Commentary: Why Better Numbers Really Matter

By Peter Coy

In finance textbooks, markets are perfect and stock prices reflect all available financial information. If that were the case, the plan by Standard & Poor's to define "core earnings" in a standard way would be unnecessary. Analysts and investors would already see through companies' "pro forma" earnings concoctions.

In the real world, though, they often don't. That's why the changes proposed by S&P (like BusinessWeek, a unit of The McGraw-Hill Companies) are a big deal. If widely adopted, they would provide a common frame of reference that would make it far more difficult for companies to hoodwink investors by playing games with their earnings. And ultimately, that would help ensure that individual stocks--and the market overall--are more accurately valued. Says University of Pennsylvania economist Jeremy Siegel, author of Stocks for the Long Run: "I think this has enormous import. This is exactly what investors are looking for."

Why is that? For starters, S&P core earnings are almost certain to be lower than the operating earnings now widely cited, though S&P executives don't know yet by how much. That means genuine earnings from core operations are likely lower than investors have been led to believe. And lower earnings per share implies higher price-earnings ratios, so the market may be more richly valued than expected.

How will investors react? They might be frightened and sell. Or they might be willing to pay more for stocks if they feel that p-e ratios are more trustworthy.

Of course, it's possible that the new measure won't affect the market any more than switching from Fahrenheit to Celsius affects how cold it is outside. Unless, of course, it becomes clear that accounting tricks have caused operating earnings to grow at an artificially rapid rate. "It might help deflate the market," says Yale University economist Robert J. Shiller, author of Irrational Exuberance. But S&P execs say they have no evidence yet that operating earnings have risen faster than core earnings.

So in the short run, will S&P's new standards have any impact? Yes, primarily on individual stocks. Short-term volatility could increase if core earnings fluctuate more than pro forma results. On the plus side, big crashes might occur less often, because bad news will dribble out instead of being exposed abruptly. S&P's new standards could even help the U.S. economy if they steer investors toward companies that can generate real value and away from those with artificially inflated results.

But perhaps the biggest reason core earnings matter is their potential impact on stock analysts. Analysts are supposed to be the ones who steer investors straight, but too often they parrot whatever earnings number a company highlights. The problem isn't necessarily conflicts of interest, although those are common enough, especially for sell-side analysts working for investment banks.

Many analysts simply lack the time to study companies in depth. Eric Hirst, an accounting professor at the University of Texas at Austin, notes that a typical buy-side analyst for a pension fund follows about 40 companies in a portfolio and keeps tabs on roughly 40 others. Analysts work about 50 hours a week and spend only about half their time on stock research. Says Hirst: "Analysts are very busy people and they tend to use data as it's presented to them."

In an experiment, Hirst and two other researchers tested 80 buy-side analysts, who presumably don't have the conflicts that sell-side analysts do because they're paid to recommend stocks for the portfolios of the investment firms that hire them. Hirst asked the analysts to study the statements of a pair of imaginary banks. There was a red-flag distinction between the banks, but the analysts didn't spot it unless it was highlighted in a format that's far more explicit than the one companies now use. Hirst also asked about 100 analysts to study a fictitious company. Most overlooked an obvious profit game.

Companies that resist standard earnings measures like S&P's new benchmark invariably argue that their own pro forma versions are superior because they exclude items that have no bearing on long-term earnings potential. But in a recent study, Russell J. Lundholm of the University of Michigan found that the more companies excluded from pro forma earnings, the lower their future cash flows were, holding other factors constant. The stocks of those with the most exclusions underperformed ones with the fewest exclusions by 17% over three years.

Economist Siegel calls S&P's core earnings plan "very, very positive." It's no panacea for shady accounting, but if it takes hold, pro forma statements will begin to look strictly amateur.

Coy is economics editor.

    Before it's here, it's on the Bloomberg Terminal.