232 | Amy Finkelstein on Adverse Selection and Hidden Information

0:00:01.0 Sean Carroll: Hello, everyone. Welcome to The Mindscape podcast. I'm your host, Sean Carroll, and today's topic is insurance markets. Now, I know in the history of Mindscape, we've talked about the meaning of life, the fundamental laws of physics, the way that giant philosophical scientific ideas fit together, as well as some more fun arts and cultural kind of topics. Compared to our usual affair, insurance markets seem pretty down to earth, maybe not the sexiest topic, depending on what you got interested in, I get that. But let me tell you, when you dig into what an insurance market is and why it might not work, it's fascinating, there's a lot of intellectually interesting things going on here, especially if you care and or are interested in the fundamental workings of capitalism and the market.

0:00:50.2 SC: So not to sort of spoil too much of the conversation to come, but the basic idea is very simple, like what is insurance? Insurance is a trade, you give somebody money now for peace of mind, basically for knowing that if something goes terribly wrong, you'll be able to protect yourself to cover your debts or whatever it is, depending on the kinds of insurance. So it's already intellectually interesting because it's trading something right now for a future thing, which is always kind of an interesting question, but there's also the question of the market for insurance and how it works and how it's a sort of counter-example to many things that capitalism is supposed to do.

0:01:33.5 SC: We know that capitalist markets can fail for various reasons, that there are monopolies or monopsonies and things like that, incomplete information, but here's an insurance market failure, if you offer insurance to a bunch of people and you say, "Okay, if you got sick, I will pay for your healthcare," and you set a certain rate and guess what, who buys your insurance, the people who are likely to get sick. [chuckle] The people who are more healthy are less likely to buy the insurance, so you have had been very careful. You had done some actuarial tables, you figured out how often people get sick, you knew that you would make money if you bought your insurance, you failed to make money 'cause only the sick people bought your insurance.

0:02:15.3 SC: So you go, "Ah, okay. So it's only the sick people that who are gonna buy it. What I need to do is to raise my prices so that I can still make money, even if only the sick people buy my insurance," but then guess what, only the really sick people buy your insurance. [chuckle] So there is a race to failure built into the very nature of an insurance market, and it's not because people are greedy, it's not because the giant establishment is out to get you, it's just that supply and demand isn't really working, it's not about fairness, it's just about the mechanism, not quite working.

0:02:49.6 SC: How do you fix that? Right, there's a whole literature on this phenomenon about adverse selection, where buyers and sellers have different information, and buyers in particular have private information about how sick they are. For example, what bad thing is going to happen to them? Still to this day, there's no such thing as divorce insurance, because the buyer is gonna know much better than the seller whether they're likely to get divorced very soon. So today's guest is Amy Finkelstein, who is who's a very well-respected economist at MIT, and she and her collaborators recently wrote a book, Risky Business: Why Insurance Markets Fail and What to Do About It. So I like a the topic because it's at the intersection there of real world stuff that really matters for legislation and things like that, and how we organize our system, but also there's interesting intellectual questions about present versus future, how to find an equilibrium and a bunch of people with different desires. All that sort of good stuff. So let's go.

[music]

0:03:52.1 SC: Amy Finkelstein, welcome to The Mindscape podcast.

0:04:09.3 Amy Finkelstein: Thanks a lot for having me, Sean. I'm delighted to be here.

0:04:12.0 SC: I wanted to mention right at the start that in your new book, Risky Business, a delightful book, happy to recommend to everybody. One of the delightful things about it is right up at the front, you are honest, you're candid, and you admit to your audience that insurance markets might not seem like the sexiest topic to read or write a book about. [chuckle] So I thought I would at least just open just in case someone listens to the podcast for just five minutes, give us your sales pitch for why insurance markets are sexier than people might think.

0:04:43.1 AF: That's a pretty low bar.

[laughter]

0:04:44.7 AF: I'll actually go for their sexy on an absolute scale.

0:04:47.5 SC: Oh, good.

0:04:47.5 AF: Although, we all have our different taste, so the reason I think insurance is so exciting and the reason I've spent most of my career studying it is that Insurance offers the possibility of doing something that I think is quite magical. It offers the hope for a measure of security and certainty in a dangerous and uncertain world, insurance touches nearly every aspect of our lives. We insure our cars, our houses, our pets, and literally our lives, and when you think about all the risks we face in the world from disease to death, to natural disasters. There's this tantalizing prospect that we can eliminate some of, not all, but some of the vicissitudes and vagaries that plague our lives. So that's why I think it's exciting, it's the tantalizing what economists call a free lunch, everyone can benefit. But what our book is about is the fact that despite I think this very real and potentially valuable service, insurance often fails to deliver on its promise.

0:06:08.4 S3: And few people understand why, and the results of that lack of understanding, as we describe in the book, can be ruined businesses ruined policies and ruined lives, and so our hope in writing this book is to help people understand some of the ostensibly frustrating features of many insurance markets, why dental insurance is so crummy, why pet insurance is so expensive. Why, if you've try tried, you'll have discovered you can't buy divorce insurance for love or money.

0:06:32.0 SC: As it were.

0:06:34.9 AF: And also why there's a real possibility for the government to step in and try and improve insurance markets, but why that's hard also, and how it sometimes fails as well.

0:06:45.5 SC: Well, you touched on the reason why. As a physicist/philosopher, I did... I was a soft target for your introduction because insurance does, when you get right down to it, the nitty-gritty of personal finance, I'm terrible at, and I have no interest in getting better at it, but there are these big picture questions about sort of trading present value for future certainty, it's a bargain between Your present self and your future self is sort of intellectually a fascinating notion to pursue.

0:07:15.3 AF: Yeah, and I think economists tend to think, and there's a lot of evidence that most people are what we call risk-averse, that they're willing to give up a small amount of money for certainty, for certain to avoid the perhaps small possibility of a much, much larger loss. So we'd rather pay a few hundred dollars a year in premiums, and then in the unfortunate event that say our car is totaled, we then get the money to repair the car or buy a new car.

0:07:50.5 SC: And not only would we prefer to do that, but it's not irrational to do that, I actually did recently a podcast with the Floss for Lara Buchak, who studies risk and rationality, and she makes the case that even if you're... She frames it as taunting economists a little bit because there's something we do in our rational decision-making other than maximizing our average expected utility, sometimes we are willing to forego a little bit of utility if we can lower our volatility, the lower our risk a little bit.

0:08:24.2 AF: Yeah. So I would say though, there are many things to taunt economists about, but that isn't one of them.

