On Dec. 11, 1996, Boeing (BA) Chairman Philip M. Condit closed the deal of his career. After a relentless three-year courtship, he persuaded the initially reluctant directors of defense giant McDonnell Douglas Corp. to agree to a merger. The combination would create the world's largest aerospace company--the first manufacturer ever with the ability to build everything that flies, from helicopters and fighter jets to space stations. "This is, I believe, a historic moment in aviation," Condit proclaimed at a Dec. 15 press conference.
But far from the glare of TV cameras, a disaster was quietly unfolding inside Boeing Co.'s sprawling factories--one that would ultimately wind up costing billions of dollars, cause several executives to lose their jobs, and lead to claims of accounting fraud. Facing an unprecedented surge in orders because of a booming economy, workers were toiling around the clock, pushing the assembly line to the breaking point. At the same time, the company was struggling to overhaul outdated production methods.
These pressures were building up to what was, in essence, a manufacturing nervous breakdown. In the weeks after the merger announcement, parts shortages and overtime approached all-time highs--triggering red warning signals in Boeing's color-coded system for monitoring factory health. As costs went through the roof, the profitability of airliners such as the 777 swooned. A special team formed to study the crisis in May, 1997, issued a report with a blunt conclusion: "Our production system is broken."
If investors had understood the scope of the problems, the stock would probably have tumbled and the McDonnell deal--a stock swap that hinged on Boeing's ability to maintain a lofty share price--would have been jeopardized. But shareholders never got the full picture until well after the merger was completed on Aug. 1, 1997. Top executives "were hoping against hope that none of the problems would bubble up before they got the deal done," says a Boeing ex-official.
Their wish came true, and the deal sailed through. Recognizing that the bad news would have to come out eventually, company leaders started debating how it should be released. One public-relations manager lobbied for disclosure on Sept. 6 or Sept. 13--the dates of the funerals for Princess Diana and Mother Teresa--when a grieving public would presumably overlook the story. Those proposals were nixed. On Oct. 8, former McDonnell CEO Harry C. Stonecipher, by then Boeing's president and chief operating officer, shot an e-mail to Condit. "We do know for certain that there is a big surprise coming, and I think we owe the Street a heads-up. We have an unmitigated disaster on our hands and need some very candid damage control," he wrote in an e-mail that was disclosed in a securities-fraud suit filed against Boeing in Seattle.
Condit, who in his 32 years at Boeing had yet to experience a failure, responded that the disclosure should be delayed. "My bias is to soften the third-quarter hit with some warning," he wrote. "Assuming the scale of the problem remains, use the fourth quarter to prepare the Street to take the real hit then."
On Oct. 22, Condit made the bombshell announcement: The company's massive production problems would force it to write off $2.6 billion--by far the biggest charge in Boeing's history. Overnight, shares fell 8%, to $49.88, wiping out about $4.3 billion in value. As investors digested the scope of the mess, the company lost years of hard-earned credibility and the stock fell a further 12%, to 43, by Oct. 27. The stock stayed in a trough throughout 1998 and won the dubious distinction of being dog of the Dow that year.
The story of how Boeing kept its 1997 production disaster secret has never been made public. Although the company subsequently fixed the assembly-line problems--and it has recently received praise for its diversification efforts--the tale provides a sobering view of how easily management can keep investors in the dark. It also sheds light on the little-known "program-accounting" method used in aerospace to this day. A controversial system that many analysts criticize for its lack of transparency, it continues to give Boeing broad leeway to goose earnings--and to make it one of the toughest companies in America to evaluate. In a deal approved by U.S. District Judge Thomas S. Zilly on Feb. 20, 2002, Boeing settled a private securities-fraud suit over the 1997 episode for $92.5 million. The company did not admit guilt. Although some of the evidence uncovered by the plaintiffs' lawyers was revealed in court documents, the vast majority was locked under seal at Boeing's request.
