Predicting Consumer Product Safety Commission Fines: An Analytical Approach

So, you are a manufacturer, distributor, or retailer of consumer products. Most likely, this means your company is regulated by the United States Consumer Product Safety Commission (“CPSC”), and specifically by its rules requiring companies to report potential defects, non-compliances, and hazards in their products. ¹15 U.S.C. §2064(b). That means the CPSC expects you to report these events even if there are very few reported injuries, and in some situations, even if there are no reported injuries at all. ²16 C.F.R. 1115.6.

Companies have an understandable but unfortunate tendency to take the long view on what constitutes a reportable event. CPSC tends to take strong exception to these delays. In the short term, these disagreements often result in product recalls at the “strong encouragement” of CPSC.

A few years later, these same companies will often receive a letter from CPSC’s Office of General Counsel. These letters announce that the company is being investigated for possible civil penalties under the Consumer Product Safety Act (“CPSA”), which is the CPSC’s primary enabling statute.

These days, such letters further tend to demand several thousand pages of documents and emails surrounding the company’s conduct, and sternly advise that the minimum fine for violating the CPSA is $100,000, with the maximum fine being $15 million. Because each individual product sold constitutes a separate violation of the law, ³U.S. v. Mirama Enters., 387 F.3d 983, 987 (9th Cir. 2004). any company that has sold at least 15,000 of an offending product line is theoretically subject to the maximum, $15 million fine. Because much of my practice involves defending companies in CPSC investigations, I have seen many of these letters and unfortunately expect to see many more.

Appreciating the Risk,
and Benchmarking the Outcome

The good news is that no one has (yet) been fined $15 million. The bad news is during his keynote lecture on February 25, 2015, at the annual ICPHSO ⁴International Consumer Product Health and Safety Organization. conference in Orlando, Fla., CPSC Chairman Elliot F. Kaye vowed to use the top end of the penalty range more aggressively.

But even before Chairman Kaye’s directive, in-house counsel and chief financial officers have had little, if any guidance on what penalty to expect beneath that enormous ceiling. This can have a paralyzing effect on a company subject to investigation: It is difficult to reserve funds on such a wide range of possible outcomes, and civil penalties typically are not covered by insurance.

Beyond the problem of reserves, there is the larger problem of effective resolution. A negotiation that begins with the agency having unlimited discretion and the company holding its rosary is not a negotiation that ends well. Companies being investigated by the CPSC therefore don’t just need guidance: They need benchmarks.

The best possible result, of course, is to have the investigation closed without any penalty at all. I’m pleased to say this does sometimes happen. But if push comes to shove, and some penalty appears to be inevitable, companies need to be able to show how their conduct compares to that of past companies, and thereby attempt to adjust the agency’s penalty expectations downward.

Substantive Legal Analytics

How do we get the legal benchmarks that companies need? My recommended answer is Substantive Legal Analytics. Most of you are probably familiar with the field of analytics, commonly defined as “the discovery and communication of meaningful patterns in data.” ⁵http://en.wikipedia.org/wiki/Analytics

Analytics are used by Google to derive consumer preferences, by marketing professionals to calculate the payback of certain advertisements, and by sales departments to target the most relevant customers.

It is long past time that similar analytical tools were incorporated into modern legal practice and decision-making. When those tools are used to forecast or actually improve a legal outcome—as opposed to simply manage logistics or costs—they become Substantive Legal Analytics. These analytics will help answer our question here.

Certainly, one doesn’t need analytics to get a rough idea of what the CPSC is doing: It is easy to total the number of fines handed out each year, or the average fine imposed, or something similar.

But this information is neither sophisticated nor at all useful in a particular negotiation: The average experience of others has no persuasive application to your situation, particularly for regulators focused on your company’s conduct.

With Substantive Legal Analytics, though, we can discover not only where government fines come from, but determine the individual factors that seem to matter most. Equally important, we can isolate those which seem to have little to no effect at all—despite regulators’ frequent insistence to the contrary.