[laughter]

0:08:29.3 AF: I think that's actually central to Economics, it's the idea going back to Von Neumann and Morgenstern of maximizing expected utility, so you would be happier having... Say you have a 10% chance, you might lose $100, that means on average, you're only gonna lose $10, so you might be willing to pay $10 for sure, and then 90 times out of 100, you're actually not gonna lose any money and you just paid. You get your $90, but the one time you do you also get your $90, so you've never risk losing the whole $100, in fact, maybe you'd even be willing to pay like $11 bucks for that.

0:09:14.1 SC: I think that's the point, yeah.

0:09:15.1 AF: And that's not irrational. That's the bread and butter of a lot of economics.

0:09:22.4 SC: Yeah, Now, well, but that's right, but that paying $11 in that case that you just raised, is lowering your expected utility, but it's still rational, right?

0:09:32.0 AF: No, it's raising your expected utility.

0:09:32.9 SC: You're giving up $11.

0:09:35.4 AF: Because you're gonna have... You're gonna have $89 for sure, both periods, say or you have $89 for sure, rather than facing a 90% probability of having a $100 and a 10% probability of having zero dollars, and you may be happier with that. There's nothing irrational about it. Of course, exposed if you don't... If you don't experience the loss, you'll wish you haven't paid the premium.

[laughter]

0:10:02.2 AF: But that's not what insurance is about.

0:10:03.9 SC: Well, actually, yeah, that's another good point you make in the beginning is that part of the utility you get from paying insurance is the peace of mind that you're covered a little bit, and that has to be factored in.

0:10:14.0 AF: For Sure, for sure.

0:10:15.5 SC: I was also surprised to learn how deep into the past it goes, and I recently wrote a book that had an anecdote in it about Edmond Halley, the comet guy, and so did you. So tell. I had no idea that Halley had anything to do with insurance.

0:10:32.2 AF: Oh, yes! Neither did we until... Or neither did I, at least I can't speak for my co-authors until we started writing this book, and you you've mentioned that you're a physicist, and my understanding is that physicists like to make fun of economists and mathematicians like to make fun of the physicists and that's kind the of intellectual pecking order.

[laughter]

0:10:55.4 AF: So from that perspective that I've lived with my whole life. It was fun for us to discover that Edmond Halley, the comet guy, as you said, and also in the distinguished 17th century mathematician, Abraham de Moivre were both involved in coming up with actuarial tables, calculating mortality risk rates for newer death rates calculations and using those to price survival insurance, something known as annuities. And while they may be... And I think they were excellent scientists, they were not very good social scientists.

[laughter]

0:11:30.3 AF: So sometimes the pecking order doesn't hold, in the sense that they calculated the mortality tables correctly for the population in that they had data for, but what they weren't thinking of is the fact that once you start selling these survival contingent streams of payment, these things called annuities, which actually, the French and Dutch governments were doing in the 17th century to finance war as much as the way modern governments issue bonds or treasury bills, so they'd say to speculators, you give me a bunch of money and we'll pay you every year until someone that you've nominated dies. So what Halley and de Moivre didn't realize is that the set of people whose...

[chuckle]

0:12:20.1 AF: On whom those bets are placed are not randomly drawn from the population, so for example, in the French case, they looked for young French girls who had already survived smallpox. So those people had quite a much longer survival prospects than, say, older men, and then after a couple of hundred years, and arguably the French Revolution, 'cause this did bankrupt Louis XIV, these mis-priced annuities. Eventually, people caught on and thought, Okay, well, we'll say set price the annuities differently if you're taking them out, a bet on the survival of a 70-year-old male as opposed to a five-year-old girl, but even then, it turned out that speculators found the really, really hell and Hardy, old man in Scotland and placed bets on them, and even in modern times, we continue to see these types of issues coming up.

0:13:18.8 SC: This isn't an aside, but let me just get on the record that I think that the pecking order that you refered to is essentially, never true. Whenever I see people in highly technical fields think that other fields are easier, when you see them try to do it, they're not really very good at it, there's a whole other set of skills that are required and we should all recognize that. But poor Halley, he did okay. In other areas, he got the comet named after him, so that's okay. [chuckle]

0:13:42.3 AF: Absolutely. And to your point, I'm sure I would never have discovered that comment.

[laughter]

0:13:45.2 AF: That comet. Sorry, I would never have discovered that comet, even if I might have, with the benefit of a few hundred more years of economics figured out the selection problem.

0:13:55.2 SC: But yeah, his mistake that you're pointing to is sort of the center of what we have to talk about here because... Well, you're the world's expert. I should let you describe what's going on here, but there's a thing called adverse selection, there's a thing in insurance, unlike selling cars or smartphones, where you care about who is buying it, you do not get the same value for any different kind of buyer.

0:14:19.9 AF: Exactly, in fact, you made the comparison, which is perfectly accurate to a smartphone or a car, the comparison we love is one of the things that prompted us to write this book, which was listening to the Supreme Court oral arguments over whether or not the mandate to require everyone to buy health insurance under ObamaCare, it was constitutional. Now, we have nothing to say on constitutional law, but what intrigued us was hearing then Supreme Court justice Antonin Scalia asked ask during oral arguments, Well, if the government can force people to buy health insurance, can it also force them to eat their broccoli? Now, he was obviously being rhetorical and trying to conjure up this image of an undesirable intrusive state, but I think to many people, if not to Scalia, the distinction, at least the distinction we're interested in between insurance and broccoli is not clear.

[laughter]

0:15:20.7 AF: And the reason it might make sense, we jokingly thought of calling our book Why is insurance different than broccoli, but we thought putting two unpopular things in the title was probably not a good idea.

[laughter]

0:15:29.1 AF: But the reason we think there may be a role for the government and in mandating the people buy insurance, be it auto insurance, flood insurance, health insurance, but not for eating their broccoli Is precisely what you said. When the supermarket sells it's broccoli, the profits and the profitability of that broccoli depend only on the price it charges and at how many customers buy it, it does not depend on which customers buy it, that is not true for insurance, how much it costs an insurer to ensure automobiles or housing or our lives depends on whether the customers who buy it are high-risk, accident-prone customers or low-risk ones, and once you realize that, what one will realize or economic theorists in the 1970s, going back to George Akerlof realized is that this can wreak havoc on the market that if you offer an insurance product, the people who are going to flock to it are the people who know that they're risker higher expected costs than the typical buyer.

0:16:37.9 AF: This happened to Edmond Halley in the hundreds of years ago, and it happens today, we give examples in the book with divorce insurance and lay-off insurance, that entrepreneurs tried to offer these products, and they found very quickly that the customers that were buying them were not the customers they wanted.

[chuckle]

0:16:57.3 AF: And so they had to raise the price to cover the higher than expected rates of divorce among people who had divorce insurance or rates of lay-off among people buying lay-off insurance. Well, unfortunately, what raising the price does, it sounds like the sensible solution to my costs are higher than my revenues, that might be business school lesson 101. The problem in a selection market, like an insurance market, is when you raise your price, you just exacerbate the selection problem, now it's only the really, really high risk.

