For more than a decade, airbag supplier Takata manufactured what turned out to be ticking time bombs – airbag inflators with a volatile propellant called Phased-Stabilized Ammonium Nitrate (PSAN) – that were placed in vehicles worldwide by the hundreds of millions.
Takata was churning them out at several dollars a bag less than airbags manufactured by rivals, such as Autoliv. According to an August 16 New York Times story, in the late 1990s, Autoliv engineers examined Takata airbags manufactured with PSAN, and concluded that the company could not and would not replicate them, because PSAN was too volatile and dangerous for that application. Autoliv reportedly informed General Motors, which had asked the Swedish supplier to match Takata’s price, and warned other OEMs as well.
In fact, the dangers of PSAN were common knowledge in the industry – enumerated in patents by airbags makers, including Takata – and in the scientific literature. Later, there were the ruptures that made this knowledge manifest to at least some of the Original Equipment Manufacturers (OEMs), in real-world incidents – a BMW in 2003 and a Honda in 2004. The 2006 explosion that rocked Takata’s plant in Monclova, Mexico, was another flare fired into the industry’s consciousness about PSAN’s stability and Takata’s ability to handle it safely. By 2008, when Honda launched its first mini-recall of 4,000 vehicles, OEMs using Takata airbags learned of two more ruptures through Honda’s Notice of Defect and Noncompliance. By 2009, Honda was aware of 11 ruptures and one death. On Christmas Eve of that year, another woman died in an airbag rupture after a fender-bender. In the years to follow, the rolling recalls and changing root cause explanations provided ample warning about Takata’s systemic manufacturing and quality control issues.
What did the OEMs do with this knowledge? Nothing. They continued to buy and install Takata airbags with PSAN propelled inflators. By 2015, Takata had captured 22 percent of the global airbag market and had its biggest year since 2007, with sales topping out at 254 billion yen.
But that is not the story told by Attachment B of the $1 billion plea agreement struck by the Department of Justice and Honda – dumped out on a Friday before a holiday weekend. Takata pled guilty to one count of wire fraud and agreed to pay hefty fines and compensation, submit to the oversight of an independent monitor (who needs no airbag experience) and initiate a corporate compliance program. Other Takata entities escaped charges, including Takata Holdings, the U.S. arm, which performed the testing, investigated the ruptures, and participated in the fraud.
This plea appears to be more obviously designed to preserve Takata’s ability to enter bankruptcy without the costs of the massive recall discouraging would-be buyers of the ailing manufacturer. So far, Takata’s suitors have filed plans requiring that the company rid itself of its liabilities. But the OEMs did not want to bear the brunt of the recall costs. The lion’s share of the $1 billion penalty – $850 million – goes to those “victims.”
In the process, the agreement trades the whole truth for a Reader’s Digest version in which the OEMs hadn’t a clue that Takata had quality control problems until 2015 and likely continues to grapple with today. (Take out your hankies.) According to the plea agreement:
“…had the OEMs been provided with the true and accurate test information and data, the OEMs either would have: (a) insisted that any problems with the PSAN inflators be resolved prior to installation into their vehicles; or (b) refused to put the airbag systems containing the faulty or problematic PSAN inflators into their vehicles.”
To this, we can only say: Nonsense!
It is true that Takata falsified test data showing that inflators using PSAN were not meeting targets for effluent gas and ballistics. The plea agreement described the scheme as the work of a cabal composed of Shinichi Tanaka, 59; Hideo Nakajima, 65; and Tsuneo Chikaraishi. For 15 years, the trio and others hid failures and ruptures during production testing, using the code term “XX” in written communications, as a verb for falsifying data. This mostly involved trimming the extreme ends off the test results to make it seem like there was less variability in the data.
But was this really the missing bit of information that would have led the OEMs to insist that the problems be resolved, or to refuse to buy Takata airbags? A 2015 Takata Engineering Integrity Audit performed at Honda’s behest, by an independent team, suggests it wouldn’t have made much of a difference:
Based upon the information provided by Takata and Honda, and despite the numerous examples of data manipulation, the audit team did not identify any potential safety risks from the audited test reports and test data for three of the four inflator designs (PSDI-5, PSDI-X and PSPI-X) as installed in Honda vehicles. For the PSPI-6 inflators, however, recent information from Takata indicates that after the PV program was completed there were a number of instances of closure burn through or closure erosion in tests of production inflators. This suggests that the burn through problems identified in this audit had not been resolved in the production inflators. This audit did not find any test results with peak pressure results in either DV or PV test programs that could have caused explosive ruptures.
