“I think he’ll fight us,” Haueter said. “For some reason, Bob does not like to be controversial with industry. He just backs away. He says he is concerned with the credibility of the agency and wants to make sure that we work with people and try to get them to come along. He thinks we should do more things with gentlemen’s agreements. Unfortunately, gentlemen’s agreements don’t work.”
Stilly Francis had a history of ultimately voting with the other board members, even when he disagreed. “While on one hand he seems to favor Boeing over us,” Haueter said, “in the final analysis, he has always signed the bottom line to go with the staff.”
John Goglia, the most colorful character on the board, had decided not to vote in the Flight 427 case. A burly Bostonian, Goglia was a former USAir mechanic who had represented his union in the investigation until he was appointed to the safety board by President Clinton. He said he wanted to avoid any appearance of impropriety. “I’m controversial enough as it is,” he said.
Haueter was worried that the board might reject his report or go with a weak probable cause. If that happened, he had a secret weapon: the families.
The 1996 ValuJet crash had shown how powerful they could be with the safety board. At the final meeting on the crash, the board was minutes away from adopting a probable cause that blamed the FAA and a maintenance company when Goglia suddenly recommended blaming the airline too. The family members—many holding big photographs of the crash victims—applauded and cheered. The emotional demonstration pressured the board to adopt Goglia’s recommendation. Haueter figured he could use the USAir families the same way—if necessary.
If the board still balked, he planned to quit. He’d gotten three job offers from engineering and aviation companies recently and had turned them all down. He and Trisha didn’t want to move away from Washington, and he did not want to desert the NTSB during its toughest investigation. But the offers had shown that he was valuable, and he was willing to take a new job if the board snubbed him.
23. DELIBERATIONS
Boeing and ALPA got one last chance to take shots at each other in their final submissions to the NTSB. The submissions were like closing arguments in a criminal trial. Once they were turned in, Haueter and his bosses would largely cut off communications with the parties. The report would be written in private, and the parties would probably not see it until the final board meeting. The investigators were like cardinals meeting in the Vatican to pick a pope. No one would know the outcome until white smoke poured from the NTSB chimney.
In previous crashes, the submissions had been long, boring letters written in the language of engineers. But Flight 427 was different. The parties were targeting not the NTSB investigators, who had pretty much decided to blame the airplane, but the five board members who would vote on the case. The submissions also had a broader audience: the public and the courts. Each side wanted to reassure customers and juries of its innocence.
Haueter viewed the submissions as an important part of the contentious party system. He knew they would be filled with bias, but he felt they brought a healthy discussion and helped the NTSB sort out the facts. He was eager to see them because he had heard there were internal fights at Boeing and at ALPA.
His moles at Boeing said there were two distinct camps within the company. One wanted to lay the blame squarely on the pilots, saying there was no evidence that the plane had malfunctioned. The other camp was more cautious and wanted to admit that there was no conclusive proof either way.
He also heard of a similar fight within ALPA about whether to mention the mistake the USAir pilots had made in pulling back on the stick. One group wanted to ignore that fact to avoid conceding that the pilots made a big error. The other camp wanted to acknowledge it and then show that the pilots were doing their best, given that they had no training about rudder problems or the crossover point.
The sixty-eight-page ALPA submission had no charts or drawings, but it presented a clear, point-by-point case to show why the plane was at fault. It portrayed Emmett and Germano as helpless victims of a sudden malfunction. It said they fought to regain control but could not overcome the problems of the 737. The union’s report made a circumstantial case that the valve had jammed and reversed. It conceded that there were no marks to prove a jam but said the thermal shock test showed that some jams did not leave a mark.
“The situation was perilous,” ALPA wrote. “The more the aircraft turned to the left, the stronger the first officer’s tendency to apply increased right rudder pedal pressure; the harder he pushed on the right rudder pedal, the more certain it became that the jam would not clear.”
In building its case, ALPA agreed with Brenner’s interpretation of the first officer’s grunts, saying there was a distinct correlation between the grunts and when the rudder would have reversed. But ALPA stopped short of saying the plane should be grounded. Cox and other ALPA pilots continued to say that a rudder hardover was “a low-probability event.” The union also dodged the touchy question of the pilots’ pulling back on the stick. It was mentioned only briefly in the report and was not characterized as a mistake.
The submission from USAir also blamed the plane. Relations between the airline and Boeing had been strained, especially since USAir had decided to buy a fleet of new planes from Airbus, Boeing’s archrival. In its submission, USAir complained that Boeing had not warned anyone about the risks of the crossover point: “Under any circumstances then known to the airline industry, the actions of the crew of USAir Flight 427 were reasonable and correct…. The crew had seven seconds, at most, in which to recognize, analyze and recover from a previously-unknown malfunction.”
The Boeing submission came in a glossy spiral binder and was professionally typeset with headlines and subheads. There were foldout charts of the flight data, drawings of the rudder system, and a table that compared the evidence for pilot error against the possibility of a PCU jam and reversal.
The eighty-three-page report appeared to be a compromise of the two camps within Boeing. It was not as absolute and righteous as the company had been a year earlier with the “Boeing Contribution,” the blatant lobbying attempt that had been such a flop. Whereas that document had been unwavering in its conviction that the pilots were to blame, the new report conceded that there was no proof: “In Boeing’s view, under the standards developed by the NTSB, there is insufficient evidence to reach a conclusion as to the cause of the rudder deflection.”
One section cautioned the NTSB about trying to name a probable cause when there was insufficient evidence. It said there had been a “clamor for a definite and expeditious explanation” and urged the safety board to be careful. “In order to avoid the wrong answer, it is essential that any cause identified by the Board in this accident investigation be supported by facts and evidence. Mere suspicion, inference and conjecture must not suffice.” It quoted Hall as saying that the only thing worse than not being able to solve a case would be to come up with the wrong solution.
The Boeing submission said there was no connection between 427, the Colorado Springs crash, and the harrowing incident on the Eastwind plane. According to Boeing, a new computer analysis of Colorado Springs confirmed what the company had maintained all along: A powerful wind had thrown the plane to the ground. Boeing dismissed Eastwind as a minor event caused by a rudder device that was misrigged and said there was no proof that the Eastwind valve had jammed, just as there was no proof in the USAir valve.
The Boeing report seemed to have been written largely by the engineers who wanted to blame the pilots. It said the NTSB did not have enough evidence to assign a probable cause—unless the safety board wanted to blame the pilots. In that case, there was plenty of evidence.