January 9, 2015
Honda had its turn on the ducking stool yesterday. The Japanese automaker, which had previously disclosed that a data entry glitch led to a failure to report some 1,729 death and injury claims to the National Highway Traffic Safety Administration’s Early Warning Reporting system, got held underwater until it agreed to pay $70 million in fines.
NHTSA’s press release blamed Honda “for failing to comply with laws that safeguard the public.” But it’s far more important to examine the longstanding systemic problems behind the headline. In an audit prepared by the defense litigation firm Bowman & Brooke, Honda explained it as an institutional failure to populate or correctly code certain fields on a form that ultimately went to the Product Regulatory Office, which was responsible for submitting EWR data. Honda learned about the problem from NHTSA’s Office of Defects Investigation (ODI) in late 2011. Yet, the automaker did not correct this administrative error and the regulator did not impose a penalty until three years later — when both were in the teeth of a Category 5 sh-tstorm over exploding Takata-supplied airbags.
Indeed, the EWR system, a key component of the 2000 Transportation Recall Enhancement, Accountability, Enforcement and Documentation (TREAD) Act, was intended to sharpen the agency’s defect detection capabilities and “safeguard the public.” But it hasn’t always worked out that way. NHTSA and the industry regard EWR as little more than a bureaucratic pain-in-the-neck: The first thing the industry did during the EWR rulemaking process was fight to keep the data secret, and the agency’s first order of business was to pare down the amount of information coming to it.
Instead of giving investigators a heads-up on emerging defects, it has functioned as an identifier of failed recalls. Manufacturers report – or not. There’s no robust process for auditing participation. There are no benchmarks for sanctions. In addition to Honda, Congress caught Ferrari failing to report a single claim in a decade. In 2013, Safety Research & Strategies caught a tire maker and two child seat manufacturers that did not report injuries and death claims to EWR; The Safety Record reported that Mercedes is likely severely under-reporting physical damage claims. Honda and Ferrari were fined, because the Hill has taken an interest. But it’s clear that they aren’t the only offenders.
In official publications, NHTSA touts EWR’s role in identifying defects or supporting defect investigations. But in more intimate settings, NHTSA has taken a less sanguine view. To the National Academies of Science committee on electronic throttle controls, representatives of the ODI said EWR was too vague to be really useful:
“In briefings to the committee, ODI analysts noted that the EWR data lack the detail needed to be the primary source for monitoring the fleet for safety defects and that the main use of these data (especially the field reports) has been to support defect monitoring and investigations by supplementing traditional ODI data,” according to the Safety Promise and Challenge of Automotive Electronics: Insights from Unintended Acceleration.
At a recent legal conference for the defense bar, NHTSA’s Chief Counsel O. Kevin Vincent conceded: “Frankly [EWR]’s not early and it’s sometimes really not a great warning, because you have to dig down into the data.”
What Honda’s Fine Tells Us about NHTSA: No Follow-Up
Once a defect reaches the public crisis stage, and everyone starts ripping at the veils of secrecy that have draped a particular problem heretofore from widespread notice, one key factor in the run-up pops up regularly: NHTSA knew pretty early on and did not respond appropriately. The last five years has seen three whoppers: Toyota Unintended Acceleration, loose GM ignition switches and Takata airbags. In each case, the release of documents from entities outside of the agency shows that NHTSA saw the iceberg coming but, for various reasons, did not adjust its course.
In the case of Toyota, NHTSA was hampered by its failure to keep up with the technological changes in automotive safety-critical systems. It didn’t know much about automotive electronics because it had never regulated them and had not laid an institutional foundation of knowledge. If you don’t know how something works, how can you investigate it? Instead, it cowered in its comfort-zone of mechanical defects and driver error, and every ounce of agency energy has been expended in preserving that view to this day.
In the other two defects, ignorance is a little harder to claim. In the case of GM, NHTSA defended its decision by saying that it didn’t understand that the airbags were turned off when the ignition was in the accessory position. But, the connection had already been made by a Wisconsin State Trooper who investigated the October 2006 crash involving a 2005 Cobalt, which killed the 15-year-old driver, Megan Phillips, and passenger Amy Rademaker – and by NHTSA’s own SCI team, which also investigated that crash.