[chuckle]

0:17:27.5 AF: The ones who like already are bickering with their spouse or have already seen the pink slip, who are gonna buy this, protection and the market, and we give examples in the book, can just literally spiral out of control. And would have what's called the death spiral it that can completely disappear.

0:17:45.1 AF: And so going back to why insurance is both exciting but also important to understand. I started by saying it's exciting 'cause it offers a measure of security, but this type of selection problem can destroy that potential, and that's a problem not just for the people who already know they're about to get laid off or about to get divorced, but for all of us, because we all, I think, value having some security, even though I believe I am happily married and at no risk of divorce or a very low risk, and likewise, have secure employment, it still would be great to know that in the unfortunate event that I lose my job or my marriage, that I had some financial resources to help cover the costs of those events, so everyone can benefit from insurance, the low risk and the high risk. The problem is that selection can make it impossible for either everyone or most people to benefit from that insurance.

0:18:40.7 SC: And it is intellectually interesting to someone like me because it is an example of a market failure. And I know that economists are well aware that markets do fail, like that is a bad stereotype to think that economists think that markets never fail. But there's sort of standard ways in which markets can fail, either by exacerbating inequality or by monopolies or things like that. And this is kind of a different way. There's nothing mistaken about it. Everyone is doing the individually rational thing, but that the market is just driving you to an equilibrium that no one wants to be in.

0:19:14.7 AF: Yes, and that's what is so, to me, so fascinating about it is that you could have a situation in which everyone values insurance. They're actually willing to pay a price that's actuarially fair based on their own risk, which means they're willing to go to go back to our example of the 10% chance of losing $100. The actuarial cost, the expected cost to the insurance company is $10, and I might be willing to pay $11 to avoid that risk. And so even if everyone were willing to do that so that everyone gets gains from buying insurance at a price that allows the insurance company to break even or even make a tiny bit of money, and yet that market may not exist. That's what's so both frustrating and intriguing about insurance markets.

0:20:06.5 SC: Yeah, and it's worth emphasizing, you said it several times, but I just want to emphasize it just so that no one misses it, that the idea of insurance by itself is not a scam.

0:20:16.0 AF: It's Cockley.

0:20:16.1 SC: It is true that the insurance company is going to make money that if everything goes well, you will be spending more money than you get back on average. But it's worth it because of this lower risk, higher certainty, the peace of mind you get along the way.

0:20:31.1 AF: Yes, to most people, I mean, there are some people the skydivers and bungee jumpers among us who may love risk and not. But I think so I'm not saying that it's irrational for everyone to not buy insurance at a fair price, but most people value it. And yes, this is how markets work. When I value the car more than it costs Volkswagen to make it, then that's what we call gains from trade. And we can argue about how to split the surplus should they make no profits, a tiny bit, a lot but that's a good thing. That's the economy functioning. And what's amazing about insurance markets is there can be that surplus on the table, and yet the market doesn't function. Or for people who insurers, I think, have a bit of a bad rap. And for sometimes with reason one of the things is that listeners may think, Oh, come on, but insurance is priced exorbitantly. It's not just priced in my example, like, Oh, I have an expected loss of $10 this one in 10 chance of losing $100. And so fine, maybe I'd be willing to pay $11. They're charging me, they're charging a price of $50.

0:21:43.4 AF: And that is empirically correct. So if we talk in the book, for example, if you look at the market for pet health insurance, which has been growing over the last couple decades, the prices really are exorbitant relative to the risk. So for example, a 12-year-old bulldog, if you want to insure him, the annual premium is $4300. And here's the kicker, the maximum payout is $5000. So that is a very, very high price. But one of the reasons the price is so high is because the people who are insuring their 12-year-old bulldogs are the people who know not only that Rover is going to about to need a lot of medical care, but that they're the type of owner who if Rover does need that medical care are not going to give him some pain relief and sadly say goodbye, they're going to do the high tech invasive surgery on their 12-year-old dog. And so one of the reasons insurance looks so bad and gets such a bad rap is because it often is high priced and also limited in all kinds of annoying ways with waiting periods and deductibles. But that's there precisely because of the selection problem. It's the insurers trying to figure out how they can insure this risk that is so adversely selected.

0:23:06.5 SC: Yeah. And the example of pet insurance prefigures something I want to get to later, but it's not just pets we have this difference of strategy about, it's ourselves, right? You know, the topic of insurance brings up injury and sickness and death and a lot of bad things that can happen to us and different among us are willing to fight and fight to stave off death as long as possible. And that might affect how much we're willing to pay for the medical care and therefore the insurance.

0:23:32.4 AF: Totally.

0:23:35.5 SC: And so the selection thing about the spiral where you only get the worst customers who want to pay for your product, that it has a crucially interesting connection to the information that different people have. It's an example where the people who want to buy your insurance know they're in trouble. They know they're going to get divorced and therefore they're buying it. And presumably the insurance company doesn't know that.

0:24:03.3 AF: Yes. And you've hit on, I think, one of the really puzzling aspects of this whole problem that we spend some time grappling with in the book, which is how can it be that you have on the one hand these massive insurers who face major financial incentives to figure out just how risky their customers are and price it accordingly. They have mounds of data. They have sophisticated actuaries and actuarial scientists and their data algorithms. In the world of big data today, how is it possible that there are things that people know about themselves that the insurance company doesn't know? So we talk about this in the book. We give some anecdotes, which I'm happy to share to illustrate. We also provide some of the empirical evidence. In some case, clever researchers have been able to document specific forms of this private information, such as knowing, for example, whether if you have a parent with the genetic disease for Huntington's chorea, there's evidence. So you have a 50/50 chance of having it yourself.

0:25:16.6 AF: It's a dominant gene. There's evidence that people at risk for Huntington's chorea have much higher rates of purchasing nursing home insurance than people who don't have that gene, the risk of the gene. And then even among the people at risk, once they do the genetic tests, the people who test positive tend to buy the insurance at higher rates than the people who don't test. So in some cases, we've been able to pinpoint the source of it. In other cases, it's genuinely surprising. In the case of life insurance, for example, an economist named Daifeng He has documented that if you look at people who look the same on every possible dimension of their health and their family history that you can think of, and you look at who buys life insurance and who doesn't, and then follow them prospectively over time, the people who buy life insurance are more likely to die in the next 10 years. And so we don't when we try to think what is it that the life insurer isn't figuring out, well, do they not know your your health?

