Value factors rely, for the most part, on some comparison between current price and some fundamental measure of a company’s production like sales, earnings, or cash flow. Dividend yield—dividends paid divided by price—has been in the stable of value factors for a long time. But does a high dividend yield indicate cheapness for a U.S. company? Or do we need to incorporate other value factors when identifying attractive high yield companies? Let’s investigate.
Here is the monthly dividend yield for the S&P 500 since 1900.
This chart is a little unfair, because it ignores return of capital to shareholders through share repurchase programs, which have become the more popular method. I’ve written a lot about buybacks here, here, and here.
Still, to get the capital return from a buyback program, you need to sell shares. Dividends are easier, steadier, and more reliable—and many investors prefer them. So the paltry yield available from the S&P 500 raises an interesting question. Is yield a good measure of cheapness? Short answer, for U.S. companies, is not anymore[i].
To see how that has the dividend yield factor has changed drastically since 2000, we can break all large U.S. companies[ii] into four distinct buckets by their dividend yield. Those buckets are 1) a dividend yield of zero, 2) a yield between 0-2%, 3) a yield between 2-4%, and 4) a yield above 4%. First, as we did with buybacks, let’s see how much of the total dividend pie is paid out by companies in these four buckets back to 1972 (when our data begins).
You can that in the 70’s—when the overall market’s yield was much higher—that most of the dividends being paid came from companies with a yield higher than 4%. But as time progressed, those high yielders represented a smaller piece of the overall pie. Visualized a little differently, here is the percent of companies in each yield bucket. The percent of companies without a dividend has risen, the percent with a higher yield (4%+) has fallen, and more companies pay a small yield (between 0-2%).
Now let’s turn to the valuations of these buckets through time. To evaluate cheapness, we calculate an independent value factor using sales, earnings, EBITDA, and free cash flow compared to prices and enterprise values. We calculate value it two ways: 1) comparing each stock to all other stocks, and 2) comparing each stock to other stocks in the same sector. This leaves us with a percentile score 0-100, where 0 means extremely cheap and 100 means extremely expensive relative to other stocks. Here is the rolling valuation of the various dividend buckets since 1972.
One trend is somewhat stable: companies without a yield tend to be more expensive than other stocks. But the more interesting trend is in the high yielders. Their valuation advantage has collapsed. Prior to the financial crisis bottom (pre-2009), the higher yielders were cheaper than the next bucket (2-4% yield) 92% of the time. Since 2009, they’ve only been cheaper 30% of the time. This is impossible to confirm, but the easiest explanation is that investors (especially older ones) have needed more income in this low rate environment, and have bid up the valuations of high yielders.
The story is quite similar (although not quite as extreme) if you calculate value relative to stocks in the same sector only.
This means that since 2000 or so, the highest yielding stocks (4%+) in each sector haven’t had a valuation advantage relative to other stocks in their sector, which they did on average prior to 2000.
Finally, we can chart the correlation between our two key data points: current dividend yield and current valuation percentile (market relative). Here it is since the peak in interest rates in 1981. Since 2001, it has been in a clear downtrend. Since the market peak in 2007, there has not been a meaningful relationship between yield and other measures of cheapness.
The implications for investors are as follows:
- Don’t assume that high yield means a company is cheap here in the U.S.
- If you are a value investor considering yield payers, use independent measures of valuation in your screening process.
- Consider looking beyond just the U.S.—which only represents about 1/3 of the global dividend pie—when searching for higher yielding companies.
More generally, you should never rely on a single factor when making decisions about a company. I’ve discussed valuation, momentum, quality, and yield extensively on this site. Using all or some of these factors in combination has been a much more effective and consistent approach to stock selection through history than using a single factor. Yield can be good, so long as you are mindful of other important attributes of the companies in question.
P.S. thanks to Meb Faber (@mebfaber) for suggesting I write a piece on dividends similar to the one I wrote on buybacks.
[i] The story is quite different in international markets, but that will be a different write-up
[ii] All U.S. companies that are both listed and domiciled in the U.S. so no ADRS, with market caps larger than the current market average. Today that is a floor of $7.5B or so, leaving roughly 500 companies.