Global Asset Allocation

1
Global Asset Allocation

• Use of momentum in trading across industry sectors

Yuri Krapivin Yuk Ping Ng Pierre
Oustinow Jonathan Steinmetz Terence Tong
2
Agenda

• Objective
• Implementation Methodology
• Results
• Improvements

3
Objective

• Develop and test a momentum based strategy to
  trade across Industry Sectors

4
Implementation

• Data
• 130 SP500 sectors
• Monthly returns from Jan 83 to Mar 00
• Principle
• Sort according return (univariate model)
• Form BUY and SELL portfolios
• Take new position every month

5
Methodology - Portfolios

• BUY Portfolio
• Top x performers in terms of total return for
  periods varying between 1 month and 1 year in the
  past
• SELL Portfolio
• Worst x performers

6
Methodology Trading positions

• Monthly trading positions
• Long BUY portfolio for a period varying between
  1 month and 1 year
• Short SELL portfolio for a period varying
  between 1 month and 1 year
• No cash investment Short position finances Long
  position

7
Results Portfolio composition

• Sample Jan 84 - Dec 94
• Higher total returns achieved for
• Portfolios top / bottom 5
• Trading position (backward / forward)
• 1/1, 12/1, 12/6
• Consistent across portfolio (fractile)
  composition

8
Results Portfolio performance
9
Annual Results 12/6 W/O Knockout
10
Results Returns w/o Knockout
11
Results Returns w/ Knockout
12
Results

• In Sample Knock out does not substantially
  improve performance
• Out of Sample Knock out significantly decreases
  returns performance
• Knock out should not be used in conjunction with
  a momentum strategy

13
Theory

• Momentum effect is not sector specific but moves
  across sectors
• A momentum based trading strategy should
  therefore account for momentum shifting across
  industry sectors

14
Improvements

• Calculate market exposure
• Consider other time periods
• Extend to other asset classes and/or other
  markets
• On / Off Trigger
• Optimize in conjunction with other asset classes