0:26:12.3 AF: No, they actually come and do a medical exam. Do they not know your family medical history? Nope, they ask about that. Maybe it's your lifestyle choices. Nope, they ask about your smoking and your skydiving and whether you travel to dangerous places. So there are cases where we don't know what it is, but the data are very clear that there is something that people know above and beyond what the insurance company is able to figure out about them. So the idea that oh, in the world of big data, we'll somehow get rid of this selection problem because big brother will know everything about us turns out not quite everything.

0:26:52.1 SC: Is there any chance that the causal flow is that buying health insurance or life insurance makes you die earlier?

0:27:00.0 AF: So, yeah, so this is one of the central empirical challenges to researchers in this area. The problem we've been talking about is one of selection that people who are higher risk are more likely to buy the insurance in the first place. The other possibility that you refer to is known as moral hazard. It's the idea that even if identical people, there are a bunch of people who all are the same risk and then we just randomly give some of them automobile insurance and other people not, the people who have the insurance will now put their foot on the gas pedal a little more and drive a little less carefully. In health insurance, there's a huge body of evidence, some of which I've contributed to documenting precisely this phenomenon that once people get health insurance, including randomly assigned health insurance, we did a randomized evaluation of this in the state of Oregon. People take a group of people and you randomly give half of them insurance and Medicaid and half of them remain uninsured.

0:28:09.0 AF: So, on average, those two groups are the same in terms of their health, their propensity to use the doctor, but the ones who get insurance do, in fact, then go to the doctor more. So, moral hazard is alive and well, but we also know that there's selection. And one of the ways you can see this is imagine now you randomize not whether you have insurance, but you randomize, this is other work we've done, effectively randomize the price at which people are offered insurance. So, a higher price or a lower price. One of the things you see in the data is you can then look at the people who choose to buy insurance at the higher price and the lower price. So, everyone has insurance. So, any effects of insurance making me more likely to go to the doctor, that's the same. What's different between the two groups is one consists of a group of people who are willing to buy health insurance at a low price. The other consists of a group of people who are willing to buy it not only at a low price, but also at a higher price.

0:29:16.4 AF: There's fewer of them, right? Because we know people don't like high prices, but you see the set of people who are willing to buy when the price is higher have much higher healthcare utilization after the fact than the people who bought when the price was lower. So, raising the price gets you a riskier set of customers. Again, you're comparing two groups of people, both of whom have insurance. So, that behavioral effect of the insurance on your healthcare use is operating for both of them, but the group that was willing to buy at an even higher price has even more healthcare use coming from the fact that they're only willing to buy it because they expected to have higher use.

0:29:55.3 SC: This is, I mean, you are bringing up the fact that this is a nice case where interests align, arguably, right? I mean, the insurance company wants you to be healthy and not use the doctor a lot. And so, maybe their best interests are served by encouraging you to adopt healthy behaviors.

0:30:13.1 AF: So, this is a great segue to another example we talk about in the book that also gets into this. Is it about selecting the customers or changing their behavior? So, something that we've observed many times in the world, perhaps some of the listeners have actually experienced this, is that some health insurance companies will offer discounts for a gym membership, bundle that along with their health insurance. And you might think, going back to your, "Oh, incentives are aligned point," that that's because it's great for the health insurer if you go to the gym and get healthier, just like it's good for you too. But the data show that that's not what's going on. So, there's fascinating work by the researchers Cooper and Trivedi who show that when insurers start offering these discounted gym memberships bundled with the health insurance product, they attract customers who are already healthier. You can measure their health before they buy the insurance. Moreover, there's other excellent work done by three economists, Damon Jones, David Molitor, and Julian Reif, in which they actually ran a randomized control trial of offering employees a workness wellness program at their university, at the University of Illinois at Urbana-Champaign.

0:31:36.6 AF: And what they found by comparing people offered the workplace wellness program and people who weren't is that it didn't make you any healthier, didn't make you any more productive, all the things that employers or workplace wellness aficionados will tout. But they also found that there's a reason that employers still offer it, which is the types of employees who like to sign up for workplace wellness programs are healthier at baseline. It's the same reason that Cooper and Trivedi find that health insurers offer gym memberships. It's because they want to attract the healthy customers. At least the relatively more healthy and fit among us can at least delude ourselves into thinking we're really going to use that gym membership. But it turns out it doesn't actually make you healthier. It just gets you healthier customers.

0:32:27.7 SC: Right. So, even if I'm not the kind of person who literally goes to the gym by the fact that I would want to be a member of the gym, I'm talking about friends of mine, not myself, but that's a hint that correlates with other healthy behaviors. And so, I want those people in my insurance pool.

0:32:43.3 AF: Exactly. Exactly.

0:32:44.2 SC: This is very complicated. This is why physicists would not be very good at it. Physicists are much better at simple things. And there's a lot of confounding variables coming in here. But let me back up again to once again sort of understand what's going on with the idea of insurance and how it feeds into the aspects of it we just mentioned. The idea is to spread risk. When you buy the insurance, presumably you're not buying it from your friend. You're buying it from a much larger company who can live through the fluctuations up and down.

0:33:18.1 AF: Exactly. Exactly. Again, the idea is you take 100 people or a million people, each of whom has a one in 10 chance of losing that $100 and you spread that risk around. The law of large numbers will work. On average, one in 10 of us will lose that $100. But if we all pool together, then not all of us will lose it. And if we all give a little bit of money, that can cover the losses for the few, the one in 10 who do lose it. So, it's exactly this sort of making it a communal rule, smoothing the risk or spreading the risk, as you nicely said, across the community.

0:34:00.4 SC: Am I misremembering or was I misinformed that there were companies in the stock market crashes or financial crisis that had insured themselves, which is sort of almost a contradiction in terms?

0:34:14.9 AF: Yes. So, there is something called reinsurance, which is somewhat bizarre at first glance and maybe even at second glance, which is insurers in turn reinsure their risk. Sometimes the reasons are because there is some risk of a correlated shock. So, the example I gave with the 10% chance of losing $100 where it's just totally each person has a one in 10, 10-sided coin flip chance, 10-sided die roll, what have you. In that case, it's going to be very unlikely with a large enough pool that, say, everyone gets wiped out. But on the other hand, if you're, say, an insurer who's insuring against natural disasters, earthquakes, floods, et cetera, and you're sort of geographically concentrated in, say, Florida or California, some years you may have no costs and some years you may have 80% of your customers have a claim. So, then in some sense, you may want to lay off some of that risk to a larger pool that's more diversified across either other types of insurance or geographically or has a larger capital stock and can bear that risk.

0:35:28.5 AF: And so, it's clear why as the seller of insurance, as the insurance company, we really do want to have a good statistical grasp on what the pool is that we're selling to, what the risks are, et cetera. I mean, how much like super fancy math goes into that? Is there a lot of power laws and long tails and heavy duty statistical techniques here or is it more or less just don't make the mistakes that have been highly make and you'll be fine?