Through a statistical model, Substantive Legal Analytics generates a prediction for each combination of relevant factors—including future cases—and thereby estimates what the appropriate penalty range should be. The essential combination of legal acumen with statistical expertise allows us to separate the signal from the noise, and better forecast what is going on, and what effect it should have on you.

Estimating a Company’s Potential Fine

To do this, I constructed a predictive model consisting of all CPSC penalties handed down over the past 10 years for late or improper hazard reporting. For each fine the CPSC imposed, I noted the amount of the fine (adjusted for inflation), and used accepted modeling techniques to rule in—and rule out—the factors that seemed to be affecting the CPSC’s penalty assessment.

The final model revealed what has traditionally mattered, and not mattered, to the CPSC in imposing fines. Indeed, despite a possible penalty range of $14.9 million, on average the model predicted the actual fine imposed within about $225,000.

This is highly useful information, both for company reserves and for meaningful negotiations with CPSC. The model thus not only provides a reasonable basis upon which companies can estimate their exposure, but also provides a useful tool for negotiations. Instead of simply complaining about CPSC proposals, companies can make a substantive counter-offer, one justified by rigorous statistical analysis, rather than cherry-picked individual cases.

Here are some representative predictions from two fairly typical fact patterns, arising both before and after the passage of the CPSIA:

Some of the things the model reports will not surprise you. For example, the most significant factor by far is whether the conduct being investigated occurred before the passage of the Consumer Product Safety Improvement Act (“CPSIA”), which significantly raised the available penalties. All other things being equal, the same conduct occurring after August 2008 (the date CPSIA was passed) tends to be penalized about $1 million to $1.5 million more than the same wrongdoing before that time. Other important factors include the number of products sold, the number of incidents reported, and the number of injuries. The higher your numbers in any of these categories, the higher you should expect the likely fine to be. For these reasons, it is important to pay close attention to what is being categorized as an “incident” or relevant “injury,” starting back with the initial recall process.

On the other hand, it was quite interesting to see which factors seemed to have very little effect at all. For example, the CPSC has long declared itself highly sensitive to unreported design changes manufacturers make to address safety issues in their products. CPSC tends to view design changes as “stealth recalls” — an attempt to conceal the problem without alerting consumers about possible problems in the existing design.

Yet, the model currently finds the presence or absence of a design change to be almost entirely irrelevant to the size of the fine eventually imposed — even when CPSC specifically reports it as an aggravating factor. It is possible that such design changes do have an overall effect on the CPSC’s decision to pursue a penalty investigation in the first place. But, if a CPSC attorney argues that a design change should drive up the amount of your company’s penalty, the model will support your argument that this should not much matter.

Somewhere in the middle we have the length of time the manufacturer took to report the alleged hazard (“the delay”), which seems both to matter and not matter. The delay certainly matters overall, because once again, if the CPSC did not think a company had delayed unduly in reporting, there would not be a penalty investigation at all. But the significance of the length of that delay is somewhat conflicting.

The sweet spot, if you will, for delay-related penalties appears to be around the 10-month mark; those who have delayed for longer definitely get penalized more, but so do those who delayed only a few months. It’s noteworthy that companies who were heavily penalized after only a short delay tended to have some extreme circumstances: poisonings, deaths, or related aggravated incidents. In this and other regards, then, sooner remains better than later, and companies facing reports of serious injuries should not hesitate to consult their counsel about putting the Commission on notice of the situation.

Analytics: Prospects in Court

The CPSC has no authority to impose civil penalties under the CPSA on its own. Rather, it is required to file suit against the offending company and seek imposition of a penalty in a United States District Court. ⁶Advance Mach. Co. v. Consumer Product Safety Commission, 666 F.2d 1166, 1168 (8th Cir. 1981).

