Short Term Asset Management

1
Short Term Asset Management
Tactical Asset Allocation Using a Sector
Prediction Model

• Pedro Candas
• Stepan Karpukhin
• Yoshinobu Kataya
• Robert Lee
• Luis Shirley

2
Agenda

• Objective
• Intuition
• Methodology
• Some of Our Models
• Data Analysis
• Comparing Our Model
• Results
• Conclusion

3
Objective

• Build a MSCI sector index prediction model using
  lagged MSCI sector index returns.
• Build a tactical asset allocation strategy using
  this prediction capability to outperform a
  strategic asset allocation of different MSCI
  sector indices

4
Intuition

• Markets are highly inefficient
• Sectors are interrelated
• Chemical sector depends on the Energy sector in
  their production process.
• Japanese industries are reliant on U.S. market

5
Methodology

• Gathered monthly MSCI sector index returns for
  Japan, U.K., and U.S. for 1990-2000
• Identified industry sectors for analysis.
• Developed prediction models with lagged industry
  data and other economic variables.

6
Some of Our Models
Dependent variables
Independent variables

• Japanese Financial Services Sector
• Japanese Chemical Sector

• Japanese Energy Sector(lagged)
• Japanese Bank Sector(lagged)
• Japanese Energy Sector(lagged)
• U.S. Energy Sector(lagged)
• Oil Prices(lagged)

7
Data Analysis

• Prediction models have up to a 64 hit rate
• Accuracy may have improved with different
  economic indicators
• Accuracy may have improved with different time
  frames on lagged data

8
Comparing Our Model

• Benchmark- Mean Variance Portfolio of the seven
  MSCI sector indices described in our
  models(Strategic Asset Allocation)
• Who will perform better?
• Strategic Asset Allocation
• versus
• Our Tactical Asset Allocation Model

9
Results

• At the same standard deviation (.034)
• Benchmark 11.1 annualized return
• Our model 23.0 annualized return

10
Results
11
Conclusion

• Predictive Models based on lagged sector returns
  may prove significant
• Greater accuracy using different timeframes for
  lagged returns data and better economic indicators