Back to Community
Classical Asset Allocation: Combining Markowitz and Momentum

Based on a great paper: Momentum and Markowitz: A Golden Combination by Wouter J. Keller, Adam Butler and Ilya Kipnis, and also a well developed An Open-Source Implementation of the Critical-Line Algorithm for Portfolio Optimization by David H. Bailey and Marcos Lopez de Prado, and also a published strategy by David Edwards.

Which impress me about CAA is in the paper it was tested by 100 years data, and did well in very serious situations, for instance, 1929 and 1987, and sustained a stable growth about 10%, but much less volatility than market. I believe it will do well in the following possible bear market.

The code is quite simple and I strictly follow the strategy description in the paper about CAA strategy.

May someone help to improve the strategy in following aspects but not limited:
1, Find out any problem or polish it to a better performance;
2, Extend to N = 16 or even N = 39 assets to see what will happen;

David Qi

Clone Algorithm
Backtest from to with initial capital
Total Returns
Max Drawdown
Benchmark Returns
Returns 1 Month 3 Month 6 Month 12 Month
Alpha 1 Month 3 Month 6 Month 12 Month
Beta 1 Month 3 Month 6 Month 12 Month
Sharpe 1 Month 3 Month 6 Month 12 Month
Sortino 1 Month 3 Month 6 Month 12 Month
Volatility 1 Month 3 Month 6 Month 12 Month
Max Drawdown 1 Month 3 Month 6 Month 12 Month
# Backtest ID: 59670b85f0bb97525285b086
There was a runtime error.
2 responses


There is no symbol EW in N = 8 universe, but there is EWJ.
That substantially change performance metrics.

Thanks Vladimir for you point out the issue. I have corrected the code and updated. Although the curve is still stable, the performance seems much poor. Maybe we can find a better asset setup.