How To Choose A Quant Strategy

Following up on my last post re-introducing quantitative investing in individual stocks, I wanted to address probably the most important part of investing in individual stocks with quantitative methods, and that is how to choose which strategies to invest in? I’ll start out with a very obvious method but in my opinion flawed, show some examples of what can happen to strategies over the short term, and finally propose a quantitive method to choose the strategies themselves. Let’s get started.

The easiest, most obvious method to choose quant strategies is simply to use historical performance. The three strategies I included in my last post, Microcap, TV2, and Momentum are among the top performing strategies historically (as documented in What Works on Wall Street) and have continued to be among the top performing strategies in the last 14 years since those results were first published. They are also based on time proven investment factors; momentum, size, value, and combinations of those factors. Honestly, this is not a bas method. Obviously, there are potential pitfalls to be avoided, such as not using enough historical data, not using time proven factors, etc. but this has proven to be a pretty robust method. And obviously, it is super easy in principle. The problem with this approach can occur, and often does over the short term.

I’ll use TV2 as a great example. Lets based on the incredible historical performance of this strategy, you launched into it exactly 8 years ago, late April 2016. What would your results have been? Well, the good news is that as expected it has outperformed vs the SP500. The chart below shows the results of my implementation of the TV2 strategy from April 19, 2016 through April 19, 2024. You would have handily beat the SP500 by almost 50% points of total return over that time period. But if you look at the path to get to that result it sure wasn’t easy.

That time period started off great for TV2 and after about 10 months the model was beating the SP500 by 20 points. Then came several periods of underperformance and outperformance and after 4 years you would have found that your returns were about 0% and the SP500 was up about 40% or so. How do you think you would have handled this situation? Most investors would have abandoned TV2 at this point or even much earlier. Of course, as often occurs with these things, that was the worst time to abandon the strategy. Subsequent performance is outstanding and in fact is the best performing strategy over the last 3 years. This happens to all type of strategies, not only the most aggressive. Let’s look at an example of a more conservative, pure value strategy, based on shareholder yield that I developed a while back (in 2014) as an alternative to just buying the large cap index and beating its results in the process.

Shareholder yield is one of the classic value strategies. It is the modern equivalent of a dividend yield strategy. Shareholder yield is one of the top performing factors over time. It has excellent historical performance that is also quite consistent over various time periods. The chart below shows a snaphot of the strategy’s performance since the end of 2008. Returns and drawdowns are about the same if you take the results back to 1999.

Great results over the long term but you can surely spot the challenge just buy eyeballing the chart. That outperformance comes with a bumpy ride. Let’s look at a zoomed in chart over the last 5 years for this strategy.

Here you would have started out underperforming right away, had a massive run of outperformance until about the middle of 2021, and then basically underperformed since then. Over this 5 year period the model underperformed SPY by 65 points. Challenging to say the least despite it being a great long term strategy with great fundamentals. Is there away to minimize the risk of these underperformance periods?

The classic way to try and avoid these kinds of scenarios is to use a metric called base rates. You can read about it in an old post here. Basically, it is a measure of a strategy’s consistency in beating benchmarks. The higher the base rate, i.e. the higher the consistency of beating the benchmark, the better. It is a good approach. The best 3 strategies I’ve discussed, microcap, pure momentum, and TV2 also happen to have some of the highest base rates among all quant strategies. But base rates don’t solve the entire challenge either. As in the TV2 example I used above, towards the end of the period the strategy comes back and outperforms but it sure takes a while, about 5 years. By the way, TV2 has a 100% base rate at 5 years or greater, as does microcap. Most investors will not stick around that long. Is there a better solution? I think so.

My solution, and the one I use for my subscribers to QuantPulse, is to use momentum among the strategies themselves to pick which strategies to invest in. Basically, measure the performance of all the strategies over a certain period, and only allocate to the strategies that have performed the best. Turns out this works quite well. Performance is good, drawdowns and turnover are manageable, and this also diversifies the portfolio across strategies, market cap, and sectors. I use the top 3 strategies ranked by momentum updated monthly with each strategy consisting of 10 stocks for a total quant portfolio of 30 stocks. Results are shown below.

Overall, I think this is a better solution than picking strategies in a discretionary manner even if based on quantitative metrics. You just don’t know how the strategies that are picked are going to perform out of sample over the short term, that is anything less than 5 years. Using momentum among the strategies makes the selection of which strategies to use a quant metric in and of itself which in my experience is a huge benefit.