Once the variables are selected, the next step is to conduct a rigorous decile analysis of returns within a defined stock universe, such as the S&P 500. Decile analysis involves sorting the universe of stocks into ten equal parts based on a specific variable and then evaluating the historical performance of each decile.
This analysis should be performed over multiple time periods (e.g., 1 year, 3 years, 5 years) to assess the predictive power of the variables over different market cycles. This will help determine:
-
- If the top decile consistently outperforms the bottom decile.
- Whether the predictive power of the variable is stable over time or has weakened or improved during specific periods.
The aim is to identify whether a variable like FCFY or trailing ROE has exhibited strong performance differentiation between deciles, with a clear and consistent gradient. If such a relationship exists, it increases the likelihood that the variable holds predictive value in future stock performance.
Initial Selection of Variables
Combining Variables for Multi-Factor Analysis
Stock Selection Based on Ranking System
Incorporation of Additional Buy-Sell Rules
Determining Buy-Sell Frequency
Rebalancing and Trim Frequency
Setting Trimming Boundaries
Determining the Number of Stocks in the Strategy
Back-Testing the Strategy
Turnover Management: Reducing Excessive Trading
Frequency Adjustment for Portfolio Trimming and Buy-Sell List
Portfolio Size Adjustment
Stress Testing for Robustness
