A midlevel General Motors engineer named Ray DeGiorgio has emerged as the central figure in GM's decadelong failure to fix a defective ignition switch.
In a Senate subcommittee hearing last week, Sen. Claire McCaskill, D.-Mo., held up a copy of a document signed by DeGiorgio purportedly showing that he authorized a redesign of the faulty switch on April 26, 2006. Seven years later, in a sworn deposition, DeGiorgio denied having known anything about the change.
"He lied," McCaskill said during GM CEO Mary Barra's appearance before the Senate Commerce Committee panel last week, without naming DeGiorgio.
Yet the one-page form that the senator held up in front of Barra and a crush of TV cameras raises many more questions about how that redesign was executed and who else might have had a say in it, according to three former high-level GM engineers who reviewed the document.
The form signed by DeGiorgio is a "validation sign-off," not the document that would have authorized a part redesign under GM's protocols, say the former engineers, who didn't work on the ignition system and didn't want to be identified discussing internal company policy.
The document is an acknowledgment from the supplier, Delphi, that the part met the technical requirements that GM authorized earlier, the sources say. It's one element of a more rigorous process to execute a part redesign -- referred to inside GM as an "engineering work order" -- that would have included input from at least a dozen people across multiple GM departments, they say.
"That validation signoff is just one step at the very end. It puts a bow around everything," one of the former GM engineers told Automotive News. "By itself, it isn't an authorization by GM to change anything."
The former GM engineers say the document raises questions about whether a more extensive paper trail exists that would shed light on the 2006 part redesign, which has emerged as a key turning point.
Research by Automotive News shows that at least nine of the 13 deaths linked to the defective switch happened after the redesign, which fixed the problem of the switch's flimsiness and was first used in 2007 model cars. Had GM recalled the older cars at the time of the design change, those victims likely would have had the chance to get their cars fixed.
A timeline submitted by the company to regulators in February says the "design engineer responsible for the ignition switch ... signed a document approving changes" to the part on April 26, 2006.
The change apparently was made with such scant documentation, including the failure to assign a new part number, that years later, GM engineers investigating crashes and complaints couldn't figure out why the post-2006 switches were different.
The redesign, and DeGiorgio's role in it, drew close scrutiny last week from lawmakers during separate House and Senate subcommittee hearings. At one point, Rep. Steve Scalise, R-La., advised Barra to pull DeGiorgio aside and ask him: "Why did you approve a change in the ignition switch and not bring it to the level of recall?"
Yet DeGiorgio or any other engineer would have faced enormous hurdles to push a part redesign quietly through both GM and Delphi, the former engineers say.
It would have been "unheard of," the second former GM engineer says, for Delphi to have issued that validation form signed by DeGiorgio without there having been an engineering work order authorizing those changes.
That process would have required the OK from DeGiorgio's supervisor, from ranking engineers on several vehicle programs that share the switch and from a team of engineers and purchasing and manufacturing reps, the sources say.
Likewise, if a Delphi engineer had quietly tried to push through a change without GM's authorization "to hide a mistake," the engineer would have faced similar checks and balances within his company, the second engineer says.
Suspicions of perjury likely will dog DeGiorgio whether or not the document he signed represented an authorization. In his April 2013 deposition, he says at least eight times that he knew nothing about the change to the switch.
Barra assured lawmakers that answers to the mystery surrounding the redesign and other questions would be forthcoming from an independent probe commissioned by GM. But so far, GM has said nothing that points to anyone besides DeGiorgio in the 2006 redesign.
Pressed by McCaskill at the Senate subcommittee hearing on whether DeGiorgio lied, Barra answered: "The data that's been put in front of me indicates that."
Barra said DeGiorgio is still employed by GM, but the company has declined to make him available to comment.
Other documents that surfaced amid the hearings last week reinforced how seriously engineers were considering the potential trouble with the now-recalled ignition switch.
A 2005 internal GM e-mail distributed at the House subcommittee hearing referenced problems with the ignition switches and indicated a sense of urgency on the part of Lori Queen, who was GM's vehicle line executive for small cars at the time.
"I'm not sure it's OK to wait," Queen wrote to 17 employees, including DeGiorgio, on Sept. 29, 2005.
The document shows she was responding to the vehicle system engineer for the Chevrolet Cobalt, John Hendler, who said he was "very aware of an issue with 'inadvertent ignition offs' due to the low mounted ignition switch in the steering column and the low efforts required to rotate the ignition."
Hendler wrote that he wanted to use a "new, more robust, increased effort design" being developed for the 2007 Chevrolet Equinox in the Cobalt and Saturn Ion as well but that others working on those vehicles thought the cost of doing so was too high.
In her testimony last week, Barra said any analyses done that used cost as a reason for not fixing a safety issue were "inappropriate." She said GM used to have more of a "cost culture" when the recalled vehicles were developed.
Some GM officials were saying similar things back then, as well. Queen, who is retired and has declined to comment on the recall, told Automotive News in a May 2004 interview that GM wasn't as concerned with pinching pennies on the Cobalt as it had been while designing the Ion several years earlier.
Queen said: "The decisions for that product were made in a different kind of environment."