Statistical arbitrage and pairs trading
Nikos S. Thomaidis, PhD1
Dept. of Economics, Aristotle University of Thessaloniki, GREECE Dept. of Financial Engineering & Management University of the Aegean, GREECE email:
[email protected] Dept URL: http://labs.fme.aegean.gr/decision/ Personal web site: http://users.otenet.gr/~ ntho18
in collaboration with Nicholas Kondakis, Kepler Asset Management LLC, NY (http://www.keplerfunds.com)
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Outline
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Outline
What is pairs trading?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Outline
What is pairs trading? Developing a pairs trading system from scratch
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Outline
What is pairs trading? Developing a pairs trading system from scratch Empirical study: statistical arbitrage between Dow Jones Industrial Average (DJIA) stocks
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Outline
What is pairs trading? Developing a pairs trading system from scratch Empirical study: statistical arbitrage between Dow Jones Industrial Average (DJIA) stocks Conclusions
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Outline
What is pairs trading? Developing a pairs trading system from scratch Empirical study: statistical arbitrage between Dow Jones Industrial Average (DJIA) stocks Conclusions Trading risks
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Outline
What is pairs trading? Developing a pairs trading system from scratch Empirical study: statistical arbitrage between Dow Jones Industrial Average (DJIA) stocks Conclusions Trading risks Opportunities
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Outline
What is pairs trading? Developing a pairs trading system from scratch Empirical study: statistical arbitrage between Dow Jones Industrial Average (DJIA) stocks Conclusions Trading risks Opportunities Future challenges
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Pairs trading: the history
2 See [Pole, 2007, Vidyamurthy, 2004] and http://www.pairtradefinder.com/forum/viewtopic.php?f=3&t=14 for interesting facts and information on the history of the topic. Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Pairs trading: the history
Pairs trading has at least twenty-ve years of history on Wall Street.
2 See [Pole, 2007, Vidyamurthy, 2004] and http://www.pairtradefinder.com/forum/viewtopic.php?f=3&t=14 for interesting facts and information on the history of the topic. Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Pairs trading: the history
Pairs trading has at least twenty-ve years of history on Wall Street. Already in the mid 80s, Morgan Stanley - and perhaps other investment companies - have started developing programs that could buy/sell stocks in pair combinations2 .
2 See [Pole, 2007, Vidyamurthy, 2004] and http://www.pairtradefinder.com/forum/viewtopic.php?f=3&t=14 for interesting facts and information on the history of the topic. Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Pairs trading: the history
Pairs trading has at least twenty-ve years of history on Wall Street. Already in the mid 80s, Morgan Stanley - and perhaps other investment companies - have started developing programs that could buy/sell stocks in pair combinations2 . These strategies were strongly quantitative (generating trading rules using statistical/mathematical techniques, executing trades through an automated computer-based system).
2 See [Pole, 2007, Vidyamurthy, 2004] and http://www.pairtradefinder.com/forum/viewtopic.php?f=3&t=14 for interesting facts and information on the history of the topic. Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Pairs trading: the history
Pairs trading has at least twenty-ve years of history on Wall Street. Already in the mid 80s, Morgan Stanley - and perhaps other investment companies - have started developing programs that could buy/sell stocks in pair combinations2 . These strategies were strongly quantitative (generating trading rules using statistical/mathematical techniques, executing trades through an automated computer-based system). Cross-disciplinary work (mathematicians, statisticians, physicists, computer scientists, nance experts).
2 See [Pole, 2007, Vidyamurthy, 2004] and http://www.pairtradefinder.com/forum/viewtopic.php?f=3&t=14 for interesting facts and information on the history of the topic. Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Pairs trading: main idea
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Pairs trading: main idea
Capitalise on market imbalances between two or more securities, in anticipation of making money when the inequality is corrected in the future [Whistler, 2004]
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Pairs trading: main idea
Capitalise on market imbalances between two or more securities, in anticipation of making money when the inequality is corrected in the future [Whistler, 2004] Find two securities that have moved together over the near past
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Pairs trading: main idea
Capitalise on market imbalances between two or more securities, in anticipation of making money when the inequality is corrected in the future [Whistler, 2004] Find two securities that have moved together over the near past When the distance (spread) between their prices goes above a threshold, short the overvalued and buy the undervalued one
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Pairs trading: main idea
Capitalise on market imbalances between two or more securities, in anticipation of making money when the inequality is corrected in the future [Whistler, 2004] Find two securities that have moved together over the near past When the distance (spread) between their prices goes above a threshold, short the overvalued and buy the undervalued one If securities return to the historical norm, prices will converge in the near future and you will end up with a prot
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
So what is pairs trading?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
So what is pairs trading?
a market-neutral trading strategy: generates prot under all market conditions (uptrend, downtrend, or sideways movements)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
So what is pairs trading?
a market-neutral trading strategy: generates prot under all market conditions (uptrend, downtrend, or sideways movements) a statistical arbitrage trading strategy: prot from temporal mispricings of an asset relative to its fundamental value.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
So what is pairs trading?
a market-neutral trading strategy: generates prot under all market conditions (uptrend, downtrend, or sideways movements) a statistical arbitrage trading strategy: prot from temporal mispricings of an asset relative to its fundamental value. a long/short equity strategy: long positions are hedged with short positions in the same or related sectors, so that the investor should be little aected by sector-wide events
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
So what is pairs trading?
