WASHINGTON -- A lone state trooper made the connection in 2007 while studying the wreckage of a Chevrolet Cobalt in a thicket of trees in St. Croix County, Wis.
So did a team of researchers from Indiana University. And so did General Motors' outside attorneys, who advised GM that the circumstances of an August 2012 crash death had a "pretty straightforward" explanation: The ignition switch in the Cobalt was prone to moving out of position, which then cut power to the airbags.
Time and again over those years, engineers and attorneys at GM dismissed or ignored a link that was obvious to experts around them, according to the GM-commissioned report by former U.S. Attorney Anton Valukas.
The report cites this as a pivotal failure because it put the faulty switch on a less urgent track of action, where it stayed until late 2013. Many witnesses interviewed by Valukas said they would have responded differently had they understood the connection.
Even as the death toll mounted, and GM settled lawsuits that shone a light on the link between ignition switches and airbags, the critical connection went undetected within GM, while company investigators pursued other "root causes" for airbag nondeployments, such as electrical malfunctions.
"In short," GM CEO Mary Barra said, "we misdiagnosed the problem from the beginning."
How the company could fail to understand a seemingly simple function in the cars it had engineered and built is one of the most puzzling questions raised by Valukas in the 325-page report -- and it's likely to play a central role in determining how GM gets treated by the government in the years ahead.
It is already prompting a new level of introspection by powerful government officials such as U.S. Rep. Fred Upton, R-Mich., who as chairman of the House Energy and Commerce Committee is a gatekeeper for major reforms.
"It has been more than a decade since we put tough new standards in place so automakers and regulators could quickly spot patterns and fix safety risks, yet this devastating design flaw slipped through the cracks," said Upton, lead sponsor of the last major reform of U.S. auto-safety law, the TREAD Act, in 2000. "It's unacceptable."
From the moment the Cobalt went on sale, it was clear to GM's engineers and safety experts that drivers could cut power to the engine by bumping the key with their knee.
Journalists complained about the problem in 2005, so two top GM safety officials, including Gay Kent, who was among the 15 employees fired last week, took a Cobalt to the company's test center in Warren, Mich. Kent, driving with a long, heavy key chain, knocked the ignition out of position simply by rubbing her jeans against the key fob.
To GM, this tendency represented a customer satisfaction problem, not a safety issue, making it subject to a cost-benefit analysis. When GM engineers considered a fix in 2004 and 2005, they decided the "tooling cost and piece prices are too high," and that "none of the solutions represent an acceptable business case," as documents released by Upton's committee in April first showed.
Over the following years, GM's investigators ignored the risk of the airbags being turned off, which would have made the safety implications of the switch flaw clear-cut. Valukas' report shows that between 2007 and 2012, they received that warning on at least three occasions, and missed or dismissed it each time.
The first warning came in February 2007 from Wisconsin State Trooper Keith Young, a veteran crash investigator who wrote a report on the Oct. 24, 2006, crash of a Cobalt that killed two teenagers.
"The ignition switch on the ... vehicle appears to have been in the accessory position when it impacted the trees, preventing the airbags from deploying," the report said.
GM received the report that March and provided it to NHTSA as part of a quarterly submission. But forensic analysis showed that none of GM's attorneys or safety experts opened it until the Valukas investigation.
Then, in April 2007, in a report to NHTSA on the same crash, researchers at Indiana University raised a similar theory, saying: "It is not known what role, if any, this may have played in the non-deployment of the air bags." GM received the report in June 2012 during litigation, but investigator Brian Stouffer dismissed it.
In October 2012, law firm Eckert Seamans, defending GM in a lawsuit, came to a simple explanation for a crash: "Somehow the key cylinder put the vehicle into 'accessory' mode which would cause the air bags not to deploy."
GM settled the case in December, but Stouffer dismissed the hypothesis for several more months, saying in February 2013 that his long search for a root cause was "stuck."
Bill Kemp, a high-ranking attorney also among the 15 employees fired last week, gave a blunt assessment in Valukas' report. It is "always disappointing," he was quoted as saying during an interview, "when someone outside the company knows more about your product than you do."
• February 2007: In a report on a Wisconsin crash, a state trooper concludes: “The ignition switch on the ... vehicle appears to have been in the accessory position when it impacted the trees, preventing the airbags from deploying.”
• March 2007: GM saves the report in its files and later provides it to NHTSA.
•April 2007: In a report to NHTSA, Indiana University researchers point to complaints about the ignition switch, adding: “It is not known what role, if any, this may have played in the non-deployment of the air bags.”
• June 2012: A plaintiff's expert gives GM the report, but GM's investigator dismisses it.
• October 2012: Law firm Eckert Seamans, defending GM in a lawsuit, says there is a “pretty straightforward” explanation for a crash fatality: “Somehow the key cylinder put the vehicle into 'accessory' mode which would cause the air bags not to deploy.”
• December 2012: GM settles the case. For the next 5 months, GM investigator Brian Stouffer dismisses this hypothesis, saying in February 2013 that his inquiry is “stuck.”
Source: Anton Valukas report to GM board
You can reach Gabe Nelson at firstname.lastname@example.org