0:35:54.8 AF: I think it's both. I think for low probability, right tail events, this came up a lot with terrorism insurance after 9/11. There's also the sort of weird, it was some of these things where there's uncertainty about the risk. We know a lot about mortality risk. We know a lot less about, say, terrorism risk. And therefore, if an event occurs, you as the insurer or the public would update your probabilities of a future event occurring. So, some of that actuarial science, some of it's quite straightforward and the pricing mortality tables at this point we've been doing for a while, some is much more complicated. But all of it, it's not enough to get the actuarial science, right? You then have to think about who's going to be buying that product and how might they differ from the typical population on which you've calculated those probabilities.

0:36:53.4 SC: So, clearly, as I said, the insurance company is going to want to know its customers as well as it can. And you already mentioned this or alluded to this, but are there any circumstances in which the insurance companies know more than their customers about their customers' own risks?

0:37:13.0 AF: That's a really great question. I don't know for sure, but we've thought about this a lot, and the way we think about it is it's not a less-versus-more thing, but at some level, of course, a life science actuary is going to have a much more precise prediction of my mortality than I am. Looking up a life table for someone of my age and gender and maybe a few other characteristics would probably do better than I could do off the top of my head. So in that sense, you could say, yeah, the insurance company or the actuary "knows more." On the other hand, the way my informational advantage comes in is whatever the insurer knows, they use to offer the insurance.

0:38:06.4 AF: And they say, okay, here's the price at which I'm going to offer insurance to middle-aged female economists with a great sense of humor and two bratty kids or what have you. I might not have known how to price that, but then once they give me the price, I know that among that group of middle-aged female economists with great senses of humor, I know something about whether I'm more or less risky than the typical. That's where my informational advantage comes in. It's not like if we had a competition of who could get the actuarial tables right that I do better than they would. They move first and then I get to react.

0:38:52.9 SC: Yeah. Okay, good. Maybe we should think about the obligations of the person to reveal some of these things, or maybe the obligations go both ways. My understanding is that if the insurance company asks you about your medical status, you have to be honest with them. You can't lie, but you don't need to volunteer things that they don't ask you. Is that roughly the idea?

0:39:19.8 AF: That's correct. And just to be clear, you could lie. I wouldn't advise it because it's very expensive to verify all this information. My understanding is, for example, with life insurance, they do do a medical exam, but for many things, they will take your self-reported word for it. But then if you actually have a claim, then they'll investigate. And so it's a very, very bad idea to try and shade the truth because you'll end up paying all these premiums. Then in the event that you actually need the insurance, then you could be denied. I guess, even if you... As an ethical matter, I don't know why it's my responsibility to tell someone more about my risk than they ask. I need to be truthful. But again, as you were talking about earlier, it's privately rational for us all to do that. It just messes things up in aggregate. One of the examples we give in the book is with auto insurance. Now, the saying is 90% of people think they're above average drivers. Well, I'm one of the 10% who know I'm a below average driver.

0:40:36.1 AF: But when I first got a car, which was in my late 20s in graduate school and I needed to get automobile insurance, I didn't think it was my job to tell them that. I answered all their questions truthfully. They asked a bunch of questions like the make and model of my car. Then they asked how long I'd had my license and whether I had any accidents or speeding tickets. The truthful answer was that I had had my license for 10 years and I had an absolutely spotless driving record. What the insurance company didn't know that I knew and anyone who spent any time in a car with me would quickly learn is that I'm a driver and the reason I had a spotless driving record is because I never drove. I had gotten my license as a 17-year-old growing up in Manhattan because my mom told me I really should get it. It would be useful to have the ID on the driver's license. I got the license and I never drove. I answered every question truthfully. It was just something about me that was kind of idiosyncratic that they didn't know.

0:41:40.4 SC: Yes, but they do know that you live in Massachusetts. I bet that they assume that you're a below average driver just on the basis of that data.

0:41:48.5 AF: Exactly, but I knew I was a below average driver even among Massachusetts drivers. That's exactly my point. They probably knew much better than me what the empirical rate of accidents are in Massachusetts, but I knew mine would be higher.

0:42:02.5 SC: Speaking of the ethical obligations, those are the ethical obligations of the insured. For the insurers, how much are they allowed to just discriminate, to just say, "Well, look, I know that you're a very, very risky group, so your premiums are much higher." That's good business sense. How legal is it?

0:42:19.0 AF: That's a really good question. Again, with the caveat that I am not a legal expert, our understanding, and we talk about this in the book, is that what insurers are allowed to price on varies a lot both state to state and also insurance to insurance. Some insurance are more heavily regulated than others. The thing that we talk about a lot in the book is not the legal matters and not even really the ethical matters, although we try and talk about that a bit, but more just from an economic perspective, what are the consequences of saying, "No, you can't price health insurance on the basis of preexisting conditions." Let's take that as an example. Now, on the one hand, we totally understand where this comes from, that it seems unjust or unfair that people who are unfortunate enough to have some chronic health condition, that in addition, they have to pay more for their health insurance. On the other hand, what we have seen happen, and we talk about this as well in the book, is that when states come in, for example, New York and New Jersey did this in the 1990s, and in the private health insurance markets, the ones that aren't through your employer, they said basically, "Yeah, we don't think it's fair, and you can't price health insurance differently for people based on their preexisting conditions. You actually can't charge higher prices based on anything.

0:43:47.6 AF: You have to offer the same price, it's called a community rate, to everyone in the community. It's so-called guaranteed issue. You also have to sell to everyone who wants the insurance. You can't say, "Well, here's the price, but only for you, Sean, and not for you, Amy. We don't have insurance for you." Well, this guaranteed issue is guaranteed to create issues. What we saw happen there is exactly what you'd expect. If everyone has to be charged the same price and offered insurance, then on average, the people who buy the insurance are the older, sicker individuals. That drives up the price of that insurance, because insurance companies have to be getting in enough to cover their claims. That drives the healthiest of those older, sicker individuals out of the market and drives the risk up more and the price up more. In the case of New York and New Jersey, it drove the market completely out of existence.

0:44:43.0 AF: So, yes, it was fairer. No one was being treated differently, but everyone was in the same poor situation of having no access to this health insurance product. Because economists always love, on the one hand, on the other hand. Let me also give you the other hand, which is it sounds, if you're trying to deal with problems of private information and selection, from a pure economics point of view, it can sound nutty to say, "We're going to restrict what the insurance company can use in pricing."

