Judge Unbias Thyself: Judges Must Perform the Gatekeeper Function and Apply Rule 702 Even When They Are the Fact-Finders, Per the Third Circuit
A primer on Federal Rule of Evidence 702, and relevant Supreme Court decisions from Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993) on, and what it means for a federal court to function as gatekeeper; an example of an abject failure to pass muster as a qualified expert opinion; an object lesson on what a federal court should and should not do when confronted by a proffered expert opinion; a ruling of law on the applicability of Rule 702 to trials to the bench; and a strong hint on whether the standards for evaluation of expert opinion are relaxed in a bench trial. All this packed in 19 pages of a 23-page precedential slip opinion by a panel of the Third Circuit: UGI v. A Permanent Easement, Nos. 18-3126 and 18-3127 (September 19, 2019). And as a bonus, the Court cites to its own recent opinion that decides the issue of the source of law, federal common law or state, for awarding just compensation for private condemnations under the Natural Gas Act. (Spoiler: state law.) The opinion is worth reading and keeping in a file folder on your computer or in a printed version near at hand. Or both.
The substantive issue is a condemnation under the Natural Gas Act (“NGA”) for which just compensation must be calculated. The easements were awarded to UGI Sunbury under the NGA, authorizing it to construct and operate an underground pipeline along a 34.4 mile stretch of land in Pennsylvania. The two property owners in these cases hired the same valuation expert, Don Paul Shearer, who testified in the same way before the same judge.
Shearer’s testimony reads as a parody of expert opinion. It was his view that compensation should be based on his invented “damaged goods theory,” which “holds that markets treat real estate near actual, or even perceived environmental contamination, as a damaged good.” These damaged goods will suffer, he believed, “some long-term stigma on the overall marketability and market value.” Shearer concluded that “any serious potential purchaser is going to discount the price(s) offered for the property based on the stigma and damaged goods aspect after the taking and existence of the natural gas pipeline[.]” He concluded that one of the properties suffered a whopping 40% reduction in value by virtue of the condemnation, while the other property went down by 60%.
How did Shearer arrive at his theory? By working at his grandfather’s appliance shop and seeing that scratch-and-dent merchandise sold for less than appliances without surface damage. Shearer “rounded out his model” by studying the impact on real estate values from the Three Mile Island nuclear incident in 1979 and the Exxon Valdez Alaskan oil spill in 1989, even as Shearer agreed that connecting these mammoth environmental catastrophes to laying an underground natural gas pipeline required “at least some leap of logic.”
And how did he test this theory? Not at all. When asked whether he had any data either way about the purported effect on the value of a property by a pipeline going across it, he stated: “Not yet. I’m prognosticating.” He admitted that there was nothing in his expert report to support his notion that the entire property is affected by the pipeline. His report, in fact, contained not a single example of a property whose value decreased after installation of a natural gas pipeline. Not one. And when it came to the actual valuation, he came up with this lulu: Buyers are simply “going to pay less. How much less? Who knows.” So how, with all this speculation surrounding his theory and the issue of what (if any) loss of value was to be found, did he come up with his figures? “I said to myself five or ten? I thought more than that. Forty? Nah. Nah, it’s—25? I know this sounds kind of crazy.” Crazy, yes. Mr. Shearer could have used a roulette wheel to determine the final figure.
Without any data to support his theory based on anecdotal evidence from appliance pricing and environmental disasters, even Shearer couldn’t trust his opinion: “Come back five years from now and I may not agree with my own opinion. If we find properties that sell with pipelines down the middle of a farm and sold for the same as an identical farm down the road, I’m wrong today. But we don’t have the data.” Indeed, he stated that his theory of permanent stigma “can’t be proven. That’s the problem.” It sure is.
Apart from everything, what is wrong with this cracker barrel wisdom that sounds like it came from Pa Kettle? The clue can be found in Shearer’s description of his method of analysis: “I put this all in my little mixing bowl and I come up with what I thought was common sense reasonable[.]” Shearer is guilty of employing the fallacy known as appeal to common sense. But common sense is often wrong: the world is not flat, heavy objects do not fall faster than light objects, God does play dice with the universe, and an electron goes through two slits at the same time (as long as no one is watching). As the Sage of Baltimore wrote: “There is always a well-known solution to every human problem—neat, plausible, and wrong.” H.L. Mencken, “The Divine Afflatus,” New York Evening Mail (16 November 1917).
Incredibly, the trial judge bought Shearer’s snake oil: “[I]t was ‘inclined to agree with Mr. Shearer that some form of ‘stigma’ attaches to the property as a whole.’” The trial judge further stated that the court “‘agree[d] with Mr. Shearer that there was an overall decrease to the value of the property . . . at least in part . . . due to the ‘stigma’ of being located so close to a natural gas pipeline.’” The Court of Appeals wrote that “that qualified, tentative endorsement of Shearer’s theory was the only factual analysis supporting the valuation awards.” The trial judge awarded a reduction in value of 15%, not 40, for one property, and 30%, not 60, for the other.
