A friend recently told me that his mother-in-law had been looking back at her and her husband’s financial decisions, identifying the good ones and the bad ones. What, in her view, was one of their worst decisions? “We bought too much insurance,” she said.

Now, it is possible to buy too much insurance. If you buy a $40 item at Best Buy and pay $3 to insure it, you probably bought too much insurance. But that’s not what my friend’s mother-in-law meant. Instead, she was talking about the various insurances on their house and their lives that they had never collected on. That, to her, meant that they had overinsured.

In Insurance and Behavioral Economics, Wharton School economics professors Howard Kunreuther and Mark Pauly and Urban Institute research associate Stacey McMorrow tell similar stories—not from random anecdotes but from market behavior—that show that the sentiment of my friend’s mother-in-law is not unusual. The authors also lay out how insurance markets should work when insurance buyers are trying to maximize their utility and insurance sellers are trying to maximize profits. In many cases, they note, insurance markets work well. In one of the best sections of the book, they put to rest the idea that there is rampant adverse selection in the market for health insurance; this mistaken belief seems to have been behind many health insurance economists’ support of the Patient Protection and Affordable Care Act, better known as Obamacare. But they also find anomalies. On the buyers’ side, they find systematic overinsurance and underinsurance. On the sellers’ side, they find for-profit insurance companies failing to maximize profits. They attribute many of those problems to the systematic mistakes people make in their thinking, mistakes that have been detailed by so-called “behavioral economists.” Adding to the problems, in many cases, are insurance regulators, mainly at the state level, but increasingly—especially with Obamacare—at the federal level.

The authors, I hasten to add, are not as strongly pro-free-market as I am. But theyare sharp, well-informed economists who are experts on insurance markets. So, while in some ways I was disappointed by the lack of a clear bottom line in many of their discussions, their book is full of nuggets of economic wisdom.

Adverse selection problems? | Consider, first, the issue of adverse selection in health insurance. A standard argument by economists is that when insurance companies do not know nearly as much about the health of people they insure as the insured themselves know, they will charge premiums that fail to reflect risk. What’s wrong with that? This asymmetry in information means that premiums for high-risk people are too low and premiums for low-risk people are too high, with the result that high-risk people overinsure and many low-risk people drop out of—or never get into—the market.

But Kunreuther, Pauly, and McMorrow show that when insurance regulators themselves don’t cause adverse selection, it tends not to happen. They write:

Where adverse selection does potentially occur, and to a serious degree, is in markets where regulation prevents insurers from taking into account risk information they surely could have. This “artificial” or “non-essential” adverse selection seems to be most characteristic of health insurance and property insurance markets where “risk rating” is prohibited by law (as in some states in the United States and in all group health insurance) or regulators depress premiums for high-risk exposures for political reasons (as in hurricane insurance in Florida).

In an article I wrote in 1994, I called this “adverse selection by law.” By contrast, they note, “in individual health insurance markets in the United States where risk rating is permitted, adverse selection is absent.” [Emphasis mine.]

Another nugget in their book is their discussion of guaranteed renewability of health insurance. It makes sense that there would be consumer demand for health insurance policies that are renewable at a price that reflects the buyer’s risk profile going in. That way, if something unfortunate happens during the first year of being insured—getting diabetes, for example—that person’s rate does not increase simply to reflect his new circumstances. Where there’s consumer demand, one would expect in a well-functioning market that producer supply will follow—at a price. And it does. The authors point out that individual health insurance “was sold for many years with guaranteed renewability protection, even before being required by law.”

Real market failures | The authors note, though, some systematic anomalies in buyers’ purchase of disaster insurance. Shortly after a disaster, they report, the demand for disaster insurance increases. They give the example of Californians after the 1989 Loma Prieta earthquake. Before the earthquake, 22.4 percent of homeowners in the affected counties had earthquake coverage. Four years later, 36.6 percent had coverage. According to the authors, seismologists say that the probability of another severe quake actually falls after a big earthquake because the stress on the fault has been relieved. So if buyers are informed and rational, the percentage of homeowners with earthquake coverage after the 1989 quake should have been lower, not higher.

Related to this anomaly is the fact that many buyers of flood insurance cancel their insurance after they have gone a few years without a flood. The authors suggest two possible reasons: (1) the longer homeowners go without experiencing a flood, the lower they estimate the probability of a flood, or (2) they think (like my friend’s mother-in-law) that the money they spent on flood insurance during dry years was wasted.

