Some Perspective on Yield-Gap Comparisons

Are stocks a good buy? With 13.5 months elapsed since last year’s bear market started—and the rebound since October 12 retracing little more than half the decline thus far—in a coldly emotionless environment, you might think the universal answer would be “yes.” But stocks are funny—when they are on sale, it often seems relatively few want to buy them. That is extra-true when bonds are also on sale since, when bond prices drop, yields rise. Such is the case today. With Treasury bills yielding around 5% and long-term bonds paying about 4% while earnings growth slows, headlines are starting to question stocks’ allure today. In our view, though, the comparison methodology has holes. And some historical perspective suggests it isn’t hard to see stocks’ appeal.

To compare stocks and bonds, some pundits use a metric called the earnings yield—the inverse of the S&P 500’s price-to-earnings (P/E) ratio. For the vast majority of the 2000s, it has been comfortably above Treasury bond and bill yields. Given the earnings yield is a rough approximation of stocks’ long-term expected price return, this means someone buying stocks could theoretically expect to return more than bond investors would receive in interest. These days, however, the gap is narrower. As Bloomberg noted this week, earnings yields now exceed 6-month Treasury bill yields by only about half a percentage point, the lowest since early 2001.^[i] The spread over 10-year yields, though higher, is also down lately—to its lowest level since 2007. Hence the questions about stocks’ attractiveness.

The comparison here isn’t entirely useless. Comparing companies’ borrowing costs to expected earnings yields gives a rough approximation of whether interest rates support investment. The same comparison can also show whether borrowing to fund a stock buyback is profitable (yes, after accounting for the US’s new 1% buyback tax). It can even be a decent snapshot of sentiment if the earnings yield isn’t skewed by short-term moves.

But the theory that it approximates long-term returns is hazy and general. In reality, stocks also pay dividends, businesses invest in new growth avenues and innovation turbocharges everything. The earnings yield doesn’t really hint at that. So, we don’t think comparing earnings yields to bond yields shows whether stocks or bonds are a better investment (and that is without factoring in all the things we think should underpin asset allocation decisions, including investors’ long-term goals, cash flow needs, time horizon and comfort with volatility). And after comparing stocks’ earnings yield to long-term bond yields for years and years, we have come to the conclusion that meaningful gaps between the two might be interesting but don’t predict future returns.

As Exhibits 1 (10-year Treasury bonds) and 2 (6-month Treasury bills) show, the comparison just doesn’t have much predictive power. Spreads were low and even negative throughout the 1990s—a fantastic time to own stocks. They were all over the map during the 2002 – 2007 and 2009 – 2020 bull markets. Interestingly, spreads rose during the 2000 – 2002, 2007 – 2009 and 2020 bear markets. They fell during last year’s, but we don’t think this says much. Rather, we think it is an after-effect of the fact that the Fed cut rates during those prior three and raised them during last year’s. That ancient and beyond widely known information is highly, highly unlikely to predict stocks, which are forward-looking and have long since priced Fed activity.

Exhibit 1: S&P 500 Earnings Yield Vs. 10-Year Treasury Yield

Source: FactSet, as of 2/16/2023. S&P 500 12-month forward earnings yield and 10-year US Treasury yield (constant maturity), 12/31/1995 – 2/15/2023.

Exhibit 2: S&P 500 Earnings Yield Vs. 6-Month Treasury Yield

Source: FactSet, as of 2/16/2023. S&P 500 12-month forward earnings yield and 6-month US Treasury yield (constant maturity), 12/31/1995 – 2/15/2023.

We also hesitate to read into the earnings yield at the moment. Now, we think it is still too early to say for sure that October 12 was the bear market’s low. Turning points are clear only in hindsight. But it is true that stock prices have bounced while earnings have fallen, which tends to happen in new bull markets. Earnings usually don’t recover until a quarter or two (in), registering the improvement that the new rally pre-priced. So, it is normal for P/E ratios to spike early in a bull market—which means earnings yields get temporarily depressed early on. It has the perverse effect of making stocks seem expensive when they are actually cheap, and the effect usually doesn’t even out for a few months.

If October 12 does indeed prove to be the low, then we would now be in the phase where stocks have bounced ahead of earnings. With about four-fifths of S&P 500 companies reporting, earnings fell nearly -5.0% y/y in Q4.^[ii] Analysts presently project a slight Q1 decline as well. That means the P is up while the E is still falling—inflating the P/E and shrinking the earnings yield. We saw similar conditions after each of the prior three bear markets. And those subsequent bull markets did a fine job of compensating investors above and beyond Treasury returns. Needless to say, low earnings yields then didn’t predict low returns or inferiority to bonds. We doubt they do now.