April 14, 2003

Don't TREAD On Me:

Automakers and consumer advocates argue over public disclosure of early warning report data.

by Eric Arnum, Editor

The Transportation Recall, Enhancement, Accountability and Documentation (TREAD) Act is proving to be a tremendous lightning rod for consulting and business process re-engineering work. But who thought it also would create whole new areas of practice for attorneys?

Here it is, barely two weeks since auto manufacturers and suppliers began collecting data for their quarterly reports to the National Highway Traffic Safety Administration, and depending upon how certain words and phrases are interpreted, the TREAD Act might or might not result in one of the biggest competitive data bonanzas in the age of motor vehicles. If manufacturers submit comprehensive reports on their manufacturing defect rates and warranty claims to NHTSA as part of the Early Warning Requirements (EWR), and if NHTSA then makes that information public, then everyone from AMR Research to J.D. Power and Associates will know their quality and warranty secrets to three decimal places. But if NHTSA gives in to the unsurprisingly vocal objections of those same manufacturers, we might end up with only slightly more information than is typically public now.

Some of the major automakers are already using longer warranties and claims of increased quality to their own competitive advantage. Others suffer from images of low quality, and still others offer relatively shorter warranties, with lower prices or other incentives compensating for the difference. We suspect they all perform quite a bit of data collection, data mining, and competitive analysis to help them understand where they stand in terms of customer satisfaction, warranty costs, safety, reliability, and quality. But they don't make that information public. Good data on warranty cost per unit of production is hard to find.

But will the TREAD Act change that reality? Will the data in those EWR reports be used by companies such as J.D. Power to put a new face on their quality reports? Or will the data be submitted to NHTSA and disclosed only in cases where recalls are required or fraud is alleged?

Innocent Until Proven Guilty?

Put it another way. NHTSA in this case is functioning as something like a law enforcement agency. Like a good detective, the agency takes lots of notes and makes a lot of comparisons. The automakers that the agency questions about their EWR data are something like precinct commanders called into police headquarters to account for spikes in crime in their neighborhoods. The EWR reports are their confidential notes, providing them with links to the details of each incident, and containing the names and addresses of the people involved. But it's not what the public will see in the daily news reports about crime rates.

The question here is whether we're looking at police investigation records, which would not traditionally be released to the public, or if we're looking at court records, which traditionally would be released unless the judge decides to keep them confidential. If we're looking at police records, then we're looking at data that may or may not allege a crime, and may or may not result in an accusation and/or arrest. In the real world, a person or a company is innocent until proven guilty, or at the very least until they're accused.

Secondarily, one has to worry about how the EWR data might be used. Warranty Week, for one, would not hesitate to derive its own spreadsheet from the data, plotting a detailed graph of warranty claims vs. production data that up until now has not been possible. We'd tell you about the number of warranty claims per thousand units produced, and we'd compare not only manufacturers, but also sedans vs. sports cars, and light trucks vs. SUVs. We'd detail how the four-cylinder engines in Brand X seem to have a problem, or how Brand Y was able to raise the reliability of their anti-lock brake systems in this past model year.

Opportunities for Misuse

However, marketing types might use the data to make advertising boasts (my brand is better than yours), and legal types might use it to bolster their own position in lawsuits already under way against a manufacturer. Indeed, the EWR reports would be required reading for "ambulance chasers" looking for the next big safety issue to exploit. With VIN numbers in hand, it only takes a Carfax lookup to get into a situation where some serious violations of a person's right of privacy are possible. Imagine the cold call from an attorney: "Hi, we understand your husband was incapacitated in an accident when his brakes failed last month."

Warranty Week unequivocally supports public disclosure of early warning data, but at least we're not afraid to disclose our bias. We're not lawyers, and we certainly wouldn't presume to understand the intricacies of the TREAD Act. But we do understand warranties, and we understand that warranty claims data is currently treated as confidential information by auto manufacturers. What little data is released is usually presented in the form of boilerplate paragraphs in financial statements. The only time significant and meaningful financial numbers for warranty expense are released is when something goes wrong, and the accountants determine that the remedy has a material impact on those financial statements.

