Yesterday, I started this three-part series with “Insurance Experts Can Misread an Insurance Policy: What a Remarkable Virginia Law Review Study Really Shows.” The article examined the fascinating Virginia Law Review study by Professors Daniel Schwarcz, Brenda Cude, Kyle Logue, and German Marquez Alcala, and particularly the remarkable fact that the researchers themselves initially misread one of the insurance coverage scenarios they were using to test consumer comprehension. This is the second article in the series. The third will be “AI Found an Insurance Coverage Issue the Researchers Missed: What That Means for Legal Scholarship.”

The first article asked what this research really shows. This one asks: What should judges and insurance regulators do with it? My answer is that they should recognize something important. Much of what this research is now demonstrating empirically is something courts have been wrestling with for generations.

Insurance policies are different from most contracts. They are drafted by insurers and insurance organizations with specialized knowledge. Consumer policyholders generally have no role in drafting the language, little practical ability to negotiate it, and vastly less experience understanding how grants of coverage, exclusions, exceptions, definitions, conditions, and endorsements interact.

This is one reason insurance law developed rules of policy construction that often favor the policyholder. Those rules are sometimes attacked as judicial favoritism toward insureds. I argue that criticism misses the point.

The Old Rules Were Developed for a Reason

Contra proferentem, the rule that genuine ambiguities are construed against the drafter, is hardly unique to insurance. But courts have long applied it with particular force in insurance disputes. The new Virginia Law Review article itself notes that construing ambiguities against the insurer is a primary rule of insurance policy interpretation.

Judge Learned Hand explained why almost eighty years ago:

An underwriter might so understand the phrase, when read in its context, but the application was not to be submitted to underwriters; it was to go to persons utterly unacquainted with the niceties of life insurance, who would read it colloquially. It is the understanding of such persons that counts; 1

There is something remarkably modern about that 1947 opinion. Judge Hand did not have an empirical study of 2,500 homeowners. He did not have eye-tracking data, cognitive interviews, behavioral economics, or artificial intelligence to test alternative interpretations.

He had experience and judgment. He understood that the insurer’s technical ability to explain what its language meant did not necessarily establish what the insurance buyer would understand those words to mean.

More than half a century later, the New Jersey Supreme Court discussed the same concern in Zacarias v. Allstate. 1 The court recognized that insurance policy language may require special treatment where it is overly technical, contains hidden pitfalls, depends upon subtle legalistic distinctions, or requires strenuous study to understand. Importantly, Allstate ultimately won that case because the majority concluded the exclusion was sufficiently clear. That makes Zacarias more useful to this discussion, not less. These interpretive rules are not automatic tickets to coverage. They are principles governing how courts determine what the contract reasonably communicates.

Different states draw these lines differently. Some recognize versions of the reasonable expectations doctrine. The Virginia Law Review authors appropriately acknowledge these substantial jurisdictional differences.

I am not suggesting that this new study magically transforms every difficult insurance provision into an ambiguity. Legal ambiguity and consumer incomprehension are not identical concepts. But the research does provide empirical support for why courts developed special interpretive rules in insurance law in the first place.

Disclosure Is Not Comprehension

There is a recurring argument made by insurance company counsel that sounds perfectly reasonable until you examine it closely. They argue that the policyholder received the policy, the provision was printed in it, and therefore the policyholder was on notice of what it meant.  Notice of words and comprehension of their meaning are two different things.

The Virginia Law Review study gives that distinction teeth. In three of the seven principal scenarios, homeowners given the relevant policy language became less accurate about coverage. In another scenario, providing the language made no significant difference. Worse, participants who received policy language tended to become more confident in their answers, including incorrect ones. The authors believe the structure of some provisions may contribute to the problem because readers encounter language pointing toward one conclusion before later language changes the result.

This finding should matter to judges. A trial judge deciding a summary judgment motion may have two excellent insurance lawyers explaining a provision through carefully researched briefs. An appellate panel may have the benefit of oral argument, law clerks, precedent, months of consideration, and familiarity with legal interpretation.  That is not how a homeowner or businessman encounters an insurance policy.

The danger is assuming that because trained lawyers and judges can eventually extract a meaning from a provision, an ordinary policyholder had an equally meaningful opportunity to understand it. The authors make this point directly. They note that judges, because they are experts in contract interpretation, may fail to appreciate the limitations faced by nonexperts. They even suggest that courts might someday consider empirical survey evidence concerning whether typical consumers can understand particular contractual language. The authors candidly acknowledge the objections: surveys can be manipulated, properly designing them is difficult, litigation could become more expensive, and disputes could become harder to resolve early.

I am not convinced we need a consumer survey in every insurance coverage lawsuit. I am convinced judges should appreciate what these surveys are telling us.

