Which Number Is The Real Mc Coy?

It's hard to peg earnings when companies calculate the bottom line in different ways

When Amazon.com reported its second-quarter loss on July 21, you might have heard any one of the following numbers: 86 cents, 51 cents, or 42 cents a share. Officially, the online retailer recorded a net loss of $138 million, or 86 cents a share. But Wall Street analysts overlooked that figure in favor of a pro-forma net loss of $82.8 million, or 51 cents a share. The Seattle merchant also disclosed a third figure--a pro-forma operating loss of $67.3 million, or 42 cents a share. Like the other pro-forma result, this number ignored the costs associated with Amazon's recent acquisitions binge.

Confused? You're not alone. As more companies include unorthodox earnings figures in news releases and comments that accompany their official accounting disclosures, even pros admit they're bewildered. "There is no single more important number for an investor than a company's earnings per share. But the fuzziness around what actual earnings are is greater than 5 or 10 years ago," says Edward Keon Jr., director of quantitative research at Prudential Securities.

True, investors can always fall back on the net-income figure sanctioned by the Financial Accounting Standards Board (FASB), which sets accounting rules in the U.S. These official bottom-line numbers are found in the income statements publicly traded companies are required to release once every three months. The problem is that Wall Street analysts generally bypass these figures because many have decided they are not the most accurate barometer of a company's health. The reason? Net income often includes big charges for one-time events, such as mergers and plant closings.

DEEPER INSIGHT. Instead, most professionals rely on operating earnings--or profits minus unusual losses and gains. "The investment community always wants to look forward to what's coming down the road," says Charles Hill, research director at Thomson Financial's First Call, which polls analysts and publishes a consensus estimate of earnings. "Taking out unusual items makes sense because it gets you to the underlying earnings capability of a company."

But because there is no standard definition of operating earnings, First Call has found that even analysts covering the same company deliver earnings estimates that are not comparable. The problem has gotten worse as companies and analysts have become more creative about what they include and exclude from their calculations. "They are pushing the envelope more," Hill adds.

For example, Amazon, like many companies, is highlighting cash earnings, which ignore the bottom-line impact of noncash charges, including those related to mergers. According to J.P. Morgan managing director Frederic Escherich, 64 companies announced some variation of cash earnings from June to August, up from 31 in the same period of 1998.

Amazon's second-quarter earnings release provides deeper insight into its treatment of cash earnings. The Internet retailer ignored the cost of stock-based compensation picked up in mergers. Most analysts who follow the company have adopted the same practice, which is also prevalent among those covering eToys, TheStreet.com, and uBid, among other Internet and high-tech companies, First Call says.

OUT OF STEP. This version of Amazon's earnings also disregards goodwill, one of the biggest costs associated with mergers. Goodwill is the difference between the price Amazon paid for its acquisitions and the value accountants assign those companies by subtracting their liabilities from their assets. FASB rules require Amazon to regularly deduct from its earnings a portion of the goodwill sitting on its balance sheet. This will depress the official bottom line for up to 40 years.

Besides the obvious desire to make earnings look better, the justification for ignoring goodwill is that--like many other accounting rules--it was created with traditional industries such as manufacturing in mind. Therefore, critics argue, it is out of step with companies at the forefront of the New Economy. To see why, compare Amazon.com with General Motors. Among the carmaker's most valuable assets are the factories and other physical properties listed on its balance sheet. But a high-tech company's best assets are usually intangible, such as a creative workforce and a strong brand name. These do not appear on a balance sheet. For this reason, when New Economy companies buy other outfits, they generally pay steeper premiums over their targets' accounting values than old-line companies do, inflating goodwill. In addition, New Economy companies usually deduct goodwill over much shorter time frames, causing such big hits to net income that many analysts disregard it entirely.

Pioneered by Internet companies, the practice of overlooking goodwill has become widespread. Now, FASB has "sprinkled holy water" on it, Hill says. Specifically, the rulemaking body has floated a proposal that would allow companies to disclose two income figures--a new one that ignores goodwill and the traditional one, which does not. FASB's goal is to "provide a standardized measure of performance for those who would like to discount goodwill," says FASB research director Timothy Lucas.

But some fear that the new method, if adopted, could heighten confusion by creating the impression that FASB sanctions alternative approaches. "It will be a big misstep towards creating all kinds of nonstandard measures," says Jack Ciesielski, publisher of The Analyst's Accounting Observer (www.aaopub.com).

To bring some uniformity to operating earnings, Standard & Poor's is developing guidelines for the companies in the S&P 500-stock index. S&P, which, like BUSINESS WEEK, is a unit of The McGraw-Hill Companies, has long served as the unofficial arbiter of reported net income for the index's 500 companies. S&P plans to circulate a proposal to Wall Street firms early next year, says James Branscome, who oversees index operations.

Whether or not analysts accept whatever S&P's benchmark turns out to be, many applaud the effort. "A set of standards would be better than where we're headed now, which is that the rules are whatever the companies want them to be," says Prudential's Keon.

ASK THE EXPERTS. In the meantime, where can investors turn if they have concerns about official net income and management's carefully massaged earnings figures? Those with accounting backgrounds may be able to pick through the numbers themselves. But even experts can be foiled by companies that choose to lump together a slew of charges--as Amazon has done with several acquisitions and investment-related costs.

If you wouldn't know an inventory writedown from an in-process research-and-development charge, you can fall back on the earnings estimates published by services such as First Call and I/B/E/S. These firms do not mandate a standard method. Instead, they defer to the earnings recipe favored by a majority of analysts covering each company. When compiling consensus figures, they require dissenters to conform.

Research shows it can be useful to compare the services' estimates. In a recent study, Prudential Securities identified 325 discrepancies between the First Call and I/B/E/S estimates of the S&P 500 companies' earnings over the past five years. Subsequently, those stocks lagged the index, although they rebounded this year.

So, should you avoid companies whose earnings are in dispute? Not necessarily, says Keon, the study's author. Sure, discrepancies "could mean that some fiddling of the books is going on," he says. "But they could also mean the market doesn't understand the company, and when it does, the stock will rebound sharply." In other words, unorthodox accounting doesn't necessarily signal that someone is trying to pull the wool over your eyes. But to protect yourself, make sure you don't take any earnings figure--conventional or not--at face value.