At the end of August, when the first EWR reports are filed, we could find out production and warranty claims data for the second quarter. In September, we'd get the same data for the past three years. By October, we'd have all manner of formerly proprietary data loaded into an Excel spreadsheet of our own, ready for analysis and output as charts, graphs, and comparisons between models and manufacturers. In short, it could become a jackpot for competitive analysis in the field of warranty claims and cost comparisons. We can't wait.

Until then, we'll have to remain satisfied with an occasional glimpse of warranty cost-cutting successes. Besides believing that manufacturers should release all their warranty data so we can write better articles, we also believe that if warranty is treated as a business process, and analyzed with the assistance of computers and spreadsheets, that manufacturers can boost their bottom line returns. In other words, TREAD Act compliance could turn out to be a public performance of the business process changes already going on now in private among the major carmakers.

Business Intelligence Approach

Subhash Solanki, regional manager of Infogain Corp., said the TREAD Act compliance system his company and client Kia Motors America Inc. tested earlier this year with NHTSA is in reality the new government-facing end of a business intelligence application that was already in place.

He said TREAD Act compliance is much more than just filling in the fields in a few spreadsheets. It has to be done programmatically -- not manually -- because the executives who sign off on the quarterly NHTSA reports have to be in a position to understand, validate, and stand behind the data they submit. Otherwise, they and their corporations might be left open to the significant civil and criminal penalties available to NHTSA under the TREAD Act.

"NHTSA insists that these spreadsheets be signed off at the highest levels," he said, "and for the president or other executives to sign off that the data is accurate, and for each of the business managers to sign off, people need to validate the raw numbers. So if there's a 300 in one of those cells, how do they know it's correct? They need to go back to systems that will help them quickly drill down on those 300 cases."

"When it's done manually, there's a risk," he said, that some call center agent might have picked the wrong code, and the error is being reported to the government as a jump in claims. Perhaps they chose category #19 for tires when they should have picked code 20 for wheels. Managers that sign off on those numbers face the wrath of NHTSA when the error is discovered.

Earlier this year, Kia and Infogain tested their TREAD Act compliance system, submitting dummy data derived from the real data already in Kia's business intelligence application to NHTSA as if it was a quarterly EWR report. Jackie Magnuson, Kia's project manager and its representative in the NHTSA workshops, said the test went rather smoothly.

"We submitted test data in the required format," she said. Although it was dummy data, it was formatted for the Excel spreadsheet templates that NHTSA has published on its Web site.

It worked. Kia Motors uploaded the Excel spreadsheet to the NHTSA file transfer site in February, and by all accounts it passed the syntax editing test. Now it's up to NHTSA to analyze the submitted data.

"We haven't heard back on that piece yet."

Other manufacturers, especially those who make or sell only a few thousand vehicles per quarter in the U.S., might find it just as easy to fill in the spreadsheets by hand, especially if they don't have a business intelligence system in place already on the lookout for spikes in warranty expense or quality issues. The EWR spreadsheets, while they might contain seven pages, are not all that complex. The first page asks for manufacturer contact information. The second page asks for production figures. The next three pages ask for details regarding deaths, injuries, property damage, and consumer complaints. The sixth page asks for the number of warranty claims in each of 24 categories (shown below in a sideways representation for easier reading), The seventh page asks for the number of field reports in the same 24 categories.


EWR Spreadsheet for Light Vehicles

Page Six: Warranty Claims


For automakers who have already invested in testing centers and computer applications that analyze manufacturing defects, consumer complaints, and warranty claims, filling in the EWR spreadsheets will become just another instance of running a quarterly report. No doubt, the reports they already create for internal consumption are already far more detailed than anything NHTSA is asking them to file. For one, they probably attach dollar amounts to each of the repairs or replacements, so executives can quantify the difference between a Code 18 problem with cruise control and a Code 10 problem with the powertrain. But if warranty claims data in general is considered to be proprietary, then monetary expressions of that data must be considered top secret.

Quality Engineering Approach

Take DaimlerChrysler AG as an example. The company opened its Quality Engineering Center five minutes away from its Auburn Hills headquarters in 1997. According to spokesman Sam Locricchio, the QEC functions as both a hospital and a morgue for analysis of quality and warranty issues. When a Chrysler executive, driving around an early pre-production vehicle, brings the vehicle in for something simple such as an oil change, it's hooked up to all kinds of tests and sensors to see what's going on inside the engine.

