changing notions of risk management in financial markets
TRANSCRIPT
Rajib Ranjan BorahCo-Founder & Director,
iRageCapital Advisory Pvt Ltd; QuantInsti Quantitative Learning Pvt
Ltd
Changing Notions of Risk Management in Financial
Markets
Impact of Proliferation of Automated Trading Systems and Technology on Financial Markets
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Table of Contents
• Changing Trends in Trading
• Major Automated Trading Risk Failures
• Changing Trends in Trading Risk Management
• Regulatory framework in India
• Risk Management Process
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Table of Contents
• Changing Trends in Trading
• Major Automated Trading Risk Failures
• Changing Trends in Trading Risk Management
• Regulatory framework in India
• Risk Management Process
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Trading in the marketsIf you have a profitable trading strategy, then …
1) Do it as frequentlyDon’t miss any opportunityComputers are always at their seatsRespond to opportunities in microseconds
3) Don’t let emotions affectGreed & fear are traders’ biggest enemies
Computers have no emotions
2) Scale it up Trade as many financial instrumentsHuman eye can monitor 10-15 stocksComputers can track thousands simultaneously
Trading is all about computations and computers do calculations faster.
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Trading Today
… and even more computers
Trading shifted from pits …
… to computers
Inevitably, machines have taken over human beings
1 2
43
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Effect of Algo-TradingOptions
FX
Equity
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
India woke up recentlyAlgorithmic trading is in nascent stage; and is picking steam super-fast
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Pros and ConsTrading algorithmically is generally more profitable
Pros• Less downtime
• No emotions (Greed & Fear)
• React faster
• Higher scalability
• Accurate and faster
calculations
Cons• Complicated systems
• Increasing errors
• More riskier
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Table of Contents
• Changing Trends in Trading
• Major Automated Trading Risk Failures
• Changing Trends in Trading Risk Management
• Regulatory framework in India
• Risk Management Process
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Major algorithmic trading incidents - I
• Credit Suisse, Nov 2007– Incident:
• Hundreds of thousands of cancel orders sent to the exchange
• Orders clogged NYSE and affected trading of over 900 stocks
– Reasons:• Trader implemented code which could change parameters
on clicking on spin button
(without any need for confirmation)
• With each click, orders were cancelled and resent
– Fine/ Losses:• $150,000 fine
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Infinium Capital, Feb 2010– Incident:
• 4612 trades on crude oil futures in 24 seconds
– Reasons:• Strategy was designed to trade energy ETFs on the basis of
crude prices
• Trader configured crude oil futures on the basis of energy ETFs
• Moreover, RMS was designed on the basis of ETF prices, not crude prices
– Fine/ Losses:• $850,000 fine by CME
Major algorithmic trading incidents - II
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Deutsche Bank, June 2010– Incident:
• Sent orders for 1.24 million Nikkei 225 Futures & 4.82 million Nikkei 225 mini-futures in first few minutes
• More than 10 times normal volume
• Market dropped 1% on orders
– Reasons:• Pair trade strategy used value of mini-Nikkei to quote
Nikkei. At start of day, there was no liquidity in mini-Nikkei
• Error recognized immediately, 99.7% orders cancelled
– Fine/ Losses:• Forced to close Algorithmic trading desk in Tokyo
Major algorithmic trading incidents - III
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• BATS listing, Mar 2012
– Incident:
• On the day of listing, stock price dropped 99%
– Reasons:
• Software bug in newly installed exchange matching engine - orders placed during auction session became inaccessible for stocks whose ticker symbols began with letters A to BFZZZ
– Fine/ Losses:
• IPO withdrawn
Major algorithmic trading incidents - IV
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Knight Capital, Aug 2012– Incident:
• Traded 154 stocks at bizarre prices (4 million trades for 397 million shares in 45 minutes): alternately bought at higher prices and sold at lower prices
– Reasons:• Accidentally installed test software which incorporated an
old piece of code designed 9 years ago
• In one out of 8 production servers, new code was not installed by a technician
• No process for second technician to review
– Fine/ Losses:• Trading loss of $460 million in 45 minutes. Fine of $12
million
• Knight Capital had to be rescued by Getco
Major algorithmic trading incidents - V
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Goldman Sachs, Aug 2013
– Incident:
• Traded stock options at very erroneous prices at the exchange
– Reasons:
• Indication of interests were sent as actual orders to the exchange
– Fine/ Losses:
• Trading loss of $100 million
Major algorithmic trading incidents - VI
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Tel Aviv Stock Exchange, Aug 2013
– Incident:
• Shares of Israel Corp. country's largest holding company fell sharply from 167,200 Israeli Shekels to 210 Shekels.
