Comparing Randomized Portfolio Allocation with Mean-Reversion Pair Trading

Summary

This project compares a Monte Carlo portfolio allocation approach with a mean-reversion pair-trading strategy on Indian equities. It uses three stocks from each of five sectors and evaluates portfolios with Sharpe, Sortino, and Calmar ratios. For allocation, the author randomizes stock weights, estimates annualized return and volatility from historical returns and covariance, and selects portfolios with the highest Sharpe ratio or lowest volatility. Sortino uses downside variation, while Calmar relates return to maximum drawdown. The pair strategy uses adjusted prices and a hedge ratio; the report says it proceeded despite the pair failing the Augmented Dickey-Fuller test.

The reported sector comparisons favor allocation in pharmaceuticals and financial services, and pair trading in technology, automobiles, and private banks. These findings use historical data from May 2017 through June 2020, with a zero risk-free rate assumption. The article warns that ratio choice affects conclusions, pairs may lack statistical support, short selling constraints and margin matter in Indian cash equities, and portfolio allocation may fare worse in bear markets. Its methodology and results are a student project, not broad evidence of future performance.

Key ideas

• The project compares three-stock allocation portfolios with a three-stock mean-reversion pair strategy.
• Sharpe, Sortino, and Calmar ratios capture different aspects of risk, including downside variation and drawdown.
• Randomized portfolio weights and historical return covariance are used to search for allocation candidates.
• The reported sector preferences are based on a limited historical sample and assumptions including a zero risk-free rate.
• The pair strategy was used despite failing the stated stationarity test, and short-sale rules affect implementation.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.