Courts do not have to invent an entirely new body of insurance law to respond to the study. Many existing doctrines already reflect the same concern. Ambiguities are generally construed against the drafter. Policy language is ordinarily evaluated from the standpoint of how a reasonable insured would understand it, subject to jurisdictional law. Exclusions are often subject to demanding rules of clarity. Some jurisdictions recognize reasonable expectations in varying forms. Courts have historically been wary of hidden limitations, technical traps, and language whose practical effect is different from what the coverage grant initially appears to promise.

Those doctrines were not created because policyholders deserve to win. They exist because insurers write the contract and possess vastly greater expertise concerning its architecture and consequences. There is an important difference between favoring a policyholder and placing the consequences of unclear drafting upon the party with the power and expertise to draft clearly.

Courts sometimes write as though policyholders not only have a duty to read insurance contracts but should somehow understand them as well. The new research makes it increasingly difficult to defend that assumption as an empirical reality.

Insurance Regulators Should Pay Attention Too

Judges see insurance policies after a dispute. Insurance regulators often see the forms before those disputes ever happen. That makes this research just as important for insurance commissioners and departments of insurance.

The Virginia Law Review authors note that a substantial majority of states impose readability requirements on insurance policies. But they also explain why conventional readability standards may not solve the problem. Making sentences shorter and words simpler does not necessarily cure a policy whose structure leads readers toward one answer before quietly taking them someplace else. Research discussed in the article suggests that improved readability scores do not necessarily produce improved comprehension.

Readable does not necessarily mean understandable. A policy can pass a readability formula and still be cognitively treacherous. Regulators therefore should not view form approval as the end of the inquiry. If empirical evidence repeatedly shows that a particular provision misleads ordinary consumers, the appropriate regulatory question should not merely be, “Does this provision satisfy our readability statute?”

It should also be, “Why are we allowing insurers to sell millions of policies containing language that people predictably misunderstand?”

The authors suggest empirical testing, better disclosures, targeted warnings, stricter comprehension standards, and other possible regulatory responses. Kyle Logue went further in his University of Michigan interview, suggesting that regulators confronted with language shown to be problematic and misleading could simply require insurers to change it.

That makes sense to me. Insurers test pricing. They test underwriting models. They test customer acquisition. They study retention, severity, frequency, catastrophe exposure, and virtually every measurable component of the insurance business.

Why shouldn’t important policy language be tested to determine whether the people buying the product can understand what it means? If a provision repeatedly causes people to reach precisely the wrong conclusion, perhaps the problem is not the people. Perhaps the provision needs to change.

The greatest contribution of this new research may therefore be that it provides empirical support for an insight judges developed through decades of deciding real insurance disputes. Insurance contracts cannot always be treated as though they were carefully negotiated agreements between equally knowledgeable parties.

What I am arguing does not mean courts should manufacture coverage. It does not mean every exclusion is ambiguous. It does not mean judges should disregard clear language merely because the result seems harsh. Instead, it means that the traditional rules protecting policyholders deserve more respect than they sometimes receive.

Those rules recognize who drafted the language, who possesses the specialized knowledge, who could have said it more clearly, and whose understanding should matter when an insurance product is sold to the public. The Virginia Law Review researchers are beginning to measure what generations of judges encountered one disputed case at a time.

With this understanding, judges and regulators should think carefully before saying, “The policyholder should have understood the policy.” The better question may be, “Could the policyholder reasonably have understood it?”

The final article in this series will examine an entirely different lesson from this research. An artificial intelligence system identified an insurance coverage issue the researchers themselves initially overlooked. That raises a fascinating question extending well beyond insurance law: Can AI become a useful second reader that challenges lawyers, professors, researchers, and other experts before we become too confident that our first interpretation is correct?

Thought For The Day

“Our whole system of insurance regulation assumes that if we make companies disclose things clearly, consumers can protect themselves.”
—Kyle Logue (University of Michigan Law School)

1 Gaunt v. John Hancock Mut. Life Ins. Co., 160 F.2d 599, 601 (2d Cir. 1947).

2 Zacarias v. Allstate Ins. Co., 168 N.J. 590, 594–95, 775 A.2d 1262, 1264–65 (2001). (“We give special scrutiny to insurance contracts because of the stark imbalance between insurance companies and insureds in their respective understanding of the terms and conditions of insurance policies… In the first instance, the words of an insurance policy are to be given their plain, ordinary meaning. ‘In the absence of any ambiguity, courts should not write for the insured a better policy of insurance than the one purchased.’ ..However, ‘[i]nsurance policies are contracts of adhesion and as such, are subject to special rules of interpretation.’ …When there is ambiguity in an insurance contract, courts interpret the contract to comport with the reasonable expectations of the insured, even if a close reading of the written text reveals a contrary meaning….‘The objectively reasonable expectations of applicants and intended beneficiaries regarding the terms of insurance contracts will be honored even though painstaking study of the policy provisions would have negated those expectations.’ ”).