FORTUNE--On CNBC’s Squawk Box on June 6, former Fed chairman Alan Greenspan made two points in uncharacteristically clear, blunt language that got Wall Street buzzing. Appearing over a Chyron billing him as “The Maestro of the Markets,” Greenspan came close to declaring that bonds are in a bubble. “Bond prices have to fall and long-term rates have to rise,” he warned. “We haven’t a clue as to how rapidly that’s going to happen.” Rates, he added, “are still well below” the level they would normally be at this stage.
Greenspan followed that reasonable alert on bonds with a mystifying endorsement of stocks -- which equity optimists have been parroting ever since. “The most positive thing in the economy is the fact that equity premiums are so high,” Greenspan said on CNBC, “which means the downside to stocks is quite limited.” He went on to profess that “if we can get stock prices to rise, which they should if this thing stabilizes,” then investors’ newfound wealth will prove a tonic for the economy. By “this thing,” Greenspan seemed to mean the shift he’s advocating in Fed policy: a gradual reduction in its bond purchases that, if started right away, may not rile the markets.
Greenspan is hanging his bull case on the concept of “equity premiums,” or the “equity risk premium” (ERP), as it’s known in academic finance. Ever since he uttered those wonkish words, CNBC anchors and their money manager guests have seemed entranced by the term. Indeed, the ERP is perhaps the single most important idea in portfolio theory -- the “holy grail” of investing, in the words of Kenneth French of Dartmouth’s Tuck School of Business.
MORE: Why Bernanke can't win
The notion is simple. The equity risk premium is the extra return that stocks must generate above those of risk-free bonds to entice investors to buy them. It makes sense: Investors want plenty of extra juice to compensate for the chance that their stock portfolios will take a deep dive, a regular occurrence in the quicksilver equity markets.
Greenspan is simply saying that stocks are offering a huge edge over bonds right now. That’s true -- but totally misleading. His argument that today’s warped, unreliable ERP bodes well for the future is seriously flawed. Greenspan, who did not reply to an emailed request for comment, implies that the ERP will remain large as interest rates rise -- but the very opposite will occur: The ERP will fall precisely because rates are increasing.
Let’s calculate the ERP, and see why it will inevitably shrink. Today, the stocks in the S&P 500 (spx) are trading at an average of 18.8 times trailing 12-month GAAP earnings (through the first quarter of 2013). If the current price/earnings ratio remains constant, investors should expect a return equal to the “earnings yield” obtained by flipping the P/E to arrive at the ratio of earnings to the stock price. That number is 5.3%, not including inflation.
Normally, a 5.3% real return, or in the 7%-plus range including inflation, would hardly be exciting. An 18.8 P/E is pretty expensive, and never, ever a reason to conclude that stocks are fabulous buys. But we’re in a strange new world, orchestrated by the Fed.
What makes the 5.3% number look good is the gigantic spread over bonds. Right now, the 10-year treasury is yielding just 2.2%; with inflation running at around 1.4%, the “real yield” is .8%. That means the ERP stands at 4.5% (the 5.3% yield on stocks minus the .8% yield on bonds). That’s extremely high by historical standards, around 1 point over the long-term average of approximately 3.5%.
Those conditions can’t last. Eventually, financial numbers revert to the mean. In fact, that has already begun happening. The ERP was at 5% as recently as May 2, and it has dropped a half a percentage point over the past six weeks as 10-year treasury rates have risen.
The trend will continue. Over the past quarter century, the 10-year treasury yield has averaged 5.4%, and inflation has run at 2.6%, giving a long-term real rate of 2.8%. Once rates get back to those levels, today’s ERP will shrink. The 5.3% expected return, less the 2.8% real rates, leaves a paltry ERP of 2.5%, a full point below what investors have expected for a half-century.
Soon, stocks will need to compete for money with bonds that yield far more than they do today. The 2.5% ERP won’t be nearly enough to lure wary investors. They’ll want the 2.8% risk-free return, plus the normal 3.5% premium, or around 6.3% before inflation. An earnings yield of 6.3% translates into a P/E of 15.9, far below today’s 18.8. To get there, stock prices would need to drop by 15%.
Of course, the bulls could advance a more optimistic scenario, in which earnings grow at a feverish pace, producing the fatter returns necessary once bond prices start rising. That would allow stock prices to keep surging from their current levels instead of first dropping hard. In this case, profits would need to expand at around 11% annually for several years to generate the required 6.3% inflation-adjusted gains. That’s just what Wall Street is predicting, and then some. The analysts’ “consensus forecast” projects a 23% increase in S&P 500 earnings over the next year. This argument holds that stocks aren’t overpriced, but that the high P/E simply augurs big profit growth in the future -- from, I might add, already record or near-bubble levels.
Rob Arnott, founder of Research Affiliates, a firm that oversees strategies for $142 billion in investment funds and one of the best minds in finance, isn’t buying that earnings will grow remotely as fast as Wall Street always predicts. “The risk premium argument only works if you believe earnings aren’t at peak and not vulnerable to retreat, and they are both,” says Arnott. “It also only works if you believe near-zero real interest rates are sustainable and normal.”
Arnott highlights the Fed’s role in creating the illusion that stocks are a bargain. “It’s not that stocks are attractively priced, it’s that they look good compared with bonds that are driven by financial repression. The large ERP would disappear if Quantitative Easing goes away. It’s a phony risk premium.”
Another problem is the low dividend payout ratio. That, along with high prices, produces low dividend yields. “The bull argument is that all of those retained earnings will be reinvested at high rates of return,” says Arnott. “But my research with Cliff Asness [co-founder of hedge fund AQR Capital] proves that a lot of those retained earnings are frittered away.” Arnott maintains that real earnings per share grow far more slowly than advertised, around 1.5% a year. Add that number to the 2% from dividends, and total return comes to around 3.5%, or around 6% including inflation. And Arnott insists even that number is optimistic.
So Greenspan is right on bonds, but wrong that today’s phantom equity risk premium bodes will for investors. As Arnott notes, “It’s fascinating that we’re in the only business where the more expensive something is, the more people want it. That goes against common sense but not against human nature.” People on Wall Street are paid handsomely for feeding our instinct for herding, and ignoring the numbers.