a market-neutral trading strategy: generates prot under all market conditions (uptrend, downtrend, or sideways movements) a statistical arbitrage trading strategy: prot from temporal mispricings of an asset relative to its fundamental value. a long/short equity strategy: long positions are hedged with short positions in the same or related sectors, so that the investor should be little aected by sector-wide events relative-value trading,
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
So what is pairs trading?
a market-neutral trading strategy: generates prot under all market conditions (uptrend, downtrend, or sideways movements) a statistical arbitrage trading strategy: prot from temporal mispricings of an asset relative to its fundamental value. a long/short equity strategy: long positions are hedged with short positions in the same or related sectors, so that the investor should be little aected by sector-wide events relative-value trading, convergence trading,
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
So what is pairs trading?
a market-neutral trading strategy: generates prot under all market conditions (uptrend, downtrend, or sideways movements) a statistical arbitrage trading strategy: prot from temporal mispricings of an asset relative to its fundamental value. a long/short equity strategy: long positions are hedged with short positions in the same or related sectors, so that the investor should be little aected by sector-wide events relative-value trading, convergence trading, and so on...
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
So what is pairs trading?
a market-neutral trading strategy: generates prot under all market conditions (uptrend, downtrend, or sideways movements) a statistical arbitrage trading strategy: prot from temporal mispricings of an asset relative to its fundamental value. a long/short equity strategy: long positions are hedged with short positions in the same or related sectors, so that the investor should be little aected by sector-wide events relative-value trading, convergence trading, and so on... Pairs trading group trading
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Why pairs work: the drunk and his dog
A humorous metaphor adapted from [Murray, 1994] to the context of pairs trading.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Why pairs work: the drunk and his dog
A humorous metaphor adapted from [Murray, 1994] to the context of pairs trading. A drunk customer sets out from the pub (Gin Palace) and starts wandering in the streets (random walk, unit-root, integrated stochastic process)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Why pairs work: the drunk and his dog
A humorous metaphor adapted from [Murray, 1994] to the context of pairs trading. A drunk customer sets out from the pub (Gin Palace) and starts wandering in the streets (random walk, unit-root, integrated stochastic process) The accompanying dog thinks: I cant let him get too far o; after all, my role is to protect him!
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Why pairs work: the drunk and his dog
A humorous metaphor adapted from [Murray, 1994] to the context of pairs trading. A drunk customer sets out from the pub (Gin Palace) and starts wandering in the streets (random walk, unit-root, integrated stochastic process) The accompanying dog thinks: I cant let him get too far o; after all, my role is to protect him! So, the dog assesses how far the drunk is and moves accordingly to close the gap
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
The drunk and his dog: the story continues
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
The drunk and his dog: the story continues
Rory and Gary, two regular customers, look outside the pubs window and bet on the drunks and the dog s position
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
The drunk and his dog: the story continues
Rory and Gary, two regular customers, look outside the pubs window and bet on the drunks and the dog s position They observe the drunk and the dog individually but their course looks no dierent than a random walk (growing variance in location, lack of predictability)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
The drunk and his dog: the story continues
Rory and Gary, two regular customers, look outside the pubs window and bet on the drunks and the dog s position They observe the drunk and the dog individually but their course looks no dierent than a random walk (growing variance in location, lack of predictability) Suddenly, Gary throws the idea: Well, its all a matter of nding the drunk, the dog must not be far away
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
The drunk and his dog: the story continues
Rory and Gary, two regular customers, look outside the pubs window and bet on the drunks and the dog s position They observe the drunk and the dog individually but their course looks no dierent than a random walk (growing variance in location, lack of predictability) Suddenly, Gary throws the idea: Well, its all a matter of nding the drunk, the dog must not be far away He is right because the gap between the two fellows should occasionally open and close but never being out of control (co-integration)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
The drunk and his dog: the story continues
Rory and Gary, two regular customers, look outside the pubs window and bet on the drunks and the dog s position They observe the drunk and the dog individually but their course looks no dierent than a random walk (growing variance in location, lack of predictability) Suddenly, Gary throws the idea: Well, its all a matter of nding the drunk, the dog must not be far away He is right because the gap between the two fellows should occasionally open and close but never being out of control (co-integration) Rory and Gary eventually agree to play the following game:
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
The drunk and his dog: the story continues
Rory and Gary, two regular customers, look outside the pubs window and bet on the drunks and the dog s position They observe the drunk and the dog individually but their course looks no dierent than a random walk (growing variance in location, lack of predictability) Suddenly, Gary throws the idea: Well, its all a matter of nding the drunk, the dog must not be far away He is right because the gap between the two fellows should occasionally open and close but never being out of control (co-integration) Rory and Gary eventually agree to play the following game: Why not betting on their relative distance rather than their absolute positions?
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
An actual traded pair
1.05 1 0.95 0.9 0.85 0.8 0.75 0.7 0.65 Mar93 Dec95 Sep98 May01 Feb04 Nov06 Aug09 May12 GT HPQ
Figure 1: Normalised price paths of Goodyear (GT) and Hewlett Packard (HPQ).
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Why pairs trading is successful?
