FORTUNE—No basic measure of whether stocks are under or overpriced gets more respect in academic circles than Robert Shiller’s Cyclically Adjusted Price-Earnings ratio, or CAPE. The CAPE ratio removes a factor that frequently makes shares look artificially cheap or pricey, and hence wildly misleads investors—big, temporary shifts in earnings. Instead of simply plugging the last four quarters results in the denominator, Shiller calculates a 10-year average of inflation-adjusted profits that smooths out the summits and valleys to arrive at a more accurate price-to-earnings ratio. What’s so compelling about the CAPE is its power to predict. When the CAPE is well below average, as it was in 2009, investors can expect sumptuous returns over the next decade. When it’s lofty, as it is today, future gains will probably fall far short of what investors expect.
It’s hardly surprising that Wall Street, a venue that’s dominated by bulls, generally dislikes the CAPE. That’s never been truer than today, when the model is pointing to a dreary future of mid-single digit gains…at best.
Now, one prominent market watcher claims that Shiller’s CAPE is seriously flawed. In a video interview that debuted Wednesday on the yahoo.finance website, Tobias Levkovich, chief U.S. market strategist for Citigroup, states that his own methodology effectively points in the opposite direction from the CAPE, showing stocks to be substantially underpriced. “The market has a 91% probability of being up 12 months from today,” declared Levkovich.
Levkovich agrees that adjusting for the fluctuations in earnings is legitimate. “By taking average earnings over time,” he says, “you’re really saying, ‘We don’t know where we are in the cycle.’” But the Citi strategist argues that the Shiller P/E fails to make another essential adjustment: incorporating today’s interest rates into the formula. He’s reminding us about the “time value of money,” and says “you’ll pay very different present values based on what the discount rate is. It gives you very different outcomes on those P/Es.”
He’s right, to an extent. The value of a stock does equal its future cash flows, discounted at rates that regularly rise and fall, just as earnings do. What’s not convincing is the whole idea that low interest rates, which usually are accompanied by extremely expensive P/Es, predict high stock returns. Levkovich uses, at least in one model, the swap rate that predicts where the 10-year treasury yield will be in five years. Today, that five year forward swap rate is around 2.3%, even lower than the current yield on the 10-year.
Sure, when you discount earnings at that fantastically low rate, today’s valuations look cheap. No mystery there. He appears to be saying that today’s bargain yields more than justify the market’s overall price. (I sent an email to Levkovich requesting comment, but haven’t heard back yet. When I do, I promise to update this story.) But very low real rates also point to something else: an expectation of poor economic growth. The expectation in such an environment is that earnings won’t grow as fast as normal, let alone faster. That doesn’t exactly sound like a great basis for soaring share prices.
Shiller’s model, on the other hand, shows that the market is borderline expensive. He’s not saying that the market can’t get yet more exorbitant, of course. It can: Traders are not always rational. But right now, it’s darn pricey–a precarious place to be if you’re an investor long on stocks.
So who’s right? Well, in my view, it’s Shiller.
Levkovich’s argument is really a variation on the highly popular “Fed Model” formula. The Fed model compares the earnings yield on stocks, the inverse of the P/E, to the yield on 10-year bonds.
Again, it seems to make sense: When treasuries are paying you around 2.5%, and the S&P earnings yield is 5.3%, it’s a no brainer: stocks are the best deal.
The Fed Model does have plenty of influence, all bad. Investors irrationally drive earnings yields down, and P/Es up, making equities far more expensive, whenever interest rates fall. Earnings and government bond yields do move together. But when you get to a trough in both, the bulls argue that stocks (lousy yields or not) look great…ahem…because bonds look even worse. That’s, in essence, what they’re saying today.
In theory, the Fed Model is wrong because lower rates are typically the product of low inflation. When inflation falls, companies don’t raise their prices as fast as in the past; their price trends are inflation. So the lower profit growth from soft prices offsets the low discount rate. Let’s say that real interest rates, after subtracting for future inflation, remain unusually low. That would be a big boon if investors started by paying bargain prices. But the fall in rates had the effect of lifting stock prices, as we know. Which means that investors (who bought in at already “inflated” valuations, so to speak) are likely to get lower-than-normal returns.
“Even without the real rate reverting to the mean, unless real profit growth exceeds historical norms, you get lower-than-normal returns just because the starting price is higher,” says Cliff Asness, co-founder of AQR Capital who received his Ph.D. from the University of Chicago under Eugene Fama, a legendary economist who shared the 2013 Nobel with Shiller.
How high are today’s prices, really? The Shiller P/E stands at 25.2, far above its historical average of around 16. It’s so much higher than the 18.8 multiple on the last four quarters of earnings for a simple reason: Profits have never, ever remotely been at today’s levels. That’s true whether you measure them as a share of GDP, sales or by any other metric.
That would strongly suggest that the market bash can’t continue. And the only trend that can produce the better-than-average returns that the bulls predict is real profit growth that exceeds historic norms. Great if that can happen. But given that we’re already in record territory, do you really want to bet your nest egg on it?