) the nature and degree
of liquidity in each market. In addition, the transparency provided by
these metrics may better enable market participants to manage their
ATSs in ways that further promote market stability and integrity.
The Commission is interested in receiving comment on the usefulness
of various market indicators that could be prepared for each contract.
The list of indicators would, for a given product and tenor, include
measures of: (1) Effective spreads; (2) order-to-fill ratios; (3)
execution speeds by order type and order size; (4) average
aggressiveness imbalances; (5) price impact for given trade sizes;
\104\ (6) average order duration; \105\ (7) order efficiency; \106\ (8)
rejection order ratios; (9) net position changes versus volume; \107\
(10) branching ratios; \108\ (11) volume imbalance and trade intensity;
\109\ (12) Herfindahl-Hirschman Indexes based on market share of open
positions under common control; and (13) metrics on the number of price
changing trades involving ATSs.\110\ Calculation methodologies for each
of the measures would be consistent across exchanges in order to ensure
compatibility and comparability across market venues.\111\
---------------------------------------------------------------------------
\104\ The size of the price change that would occur if specific
sizes of market orders were executed at that instant.
\105\ Average length of time that orders for a specific
instrument remain in the book before being modified, filled, or
cancelled.
\106\ Notional value executed vs. notional value entered or
modified.
\107\ See CFTC Net Position Changes Data, available at http://www.cftc.gov/MarketReports/NetPositionChangesData/index.htm.
\108\ See Vladimir Filimonov, David Bicchetti, Nicolas Maystre,
& Didier Sornette, ``Quantification of the High Level of Endogeneity
and of Structural Regime Shifts in Commodity Markets'' (Mar. 20,
2013), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2237392.
\109\ See David Easley, Marcos M. Lopez de Prado & Maureen
O'Hara, ``Flow Toxicity and Liquidity in a High Frequency World''
(Feb. 20, 2012), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1695596.
\110\ For a given market, such metrics would be calculated by
identifying the relevant category of trader on trades that result in
a price move from a previous trade and determining the percentage of
those trades where an ATS was on one or both sides of the trade.
\111\ SEC Rules 605 (Disclosure of Order Execution Information)
and 606 (Disclosure of Order Routing Information) of Regulation NMS
respectively require market centers (as defined in the rules) to
make publicly available standardized, monthly reports of statistical
information concerning their order executions and broker-dealers to
make publicly available quarterly reports that, among other things,
identify the venues to which customer orders are routed for
execution. See 17 CFR 242.605 (formerly Securities Exchange Act Rule
11Ac1-5) and 17 CFR 242.606 (formerly Securities Exchange Act Rule
11Ac1-6).
---------------------------------------------------------------------------
Several of the measures described in this Concept Release would
provide additional information about market quality that market
participants cannot derive exclusively from real-time order book
information provided by each exchange. The Commission expects that
market participants could use this additional information, together
with information currently available in the order book, in order to
better inform their trading efficiency and strategies and to mitigate
adverse effects of their actions and other market participants' on the
market. Further, the Commission expects that these measures could be
used to help understand changes in market quality. In addition, the
Commission believes that providing consistent measures of market
quality across exchanges would promote market efficiency through
transparency and market competition.
To clarify what costs and benefits these market metrics may provide
to participants, the Commission requests comment to the questions
below, including that, if these metrics are beneficial, the appropriate
frequency of publication.
89. What market quality indicators are in place today? Please
describe the metrics, how and where they are deployed, and how market
participants access these indicators and at what cost.
90. What value would each of the market quality metrics described
above provide to market participants receiving them? If possible,
please be specific about how each market quality measure could be used
to enhance reliability and risk management of ATSs.
[[Page 56562]]
91. Conversely, could any of the market quality metrics described
above be used by market participants to manipulate the order book,\112\
to identify competitors' trading strategies, or to engage in other
trading activities that do not contribute to effective risk management
and efficient discovery the traded asset's economic value? If so,
please provide specific information regarding how such information
could be misused. If possible, please provide recommendations regarding
steps the Commission could take to prevent misuse.
---------------------------------------------------------------------------
\112\ Meaning, behaviors that, while not strictly illegal, are
used to advantage one's own orders in ways that do not contribute to
efficient price discovery.
---------------------------------------------------------------------------
92. Are there additional market quality metrics that the Commission
should contemplate requiring exchanges to provide? If so, what value
would they provide and how would they be used?
93. If the Commission determines that measures should be calculated
in the same way by various exchanges in order to provide comparable
measures of market quality, then how, specifically, should each of the
above mentioned metrics be calculated in order to ensure that they are
most valuable to market participants?
94. What timing and mode of dissemination is appropriate for each
metric? For example, should measures be provided as daily averages?
95. Does the liquidity of a given market impact which market
quality metrics would be reliable and useful when calculated for that
market? If so, which metrics are inapplicable in less liquid markets,
and why? What liquidity measures and thresholds are relevant to
determining which metrics should apply to a given market?
3. Market Quality Incentives
The impact of ATSs, and particularly those implementing HFT
strategies, is a topic of ongoing interest among researchers, market
participants and others. Several studies have found that increases in
automated trading are associated with improved market quality.\113\
Some researchers and market participants, however, have also noted that
the presence of HFT has the potential to shape the types of liquidity
providers available in a market,\114\ may discourage ATSs from
submitting resting orders that remain in the order book long enough for
humans to react, and may also be associated with undesirable trading
practices that are more easily implemented by automated systems.\115\
Various recommendations have been advanced to promote the benefits of
HFT while simultaneously disincentivizing trading strategies that do
not contribute to efficient price discovery.\116\
---------------------------------------------------------------------------
\113\ See Jonathan Brogaard, Terrence Hendershott & Ryan
Riordan, ``High Frequency Trading and Price Discovery'' (Apr. 22,
2013), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1928510; Hasbrouck & Saar, supra note 18;
Terrence Hendershott, Charles Jones & Albert Menkveld, ``Does
Algorithmic Trading Improve Liquidity?'' Journal of Finance, Vol. 66
at 1-33 (August 30, 2010), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1100635.
\114\ See J. Doyne Farmer & Spyros Skouras, ``An Ecological
Perspective on the Future of Computer Trading,'' Quantitative
Finance (2013); IOSCO Report on Regulatory Issues Raised by
Technological Changes, supra note 4; William Barker & Anna
Pomeranets, ``The Growth of High-Frequency Trading: Implications for
Financial Stability,'' Bank of Canada Financial System Review (June
2011), available at http://www.bankofcanada.ca/2012/01/publications/periodicals/fsr-article/the-growth-of-high-frequency-trading/.
\115\ See Farmer & Skouras, supra note 114; Eric Budish, Peter
Cramton & John Shim, ``The High-Frequency Trading Arms Race:
Frequent Batch Auctions as a Market Design Response'' (July 7,
2013), available at http://faculty.chicagobooth.edu/eric.budish/research/HFT-FrequentBatchAuctions.pdf; John McPartland,
``Recommendations for Equitable Allocation of Trades in High
Frequency Trading Environments'' (July 25, 2013), available at
http://www.chicagofed.org/Webpages/publications/policy_discussion_papers/2013/pdp_1.cfm.
\116\ See McPartland, supra note 115.
---------------------------------------------------------------------------
Those recommendations include for example, utilizing a trade
allocation formula that is an intermediate between a cardinal ranking
(time-weighted), Pro Rata allocation formula and a Price/Time
allocation formula. This would be intended to reward market makers for
leaving resting orders in the order book for a longer period of time,
rather than simply for being first in the order book at a given price.
Second, create a new limit order type that would prioritize orders that
remain resting in the order book for some minimum amount of time.
Third, require orders that are not fully visible in the order book
(e.g., iceberg orders) to go to the end of the queue (within limit
price) with respect to trade allocation. Fourth, aggregate multiple,
small orders from the same legal entity entered contemporaneously at
the same price level and assign them the lowest priority time stamp of
all such. Fifth, require exchanges to use batch auctions once per half
second at random times rather than use continuous trade matching.\117\
Lastly, limit visibility into the order book to aggregate size
available at a limit price. This would help to ensure that automated
traders are placing orders based on their knowledge of the economic
value of the asset being traded rather than their knowledge of order
book dynamics or of other market participants' trading patterns.
---------------------------------------------------------------------------
\117\ See Budish, supra note 115; J. Doyne Farmer & Spyros
Skouras, ``Review of the Benefits of a Continuous Market vs.