A behavioural-nance explanation: New information is rapidly impounded in stock prices through investment activity (market eciency) Stock price movements reect all publicly available information (future earnings prospects, corporate news, political events) Two securities that are close substitutes for each other respond similarly to incoming news Overreaction and herding behaviour of uninformed and noisy investors often drives prices apart But, any deviation is temporary and rational traders are expected to close the gaps in the long run
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Basic steps in developing a pairs trading system
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Basic steps in developing a pairs trading system
Group formation
Pick closely-related stocks and detect stable relative price relationships
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Basic steps in developing a pairs trading system
Group formation
Pick closely-related stocks and detect stable relative price relationships
Group trading
Determine the direction of the relationship (divergence, re-convergence) Find suitable trade-open and trade-close points
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Basic steps in developing a pairs trading system
Group formation
Pick closely-related stocks and detect stable relative price relationships
Group trading
Determine the direction of the relationship (divergence, re-convergence) Find suitable trade-open and trade-close points
Risk management
Minimise divergence risk (the gap between stocks further widens) Fine-tune parameters with respect to a trading performance criterion (maximise expected return, maximise a reward-risk ratio, etc)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Group formation strategy
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Maximum price correlation (MPC)
1 2
Choose a charting time-frame Compute the correlation of historical price series, e.g. Correlation coecient 0.91 0.87 0.81 0.76 0.26 0.17
Pair Pair Pair Pair ... Pair Pair
3
1 2 3 4 19 20
Stock Stock Stock Stock ... Stock Stock
1 1 2 8 13 26
Stock Stock Stock Stock ... Stock Stock
3 5 4 10 26 27
Pick the top 20% of pairs (i.e 4 pairs) with the highest historical correlation Formed groups: {1, 3, 5}, {2, 4}, {8, 10}
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Minimum normalised price distance (MNPD)
Popular in literature [Gatev et al., 2006, Andrade et al., 2005] Construct a cumulative total return index for each stock over the formation period
t
crt,i
(1 + r,i ), t = 1, 2, ..., T
=1
where cr0,i = 1 and rt,i is the t-periods return on stock i . Introduce a distance measure: e.g. Euclidean distance
T
d(i , j) |cr
,i
cr ,j |
t=1
(crt,i crt,j )2
Rank stock pairs based on increasing values of d - pick the top a% of the list for group formation
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Identify stationary relationships (1/5)
Applying techniques from co-integration analysis [Engle and Granger, 1987, Burgess, 2000, Vidyamurthy, 2004] Assume that a group of stocks with price vector Pt = (Pt1 , Pt2 , . . . , PtN ) satisfy the relationship Pt1 = c + 2 Pt2 + + n PtN + Zt where Zt is the mispricing index (captures temporal deviations from equilibrium) The coecients of the relationship can be estimated using Ordinary Least Squares (OLS)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Identify stationary relationships (2/5)
Construct a portfolio as follows: Stocks Positions 1 +1 2 -2 3 - 3 N - N
where i is the OLS estimate of i and + (-) indicates a long (short) position The portfolio value Zt Pt , where (1, 2 , 3 , . . . , N ) is by construction mean-reverting (uctuates around c , the OLS estimate of c)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Identify relationships with OLS (3/5)
15
Stock 1 Stock 2
Prices
10
5 0
50
100 150 Group formation sample
200
250
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Identify relationships with OLS (4/5)
14.5 14 13.5 Stock 2 13 12.5 12 11.5 11 5.5 6 6.5 7 Stock 1 7.5 negative mispricing
Equilibrium relationship: P2 = 14.843 0.257 P1
actual price pairs equilibrium relationship positive mispricing
8.5
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Identify relationships with OLS (5/5)
16.5 16 Relative mispricing 15.5 15 14.5 14 13.5 13 0
Zt=P2 + 0.257 P1 Stock 2 underpriced relative to Stock 1 Stock 2 overpriced relative to Stock 1
50
100 150 Group formation sample
200
250
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Conditions for meaningful capital allocations
The average capital invested on each stock (average price number of shares) must be below 80% and above 5% The ratio between the maximum and the minimum number of shares held from each asset should not exceed 10. etc These place restrictions on the beta coecients (stock holdings) restricted OLS estimation
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Group trading
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading strategya
a
See also [Thomaidis et al., 2006, Thomaidis and Kondakis, 2012]
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading strategya
a
See also [Thomaidis et al., 2006, Thomaidis and Kondakis, 2012]
Open a position in a group, over the trading period, when the mispricing index diverges by a certain threshold
Buy the portfolio, if Zt < ZtL, tH, Sell the portfolio, if Zt > Z
where ZtL, , ZtH, is a 100 (1 2)% condence envelope on the value of the mispricing.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading strategya
a
See also [Thomaidis et al., 2006, Thomaidis and Kondakis, 2012]
Open a position in a group, over the trading period, when the mispricing index diverges by a certain threshold
Buy the portfolio, if Zt < ZtL, tH, Sell the portfolio, if Zt > Z
where ZtL, , ZtH, is a 100 (1 2)% condence envelope on the value of the mispricing. Unwind the position after h periods of time
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading strategya
a
See also [Thomaidis et al., 2006, Thomaidis and Kondakis, 2012]
Open a position in a group, over the trading period, when the mispricing index diverges by a certain threshold
Buy the portfolio, if Zt < ZtL, tH, Sell the portfolio, if Zt > Z
where ZtL, , ZtH, is a 100 (1 2)% condence envelope on the value of the mispricing. Unwind the position after h periods of time unless
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading strategya
a
See also [Thomaidis et al., 2006, Thomaidis and Kondakis, 2012]
Open a position in a group, over the trading period, when the mispricing index diverges by a certain threshold
Buy the portfolio, if Zt < ZtL, tH, Sell the portfolio, if Zt > Z
where ZtL, , ZtH, is a 100 (1 2)% condence envelope on the value of the mispricing. Unwind the position after h periods of time unless the mispricing index continues to diverge (does not cross up the lower bound or cross down the upper bound)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading strategya
a
See also [Thomaidis et al., 2006, Thomaidis and Kondakis, 2012]
Open a position in a group, over the trading period, when the mispricing index diverges by a certain threshold
Buy the portfolio, if Zt < ZtL, tH, Sell the portfolio, if Zt > Z
where ZtL, , ZtH, is a 100 (1 2)% condence envelope on the value of the mispricing. Unwind the position after h periods of time unless the mispricing index continues to diverge (does not cross up the lower bound or cross down the upper bound) Close the position earlier and open a new position if the synthetic re-converges and crosses the opposite bound
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading strategya
a
See also [Thomaidis et al., 2006, Thomaidis and Kondakis, 2012]
Open a position in a group, over the trading period, when the mispricing index diverges by a certain threshold
Buy the portfolio, if Zt < ZtL, tH, Sell the portfolio, if Zt > Z
where ZtL, , ZtH, is a 100 (1 2)% condence envelope on the value of the mispricing. Unwind the position after h periods of time unless the mispricing index continues to diverge (does not cross up the lower bound or cross down the upper bound) Close the position earlier and open a new position if the synthetic re-converges and crosses the opposite bound ZtL, , ZtH, is of the form c z Z , where c , Z are the sample mean and standard deviation of the synthetic value over the formation period and z is the critical value from a N(0, 1) distribution.