In a three-month investigation, BusinessWeek has reconstructed this hidden chapter in the company's history--and analyzed its current implications. New details supplied by several inside witnesses indicate that Boeing did more than simply fail to tell investors about its production disaster. It also engaged in a wide variety of aggressive accounting techniques that papered over the mess. Critics say the company should have taken charges for the assembly-line disaster in the first half of 1997, even if it meant jeopardizing the McDonnell merger. They also claim that Boeing took advantage of the unusual flexibility provided by program accounting--a system that allows the huge upfront expense of building a plane to be spread out over several years--to cover up cost overruns and to book savings from efficiency initiatives that never panned out. "Boeing managed its earnings to the point where it got caught," says Debra A. Smith, a partner at Constraints Management, a Seattle-area manufacturing consultancy, and a former senior auditor at Deloitte & Touche who worked on the company's account during the early 1980s. "Boeing basically decided in the short run that [managing earnings] was a lesser evil than losing the merger," adds Smith, a onetime accounting professor at the University of Puget Sound in Tacoma, Wash.
At a time when investors are asking themselves how far Corporate America can be trusted, the Boeing saga provides rich new evidence that companies have much greater leeway to manipulate their numbers than most people suspect. The aerospace giant was a widely held blue chip that had a huge short-term incentive to prop up its stock price. Taking advantage of an investment community willing to tolerate the company's opaque reporting system, executives managed to conceal fundamental operational problems for nearly a year. Some of these officials, including Condit, are still at Boeing. To this day, they insist that they had no obligation to disclose cost overruns when they occurred in the first half of 1997--which raises the question of how swiftly they would let investors know if a similar problem arose today.
As is often the case, none of the outside watchdogs ever barked. The board never forced Condit to come clean about the company's production problems. Stock analysts and business journalists underestimated them. And although the company's auditor, Deloitte & Touche, raised red flags about Boeing's troubles, it doesn't seem to have put much pressure on its big client to share this information with investors. As a result, Boeing's financial reporting in early 1997 bore little relationship to its business reality. When the company finally disclosed its problems, "I was stunned," recalls Richard J. Glasebrook II, managing director of Oppenheimer Capital, owner of 5% of McDonnell at the time. "I thought that Boeing had the building of commercial aircraft down cold."
For its part, Boeing insists it never misled shareholders at all--and that critics and plaintiffs' lawyers are holding the company up to unreasonable standards. Noting that the construction of planes is an incredibly complex and expensive endeavor, it says that managers told the public about the 1997 production problems as soon as they were legally required to do so. It also points out that the Securities & Exchange Commission never filed charges. Boeing defends program accounting as a legitimate way to report performance and notes that its accounting decisions were approved by Deloitte--which was not named in the fraud suit. All of the key players in the drama, including Condit, Stonecipher, and former Chief Financial Officer Boyd E. Givan, declined to comment. "Boeing and its senior management always tried to do the right thing," the company wrote in a letter to BusinessWeek. "Boeing and its executives frequently disclosed Boeing's production difficulties." Deloitte also insists it acted properly.
So why did Boeing spend $92 million to settle fraud claims? General Counsel Douglas Bain says the company did not want to risk litigating complex accounting issues in front of a lay audience. He adds that because the case would have been tried in Seattle, which felt stung after Boeing moved its headquarters to Chicago on Sept. 1, 2001, company lawyers feared a vindictive jury. Instead, it was willing to accept the settlement--which Bain says is totally covered by the company's insurance policies.
THE ROOTS OF THE PRODUCTION CRISIS
If the Boeing lawsuit had gone to trial, jurors would have heard a story that begins around 1996, when Boeing was facing a strategic crisis because of shrinking defense business. Condit had a choice: either acquire a bigger share of the market or drop out of it. Since defense provided stable income to offset roller-coaster commercial plane sales, Condit chose to grow, buying Rockwell Aerospace & Defense in August, then striking the McDonnell deal just four months later. The ambitious chairman, always admired more for his strategic shrewdness than his operational expertise, was hoping to seal his legacy. But this bold expansion came at a time when the company faced mounting internal problems.
