A Failure to Start: Lessons Learned from the Valukas GM Report

As a result of the tsunami of public and congressional criticism concerning its asserted failure to recall promptly automobiles that had a defect with the ignition, General Motors commissioned Anton Valukas of Jenner & Block to conduct an internal investigation (the “Valukas Report”). On May 29, 2014, Mr. Valukas issued his report. It was not only critical of GM’s design and engineering departments, but it also attacked the in-house legal department for a number of failings.

Given the high profile nature of the GM problem and Mr. Valukas’ esteemed reputation, the report’s analysis and recommendations will have reverberations for in-house counsel for years to come. This article reviews the findings and recommendations in the report.¹ Some inspire lessons that in-house counsel should take into account for the future. Others, however, reveal misperceptions of the role and abilities of in-house counsel that must be challenged.

1 All assertions of fact contained in this article concerning the ignition switch problem are based solely on the information contained in the Valukas Report and have not been independently verified. This article takes no position on the accuracy of those facts and is not intended to constitute any judgment concerning GM’s conduct in general or the merits of any particular claim.

Context of the investigation

In 2002, GM introduced a design change to the ignition systems of some of its cars. This change led to loss of power under certain circumstances, resulting in moving stalls and circumstances in which airbags would not deploy in a crash. GM now asserts that at least 54 crashes, resulting in the deaths of at least 12 people, occurred when the airbags did not activate due to this problem. GM did not issue a recall until 2014.²

Unsurprisingly, this led to substantial criticism of GM.³ In addition to pervasive media coverage, GM was the subject of congressional hearings, during which both the CEO and the general counsel were treated quite harshly. More recently, GM came under attack from the National Highway Traffic Safety Administration, the entity that is charged with ensuring automobile safety. As of this writing, GM was still struggling with its legal and public relations response.

As part of its strategy, GM retained Mr. Valukas and his firm to do the following: “determine how and why it took so long for GM to recall the Cobalt.” The law firm was asked to complete the task on an expedited basis. Nonetheless, Mr. Valukas managed to issue a 315-page report just two and a half months later, following “hundreds” of witness interviews and the review of “millions” of documents. Even after that Herculean effort, the Valukas Report acknowledges the existence of substantial relevant material that was not reviewed.

Accordingly, a realistic assessment of this endeavor would label it as a substantial, but not necessarily authoritative, analysis. Comparable efforts – for example, the analysis of the problems with the Challenger O-rings — took years, not months, to complete. Nonetheless, it has been generally treated as if it were authoritative by the media, by Congress, and, reportedly, by GM senior management. As a result, it must be taken seriously.

2 In the interests of full disclosure, the author notes that his automobile is under recall for this problem. At the time of writing, however, it has been relatively crash-free for seven years so this fact is unlikely to affect the analysis herein.

3 See, e.g., R. Nader, Unsafe at Any Speed (Grossman Publishers 1965).

Overview of findings concerning in-house counsel

After a lengthy recitation of how the problem came to exist, the Valukas Report turns its attention to the legal department in a section titled: “2006-2014: The Long Investigation Into Root Cause.” The report asserts that in the beginning of 2006, GM attorneys began to see accidents and claims relating to the ignition switch. At the time, the GM legal department had a group of lawyers in charge of safety issues and another group in charge of product liability litigation. Representatives of both groups served on committees charged with the decision to settle product liability cases. The lawyers were supported by engineers and claims administrators. The report generally examines the conduct of the inside counsel during and after various settlement committee meetings, reviews interactions with outside counsel and investigators (including a lengthy exegesis of an external analysis of one particular accident as well as the details of several cases in which outside counsel warned of the possibility of punitive damages), and discusses the internal policies and procedures of the legal department itself.