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Example: trading a group of 2 stocks (1/2)
GOODYEAR (GT) vs HEWLETT PACKARD (HPQ) 42 40 Price ($) 38 36 34 0 GT HPQ
20
18
16
14
20
40
60 Trading period
80
100
120
12
24 23 Mispricing 22 21 20 19 18 0
Zt=PGT 1.06 PHPQ
20 40
Mispricing index
Confidence bounds 80
Long positions 100
Short positions 120
60 Trading period
Figure 3:
Mispricing index: Zt = PGT 1.06PHPQ , Trading parameters: HOP = 1day , L = 10%, H = 5% .
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Price ($)
Example: trading a group of 2 stocks (2/2)
24 23 Mispricing 22 21 20 19 18 0 20 40 Mispricing index 60 Trading period Confidence bounds 80 Long positions 100 Short positions 120
8 Cumulative return (%) 6 4 2 0 2 0 20 40 60 Trading period 80 100 120
Figure 4: HOP=1 day, L = 10%, H = 5% .
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Example: trading a group of 4 stocks (1/2)
1.6 Normalised prices 1.4 1.2 1 0.8 0 0 1 2 3 4 0 AA AXP CAT IBM Long positions Short positions
50
100 Trading period
150
200
250
Mispricing
Mispricing index 50
Confidence bounds 100
Long positions Trading period
Short positions 150 200 250
Figure 5: HOP=1 day, L = 20%, H = 20% .
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Example: trading a group of 4 stocks(2/2)
0 1 Mispricing 2 3 Mispricing index 4 0 30 Cumulative return (%) 20 10 0 50 100 Trading period Confidence bounds Long positions 150 Short positions 200 250
10 0
50
100 Trading period
150
200
250
Figure 6: HOP=1 day, L = 20%, H = 20% .
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
System performance measurement
Are there truly successful rules that deliver consistent return or risk-adjusted return? Performance indicators (mean, std, downside std, information ratio (IR), downside IR) How does performance vary with dierent market conditions? Can high returns be explained by specic exposure to industry and other systematic risk factors? Are we capturing other patterns of stock movements (price reversals)? How skillful is our system in terms of picking the right pairs/nding price equilibriums? How able is our system to early detect price divergence and predict re-convergence points? Do our strategies require too much trading? Do our strategies maintain their performance ranking over time? Do the best remain the best and the worst remain the worst?
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Experimental set-up
Daily prices of 30 stock members of Dow Jones Industrial Average (DJIA) index (with dividends reinvested) Sample period: 3 Jan 1994 to 24 Feb 2010 Group formation:
Window length (WL) {125, 250} days Screen out DJIA stocks with one or more days without a trade (identify relatively liquid stocks and facilitate pairs formation) Choose matching stocks based on MNPD and MPC criteria (form groups from the 5%, 20% or 50% highest-ranking pairs of the list)
Trading strategy
Trading period: subsequent {50, 125, 150} days Hold-out period (HOP): {1, 5, 10, 25} days L , H {1, 5, 10, 20, 40}%
A total of 3, 600 parametrisations
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Best trading strategies
Design parameters3 WL TP GFC HOP L (%) H (%)
Best strategy (Mean return) Sample: 1994-2010 125 150 MPC - 5% 25 40 1
Best strategy (IR) 125 150 MPC - 20% 25 10 1
3 WL: Length of moving window, TP: Trading period, GFC: Group formation criterion, HOP: Position holdout period. Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Performance of best trading strategies
Trading measures
Mean(%) Stdev(%) DStdev(%) IR DIR
Best strategy Best strategy (IR) Buy & hold portfolio (Mean return) Sample: 1994-2010 (784 observations) 11.65 7.78 5.92 26.44 9.94 22.00 23.75 6.48 16.54 0.44 0.78 0.27 0.49 1.20 0.36
Table 1: Average weekly performance (annualised measures).