As the global economy emerged from recession in the early 1990s, aerospace companies enjoyed skyrocketing orders, which rose worldwide from 15 aircraft in 1993 to 898 in 1996. But Boeing's antiquated parts-tracking system couldn't keep up with the increased production volume. And supply problems prevented it from getting enough seats and electronic gear on time.
This problem was compounded in late 1994 when Boeing realized that rival Airbus Industrie, the European consortium, was undercutting it on price, thanks to lower manufacturing costs and government subsidies. By that year, Airbus had grabbed 30% of the global jet-plane market--up from less than 3% two decades earlier. It was a potentially devastating development, since lost customers in the airliner industry are hard to win back after they've spent a fortune training pilots and mechanics on rivals' equipment.
Boeing was forced to knock down costs across the board. It made early-retirement offers to 9,500 workers in 1995, slashing its staff of veteran mechanics and engineers. Execs also rolled out a bug-ridden new computer system for tracking parts, known by the unwieldy name of Define & Control Airplane Configuration/Manufacturing Resource Management (DCAC). As a consultant pointed out in a report to factory execs in the summer of 1997, the prospect of doubling production rates in the face of such change was like attempting "a four-and-a-half somersault off a 50-foot board into a pail of water."
By early 1997, warning signs were everywhere that Boeing's overheated factories were boiling over. On Dec. 2, 1996, Deloitte & Touche had alerted the board's audit committee "that production performance metrics began to trend unfavorably--overtime, parts shortages, rework, defective parts, and out-of-sequence work increased." About a dozen alarmed managers at the company's giant Renton (Wash.) facility met in May for a "reality check" meeting. They calculated that Boeing's production system was more than $1 billion over initial cost projections. One manager taking notes at the get-together concluded that "we have a real financial crisis on our hands" with "no relief" in sight.
While top execs did acknowledge publicly that the company was having some production problems in the first half of 1997, they consistently underplayed the issue. In fact, the company developed a Production Issues Communication Plan on Mar. 27, 1997, to stave off "focus[ed] media attention on the issue of our ability to meet customer commitments." The document, among other things, coached executives on how they should respond to reporters' questions, according to plaintiffs' lawyer Steve W. Berman, who brought the fraud suit against Boeing. Talking to reporters after the company's annual meeting in April, 1997, Condit said that the ramp-up in demand "has resulted in a near-term decline in productivity at company facilities and some supplier locations." With characteristic confidence, he said that the first quarter's inefficiencies "would not be repeated during the remaining quarters of the year" and that the company was not having "systematic" assembly-line malfunctions.
But already many managers were predicting late deliveries--a catastrophe that aircraft makers avoid at all costs, since it triggers enormous late fees and wreaks havoc on customers' business plans. As Boeing's Gary Scott, vice-president for 737/757 production, wrote in a memo on Mar. 20, 1997: "Late deliveries due to late [engineering] releases are becoming a significant problem." As the weeks wore on, the delays grew worse. On June 27, three days before the end of the second quarter, about a dozen Boeing commercial airplane executives met in a windowless meeting room in Renton for their second-quarter financial plan review. They learned that at least 15 aircraft would probably miss their delivery dates, according to the plaintiffs' complaint.
INSIDE CONCERN, OUTSIDE ASSURANCES
At this point--about a month before the McDonnell merger vote--Boeing should have started sharing some of the bad news with investors, according to Smith and other independent accounting experts. Under special aerospace industry rules that are part of generally accepted accounting principles (GAAP), companies are required to accrue reserves for probable late-delivery penalties and disclose them in the quarter they become known. Additionally, "abnormal" costs must be recognized in the quarter in which they are incurred. Expenses stemming from the company's production problems in the first half of 1997 "should have been booked in the quarters they were incurred and reflected in the profit numbers for those quarters," says Eugene A. Imhoff Jr., an independent expert on program accounting at the University of Michigan.