As a practical matter, this distinction makes little difference. Virtually all of CPSC’s penalty investigations, like its “voluntary” recalls, are resolved out of court. Particularly in the era of social media, few companies are interested in a public spat with a federal safety agency, a situation that screams “no win.” The CPSC does occasionally file penalty lawsuits through the Department of Justice, but these cases virtually always seek entry of pre-negotiated consent decrees, and are generally reserved for offenders the CPSC does not trust to obey an out-of-court settlement agreement.

However, there are some situations where actual litigation could occur, such as when CPSC is seeking a particularly extreme penalty, or the penalty is being sought against a company (or successor to that company) with a low public profile. In these situations, Substance Legal Analytics can play a further role.

Let’s start with some first principles. The general rule, as established by the Supreme Court decades ago, is that “mere unevenness in the application of the sanction does not render its application in a particular case unwarranted in law.” ⁷Butz v. Glover Livestock Commission Co., Inc., 411 U.S. 182, 187 (1973). It is understood that agencies can and must have some discretion, so simply comparing one’s penalty to some other case, in isolation, is not a winning strategy, particularly when the agency can itself point to some other case it feels is more representative.

However, federal appellate courts have been receptive to well-supported presentations against unreasonable penalties. The Seventh Circuit sliced in half one penalty imposed by the Commodity Trader Futures Association. ⁸Monieson v. Commodity Futures Trading Com’n, 996 F.2d 852 (7th Cir. 1993). The Eighth Circuit vacated a penalty that would have imperiled the appellee’s ability to remain in business. ⁹Syverson v. U.S. Dept. of Agriculture, 601 F.3d 793 (8th Cir. 2010). And as the D.C. Circuit held, in reprimanding the Securities and Exchange Commission:

[The] Commission must do more than say, in effect, petitioners are bad and must be punished. Petitioners do not stand alone; they are, alas, only two in a long line of enterprises and individuals who have seen fit to conduct themselves in violation of the law of the land… . Faced with a task of such gravity, the Commission must craft with care. ¹⁰Blinder, Robinson & Co., Inc. v. S.E.C., 837 F.2d 1099, 1113 (D.C. Cir. 1988).

Against this backdrop, analytics have the potential to be highly effective in the appropriate case. From a statistical standpoint, the question would be whether the proposed fine is an “outlier”—a penalty so far outside what the model would predict, based on past cases, that it is highly unlikely to be the result of consistent exercise of agency discretion. Various statistical tests exist to help answer this question. A proposed CPSC penalty that flunked these tests should have a much more difficult time surviving scrutiny, thereby motivating district courts responsible for imposing the actual penalty to be more measured in their assessments.

Of course, the goal for most companies is to resolve these allegations outside of court. Here, too, an analytically-sound settlement proposal provides the strongest leverage. The Department of Justice, like most sophisticated litigants, takes pride in being selective about the quality of its cases. A proposed fine that could persuasively be characterized as a statistical outlier may be less attractive than one that conforms to traditional expectations, regardless of what CPSC’s chairman would prefer to see.

Conclusion

Although this article applied Substantive Legal Analytics to CPSC penalty investigations, there is little reason the same approach could not be extended to (a) potential fines imposed by other government agencies (whether federal, state, local, or international), (b) a company’s litigation record involving particular products, jurisdictions, or opposing counsel, or (c) a series of class action settlements or exposures. The potential usefulness of Substantive Legal Analytics is limited only by the acumen of the practitioner and the nature of the issue facing the company.

In the meantime, the takeaways are these:

• Whether facing a government investigation or systematic litigation, look when possible for past patterns instead of relying on intuition.

• When past patterns are available, use sophisticated statistical methods, not raw averages or individual experiences to guide your decisionmaking process.

• Understand that legal analytics require both legal acumen and statistical expertise. A statistician, by herself, does not understand legal risks; a lawyer without statistical expertise will be challenged to provide the right guidance to a statistician. The best path is for the company to work with counsel who are proficient in both areas.