"If they find any issues, they have a network to get right back to the development team of engineers to let them know," Locricchio said. "If there are any issues, they are caught way ahead of the standard production curve -- before any customer touches the vehicles."

Later on, once a vehicle is in full production, whenever a part is brought into a dealership for replacement under warranty, it ends up at the QEC. All the parts are barcoded and scanned in. A record is created that contains all the relevant data: when was it made, when did it come back, what region was it returned in, etc. Then an "autopsy" is performed, to determine the cause of the failure. Once the engineers spot a trend, they try to narrow the cause down to a supplier, or to a problem in a plant, or to a design issue. All that data is entered into a computer, and ends up in reports that are required reading not only in Auburn Hills, but also over at the home office in Stuttgart-Möhringen.

"Information is king. You can't fix or prevent a problem without a network of data," Locricchio said. "As far as being able to track quality issues, or to try and be proactive as it relates to quality and warranty, certainly we're doing that. And the Quality Engineering Center is a big part of that."

"We want to be best in class by 2007," he said. "I think where we have begun to pull away -- even from the domestics in that regard -- has been in having facilities like this in place, and having a process in place. I think what you hear a lot of vendors talking about in regards to that data is management -- where does it go? It's great to find it, but you better have something in place to do the fixing."

Warranty Boosting the Bottom Line?

Last week at the DaimlerChrysler annual meeting in Berlin, company executives said a 20% cut in warranty expense is going to help the company meet its goal of $2 billion profits this year. While that doesn't give a dollar amount in either warranty cost or savings, it does help one decide how many zeros might be involved in the figure. This is a big issue that for just one manufacturer must run into the hundreds of millions of dollars.

Chrysler vice president of corporate quality Donald Dees previously announced a 21% drop in warranty expense from the model years 2001 to 2002. Those costs are down 50% since 1996, he added. The gains are largely due to improvements in powertrain manufacturing quality and increased testing, he said, and are largely the reason the company was able to offer its seven-year, 70,000-mile powertrain warranty in the first place.

But ask them for an expression of the figures in dollar amounts, or, for that matter, in euros. They won't disclose that, and they won't announce the expense per unit or the number of warranty claims per thousand vehicles sold.

"You want to be best in class," Locricchio said. "But you don't want to give somebody a playbook. We can talk in percentages about how we've improved. But we don't give out what that translates in dollar amounts. I mean, that is competitive data."

It is now, but it might not remain that way.

In the document TREAD Milestones, NHTSA lays out the EWR final rule as it now stands:


Section 3(b) Early Warning Reporting Requirements

    (4) Handling and Utilization of Reporting Elements.
      (C) Disclosure. None of the information collected pursuant to the final rule shall be disclosed pursuant to section 30167(b) of the vehicle safety statute unless the Secretary determines the disclosure of the information will assist in carrying out sections 30167(b) and 30118 through 30121 of the statute.

Legal teams are now carefully analyzing the meaning of those words, and using them to bolster their views that the data either should or should not be released to the public. It's no surprise to find out where they stand. Automakers want NHTSA to keep the data confidential, and consumer advocates want it released to prevent another Firestone/Ford Explorer tire debacle.

Against Public Disclosure

The Association of International Automobile Manufacturers Inc., representing the major automakers, is against the routine release of EWR data. In a July 1, 2002 letter to NHTSA, AIAM director of safety Michael X. Cammisa takes the none/unless conditional statement in Section 3(b)(4)(C) to mean that public disclosures should always be the exception, not the rule.

According to AIAM's analysis, the above-cited Section 30167(b) authorizes the release of defect or noncompliance information obtained by NHTSA to assist the agency in carrying out the safety defect and noncompliance recall program or to assist in the program for maintenance of records of first purchasers of vehicles or tires. But AIAM believes Section 30166(m)(4)(C) of the vehicle safety statute reverses the presumption of information-release.

"In other words, under the existing statute, there is a presumption that the agency will release information as necessary for it to carry out specified programs. However, with the regard to the information submitted to the agency in the early warning reports, the presumption created by the TREAD language is that the early warning information should not be released unless necessary to do so." (emphasis added)

AIAM said NHTSA should presume that EWR data is confidential business information, not to be routinely released to the public. It said carmakers don't release warranty claims statistics to the public now. Instead, it says, "This information is collected for business purposes and is kept secret within the company due to concerns regarding the adverse competitive impacts that would be associated with public disclosure."

