“Insurance Is Sexy.” Discuss.

Amy FINKELSTEIN: Insurance markets offer an incredibly tantalizing and, dare I say, sexy prospect of providing a measure of certainty in a dangerous and uncertain world. 

Amy Finkelstein is an economist at M.I.T. She didn’t always think of insurance as sexy. But once she converted to this position, Finkelstein — like most converts — has become a true believer. She is the co-author of a new book called Risky Business: Why Insurance Markets Fail and What to Do About It. Today on Freakonomics Radio: the latest installment of the Freakonomics Radio Book Club. We will find out whether insurance is actually sexy; along the way, we explore the many different forms of insurance that do exist — and why some that don’t exist perhaps should.

FINKELSTEIN: I say this with all respect and love for my husband, I’d love divorce insurance.

Everything you always wanted to know about insurance but — if you’re like most people — just didn’t care enough to ask.

*      *      *

Amy Finkelstein wrote her new book, Risky Business, with two other economists: Liran Einav from Stanford and Ray Fisman from Boston University. Finkelstein says her “love affair with insurance” — her words — began when she was a junior staffer on the Council of Economic Advisors.

FINKELSTEIN: It was in 1997, 1998, during the Clinton administration, and Janet Yellen was the chair. One of the really fun things about the Council of Economic Advisors is, it’s quite small — you know, there’s 20, 30 people there — but it has the same scope of things it needs to advise the president on as the Treasury Department, which has tens of thousands of employees. So if you’re a eager young staffer with no life — and I qualified on all those dimensions — you get to work on all kinds of things. When I looked back at what were the things that really excited me, I had gotten to work on natural-disaster insurance, on automobile insurance, on unemployment insurance. The common denominator, it turned out, was insurance. 

From there, Finkelstein got her Ph.D. in economics from M.I.T. She has since won major awards for her research, most of which is about health insurance. If her name or voice sound familiar, you may remember a Freakonomics Radio episode from 2015 called “How Do We Know What Really Works in Healthcare?” Now I asked Finkelstein to name her favorite kind of insurance, or perhaps the most unusual type she’s come across.

FINKELSTEIN: They’re pretty idiosyncratic. So Lloyd’s of London makes a lot of money doing sort of boutique insurance. Betty Grable in the 1930s wanted to insure her legs. The lead singer of KISS insured his remarkably elongated tongue. There was space-satellite insurance several years before the first moon landing. So I find those types of boutique insurances quite fun and funny. 

Stephen DUBNER: And if Amy Finkelstein were to take out a boutique insurance policy that only Lloyd’s could underwrite, what would it be? 

FINKELSTEIN: I always worry I’m going to lose either my sense of humor or my ability to write economics papers. Or maybe the two go hand in hand. When people stop finding me funny, I’d like a big payout. Now, perhaps that day has already come, so I can’t buy the insurance. 

DUBNER: How many types of insurance do you have? 

FINKELSTEIN: I have health insurance, I have life insurance. It’s on my to-do list to buy long-term care insurance. I have auto insurance and I have homeowner’s insurance. 

DUBNER: Okay, that’s it?

FINKELSTEIN: I believe so.

DUBNER: You don’t have any umbrella, or excess liability?

FINKELSTEIN: Ah, I do have — yes, I have an umbrella policy.

DUBNER: Do you also have perhaps some forms of insurance that are kind of automatic, that you don’t even think about? Any unemployment, disability? 

FINKELSTEIN: Well, we all have government-provided unemployment insurance. Through my employer, M.I.T., I also get additional disability insurance. 

DUBNER: How about social security insurance?

FINKELSTEIN: Yup. We all get that. 

DUBNER: Medicare. 

FINKELSTEIN: I’ll have it. But I’m not that old yet. 

DUBNER: Right, but you’re paying into it now. 

FINKELSTEIN: For sure. 

DUBNER: So really, you started by naming maybe five or six and it got us to eight or nine. Do you have a pet, and/or pet insurance? 

FINKELSTEIN: We have two guinea pigs and a gecko, but no insurance for them. My coauthor Liran has a dog and didn’t buy pet insurance at first because it’s so expensively priced, precisely because it’s priced for the people who are planning to get a lot of care. When he discovered, when their dog got ill, just how much his wife was willing to spend to try to save the dog, he thought maybe we should get insurance, but because the dog now had a preexisting condition, the insurer wouldn’t let them. 

DUBNER: So, insurance is one of those topics that makes the average person’s eyes just glaze over. Did you used to be like that before you fell in love with insurance? 

FINKELSTEIN: Probably yes, because I fell in love with it for a very particular reason. The sort of cartoon version of economics is the magic of the markets, Adam Smith’s invisible hand. And of course, most people realize that’s a little extreme and cartoonish. We understand that monopoly firms can set exorbitant prices, there can be bank runs, polluting firms can pollute, and we have government policies to deal with all of those. But one thing I hadn’t realized ‘til I started working in economics is there’s another type of market frailty that’s really important, that’s the subject of a lot of government policy, but that most people just don’t seem to be as aware of. And that’s the problem of selection. And it’s front and center in insurance markets.

