Choosing the Optimal Model-Based Investment Portfolio from Validea
In this post, I analyze public data from Validea to find the optimal investment portfolio from their collection.
About a month ago, a friend of mine reached out to me asking for my opinion on Validea and their investment quantitative models. I volunteered to crunch some numbers to find portfolios with the best performance.
Validea is analytical service that offers access to portfolios that follow established investment strategies, including strategies of Warren Buffet, Peter Lynch, Benjamin Graham, etc. It is important to understand that these portfolios are based on Validea’s own quantitative models that attempt to approximate these strategies. Validea is transparent about the accuracy and performance rates, and I used that data from their website in this analysis.
I started by scraping the data for the past 12 years (2009-2020) into CSV file. For the sake of simplicity, I downloaded performance indicators only for the optimal portfolio sizes and rebalancing periods. At the end, I had the following data points for each of the 22 portfolios:
- portfolio name
- optimal number of stocks
- optimal rebalancing frequency
- annual return (return)
- amount by which portfolio’s return exceeded S&P return that year (spbeat)
- year
I also included portfolios’ beta and accuracy rates, but I did not use them for anything.
Naturally, I first looked at the average returns for each portfolio:
The following models came in top 3 based on their average annual return:
- P/B Growth
- Twin Momentum
- Value Composite Investor
We also see some porfolios that clearly underperform S&P on average.
I then looked at the average “over-returns” - the amount by which portfolio beat the S&P:
The top 3 winners were the same three portfolios.
Next, I looked at how often each of the portfolios beat the S&P in the past 10 years:
The top 3 models in this category were:
- Twin Momentum
- Quantitative Momentum
- P/B Growth
Finally, I compared portfolios from the “wealth preservation perspective”: forget beating S&P, we just want to get positive returns as often as possible.
This time, the top 3 winners were:
- Twin Momentum
- P/B Growth
- Momentum
It looks like we have two models that consistently come on top. Dashan Huang’s “Twin Momentum” focuses on stocks that gain momentum in both price and seven fundamental indicators. This model has returned on average ~ 28% annualy, has beaten S&P over 80% of the time, and yielded positive returns nearly 9 out of 10 times. Partha Mohanram’s “P/B Growth” model has similar performance. Perhaps, it’s not a coincidence that both portfolios are offered as a part of Validea’s premium subscription plan.
It important to understand that all there results are based on models. In real-life trading, factors like trading fees, taxes, and timing can impact the results. The models’ accuracy (~58% each model) should be taken into account. The past performance is never a guarantee of the future performance. Finally, this post should not be considered as a review of Validea. I am currently not affiliated with Validea, but may consider using them in the future.