The Texas Supreme Court Issues an Unsettling Opinion on Causation
BIC Pen Corp. v. Carter has been decided, again, and seemingly for the last time. The court previously disposed of the design defect claims and this time the tragic case of a little girl severely burned as a consequence of her five-year-old brother playing with a lighter was before the court on the issues of manufacturing defect and causation.
Given that BIC's own tests of the lighter involved in the accident showed that a flame was produced more readily (because less force was required) than designed the court steamrolled over BIC's manufacturing defect arguments. It gave especially short shrift to the manufacturing variance argument (which actually has a lot of merit). Thus, by the time the opinion gets to causation, a defect for which liability may be imposed having just been found, to the reader things look pretty grim for BIC.
The court focused immediately on the issue that decides causation. Since lighters are designed to be child-resistant and not child-proof (as otherwise lots of adults could not operate them) how do we know that Carter's brother wouldn't have been able to light it anyway? Which is to ask how do we know that but for the defect Carter's brother would not have started the fire? Clearly Carter has to produce some evidence that her brother was enabled by the defect since even if a lighter meets the CPSC standard for lighters it may still be usable by 15% of children his age.
First the court looks at evidence that the brother was developing more slowly than a typical boy his age. The idea is that if the boy was cognitively more like a four-year-old than like a five-year-old then it's more likely that he was indeed enabled by the defect. The court quite rightly decided that it was an apples to oranges assertion because the defect was in a feature meant to physically deter children whereas there was no defect found in the mechanism meant to confound them. Next the court turned to Carter's contention that she could prove her brother was enabled by the defect using the reasoning of Havner.
Carter points to data indicating that low-force-to-operate levels in lighters such as the one in question quadruple the likelihood of a child being able to produce a flame. If Havner requires a doubling of the risk isn't a quadrupling more than enough? Before we get to how the court decided this issue let's run the numbers. As best we can tell there is at least one good test showing that the number of children able to active the lighter goes from 4% to 16% when the physical deterrent to ignition (force) is reduced. Now, assuming that the lighter in question produced a reduction in force necessary to produce a flame similar to the one used in the test then we can calculate how likely it was that Carter's brother was enabled by the defect. Using Bayes' Rule, the likelihoods of ignition for as-designed lighters and reduced force lighters and our knowledge that the lighter in question was in fact defective we can calculated that the likelihood that Carter's brother would not have been able to start the fire (i.e. was enabled by the defect) but for the defect is 75% - a result easily clearing the "more likely than not" hurdle.
Now, the court could have decided that the quadrupling of risk of ignition involved a different level of force (which, in fact, it did) and that plaintiff had zip, zero, nada evidence that the reduction in force at issue (a very tiny one) has ever been shown to double, much less quadruple, risk (which, apparently, it has not). That would have preserved and extended to manufacturing defect cases the rational decision-making approach to uncertain cases established by the court in Havner. Instead, the court decided that probabilistic reasoning can't be used in manufacturing defect cases and gave an extraordinarily weak rationale for doing so.
The court held that "[t]he nature of the injury-causing activities and testing that would have to be done to show causation in this case are not similar to, nor do they pose the practical difficulties posed by, those we considered in Havner. In this case, testing of J-26 lighters posed no unreasonable risk of injury to the test subjects as would have been the case if testing of the drug on humans had been performed under the facts of Havner ... In such instances, tests are done with surrogate lighters that do not pose a risk of harm to the participating children. And testing was also performed on the Subject Lighter itself that posed no risk of injury. Thus we decline to adopt a Havner-type analysis as to causation in this case where manufacturing defects are the basis for the liability claim."
So, because the test subjects tested lighters that didn't contain butane and because the lighter in question posed no risk of burning children when it was tested in a lab post accident, modern techniques for inferring causation aren't appropriate? With all due deference that makes no sense whatsoever.
Maybe the court was worried about permitting Havner-esque analysis in manufacturing defect cases because eventually it would be deployed in a case where the allegedly defective product had been lost or destroyed. If so, a defendant's product could indeed be found to have been more likely than not the cause of the accident even though the product was lost or destroyed. But for that to happen one of two things would have to hold true. Either the defect was responsible for a huge percentage of all such accidents or a huge percentage of the product had been manufactured defectively. In this case, for example, had the lighter been destroyed in the fire plaintiff could not show "more likely than not" (even assuming a quadrupling of the risk) unless she could also show that at least two-thirds of all lighters had the reduced-force defect. Either way, imposing liability when a defect is almost always the cause of an accident or when almost all of the product involved has been defectively manufactured hardly seems an injustice.
Case by case for more than a decade the Texas Supreme Court has worked to produce a coherent and just approach to causation by demanding sound science and utilizing modern decision theory approaches when reasoning out the cases before it. Hopefully, given our fondness for Havner, the true meaning of BIC Pen Corp. v. Carter is exceptio probat regulam in casibus non exceptis.