Okay, let’s define some terms here. When economists talk about “selection,” they’re inviting us to look beneath the surface of a given data set, to see how the data were put together — or selected. And, just as important, to see what was left out. Imagine you are an attorney defending a securities firm accused of fleecing its investors. You might like to select an entire jury made up only of securities-firm C.E.O.s. The prosecutor, meanwhile, might like to stack the jury with small-time investors. The job of the judge and the judicial system is to make the jury pool more representative of the general population. When it comes to insurance companies, you can see why selection is so important. Let’s say you are selling health insurance. You’ll make more money if you can sign up healthy people and avoid sick people — but this gets tricky, because sick people have more incentive to buy your product.

This is what economists call “adverse selection” — when your selection process gets overrun by exactly the people you don’t want. If you’ve ever lived in a New York City co-op apartment building, you are familiar with adverse selection: the people who most want to serve on the co-op board are the exact people you probably don’t want on the board: bossy, nosy, aggressive, self-interested. As Amy Finkelstein said, selection is “front and center” in insurance markets. It becomes a cat-and-mouse game where insurance companies try to pick the low-risk customers, while the high-risk customers try to persuade the insurers that they’re low-risk. So how do you solve that? With health insurance, at least, Finkelstein cites a famous paper by the economist George Akerlof, called “The Market for ‘Lemons’: Quality Uncertainty and the Market Mechanism.” Here’s Finkelstein reading a passage from her book Risky Business: 

FINKELSTEIN, EXCERPT: Absent a requirement to be insured, Akerlof argued, sick customers (lemons) would flock to the market for health insurance, driving up prices for healthy individuals, who would then opt out of buying coverage. Akerlof gave a shout-out to Medicare — what was then a brand-new compulsory public health insurance program for the elderly. He argued that this type of mandate was exactly what the doctor ordered to fix the adverse-selection problem in health insurance. It is this sort of reasoning that led most economists, including ourselves, to the intuition that mandates are the best and most straightforward solution to selection problems. From this perspective, the health insurance mandate in the 2010 Affordable Care Act offered the canonical approach to addressing selection.

The Affordable Care Act, as you likely know, is sometimes called Obamacare.

FINKELSTEIN: I think if you’d asked me before we had the experience of the Obamacare mandate, I would have said, “Mandates are the classic solution to the selection problem. We’re all done.” One of the motivations for this book came about listening to the oral arguments over the Obamacare mandate, and hearing then- Supreme Court Justice Antonin Scalia ask, “If you can mandate that people buy health insurance, can we also mandate that they eat their broccoli?” He was obviously being rhetorical and conjuring up a paternalistic, intrusive state. But I think for many people, the distinction between insurance and broccoli is lost. The key difference from our perspective is the supermarket doesn’t care which customers buy its broccoli. Its profits are the same whether people buy it and put it to good use or throw it out. Whereas the health insurer or any insurer cares greatly who its customer is, and they’re worried about getting the high-cost customers, or making sure that they can price it accordingly. So the canonical solution is a mandate. And that’s what I taught for decades to my students. What happened in real life was both a win but also a cautionary tale for that theory. So mandates do work the way economists like George Akerlof, who wrote about this in his Nobel Prize-winning work, “The Market for Lemons,” you know, predicts. They do get the healthy into the market and help with the adverse-selection problem. The main problem with a mandate — let’s go back to the broccoli example. Anyone who has a young child knows that telling them to eat their broccoli doesn’t make it so. And the same is true with the mandate. It turns out in the real world, unlike in our textbooks, simply saying that everyone has to have health insurance or auto insurance doesn’t mean that everyone buys it. And really, if you want to get the healthy or the low-risk into the market, you have to either put a big penalty on it to get people into the market. Or the opposite of that — if you don’t want to give them a stick, give them a carrot, and subsidize them in.

DUBNER: And how successful would you say overall that combination of stick and carrot has been? 

FINKELSTEIN: The glass is both half-full and half-empty. On the one hand, the share of people in the United States without health insurance has fallen in half. On the other hand, there’s still 30 million Americans without health insurance. 

DUBNER: And what about the cost of healthcare overall? I mean, for the past, gosh, 10 or 15 years, I’ve been hearing that the whole industry is approaching 20 percent of G.D.P. I don’t think we’ve actually hit that yet. It’s obviously very hard to say that X causes Y in a sphere as complicated as healthcare. But how do you think the mandate and the Affordable Care Act generally have done at, if not decreasing prices, at least reaching some predictability or equilibrium in pricing?

FINKELSTEIN: Yeah. I think this is something that gets very confused in policy debate, that whenever anyone says, “We want to expand health insurance coverage,” they say in the same breath, “and rein in health care costs,” because those are the two big problems in the U.S. healthcare sector — the difficulties with insurance and the sky-high costs. But unfortunately, anything you do to increase health insurance coverage is going to increase healthcare spending. Because what people do when they have health insurance is they use more healthcare. So in some sense, I think we have to separate those problems entirely. 