Although the Valukas Report seems to exonerate the GM general counsel, it is clear that it explicitly or implicitly criticizes the lawyers of the following:

• Failure to identify problematic “trends” or “patterns”;
• Failure to elevate important safety or technical issues to more senior attorneys;
• Tolerance of inappropriate delays in convening meetings on safety issues;
• Failure to recognize the implications of outside investigator conclusions; and
• Construing their roles too narrowly.

These are relatively grave charges. Individually and together, they suggest that at least some of the responsibility for the crashes and fatalities lay with the GM legal department. But are they fair? If so, what lessons can be learned?

Failure to identify problematic “trends” or “patterns”

The most serious allegation in the Valukas Report is that GM’s in-house lawyers failed to identify and report the trend of ignition switch litigation. Had they done so, the report implies that the underlying problem would have been identified earlier and lives saved. The principal job of the various committees on which in-house legal staff served was to generate better claim evaluations and settlement forecasts; nonetheless, some — but not all — lawyers also believed that they had a second function to spot “trends” indicating safety issues. Thus, if the report is correct, this could be viewed as a serious failure on their part. It is not clear, however, that the Valukas Report adequately documents either the objective existence of such a trend, or the subjective ability of the in-house counsel to detect it.⁴

This documentation failure is no mere cavil. A “trend” (or, more generically, a pattern) is not a singular event that can be directly observed; its existence can only be inferred. There is a natural human tendency to see patterns when looking at facts in hindsight when the analyst already knows which events turned out to be important and which were not. It is a bit like reading the last chapter of a murder mystery before reading the book from the beginning; knowing who the culprit is makes it much easier to distinguish real clues from red herrings and important events from distractions.⁵ Accordingly, any retrospective investigation which assigns blame based on non-detection/non-reporting of a pattern or trend should demonstrate that it is not engaging in such retrospective cherry picking.

In addition to this generic bias, there are a number of underlying statistical and analytical problems with the report’s specific assertions that there was, in fact, a “trend” that was detectable at the time.⁶ To be valid, the accusation that in-house counsel failed to identify a trend must do two things: (1) it must demonstrate that there was an actual trend, and (2) it must also demonstrate that the in-house counsel had the ability to perceive that trend if they were acting properly. The report, however, really does not satisfy either of these tasks.

First, the readers of the report are never told how dramatic the pattern of ignition switch accidents was. For example, if the pattern was 5, 25, 125, 625, 3,125 then it would be easy to agree that the trend existed and that the lawyers should have seen it.⁷ However, this does not appear to be the case. Instead, the claims were more like this pattern: 3, 5, 7, 8, 8, 10, 12.⁸ Standing purely on its own, such a pattern might or might not even be an actual trend. Such a sequence might be consistent with purely random events and only look like a pattern to someone looking for one. The report neither discusses this possibility nor provides sufficient raw data to allow for independent analysis. Second, and perhaps more important for the report’s conclusion, even if the latter were a statistically valid pattern, it is highly doubtful that anyone other than a statistician would be able to identify it as such. If so, faulting lawyers for not recognizing and reporting such a trend is misplaced. Unfortunately, a reader of the report cannot discern how well the data supports this profound criticism.

A related problem for the Valukas Report is that the information about the ignition lock issue was embedded in an enormous amount of other data. GM, like any large consumer products company, gets numerous claims all the time. Thus, in any given year it may face hundreds, perhaps thousands, of claims relating to every conceivable aspect of their automobiles. In some years, either for purely random reasons or because of the vagaries of the production process, the total number of claims that GM receives will go up and down. So, for example, if during the relevant period of time there was a general upward trend in the total number of claims received by GM, the increase in any particular type of claim could have been entirely masked. For example, the claims could have looked like this:

Faced with this matrix, it becomes much less intuitive that a lawyer should focus on the ignition lock as the burgeoning problem. Indeed, even if there were not a general upward trend in claims, year to year random fluctuation would suggest that any given type of claim might increase or decrease. Thus, lawyers in the GM legal department would routinely see claims go up, sometimes over several years, not because of any deep-seated component flaw, but merely because such things would, on occasion, be expected to happen. It is the same way that a poker player might occasionally get a sequence of hands that contained a low pair, a high pair, two pair and then three of a kind without assuming that the dealer was cheating. If you play poker long enough, sometimes those things just happen. Given the number of components in a modern car, pure chance would almost mandate that claims arising from some of them would go up for several years in a row.⁹