The OEMs – Shoulda, Coulda, Woulda
The three executives, all Japanese citizens, were indicted separately on fraud and conspiracy counts. Bringing criminal charges against auto-manufacturing executives is a relatively new phenomenon. In 2015, some auto-parts higher-ups were indicted on price-fixing charges. Last year, the DOJ indicted six Volkswagen managers in the diesel emissions cheating scandal. Serious consequences for corporate criminality in the auto sector represents some progress. But, as long as there is a financial incentive to ignore or cover up problems, they won’t be the last to be indicted.
And this brings us around to Takata’s co-conspirators – the OEMs. Despite industry knowledge in the late 1990s that Phase-Stabilized Ammonium Nitrate was a risky choice for an airbag inflator – vulnerable to explosions when subjected to temperature changes and moisture exposure – it sure was cheap. Ultimately, GM, Honda, Toyota, and others chose to use it as a significant cost-saving.
And in order for Takata to succeed in maintaining the fiction that PSAN posed no problems, the automakers had to look the other way as each red flag was hoisted and ignore their own supplier quality control procedures. Honda, is perhaps, the most egregious example – as we pointed out in May (Honda Finds Convenient Scapegoat in Takata)
Honda experienced its first rupture in an inflator using PSAN as a propellant in 2004, but passed it off as an anomaly, and waited a year to tell Takata. Even before that, Honda had a rupture with a Takata Unibody-style inflator, which was determined to be the result of a bad weld, indicating a manufacturing problem. By 2009, Honda was aware of 11 ruptures and at least one death – which it concealed from the National Highway Traffic Safety Administration.
A more recent report Takata was compelled to submit to NHTSA under the terms of a 2015 Consent Order, illustrates how involved Honda was in Takata’s design and manufacturing decisions going back to 2000.
Around that time, Takata and Honda were performing sled tests using modules containing PSDI inflators with propellant wafers in the shape of a shark fin and others in the batwing configuration. When a module with a shark fin wafer failed to satisfy one of the test criteria, Honda insisted that Takata use batwings, even though Takata argued that the problem could be remedied and that the shark fin design would produce more consistency in the wafer density, and therefore less variability in ballistic performance. Honda’s February 2000 decision to go with the batwing design came just four months before production was slated to begin, rushing the normal design and production validation processes and leaving Takata insufficient time to test and set up a new production line at its inflator assembly facility in LaGrange, Georgia.
This caused Takata to perform its initial production validation testing using pre-production batwing wafers pressed on a non-production press. And to meet Honda’s volume needs, Takata had to ship a Stokes 340 press – a machine intended for research and development, not mass production – to its Lake Moses, Washington plants to augment the work of its Gladiator presses. The problems with using the Stokes press were later implicated as a possible root cause for the first Honda recalls.
In 2010, Honda and Takata hired outside scientists from Pennsylvania State University and University of California Berkeley to investigate the inflator ruptures. PSU’s analysis of Takata’s PSAN propellant between 2011 and 2014, concluded that it is susceptible to dynamic burning – when it is exposed to sudden pressure increases, such as that created by low-density wafers, it may burn at a much faster rate and at higher temperatures than expected, leading to over-pressurization. Penn State believed this dynamic burning effect contributed to the ruptures and published a paper to that effect in 2012, warning that “the effect of dynamic burning behavior of the propellant needs to be accounted for when designing or analyzing systems that subject the PSAN propellant to high pressurization rates.” Honda was an active participant in all discussions.
Honda’s close relationship with suppliers is legendary – written up regularly in the business press. And even when business professors are not lauding your super-tight collaborations with suppliers, every automaker has a system in place to communicate with and check on component manufacturers, from design to post-production. OEMs, which are ultimately responsible for the end product, can and do require proof of a product’s safety. For example, Design Failure Modes and Effects Analyses (DFMEA) identify potential failures, rank them by severity, and identify fail safes to mitigate them. DFMEAs and other analyses are the bases of Production Parts Approval Process (PPAPs), packets which OEMs use to choose their suppliers. These documents would have predicted that ruptures could happen as a result of over-pressurization and given them a severity 10 – which engineers consider a red flag that the design needs to change. Significantly, according to the plea agreement, Takata had minimal internal controls and compliance systems until 2006. How did that escape the OEMs’ notice when they audited Takata before selecting it as a supplier?