Later in the letter, it states, "The comprehensive nature of the TREAD information facilitates manufacturer-to-manufacturer comparisons, readily enabling extrapolations from one manufacturer�s costs and quality levels to those of a competitor. The comparative information would enable one company to use the experience of another to select optimal product design, production process, and pricing strategies, while avoiding the cost and risk that would otherwise necessarily be encountered through trial-and-error."

Others say this is all a red herring meant to distract people from the reason the TREAD Act was written: to prevent another tire safety debacle. They say that section 30167(b) applies only to information that was presumed to be confidential when it was submitted. In other words, when confidential data is submitted, no disclosures should be made of that data unless some greater purpose demands NHTSA break the promise to keep a secret. However, nowhere in the TREAD Act is a promise made that EWR data will remain confidential. Also, in its history, little information has ever been released by NHTSA pursuant to section 30167(b). Besides, there are overriding disclosure rules in the Freedom of Information Act that NHTSA must follow, and these supersede anything in section 30167(b). And manufacturers still have the right to request confidentiality.

In Favor of Public Disclosure

In a Feb. 14 filing regarding a related NHTSA matter under which the agency plans to modify its rules governing the treatment of confidential data, the consumer watchdog group Public Citizen said the clear intent of Congress in passing the TREAD Act was to mandate the release of auto safety data to both the government and the public.

"The law followed upon shocking media and Congressional revelations of secret company memoranda and actions ... The public availability of information in that case would have saved lives and prevented a catastrophic loss of faith in both the industry in general and the reputation of Ford and Firestone specifically. It would be a profound and devastating evisceration of the law passed by Congress to solve this problem of industry cover-up to seal up early warning records, as the manufacturers now are, outrageously, pressuring the agency to do."

The filing goes on to say that at the time the law was being debated, Public Citizen president Joan Claybrook (herself a NHTSA veteran) sought to clarify in telephone conversations with certain members of Congress that they intended for NHTSA to release the early warning data they collected. She also cites comments made by then-President Bill Clinton while signing the bill into law in November 2000 which clearly show his preference for "maximum public availability of information."

Get It In Writing

Anyone who has ever bought or sold an automobile knows not to rely on verbal assurances. Perhaps in politics verbal assurances and telephone conversations matter more when it comes time to clarify the meaning of the written word. And perhaps those who supported TREAD thought those verbal assurances would survive first contact with opposing attorneys. But in matters pertaining to the purchase of an automobile, any promises the salesman says on the dealership floor might be quite reassuring, but they're not worth much unless they're "in writing."


As always, it's up to agencies such as NHTSA to interpret the law, make the rules, and decide what was intended by all these conflicting verbal and written statements. Some clues to NHTSA's leanings on the matter of EWR data disclosure are contained in the Federal Register for July 10, 2002 (Volume 67, Number 132):

Page 45865

VII. How NHTSA Plans To Handle and Utilize Early Warning Information

    A. Review and Use of Information
      We are developing an enhanced data warehouse and data processing system called ARTEMIS--Advanced Retrieval (Tire, Equipment, Motor vehicles) Information System. ARTEMIS will provide for centralized storage of information, include a document management system, use data analysis tools, and facilitate the provision of appropriate information to the public.
Page 45871-72

X. Rulemaking Analyses

    Data Quality Guidelines
      The information that NHTSA is mandated to collect may be made available to the public via the agency's website. The distribution of such data via the agency's website may constitute "information dissemination" as that term is defined ... by the Department of Transportation.
      If a determination were made that the public distribution of the early warning data constituted information dissemination ... then the agency would review the information prior to distribution to ascertain its utility, objectivity, and integrity (collectively, "quality"). Under the Guidelines, any affected person who believed that the information ultimately disseminated by NHTSA was of insufficient quality could file a complaint with the agency.

Back to Part Two    





Sign up for a free subscription to Warranty Week:
     subscribe     change of address     unsubscribe
Value Added Services
ServiceBench
Assurant Solutions Shed The Light
Fulcrum Analytics
Camstar