0:45:14.5 AF: That's like creating private information for the customer. On the other hand, if what you want is not insurance for my claims this year, given that I'm a relatively unhealthy individual, but I want insurance against being an unhealthy individual, behind the Rawlsian veil of ignorance, then it makes perfect sense to have community rating charge everyone the same. Because that's to reflect the fact that if we charge higher premiums to people who are higher cost, to people who are worse drivers or less healthy, then we lose the insurance or the potential for insurance against the possibility of being unlucky enough, such as myself, to have been born a terrible driver or someone who's born with a genetic disease or a predisposition to certain chronic conditions. And that's one of the reasons you can think that the government may get involved to say, look, we don't know how to, or I'm not a philosopher, but I don't think we know how to literally buy insurance behind the veil of ignorance in utero, as it were, but that's actually a role that the government can play.

0:46:22.1 SC: Good. So that's the perfect place to get into that. So just to summarize, we could imagine letting the insurers use all the information they have and discriminate as much as they want, but that seems a little unfair. The people who need the insurance products the most are being charged or the least able to afford it are being charged the most. So we could try to make it fair by insisting that everyone be charged the same, but then we have the selection effect where only the riskiest people are buying it. So the next obvious twist to put on it is to say, now everyone has to buy insurance, right?

0:46:54.0 AF: Correct. So rather than just require that the insurers offer it to everyone at the same price, right, which gets you your guaranteed issues, you can instead also insist that everyone buy it at that price. So whether that's the government providing it directly or as in the case of automobile insurance and most recently health insurance mandating that everyone provided. And that's why going back to the Scalia question about, can you force everyone to eat their broccoli if you can force them to buy health insurance, those are two very different situations. But one thing I learned or relearned I should say from watching the experience of the health insurance mandates play out is both that the economic theory really does apply in practice, but also perhaps not surprisingly, it is true that the real world is more complicated than what's dreamed of in our philosophy. And so what we saw, so I think two main lessons came out of thinking about these mandates. So the first is the textbook model is that mandates are a way to solve the selection problem because if everyone has to have insurance, then by definition, there's no selection. Everyone's in, you've got your random or representative pool.

0:48:09.7 AF: What I hadn't really thought about until I watched this play out in practice is, again, to use the broccoli analogy, anyone who has young children knows that saying you must eat your broccoli to your toddler doesn't make it so. And similarly, saying that people must have health insurance, mandating it, also it turns out doesn't make it show. And what turns out to matter a lot, and there's evidence of this from Massachusetts, is what kinds of penalties are put in place for people who don't buy insurance. And when those penalties are in place, you see that that's much more effective at getting the healthy customers into the market than simply saying you must buy insurance. So there's the old Robin Williams joke about the cops in England, the Bobbies as they're called, they don't carry guns. And so he deadpans. Well, what do they do when they're arresting a criminal? They shout, "Stop or I'll say stop again," right? You know, like it's not that effective. And similarly, saying you must have health insurance without putting some teeth behind that is similarly not as effective.

0:49:19.2 AF: So, again, it's not just enough to mandate, you have to enforce the mandate, which may sound obvious, but to get quickly back to the types of issues that we were saying what is and isn't, what is or isn't fair, one way to enforce a mandate is to fine people for not complying. The other way, instead of the stick is the carrot, it's to subsidize people for complying. And that's a lot of what we've done effectively under the Affordable Care Act is especially for lower income individuals offer subsidies for people to buy health insurance. One of the things economists have pointed out, which is true, but also you can see why it would be unpopular is if you're trying to design those selections, excuse me, if you're trying to design those subsidies to deal with the selection problem, the people that you should be subsidizing are the people who, in some sense, you feel like might need the subsidies the least, right? It's not the old sick people. You want to subsidize the young, healthy people. That's actually how you'd solve the selection problem. Not clear that's going to fly past the sniff test with Joe Public.

0:50:31.9 SC: Well, Joe Public does get in the way a lot, but Joe Public also there's a folk wisdom there. I mean, one of the off-putting aspects of the Affordable Care Act, even though I thought it was a great step forward myself was, oh my goodness, there were a lot of complications and bureaucracy and paperwork. And that is both intrinsically off-putting, but it also just opens up new venues for malfeasance, right? I mean, you mentioned in the book about what a huge amount is spent on lobbying the government by insurance companies and people in that sector. And this is part of the reason why, right? When there's a lot of rules, there are a lot of loopholes.

0:51:11.6 AF: Absolutely. This is another thing we talk about in the book that because sometimes the selection problem, when the government gets in and tries to fix things, the selection problem can cease being a game between the customer and the insurer and can become a game between the government and the insurer. And then the government is often on the losing side of this. So right again, so one example that we discuss in the book is the public Medicare program, which provides health insurance to individuals 65 and over automatically through the tax system. There's an option instead that was created in the '80s and has become much bigger in recent decades to try to allow people more choice in the type of health insurance they have and to also allow the private market to maybe reap some of its potential efficiencies, which is you can let private insurers offer Medicare and people may prefer the product that they're offering. It has to meet some minimum requirements of the regular Medicare program, but then they can also provide additional benefits.

0:52:21.3 AF: And then the idea is that the government, instead of paying for your Medicare, pays the private insurer to insure you instead. So in other words, rather than the government bearing all of our medical costs, it sells some of our risk to a private company who thinks they can manage your risk better. So that sounds perfectly fine. Until then, you have to think about, well, how much should the government pay a private insurer to take a Medicare customer off its hands if the customer wants to go to that insurer? And what the government would like to do is pay them what they would have spent to cover this person with Medicare, or maybe even a little less if the government wants to save money. Okay. So it sets a price. It says, look, older people are more expensive. Maybe women are more expensive than men. And then guess what? That the private health insurer is going to want to find the healthiest people by age and gender. And they may do it by offering. In fact, they do it by offering things like gym memberships and other strategies like that. And so then the government, when we talk about this in the game, we talk about it as sort of a game of whack-a-mole.

0:53:32.3 AF: Well, the government got smart to that and started collecting data on your past medical claims. Because one thing that's going to predict your future medical claims is your past ones. If I've been to the doctor and hospital a lot over the last year, I'm likely to be high cost this coming year. So they start pricing then on additional things. And that can help with the problem, but it doesn't solve it entirely. And round and round you go with the insurer always kind of staying one step ahead of the government. Whenever the government does, the insurer then spends some time trying to think about how it can, as we say, cream skim the best customers.

0:54:13.2 SC: Do other countries have to put up with all this? We've been talking about the United States. We both live there. But we have international listeners here on Mindscape. Is this just a global phenomenon that is a result of the structure of insurance and markets? Or is the United States a special outlier?