Randomised Stop Auctions and of Alternative Priority Rules (Policy
Options 7 and 12),'' Foresight U.K. Government Office for Science,
Economic Impact Assessment (2013), available at http://www.bis.gov.uk/assets/foresight/docs/computer-trading/12-1072-eia11-continuous-market-vs-randomised-stop-auctions.pdf.
---------------------------------------------------------------------------
96. Should exchanges impose a minimum time period for which orders
must remain on the order book before they can be withdrawn? If so,
should this minimum resting time requirement apply to orders of all
sizes or be restricted to orders smaller than a specific threshold? If
there should be a specific threshold, how should that threshold be
determined?
97. The Commission seeks to understand where time-weighted Pro Rata
trade allocation is currently being utilized and what the effects have
been. Please note examples from exchanges and, to the extent possible,
please comment on the impact that such matching algorithms have had on
the amount of time resting orders are left in the order book, as well
as on other aspects of market quality.
98. If exchanges aggregated multiple, small orders entered by the
same entity with the intent of abusing rounding conventions to gain a
disproportionate share of allocations, what criteria should exchanges
use to distinguish such orders from those that are entered by the same
legal entity for legitimate trading purposes? Are there empirical
patterns that could be used to reliably identify such manipulative
intent?
99. Would batched order processing increase the number of
milliseconds that are necessary for correlations among related
securities to be established? If so, what specific costs would result
from this change and how do those costs compare to the potential
benefits described in recent research?
100. What costs and benefits result from providing market
participants with real-time access to information about the order book
that extends beyond aggregate size available at a limit price? Is there
a legitimate economic benefit that results from market participants
(both human participants, and ATSs) accessing such information? Is it
possible for market participants to use such information to manipulate
the order book?
101. The Commission seeks to understand whether any of the
recommendations above are inapplicable or irrelevant to markets subject
to the CEA. If so, please indicate which recommendation(s) and what
makes it inapplicable or irrelevant to those markets.
[[Page 56563]]
4. Policies and Procedures To Identify ``Related Contracts''
Rule 38.255 of the Commission's regulations require DCMs to
establish and maintain risk controls for trading.\118\ Appendix B to
the Part 38 regulations provides the following guidance on such risk
controls: If a contract is linked to, or is a substitute for, other
contracts, either listed on [the DCM's] market or on other trading
venues, the designated contract market must, to the extent practicable,
coordinate its risk controls with any similar controls placed on those
other contracts.\119\ The guidance contained in the appendix further
provides that, to the extent practicable, DCMs should coordinate not
only with other DCMs, but national security exchanges as well.\120\
These measures could protect against market disruptions cascading from
one trading platform to the next.
---------------------------------------------------------------------------
\118\ See 17 CFR 38.255.
\119\ See DCM Final Rules, 77 FR at 36718.
\120\ See id.
---------------------------------------------------------------------------
102. If you are a DCM, please address whether you have (i)
identified all contracts that are linked to, or are a substitute for,
other contracts either listed on your market or on other trading
venues; and, if so, (ii) coordinated your risk controls with any
similar controls placed on those other contracts. If you have not
identified such contracts and coordinated risk controls on such
contracts, please address any other means by which you are addressing
risk controls applicable to contracts that are linked to, or are a
substitute for, other contracts listed on your exchange or on other
trading venues.
103. Please explain whether it would be beneficial for exchanges to
develop and document policies and procedures for regularly reviewing
contracts on other exchanges in order to identify those that are
``linked to'' or that are ``a substitute for'' contracts listed on its
own market.
5. Standardize and Simplify Order Types
This Concept Release inquires about the possible standardization
and simplification of order types that have complex logic embedded
within them. A proliferation of order types, both within and across
exchanges, can result in a similar increase in both the expected and
unexpected responses of automated systems to order and trade signals.
As of November 2012, for example, it was reported that BATS Global
Markets alone listed more than 2,000 order types.\121\ A review of
current and proposed order types could be performed with the goal of
consolidating and simplifying order types.\122\ A proliferation of
complex order types leads to complex testing scenarios. Therefore, it
is possible that consolidation of order types could reduce the
potential for instability resulting from unexpected interactions of
multiple ATSs using multiple means of execution within the order
book.\123\
---------------------------------------------------------------------------
\121\ See Peter Chapman, ``Too Many Order Types, Traders Fret,''
Traders Magazine (Nov. 2012), available at http://www.tradersmagazine.com/issues/25_344/order-types-equities-structure-110515-1.html.
\122\ The SEC is currently in the process of reviewing order
types within securities markets. See Scott Patterson & Jean
Eaglesham, ``Exchanges Retreat on Trading Tools,'' Wall St. J. (Oct.
24, 2012) (quoting former Chairwoman of the SEC, Mary Schapiro: ``I
worry about the complexity in the market, I worry about the
profusion of order types, I worry about the fragmentation.''),
available at http://online.wsj.com/article/SB10001424052970203400604578074963881803302.html. See also SEC
Roundtable Transcript, supra note 2, at 96-99.
\123\ See SEC Roundtable Transcript, supra note 2, at 96 (``It
is the proliferation of all these order types and the complexity of
these order types that is adding unnecessary complexity to the
market, which is already an extremely complex system as it is . . .
when you have complex order types, it leads to extremely complex
testing scenarios, and you are not going to pick up all the things
you could or should because you don't know what that actual matching
engine logic is in general.'').
---------------------------------------------------------------------------
104. Please explain whether the standardization and simplification
of order types that have complex logic embedded within them would
reduce the potential for instability and other market disruptions. If
not, what other measures could achieve the same effect?
105. If the Commission were to consider the standardization and
simplification of order types in a future rulemaking, please identify
who should conduct this review (i.e., the Commission, trading
platforms, or other parties).
G. General Questions Regarding All Risk Controls Discussed Above
Finally, the Commission requests comment on the following general
questions, with respect to each of the risk controls discussed above:
106. For each of the specified controls described above [see
sections III.C-F], please indicate whether you are already using the
control on customer and/or proprietary orders. If applicable, please
also indicate how widely you believe the control is currently being
used in the market, and how consistent the application of the control
is among firms.
107. If possible, please indicate specific costs associated with
implementing each of the risk controls described above [see sections
III.C-F]. Please include detailed estimates, distinguishing between the
cost of developing the functionality, the cost of implementation, and
the cost of ongoing operations.
108. Please describe the specific benefits associated with each of
the risk controls. Where possible, please indicate the market
participant category(ies) to which the benefit would accrue.
109. Please comment on the appropriate order of implementation and
timeline for each risk control, including any distinctions that should
be made based on the category of registrant or market participant
implementing the same or similar control, whether the market
participant is using DMA, and whether implementation is already in
place for certain categories.
110. Are any of the risk controls unnecessary, impractical for
commercial or technological reasons, or inadvisable? If so, please note
the control and provide reasons why.
111. A number of the pre-trade risk controls contemplated above are
similar protections at distinct points in the life of an order.
a. Please comment on the utility of redundant pre-trade risk
controls and the desirability of risk control systems in which controls
are placed at one or more than one focal points.
b. If pre-trade risk controls should reside at one or more than one
focal point, then please identify, for each risk control, what that
focal point should be?
112. Are there risk controls that should be implemented across
multiple entity types? If so, which controls and for which types of
entities should they apply? Also, please comment generally on the
factors the Commission should consider when determining the appropriate
entity(ies) upon which to place a risk control requirement that could
pertain to more than one entity.
113. Are there controls that should not be considered for
overlapping implementation across exchanges, clearing members and
market participants? If so, please explain which ones and why.
114. Each of the risk controls is described in general, principles-
based terms. Should the Commission specify more granular or specific
requirements with respect to any of the controls to improve their
effectiveness or provide greater clarity to industry participants? If
so, please identify the relevant control and the additional granularity
or specificity that the Commission should provide. Are any of the
controls, as
[[Page 56564]]
currently drafted, inadequate to achieve the desired risk-reduction?
115. To the extent that there is any need to standardize or provide
greater specificity regarding any measures discussed in this Concept
Release, including those that reflect industry best practices, please
describe the best approach to achieve such standardization (i.e.,
through Commission regulation, Commission-sponsored committee or
working group, or some other method).
116. How should risk control monitoring be implemented? Should
compliance be audited by internal and external parties? For each
control, please identify the appropriate entity(ies) to monitor
compliance with the control. Also, please describe what an acceptable
compliance audit would entail for each control.
117. Are there additional controls that should be considered, or
other methods that could serve as alternatives to those described above
[see sections III.C-F]? If so, please describe the control, its costs
and benefits, the appropriate entity(ies) to implement such control,
and whether there is any distinction to be drawn in the case of DMA.
118. Would any of the risk safeguards create a disincentive to
innovate or create incentives to innovate in an irresponsible manner?
