DETROIT — General Motors’ investigation into its mishandling of a fatal defect spreads blame among dozens of employees at various levels, but it bluntly says the problem began — and remained unfixed for years — largely because of “a single engineer.”
The report condemns GM’s infamously slow-moving bureaucracy and lack of individual accountability, while also explaining how one relatively low-level employee can precipitate a crisis that now threatens to cost the automaker billions of dollars.
The engineer, Ray DeGiorgio, who has been fired, approved an ignition switch that he knew to be flimsy, secretly authorized changing it years later, then “misled” coworkers who were looking into crashes and customer complaints, the inquiry found. DeGiorgio, a 23-year GM veteran, is mentioned 207 times in the 325-page report by former federal prosecutor Anton Valukas, including in the very first sentence.
He is one of 15 employees dismissed by GM over what CEO Mary Barra last week called a “pattern of misconduct and incompetence.” DeGiorgio’s poor judgment, subpar engineering work and apparently dreadful memory are directly blamed for “catastrophic results,” including 13 deaths that GM has linked to failures of a switch he designed with little oversight.
Ironically, DeGiorgio also may have saved some lives by ultimately fixing the part he had once called “the switch from hell.” But the way he did so also delayed detection of the defect, as deaths and injuries in Chevrolet Cobalts and Saturn Ions continued to mount.
“He actually changed the ignition switch to solve the problem in later model years of the Cobalt, but failed to document it, told no one, and claimed to remember nothing about the change,” Valukas wrote. The report says DeGiorgio made a “deliberate” decision to keep the part number unchanged, “an act that violated GM’s policies and which would throw GM investigators off the track for years.”
DeGiorgio, who was suspended with pay for two months before being fired, hasn’t spoken publicly since the recall and couldn’t be reached for comment last week.
The Valukas report says there is “no question” that DeGiorgio, the switch’s design release engineer since 1999, knew it fell short of GM’s specifications when he approved it in 2002. DeGiorgio, 61, told investigators who interviewed him in March and over two days in May that he didn’t realize it was a safety risk because it worked without incident during testing.
There’s no evidence that anyone else knew the switch was out-of-spec at the time, the report says; neither did DeGiorgio divulge that fact when issues with the part were brought to his attention multiple times, even when directly asked.
Instead, DeGiorgio was consumed by a problem in which cars with the switch were failing to start in cold weather, something the report says was “a personal embarrassment to DeGiorgio.”
While working on a fix for that issue, he also began discussing options for increasing the torque with employees at Delphi Automotive, which manufactured the switch in Mexico. He ended up signing off in April 2006 on a new design that Delphi had created five years earlier.
The work order he created for the redesign listed an electrical change to fix the failure-to-start problem but not the stronger spring and plunger that created more torque. Moreover, the report says DeGiorgio repeatedly told coworkers in 2009 and later that there had been no change that affected the torque in the switch.
“For his part, DeGiorgio remembers none of this and offers no explanation for why he did not change the part number,” the report says.
A little more than a year ago, a deposition of DeGiorgio produced what a lawyer working for GM called a “bombshell” — evidence presented by a plaintiff’s lawyer showing the specific switch modification that DeGiorgio had approved in 2006. Under oath, DeGiorgio said he had “absolutely” never known about or authorized the change.
GM eventually agreed to pay $5 million to settle the case, brought by the parents of a Georgia woman who died in 2010 when she lost control of her Cobalt and skidded in front of another car. Black box data showed that the ignition cut off seconds before the crash.
GM hired DeGiorgio to work at its technical center in June 1991, right after he received a mechanical engineering degree from Lawrence Technological University in suburban Detroit. Engineering was apparently a second career for DeGiorgio, who was 38 at the time. A Detroit native, DeGiorgio had earned a bachelor’s degree in fine arts from Eastern Michigan University in December 1977, but it’s unclear what type of work he did in the subsequent decade.
During the 2013 deposition, DeGiorgio said he had a master’s degree in management from Central Michigan University. In two inquiries by Automotive News, the university said it was unable to locate any record that DeGiorgio had attended or received a degree.
In September 2006, shortly after assembly workers at GM’s plant in Lordstown, Ohio, began using the modified switches, DeGiorgio bought his son a new Cobalt to drive to and from college. It’s unclear whether he got a Cobalt with an old switch — evidence he didn’t believe it to have any serious problems — or one of the first cars with the safer version. In a footnote, the report says investigators don’t know the answer to that either.
The Valukas report sharply criticizes engineer Ray DeGiorgio, who was fired by GM for his role in creating, compounding and prolonging the ignition switch problem.
– Approved substandard switch: “According to DeGiorgio, in 2002, he had the authority to approve the Ignition Switch for production — and did in fact approve it — even though it did not comply with all specification requirements, without needing anyone else to sign off on his approval decision.”
– Didn’t notify colleagues: “No evidence has been uncovered that suggests that any GM employee, other than DeGiorgio, knew, prior to production, that the Ignition Switch approved for the Ion and Cobalt was below specification.”
– Authorized change in design: “On April 26, 2006, DeGiorgio approved the redesigned Ignition Switch by signing what is called a Form 3660. … But the work order to which it was linked … did not include the change to the spring and plunger.”
– Didn’t change part number: “An internal Delphi document dated May 27, 2006, expressly stated that ‘Ray DeGiorgio agree[d] to implement change without changing GM [part number]. He provides his approval.’”
– Diverted investigators: “DeGiorgio told [GM investigator John] Sprague that the MY 2008 change to the Cobalt ignition switch was a change only to the antitheft system. DeGiorgio said there had been no change to the switch that would have affected the power mode shutting off.”
– Does not recall: “To this day … DeGiorgio claims not to remember authorizing the change to the ignition switch or his decision … not to change the switch’s part number. DeGiorgio’s deliberate decision not to change the part number prevented investigators for years from learning what had actually taken place.”
Source: Anton Valukas’ report to GM board
You can reach Nick Bunkley at email@example.com. — Follow Nick on