Using the Price to Sales Ratio

Price to sales is a very simple valuation ratio. It has the tendency to bias you towards lower margin and higher debt companies, all else equal, but it has still been a very effect measure of cheapness and a fine standalone factor for stock selection. Having explored the history of the ratio, let’s now turn to its measurement and usefulness in stock selection.

Here is a summary of the finding which will be explored below:

  • The largest return spread for best minus worst is in the simplest factor: plain old price-to-sales, un-adjusted for debt, sector, or size.
  • Sales-to-enterprise value—which controls for the bias towards high debt companies—works well.
  • Adjusting for sector does not materially change results, but it does control for large potential sector exposures that result from using the un-adjusted factor.
  • Like all value factors, the various iterations of price-to-sales work better in small stocks than in large stocks.

Our universe is all U.S. stocks trading at market caps greater than $200MM back to 1963. Everything is rebalanced as usual, on a rolling annual basis (although the signal can last longer than 12 months).

Here are some high level results.

sp large and all

Broken value investing record: the factors have worked! Here is how sales to price has done in rolling five year periods (versus equal weighted benchmarks of all stocks and large stocks).  The factor has worked at roughly the same times in the large and all stock universes.

rolling excess price to sales

One notable feature of this factor is that the best decile is 70% industrial and consumer stocks on average.sp allocations

As we did with price-to-book, we can adjust the calculation of price to sales to be relative only to other stocks in the same sector, which results in the following balanced allocations and results.

sp adjusted allocations sp adjusted gics large and all

Finally, we saw that price to sales biases you towards companies with higher debt and lower margins. One way to adjust for the debt bias is to use the sales to enterprise value multiple instead. Here are the results for that method (quite similar).

sev large and all

And here are the results through time comparing sales-to-price and sales-to-enterprise value. I overlay the trend in interest rates not to show that one causes the other but merely that the declining relative advantage of Sales/EV over Sales/Price has coincided with a steady decline in interest rates.

sev vs sp

The bottom line is that like all other value factors, price to sales works. Be aware of the biases inherent in the factor (and ideally use it in combination with other factors, we will get there). If value works by helping to identify stocks for which the market has grown overly pessimistic, we should be careful not to invest in a company that looks cheap (i.e. out-of-favor) but is really just a lower margin business that uses more debt than others.