But Boeing never did so. On July 21, 1997, it reported second-quarter earnings of $334 million--a number that understated its production costs, the suit charges, by $292 million. What's more, investors still had no idea that the company had run up $253 million in unexpected costs in the first quarter or that it would soon would be facing more than $200 million in late-delivery penalties.
Boeing says it would have been a violation of GAAP to recognize any increased production charges in the first half of 1997. Of the 15 planes viewed as potential late deliveries on June 27, it says that "nearly all" shipped on time. Noting that costs always increase sharply during growth periods, lawyer Bain says the unplanned escalation in expenses that the company suffered in early 1997 was consistent with historic patterns. Therefore, he argues, they were not "abnormal"--and did not meet the test, under GAAP, for an immediate write-off. Boeing argues that nothing "abnormal" happened until it had to halt its production lines on Oct. 3. The lawsuit was primarily "a hindsight dispute about the timing of a write-off," says Ronald W. Stahlschmidt, a program-accounting expert hired to be a company witness.
But a Deloitte working paper written in the second half of 1997 offers evidence that some members of the auditing team took the opposite view: that the cost overruns should have been charged in the first half of the year. In the memo, Deloitte called the production disruptions in the first and second quarters "so abnormal as to require treatment as current-period charges." The auditor then said that Boeing's decrease in third-quarter earnings was "retroactively applied" to the first and second quarters of 1997. Boeing says that "the work paper's reference to `retroactive' does not suggest that Boeing should have restated its first- or second-quarter 1997 earnings. It means only that, in taking a charge correctly in the third quarter, Boeing estimated that charge by reference to year-to-date data."
BAILING OUT THE 777
The 1997 assembly-line meltdown afflicted every type of plane Boeing made. But it created a particular problem for the prestigious 777 line, according to former Boeing employees. To understand why, it's necessary to look closely at the unusual accounting system used by the aerospace industry. The program-accounting method was developed by Boeing and others in the industry in the 1960s to deal with a central problem: most of the costs of creating a new plane are incurred in its early years, while revenues roll in during later years.
To smooth out costs and revenues, aerospace companies are allowed to average them over the entire duration of an airplane "program"--usually defined as an initial production run of 400 aircraft. They do this by establishing a projected profit margin up front--say, 10% for the entire line. This number, which is continuously updated, is based on Boeing's estimates of the average costs and revenues over the remainder of the program. Every quarter, the "profit" the company reports is based on these projected averages, rather than its actual costs or revenues. The whole system is built on faith that aerospace companies can come up with accurate long-term forecasts. To a degree unmatched in nearly any other industry, aerospace companies' disclosures are based on their own private estimates. In this way, companies such as Boeing can absorb the ups and downs that characterize the industry.
But that comes at a considerable price: little transparency for investors. "The problem with program accounting is that it is virtually impossible to audit," says Lynn E. Turner, former chief accountant at the SEC and now director of the Center for Quality Financial Reporting at Colorado State University. "No one really knows whether the company will produce as many planes as [are] needed to recover the costs."
To mitigate this problem, the rules require companies to take an immediate charge as soon as they have evidence that a line's long-term profit margin will disappear--or, in industry lingo, that the program will be in a "forward-loss" position. And that's just what appears to have been happening to the 777 line in early 1997. Launched with great fanfare in the early 1990s, it had a development budget of $5 billion to $7 billion for initial design, production tooling, and flight-testing. By 1995, it had quietly overrun this budget by nearly 100%, according to two former high-ranking Boeing managers.
The prospect of a forward loss in the 777 was galling to Boeing, since it was the newest model--the plane that boasted the most advanced technology, that was to drive the company's performance in the next decade, and that carried Condit's reputation. Downgrading the 777's forecast would have been not only an embarrassment but also a threat to the merger. So it was bad news when, in its 1996 audit, Deloitte said the "low gross margin" projected for the current block of 777 aircraft risked a decline in "cost performance that would place the program in a forward-loss position."