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Portfolios of good strategies
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Portfolios of good strategies
No investor would risk putting all his money in a single strategy
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Portfolios of good strategies
No investor would risk putting all his money in a single strategy Mixing-up dierent parameter combinations
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Portfolios of good strategies
No investor would risk putting all his money in a single strategy Mixing-up dierent parameter combinations Bundles of trading strategies:
Distribute your capital evenly between the top-a % of the parameterisations
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Performance of mixtures - Mean return
Strategies Mean(%) Stdev(%) DStdev(%) IR DIR
Percentage of trading strategies 100 90 65 35 10 1.98 2.46 3.42 4.64 6.48 3.65 3.63 3.69 4.00 6.19 2.26 2.19 2.10 2.16 2.71 0.54 0.68 0.93 1.16 1.05 0.88 1.12 1.63 2.15 2.39
Best Buy & strategy hold 11.65 5.92 26.44 22.00 23.75 16.54 0.44 0.27 0.49 0.36
Table 2: Average weekly performance on the full sample period (annualised measures).
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Performance of mixtures - Information ratio (1/2)
Strategies Mean(%) Stdev(%) DStdev(%) IR DIR
Percentage of trading strategies 100 90 65 35 10 1.98 2.46 3.42 4.43 5.93 3.65 3.63 3.66 3.65 4.31 2.26 2.19 2.09 2.18 2.55 0.54 0.68 0.93 1.22 1.37 0.88 1.12 1.64 2.03 2.32
Best Buy & strategy hold 7.78 5.92 9.94 22.00 6.48 16.54 0.78 0.27 1.20 0.36
Table 3: Average weekly performance on the full sample period (annualised measures).
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Performance of mixtures - Information ratio (2/2)
IRmaximising strategies 350 300 250 Cumulative return (%) 200 150 100 50 0 50 top100 top90 top65 top35 top10 best strategy buy & hold
Dec95
Sep98
May01
Feb04
Nov06
Aug09
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Systematic risk exposure
200 Market SMB HML Top10%(IR)
150
cumulative return (%)
100
50
50 Mar93
Dec95
Sep98
May01
Feb04
Nov06
Aug09
May12
Figure 7: Historical performance of the top-10% portfolio (IR) and systematic factors of risk.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Systematic risk exposure
Strategies Alpha MKT SMB HML MOM LTR STR Consumer Durables Manufacturing HiTec Health Other 100 0.00 (0.12) -0.15 (0.01) 0.00 (0.83) -0.00 (0.98) -0.04 (0.00) -0.02 (0.36) 0.04 (0.00) 0.04 (0.00) 0.00 (0.95) 0.03 (0.03) 0.02 (0.04) 0.01 (0.39) Percentage 90 0.00 (0.04) -0.15 (0.01) -0.00 (0.89) -0.00 (0.82) -0.04 (0.00) -0.01 (0.52) 0.03 (0.00) 0.04 (0.00) 0.00 (0.96) 0.03 (0.03) 0.02 (0.04) 0.01 (0.36) of trading 65 0.00 (0.00) -0.13 (0.01) -0.01 (0.53) -0.01 (0.50) -0.03 (0.00) -0.00 (0.87) 0.03 (0.00) 0.04 (0.00) 0.00 (0.88) 0.03 (0.05) 0.02 (0.06) 0.01 (0.44) strategies 35 0.00 (0.00) -0.11 (0.04) -0.00 (0.82) -0.01 (0.49) -0.03 (0.00) 0.00 (0.81) 0.03 (0.00) 0.03 (0.01) -0.01 (0.69) 0.03 (0.08) 0.02 (0.10) 0.01 (0.55) 10 0.00 (0.00) -0.15 (0.03) 0.00 (0.89) -0.03 (0.10) -0.03 (0.00) -0.00 (0.89) 0.01 (0.09) 0.04 (0.00) -0.01 (0.61) 0.03 (0.14) 0.02 (0.14) 0.04 (0.10) Best strategy 0.00 (0.00) -0.46 (0.00) 0.02 (0.40) -0.11 (0.01) -0.04 (0.09) -0.07 (0.06) -0.02 (0.30) 0.15 (0.00) 0.03 (0.56) 0.09 (0.04) 0.03 (0.20) 0.14 (0.00)
Table 4: OLS estimates of the regression equation.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Pairs trading is a cost-sensitive strategy
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Pairs trading is a cost-sensitive strategy It involves
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Pairs trading is a cost-sensitive strategy It involves
Frequent re-balancing of trading positions
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Pairs trading is a cost-sensitive strategy It involves
Frequent re-balancing of trading positions Multiple openings and closings of trades
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Pairs trading is a cost-sensitive strategy It involves
Frequent re-balancing of trading positions Multiple openings and closings of trades Short-selling
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Pairs trading is a cost-sensitive strategy It involves
Frequent re-balancing of trading positions Multiple openings and closings of trades Short-selling
Transaction costs, margin requirements, etc
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Pairs trading is a cost-sensitive strategy It involves
Frequent re-balancing of trading positions Multiple openings and closings of trades Short-selling
Transaction costs, margin requirements, etc How the strategies are expected to perform in a more realistic market environment?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading costs
Pairs trading is a cost-sensitive strategy It involves
Frequent re-balancing of trading positions Multiple openings and closings of trades Short-selling
Transaction costs, margin requirements, etc How the strategies are expected to perform in a more realistic market environment? Can generated prots oset trading costs?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Descriptive statistics (1/2)
Top-10% (IR) portfolio of strategies
Sample period Total days in sample: Total trading days in sample: Total number of traded stocks: Group formation Total number of formed groups: Average size of groups: Group trading Total number of group openings during study: Number of groups that never open: Average number of active groups per trading day: Fraction of trading time groups are open: Average number of times a group is opened over the trading period: Average duration of positions (days): Average duration of long positions (days): Average duration of short positions (days): 4065 3865.7 35 71.43 4.51 (1.59) 195.76 4.19 1.17 (0.45) 0.88 3.32 (2.24) 27.59 (28.94) 24.50 (30.66) 30.15 (27.05)
Notes: (1) Averages over all parametrisations, (2) Standard deviation in parentheses.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Descriptive statistics (2/2)
Top-10% (IR) portfolio of strategies
Divergence risk Percentage of groups that never open: Percentage of groups opened once but never converging in the trading period: Percentage of groups that have multiple round-trip trades and a nal divergent trade: Percentage of groups with no nal divergent trade:
Note: Averages over all 360 parametrisations.
