Statistical Arbitrage Trading

1
Statistical Arbitrage Trading
Opportunities Outside the Laboratory for Ph.D.s
(IAP 2001)

• Michael Biafore (Ph.D. Course 8, 94)
• Strategist (Long-Term Capital Management Tokyo,
  97-99)
• Stat Arb Trader (NMS Services (Cayman),
  99-present)
• What is Statistical Arbitrage?
• What mathematical tools do we use?
• How is Statistical Arbitrage different from other
  Wall Street jobs for Ph.D.s?
• Some parting words

2
Statistical Arbitrage Trading

• What is arbitrage?
• In what sense is it statistical?
• What sorts of trading do we do?
• What are examples of things we do not do?

3
Tools That We Use

• Classical time series techniques (autoregression,
  vector error correction, cointegration)
• Statistical pattern recognition techniques
• Extreme value theory
• Statistical decision theory
• Game theory

4
(most of)
M.I.T.
Wall Street
trading
You are here
5
Statistical Arbitrage
research

trading
6
Risk Management

• Some interesting cross-disciplinary work, esp.
  with statistical physics (e.g., see Theory of
  Financial Risks From Statistical Physics to Risk
  Management, BOUCHARD and POTTERS (2000),
  Cambridge U. Press)
• Several Wall Street firms are actively hiring
  Ph.D.s in this area

7
Work environment

• Degree of autonomy
• Stress level
• Office politics
• Managers
• Compensation
• Physical work environment

8
Parting words

• On most of Wall Street, research is not the
  central business function (yet)
• Money alone cannot begin to compensate you for
  work that does not excite you
• Look before you leap (some suggestions)

9
References

• Time Series Analysis, J.D. Hamilton (1994)
• Statistical Pattern Recognition, A. Webb (1999)
• A Probabilistic Theory of Pattern Recognition,
  L. Devroye et al., (1996)
• Statistical Learning Theory, V.N. Vapnik (1998)
• Modelling Extremal Events, P. Embrechts et al.
  (1998)