Robert Shiller is rightly revered for developing what’s arguably the gold standard for gauging if stocks are cheap, pricey, or somewhere in-between. The Yale University’s economist’s celebrated measure is the Cyclically-adjusted price-earnings ratio, or CAPE. The CAPE eliminates the distortions caused by big swings in profits that can make P/Es at any moment look artificially inflated or depressed. Think right now how P/Es appear outrageously high simply because the COVID has temporarily crushed earnings that are sure to rebound. Shiller smooths those peaks and valleys by averaging S&P 500 profits over the trailing ten years; he then adjusts that stream for inflation. His numerator, the “P” or price is simply the current reading for S&P 500 index. The “E,” the denominator, is that adjusted figure for earnings per share.

The CAPE has no peers as a predictor of where big cap shares are headed. And right now, it’s flashing bright red. The current reading, based on the S&P’s record close of 3853 on January 21, is 35.13. It’s only been that high in one period over the 140 years: in the run-up to the tech bubble that burst in 2000, sending the S&P down 44%. In the months before the 1929 meltdown, the CAPE peaked at 33, and stood at 27 prior to the collapse in the Great Financial Crisis. Usually, the CAPE ascends to its highest points when the economy is great. It’s especially worrisome this time that valuations keep planting the flag at higher and higher peaks when the economy is struggling to regain its output at the end of 2019, and according to the CBO’s projections, won’t get there until 2022.

A CAPE of 35-plus seems so high that even the bulls would strain to portray that equities as a bargain. But Shiller recently introduced a new yardstick that give the optimists a fresh narrative, and they’re mining it with relish. It’s called the Excess CAPE Yield. The formula takes the inverse of the CAPE, which is really Shiller’s earnings yield––his measure of the profits the S&P is delivering for each dollar investors are paying. He then subtracts the “real” or “inflation-adjusted” yield on the 10-year Treasury. That number represents the margin that stocks are paying over bonds.

What puts the gloss on the Excess CAPE yield, and hence on equities, is that the real yield on Treasuries is now famously negative, an extremely rare occurrence. As of January 21, the regular CAPE yield was a paltry 2.85% (the inverse of the CAPE at 35.13). That looks pretty bad. But Shiller puts inflation at 1.68%, which is 0.56% higher than the long bond rate of 1.12%. So the “real” yield on the 10-year is a negative 0.56%. The Excess CAPE stands at 3.41%. That means stocks, which are extremely pricey, still beat bonds by a wide margin, mainly because bonds are offering lousy, less-than-zero returns far into the future.

But does the Excess CAPE really have anything like the predictive power of the regular CAPE? It’s true that in previous periods, investors have done fairly well when buying in at an Excess CAPE of around 3.5%. It stood at that level in June of 2008 and August of 1988, and shareholders got 8.2% and 14.5% real annual returns over the next decade. But the evidence is muddled. Those who purchased at the 3.5% mark in October of 1973 lost 2%, and buyers in October of 1971 barely broke even. Folks reaped mediocre gains buying at 3.5% in the 50’s and 60’s.

It’s also significant that in recent months, the Excess CAPE has been pointing to lower and lower future returns. It’s dropped steadily from 4.88% in March of last year to today’s 3.41% as the S&P has soared, and it’s also been hit more recently by the spike in the long bond––as fixed income pays more, equities’ edge, as measured by the Excess CAPE, shrinks.

The argument for using the original CAPE, and ignoring the newcomer, the Excess CAPE, is three-fold. First, the Excess CAPE is really signaling that stocks are extremely expensive, and bonds are even more outrageously overpriced––so much so that they won’t even keep pace with inflation. “Saying that stocks are pricey and bonds are worse is not a reason to go out and buy stocks,” says Rob Arnott, chief of Research Affiliates, a firm that oversees investment strategies for ETFs and mutual funds.

Second, the Excess CAPE is even higher for European and Japanese stocks than for the S&P. That’s because they’re both much cheaper––the Shiller P/E is a lot higher––and their real rates are even lower. “That’s an argument for buying Japanese and European stocks instead of U.S. equities,” says Arnott. “If that measure really moves markets, those two will way outperform the U.S.” It’s also been that way for a long time, yet Japanese and European shares have proven poor performers despite their elevated Excess CAPEs.

Third, there’s nothing normal about negative real rates, and little reason to think they’ll stay negative for a period of years. Let’s look at a scenario where rates “revert to the mean” or go even higher. Say the real yield on the long bond rises to 3%, and regular CAPE yield waxes to 6%. That means bonds would drop 20%, and stock fall by 50% (because the CAPE would drop from 35 to 16.6, where it’s been many times before). Then, the Excess CAPE would still be 3%, just below where it is now (the 6% regular CAPE yield minus the 3% real yield on the long bond). In other words, the Excess CAPE barely changed, but stocks lost half of their value. It’s sending the same prediction now as when stocks were twice as pricey. How’s that for a forecast?

The Excess CAPE would have misled us, while the regular CAPE was right in displaying the real “CAPE fear.” Smart investors should stick with the original.