If so, please identify the control, the concern raised, and how the
control should be amended to address the concern. Responses should
indicate how an amended risk control would still meet the Commission's
objectives.
119. Should the Commission consider any pre-trade risk controls,
post-trade reports, or system safeguards appropriate exclusively to
market makers or to ATSs used by market makers? If so, please describe
such controls or safeguards.
120. Should the Commission or Congress revisit its approach to
issuing civil monetary penalties for violations of the Act,
particularly as they relate to automated trading environments?
Currently, the maximum civil monetary penalty the Commission may issue
is capped at $140,000 ``per violation.'' Is such a civil monetary
penalty sufficient to deter acts that constitute violations of the Act,
given that an individual violation could impose costs to the market and
the public well in excess of $140,000?
121. Please describe the documentation (or categories of documents)
that would demonstrate that a market participant operating an ATS has
implemented each risk control addressed in this Concept Release,
including, for example, computer code, system testing results,
certification processes and results, and calculations.
122. Would a fee (collected by, for example, the DCM or SEF) on
numbers of messages exceeding a certain limit be more appropriate than
a hard limit on the number or rate of messages?
123. Should such a penalty be based on a specified number or rate
of messages or on the ratio of messages to orders filled over a
specified time period?
124. Recent disruptive events in securities markets illustrate the
importance of effective communication between exchanges' information
technology systems. The Commission requests public comments regarding
relevant systems in its regulated markets, including both DCMs and
SEFs. What data transfers or other communications between exchanges are
necessary for safe, orderly, and well-functioning derivatives markets?
What additional measures, if any, would help promote the soundness of
such systems (e.g., testing requirements, redundancy standards, etc.)?
IV. List of All Questions in the Concept Release
Listed below are all questions raised in the preceding sections of
this Concept Release.
High Frequency Trading
1. In any rulemaking arising from this Concept Release, should the
Commission adopt a formal definition of HFT? If so, what should that
definition be, and how should it be applied for regulatory purposes?
2. What are the strengths and weaknesses of the TAC working group
definition of HFT provided above [see section II.A.1]? How should that
definition be amended, if at all?
3. The definition of HFT provided above uses ``recurring high
message rates (orders, quotes or cancellations)'' as one of the
identifying characteristics of HFT, and lists three objective measures
((i) cancel-to-fill ratios; (ii) participant-to-market message ratios;
or (iii) participant-to-market trade volume ratios) that could be used
to measure message rates. Are these criteria sufficient to reliably
distinguish between ATSs in general and ATSs using HFT strategies? What
threshold values are appropriate for each of these measures in order to
identify ``high message rates?'' Should these threshold values vary
across exchanges and assets? If so, how?
4. Should the risk controls for systems and firms that engage in
HFT be different from those that apply to ATSs in general systems? If
so, how?
Reductions in Latency
5. Discussions on latency often focus on the how quickly an
exchange processes orders, the time taken to submit orders, and how
quickly a firm can observe prices of trades transacted on the exchange.
The Commission is interested in understanding whether there are other
types of messages transmitted between exchanges, firms and vendors
wherein differences in latency could provide opportunities for
informational advantage. Recent press reports have highlighted such
advantages in the transmission of trade confirmations by a specific
exchange. Are there other exchanges and trading venues where similar
differences in latency exist? The Commission is interested in
understanding whether the extent of latency in any such message
transmission process can have an adverse impact on market quality or
fairness. Should any exchanges, vendors and firms be required to audit
their systems and process on a periodic process to identify and then
resolve such latency?
Financial Integrity of the DCO
6. Are there distinct pre-trade risk controls, including measures
not listed below, or measures in addition to those already adopted by
the Commission, that would be particularly helpful in protecting the
financial integrity of a DCO?
Risk Controls Applicable in the Case of DMA
7. Are there distinct pre-trade risk controls, including measures
not listed below [see section III.C.], or measures in addition to those
already adopted by the Commission, that should apply specifically in
the case of DMA?
Message and Execution Throttles
8. If, as contemplated above [see section III.C.1], maximum message
rates and execution throttles were used as a mechanism to prevent
individual entities or accounts from trading at speeds that are
misaligned with their risk management capabilities, how should this
message rate be determined?
9. Message and execution throttles may be applied by trading firms
(FCMs and proprietary trading firms), clearing firms, and by exchanges.
The Commission requests public comment regarding the appropriate
location for message and execution throttles.
a. If throttles should be implemented at the trading firm level,
should they be applied to all ATSs, only ATSs employing HFT strategies,
or both?
b. What role should clearing firms play in the operation or
calibration of
[[Page 56565]]
throttles on orders submitted by the trading firms whose trades they
guarantee?
10. Should the message and execution throttles be based on market
conditions, risk parameters, type of entity, or other factors?
11. What thresholds should be used for each type of market
participant in order to determine when a message or execution throttle
should be used? Should these thresholds be set by the exchange or the
market participant?
12. Are message and execution thresholds typically set by contract,
or by algorithm? What are the advantages and disadvantages to each
method?
13. Who should be charged with setting message rates for products
and when they are activated?
14. Would message and execution throttles provide additional
protection in mitigating credit risk to DCOs?
Volatility Awareness Alerts
15. The Commission is aware that alarms can be disruptive or
counterproductive if ``false alarms'' outnumber accurate ones. How can
volatility alarms be calibrated in order to minimize the risk that
false alarms could interrupt trading or cause human monitors to ignore
them over time?
Self-Trade Controls
16. What specific practices or tools have been effective in
blocking self-trades, and what are the costs associated with wide-
spread adoption of such practices or tools?
17. Please indicate how widely you believe exchange-sponsored self-
trading controls are being used in the market.
18. Should self-trade controls cancel the resting order(s)? Or,
instead, should they reject the taking order that would have resulted
in a self-trade? If applicable, please explain why one mechanism is
more effective than the other.
19. Should exchanges be required to implement self-trading controls
in their matching engines? What benefits or challenges would result
from such a requirement?
20. Please explain whether regulatory standards regarding the use
of self-trading control technology would provide additional protection
to markets and market participants.
21. If you believe that self-trading controls are beneficial,
please describe the level of granularity at which such controls should
operate (e.g., should the controls limit self-trading at the executing
firm level? At the individual trader level?) What levels of granularity
are practical or achievable?
22. If you believe that self-trading controls are beneficial,
please explain whether exchanges should require such controls for
market participants and identify the categories of participants that
should be subject to such controls. For example, should exchanges
require self-trading controls for all participants, some types of
participants, participants trading in certain contracts, or
participants in market maker and/or incentive programs? What benefits
or challenges would result from imposing such controls on each category
of participant?
Price Collars
23. The Commission is aware that some exchanges already have price
collars in place for at least a portion of the contracts traded in
their markets. Please comment on whether exchanges should utilize price
collars on all contracts they list.
24. Would price collars provide additional protection in mitigating
credit risk to DCOs?
Maximum Order Sizes
25. Are such controls typically applied to all contracts and
customers, or on a more limited basis?
26. Do exchanges allow clearing members to use the exchange's
technology to set maximum order sizes for specific customers or
accounts?
27. Would additional standardization in the capabilities of this
technology or more uniform application of this technology to all
customers and contracts improve the effectiveness of such controls?
28. To what extent are clearing firms and trading firms conducting
pre-trade maximum order size screens? Please explain whether firms are
conducting such screens by utilizing: (1) Their own technology; (2) the
exchange's technology, or (3) a combination of both.
29. Would regulatory standards regarding the use of such technology
provide additional protection to the markets?
Trading Pauses
30. Trading pauses, as currently implemented, can be triggered for
multiple reasons. Are certain triggers more or less effective in
mitigating the effects of market disruptions?
31. Are there additional triggers for which pauses should be
implemented? If so, what are they?
32. What factors should the Commission or exchanges take into
account when considering how to specify pauses or what thresholds
should be used?
33. How should the re-opening of a market after a trading pause be
effected?
Credit Risk Limits
34. What positions should be included in credit risk limit
calculations in order to ensure that they are useful as a tool for
limiting the activity of a malfunctioning ATS? Is it adequate for such
a screen to include only those positions entered into by a particular
ATS or should it include all the firm's positions?
35. Should pre-trade credit screens require a full recalculation of
margin based on the effect of the order?
36. In light of your answers to the previous two questions, where
in the lifecycle of an order should the credit limits be applied and
what entity should be responsible for conducting such checks?
37. If credit checks are conducted post-trade, what should be done
when a trade causes a firm to exceed a limit?
38. Please describe any technological limitations that the
Commission should be aware of with respect to applying credit limits.