The authors also find that sometimes people overinsure. They point out the case of Aflac’s cancer-only policy, which pays $300 for every day the cancer patient is in the hospital. Using the probability of getting cancer, the authors estimate that, in return for an annual premium of $408, people receive an expected payout of only $77. That’s expensive insurance, and I’ll never again look so fondly on that duck.

Insurance companies are supposed to maximize profits, so you’d think they would make good decisions based on probabilities and the size of payouts. That’s certainly what I thought before reading this book, and it’s true in many, and probably most, cases. But what is striking are the anomalies and how extreme some of them are. Exhibit A is the price of terrorism insurance after the September 11, 2001 terrorist attacks. Before the attacks, for example, Chicago’s O’Hare Airport paid an annual premium of $125,000 for $750 million of terrorism insurance. But after the attacks, the best deal the airport could get was $150 million in coverage for a whopping annual premium of $6.9 million. Even if the loading fee (transactions costs and costs for advertising, employees, building, etc.) was a relatively large 50 percent of the premium, the implied probability of an attack in a year was 1 in 43. And who believed it would be that high? It seems as if insurance sellers, like insurance buyers, can be subject to panic and bad thinking. Of course, if O’Hare actually bought that high-price policy—and neither the authors nor the source they cite make clear whether the airport did buy—the insurance company would have made lots of money, so the transaction doesn’t necessarily reflect panic on the selling company’s side. But the fact that that was the best deal O’Hare could get does suggest that competing insurers were panicked.

Deregulating principles | In a chapter titled “Design Principles for Insurance,” the authors lay out three principles for insurance regulation. Interestingly, although they never come out and say it, all three principles, if followed, would lead to less regulation. The principles are:

  • Avoid premium averaging.
  • Do not mandate insurance benefits not worth their cost.
  • Examine the impacts of crowding-out effects on behavior.

The first principle, if followed, would end one of the key features of Obamacare: its prohibition of insurance companies pricing for risk. That means, in essence, that Obamacare prohibits insurance, and replaces it with socialized financing of health costs that uses private companies as intermediaries.

The second principle, if followed, would mean getting rid of Obamacare’s requirement that health insurance companies pay for various tests—mammograms, Pap smears, and prostate PSA tests, for example—without charging a co-payment to beneficiaries. The authors do not point out either of those two implications of their principles. That’s a disappointment because the book was published almost three years after Obamacare was passed. It seems as if the authors decided to pull their punches in precisely the area where their work could have had one of its biggest effects.

The third principle, “examine impacts of crowding-out effects on behavior,” is not really a principle, but it is a good idea. The authors cite a 2007 paper by Jeffrey Brown, Norma Coe, and Amy Finkelstein that shows that Medicaid’s coverage of long-term care crowds out private purchase of such coverage.

Suggestions | Although it comes out implicitly in the book, it would have been nice if the authors had, upfront, made explicit the three factors that make something an “insurable event.” Those three factors are:

  • The event has a high cost.
  • The event has a low probability.
  • Having insurance does not much influence the probability of the event.

Had they made those three factors explicit, it would have helped not only the reader, but also the authors, identify insurance anomalies. Consider, for example, the Medicare drug plan initiated by President George W. Bush. The authors point out that Medicare’s website “supplies an online decision tool that will tell the beneficiary how much out-of-pocket expense to expect under each of the Medicare drug plans, given information the beneficiary inputs about the drugs he or she is now taking.” I expected the authors to then inform the reader that this certainty of collecting means that the drug benefit is not insurance. The essence of insurance, as I noted, is that it pays for items that people have a small probability of using. But the authors don’t even mention that this is not really insurance. Their exposition is accurate, but what it shows is that the Medicare drug benefit is largely pre-payment for drugs with a large dose of taxpayer funding thrown in.

Probably reflecting the fact that one of the authors, Mark Pauly, earned his doctorate under James Buchanan, one section of the book attempts to put insurance regulation into a “quasi-constitutional” framework. But the attempt is half-hearted. The authors’ plea seems to be more of a plea for “good government” than for any hard-nosed constitutional limits on insurance regulators.

I’m not sure, even after reading the book thoroughly, what the authors’ main goal with the book is. What does come across, though, is the ways in which insurance regulation alters incentives and distorts the economy.