0:54:30.3 AF: So selection is a global problem that's a function of markets. And not only insurance markets, although it's one of the biggest examples and we try and convince our readers one of the more interesting ones, selection markets can occur whenever the different customers are going to impose different costs on the seller and the customer knows something that the seller doesn't. So we give examples in the book of saying all you can eat restaurant or education loans that universities have tried to offer in which the so-called income sharing agreements in which the amount you have to pay back depends on how much you make after college. And what you see, and the economists Nathan Hendren and Daniel Herbst have shown this very nicely, is that even when you look at what college people are going to and what their major is, people still have private information about how much they're going to make after college. And guess who wants to buy the income sharing agreement?

0:55:32.3 AF: The ones who aren't going to make very much. So it's a pervasive problem that's not about the country or even the particular market. That being said, although this is probably more a topic for a different podcast, it's also the subject of our next book, in the particulars of health insurance, the way the United States has tried to deal with this problem is unique. And we have come to conclude uniquely problematic relative to how other high-income countries have tried to deal with this.

0:56:05.0 SC: Yeah, that's a topic for another podcast but an extremely believable claim given our empirical information. But with all that in mind that we've been exploring here, is there... This is probably an unfair question. Is there a simple set of things that we should obviously be doing better here in the United States for the world than we are doing?

0:56:25.5 AF: We being the government or the customer?

0:56:27.8 SC: Society. [chuckle]

0:56:30.3 AF: I'm sorry.

0:56:30.7 SC: Society. The whole. The totality. The system.

0:56:32.9 AF: Society. Yeah. So I would say for society, no. I think one of the messages of our book, which may make it not a popular sound bite on late night TV but is an important message, is there aren't facile solutions. There really are trade-offs. Take even the case of mandate that we've talked about already and think about perfect enforcement of the mandate, so we don't have to deal with that issue. There's still the question of how much insurance should the government mandate? And you see, for example, in automobile insurance, the minimum requirements vary drastically across states. Now, if you mandate a very, very low amount of insurance coverage and most people want more than that, you're just kind of recreating the selection problem, but now it's on the margin of how much insurance you buy rather than whether you buy it at all.

0:57:29.7 AF: On the other hand, if you set very high minimum requirements as some states do, then you have the problem that maybe not everyone actually wants that amount of insurance. So really these are not issues that sort of are straightforward. Another example that I think is kind of an interesting and important one is the problem with selection and with insurance markets is that people don't have choice of the products that we think they should have available to them. And yet, one of the ways we deal with that problem is to try to restrict people's choices. So [chuckle] we give the example in the book, there was a sort of scathing New York Times expose that got our blood boiling where it talked about a very unfortunate case of a man who when he was 65 and getting on to Medicare at first was quite healthy and therefore was attracted to those gym memberships and the private plans and got one.

0:58:31.5 AF: And then unfortunately, about seven years later got quite ill and wanted access to additional physicians and other things that his plan didn't cover and tried to go back to the regular public Medicare program and buy a supplemental private policy to cover some of the gaps, and he couldn't. No one would sell him one. And the reason was is because when you're 65, the regulation says that... Again, you have to offer these plans to everyone, but the insurer. But then afterwards, you don't. And that's a way of trying to not make the market collapse, have the always-guaranteed issue like New York and New Jersey had but also prevent rampant selection. And so yes, it was very unfortunate that when he was sick, he could not get a more comprehensive policy, but that's also how you prevent selection from wreaking havoc on a market. If you let people wait till the risk occurs to buy the insurance, then you can destroy the market for everyone.

0:59:32.0 SC: Is there some argument that the government should just give a basic insurance plan to every person?

0:59:39.2 AF: There is. That's the subject of my next book that's coming out this summer with my co-author Liran Einav from this one. But spoiler alert, the argument it turns out really isn't about selection. As much as we went into it thinking it would be 'cause that's what we think about, it's a topic for a different day. It's a very different...

1:00:00.5 SC: Okay. We'll look at that one. [chuckle]

1:00:00.6 AF: I'm sorry.

1:00:01.1 SC: We'll look for that book. It's a very interesting question to think about. So there's two more things to get to before we end. One is, we were just hinting at it with the example you gave of the older gentleman trying to get insurance. We're not very good as a society or maybe even as a species in planning for our last days. We don't wanna die, most of us. We're willing to spend the money we have to do it. There seems to me as someone who is not there yet, but eventually will be, that we have a system in the United States where there are ways that all of your wealth disappears when you get old and sick because you're willing to spend as much as you can to stay alive and hospitals and doctors insurers are willing to take it. And we haven't quite figured out how to deal with that.

1:00:52.4 AF: Yes, and without denying that that is a real problem, let me give you though a counterpoint which in some work I've done also with Liran Einav who's my co-author on this book as well as some other economists, which is to try to look into this concern that comes up a lot about all of the healthcare spending that occurs at the end of life. So the fact that you might have heard that's often quoted is that a quarter of all medical spending on the elderly occurs in the last 12 months of life, and it turns out that is actually true. So the first thing whenever I hear something a lot, it sounds inflammatory, I wonder if it's true. We've checked, and that is a true fact. But the next thing is, if we ask... Well, there's two reasons for that. One is because you spend more when you get sick, and people who are sick have higher death rates. They're more likely to die.

1:01:53.4 AF: And that just seems plausible. Like of course, we spend more healthcare money on sick people than healthy people, but if you... But what's behind the concern about that fact is the assumption or sometimes assertion that we're spending money on people who we know are about to die. So they're in intensive care, bedridden, hooked up to many machines and clearly going to die. It's only a question of, "Is it today or next week or next month?" And we're spending all this money to keep people alive for a few more days. It turns out if you look in the data and you use fairly sophisticated machine learning techniques and all their past health records to predict their mortality, it's very, very hard to know at the time that most of this medical care is occurring, that people have a very, very high mortality.

1:02:49.1 AF: So in other words, even if you look... And we did this among people who have been diagnosed with cancer and are at a fairly late stage of cancer, right? Even if you look at 80-year-olds with cancer, there's not many of them who at the time they're getting their medical treatment have a 70% plus annual mortality rate. So just it's... It's very, very hard to identify a group of people that we're spending a lot of money on who are actually going to die with 90-95% probability. So yes, death probabilities are elevated among the sick, and yes, we spend a lot of money on the sick, but we're not spending money on people that we... On a large group for people who we "know" with high probability are going to die.

[chuckle]

1:03:36.4 AF: The example I always think of in trying to explain this, which obviously is a little silly but I think kind of drives home the point is when my kids were learning to ski and they went to ski school, one of the things the ski instructor always would say is, "You really need to be careful towards... Because most ski accidents occur on the last run of the day." [laughter] To which I thought, "Well, of course they do because once you have an accident you stop skiing." [chuckle] So again, it doesn't mean that your probability of a ski accident is necessarily higher at 4 o'clock.