39. The Commission is particularly interested to receive public
comment on the ``hub'' model and its applicability to different types
of pre-trade risk controls. What are the strengths and weaknesses of
this approach relative to other pre-trade or post-trade approaches to
checking trades against credit limits? How would the latency between
the ``hub'' and the exchanges be managed to provide accurate limits for
high frequency ATS?
40. If you believe that post-trade credit checks would be an
effective safeguard against malfunctioning ATSs, what is the maximum
amount of latency that should be allowed for conducting such checks?
What technological or information flow challenges would have to be
addressed in order to implement post-trade checks with that degree of
latency?
41. With respect to any entity that you believe should be
responsible for applying credit risk limits, please describe the
technology necessary to implement that risk control and the cost of
such technology.
Order, Trade and Position Drop Copy
42. What order and trade reports are currently offered by DCMs and
DCOs? What aspects of those reports are most valuable or necessary for
implementing risk safeguards? Please also indicate whether the report
is included as part of the exchange or clearing service, or whether an
extra fee must be paid.
43. If each order and trade report described above were to be
standardized, please provide a detailed list of the appropriate content
of the
[[Page 56566]]
report, and how long after order receipt, order execution, or clearing
the report should be delivered from the trading platform to the
clearing member or other market participant.
Trade Cancellation or Adjustment Policies
44. Is a measure that would obligate exchanges to make error trade
decisions (i.e., decisions to cancel a trade or to adjust its price)
within a specified amount of time after an error trade is reported
feasible? If so, what amount of time would be sufficient for exchanges,
but would be sufficiently limited to help reduce risk for
counterparties to error trades?
45. Should exchanges develop detailed, pre-determined criteria
regarding when they can adjust or cancel a trade, or should exchanges
be able to exercise discretion regarding when they can adjust or cancel
a trade? What circumstances make pre-determined criteria more effective
or necessary than the ability to exercise discretion, and vice versa?
46. Do error trade policies that favor price adjustment over trade
cancellation effectively mitigate risk for market participants that are
counterparties to error trades? Are there certain situations where
canceling trades would mitigate counterparty risk more effectively? If
so, what are they and how could such situations be identified reliably
by the exchange in a short period of time?
47. Should error trade policies be consistent across exchanges,
either in whole or in part? If so, how would harmonization of error
trade policies mitigate risks for market participants, or contribute to
more orderly trading?
Order Cancellation Capabilities
48. The Commission's discussion of kill switches assumes that
certain benefits accrue to their use across exchanges, trading and
clearing firms, and DCOs. Please comment on whether such redundant use
of kill switches is necessary for effective risk control.
49. What processes, policies, and procedures should exchanges use
to govern their use of kill switches? Are there any different or
additional processes, policies and procedures that should govern the
use of kill switches that would specifically apply in the case of DMA?
50. What processes, policies, and procedures should clearing firms
use to govern their use of kill switches when using such a safeguard to
cancel and prevent orders on behalf of one or more clients?
51. What objective criteria regarding kill switch triggers, if any,
should entities incorporate into their policies and procedures?
52. What benefits or problems could result from standardizing
processes, policies, and procedures related to kill switches across
exchanges and/or clearing firms?
53. Please explain how kill switches should be designed to prevent
them from canceling or preventing the submission of orders that are
actually risk reducing or that offset positions that have been entered
by a malfunctioning ATS.
54. The Commission requests comment regarding whether kill switches
used by clearing firms already have or should have the following
capabilities: (a) Distinguish client orders from proprietary orders;
(b) distinguish among orders from individual clients; and (c) cancel
working orders and prevent additional orders from one or more of the
clearing firm's clients, or for all the clearing firm's proprietary
accounts, without cancelling and preventing all orders from the
clearing firm.
55. The Commission is aware of proposals that would enable FCMs to
establish credit limits for customers that are stored at a central
``credit hub'' for the purpose of pre-trade credit checks. If such a
model were implemented, is it possible that it could also be enabled
with kill switches that cancel existing working orders and prevent
additional orders from being submitted by one or more market
participants? Should such an approach be designed to complement kill
switches that are controlled by exchanges, clearing members, and
trading firms, or to replace these kill switches? What benefits and
drawbacks would result from each approach?
ATS Testing
56. Please describe the necessary elements of an effective ATS
testing regime, in connection with both the initial deployment and the
modification of an ATS.
57. With respect to testing of modifications, how should the
Commission and market participants distinguish between major
modifications and minor modifications? What are the objective criteria
that can be used to make such distinctions? Should any testing regime
applicable to ATS modifications distinguish between major and minor
modifications, and if so, how?
58. What challenges or benefits may result from exchanges
implementing standardized procedures regarding the development, change
management and testing of exchange systems? Please describe, if any,
the types of standardized procedures that would be most effective.
Crisis Management Procedures
59. Should basic crisis management procedures be standardized
across market participants? If so, what elements should be addressed in
an industry-wide standard?
60. Are there specific, core requirements that should be included
in any crisis management procedures? Similarly, are there specific
types of crisis events that should be addressed in any crisis
management procedures? If so, please identify such requirements and/or
crisis events and the level of granularity or specificity that the
procedures should have with respect to each.
Self-Certification and Clearing Firm Certification
61. How often should a market participant certify that their pre-
trade risk controls, post-trade reports and other measures, and system
safeguards meet the necessary standards?
62. Which representative of the market participant should be
required to attest that the certification standards have been met?
Should it be the market participant's chief executive officer, chief
compliance officer, or similar high-ranking corporate official, or some
other individual?
63. Which entity(ies) should receive certifications from market
participants? For example, should it be the market participant's
clearing firm, its designated self-regulatory organization (if
applicable), one or more trading platforms, a registered futures
association, the Commission, or other entity?
64. Should DCMs, SEFs or clearing member firms be required to audit
market participant certifications? What would be covered in an audit
and how often should these audits occur? Should the same entity that
receives the certification be required to perform the audit?
Risk Event Notification Requirements
65. Do commenters believe that risk event notifications would help
to better understand and ultimately reduce sources of risk in automated
trading environments? What information should be contained in a risk
event notification to maximize its value?
66. What types of risk events should trigger reporting
requirements, and what entities should receive risk event notifications
from market participants operating ATSs?
67. Which entities should receive risk event notifications?
[[Page 56567]]
ATS or Algorithm Identification
68. Should the Commission define ATS or algorithm for purposes of
any ATS identification system that may arise from this Concept Release?
If so, how should ATS or algorithm be defined? Should a separate
designation be reserved for high frequency trading algorithms and if
so, what is the threshold difference?
69. What are the existing practices within trading firms for
internally identifying ATSs or algorithms and for tracking their
performance, including profit and loss? What elements of existing
practices could be leveraged in any ATS or algorithm identification
system proposed by the Commission in the future?
70. The Commission understands that an ATS may consist of numerous
algorithms, each of which contributes to a trading decision. If an
algorithm-based identification system is proposed, which of the
potentially multiple algorithms that constitute an ATS should carry the
ID? In addition, what degree of change to an algorithm should
necessitate the use of a new ID, and how often does this change
typically occur? What is the appropriate definition of ``algorithm''
for purposes of an algorithm identification system?
71. If the identification system resides at the ATS level, how
should such IDs be structured to ensure that they are nonetheless
sufficiently granular to identify components that may be leading or
have led to unstable market conditions?
72. What message traffic between an ATS and a trading platform
should include the ATS or algorithm ID (all messages, orders only,
etc.)?
73. What relationship should this ATS ID have to the legal entity
identifier (LEI)?
Data Reasonability Checks
74. Please describe existing practices in the industry concerning
how and the extent to which ATSs use (1) market data; and (2) news and
information providers, including social media, to inform trading
decisions.
75. The Commission requests comment regarding any risk controls,
including reasonability checks, currently being used by market
participants operating ATSs to review market data and news and
information providers, including social media. Please describe the risk
control, including the purpose of the control, the extent of its use
among derivatives market participants, and any other aspects of the
risk control that you believe would be helpful for the Commission to
understand.
76. The Commission requests public comment concerning the lock-up
process for government economic reports, and any additional measures
that might be taken to protect against inappropriate disclosure.
77. Please describe the extent to which potentially market-moving
data from non-governmental economic reports can be obtained prior to
its public release for a fee. Are there specific reports or types of
reports for which early disclosure should not be permitted? What
process should be used for identifying non-governmental economic
reports whose early release should not be permitted? Should the data
release process for such reports be similar to the data lock-up process
implemented for the release of government economic data?
