Hansen's Superior Predictive Ability (SPA) Test

This Python project implements Hansen's Superior Predictive Ability (SPA) Test, a statistical test for evaluating the performance of multiple strategies against a null hypothesis. The SPA test adjusts for the correlation structure between strategies and accounts for multiple comparisons, identifying statistically significant outperforming strategies. This version doesn't use bootstrapping and uses a loglog correction.

Features

• Calculates Sharpe ratios for multiple strategies based on their log returns.
• Transforms Sharpe ratios using Hansen's adjustment to account for correlation effects.
• Computes significance thresholds and identifies strategies that outperform a given null Sharpe ratio.
• Implements multiple-comparison correction to control for false positives.

Installation

Requirements

This project requires Python 3.8 or later and the following libraries:

• numpy
• pandas
• scipy

To install the required libraries, run:

pip install numpy pandas scipy

Files

• superior_predictive_ability.py: Main Python script implementing the SPA test.
• log_returns_matrix.csv: Sample dataset containing log returns for various strategies.

Usage

1. Prepare a CSV file containing log returns for multiple strategies. Each column should represent a strategy, and each row should represent a time period. Ensure the index is a timestamp or sequential period identifier.
2. Use the following code to run the SPA test:

   python superior_predictive_ability.py
   

   The script reads the log returns data from log_returns_matrix.csv and performs the SPA test with default parameters.
3. The output will list statistically significant strategies or indicate if none were found.

References

• Pav, S. E. (2019). "Conditional inference on the asset with maximum Sharpe ratio".
• Hansen, P. R. (2005). "A Test for Superior Predictive Ability". Journal of Business & Economic Statistics.