Questionable Accounting Rule
Draws Scrutiny From Analysts

By ELIZABETH MACDONALD
Staff Reporter of THE WALL STREET JOURNAL

Investors in stocks of high-tech and pharmaceutical companies, beware: A study warns that a growing number of the companies are using an obscure accounting rule that some critics say can inflate subsequent earnings artificially.

The authors of the report say that 3Com, Novell and Exar have already benefited from the rule. While all three companies acknowledge using the bookkeeping method, they say it is mandated by accounting rules, and hasn't distorted their profits.

The March 1997 study by Baruch Lev, an accounting and finance professor at New York University, and Zhen Deng, an NYU graduate student in accounting, found an increasing number of companies using the 22-year-old rule.

The rule in question lets an acquiring company set a value for the "in-process" research and development assets at an acquired company, and immediately write off that amount. The higher the value, the more acquirers can avoid hits to future earnings from goodwill. That's because any goodwill -- the premium of the purchase price over the acquired company's book value -- is supposed to be deducted from the acquirer's profits over periods as long as 40 years.

Jack Ciesielski, editor of the Analyst's Accounting Observer, a Baltimore publication for stock analysts, warns, "Investors should be careful of subsequent earnings posted by acquirers using this rule, because they are a bit jazzed."

Gabrielle Napolitano, a securities analyst who follows accounting issues at Goldman Sachs, holds a similar opinion: "Acquiring companies may be assigning too high a value to this in-process R&D, distorting subsequent earnings."

The practice is on the rise. The study says that only three companies wrote off part of their acquisitions as "in-process" R&D during the 1980s. But 389 have done so in the 1990s, a record 156 last year alone. Nearly 40% of the acquisitions occurred at computer-software and equipment companies, the rest at pharmaceutical and biotech concerns.

To be sure, companies must take such write-offs to comply with accounting standards. "The companies are not doing anything wrong," says Mr. Lev. "But what is wrong is the accounting system that allows these immediate write-offs." The authors say investors should watch out for euphemisms for this acquired R&D, such as "incomplete technology," "in-process engineering and development," or "acquired technology rights."

A computer search by Mr. Lev and Ms. Deng found 392 acquisitions using such a write-off between 1980 to 1996. The average R&D write-off was a whopping 72% of the entire purchase price. The size of the write-offs, the authors say, caused more than 75% of companies in the sample to post a loss for that quarter.

After that, the authors calculate, the write-offs gave a temporary 22% kick on average to the acquiring companies' earnings in the fourth quarter after the acquisitions. Reason: If they hadn't taken the immediate write-offs for the R&D, the acquirers would have had to count the R&D costs as expenses over four years -- which Mr. Lev estimates is about the average life of R&D software -- which would have reduced their earnings over that period. "Managers love this write-off because it inflates future earnings," says Mr. Lev.

Moreover, the authors estimate the write-offs temporarily overstated the acquiring companies' return on equity by an average 37% in the year after the write-off. That can put a smile on the faces of corporate executives, because return on equity is a yardstick used by compensation committees to set executive pay, according to William M. Mercer Inc., a New York compensation consulting firm.

The impact on earnings can be substantial.

In 1994, 3Com, Santa Clara, Calif., wrote off $192.9 million, or 88% of its total $220 million acquisition price for four companies, including Synernetics and Nice Com. Mr. Lev estimates that by the fiscal year ended May 1996, 3Com's write-offs had boosted its annual earnings by 23% over the level that would have been reported if the acquired R&D had been written off over four years.

Alan Groves, 3Com's corporate controller, said Mr. Lev used "highly fallacious" assumptions, and that 3Com's accounting was "conservative." But he wouldn't elaborate.

Another case in point: Novell of Provo, Utah. Between June 1993 and 1994, this maker of network software wrote off as "in-process" research and development $425.9 million, or 80% of its total $529.7 million price tag for four acquisitions -- Unix Systems Lab, Serius, Fluent and Borland International's Quattro Pro, a spreadsheet program. Novell eventually sold much of Unix's software line and Quattro Pro.

Mr. Lev says that if Novell had written off the R&D expenses over four years, its net income would have been $19.5 million in the year ended October 1996, less than one-sixth of the $126 million it actually reported. Novell's director of investor relations, Peter Troop, says, "Novell accounted for its acquisitions as required by generally accepted accounting principles."

Exar also wrote off $16.9 million in acquired R&D for two companies it bought between May 1994 and June 1994. By the end of the fiscal year ended in March 1996, this Fremont, Calif., telecommunications company reported $13.6 million in net income. But Mr. Lev says if it had written the R&D off over four years, its net would have been $10.4 million. So the R&D charges boosted net by 31%. Ron Guire, chief financial officer of Exar, remarks, "It's a nice study for a college, but in practice I think it's crazy."

Mr. Ciesielski says the sheer size of some of the R&D write-downs indicates that some acquiring companies are abusing the rules by including too much goodwill. But most companies don't release the details of the calculations they use to arrive at the write-down amounts. Companies wouldn't have to worry about the size of their goodwill write-downs if they used "pooling-of-interest" accounting; but companies that use that method for an acquisition can't quickly dispose of unwanted assets.

So far, both the Financial Accounting Standards Board, the chief accounting rulemaking body, and the Securities and Exchange Commission say they are looking into the controversy. But little action has been taken to date.

Determining whether R&D write-offs were appropriate in hindsight is complicated by the fact that, as Bear Stearns accounting expert Pat McConnell notes, "no reliable data on R&D cash flows postacquisition exist because such revenue is typically commingled with a firm's other earnings."

Adds Mr. Ciesielski: "If these write-offs are in fact all R&D, the companies should be generating phenomenal revenues and new products -- but are they?"

* * *

BUFFETT AT THE BAT: At the Berkshire Hathaway Inc. annual meeting Monday, billionaire investor and Salomon Inc. stockholder Warren Buffett said the "odds are overwhelming" that he'll convert some of his Salomon preferred-stock holdings into the common stock of the securities firm this year, rather than cash out.

In 1995, in what was seen as a no-confidence vote for Salomon's management, Mr. Buffett chose to take cash for some of his preferred stock, but, last year, Mr. Buffett decided to convert a $140 million chunk of preferred into Salomon common stock. Mr. Buffett can convert his preferred stock into common shares at $38 apiece, well below Salomon's current price of $52.625.

When asked about his plans, Mr. Buffett at first hedged a bit, saying, "We see no reason to swing at the ball when it's still in the pitcher's glove." Mr. Buffett doesn't have to decide whether to convert or not this year until the fall.