Registration of Firms Operating ATSs
78. Should firms operating ATSs in CFTC-regulated markets, but not
otherwise registered with the Commission, be required to register with
the CFTC? If so, please explain.
79. Please identify the firm characteristics, trading practices, or
technologies that could be used to trigger a registration requirement.
80. Should all firms deploying ATS be required to register, and
should there be different standards for firms deploying HFT strategies?
What are the appropriate thresholds levels below which registration
would not be required?
81. Since the floor trader distinction only addresses proprietary
traders, please explain whether there is any other category of market
participant, such as those deploying ATS or HFT strategies and trading
on behalf of clients (aside from market participants already subject to
Commission jurisdiction, such as Introducing Brokers and FCMs) that the
Commission should consider with respect to potential registration
requirements.
82. Should software firms providing algorithms be required to
register, and under what authority? What standards should apply to such
firms?
83. Please identify the functionalities discussed in this Concept
Release that could be applied to floor brokers that operate ATSs. Are
there any other controls not mentioned in this Concept Release that
should be under consideration?
84. Please supply any information or data that would help the
Commission in deciding whether firms may or may not meet the definition
of ``floor trader'' in Sec. 1a(23) of the Act.
85. Do you believe that the registration of such firms as ``floor
traders'' would effectuate the purposes of the CEA to deter and detect
price manipulation or any other disruptions to market integrity?
86. Considering the broad deployment of automated trading systems
across both equities and derivatives markets, the Commission seeks to
understand the appropriate level of coordination between itself and the
SEC in defining and applying possible standards to the ATS and HFT
trading space. How closely should the CFTC and SEC coordinate on
possible rules and requirements for trading firms? The Commission also
seeks public comment on the appropriate level of coordinated oversight
between itself and relevant Self-Regulatory Organizations such as
National Futures Association and FINRA.
87. Using the Flash Crash as an example, is it important to have
identical definitions and remedies in the case of ATS and HFT
registration requirements or do the existing market controls, such as
circuit breakers, provide the necessary market protections in both the
equities and derivatives markets? If the rules are not coordinated,
what impact would this have on market interaction and oversight?
88. If trading venues apply mandatory functionalities to access
derivatives markets, what benefit would a registration requirement
provide to the Commission?
Market Quality Data
89. What market quality indicators are in place today? Please
describe the metrics, how and where they are deployed, and how market
participants access these indicators and at what cost.
90. What value would each of the market quality metrics described
above [see section III.F.2] provide to market participants receiving
them? If possible, please be specific about how each market quality
measure could be used to enhance reliability and risk management of
ATSs.
91. Conversely, could any of the market quality metrics described
above [see section III.F.2] be used by market participants to
manipulate the order book, to identify competitors' trading strategies,
or to engage in other trading activities that do not contribute to
effective risk management and efficient discovery the traded asset's
economic value? If so, please provide specific information regarding
how such information could be misused. If possible, please provide
recommendations regarding steps the Commission could take to prevent
misuse.
[[Page 56568]]
92. Are there additional market quality metrics that the Commission
should contemplate requiring exchanges to provide? If so, what value
would they provide and how would they be used?
93. If the Commission determines that measures should be calculated
in the same way by various exchanges in order to provide comparable
measures of market quality, then how, specifically, should each of the
above mentioned metrics be calculated in order to ensure that they are
most valuable to market participants?
94. What timing and mode of dissemination is appropriate for each
metric? For example, should measures be provided as daily averages?
95. Does the liquidity of a given market impact which market
quality metrics would be reliable and useful when calculated for that
market? If so, which metrics are inapplicable in less liquid markets,
and why? What liquidity measures and thresholds are relevant to
determining which metrics should apply to a given market?
Market Quality Incentives
96. Should exchanges impose a minimum time period for which orders
must remain on the order book before they can be withdrawn? If so,
should this minimum resting time requirement apply to orders of all
sizes or be restricted to orders smaller than a specific threshold? If
there should be a specific threshold, how should that threshold be
determined?
97. The Commission seeks to understand where time-weighted Pro Rata
trade allocation is currently being utilized and what the effects have
been. Please note examples from exchanges and, to the extent possible,
please comment on the impact that such matching algorithms have had on
the amount of time resting orders are left in the order book, as well
as on other aspects of market quality.
98. If exchanges aggregated multiple, small orders entered by the
same entity with the intent of abusing rounding conventions to gain a
disproportionate share of allocations, what criteria should exchanges
use to distinguish such orders from those that are entered by the same
legal entity for legitimate trading purposes? Are there empirical
patterns that could be used to reliably identify such manipulative
intent?
99. Would batched order processing increase the number of
milliseconds that are necessary for correlations among related
securities to be established? If so, what specific costs would result
from this change and how do those costs compare to the potential
benefits described in recent research?
100. What costs and benefits result from providing market
participants with real-time access to information about the order book
that extends beyond aggregate size available at a limit price? Is there
a legitimate economic benefit that results from market participants
(both human participants, and ATSs) accessing such information? Is it
possible for market participants to use such information to manipulate
the order book?
101. The Commission seeks to understand whether any of the
recommendations above [see section III.F.3] are inapplicable or
irrelevant to markets subject to the CEA. If so, please indicate which
recommendation(s) and what makes it inapplicable or irrelevant to those
markets.
Policies and Procedures To Identify ``Related Contracts''
102. If you are a DCM, please address whether you have (i)
identified all contracts that are linked to, or are a substitute for,
other contracts either listed on your market or on other trading
venues; and, if so, (ii) coordinated your risk controls with any
similar controls placed on those other contracts. If you have not
identified such contracts and coordinated risk controls on such
contracts, please address any other means by which you are addressing
risk controls applicable to contracts that are linked to, or are a
substitute for, other contracts listed on your exchange or on other
trading venues.
103. Please explain whether it would be beneficial for exchanges to
develop and document policies and procedures for regularly reviewing
contracts on other exchanges in order to identify those that are
``linked to'' or that are ``a substitute for'' contracts listed on its
own market.
Standardize and Simplify Order Types
104. Please explain whether the standardization and simplification
of order types that have complex logic embedded within them would
reduce the potential for instability and other market disruptions. If
not, what other measures could achieve the same effect?
105. If the Commission were to consider the standardization and
simplification of order types in a future rulemaking, please identify
who should conduct this review (i.e., the Commission, trading
platforms, or other parties).
General Questions Regarding All Risk Controls
106. For each of the specified controls described above [see
sections III.C-F], please indicate whether you are already using the
control on customer and/or proprietary orders. If applicable, please
also indicate how widely you believe the control is currently being
used in the market, and how consistent the application of the control
is among firms.
107. If possible, please indicate specific costs associated with
implementing each of the risk controls described above [see sections
III.C-F]. Please include detailed estimates, distinguishing between the
cost of developing the functionality, the cost of implementation, and
the cost of ongoing operations.
108. Please describe the specific benefits associated with each of
the risk controls. Where possible, please indicate the market
participant category(ies) to which the benefit would accrue.
109. Please comment on the appropriate order of implementation and
timeline for each risk control, including any distinctions that should
be made based on the category of registrant or market participant
implementing the same or similar control, whether the market
participant is using DMA, and whether implementation is already in
place for certain categories.
110. Are any of the risk controls unnecessary, impractical for
commercial or technological reasons, or inadvisable? If so, please note
the control and provide reasons why.
111. A number of the pre-trade risk controls contemplated above are
similar protections at distinct points in the life of an order.
a. Please comment on the utility of redundant pre-trade risk
controls and the desirability of risk control systems in which controls
are placed at one or more than one focal points.
b. If pre-trade risk controls should reside at one or more than one
focal point, then please identify, for each risk control, what that
focal point should be?
112. Are there risk controls that should be implemented across
multiple entity types? If so, which controls and for which types of
entities should they apply? Also, please comment generally on the
factors the Commission should consider when determining the appropriate
entity(ies) upon which to place a risk control requirement that could
pertain to more than one entity.
113. Are there controls that should not be considered for
overlapping implementation across exchanges, clearing members and
market participants? If so, please explain which ones and why.
[[Page 56569]]
114. Each of the risk controls is described in general, principles-
based terms. Should the Commission specify more granular or specific
requirements with respect to any of the controls to improve their
effectiveness or provide greater clarity to industry participants? If
so, please identify the relevant control and the additional granularity
or specificity that the Commission should provide. Are any of the
controls, as currently drafted, inadequate to achieve the desired risk-
reduction?
115. To the extent that there is any need to standardize or provide
greater specificity regarding any measures discussed in this Concept
Release, including those that reflect industry best practices, please
describe the best approach to achieve such standardization (i.e.,
through Commission regulation, Commission-sponsored committee or
working group, or some other method).