DUBNER: On the other hand, there’s always the argument that if people are insured, there will be more preventive care, especially if healthcare providers are incentivized to provide preventive care, then there will be less procedures down the road and so on. Shouldn’t there be savings in that cycle? 

FINKELSTEIN: Yes, that’s always the argument, and hope springs eternal. But I have yet to see the evidence for it.

DUBNER: How much has health insurance in the U.S. been complicated or even perverted by the fact that it’s usually tied to employment, unlike just about every other country on Earth?

FINKELSTEIN: That’s a great question. I think there’s two answers. One is, if you go back to the root problem of selection, which may be a reason the government is involved in health insurance or any insurance, the employer suddenly sounds like a really good idea because if you just ask people to buy health insurance, it’s the sick who are going to go buy it. But if your employer is offering it — and you’re not choosing your job just based on your health status, so you’re not going to select your employer based only on whether they’re offering more or less generous health insurance.

DUBNER: Plus which, anyone who’s capable of holding down the kind of job that provides you healthcare insurance is, prima facie, a lot healthier than many people who don’t have healthcare insurance. 

FINKELSTEIN: That’s definitely true. Although, they are also covering spouses, and dependents who could be quite sick. On the other hand, having your employer double as your insurer creates all kinds of other problems. Like, you know, if you get sick and have to stop working, you would lose your coverage. So I think it’s a really mixed bag. If you look at the kinds of things your employer does with your health insurance, though, you also see that just having your employer provide it isn’t enough to deal with the selection problem. So, for example, have you ever wondered why you can only change your health insurance once a year, in the fall? 

DUBNER: Yes, I have.

FINKELSTEIN: Okay. Well, I’m here to answer that, then. It’s precisely to prevent people from waiting until they get sick to sign up for more comprehensive health insurance coverage. Because if you can only do it once a year, you kind of have to make a decision about how much coverage do I want in case something happens this year. Now, there are exceptions. I did, for example, have a son in July. And so when I was two or three months pregnant in October, I thought, “Oh, I really want to be able to deliver at the best hospital in Boston.” So I upgraded my health insurance, but most people don’t know nine months in advance when they’re going to have a major medical expenditure. 

So what happens when you can’t predict a major medical expense? In 2020, The New York Times published an article about a 72-year-old man who had a hard time buying additional coverage after he was diagnosed with cancer. In Risky Business, Finkelstein and her co-authors take issue with the article. What did the Times get wrong?

FINKELSTEIN: So in the same way that you can only choose your health insurance once a year, in the market for health insurance for the elderly, there’s a lot of regulation requiring that insurers offer everyone when they turn 65 supplemental coverage to top up Medicare. They can’t say, “Well, we’re not going to take you because you have a preexisting condition.” But after those six months of you turning 65, then they can. And so this article that we took issue with was bemoaning a person, his name was Ed Stein. When he was 65, he’d been healthy and fit and he’d forgone supplemental coverage and in fact, chosen a rather restrictive policy that had other perks that are designed to attract healthy people, like gym memberships and the like. And then about seven years later, he became unfortunately quite ill, needed expensive treatment, and wanted to then get this more comprehensive so-called Medigap policy. And they were unwilling at that point to sell it to him. And the Times was excoriating the evil insurance company. And our view was — not to at all belittle the horrible situation that this poor man found himself in — it also made total sense why insurers can’t let you just wait until you get really sick to get quote unquote insurance. Then it’s not insurance. 

*      *      *

Once you become an adult, you tend to buy more and more types of insurance. In many cases, it can be mandatory — like auto insurance if you want to drive a car; homeowners insurance if you want to get a mortgage; and, in some places, health insurance. There’s other insurance you may be buying into that you don’t really think about, or even know about — like unemployment and disability insurance, Medicare and Social Security. Those payments are automatically deducted from your paycheck. Then there are the many optional forms of insurance: life insurance, crop insurance, travel insurance, hole-in-one insurance for golf tournaments; kidnap-and-ransom insurance for thrill-seekers; identify-theft insurance for the paranoid. If there’s something you’re afraid of, someone is probably out there willing to sell you insurance against it. But not always. In their book Risky Business, Amy Finkelstein and her co-authors write about an insurance product that failed: divorce insurance. It seemed like a good idea. Divorce, besides being unpleasant, can be expensive, so a prudent couple might want to insure themselves against the expense. That’s how one entrepreneur saw it; the product he offered was called Wedlock. Here is Finkelstein reading a passage from Risky Business, about how Wedlock went out of business in just a couple of years.