– Reasons:
• Trader wrongly entered Israeli Corp as scrip name instead of some other firm
– Fine/ Losses:
• All trades cancelled
Major algorithmic trading incidents - VII
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Everbright Securities, Aug 2013– Incident:
• Rogue algorithm kept buying – index moved up 6% intraday
• Did not inform regulators, shorted the artificial bubble – banned from prop trading forever for insider trading
– Reasons:• Trader wrongly entered Israeli Corp as scrip name
instead of some other firm
– Fine/ Losses:• Banned from prop trading forever for insider trading
Major algorithmic trading incidents -VIII
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• HanMag Securities, Dec 2013– Incident:
• HanMag exercised wrong call and put options
• 36,100 trades in a few minutes
– Reasons:• Error in automated profit taking trade program
(interchanged puts with calls)
– Fine/ Losses:• Some firms returned money back to HanMag (Optiver
returned $600k trading profits)
• Eventual loss of 57 billion Korean Won
Major algorithmic trading incidents - IX
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• United Airlines
– Incident:
• On Sep. 7, 2008 United Airlines had a downward price spike
– Reasons:
• Google’s newsbots picked up an old 2002 story about United Airlines possibly filing for bankruptcy
• News Analytics based automated traders reacted to it
Major algorithmic trading incidents - X
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• United Airlines– Incident:
• On Apr 23, 2013 Markets dropped 0.8% momentarily
– Reasons:• Twitter account of news publisher hacked – false news
of White house explosion
• News Analytics based automated traders reacted to it
Major algorithmic trading incidents - XI
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Table of Contents
• Changing Trends in Trading
• Major Automated Trading Risk Failures
• Changing Trends in Trading Risk Management
• Regulatory framework in India
• Risk Management Process
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Trading Risks
Traditionally trading
operations focused
on following risks
• Market Risk
• Credit / Counter-party Risk
• Financial Risk
• Liquidity Risk
• Regulatory Risk
Automated trading
requires additional
focus on
• Operational Risk
• System Risk
• Greater focus on Natural
Disaster Risk
• Regulatory Risk (Automated
Trading related)
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Issues with Algo-Trading• Orders flow without human control
– Higher reliance on technology
– GIGO (Garbage Input → Garbage Output)
• Before a human can realize (and then respond) → tremendous damage would happen already
• Trades happen at such a fast pace → positions could become huge in no time
– Real-time monitor of positions, exposures, regulation checks
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Algo-trading system risks• System and Operational Risks specific to automated
trading can be classified into the following categories:– Access– Consistency– Quality– Algorithm– Technology– Scalability
• Such System and Operational risks have to be handled pre-order– Within the application – Before generating an order in the Order Management
System
• Moreover, it is pertinent that the trader understands the internal working of the black-box
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Table of Contents
• Changing Trends in Trading
• Major Automated Trading Risk Failures
• Changing Trends in Trading Risk Management
• Regulatory framework in India
• Risk Management Process
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Half-yearly system audit conducted only for algorithmic trading facility
• Members are required to provide following information on NSE-ENIT:– details of all algorithmic strategies in the template provided– auditor certificate
• Audit provides following reports:– Summary report: Ratings of ‘Strong’, ‘Medium’ or ‘Weak’ on each
broad areas (which is to be submitted to exchange via NSE-ENIT)– Detailed report
• In case audit report has a rating of Weak, the member is required to submit an ATR (Action Taken Report) to exchange
• Auditors to provide report on their letter heads:– List of all strategies approved
Audit Process & Requirements
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
SEBI’s broad guidelines on Algorithmic Trading
(Circular CIR/MRD/DP/09/2012 dated 30 Mar 2012):
Guideline for exchanges:
• The stock exchange shall have arrangements, procedures and system capabilityto manage the load on their systems in such a manner so as to achieve consistentresponse time to all stock brokers. The stock exchange shall continuously studythe performance of its systems and, if necessary, undertake system up gradation,including periodic up gradation of its surveillance system, in order to keep pacewith the speed of trade and volume of data that may arise through algorithmictrading.
• In order to ensure maintenance of orderly trading in the market, stock exchangeshall put in place effective economic disincentives with regard to high daily order-to-trade ratio of algorithmic trading orders of the stock broker. Further, the stockexchange shall put in place monitoring systems to identify and initiate measuresto impede any possible instances of order flooding by algorithms.
• The stock exchange may seek details of trading strategies implemented throughalgorithmic trading for such purposes viz. inquiry, surveillance, investigation, etc.