[laughter]

1:04:15.2 SC: Well, oh no, that's actually very, very helpful data or information. It's visually or viscerally, I suppose I should say, a very different idea to think that someone has a 95% mortality rate and we're spending millions of dollars versus someone even with a 50% mortality rate, like if it's only 50-50 odds...

1:04:33.5 AF: Exactly. That's a high rate, but that's not dying for sure.

1:04:37.6 SC: I'm gonna keep fighting if I have 50-50 odds. Yeah, that's especially if I could maybe live for another 10 years. That's an enormous amount of time. Okay.

1:04:46.1 AF: Exactly.

1:04:47.1 SC: These problems are hard. Again, I really should go into physics instead. Physics is so much easier and less emotionally taxing. But speaking of physics in the very broadest sense, the other question I want to bring up before we go is you've alluded a couple of times to asking these questions, which are not only intellectually interesting, but morally and politically charged and collecting data to see what the answers are because the answers aren't always what we might guess or what our theories would predict. And apparently this is not always the way that people reach their conclusions. You've given talks. You've written about the need for more empirical testing and the difficulty of empirical studies within economics.

1:05:32.2 AF: Yes, I think one way to think about it is, we all have our hypotheses, our prior beliefs, sometimes even our hope as to what would be the effect of a policy, but it's incredibly important when we're doing academic research or when the public is trying to find out what the effect of a policy is from that research that we take great care to make sure that we are getting cause and effect and no spurious correlation. And so there's been a real revolution, I would say, in empirical work. My colleague, Josh Angrist, who won the Nobel Prize a couple of years ago for some of this work, refers to it as the credibility revolution [laughter] in trying to make sure that we're isolating causal effects. And one of the most valuable tools in the work I do, which is mostly in health economics for doing that is something that to my other colleagues who also won the Noble prize. [chuckle] It can be hard on the ego hanging out here, but...

1:06:43.3 SC: Come on, Amy. [chuckle]

1:06:45.4 AF: Who won the Nobel Prize in development economics for popularizing this approach for policies designed to try and reduce global poverty, but it applies more broadly to any policy, is the randomized experiment, to literally do what we do in medicine all the time. If you wanna know if a drug worked, we run a randomized trial and we look at what the effect is of giving the drug to one group and not to another. One can do the same thing with any policy question. I mentioned earlier, we did this with the case of what is the effect of covering low income uninsured people with Medicaid? You can do a randomized trial. This was done in the State of Oregon in 2008. Kate Baker, and I studied it, in which the state randomly allocated Medicaid to some low-income uninsured adults and not to others.

1:07:37.5 AF: And then you can see what happens. And I think it's incredibly important, whatever your theories or your biases or your hypotheses or your hopes are, to look at the data and learn from it. One of the things many people had claimed was, if we expand Medicaid health insurance to previously uninsured individuals we'll get these individuals out of the emergency room where they have to be by law be treated and into cheaper and more effective primary and preventive care. And what we saw from this randomized controlled trial is that, yes, covering the low-income uninsured with Medicaid does get them into more primary and preventive care, and they have more doctor visits. Their rates of mammograms go up. Their rates of diabetes checks for... Blood sugar checks for their diabetes go up, and yet they also go to the emergency room more.

1:08:35.6 AF: So as we talked about earlier, health insurance makes medical care cheaper, and one thing that does is get people to use more of it. Whether that's a good thing or a bad thing, we could have a different discussion about, but you shouldn't be arguing for health insurance expansions on the grounds that it's gonna get people out of the emergency room.

1:08:54.6 SC: Is there in economics some kind of active resistance to emphasizing these kinds of empirical results or it's just a recognition that it's hard to tease out all the causal relationships?

1:09:05.8 AF: It's a good question. I don't know of anyone who's against... I can't think of why anyone would be against causal evidence and causal inference. I think the problem is that often it's really hard, and the challenges come in is how do you form your most educated guess when you don't have rigorous causal evidence? And there I think reasonable people can take different approaches. I think the other thing that's hard is there are many well-intentioned people in the world outside of economics, who want... Who hope for the best, who hope that expanding health insurance can be an easy decision because it'll not only improve people's lives, but save us money. And having to tell people that actually, "No, there's no free lunch. We face real trade-offs." And it may well be worth it to spend that money, but it's not gonna be free. I think there's a lot of resistance in that sense as well from well-intentioned people, but who wish that we didn't face as hard choices as we sometimes do face.

1:10:09.9 SC: Okay, for the very, very final question then, to give the audience a tangible take-home message. Given that you're the world's expert in all this stuff, what kind of insurance do you have that would be different to anybody else's? Does this actually make you change your mind about the annual enrollment at MIT when you're giving your options for life insurance and health insurance and dental insurance?

1:10:32.4 AF: Well, I will say that one thing it really prompted me to do is to try... When we bought our first home, to get a homeowner's insurance policy with as high a deductible as possible, the reason being home insurance as with most risk is most valuable for the catastrophic risk, your home burns down or a large section of it is destroyed. That's a real financial shock. And yet, despite the rather high price of homes in the Boston area, a lot of the homeowners policies have a $500 deductible, $1000 deductible. That's real money, but it's kind of trivial compared to the cost of a home. And so I searched high and low for a homeowner's policy with the largest homeowners deductible that I could find, a $10,000 deductible because the premium was a lot lower. And I would argue based on the type of research I do that the reason the premium was a lot lower is because people who know that there's a lot of problems with their home and they're gonna have a lot of claims want a low deductible.

1:11:42.3 AF: Now, going back to supposedly being an insurance expert, I will say that shortly after we bought this enormous deductible, relative to what you could find, policy, there was the winter of 2015 where if you spent any time in Massachusetts, there was massive amounts of snow. It was declared a federal disaster area. Everyone had leaks, et cetera. We got, as did everyone... We had to get contractors in to repair the damages, and we blew well past our $10,000 deductible. The next year, we had a flood in our basement, blew past the deductible again, and it did cause my long-suffering husband to observe that it could be quite costly to be married to an insurance expert. [chuckle] But actually he will admit that if you look over all the years we've had the policy, even with those two big losses, the lower premiums we're paying every year on the high deductible policy have in the end paid off.

1:12:41.3 SC: You gotta do that. I once knew a guy who won a lottery ticket, came in with a $200,000 prize, and I asked him how often he bought lottery tickets, and it's this easy calculation that he was losing money even if there had been no taxes at all, so...

1:12:55.0 AF: Wow, wow.

[chuckle]

1:12:55.9 SC: Yeah, I know. [chuckle] All right, good. Both intellectually challenging, socially important and practically useful information here. So Amy Finkelstein, thanks so much for being on The Mindscape Podcast.

1:13:07.3 AF: Thank you so much for having me.

[music]