When Honda Motor Co. said two months ago that it would no longer use Takata as supplier for its airbags, the automaker said that testing data on the airbags had been "misrepresented and manipulated."
Now, newly obtained internal emails suggest the manipulation was both bold and broad, involving open exchanges among Takata employees in Japan and the United States.
"Happy Manipulating!!!" an airbag engineer, Bob Schubert, wrote in one email dated July 6, 2006, in a reference to results of airbag tests. In another, he wrote of changing the colors or lines in a graphic "to divert attention" from the test results and "to try to dress it up."
The emails were among documents unsealed recently as part of a personal injury lawsuit against Takata and obtained by The New York Times. Takata said in a statement that the exchanges concerned only the formatting of data and were unrelated to defective airbags that are under recall.
Takata's airbags, which can explode when they deploy, have been linked to eight deaths in the United States and more than 100 serious injuries and have prompted the recalls of almost 20 million vehicles. Regulators have said that millions more airbag inflaters may need to be recalled unless the company can prove that the propellant they use — ammonium nitrate — is safe.
The National Highway Traffic Safety Administration imposed a $70 million penalty on the Japanese supplier, also citing data manipulation, in November, on the day that Honda dropped Takata.
The emails referred to the testing of airbag inflaters, which contain the propellant and can shatter, sending shrapnel flying into the car's cabin.
Honda would not comment on whether the emails were examples of Takata's misrepresentations. The automaker said it had reached its conclusions after reviewing millions of internal Takata documents. But four airbag experts asked by The New York Times to review the emails said the emails suggested an effort to misrepresent testing data.
"To have these kinds of offhand remarks shows that this is a systemic issue at Takata," said Mark Lillie, a former Takata engineer and whistle-blower.
Takata's practice of manipulating airbag test results dates to at least 2000, just as the company began to introduce a new type inflater.
At first, the U.S. engineers appeared to raise concerns about data manipulation.
An internal report prepared by U.S. employees in November 2000, for example, details discrepancies in airbag test results sent to Honda earlier in the year. In several instances, "pressure vessel failures," or airbag ruptures, were reported to Honda as normal airbag deployments, the report said.
Then in January 2005, Schubert alerted a colleague in a memo that he had been "repeatedly exposed to the Japanese practice of altering data presented to the customer," adding that such conduct was described at Takata as "the way we do business in Japan."
In the memo, Schubert warned that while the fudging of the data had initially not changed the fundamental conclusions of the data, the practice had "gone beyond all reasonable bounds and now most likely constitutes fraud."
"I cannot, in good conscience, fail to report these issues to you," Schubert said, according to the memo.
But a year later, Schubert was urging his colleagues to manipulate the test data, the newly obtained emails suggest.
Specifically, the emails point to concern about tests conducted at an elevated temperature, the experts consulted by The Times said. Together with moisture, high temperatures are known to make ammonium nitrate more volatile.
In response to Schubert's invitation — "Happy Manipulating!!!" — a colleague appeared to be more cautious. "If you think I'm going to manipulate, you really should try and get to know me better," he insisted. Yet he offered: "I would be willing to deviate for running slightly high" in tests at higher temperatures.
"Hey, I manipulated," Schubert responded. The objective, he said, was to help disguise that some of the inflaters performed differently from the rest — a dynamic referred to as "bimodal distribution."
"I showed all the data together, which helped disguise the bimodal distribution," Schubert wrote. "Nothing wrong with that. All the data is there. Every piece," he added. But then he suggested using "thick and thin lines to try and dress it up, or changing colors to divert attention."
In a statement, Takata said the emails were not examples of manipulation.
"Mr. Schubert is referring to the formatting of a presentation, not to changing data, and the emails in question are completely unrelated to the current airbag inflater recalls," the statement said. "In fact, as has previously been reported, Mr. Schubert played a significant role in raising concerns about the past testing data issues referenced in the settlement with NHTSA in early November — issues that will not be tolerated or repeated."
Takata previously said it did not dispute assertions that it had manipulated test data but that any manipulation had been unrelated to the recalls. Schubert did not respond to a request for comment; in the past, he has referred all questions to Takata.
A spokesman for Honda, Chris Martin, reiterated that the automaker was "aware of evidence that suggests that Takata misrepresented and manipulated test data." But he declined to comment specifically on the documents and said that Honda, which has been Takata's biggest customer, would be in a better position to comment, once a third-party audit of Takata's test data, requested by the carmaker in November, was complete. Takata is also the subject of a criminal investigation by the Justice Department.
Chris Caruso, an engineer for General Motors and its parts supplier Delphi from 1979 to 2006, said a bimodal distribution showed the parts being tested were not consistent — generally a requirement for meeting quality standards for automotive safety products.
(BEGIN OPTIONAL TRIM.)
Caruso, who now works as a safety consultant in litigation involving airbag issues, including lawsuits against Takata, said bimodal distribution should result in the rejection of parts by the purchaser.
"Clearly they are saying the data is not good, but if they can manipulate it, they can make it at least appear to be good data," he said. "This is really bad."
(END OPTIONAL TRIM.)
Linda Rink, a former senior staff scientist for Autoliv, another airbag manufacturer, and a longtime consultant for airbag manufacturers including Takata, cautioned that a bimodal distribution did not necessarily mean the part in question was unsafe. And it was unclear from the emails whether the data in this instance was directly related to propellant problems or inflater fragmentation, she said.
Still, she said that any obfuscation of test results was a serious concern.
"If they would disguise inflater data sent to the customer, there is a serious problem with ethics within that company," Rink said. "Having a bimodal distribution requires an explanation and a root-cause analysis, not a cover-up."
(STORY CAN END HERE. OPTIONAL MATERIAL FOLLOWS.)
The documents were unsealed as part of a lawsuit brought by a Florida woman who was paralyzed after her Takata airbag deployed aggressively during an accident in her 2001 Honda Civic in June 2014.
"Takata engineers knew of the dangers of manipulating test data," said Ted Leopold, the lead plaintiff lawyer in the case. "The only thing they did not know was the names of the individuals who were going to be injured or killed, and the date it was going to happen."