3.21
26.31
57.13 13.34
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
The impact of transaction costs (1/2)
Transaction cost4 Strategies Mean(%) Stdev(%) DStdev(%) IR DIR 0 bps Best at Zero Cost 5.93 4.31 2.55 1.37 2.32 Best 5.93 4.31 2.55 1.37 2.32 10 bps Best at Zero Cost 5.38 4.34 2.54 1.24 2.12 Best 5.78 4.30 2.54 1.35 2.27
Transaction cost Strategies Mean(%) Stdev(%) DStdev(%) IR DIR
50 bps Best at Zero Cost 4.93 4.33 2.55 1.14 1.93
Buy & hold Best 5.34 4.28 2.55 1.25 2.10 5.92 22.00 16.54 0.27 0.36
Table 5: Top-10% (IR) portfolio.
4
Fixed cost per unit of trading volume.
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
The impact of transaction costs (2/2)
160 140 120 cumulative return (%) 100 80 60 40 20 0 20 Mar93 Dec95 Sep98 May01 Feb04 Nov06 Aug09 May12 0 bps 10 bps 50 bps
Figure 8: Historical performance of the top-10% (IR) portfolio assuming dierent levels of transaction costs.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (1/2)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (1/2)
Statistical arbitrage strategies are highly parametrised
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (1/2)
Statistical arbitrage strategies are highly parametrised If we experiment with enough parameter settings, some of them are likely to beat the benchmark under any performance measures, by chance alone
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (1/2)
Statistical arbitrage strategies are highly parametrised If we experiment with enough parameter settings, some of them are likely to beat the benchmark under any performance measures, by chance alone For example, strategies that went short in DJIA stocks during the period Apr 2008 - Oct 2008, would possibly outperform the market portfolio in a longer sample
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (1/2)
Statistical arbitrage strategies are highly parametrised If we experiment with enough parameter settings, some of them are likely to beat the benchmark under any performance measures, by chance alone For example, strategies that went short in DJIA stocks during the period Apr 2008 - Oct 2008, would possibly outperform the market portfolio in a longer sample Simply because of the special characteristics of this single period
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (1/2)
Statistical arbitrage strategies are highly parametrised If we experiment with enough parameter settings, some of them are likely to beat the benchmark under any performance measures, by chance alone For example, strategies that went short in DJIA stocks during the period Apr 2008 - Oct 2008, would possibly outperform the market portfolio in a longer sample Simply because of the special characteristics of this single period Data snooping(dredging or shing): The practice of hand-tailoring the trading strategy to the data under consideration [Sullivan et al., 1999, White, 2000]
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (2/2)
Is the seemingly outstanding performance
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (2/2)
Is the seemingly outstanding performance due to genuine superiority?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (2/2)
Is the seemingly outstanding performance due to genuine superiority? or...
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping (2/2)
Is the seemingly outstanding performance due to genuine superiority? or... due to luck?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Data snooping quotations
Given enough computer time, we are sure that we can nd a mechanical trading rule which works on a table of random numbers, provided of course that we are allowed to test the rule on the same table of numbers which we used to discover the rule. [Jensen and Bennington, 1970] Even when no exploitable [trading] model exists, looking long enough and hard enough at a given set of data will often reveal one or more [trading strategies] that look good, but are in fact useless. [White, 2000] If you have 20,000 traders in the market, sure enough youll have someone whos been up every day for the past few years and will show you a beautiful P&L. If you put enough monkeys on typewriters, one of the monkeys will write the Iliad in ancient Greek. But would you bet any money that hes going to write the Odyssey next? [Taleb, 1997]5
5 Random Walk: Taleb on Mistakes that Market Traders can make, http://equity.blogspot.com/2008/11/taleb-on-mistakes-that-market-traders.html Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
How to eliminate data snooping biases?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to eliminate data snooping biases?
Using an estimation and validation (test) data set
Helps observing model performance beyond the training sample Sensitive with respect to the particular choice of sample periods (training and testing) Sensitive to market conditions
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to eliminate data snooping biases?
Using an estimation and validation (test) data set
Helps observing model performance beyond the training sample Sensitive with respect to the particular choice of sample periods (training and testing) Sensitive to market conditions
Using multiple estimation/validation periods
Reported performance is less prone to data-snooping biases Problems arise if these periods are consecutive The choice of periods can introduce further bias
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to eliminate data snooping biases?
