Momentum Investing: What happens if we boot stocks over 10x P/S?

Short answer up front–very little.⁽¹⁾

This was a simple question posed to me by one of our blog readers–what impact does excluding stocks trading at 10x P/S have on a Momentum portfolio?

A good question–especially for those who are “value” investors that are interested in momentum. For most systematic value investors, the prospect of adding stocks trading at over 10x P/S sounds ludicrous–see this article on the performance of stocks trading over 10x P/S.

Since I didn’t know the exact impact, I went and ran the tests described below.

Data

The universe is the largest 1,500 common stocks in the U.S. Data is from FactSet. The data runs from 1/1/1992 through 12/31/2021. Momentum portfolios are constructed every quarter (12/31, 3/31, 6/30, 9/30) by selecting the top quintile (300 firms) using a 12_2 Momentum calculation (the prior 12-months’ total return ignoring the last month). Returns are shown to equal-weight (EW) portfolios.

Within this top quintile of momentum stocks, I then bifurcate the sample into the following:

1. Above 10x P/S
2. Below 10x P/S

The image below shows the number of firms in each bucket.

As we can see, there are two time periods when the momentum portfolio has a decent amount of firms trading above 10x P/S (100 firms or more):

1. Post internet bubble–1/1/2000-12/31/2000
2. Post COVID–7/31/2020-6/30/2021

However, the majority of the time the momentum portfolio is under 50 stocks (out of 300) that trade above 10x P/S. The average number of firms within the momentum sample trading above 10x P/S is 44 firms (44 out of 300 = 14.67%).

So what happens if we include/exclude these firms from the momentum portfolios?

Results

To test the impact, we examine the returns of 3 momentum portfolios, and compare them to the market:

1. US MOM– ALL FIRMS EW: Top quintile of momentum firms (300). Portfolio is equal-weighted.
2. US MOM–EX-FIRMS > 10X P/S EW: Firms within the top quintile on momentum that are trading below 10x P/S. Portfolio is equal-weighted.
3. US MOM–FIRMS > 10X P/S EW: Firms within the top quintile on momentum that are trading above 10x P/S. Portfolio is equal-weighted.
4. SP500 INDEX: S&P 500 Index

All returns shown below are gross of any transaction fees and management fees.

The results above highlight a few things:

1. Comparing columns 1 and 2, we get the answer to the main question I was posed. Excluding momentum firms trading at 10x P/S has a marginal impact over the long-run (1/1/1992-12/31/2021). There is only a 6 bps a year difference over the entire time period.
2. Comparing columns 1 and 3, we see that building a momentum portfolio solely on stocks trading above 10x P/S is a bad idea (not surprising). One would lose ~ 3.5% annually compared to the simple momentum portfolio.

So is excluding stocks trading above 10x P/S a horrible idea?

No.

There are time periods when there is a pretty large divergence in returns between the naïve momentum portfolio and the momentum portfolio excluding stocks above 10x P/S.

Below are the annual returns, with a few years highlighted:

The annual returns above show that in the internet bubble and the post-COVID there were some pretty large divergences between the 3 momentum portfolios.

Summary

Overall, we see that excluding firms trading above 10x P/S has a minimal impact over the entire time sample (6 bps difference from 1992 to 2021). However, if one wanted to build a momentum portfolio and exclude these firms, it didn’t historically hurt performance. However, as many know, momentum requires a lot of turnover (see here), which can cause tax issues if not done within an IRA or another tax-efficient wrapper (such as within an ETF).