Accordingly, statistics teaches us that in order to determine whether there was in fact a detectable trend, for each of the relevant years one must (1) establish what a “normal” number of claims relating to any particular component would be expected in any given year given the total number of product liability claims received, then (2) determine how much above (or below) that “normal” number could be attributable to mere random variation, and only then (3) determine that the number of claims actually received exceeded such random variability by enough to signal a real trend.¹⁰

The Valukas Report does not contain any such analysis of the underlying GM product liability data; it does not even acknowledge the importance of such statistical reasoning. Nor does it explain whether and why the GM in-house lawyers could be expected to have the capacity of performing such analysis or had any understanding that such analysis was part of their jobs. It is entirely silent on all these topics and just asserts that a trend existed that was so obvious it was immediately apparent.

Making matters even more complicated is that the Valukas Report focuses not on the underlying accident rates themselves but on the claims made resulting from such accidents. It is well known to in-house counsel, if not to Jenner & Block, that there are fads in personal injury litigation. The personal injury plaintiffs’ bar talks to each other. When one of them successfully settles a case, that fact becomes known. Other plaintiff attorneys are more aggressive in looking for, taking on and pressing such claims.

One of many examples occurred in the 1980s and ‘90s. For decades, there were no cases alleging personal injury from adjacency to high-tension power lines. Then, a few suits were filed. Shortly thereafter, a larger number of suits were filed. They dried up only when science could not document any actual injury from the power lines. Since there was no injury, the increase in suits cannot be attributed to any actual underlying trend. It was entirely due to the somewhat lemming-like behavior of the personal injury bar.

As a result, even in the absence of an actual increase in the number of accidents resulting from a particular component, it is not uncommon for there to be an increase in the number of claims received relating to a particular component. It should also be noted that during this period both GM and one of its principal suppliers were facing financial difficulties that resulted in bankruptcies for each of them. This also could have affected the timing and mix of claims as plaintiff counsel would be expected to accelerate claims to avoid the bankruptcy bar date.

Making things even more complex, the time delay between an accident and a subsequent claim can vary. Thus, an in-house lawyer seeing a rise in claims, without more, might not inevitably draw the conclusion that there was an underlying trend of component failure. Put simply, a pattern in claims presented does not necessarily imply the same pattern of underlying component failure. It may be suggestive but it is not indicative.

All that having been said, there are still lessons to be learned for in-house legal departments. First, it is clear that there was no internal agreement whether trend-spotting was part of the lawyers’ responsibilities. If it was, then some attorneys did not get the memo. If it was not, the lack of unanimity opened the door to the assertion that it was. Accordingly, this report makes clear that the precise role and mission of in-house counsel should be clear and in writing.

Second, particularly with safety and consumer-facing issues, in-house counsel should ensure that relevant information from outside counsel is conveyed in a documented process to corporate safety engineers or similar personnel. The report is correct that outside counsel may have access to information developed in litigation not otherwise available to the corporation. Even if the report cannot justify its claim that GM lawyers did not identify a real trend, that does not mean that the information they had available to them could not have been useful.

4 The report identifies precisely zero formal policies on this latter task but it does refer to one event when a junior lawyer raised the issue of a recall and was informed that the engineers were already working on this issue. The Valukas Report solemnly notes that “[t]his lawyer got the ‘vibe’ that the lawyers had ‘done everything we can do.” There does not seem to be any legal definition of the word “vibe,” but the online Urban Dictionary defines it as a “distinctive emotional atmosphere; sensed intuitively.”