FINKELSTEIN, EXCERPT: We looked at some of the products Wedlock offered, and it was easy to see why. The prices were exorbitant. Let’s take a sample policy that required a couple to pay a little over $1,900 per year in premiums. If a couple divorced after ten years of paying premiums, they’d get $27,500 — more than enough to pay lawyers in even a relatively expensive separation. But they also would have paid in ten years’ worth of premiums, which adds up to a whole lot — around $19,000 total. The payout was sufficiently modest that a New York Times story on Wedlock suggested that couples would be better off putting the money they might have spent on divorce insurance premiums into a savings account instead. It would yield just a little less than divorce coverage to pay lawyers’ fees if the need arose. But as a bonus, if the couple stayed married, they could keep their money, plus a little interest. 

To Finkelstein, the idea of divorce insurance is sound.

FINKELSTEIN: You know, there are two million marriages a year in the United States, and unfortunately some of those don’t last. And it would be great to have insurance against that. I say this with all respect and love for my husband, I’d love divorce insurance. And I would add that the reason it doesn’t exist is out of selection concerns, that the people who want to buy it are the people who know their marriage is on the rocks — and that’s not, as far as I know, my case. I just really like security. Now I have tenure, but I would have loved to buy tenure-denial insurance before that event. I would love to insure everything. I don’t like risk. I like safety. I like security. And I think that’s true of most of us. 

DUBNER: Is there a way to recast divorce insurance — you know, you can see why it’d be very hard to work that out for a number of reasons. Not only is there adverse selection, but also I think most people in a marriage would be uncomfortable raising the point of wanting to get divorce insurance. What if you could flip it and call it, like, marriage-longevity insurance? Do you think that could work?

FINKELSTEIN: I mean, there’s a reason economists make for lousy entrepreneurs, right? My response is, “Well, if it could work, someone would have tried it.” I will say, though, you see that happening with life insurance, right? Life insurance, you might say, “Oh, well, people don’t want to think about dying.” We sell the reverse of that, longevity insurance, those are called annuities. They pay out for every month or year that you live. And both life insurance and annuities are plagued by selection problems.

What does Finkelstein mean when she says life insurance and annuities are “plagued by selection problems”? Think about it from the perspective of the insurance companies. You could imagine that a life insurer, for instance, would prefer to sell policies to younger, healthier people rather than older, sicker people. And since these companies are full of actuarial scientists who are trained to sift through the data and predict how long a given person will live, you’d think they’d be very good at selecting healthier people whenever possible — and, when that’s not possible, charging the unhealthy people much higher prices. But that is not always the case. The economist Daifeng He analyzed data from The University of Michigan Health and Retirement Study, which has tracked thousands of Americans for decades. Here’s how Finkelstein and her co-authors describe these findings:

FINKELSTEIN, EXCERPT: What He found was startling to us, given our faith in the power of number-crunching insurance actuaries. Over a period of 12 years during which she tracked mortality, those who died were nearly 20 percent more likely to have bought life insurance relative to people who remained alive during the period, and looked similar on the many characteristics that insurers ask about. To put it a different way, they’re 20 percent less likely to continue with their premium payments in the far distant future. That is, the bad customers (from the insurer’s perspective) bought insurance, and the good customers (the healthier, longer-lived ones) stayed away.

DUBNER: It seems as though we humans are probably a little bit better at assessing our own risk than the firms who are paid to assess that risk are. Is that true?

FINKELSTEIN: It’s true. And it’s, to be honest, totally mind-boggling and more than a bit puzzling. We all know that humans, myself included, are kind of clueless and bad at thinking through probabilities. And then on the other side, you have these, deep-pocketed insurance companies with armies of actuaries, mounds of data, sophisticated algorithms, and all the incentive in the world to make sure they can figure out what your risk is, how likely you are to die.

DUBNER: And a permission to intrude in your life in a way that private citizens might not be, right? I can ask you, “Amy, when’s the last time you smoked? Have you ever smoked? When you smoked, how much did you smoke? Did you also do some drugs when you were smoking, etc.,” 

FINKELSTEIN: Yes. And not only that, but when you say, “I can ask you,” you mean if you were a life insurer. You can’t ask me on the podcast, I’ll take the fifth. But yes, not only as a life insurer, or would-be insurer, can you ask me, but more to the point, you might say, “Okay, well I can just dissemble and lie.” That would be a very bad idea because, if unfortunately I die and my family tries to collect on the insurance, they’re going to investigate very carefully and make sure that I gave truthful answers, and untruthful answers would be disqualifying. 

DUBNER: But I could also imagine that as the so-called big data revolution gets bigger, and as artificial intelligence and machine learning get more useful, that it could radically shift advantages toward the industry away from customers. What do you think of that? 

FINKELSTEIN: I think it’s a very natural intuition. It’s one we have as well — that in the age of big data, this informational advantage that customers have should disappear and markets should function better. We’re still waiting for that blessed day. Big data doesn’t seem to yet be the answer, and perhaps it never will. We actually saw after we wrote the book a fascinating example of this play out, not in an insurance market, but in a real-estate market. Do you remember Zillow Offers, when Zillow was going to get into the business of buying and selling homes, not just being a platform where others could do it?

DUBNER: That didn’t work out in the end, did it? 