How in the world could any individual, much less a judge schooled in the law and the evaluation of expert testimony, have bought into this nonsense? Part of the problem is that the judge made no evaluation of the admissibility of the expert opinion under Rule 702 or the Supreme Court cases setting forth the role of a federal judge as gatekeeper. The judge apparently thought that Rule 702, which gave him “wide discretion when deciding whether those requirements have been met,” simply did not apply because in a bench (as opposed to a jury) trial there is a strong preference for admission. All he had to do, concluded the judge, was to hear the testimony and assess its flaws, all the while admitting the testimony of all the experts. In short, the trial judge abandoned the rigor of Rule 702, ignored its process of assessment, and treated the expert testimony just like any other evidence in a trial to the court: let it in for what it was worth, which, as it turned out, was worth a lot, at least to him.
That was error, the Third Circuit held. Whatever discretion a trial court has in assessing the reliability and fit of proffered expert opinion, it has no discretion not to make that assessment under Rule 702 as guided by case law. The trial court was wrong in assuming, apparently, that because the trial was to the court, it could simply abandon the gatekeeper function. In a close textual analysis of Rule 702, which uses the term “trier of fact,” and comparing that language to Federal Rule of Evidence 403, which speaks of excluding evidence that could “mislead the jury,” the appellate court ruled that the Rule 702 assessment must be applied in all cases, no matter who the “trier of fact” is. The panel quoted the advisory committee’s note to the 2000 amendment: “The trial judge in all cases of proffered expert testimony must find that it is properly grounded, well-reasoned, and not speculative before it can be admitted.”
The appellate court then discussed whether Daubert’s requirements are “relaxed” in the context of bench trials. Two circuits have so held, the Eighth and Eleventh. While not deciding the issue explicitly in its opinion, the panel gave strong indications that it would hold to the contrary and that the standard is not relaxed, citing to the text of Rule 702 that applies, without distinction, to all triers of fact and imposes the same conditions on whether an expert is permitted to testify. The panel also quotes Justice Stephen Breyer that “we ‘judges lack the scientific training that might facilitate the evaluation of scientific claims or the evaluation of expert witnesses who make such claims.’”
This last discussion is important to the panel’s opinion and gives an important clue as to its likely future determination of this issue of relaxation. But there is further evidence to support the idea that the test should not be relaxed for a bench trial. We start with the fact that the very failure to make a Rule 702 assessment was reversible error—and it was not harmless error because the trial court relied upon it. The panel could have stopped there, reversed, and sent the case back to the trial court for further proceedings. Yet, the panel went further and set forth in considerable detail Shearer’s expert opinion. The panel concluded that “Shearer’s testimony lacked reliability and did not fit the case, contravening the mandatory requirements of Federal Rule of Evidence 702 applicable to both bench and jury trials.” But why go further, as it did? One answer could be that the panel wanted to block the use of Shearer’s opinion on remand, end its use once and for all, at least in this case.
But the other reason, I believe, is as likely, and anticipates a future decision on the “relaxation” issue and the standard of appellate review. As described by a panel that included a district judge (from a different district in Pennsylvania than the trial judge) sitting by designation, it is incredible that Shearer’s testimony could be given any credit, admitted for any purpose, or relied uoon for any reason by any trial judge sitting anywhere in the country, with or without application of Rule 702 and Daubert. If a trial judge could go this far off base, not only must the full rigor of Rule 702 apply to bench trials, but its application must be judged by a standard of review as strict as when a jury is the trier of fact: discretion yes, relaxed no. No trial judge likes to be reversed, much less in an opinion that comes close to sarcasm in treating the testimony (I don’t blame the panel for yielding to temptation), a fact I know well in having clerked for a judge in the Southern District of New York. The in terrorem effect of an “unrelaxed” standard and a comparably rigorous appellate review, like this one, may have a positive effect. To paraphrase Johnny Cochran: If the expert opinion doesn’t fit, you can’t admit.
There are other valuable parts of the opinion, new and old, which may be found at slip op. 17. Quoting from another opinion, the panel offers a useful explanation of why “it is impossible to test a hypothesis generated by a subjective methodology[:] because the only person capable of testing or falsifying the hypothesis is the creator of the methodology.” The panel also explains that Shearer’s “foundations” for his theory are nothing more than speculative and have no logical relation to each other: “But the two principles—that consumers prefer undented appliances and property values declined near the Three Mile Island catastrophe—meet only by assumption.” Nice writing, and persuasive.
Perhaps, we should not be too harsh on the trial court, even though, in the cold light of post-trial briefing and a review of the cold record, the Shearer opinion seems absurd. We all have our biases, and the list of cognitive biases, including confirmation bias as the best known, are as long as two arms. (Trust my expert opinion on this!) Training in self-awareness of cognitive biases can get even the most skeptical among us only so far. But work we must, and judges have a heavy responsibility to understand and confront their own cognitive biases.
Here, as set forth above, the bias most likely at work with the trial court involves the logical fallacy of appeal to common sense. Turning to some science writing to support Mr. Mencken’s acerbic observation quoted above, here is what Ball et al. (2018) have to say in the abstract to their “Conflict and dual process theory: The case of belief bias,” found in Current issues in thinking and reasoning (Neys, ed.):
Belief bias refers to the reliable finding that people judge believable conclusions to presented arguments as more acceptable than unbelievable conclusions, regardless of their logical validity. This belief-bias effect occurs irrespective of the strength of presented arguments and regardless of whether tasks involve formal or informal reasoning.
Don’t believe this? Well, then, I have some Reiki lessons to sell you. Just don’t mention Federal Rule of Evidence 702 to the instructor, please.