Contrary to the DOJ’s fairytale predictions about what the OEMs would have done with information about bad tests, automakers’ design analyses and audits should have identified and disqualified the use of PSAN in airbag inflators before they went into any vehicle. Later, the public record of ruptures, recalls, and revolving root causes gave them enough reason to galvanize their supplier quality control processes into action. By their own policies, the OEMs should have ceased using Takata as a supplier no later than 2009, when it was clear Takata could not control its systemic issues, with deadly consequences. But none of that actually did happen – because ammonium nitrate is cheap.
The OEMs simply gambled, and their customers lost.
How to Read the Plea Agreement
The Safety Record gets the fact that legally, the wire fraud charge necessitates that the “victim” in the transaction be the entity that paid Takata for its services, i.e, the OEMs. But, we wouldn’t want our readers to confuse the legal victims with the real victims.
Takata airbag ruptures are responsible for 16 deaths worldwide and more than 180 injuries in the U.S. The most recent U.S. death occurred in September. A 50-year-old California woman on her way to get a flu shot died of inflator shrapnel injuries after a front-end collision. The plea agreement sets aside only $125 million to compensate those who have not already resolved their claims against Takata. The payouts will be administered by attorney Kenneth Feinberg, who denied 91 percent of the claims while performing the same service for GM during the ignition switch crisis. This is a pittance, considering that the some of the injuries include serious neck and chest wounds, and facial lacerations necessitating long and costly recoveries.
This restitution fund lives for only five years – the current estimate for completing the recall. After that, any leftover monies go to the government. What happens to people injured after recalls? People are still being injured on a regular basis. And some of the airbag replacements use the same inflators, under the theory that the inflators don’t become dangerous until age six, giving the OEMs the leeway to worry about those inflators later. Even if the deterioration rate of PSAN bought the manufacturers six more years, Takata’s continuing manufacturing problems wipe that margin out. Takata airbag inflators have been subject to four recalls since 2014 for manufacturing defects that could lead to ruptures.
June 2014: GM recalled model year 2013 and 2014 Chevrolet Cruze vehicles after an investigation of a rupture in a 2013 Cruze showed that some SDI-X inflators were made using outer baffle components intended for the PSDI-X inflator. Nissan also issued a U.S. recall, and Nissan and Honda issued foreign recalls.
July 2014: Mercedes-Benz recalled 311 model year 2014 SLK and SL class vehicles because inadequate fasteners could come apart, allowing the diffusor to detach and enter the passenger compartment.
September 2015: GM recalled 395 model year 2015 vehicles after the body bore seal separated from a side airbag inflator (SSI-20 inflator) during lot acceptance testing. The root cause remains unknown.
October 2015: Honda recalled 515 model year 2016 CR-V vehicles after a PSDI-X inflator failed during lot acceptance testing. Takata concluded there was either a stamping issue or a weak metal canister.
- January 2016: Volkswagen recalled 734 model year 2015 vehicles after an SSI-20 side airbag ruptures in a 2015 Tiguan. Takata subsequently combined the GM and Volkswagen vehicles into one equipment recall for SSI-20 airbags in 1,129 vehicles, noting that it had not yet determined a root cause.
And what about the millions forced to drive vehicles with potentially deadly airbags while waiting years for a replacement? Will the recall really be done in five years? GM, for example, reacted to all these revelations about Takata’s fraud, the dangers of PSAN and Takata’s ongoing manufacturing issues by asking NHTSA for a one-year extension to comply with the agency order that recalls of certain model years must be completed by December 31, 2016. GM had flouted the Consent Order for months. At the time it requested the extension, its repair rate for Takata-related recalls was a dismal .17% (only 552 vehicles).
After initially informing customers that GM “strongly recommends that you have this safety recall repair performed immediately” and urging occupants not to sit in the passenger seat, GM’s later customer communications deemed its campaign a “preliminary recall,” and refused to actually repair the vehicles:
“GM continues to believe that there is no immediate danger of inflator rupture in the recalled GM vehicles and that these vehicles are presently safe to drive. You will be notified when there is a need to take action.”
The plea agreement’s section entitled “The Victims” only contains two points devoted to the OEMs who purchased the defective bags based on false data. But what made Takata’s deception truly criminal were the deaths and injuries that resulted long after the doctored validation test results were shoved in a file. And the DOJ spared not a word for those individuals who never had any hand in the decision to roll the dice on PSAN, or saw a ballistic test, but suffered the most severe consequence.
Since 2009, there have been 11 deaths in the U.S. caused by inflator ruptures – people who had their necks sliced open – some in crashes so minor, they should have walked away. Save one 13-year-old, we know their names, and since the government didn’t mention them in the official record, The Safety Record will:
Hai Ming Xu
Hien Thi Tran