5 See, e.g., A. Christie, Murder on the Orient Express (Collins Crime Club 1934).

6 From a statistician’s point of view, saying the legal department should have concluded that something on the engineering side was amiss from something like frequency of legal claims is equivalent to this making this statement: “The lawyers should have inferred that the number of legal actions was statistically unlikely assuming normal failure conditions.”

7 This is an exponential growth, a trend that is both easy to demonstrate exists and to detect.

8 This is not intended to be an exact recounting of the pattern of claims. It is intended only to be illustrative of the type of pattern presented during the relevant time period.

9 To see why this is true, consider the odds of rolling two dice and getting 2, 3, 4, 5, 6. For any one person, it could happen but would be relatively unlikely. On the other hand, if one observed all the craps tables in Las Vegas for a 24-hour period, there are so many people rolling dice that one would expect to see that sequence (as well as many other similarly unlikely sequences) come up at least once.

10 In fact, the problem is even more complicated than this. According to the report, this problem was not uniform across all models but only occurred in specific platforms. This means that any “true” or expected lawsuit rate must take into account all sorts of other latent variables further upstream (true failure rate of affected vehicles, how prevalent these vehicles are on the road, how much and in what manner these cars are driven by their self-selected owners compared to other cars, etc. It might, for example, be the case that the same defect in Buicks might lead to many fewer accidents than the same defect in a sports car because the demographic of Buick drivers just drive more slowly and carefully and less often leading to fewer accidents than sports car owners. Nowhere in the report is there any discussion of any of these complicating factors or how the legal department could be expected to take them into account.

Failure to elevate important safety issues

The second accusation made by the Valukas Report against the GM in-house lawyers is that they failed to appropriately elevate serious safety issues. The former general counsel of GM asserted that when lawyers became aware of a safety issue, they were to raise it by going to a specific contact person who would then raise it with the engineers. The Valukas Report uses that retrospective claim to describe the law department’s internal written procedures for elevating issues, although confusingly it refers to a document that focuses on violations of federal, state or local laws or regulations. It then returns to interviews for the understanding that the obligation to elevate applied not just to violations of laws, but also to safety issues. As a threshold matter, of course, this allegation depends on the earlier assertion that the lawyers were aware of important safety issues that needed to be addressed. For purposes of this section it will be assumed that this is accurate.

The evidence the Valukas Report brings to bear on this point is decidedly mixed. On the one hand, the report seems to rely on internal guidance that addresses violations of law and regulation rather than safety or liability. It is not clear from the report that the safety issues, at least before 2011 or so, were in violation of any law or regulation. Indeed, recent Congressional hearings have criticized the NHTSA for its ambiguous regulations and praxis. Thus, retrospectively assuming that the safety issues constituted violations of law may be over aggressive.

Even in that context, however, it does appear that there was reluctance among the lower levels of GM attorneys to escalate concerns about safety, and, just as important, concerns about the speed with which GM was addressing such safety issues. Such reluctance can derive from several sources. First, no junior attorney wants to be the “boy/girl who cried wolf.” Without near-certainty that a problem exists, junior attorneys tend to rely on the normal mechanisms of command and control rather than the extraordinary escalation that the Valukas Report asserts was necessary. Second, it appears that the GM multiple-level committee structure left much to be desired in making clear scope of authority and obligation to escalate if normal response is inadequate. While such committees may be useful for dissemination of information and for achieving corporate consensus, they have a similar consequence of diffusing personal responsibility into the ether. At the end of the day, attorneys who have reported a problem to the relevant committee might quite reasonably feel they have discharged their duty. A third problem, not unique to GM, was that the company was facing near-catastrophe for most, if not all, of the relevant period. Between the recession, the financial crisis, bankruptcy of Delphi (one of its key suppliers), and GM’s own bankruptcy and de facto temporary takeover by the government, the senior members of the GM legal department had bigger fish to fry. It is a courageous junior attorney, who, in the midst of this travail, feels emboldened enough to raise what might be perceived as a garden-variety product recall issue.