To avoid this humiliation, the company allegedly started to shift monetary reserves from healthier aircraft programs to keep the 777 on budget, according to the complaint and two former high-ranking executives. These are funds that would accumulate when Boeing overestimated the costs for some plane lines. Officers established "management reserves" to pay for, as current corporate Controller James M. Bell put it, "unknown unknowns"--unexpected expenses that historically arise because of the complexity of airplane construction.
While the establishment of these management reserves is perfectly legal, it is a violation of GAAP to shift money from one program to another. Nevertheless, a handwritten note from a Boeing employee dating from May or June, 1997, said the company had borrowed from other divisions to boost the 777. "Much of the reserve balance used to cover performance was generated by other divisions and spent on the 777 division," it ran. Specifically, the company drew on reserves from the profitable 767 line after the reserve established for the 777 "was overrun" in early 1997, says one former Boeing manager.
Boeing vehemently denies it shifted reserves. The handwritten 1997 note "did not mean what the plaintiffs' lawyers suggested it meant," says Douglas Greene, an outside lawyer at Seattle's Perkins Coie who represented Boeing in the securities-fraud litigation. Boeing did not, and does not, have any "off-balance-sheet," "cookie-jar," or "general contingency" reserves, says Judith Muhlberg, vice-president for communications. "All reserves at Boeing are related to specific items."
Another method Boeing allegedly used to stave off a 777 write-off was exaggerating the effectiveness of some of the cost-savings initiatives it had launched in the mid-1990s. Under the flexible rules of program accounting, plane makers are permitted to make projections about efficiency efforts and start tabulating the benefits immediately--even if, as a practical matter, the initiative isn't yet working.
But this practice can run afoul of the law, Berman claims, if it is intended to prevent a particular airplane program from getting into a forward-loss position, which would require the company to take a current charge against earnings. And that's just what he says was happening in 1997.
According to the plaintiffs' complaint, Boeing "arbitrarily manipulate[d] cost-savings figures upwards in order to keep the 777 gross profit estimates from falling into a [forward-] loss position" during the second quarter of 1997. The complaint quotes a Deloitte working paper that says Boeing's managers admitted the second-quarter cost-reduction figures were "a plug" to keep the 777 profit margins on target. A plug is a commonly used accounting term for a number artificially generated to produce a desired result.
It is unclear which particular savings initiatives were used to create the alleged plug. As part of its plan to catch up to Airbus, Boeing had several efficiency programs under way. But former employees assert that the estimated savings from the DCAC parts-tracking initiative, for one, were exaggerated. Launched in 1994, the program cost a total of $1 billion, they say. Boeing estimated it would pay for itself in two years--and cut the expense of building a plane by at least 25%.
Problem was, the program cost far more than had been anticipated. Almost from the start, the DCAC software was riddled with bugs. One exec involved in DCAC estimates that overruns reached hundreds of millions of dollars. An internal audit in the third quarter of 1997 revealed that estimated DCAC cost savings were being "arbitrarily predetermined" to show more of a savings benefit than actually existed. When production managers at Boeing's Wichita assembly plant were shown the estimated savings they were expected to meet in March, 1997, "no one would sign off on these numbers" because the savings benefit appeared unrealistic, according to a report written by internal auditor Susan Parker. Even without the signatures, the DCAC savings figures were included in that year's business plan, the report states. Boeing declined to discuss in detail the allegations that it used bogus cost savings to prop up the 777. "We are limited in the amount of things we want to say about it for competitive and proprietary reasons," says Muhlberg. She added that "Boeing believes that its DCAC/MRM...cost-savings estimates were reasonable when made."
CONTROVERSY OVER PROGRAM ACCOUNTING
Boeing's efforts paid off: The company never declared a forward loss on the 777 in 1997--and has not done so at any time since. Does that mean that the line met the original profitability targets? Not necessarily. Very quietly, Boeing has bought itself time to resolve the line's problems by increasing the number of planes in the 777 program. The initial block of 400 units has been extended twice--first to 500 planes, and then later to 600.