SEBI guidelines
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• The stock exchange shall include a report on algorithmic trading on the stock exchange in the Monthly Development Report (MDR) submitted to SEBI inter-alia incorporating turnover details of algorithmic trading, algorithmic trading as percentage of total trading, number of stock brokers / clients using algorithmic trading, action taken in respect of dysfunctional algorithms, status of grievances, if any, received and processed, etc.
• The stock exchange shall synchronize its system clock with the atomic clock before the start of market such that its clock has precision of atleast one microsecond and accuracy of atleast +/- one millisecond.
• Stock exchange shall ensure that the stock broker shall provide the facility of algorithmic trading only upon the prior permission of the stock exchange. Stock exchange shall subject the systems of the stock broker to initial conformance tests to ensure that the checks mentioned below are in place and that the stock broker’s system facilitate orderly trading and integrity of the securities market. Further, the stock exchange shall suitably schedule such conformance tests and thereafter, convey the outcome of the test to the stock broker.
SEBI guidelines
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Guideline to brokers:
• The stock broker, desirous of placing orders generated using algorithms, shallsubmit to the respective stock exchange an undertaking that –
– The stock broker has proper procedures, systems and technical capability tocarry out trading through the use of algorithms.
– The stock broker has procedures and arrangements to safeguard algorithmsfrom misuse or unauthorized access.
– The stock broker has real-time monitoring systems to identify algorithms thatmay not behave as expected. Stock broker shall keep stock exchangeinformed of such incidents immediately.
– The stock broker shall maintain logs of all trading activities to facilitate audittrail. The stock broker shall maintain record of control parameters, orders,trades and data points emanating from trades executed through algorithmtrading.
– The stock broker shall inform the stock exchange on any modification orchange to the approved algorithms or systems used for algorithms.
SEBI guidelines
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
SEBI later laid out additional guidelines pertaining to Audit
(Circular CIR/MRD/DP/16/2013 dated 31 May 2013):
• The stock brokers/ trading members that provide the facility of algorithmictrading shall subject their algorithmic trading system to a system audit every sixmonths in order to ensure that the requirements prescribed by SEBI / stockexchanges with regard to algorithmic trading are effectively implemented
• Such system audit of algorithmic trading system shall be undertaken by a systemauditor who possesses any of the following certifications:
– CISA (Certified Information System Auditors) from ISACA;
– DISA (Post Qualification Certification in Information Systems Audit) fromInstitute of Chartered Accountants of India (ICAI);
– CISM (Certified Information Securities Manager) from ISACA;
– CISSP (Certified Information Systems Security Professional) fromInternational Information Systems Security Certification Consortium,commonly known as (ISC)
SEBI guidelines
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Deficiencies or issues identified during the process of system audit of trading algorithm / software shall be reported by the stock broker / trading member to the stock exchange immediately on completion of the system audit.
• In case of serious deficiencies / issues or failure of the stock broker / trading member to take satisfactory corrective action, the stock exchange shall not allow the stock broker/ trading member to use the trading software till deficiencies / issues with the trading software are rectified and a satisfactory system audit report is submitted to the stock exchange. Stock exchanges may also consider imposing suitable penalties in case of failure of the stock broker/ trading member to take satisfactory corrective action to its system within the time-period specified by the stock exchanges.
SEBI guidelines
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• The audit process shall broadly cover the following aspects: – Approved features and system parameters implemented in the
trading system
– Adequacy of input, processing and output controls should be tested
– Adequacy of the application security should be audited
– Event logging and system monitoring
– Robust Password management standards
– Network management and controls
– Backup systems and procedures
– Business continuity and disaster recovery plan
– Proper Documentation for system processes
– Security features such as access control, network firewalls and virus protection should be actively managed
Audits
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• The stock broker, desirous of placing orders generated using algorithms, shall satisfy the stock exchange with regard to the implementation of the following minimum levels of risk controls at its end -– Price check
– Quantity check
– Order Value check
– Cumulative Open Order Value check
– Automated Execution check - an algorithm shall account for all executed, un-executed and unconfirmed orders, placed by it before releasing further order(s)
– Pre-defined parameters for automatic stoppage in the event of a runaway situation / execution in a loop
– All algorithmic orders are tagged with a unique identifier provided by the stock exchange in order to establish audit trail
Audits
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• System compliance requirement for CTCL on annual basis:– Members to submit to the exchange the system audit report every
year (for the year ended Mar 31) after getting the CTCL tradingfacility audited from any qualified auditor
– Report to be submitted through NSE-ENIT by April 30
• System compliance requirement for Algorithmic Trading Facility on half yearly basis:– Members to submit the System Audit Report for the half year
ended March 31 (i.e. for the period from October 01 to March 31)and September 30 (i.e. for the period April 01 to September 30),after getting the Algorithmic trading facility audited from anyqualified auditor
Audit Timelines
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• Algorithm to be executed in Mock Trading environment –logs to be certified by auditor
• Algorithm to be executed in Test market at NSE – logs to be certified by auditor
• Apply to exchange for strategy demonstration date with following documents:
– Strategy document
– Risk Management document
– Network Architecture
– Auditor certificates (both Mock market and Test market)
– Application form (signed by director/senior management)
• Algorithm to be demonstrated with exchange
Strategy Approval Process
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
• After approval from exchange, member applies for trading ids (NEAT ids)
• NEAT ids converted to CTCL ids for particular vendor. Vendor of software intimated about ids and confirmation obtained
• Member uploads location code details (12 digits) along with dealer details under CTCL ID before commencing trading
• Member can trade as either PRO or on behalf of CLIENTS.
