Also notable: Trump settled the case with the SEC, rather than fight the charges—something that Trump recently has said he never does.

In the accounting world, few practices are attracting more widespread condemnation these days than the abuse of “pro-forma” or “adjusted earnings.” In his annual letter to shareholders, Warren Buffett once again assaulted “misleading numbers that can deceive investors,” endorsed by “Wall Street analysts parroting phony…’earnings’ figures fed to them by managements.” The Berkshire Hathaway (BRK-A) CEO argues that companies frequently use pro-forma reporting to choose the numbers that make them look good, and leave out the rest. Critics have roundly vilified Valeant Pharmaceuticals (VRX) for telling investors that what matters most are results that ignore such real, routine, and recurring costs as stock options and acquisitions expenses. After years in Valeant’s thrall, investors are no longer buying the doctored rap. Since last July, its shares have dropped 74%, erasing almost $70 billion in market value.

In June of 2002, when the SEC brought the landmark action against Trump Hotels and Casino Resorts, the agency said it pursued the Trump case—its first for this type of abuse—because it provided a “stark illustration of how pro-forma numbers can be used deceptively and the mischief they can cause.” The agency accused Trump Hotels of using “fraudulent” reporting used to “tout purportedly positive results.”

Prompting the SEC’s pioneering foray was a Trump Hotels press release from October of 1999. It appeared to announce a big improvement in Trump Hotels’ earnings. According to the release, the casino operator had $403 million in revenues and $14 million in net income in the third quarter of 1999, improvements over sales of $397 million and profits of $5.3 million for the same period in 1998. The release then boasts that its earnings per share of $0.63 handily bested analysts’ estimates of $0.54. The numbers, Trump’s press release said, demonstrated substantial progress towards making its money-losing operations profitable.

Those numbers however were inflated. Had Trump had correctly reported the company’s true results, it would been hard to find improvements.

The release didn’t use the words “pro-forma” or “adjusted.” It did mention that the results excluded an $81 million charge for shuttering the antiquated Trump World’s Fair hotel and casino. Leaving that out didn’t pose a problem, since it was a genuine, one-time expense.

Trump Hotels, however, in the numbers that they reported also included a thoroughly extraordinary item in an attempt to spin its lagging performance. Unmentioned in the release was that its revenues included an extra $17.2 million gain from the termination of a lease on the All Star Café, run by Planet Hollywood, an eatery in Trump’s Taj Mahal casino. Hence, Trump Hotels was both subtracting a giant one-time expense and adding a big non-recurring gain to arrive at a phantom number. The SEC found that correct pro-forma reporting means dropping all extraordinary items rather than erasing the ones with a negative sign and adding the big pluses.

In other words, Trump Hotels was creating its own parallel accounting universe.

At first, investors bought the story, even though the release showed revenues gaining only 2.9% over the past twelve months, but its profits jumping 164%. The day of the announcement, Trump Hotels shares rose 7.8% to $4.31. That was before analysts began digging into the results, and were rightly mystified by the pro-forma rendering.

Within three days, Deutsche Bank and Bear Stearns issued reports stating that Trump management was acknowledging that the big “improvements” were really the result of the one-time gain of $17.2 million. When that item was excluded, it emerged that revenues dropped 2.9%, and that profits hadn’t more than double, but instead had fallen by almost 43%. The shares quickly fell below $4, and Deutsche Bank lowered its forecast for 1999 for the company to a loss of $1.64 per share, lower than the $1.17 a share loss it had previously estimated.

Trump Hotels quickly agreed to a “cease and desist” agreement with the SEC, and never paid a fine for the matter. Still, looking back on the case, Trump Hotels took a truly extraordinary approach to inflating its results. Today, almost all the pro-forma abuses consist of excluding commonplace costs. Trump Hotels, however, was counting what was obviously a non-operating gain in its operating income. “It’s a different slant,” says Jack Ciesielski, author of the Analysts’ Accounting Observer. “Now it’s all about what you’re leaving out, and with Trump Hotels, it was what they were putting in.” Another big difference: The folks who unearthed the real numbers—in a rare show of righteousness—were the Wall Street analysts—the same people who now sell pro-forma as the new reality.

In retrospect, it’s hard to understand how any company could judge that big gain as revenue from the normal course of business. The company’s 10-Q reports that Trump Hotels booked the $17.2 million to reflect “all improvements, alternations, and All Star’s personal property” that it kept when All Star canceled the lease, as determined by an “independent appraisal.” Clearly, Trump Hotels couldn’t count on regular revenue from tenants’ canceling leases and handing back spaces with renovated bars, fancy banquettes, and glamorous lighting.

So who was Trump Hotels’ auditor at the time? What accounting firm could possibly call randomly inheriting an upgrade in décor as operating income? It was Arthur Andersen, the now-defunct firm that was overseeing Enron’s books at the during the same period, and that notoriously counted Enron’s sale of Nigerian barges as operating income, perpetuating one of the greatest frauds in corporate history. By 2005, Trump Hotels and Resorts was bankrupt. Now that its ex-chairman is running for president, it’s worth remembering where the big fuss over bogus numbers all got started.