Under the rules of program accounting, these maneuvers directly improve the 777's reported profitability by lowering the percentage of the original development costs that are charged each quarter. Just extending the block from 500 to 600 aircraft reduced deferred production costs for the 777 program alone by $8 million per plane, according to Todd Ernst, a Prudential Securities Inc. aerospace analyst. He calculated that this saved Boeing $68 million a quarter and boosted quarterly earnings by up to 6 cents a share for 2001. "The magnitude of the impact struck me," he said. "It was a big change in the cost of the airplane."
A neat trick--and one that is perfectly legal under the GAAP rules governing program accounting. The company didn't have to announce this important development with a big press conference. To the contrary, the news was buried in a cryptic sentence deep in Boeing's 2000 10-K. For its part, Boeing claims that programs are extended not to defer production costs but to reflect more sales than originally anticipated. When Boeing has firm sales contracts or high "certainty that the market is going to absorb more airplanes than we have in the current [program], we extend it," says Controller Bell.
No doubt other corporations would love to have such financial flexibility. The wonders of program accounting give Boeing more ability to paper over any short-term disasters than is enjoyed in nearly any other industry. As a result of this situation, investors need to be able to place an unusually high degree of faith in the company's managers. Boeing says that it has always earned shareholders' confidence with its rigorous cost- and sales-estimation process. The business planning process in 1997, for example, was "methodical and analytic," the company wrote to BusinessWeek. "It represented Boeing's half-century of experience in manufacturing commercial airplanes. Hundreds of industrial engineers and accountants worked as separate teams to estimate the thousands of separate cost items in each program."
That may well be the case. But if there's a moral to this story, it's that when the stakes are high, the temptations to take advantage of any flexibility in accounting standards can be great. "If you can prove that subjective judgments are dead wrong to the tune that Boeing is willing to cough up $92 million, they must really have been exaggerating their hopes for the future," according to Imhoff, the University of Michigan program-accounting expert. The ease with which executives have the capacity to manipulate program accounting is one reason why the SEC told the American Institute of Certified Public Accountants in the early 1980s that it would not permit the extension of the system to other industries, according to Turner, the former SEC accounting chief.
After the 1997 meltdown, Condit nearly lost his job and Chief Financial Officer Givan was replaced in the position by Deborah C. Hopkins. To regain credibility with the markets, the company took definite steps in an effort to increase its own financial transparency. Former company officials say that Hopkins and Stonecipher toyed with the idea of dropping program accounting altogether. The two Boeing executives complained that the system confused investors and didn't provide managers with sufficiently detailed information about the costs involved in manufacturing an aircraft. "You cannot reduce the cost of a wing if you don't know where you are starting," Stonecipher complained in an April, 1999, interview with CFO magazine.
That idea was abandoned. But in 1999, the company did start providing investors with supplemental information based on standard unit accounting. For example, it now supplies a line item in its reports indicating what the commercial airplane division's quarterly earnings would be on a traditional unit basis.
This information, in theory, should make it harder for the company to mask a production crisis ever again. But analysts complain that the direct operating income of the commercial aircraft division is mixed in with earnings from its pension fund and leasing business. And those items, which are only disclosed annually, can be significant. In 2001, pension-fund income accounted for about $785 million, or 18.6% of Boeing's pretax earnings. "A whole mishmash of other operating items get thrown into this unit-accounting line item," complains Robert Friedman, aerospace analyst at Standard & Poor's. "It definitely clouds the quality of earnings picture."
So while Boeing is not as big a mystery today as it was in 1997, it is still much harder to evaluate than most companies. Its accounting methods are none too popular with the professionals charged with decoding Boeing's books. "You can drive a truck through what's GAAP in aircraft manufacturing," says Heidi Wood, an aerospace analyst at Morgan Stanley Dean Witter & Co. "I think everybody has grown weary of program accounting for a while."
At the moment, though, there are no plans to get rid of the system. That's not a bad thing for shareholders to keep in mind. At a time when investors are seeking the maximum in transparency, Boeing is not even close to that standard. By Stanley Holmes and Mike France