116. How should risk control monitoring be implemented? Should
compliance be audited by internal and external parties? For each
control, please identify the appropriate entity(ies) to monitor
compliance with the control. Also, please describe what an acceptable
compliance audit would entail for each control.
117. Are there additional controls that should be considered, or
other methods that could serve as alternatives to those described above
[see sections III.C-F]? If so, please describe the control, its costs
and benefits, the appropriate entity(ies) to implement such control,
and whether there is any distinction to be drawn in the case of DMA.
118. Would any of the risk safeguards create a disincentive to
innovate or create incentives to innovate in an irresponsible manner?
If so, please identify the control, the concern raised, and how the
control should be amended to address the concern. Responses should
indicate how an amended risk control would still meet the Commission's
objectives.
119. Should the Commission consider any pre-trade risk controls,
post-trade reports, or system safeguards appropriate exclusively to
market makers or to ATSs used by market makers? If so, please describe
such controls or safeguards.
120. Should the Commission or Congress revisit its approach to
issuing civil monetary penalties for violations of the Act,
particularly as they relate to automated trading environments?
Currently, the maximum civil monetary penalty the Commission may issue
is capped at $140,000 ``per violation.'' Is such a civil monetary
penalty sufficient to deter acts that constitute violations of the Act,
given that an individual violation could impose costs to the market and
the public well in excess of $140,000?
121. Please describe the documentation (or categories of documents)
that would demonstrate that a market participant operating an ATS has
implemented each risk control addressed in this Concept Release,
including, for example, computer code, system testing results,
certification processes and results, and calculations.
122. Would a fee (collected by, for example, the DCM or SEF) on
numbers of messages exceeding a certain limit be more appropriate than
a hard limit on the number or rate of messages?
123. Should such a penalty be based on a specified number or rate
of messages or on the ratio of messages to orders filled over a
specified time period?
124. Recent disruptive events in securities markets illustrate the
importance of effective communication between exchanges' information
technology systems. The Commission requests public comments regarding
relevant systems in its regulated markets, including both DCMs and
SEFs. What data transfers or other communications between exchanges are
necessary for safe, orderly, and well-functioning derivatives markets?
What additional measures, if any, would help promote the soundness of
such systems (e.g., testing requirements, redundancy standards, etc.)?
V. Appendices (Specific Measures in Bold Font)
A. Pre-Trade Risk Controls
----------------------------------------------------------------------------------------------------------------
Party(s) to implement risk
Potential pre-trade risk control control Substance of control
----------------------------------------------------------------------------------------------------------------
1a. Maximum Message Rate (Message Market Participants 1a. Market participants operating ATSs
Throttle). Operating ATSs, Trading must establish a maximum message rate
Platforms, and Clearing per unit time for each ATS. This control
Firms. should be calibrated to address the
potential for unintended message flow
(including orders) from a malfunctioning
ATS. Market participants' systems must
prevent the submission of messages in
excess of the specified rate.
Trading platforms' systems must prevent
the acceptance of messages in excess of
their own specified rates and must log
instances when each ATS attempted to
exceed such limits.
Separately, trading platforms must
establish systems enabling clearing
firms to set rate limits directly at the
trading platform. Trading platforms,
clearing firms and market participants
may set rates independently of each
other.
In all cases, human monitors must be
alerted when limits are breached.
1b. Maximum Execution Rate (Execution Market Participants 1b. Market participants operating ATSs
Throttle). Operating ATSs, Trading must establish a limit on the maximum
Platforms, and Clearing number of orders that each of their ATSs
Firms. can execute in a given direction per
unit time. The limit should be unique to
each ATS and should be calibrated to
address the potential for unintended
executions arising from a malfunctioning
ATS. Additional orders in excess of the
limit should not be submitted or
executed.
Trading platforms must establish a
maximum number of orders in the same
direction they will execute per unit
time from a uniquely identified ATS, and
must prevent execution of trades that
would violate this limit.
Separately, trading platforms must
establish systems enabling clearing
firms to set per-customer message rate
limits directly at the trading platform.
Trading platforms, clearing firms and
market participants may set rates
independently of each other.
[[Page 56570]]
2. Volatility Awareness Alerts.......... Market Participants Market participants operating ATSs must
Operating ATSs. implement automated solutions to
immediately notify system supervisors
when the prices of individual or groups
of assets relevant to an ATS's trading
strategies move either up or down by a
given percentage within a predetermined
period of time, or when the volume of
individual or groups of assets relevant
to an ATSs trading strategies over a
specific period of time increase or
decrease beyond a predetermined
threshold. This control should help
system supervisors identify market
conditions which are not appropriate to
the continued operation of a particular
ATS or algorithm. The alert should be
configurable by contract.
3. Self-Trade Controls.................. Trading Platforms and All Trading platforms must provide, and all
Market Participants. market participants must apply,
technologies to identify and limit the
transmission of orders from their
systems to a trading platform that would
result in self-trades.
4. Price Collars........................ Trading Platforms and All Trading platforms must assign a range of
Market Participants. acceptable order and execution prices
for each of their products. All orders
outside of this range would be
automatically rejected, and orders
already in the order book but outside of
the acceptable range should not be
elected by the matching engine.
All market participants must establish
similar product-specific price collars
and should implement systems to ensure
that orders outside of the collar are
not transmitted to the relevant trading
platform.
5. Maximum Order Size................... Trading Platforms, Clearing Trading platforms, clearing firms, and
Firms, and All Market all market participants must each
Participants. establish default maximum order sizes
for orders submitted, transmitted, or
processed by their systems.
A market participant's systems must
prevent the submission of orders in
excess of its internally-specified
limits. A clearing firm's systems must
prevent the transmission of customer
orders in excess of its limits for that
customer. Trading platforms must prevent
their systems from processing or
executing orders in excess of the limit
specified by the trading platform.
In addition, for DMA customers, trading
platforms must establish similar systems
enabling clearing firms to set per-
customer order size limits directly at
the trading platform.
Limits set by market participants,
clearing firms, and trading platforms
may be different from, and operate
independently, of each other.
6. Trading Pauses....................... Trading Platforms.......... Trading platforms would be required to
institute trading pauses, similar in
nature to stop-logic functionality, but
covering a wider array of adverse states
of an automated central limit order
book.
7. Credit Risk Limits................... Trading Platforms, Clearing While some trading firms and FCMs conduct
Firms and/or Market post-trade credit checks with varying
Participants Operating degrees of latency and pre-trade credit
ATSs. risk screens are already required
pursuant to Commission regulations, the
Commission seeks public comments
regarding any additional measures that
could help protect the financial
integrity of DCOs, as well as additional
input from the public regarding the
appropriate location and timing in the
order lifecycle for credit checks.
----------------------------------------------------------------------------------------------------------------
B. Post-Trade Reports and Other Post-Trade Measures
----------------------------------------------------------------------------------------------------------------
Party(s) to implement
Potential post-trade report or measure report or measure Substance of report or measure
----------------------------------------------------------------------------------------------------------------
8. Order Report (Post-order drop copy).. Trading platforms.......... Trading platforms must provide a
duplicate copy of each order to the
originating market participant and to
the market participant's clearing
firm(s) simultaneously with such order's
receipt by the trading platform.
9. Trade Report (Post-trade drop copy).. Trading platforms.......... Trading platforms must provide a
duplicate copy of each executed trade to
the originating market participant and
to the market participant's clearing
firm(s) simultaneously with such trade's
execution by the trading platform.
10. Position Report (Post-clearing drop DCOs....................... DCOs must provide net position per
copy). maturity per contract to the originating
market participant and the market
participant's clearing firm(s) as soon
as the contract is matched at the
clearinghouse.
11a. Uniform Adjust or Bust Error Trade Trading platforms and All 11a. Trading platforms must establish
Policies. Market Participants. policies for adjusting the price of
trades or breaking trades that have been
executed due to an error.
Policies must favor price adjustments
rather than trade cancellation. To the
extent possible, policies must require
decisions by the trading platform to be
made on the basis of readily available
objective criteria in order to
facilitate rapid or immediate decisions.
11b. Standardized Reporting Window for ........................... 11b. Market participants must report
Error Trades. error trades to the trading platform
within five minutes after the trades are
executed.
Trading platforms must notify market
participants of a potential adjust-or-
bust situation immediately.
Trading platforms must make a decision
and notify market participants of that
decision within a specified period of
time.