The Valukas Report does, therefore, present a cautionary tale for all senior members of medium- and large-sized legal departments. It is not enough to issue memoranda reminding junior attorneys of their obligations to elevate problems. This message must be personally reinforced from the highest levels. Thus, when issues are escalated, in the absence of true frivolity, senior lawyers must make clear that this is expected and endorsed. Escalation should be seen as part of the routine not as an exception to it. Even better, senior counsel should make time, hard as that may be to do, to interact with junior attorneys on a regular basis. This may come in the form of regular meetings, meals or educational programs. However it occurs, it permits and encourages junior attorneys to be open in communications, and to understand that senior lawyers do not view them as an impediment to the proper functioning of the law department. It appears that this type of interaction was absent from the GM legal department, rendering the senior lawyers and the general counsel functionally isolated from the activities and concerns of the people who work for them. As such, the Valukas Report may, in this regard, have been too lenient in its findings concerning the senior lawyers on whose watch these problems occurred.

Tolerance of inappropriate delays in convening meetings on safety issues

A third charge leveled against the GM lawyers is that they tolerated inappropriate delays in convening important meetings. The Valukas Report recounts several occasions when meetings did not take place as quickly as the author of that report thought they should have. There are references to “unexplained” delays (with the implicit assertion that such delays were so long that they inherently required explanation), although the report does not describe what would be considered to be a normal length of time to initiate an extraordinary inter-group meeting. Nor does it describe when the lawyers should have determined that the delay was too long. This assertion of failure has both a kernel of truth and an outer layer of unfairness.

As to the unfairness, this results from the inherent bias, discussed previously, of any retrospective report. For example, in hindsight, a six-month delay does seem to be unreasonable. For the lawyers at the time, however, they did not know that the delay would last this long for at least some part of that period. Nowhere is it alleged that they were told at the time the meeting was called that it would not occur for six months; the delay simply accrued. Further, the report seems to indicate that at GM, meetings were rarely prompt and that the corporate culture was rife with temporizing and delay. As a result, the lawyers were not experiencing something unusual or anomalous. From their point of view, the delay was business as usual. Pressing the meeting would have put them at odds with the culture of their clients, a role not entirely customary for those accustomed to be third-party advisors rather than direct drivers of the process.

On the other hand, and again assuming that in-house counsel were aware of a pressing safety issue, the report is not unfair — if not explicit — that the fiduciary duty of the lawyers may have required taking over the process rather than merely observing it. While legal counsel may not be the first choice, or even the second or third choice, to invoke corporate response, at some point necessity imposes that obligation. At GM, however, it does not appear that the lawyers understood that obligation or were prepared to fulfill it.

There is little doubt that GM engineers and safety personnel put key in-house lawyers in a difficult position. It is, however, incumbent on all lawyers to understand that, on occasion, the obligations imposed by their unique position in the corporate structure requires that difficult decisions are made, and that legal counsel must either drive the process or find someone who will. Like baseball umpires who must take control of the field when a brawl occurs, direct intervention is sometimes necessary. The reluctance of the GM legal department to step in does appear to be a deficiency, and one that is instructive to all in-house departments, large and small.

Failure to recognize the implications of outside investigator conclusions

Simply put, this is a questionable allegation at best. The Valukas Report indicts the in-house counsel for not recognizing the validity of outside investigations or studies of the ignition lock problem. In doing so, however, it not only appears to suffer from hindsight bias, but it also misperceives the way that information comes to in-house lawyers and what they do with it.

Almost as a matter of necessity, inside counsel in the United States do not directly review all the documentation available in every claim or case. That is the responsibility of outside counsel.¹¹ Further, the reality of the world of litigation is that every plaintiff ’s lawyer worth his or her salt can find some expert — usually one with impressive credentials — to opine that the cause of the accident was some patent defect of design or manufacture of the product. Further, the inherent risk-avoidance behavior of outside counsel invariably leads them to set forth a substantial risk of loss in almost every case and, if asserted by the plaintiff, the possibility of punitive damages. As a result, in-house lawyers come to expect and discount these predictions of gloom and disaster.