– For PRO trading, PRO Undertaking, PRO Location Undertaking must be submitted. PRO enablement should also be done for the particular trading id.
Strategy Approval Process
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Table of Contents
• Changing Trends in Trading
• Changing Trends in Trading Risk Management
• Major Automated Trading Risk Failures
• Regulatory framework in India
• Risk Management Process
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 1: Setting risk management
structure & policies
• Dedicated risk department
• Completely cut off from trading department
• Full autonomy & powers to risk department
• Approval process for each new product and operation introduced
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 2: Identifying sources of risk
• Market Risks
• Credit / Counter-party Risks
• Financing Risks
• Operational Risks (Systems, Mechanical, Criminal)
• Regulatory Risks
• Liquidity Risks (Exogenous & endogenous)
• Natural disasters, political, terrorism, etc
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 3: Evaluating risk components
• Market Risks :
• Sensitivity Analysis• Total Greeks, Dividend, Currency
exposures
• What-if scenario analyses
• VaR analysis
• Stress tests
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 3: Evaluating risk components
• Credit / Counter-party Risks
• Basel II IRB method
(Internal Rating Based Method)
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 3: Evaluating risk components
• Financing Risk
Probability of downgrade * interest rate hike * Size of portfolio
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 3: Evaluating risk components
• Regulatory Risk
Probabilities of new Regulations- Is estimated from News Analysis & Historical Data
Examples…
• Short Selling Ban
• Margin Increase
• Taxes Introduced
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 3: Evaluating risk components
• Operational Risks (Systems, Mechanical, Criminal)
• Robustness of a System
• System Load handling capacity
• Maximum order flow before system detects failure
• Maximum leeway in error while setting parameters
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 3: Evaluating risk components
• Liquidity Risks
• Liquidity adjusted VaR
L-VaR = VaR + Liquidty Adjusted
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 3: Evaluating risk components
• Natural Disaster, Political Risk, Terrorism
• Risk v/s Uncertainty
• News Analysis
Have the potential to wipeout portfolios
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 4: Setting risk limits
• Market Risks :
• Total cash exposure
• Exposure to geography
• Exposure to sector
• Exposure to asset class
• Exposure to assignment / delivery risks (settlement risks)
• Settlement Type (future vscash)
• Exposure to interest rates
• Exposure to exchange rates
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 4: Setting risk limits
• Credit / Counter-party Risks
• Maximum exposure to any counter-party
• Maximum exposure per credit rating level
• Financing Risks
• Maximum amount borrowed per counter-party
• Repayment period for loans
• Rho exposure
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 4: Setting risk limits
• Operational Risks (Systems, Mechanical)
• Max exposure per strategy
• Max orders per second
• Max orders in a day
• Max exposure per application
• PnL fluctuation per application
• Price Range check
• Max order size
• Max Value Traded
• Net Value of portfolio
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 4: Setting risk limits
• Operational Risks (Criminal/Fraud/ Theft, etc)
• Access Control
• Transparency of operations
• Rotation of team members
• Audit (internal & external)
• Centralized PnL reconciliation
• Independent verification of price to pricing models
• Online Infiltration & Virus Protection
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 4: Setting risk limits
• Liquidity Risks
• Maximum exposure per instruments of each liquidity category
• Total exposure per liquidity category
• Natural disasters
• Score-card approach
• Similar to one used By Insurance/ Actuaries
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
© Copyright 2010-2014 QuantInsti Quantitative Learning Private Limited
Risk Management Process• Phase 5: Designing systems with
strict adherence to risk controls
• Centralized system which summarizes net position & exposure
• Asset classes, Interest rates, Exchange rates, Volatility, Dividends, Counter parties
• What if Analysis
• Centralized control of all trading operation
• Pre trade controls
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
Thank You!