As early as 2007, the National Highway Traffic Safety Administration (NHTSA) possessed information indicating that an ignition switch defect contributed to fatal accidents involving General Motors (GM) vehicles. But the agency did not issue a recall until 2014.
These are central findings in an audit report released last month by the U.S. Department of Transportation’s Inspector General assessing NHTSA’s internal procedures. The report provides recommendations to improve four of NHTSA pre-investigation processes, concluding that deficiencies in in each step contributed to the delayed discovery of a switch defect that caused at least 124 deaths.
The audit report first proposes ways to improve the process by which NHTSA collects early warning reporting (EWR) data from manufacturers. It then makes suggestions for how NHTSA can improve its collection of consumer complaints. The report also recommends that NHTSA’s Office of Defects Investigation (ODI) revise its process for assessing EWR data and consumer complaints. It also advises NHSTA to standardize its criteria for opening formal investigations that might lead to recalls and fines.
Secretary of Transportation Anthony Foxx has named a team to advise NHTSA on implementing changes to these four pre-investigation processes, and NHTSA Administrator Mark Rosekind has reportedly pledged to restructure the agency along the lines called for in the report.
The auditors recommended that ODI create stricter standards for early warning data that manufacturers must submit to NHTSA each quarter. These data, which include warranty claims and field reports, provide a specific and technical basis for launching an investigation. By expanding data verification processes and compelling manufacturers to comply with requirements, the auditors urge ODI to improve the quality and consistency of early warning data.
According to the audit report, a lack of ODI guidance and verification processes means that manufacturer reports currently vary in usefulness. For instance, manufacturers inconsistently attach “vehicle codes” – markers that designate the type of issue present – to submitted data. Manufacturers may even mischaracterize incidents by avoiding codes like “fire” in favor of less alarming codes like “strange odor.” ODI does not verify or monitor use of codes, relying on an “honor system,” according to the audit report.
NHTSA apparently received 15,600 manufacturer reports on vehicles subject to the ignition switch recall before taking any action. One of these vehicles, the Chevrolet Cobalt, ranked second for injuries involving air bags, but NHTSA reportedly took no action. NHTSA’s failure to verify GM’s use of codes apparently contributed to this shortcoming. In once instance, the audit report says, GM used a code inconsistent with a state trooper’s report when classifying a fatal accident involving a 2005 Cobalt, even though the trooper report suggested that ignition switch and air bag issues contributed to the crash.
The audit report also advises ODI to improve the quality of consumer complaint submissions by providing detailed guidance to consumers filling out complaints, including a list of records to retain and information to include.
Consumer complaints are ODI’s primary means for identifying safety concerns. But the consumer complaint form does not allow individuals to attach supporting documentation and does not define the eighteen part categories with which consumers must label their complaint. As a result, 50-75% of complaints incorrectly identify the affected parts and roughly 25% do not provide adequate information to determine the existence of safety concerns, according to the audit report.
The audit report further states that ODI received 9,266 consumer complaints involving GM vehicles subject to recall before taking action, including three indicating a fatality. One letter stated that “the problem is the ignition turn switch is poorly installed. Even with the slightest touch, the car will shut off in motion.”
The audit report recommends that ODI also reform its process for analyzing early warning data and consumer complaints. In particular, the report directs NHTSA to increase employee supervision, fully train employees, and use standard statistical practices when determining the likelihood of a defect.
In 2014, ODI received 78,000 consumer complaints. A single initial screener reviews each complaint – an average of 330 per day – and forwards only 10% to advanced screeners, according to the audit report. The Defects Assessment Division Chief acknowledged that this initial screener’s supervision is “minimal” and that the screener receives little guidance. Advanced screeners’ supervisory review is similarly “informal.”
The audit report goes on to note that ODI staff do not follow standard statistical practices when assessing early warning data. Furthermore, staff members charged with interpreting statistical results have had no statistics training, and other staff members lack training to stay current with developments in the automotive world.
In 2007, GM provided ODI with a report pointing to the ignition switch as a possible cause of air bag non-deployment in a fatal accident. ODI staff did not note this potential link when documenting reviews because they apparently did not consider all available information in the death and injury report. This incident was not isolated either: according to the audit report, in 2012 and 2013 ODI received 13 reports on 2005 to 2010 Cobalts that GM had categorized as air bag-related, but the office did not launch an investigation.
Finally, the audit report advises ODI to develop standards for the amount of type and information needed to determine whether a potential safety defect warrants an investigation. The report also requests that ODI document its review process.
There currently exists no consensus on the amount and type of information needed for ODI’s division chiefs to open an investigation. Also, ODI does not document the review that does occur. This practice is especially problematic given that ODI does not return to proposals once they are declined for investigation.
The audit report states that these practices cloud review of NHTSA’s failure to investigate the defective switch: the agency did not sufficiently document why it did not investigate the problem in 2007, 2009, and 2010.
Last month, Senate Democrats introduced a bill that would have increased funding for defects investigation from $11 million to $30 million and raised NHTSA’s cap for fines against automakers. But until NHTSA completes its implementation of reforms, a significant increase in funding will likely face opposition. Senator Claire McCaskill (D-MO), who once introduced a bill to double funding for auto safety, has stated that Congress “can’t start throwing money” at the agency until NHTSA has restructured itself to be able to “function like it’s supposed to.” Last week, the Senate passed a version of the proposed reforms as part of a larger transportation bill. Although it includes some of the measures included in the original proposal – for example, an increase in NHTSA’s cap on automaker fines – many Democrats and consumer advocates are concerned that the new bill misses opportunities to improve vehicle safety.
The entire Inspector General audit report can be found here.
This essay is the second in a three-part series, Getting Defective Vehicles Off the Road.