----------------------------------------------------------------------------------------------------------------
[[Page 56571]]
C. System Safeguards
----------------------------------------------------------------------------------------------------------------
Party(s) to implement
Potential system safeguard safeguard Substance of safeguard
----------------------------------------------------------------------------------------------------------------
CONTROLS OVER ORDER PLACEMENT........... Trading Platforms, Clearing Trading platforms, clearing members, and
Firms, and All Market market participants must have systems
Participants. and processes in place to:
Order Cancellation Capabilities
12a. Auto-cancel on disconnect.......... ........................... 12a. Exchanges should implement a
flexible system that allows a user to
determine whether their orders should be
left in the market upon disconnection.
This should only be implemented if the
clearing firm's risk manager has the
ability to cancel working orders for the
trader if the trading system is
disconnected. The exchange should
establish a policy whether the default
setting for all market participants
should be to maintain or to cancel all
working orders.
12b. Selective working order ........................... 12b. Immediately cancel one, multiple, or
cancellation. all resting orders from a market
participant as deemed necessary in an
emergency situation.
12c. Kill switch........................ ........................... 12c. Immediately cancel all working
orders, and the ability to prevent
submission (market participant),
transmittal (clearing member), or
acceptance (trading platform) of any new
orders from a market participant, or
particular trader or ATS of such market
participant.
13. Repeated Automated Execution ........................... 13. Market participants operating ATSs
Throttle. must establish a limit on the maximum
number of orders that each ATS can
submit. When an ATS reaches that maximum
it must be automatically disabled until
a human re-enables it.
14. System heartbeats (see section ........................... 14. Trading platforms must provide, and
III.E.1.a and footnote 90). market participants operating ATSs must
utilize, heartbeats that indicate proper
connectivity between the trading
platform and the ATS. Such heartbeats
must also indicate the status of
connectivity between an ATS and any
systems used by the trading platform to
provide the ATS with market data.
If connectivity to any system is lost,
the ATS should be disabled, and resting
orders should be maintained or cancelled
based on the pre-determined preferences
of the firm that lost connectivity.
POLICIES AND PROCEDURES FOR THE DESIGN, Market Participants 15a. Market participants operating ATSs
TESTING, AND SUPERVISION OF ATSs Operating ATSs. must properly design their systems to
15a. ATS Design......................... avoid violations of the CEA, Commission
regulations, or DCM and SEF rules
related to fraud, disruptive trading
practices, manipulation and trade
practice violations. They must also
ensure that their ATSs include all
applicable pretrade risk controls and
system safeguards as described herein.
15b. ATS Development and Change Trading Platforms and 15b. Trading platforms and market
Management. Market Participants participants operating ATSs must
Operating ATSs. maintain a development environment that
is adequately isolated from the
production trading environment. The
development environment may include
computers, networks, and databases, and
should be used by software engineers
while developing, modifying, and testing
source code.
Firms must maintain a source code
repository to manage source code access,
persistence, and changes.
Firms must establish and document
procedures for communicating the
functionality and requirements of, and
changes to, their proprietary software.
These procedures must include an audit
trail of material changes that would
allow them to determine, for each
change: Who made it, when they made it,
and what the purpose was for the change.
Firms must have documented policies and
procedures that allow representatives
from trading, risk, and software
management to approve changes and to
verify internal testing before a new or
modified trading system can be enable in
production.
15c. ATS Testing........................ Trading Platforms and 15c. Market participants operating ATSs
Market Participants must test each ATS both internally and
Operating ATSs. on each trading platform on which an ATS
will operate. Relevant tests include,
but are not limited to, unit testing,
functional testing (both integration and
regression testing), non-functional
testing, and acceptance testing.
Functional testing must include all
applicable pre-trade risk controls, post-
trade reports and other measures and
system safeguards. Non-functional
testing must include testing under
stressed market conditions.
Market participants must perform such
testing on each algorithm prior to
initial deployment, and prior to re-
deployment, after certain modifications
to the algorithm.
Trading Platforms must provide test
environments that simulate the
production trading environment so that
market participants may conduct exchange-
based conformance testing on their ATSs
once they have completed internal
testing. Conformance testing must
include tests for all ATS risk
mitigation controls that are able to be
tested by the exchange.
Exchange-based conformance testing must
be done after certain modifications to
the operating code.
15d. ATS Monitoring and Supervision..... Market Participants 15d. Market participants operating ATSs
Operating ATSs. must ensure that their ATSs are subject
to continuous real-time monitoring and
supervision by trained and qualified
staff at all times while engaged in
trading.
[[Page 56572]]
Appropriate supervision includes
automated alerts when ATS order behavior
breaches design parameters or when
market conditions diverge from program
expectations. It also includes automated
alerts upon loss of network connectivity
or data feeds.
Monitoring and supervision staff must
have the ability and authority to
disengage the ATS and to cancel resting
orders when system or market conditions
require it, including the ability to
contact trading platform staff to seek
information and cancel orders. They must
also have acceptable dashboards and
control panels to monitor and interact
with the ATS.
Monitoring and supervision staff must
record the time when they assume
responsibility for an ATS and the time
when they relinquish control to others.
Recording must be achieved through
distinct log-ins to the required control
panel by each staff person. Log-in must
also be subject to access controls that
ensure the correct staff person is
identified.
15e. Training for ATS Monitoring Staff Market Participants 15e. Firms operating ATSs must develop
(see section III(E)(2)(b) and footnote Operating ATSs. training for all staff involved in
97). monitoring or designing ATSs. Training
must, at a minimum, cover design
standards, event communication
procedures, and requirements for
notifying exchange and commission staff
when risk events occur.
Additionally, each firm must develop,
document, and implement training
policies that ensure human monitors are
adequately trained for each new
algorithm that is implemented. Training
must include, at a minimum, the economic
rationale for the algorithm and
mechanics of the underlying process, as
well as the automated and non-automated
risk controls that are applicable to the
algorithm.
15f. Crisis Management Procedures....... Trading Platforms and 15f. Trading platforms and market
Market Participants participants operating ATSs must develop
Operating ATSs. and document procedures that direct the
actions of ATS supervisors, exchange
trading monitors, and support staff in
the event that an algorithm malfunctions
or responds to market signals in an
unanticipated manner.
Procedures should direct the process for
evaluating, managing, and mitigating
market disruption and firm risk. The
procedures should also specify people to
be notified in the event of an error
that results in violations of risk
profiles or potential violations of
exchange or Commission rules.
SELF-CERTIFICATIONS AND NOTIFICATIONS
16a. Self-Certification and Clearing Market Participants 16a. All firms operating ATSs must
Firm Certification. Operating ATSs. certify annually that their ATSs
individually and collectively (i.e. at
the algorithm, account, and firm levels)
comply with all Commission and trading
platform requirements regarding pre-
trade risk controls and post-trade
reports and other measures, as well as
all applicable risk controls.
Clearing firms must institute reasonable
measures to confirm that their client
trading firms implement the pre-trade
risk controls that are required.
16b. Risk Event Notification Market Participants 16b. Market participants operating ATSs
Requirements. Operating ATSs, Trading must notify the exchange, and the
platforms. exchange must notify the Commission
whenever an algorithm violates its
design parameters or whenever risk
control technologies or processes do not
function as planned even if they do not
result in destabilization of the
markets. The exchange must also notify
the Commission whenever any of its own
risk management technologies or
processes violate design parameters or
do not function as planned.
17. ATS or Algorithm Identification..... Market Participants A unique identifier would be assigned to
Operating ATSs. each ATS or algorithm, and all orders
submitted by that ATS or algorithm would
be tagged with the identifier.
18. Data Reasonability Checks........... Market Participants All firms operating ATSs must have
Operating ATSs. ``reasonability checks'' on incoming
market data and other data (including
social media).
----------------------------------------------------------------------------------------------------------------
D. Other Protections
----------------------------------------------------------------------------------------------------------------
Party(s) to implement
Potential additional protection protection Substance of protection
----------------------------------------------------------------------------------------------------------------
19. Registration of All Firms Operating Market Participants All firms operating ATSs to trade solely
ATSs. Operating ATSs. for their own account and not otherwise
registered with the Commission must
register with the Commission.
20. Market Quality Data................. Trading Platforms.......... Trading platforms must provide to all
market participants a daily summary of
market quality for each product traded
on its platform.
The feeds would include measures of
execution quality including: (1)
Effective spreads; (2) order to fill
ratios; (3) execution speed for
different types of orders and different
order sizes; (4) aggressiveness
imbalance; (5) price impact for given
trade sizes; (6) average order duration;
(7) order efficiency; (8) rejection
order ratio; (9) net position changes
versus volume; (10) branching ratios;
(11) volume imbalance and trade
intensity; (12) Herfindahl-Hirschman
Indexes based on market share of open
positions under common control; and (13)
metrics on the number of price changing
trades involving ATSs.