In the Takata airbag saga, we learn several things from that single sentence in Honda’s audit report. In 2011, NHTSA caught Honda under-reporting death and injury claims and took no action. NHTSA had a detailed spreadsheet of injuries and deaths caused by “unusual” and “energetic” airbag deployments. Yet, the Recall Management Division allowed Honda to make Part 573 Defect and Noncompliance reports without ever mentioning deaths or injuries from this defect – so the public was largely unaware. NHTSA, which had no formal open investigation into Honda’s rolling recalls since 2009, did not open a new formal investigation, despite this knowledge of injuries and deaths.
A Brief History of EWR
In 2000, Congress passed the TREAD Act in the wake of the Ford Explorer/Firestone tire fiasco. The EWR system required manufacturers to submit reams of death, injury, property damage, warranty claims and other information to the government on a quarterly basis. The information was supposed to help government investigators identify defect trends.
EWR data consists of aggregates of broad defect categories, such as “powertrain” and “airbags” but no sub-categories. So it is impossible to determine from EWR data even some of the common defect trends, such as non-deploying airbags, or inadvertently deploying airbags, or ignition switch failures. (If NHTSA investigators want information beyond the general categories code they request the underlying documents for review.) Any incident involving a vehicle older than 10 years or a tire older than five years does not have to be reported.
The TREAD Act presumed that most of the EWR data would be public, but after a court battle between the Rubber Manufacturers Association and Public Citizen over the accessibility of tire claims data, NHTSA responded by keeping warranty claims, consumer complaints to the manufacturer, field reports, common green tire information, production data for all except light vehicles, and the last six digits of the vehicle identification number in death and injury claims confidential. Only death, injury and property damage information was included in the public dataset.
In the 2007 Final Rule, NHTSA amended the definition of “fire” to more accurately capture fire related events, eliminated the requirement to produce hard copies of product evaluation reports, and required automakers to update missing vehicle identification number (VIN)/ tire identification number (TIN) or components on incidents of death or injury to a period of no more than one year after NHTSA received the initial report.
In August 2013, the agency added new reporting categories related to emerging technologies. The Final Rule mandated that automakers specify the vehicle type and the fuel and/or propulsion system type in their quarterly EWR submissions, and added new component categories of electronic stability control, forward collision avoidance, lane departure prevention, backover prevention systems for light vehicles and stability control systems for buses, emergency vehicles, and medium-heavy vehicle manufacturers.
How Significant are EWR Failures?
Early Warning Reports offer no laser-like precision, but they are not useless. In a 2012 Preliminary Regulatory Evaluation, NHTSA said: “Since 2004, EWR data have played a role in the opening of 150 Preliminary Evaluations (PEs). Among those 150 PEs, 31 PEs have been opened based on EWR data, with the remaining 119 PEs supported—but not initiated—by EWR data.”
In an undated document entitled Public Release of EWR Data, the agency said:
As of October 1, 2011, NHTSA has used the EWR data in 225 investigations; 68 were launched because of EWR data alone; 157 were prompted by other information but supported by the EWR data.
The Safety Institute, a non-profit organization founded by SRS president Sean Kane, has also demonstrated the use of EWR data through its Vehicle Safety Watch List. “The watch list is the culmination of years of hearing that NHTSA simply didn’t have the data to spot problems sooner” said Kane.
And yet, NHTSA, proudly “data-driven,” remains unwilling to make its peace with this unwanted tool. If the agency caught Honda under-reporting EWR data in 2011, it’s fair to assume this wasn’t the first time it found missing claims from a manufacturer. Honda was in no hurry to fix it. There’s no question that it is Honda’s obligation to make these reports, and it deserves to be sanctioned for breaking the law. But, we submit that but for the agency’s laissez-faire attitude toward EWR, Honda might have corrected it much sooner.
NHTSA’s new administrator, Mark Rosekind, has correctly identified some of NHTSA’s most serious gaps – a need for a formal, structured approach to vehicle defect analysis and recalls. He has vowed to seek more resources and authority from Congress.
Where is EWR in this mix? In a recent Automotive News article, Rosekind recounted a conversation with “a senior employee from NHTSA’s Office of Defect Investigations who suggested using less data could be more effective in spotting defects, rather than more.” At best, it’s a curious stance for any investigator to take. But, when you consider the record, perhaps it’s not entirely crazy. Not knowing doesn’t seem to be NHTSA’s problem so much as not doing, caused by a long history and complex combination of a lack of resources, technical knowledge, process, procedures and priorities.