As it turns out, in the case if the ignition lock problem, some of the outside investigators were correct; the GM experts were wrong. That does not, however, mean that the GM legal department — at the time — were negligent or deficient. Indeed, had they substituted their own personal views for those of the client corporation’s inside and outside experts, they would not have been doing their jobs; they would have been usurping the roles of all the technical support staff. Once legal counsel goes down that road, they change from advising the client to being the client, a bold and unaccustomed arrogation of authority.

It is important to note that the Valukas Report is not simply talking about the practices of settlement or litigation. It is asserting that the lawyers themselves should have made the technical and substantive decisions of the company on design and safety. Because the report does not justify this profound change in scope of responsibility, it suggests a failure to truly understand the job of an inside lawyer and how it is carried out on a day-today basis. Once such lawyer has made sure that the outside expert’s views have been reviewed by GM experts and taken into account, his or her job in this regard is complete.

Had the report made a more limited critique, however, it might have been on the mark. It seems clear that GM had a substantial in-house capacity to review and interpret expert reports. Even if inside counsel could not personally evaluate the expert testimony, there is no reason they could not convey those reports to the company’s technical experts. Had they done so, the information the Valukas Report found so important would have been available to those who could understand and use it. This suggests that all in-house counsel should consider whether there should be more internal interaction with adverse expert testimony, particularly when consumer or safety issues are at stake.

11 This may not be the case in every jurisdiction. In some countries, inside counsel have higher obligations although they rarely entirely duplicate the work of outside counsel.

Construing their roles too narrowly

The report asserts that the lawyers contributed to the problem because the lawyers construed their roles too narrowly. This issue has been touched on above, so only a few remaining observations remain.

In-house counsel must maintain a precarious position to be effective. On the one hand, they are part of the corporation and cannot pretend they are not. To be effective, they must be perceived as contributing to the overall goals of the company. Indeed, some would go further and assert that for in-house counsel to be effective they have to enable business strategy and must be seen to support company goals. On the other hand, much of their value comes from the ability to provide an independent viewpoint, even occasionally leaning into the wind to ensure that contrary views are fully aired and explored. In this regard, they are like umpires in baseball. They are critical to the game, but are not direct players.

Once lawyers undertake a direct role in the operation of the company, they lose this position of objectivity. Other members of the corporate team may now view their advice not as independent, but as motivated by the need to justify operational recommendations or decisions. Inside counsel become just another faction to be disputed with no special authority or responsibility.

In the short term, direct involvement may allow the legal department to accomplish some particular goal. In this case, it might have shortened the time for GM to respond appropriately to the ignition lock situation. In the longer term, it is harder to justify or to expect beneficial results. Thus, the Valukas Report may, in its limited scope, be correct, but the failure of the report to address the longer-term problems with such an expansion of responsibility leaves its analysis incomplete.

In any event, there is a lesson to be learned from the report: the legal department must always be clear about its mission and prepared to explain and justify — to the board, to regulators and to the public — why that mission and its limitations are appropriate. It is apparent from the report that the members of the GM legal department were unprepared for this explanation, thereby allowing Jenner & Block to take control of the conversation and the analysis.

Conclusion

The Valukas Report is a valuable and important contribution to the discussion of the duties and responsibilities of an in-house legal department. At the same time, it is not without its flaws, some of them significant. Whether one agrees or disagrees with any specific conclusion or recommendation, however, it must be taken into account in the internal and external dialogue about the obligations of corporate counsel and how they should be fulfilled.

---

Mr. Slavitt would like to thank China Boak Terrell and Isaac Slavitt for their insights and contributions. As always, the views contained in this article are solely those of the author and do not necessarily represent either the views of his employer or the Association of Corporate Counsel.