FINKELSTEIN: Yes, it spectacularly collapsed. I think the idea was, look they’ve got a ton of data. They can figure out how to price a house, but one way I think about it is of course insurers know more than we know at some level, and they have really sophisticated algorithms. But then you get to be the next mover, you get to say, “Oh okay, they’ve put everything they know into their offer of how much they’re willing to buy my house for. But like, they haven’t lived in my house, and they don’t know about that nasty patch of mold in the corner.” So it’s not that they’re not more sophisticated. It’s that they reveal what they know when they set the price at which they’re willing to buy your house or sell you a life insurance policy. And then you can think, “Okay, they’re much better than me at figuring out the mortality risk of a 49-year-old female economist. But what they don’t know is that I get nine hours of sleep a night or what have you. There are some things that data can never tell you. One of which is also what you’re going to want to do when a risk occurs. Like, if your pet gets sick, are you going to want to quietly euthanize him and bury him in the backyard? Or are you going to want to pull out all the stops of end-of-life care? I don’t know how big data could ever know that. 

DUBNER: So you write in the book about how, with the passage of the Civil Rights Act in the 1960s, it became illegal to set prices that depend on a customer’s race or ethnicity. And then you write that Black Americans have a lower life expectancy than white Americans, so if insurers could take race into account, white people would pay less. Instead, that’s not what happens. You say that that’s a good policy, but that there’s a cost to this trade-off. So talk about that a bit more. 

FINKELSTEIN: Yeah, I think it’s a good policy in the sense that both the available evidence and common sense — and any knowledge of American history — suggests that the reason African-Americans have higher death rates than white Americans of the same, say, age and gender is because of a history and continued systemic discrimination that has forced them into circumstances that have made their health worse. So it seems unfair and unjust to then say, “Well, you have to pay more for your life insurance because of the centuries of discrimination that you and your ancestors have faced.” On the other hand, when you don’t say that, that is going to create selection in the market. There also are cases — remember we talked about the opposite of life-insurance, annuities, policies that pay out as long as you live. Those should be priced lower for people who have higher mortality. So in this case, African-Americans should get lower prices for annuities. That’s also illegal. So it’s kind of a blunt instrument that sometimes fits with our intuition of fairness, and sometimes doesn’t. 

This is a recurring theme in Finkelstein’s book: when the government gets involved with insurance, the demands of insurance policy can come into conflict with other concerns. Consider the recent bank failures: The Federal Deposit Insurance Corporation says it only insures bank deposits up to $250,000. But faced with the possibility of a bank run that could have threatened the financial system, the F.D.I.C. guaranteed all deposits at the banks in question — essentially giving a green light to future risk-taking.

FINKELSTEIN: We all know what happens when the hurricane hits, and people don’t have insurance. Politicians find it impossible to resist the temptation to put on a hard hat, fly down, tour the stricken area and declare a federal disaster area, and send in FEMA to provide some rebuilding and support and insurance. The insurance markets don’t always work well. There are people who can’t afford the insurance. It makes sense that you want to come in and help in a disaster. But then, okay, if the government’s going to come and help me, why should I buy insurance? FEMA’s not going to be as comprehensive as a private policy, but a lot of what I’m paying for with my private policy is stuff that I’d otherwise get from the government.

DUBNER: There is a very recent working paper that you may be referring to here, I don’t know, it’s called “The Unintended Consequences of Post-Disaster Policies For Spatial Sorting.” It’s by Eunjee Kwon, Pierre Magontier, and Marcel Henkel. Here’s what they write: “Post-disaster aid from the government aims to relieve affected communities, but excessive bailouts may encourage economic activity to remain in exposed areas.” Can you say a little bit more about that problem? 

FINKELSTEIN: The paper you’re talking about is talking about natural disasters, where it’s an obvious example. It’s actually a big problem in a market I spend a lot more time studying, which is healthcare markets. It doesn’t play on T.V. quite as much, with the politician coming down. But the same thing happens when people get very sick and cannot afford the care they need. And I’m not just talking about the emergency room, which looms large in the imagination, that emergency rooms are required to treat people regardless of cost. There’s a whole network of formal policies and regulations that fund hospitals to cover non-emergency care for the indigent, uninsured; federally qualified health clinics and other community health centers that are funded and regulated to provide that care. And that’s done for a very good and very important reason, which is the same thing that prompts FEMA to step in, in the case of natural disaster, prompts us also to reach out and help people when they’re desperately ill and can’t afford the medical care they need. But that then creates this patchwork of confusing and arbitrary sets of availability of care. It doesn’t really provide good care, and it can discourage people from buying a better health insurance policy. Because, again, as with the homeowner’s insurance, a lot of what they’re going to pay for, not all of it, but some of it is just paying for what they would have gotten for free otherwise. So it’s an example of, in trying to solve a problem, we can sometimes create additional problems.