Using an estimation and validation (test) data set
Helps observing model performance beyond the training sample Sensitive with respect to the particular choice of sample periods (training and testing) Sensitive to market conditions
Using multiple estimation/validation periods
Reported performance is less prone to data-snooping biases Problems arise if these periods are consecutive The choice of periods can introduce further bias
Statistical techniques
Little sensitivity to market conditions Helps exploring new market scenarios (beyond those present in the dataset)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How would you choose your sample periods?
Buy & hold strategy
350 300 Cumulative return (%) 250 200
1998 2000 1999 2002 2003 2008 1997 2010 1996 1995 1994 2001 2004 2005 2006 2007
150 100 50 0
2009
50 Mar93
Dec95
Sep98
May01
Feb04
Nov06
Aug09
May12
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading performance comparisons (1/2)
Splitting the data set into estimation and validation periods
Sample 1 Estimation period Validation period Number of observations (validation set) 1994- 96 1997- 99 Sample 2 1997-99 2000-02 Sample 3 2000-02 2003-05 Sample 4 2003-06 2006-10
756 days
756 days
756 days
1041 days
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Trading performance comparisons (2/2)
Validation period 1 10 150 25 20 cumulative return (%) cumulative return (%) 100 15 10 5 0 0 Apr00 5 Oct99
Validation set 2
IR=0.27
cumulative return (%)
Top10% (IR) Buy & hold
IR=1.21
30 20 10 0 10 20 30 Jan03 cumulative return (%) cumulative return (%)
50 0
IR=1.28
Top10% (IR) IR=0.17 Buy & hold
Jun96
Jan97
Jul97
Feb98
Sep98
Mar99
Oct99
Apr00
Nov00
May01 Validation set 4
Dec01
Jul02
Validation set 3 5 50 20 IR= 0.55
50 Top10% (IR) Buy & hold IR= 0.78
IR=0.52
Top10% (IR) Buy & hold cumulative return (%) cumulative return (%)
cumulative return (%)
10
IR=0.05
50
5 Jul02
Jan03
Aug03
Feb04
Sep04
Mar05
Oct05
50 May06
10 Oct05
May06
Nov06
Jun07
Dec07
Jul08
Jan09
Aug09
100 Mar10
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Statistical techniques
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Statistical techniques
Random portfolios [Burns, 2006] How skillful is our strategy in terms of picking the right stocks
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Statistical techniques
Random portfolios [Burns, 2006] How skillful is our strategy in terms of picking the right stocks at the right combination?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Statistical techniques
Random portfolios [Burns, 2006] How skillful is our strategy in terms of picking the right stocks at the right combination? Monkey trading Is our trading system superior to a monkey, which opens and closes trading positions at random points?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Statistical techniques
Random portfolios [Burns, 2006] How skillful is our strategy in terms of picking the right stocks at the right combination? Monkey trading Is our trading system superior to a monkey, which opens and closes trading positions at random points? Other more sophisticated approaches:
Reality Check [White, 2000] Test of Superior Predictive Performance [Hansen, 2005] False discovery rate [Bajgrowiczy and Scailletz, 2009]
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Skillful vs lucky stock picking
1 Probability of superior group formation skills 0.8 0.6 0.4 0.2 0 0.3 90th percentile 0.2 Monthly return 0.1 0 0.1 0.2 Median 10th percentile Dec95 Sep98 May01 Feb04 Nov06 Aug09 Top10% (IR) strategy
months of consecutive out performarnce
months of consecutive under performarnce
Dec95
Sep98
May01
Feb04
Nov06
Aug09
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Group-selection skills: interesting statistics
Based on the probability of superiority Percentage of skilled months: 63.10% Percentage of unskilled months: 36.90% Average number of consecutive skillful-picking months: 2.51 Average number of consecutive unskilled-picking months: 1.47
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Do stock-picking benets accumulate over time?
300 Top10% (IR) strategy 250 Probability of outperformance: 98.20%
200
Cumulative return (%)
150
100
50
50
100 Mar93
Dec95
Sep98
May01
Feb04
Nov06
Aug09
May12
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Is my trading system as smart as a monkey?
6
6 This particular monkey-trader was recruited from http://www.free-extras.com/images/monkey_thinking-236.htm . Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
Skillful vs lucky trading
1 Probability of superior group formation skills 0.8 0.6 0.4 0.2 0 0.2
months of consecutive out performarnce
months of consecutive under performarnce
Dec95 90th percentile
Sep98
May01
Feb04
Nov06 Top10% (IR) strategy
Aug09
0.1 Monthly return
0.1 Median 0.2 10th percentile Dec95 Sep98 May01 Feb04 Nov06 Aug09
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Group-trading skills: interesting statistics
Percentage of skilled months: 66.31% Percentage of unskilled months: 32.62% Average number of consecutive skilled months: 2.88 Average number of consecutive unskilled months: 1.49
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Beating the monkey in terms of cumulative return
300 Top10% (IR) strategy 250 Probability of outperformance: 98.20%
200
Cumulative return (%)
150
100
50
50
100 Mar93
Dec95
Sep98
May01
Feb04
Nov06
Aug09
May12
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure Trade at higher frequencies (microstructure information)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure Trade at higher frequencies (microstructure information) Select stocks with similar response patterns to market disturbances
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure Trade at higher frequencies (microstructure information) Select stocks with similar response patterns to market disturbances Event-response analysis [Pole, 2007]
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure Trade at higher frequencies (microstructure information) Select stocks with similar response patterns to market disturbances Event-response analysis [Pole, 2007] Incorporate any type of prior expert knowledge
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure Trade at higher frequencies (microstructure information) Select stocks with similar response patterns to market disturbances Event-response analysis [Pole, 2007] Incorporate any type of prior expert knowledge Achieve the right balance between automation and human intervention
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure Trade at higher frequencies (microstructure information) Select stocks with similar response patterns to market disturbances Event-response analysis [Pole, 2007] Incorporate any type of prior expert knowledge Achieve the right balance between automation and human intervention Is it possible to select the best-performing rules ex ante?