[[Page 56573]]
21. Market Quality Incentives........... Trading Platforms.......... Trading platforms must implement changes
that will limit market participants'
abilities to improperly advantage their
own orders in ways that do not
contribute to efficient price discovery,
including, for example: (1) Utilize a
trade allocation formula that is an
intermediate between a cardinal ranking
(time-weighted), Pro Rata allocation
formula and a Price/Time allocation
formula; (2) Create a new limit order
type that would prioritize orders that
remain resting in the order book for
some minimum amount of time; (3) Require
orders not fully visible in the order
book to go to the end of the queue
(within limit price) with respect to
trade allocation; (4) Aggregate
multiple, small orders from the same
legal entity entered contemporaneously
at the same price level and assign them
the lowest priority time stamp of all
the orders so aggregated; (5) Require
exchanges to use batch auctions once per
half second at random times rather than
use continuous trade matching; and (6)
Limit visibility into the order book to
aggregate size available at a limit
price.
22. Policies and Procedures for Trading Platforms.......... Trading platforms must develop and
identifying ``related'' contracts. implement policies and procedures for
identifying securities or products
listed on other exchanges that would
constitute ``related'' contracts to
those that are listed on their own
exchange.
23. Standardize and Simplify Order Types Trading Platforms.......... Trading platforms must work with the
Commission to standardize order types
across exchanges, and to reduce the
overall number of order types that have
complex logic embedded within them.
----------------------------------------------------------------------------------------------------------------
Issued in Washington, DC, on September 9, 2013, by the
Commission.
Christopher J. Kirkpatrick,
Deputy Secretary of the Commission.
Appendices to Concept Release on Risk Controls and System Safeguards
for Automated Trading Environments
Appendix 1--Commission Voting Summary
On this matter, the following Commissioners voted in the
affirmative: Chairman Gensler, Commissioner Chilton (with the
concurrence set out below in Appendix 3), Commissioner O'Malia (with
the concurrence set out below in Appendix 4), and Commissioner
Wetjen. No Commissioner voted in the negative.
Appendix 2--Statement of Support of Chairman Gary Gensler
We have witnessed a fundamental shift in markets from human-
based trading to highly automated electronic trading. Automated
trading systems, including high frequency traders, enter the market
and execute trades in a matter of milliseconds without human
involvement. Electronic trading makes up over 91 percent of the
futures market. The swaps market also is moving toward electronic
trading.
In our oversight of U.S. derivatives markets, both futures and
swaps, the Commodity Futures Trading Commission (CFTC) must look to
continually adapt our regulations in these changing times. Our
mission to promote transparency, ensure for market integrity and
prohibit abuses is just as important in the fast-moving world of
electronic trading as it was when people traded over the phone, in a
pit or on a floor.
The CFTC already has taken a number of important steps to keep
pace with rapidly evolving 21st-century markets. We have adopted
rules to implement pre-trade risk filters for futures commission
merchants, swap dealers, designated contract markets and swap
execution facilities. We also have new rules to prohibit disruptive
trading practices and other market abuses.
In publishing this Concept Release, we are seeking public input
on what additional risk controls and system safeguards are
appropriate given this ever-changing technological environment.
Traditional risk controls and system safeguards, many of which were
developed according to human speed and floor-based trading, must be
evaluated in light of new market realities.
Further, as sure as computers and programs have had technical
glitches in the past, we must look to risk controls and system
safeguards to protect markets when such glitches inevitably occur
again. This Concept Release is intended to stir public discussion
and debate on how best to protect the functioning of markets for the
benefit of farmers, ranchers, merchants and other end users who rely
on markets to hedge risk--particularly in light of the reality that
the majority of the market is using automated trading systems.
Appendix 3--Concurrence of Commissioner Bart Chilton
While I concur in the concept release, am most appreciative of
the staff work, and am largely pleased at the result, this has taken
far too long to come to fruition.
In general, those involved in financial markets seem to have
blindly accepted that technology is almost always a good thing. Yet
we continue to see major technology problems, like NASDAQ shutting
down twice in as many weeks. Last year it was NYSE. In the futures
world, we see technology glitches that simply should not occur. I
acknowledge that, with the staggering volume of trading, some might
simply be astounded that--in the main--it works so well. But it
doesn't work well enough if we continue to see aberrations--
particularly if they are market missteps that could have been
avoided. That's to say nothing of the high frequency cheetah traders
who have, some I am convinced intentionally, contorted markets in a
manipulative fashion. In addition, there are a shocking number of
transactions that appear to be wash trades--that also has the
possibility of impairing the fair and effective functioning of
financial markets.
I'm pleased we are moving this concept release forward, but
given this environment it has taken way too long. If we continue at
this pace, Rip Van Winkle could keep up with any possible action we
might take. We need to understand that some of these issues are
urgent and need action now. They can't wait another year or more.
At the same time, there is one thing that can be done now. In
fact, I suggested this policy shift be included in the concept
release, but since it is a larger issue than just a technology-
related matter, it was decided to omit it. That's fine, because my
suggestion is really an action for the Congress.
As long as we have a puny penalty regime at the CFTC, we are
going to see traders risk getting caught because the potential
profits are so great. We can only impose a civil monetary penalty
(CMP) of $140,000 per violation. That's the law. Furthermore, the
case history suggests that a ``violation'' may be only once per day.
In these millisecond markets where we have seen a million change
hands in a minute, $140k is a joke--and it's not very funny.
This Agency is hampered by staffing needs due to a lack of
funding. We have hundreds of cases being investigated right now. The
least Congress can do, so that we can try and keep up--and if need
be, cage the cheetahs and others who violate the Commodity Exchange
Act--is to increase the CMPs. Specifically, I've suggested
increasing the maximum penalty levels to $1 million per violation
for individuals and $10 million for firms. That would be a
deterrent. That would stop some of the cheetahs and others out there
who are tempted to use powerful technologies in unlawful ways.
I look forward to receiving comments, and hope that we let no
moss grow on this matter.
Appendix 4--Statement of Concurrence by Commissioner Scott D. O'Malia
During my time at the Commodity Futures Trading Commission
(``Commission''), I have
[[Page 56574]]
consistently emphasized that the Commission must have a strong
understanding of today's highly automated and interconnected trading
environments in order to oversee its markets effectively. As head of
the Commission's Technology Advisory Committee (``TAC''), I have
committed considerable TAC time and resources to strengthening our
understanding of automated markets. I am grateful for all the hard
work of the TAC members as well as the efforts of the members of the
Subcommittee on Data Standardization and the Subcommittee on
Automated and High Frequency Trading, who have devoted hours of work
on issues related to automated trading systems and pre-trade
functionality. I hope that this Concept Release, and in particular
the public comments the Commission receives in response, will build
on this work.
The Concept Release asks over a hundred questions, which is
appropriate given the importance of hearing from all sectors of the
industry and benefiting from their knowledge and views of automated
trading. I would like to highlight a few questions that I believe it
would be particularly constructive to receive feedback from the
public on. The first is to establish what current protections are in
the market today and the extent to which the technology is deployed,
as well as its effectiveness. The second is an overarching question:
Whether there is a need for regulatory action with regard to any of
the measures currently in the market. In other words, should the
Commission federalize any current industry practices/standards?
Third, it would be helpful to receive public feedback on the
definitions for high-frequency trading and automated trading systems
that the TAC, after extensive effort by its Subcommittee on
Automated and High Frequency Trading, has proposed. Finally, it
would be beneficial to receive feedback on the possibility of a
registration requirement for firms operating automated trading
systems and not otherwise registered with the Commission. The
Concept Release cites the definition of ``floor broker'' as the
potential basis for such a requirement; I am interested to get
public input on whether this, or any other provision in the
Commission's statute or regulations, can serve as a valid foundation
for registration.
The Concept Release is far from perfect. For example, it could
have provided a more thorough and clear cataloguing of existing
industry practices and recommendations; a recent TAC reference
document is more clear and concise in compiling existing standards
and recommendations in the market today.\124\ Nevertheless, I
support today's issuance of the Concept Release in order to receive
input from market participants on all of the issues contained
herein. I look forward to reviewing the comments submitted in
response to the Concept Release.
---------------------------------------------------------------------------
\124\ This document is available at http://www.cftc.gov/ucm/groups/public/@newsroom/documents/file/tac103012_reference.pdf.
[FR Doc. 2013-22185 Filed 9-10-13; 8:45 am]
BILLING CODE 6351-01-P