*      *      *

DUBNER: So Amy, the title of your new book is Risky Business, which I don’t think will blow anyone’s mind, for better or worse. The subtitle, though, I think may surprise a lot of people. It is “Why Insurance Markets Fail and What to Do About It.” I think the average person, when they think about insurance — which is probably almost never, no offense — they don’t think about it failing. They think, “Oh, the insurance people win. They always beat me. I have to pay a lot of money. The bad thing that I’m afraid of almost never happens. And if the bad thing does happen, I may have to fight to get my payout. And if that happens, my premiums will probably go up.” That all sounds like an insurance company succeeding. So when you say that insurance markets fail, what do you mean, exactly? 

FINKELSTEIN: Well, first, I want to go back to something you just said about, people think, “Oh, the insurer won because I didn’t make a claim.” Now, without denying that there are cases where insurers make exorbitant profits, and there’s a reason governments can get involved there, one of the fundamental problems is people thinking, “I bought this insurance policy, and I never got a payout on it, what a waste.” No! You should be thrilled. You should be thrilled both that the bad things didn’t happen, and that you were able to lead your life with the security of knowing that if they did, you would get a payout. It’s not a investment in which you’re putting in money to get money back. You’re putting in money so that in the small chance you need it, you really get it. And when we say insurance markets fail, we mean they often fail to deliver on that promise and that potential because the market disappears entirely. Or it gets priced so expensive, like $4,300 a year to insure your 12-year-old bulldog for a policy that at most will pay out $5,000. Or it’s incredibly limited. Like, yeah, you can buy dental insurance, but it’s not going to cover anything actually expensive, like an implant or a root canal. That’s what we mean by them failing. And all of the reasons they fail is because of this selection problem, that customers know stuff about how likely they are to need payouts on their insurance, and the ones who know that they’re more likely to need it flock to the market and destroy it for the rest of us. 

DUBNER: Okay, so this gets us to the last few words of your subtitle, which are the hardest part, “and what to do about it” — what to do about this failure in the insurance markets that’s caused by this selection problem. So what are your solutions? Because one could imagine that you could equilibrate these things, right? You could create the right selection for all these markets, but it would probably take an awful lot of compulsion or fees, or taxes, or super-strong mandates that perhaps have prison sentences attached to them. So without going that extreme, how can you do that? 

FINKELSTEIN: Well, I have to confess that our original subtitle was: “And what, if anything, to do about it.”

DUBNER: Are you blaming the publisher? The publisher took out the qualifying language, you’re saying? 

FINKELSTEIN: Throw them under the bus, for sure. Because we wanted to make clear that it’s not always clear you can do something about it. 

A quick side note here: the publisher is Yale University Press, in case you want to drop them a line.

FINKELSTEIN: But I think what to do about it depends on who you are. If you’re the customer, I have several suggestions for what to do about it. They’re different than if you’re the seller or if you’re the government.

DUBNER: All right, let’s start with the customer.

FINKELSTEIN: Okay. So if you’re a would-be customer of insurance, the first thing to do is to realize that insurance needs to be bought before the risk happens. You don’t want to wait till you are sick to get health insurance or wait until your car breaks down to get auto insurance. That’s the first lesson. That’s kind of obvious. And then the second is: in deciding what kind of insurance to buy, if you realize this problem of selection, you can also realize how to get a good deal. So the way to get a good deal is to sacrifice a little bit of insurance: buy a policy with a long waiting period or a very big deductible. So, for example, when we bought our home, I searched high and low for a homeowner’s policy with the biggest deductible known to man. You know, usually you can get $500, $1,000. I mean, that’s trivial compared to the value of a home. I managed to find a policy with a $10,000 deductible. Super-cheap. Now, what happened a few years later, in I think 2015, Boston was hit with these tremendous snowstorms, ice dams, etc. We, like every other family in the Boston area, had damage to our home. We blew through the deductible. We had to pay all of it. Next year, we had a leak in our boiler and flooded the basement — there went the, deductible again, causing my husband to complain that it turned out to be a very expensive proposition being married to a so-called insurance expert. But he will agree, that all told — we’ve now had the policy for about 15 years — even despite those two really expensive events that we had to pay a lot for because we had this huge deductible, we still have ended up saving money because the premiums are so much lower because they’re priced for lower-risk people.

DUBNER: Okay. So you gave me from the customer’s perspective, what about the seller perspective? What can possibly be done about it? 

FINKELSTEIN: Well, I think the first thing is to be humble and to recognize that this is a real problem. I think the second thing is to think about how to design your policies to attract the kind of customers you want. One of my favorite examples is work that the researchers Cooper and Trivedi have done on why health insurers offer discounts for gym memberships bundled with health insurance. Now, you might think — I’m going to suspect most people would — “Oh, well, sure. The health insurer wants you to be healthy, too. So they want to encourage you to go to the gym.” Turns out that’s not why. We have a lot of good evidence that these so-called workplace-wellness programs don’t actually make people healthier. The reason employers offer workplace wellness programs, and the reason health insurers offer you a discounted or free gym membership is not because they’re hoping to make you healthier; it’s because they’re hoping to attract the customer for whom that’s appealing. Not the incredibly sick, homebound, couch-bound customer with multiple chronic conditions who can barely walk around, but the healthy and spry customer who at least can delude themselves into thinking, yeah, they’re really going to go to the gym. 