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure Trade at higher frequencies (microstructure information) Select stocks with similar response patterns to market disturbances Event-response analysis [Pole, 2007] Incorporate any type of prior expert knowledge Achieve the right balance between automation and human intervention Is it possible to select the best-performing rules ex ante?
Historical (in-sample) performance
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
How to improve your pairs trading system
Use rm fundamentals to select stocks with similar factor risk exposure Trade at higher frequencies (microstructure information) Select stocks with similar response patterns to market disturbances Event-response analysis [Pole, 2007] Incorporate any type of prior expert knowledge Achieve the right balance between automation and human intervention Is it possible to select the best-performing rules ex ante?
Historical (in-sample) performance Economic conditions (picking those rules that perform better with a particular state of the business and market cycle)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Event-response analysis
1.35 1.3 1.25 1.2 1.15 local minima 1.1 1.05 1 0.95 0 local maxima
Normalised price
20
40
60 80 Group formation period (days)
100
120
140
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Epilogue
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Epilogue
Pairs trading is a statistical arbitrate trading strategy
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Epilogue
Pairs trading is a statistical arbitrate trading strategy Performs better under limiting conditions
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Epilogue
Pairs trading is a statistical arbitrate trading strategy Performs better under limiting conditions
innitely-dimensional asset universe
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Epilogue
Pairs trading is a statistical arbitrate trading strategy Performs better under limiting conditions
innitely-dimensional asset universe innite amount of trading time, etc
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Epilogue
Pairs trading is a statistical arbitrate trading strategy Performs better under limiting conditions
innitely-dimensional asset universe innite amount of trading time, etc
Computational challenges (processing huge amounts of information, asset selection, ne-tuning, model estimation)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Epilogue
Pairs trading is a statistical arbitrate trading strategy Performs better under limiting conditions
innitely-dimensional asset universe innite amount of trading time, etc
Computational challenges (processing huge amounts of information, asset selection, ne-tuning, model estimation) Implementation challenges (high portfolio turnover, trading costs, execution risk)
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
Epilogue
Pairs trading is a statistical arbitrate trading strategy Performs better under limiting conditions
innitely-dimensional asset universe innite amount of trading time, etc
Computational challenges (processing huge amounts of information, asset selection, ne-tuning, model estimation) Implementation challenges (high portfolio turnover, trading costs, execution risk) If benets exceed costs your system is a hit!
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
References I
Andrade, S., Vadim, P., and Seasholes, M. (2005). Understanding the protability of pairs trading. working paper. Bajgrowiczy, P. and Scailletz, O. (2009). Technical trading revisited: False discoveries, persistence tests, and transaction costs. working paper. Burgess, N. (2000). Statistical arbitrage models of the FTSE 100. In Abu-Mostafa, Y., LeBaron, B., Lo, A. W., and Weigend, A. S., editors, Computational Finance 1999, pages 297312. The MIT Press.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
References II
Burns, P. (2006). Random portfolios for evaluating trading strategies. working paper. Engle, R. F. and Granger, C. W. J. (1987). Co-integration and error correction: Representation, estimation, and testing. Econometrica, 55:251276. Gatev, E., Goetzmann, W., and Rouwenhorst, K. (2006). Pairs trading: performance of a relative-value arbitrage rule. The Review of Financial Studies, 19(3):797827. Hansen, P. (2005). A test for superior predictive ability. Journal of Business & Economic Statistics, 23(5):365380.
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
References III
Jensen, M. and Bennington, G. (1970). Random walks and technical theories: some additional evidence. The Journal of Finance, 25:469 482. Murray, M. (1994). A drunk and her dog: An illustration of cointegration and error correction. The American Statistician, 48(1):3739. Pole, A. (2007). Statistical arbitrage: algorithmic trading insights and techniques. John Wiley and Sons, Inc.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading
References IV
Sullivan, R., Timmermann, A., and White, H. (1999). Data-snooping, technical trading model performance and the bootstrap. The Journal of Finance, 54:16471691. Thomaidis, N. S. and Kondakis, N. (2012). Detecting statistical arbitrage opportunities using a combined neural network - GARCH model. Working paper available from SSRN. Thomaidis, N. S., Kondakis, N., and Dounias, G. (2006). An intelligent statistical arbitrage trading system. Lecture Notes in Articial Intelligence, 3955:596599. Vidyamurthy, G. (2004). Pairs trading: quantitative methods and analysis. John Wiley and Sons, Inc.
Nikos S. Thomaidis, PhD Statistical arbitrage and pairs trading
References V
Whistler, M. (2004). Trading pairs: capturing prots and hedging risk with statistical arbitrage strategies. John Wiley and Sons, Inc. White, H. (2000). A reality check for data snooping. Econometrica, 68(5):10971126.
Nikos S. Thomaidis, PhD
Statistical arbitrage and pairs trading