So that’s the customer perspective, the insurer perspective — but let’s not forget the government perspective on insurance. What, exactly, do governments do around insurance?

FINKELSTEIN: I use this line the first day of the graduate economics course I teach on government policy. I start by saying: it’s right there in the data — that if you look at the federal budget, as this Treasury official, Peter Fisher, said, “The federal government is a giant insurance company with a sideline business in national defense.” National defense is about 20 percent of the federal budget. And I think easily 60 percent of it is Social Security, Medicare, Medicaid, disability insurance, unemployment insurance. So I think governments do several things. Sometimes they directly provide insurance, like Medicare. Other times they simply require that insurance be provided, like worker’s compensation insurance or now health insurance. Other times they offer discounts or subsidies for people buying insurance or fines if they don’t. And then the final thing they do is regulate. They say, “No, you can’t charge higher prices to this group of people.” Or: “You can’t offer a policy if you won’t sell it to everyone.”

DUBNER: Now in each of these cases — let’s say, Social Security, unemployment, disability and so on — our participation is compelled, as long as you qualify. If you’re working and so on, you pay into these things. And in many cases, the people who pay in do not get paid out in anything like the proportion they pay in. And they are helping the people who are not paying in to get coverage. So is that really insurance?

FINKELSTEIN: That’s a great question. You can think of that as redistribution, or you can think of it as insurance behind the veil of ignorance. None of us know exactly when we’re going to die or when our job is going to end, and so there is real insurance value. Take the case of Social Security, which comes in the form of an annuity. As you said, people pay taxes when they work and then when they retire and claim Social Security, they’re going to get a fixed amount that’s related to how much they paid, in every month until they die. So on the one hand, that’s really valuable to everyone because otherwise, how would you figure out how to spend your money in retirement? You’d worry that if you hoarded your money in case you lived a long time and then you ended up dying shortly after retirement, not only is it unfortunate that you die, but it’s unfortunate that you missed all that chance to enjoy your hard-earned income. Or if you decide to live life and be merry right after retirement and then are in some sense fortunate to live a long time, you can end up quite poor in old age. And so an annuity is supposed to solve exactly that problem. On the other hand, as we were talking about lower-income people, African-American people, have higher mortality. They’re paying in the same amount based on what they’re earning and they’re not likely to get the annuity for as long. 

DUBNER: The more we speak, the more it sounds to me at least like what you are describing here is a big broad set of issues or problems that are almost impossible to solve because they’re at the intersection of so many things, They’re at the intersection of markets, politics, personal behavior. And also we try to insure many, many, many things — from life and health to our homes and flights and jobs and so on. So I don’t think this is a good idea, but: if someone were to say, “Hey, you know, what this country really needs is an insurance czar or czarina, and I think Amy Finkelstein is the only person who can really do it.”

FINKELSTEIN: Definitely a bad idea. 

DUBNER: Let’s both pretend for a minute that it’s not a bad idea. Without a magic wand, right, it’s not like you could magically make things better by, let’s say, disengaging healthcare insurance from employment or whatnot. But what are a few things you could do from a high level that would start to address some of this multitude of problems you’ve been talking about? 

FINKELSTEIN: I think the first thing you have to do is decide which is the most important problem. I would pick the area that I’ve studied the most, which is healthcare and health insurance. And I actually have a solution, and it’s the subject of my next book.

DUBNER: Oh, you’re just going to give us the cliffhanger? You’re not even going to give us the answer?

FINKELSTEIN: You tell me, how do I sell more books? No, so this book that’s coming out this summer, again with Liran Einav. It’s precisely this question of, oh my God, the problems in health insurance markets seem so intractable. Everyone’s always asking me, “What would you do”? And I always say, “I don’t know what I would do. That’s why I study it.” And at some point that became too cheap and pat an answer. And so we started thinking about it, and we do have a solution. And the solution is rooted in first understanding what the problem is. The fundamental problem in U.S. healthcare is that we always step in when people are sufficiently ill and sufficiently destitute and rescue them. And without any conscious planning or thinking, we’ve enacted a series of patchwork policies. We have disease-specific insurance, special coverage if you have end-stage renal disease or if you have breast cancer. We have insurance programs if you’re a certain age or live in a certain place or a certain income. But we’ve never done it as a coherent, unified policy. And the point of our book is to say, once you recognize that that’s what we always have done — and we go back actually to the start of the Republic and the first example of nationally mandated health insurance from the 18th century in the U.S., we realize that the solution is, in fact, to recognize what we’re always going to do and do it sensibly, and both provide automatic basic coverage for everyone. But then going back to some of the other things we’ve talked about, that not everyone wants the same thing, also, allow people who want and can afford to buy more, to buy more in a well-functioning market. 

DUBNER: Now, as great as that may sound, in theory, my initial thought would be, Well, wait a minute, you’re talking about an industry, all-in healthcare, that’s roughly 20 percent of G.D.P., which is so much. I mean, on the one hand, it’s great news that we’re such a rich country that we can spend 20 of every $100 on our health. That’s, in some ways, a great indicator of a healthy society if all that spending really worked well. It doesn’t seem to work that well, is the problem. But an even bigger problem, perhaps, is that there are so many firms, shareholders, individuals, investors, etc., who get rich year after year after year. And now you want to un-entrench, even if slightly, those entrenched interests. Do you think a plan as sensible as yours stands any kind of real chance in the real world? 

FINKELSTEIN: So I have two answers to that. The first is I don’t think that’s the right question. We started the book, as I said, because everyone kept asking us, “What do you think we should do about healthcare reform?” And we kept saying, “Well, it’s complicated,” as President Trump said. That was actually true. The point of our book is not to say how to get there, but let’s, you know, give you the North Star, and then let’s figure out how we accomplish it. I’d also say, though, that universal health insurance coverage in the United States is not as impossible as the standard cocktail party narrative would have you imagine. And part of the reason is that it also wasn’t inevitable in all the other countries that currently have it. So it’s never been a walk in the park. The doctors in Canada went on strike when they first introduced universal medical insurance in the first province there. The bitter battles over the formation of the National Health Service in the U.K. in 1948 are hard to fathom, given how popular it is today in the U.K. And there are a series of near-misses that occurred in the U.S. So I don’t think we should be that pessimistic. And I think in some sense, our role is to put the idea out there so that, when the moment occurs, when the impossible becomes inevitable, we have a sensible plan in place.

DUBNER: If we wanted to be very reductive, would you say the average American is over-insured, under-insured, or by some miracle, insured at just the right level? 

FINKELSTEIN: Underinsured. 

DUBNER: So what are the areas then that are most prone to underinsurance? 

FINKELSTEIN: I’d say life insurance, disaster insurance, and then unemployment or layoff, I’d say, is a big one. 

DUBNER: As you write, the market for layoff insurance is almost nonexistent, right?

FINKELSTEIN: I think people want that insurance. It’s just that selection makes that market not function from the perspective of anyone who tries to offer it. 

DUBNER: How about this one? There are a lot of people in the world who would like to have a life partner and have a hard time successfully finding that person. Could you imagine that there’s some kind of insurance —

FINKELSTEIN: You want happiness insurance?

DUBNER: Yeah, well, I was thinking, you know, partner insurance, that if it doesn’t work out, at least, you know, let’s say this money is going into a pool so that you can put a down payment on a first home if you do get that partner. But if you don’t, you have that bounty and you can cash it out and do something that somehow compensates for not having a partner?

FINKELSTEIN: It’s funny you should mention this. I don’t know of a formal financial product like this, but I will admit that when I was a senior in college, a bunch of friends of mine were talking and apparently, there was this thing going around, where before you left college, you were supposed to reach out to some good friend and make an agreement that if in ten years neither of you were partnered up, maybe you would give it a try. They called it your safety husband. And a friend of mine from high school was walking by. So I called him over and I told him about this and I said totally jokingly, “So, Darren, will you be my safety husband?” He said, “Safety husband, Amy? I’m your reach.”

DUBNER: Poor Darren. And what about Ben, what about your actual husband? How did that meeting happen?

FINKELSTEIN: Well, of course, it involves a selection story as well. And not just the good selection he made in me. On our first date, we went out to dinner, and at the end, we’re splitting the bill as I think most modern couples do. And he pulls out this credit card that has a giant picture of the Harkness Tower at Yale. And he shows it to me. And I thought, “What a loser, trying to impress me by the fact that he went to Yale. Like, who does this?” And he said, “You know, this card is really cool. I got it because they market it at a really low interest rate if you went to Yale and I think it’s because they think where you went to college can predict being a good credit risk.” And I thought, “Ah, he’s not a tool who likes to brag about where he went to undergrad. He’s a tool who’s as obsessed with selection and economics as I am.”

That was Amy Finkelstein, award-winning economist at M.I.T.; her new book is called Risky Business: Why Insurance Markets Fail and What to Do About It; her coauthors are Liran Einav and Ray Fisman.

*      *      *

Freakonomics Radio is produced by Stitcher and Renbud Radio. This episode was produced by Alina Kulman and mixed by Greg Rippin, with help from Jeremy Johnston. Our staff also includes Zack Lapinski, Morgan Levey, Ryan Kelley, Katherine Moncure, Rebecca Lee Douglas, Julie Kanfer, Eleanor Osborne, Jasmin Klinger, Daria Klenert, Emma Tyrrell, Lyric Bowditch, and Elsa Hernandez. The Freakonomics Radio Network’s executive team is Neal Carruth, Gabriel Roth, and Stephen Dubner. Our theme song is “Mr. Fortune,” by the Hitchhikers; all the other music was composed by Luis Guerra.