John Cochrane (famous Chicago economist) gives an interesting podcast on health care with host Russ Roberts. You can find more thorough remarks in this paper. I find myself amenable to both free market critiques and lefty critiques of our health care system, because I think the U.S. is at a local minima. Moving a long distance toward more market oriented Singaporean health care or a single-payer Canadian health care would both be improvements.
I was hoping Russ would press him on one point I can't comprehend, which is central to Cochrane's argument. First, we agree that insurance should be long-lived and guaranteed renewable. Meaning, if a person gets sick, the insurance company can't drop them or raise their premiums. This makes perfect sense to me. The reason we buy insurance is to insure ourselves from risk. In this case, the risk of getting really sick in a way that will cost a ton of money for care. If I buy insurance that only lasts one year and I get cancer, I'll get the benefit for one year, but the next year when I try to renew my premiums will go sky high. What good is that? The answer is obvious. I need to buy "insurance insurance." I need insurance against the risk that my premiums rise. Of course, if insurance insurance isn't long lived, I'll need to insure the risk my insurance insurance premiums rise and this rabbit hole goes on and on. It's turtles all the way down. This logic leads to the idea that insurance should be long lived. Rather than buying an infinite number of insurance contracts, I can buy on contract that never expires.
What I can't figure out is this statement?
Suppose that two parents have a kid that is covered on their insurance, until he's twenty-five. Then he has to go to the market and buy his own insurance. Now suppose when he's twenty-four, he gets cancer. In John's world his premiums should skyrocket. I think he would argue that the parents should buy "insurance insurance" for the kid. The parents need to buy insurance against the kid's future high premiums, before he gets sick. Obviously, they can't do that after he's diagnosed. The "insurance insurance" would be too expensive. They need to do it earlier. Perhaps, when he is born. Though what if we find he has complications? Perhaps, when he's conceived? But what if the parents have genetic predispositions to certain diseases? The kid is more likely to become sick, thus the insurance on his future premiums must be higher. Maybe they'd be high enough that the two parents wouldn't get married, because of the possible complications. Oh no, two people who love each other can't get married. What if the grandparents stepped in and helped out? They could buy "insurance insurance insurance," they could hedge the risk that their son or daughter grew up wanting to marry someone who wasn't as genetically compatible.
Maybe this sounds convoluted, but I'm trying to get to the notion of complete markets. Complete markets means that all risk that can be insured is insured. Since we are risk averse, complete markets make us all better off. If one of our houses burns down, we'll all chip in $100 bucks to get us back on your feet. The Amish did it literally, but nowadays we do it through insurance companies. The premium on home insurance is just the $100 bucks that goes to some unlucky person that just watched their home go up in flames.
But in a truly complete market, ALL risk is insurable. I could insure against a risk, before I was ever born, before I ever had a medical history. In a truly complete market, we insure all risks at "time zero," before we even exist. I picture a random soul in heaven, just like all other souls. One of the risks those souls want to insure against is sickness. Maybe being born sick. Maybe being sick later in life. Basically, they don't know when it will or if it will happen, but there's some risk they'll get sick. So they get with all the other souls and agree, we'll all put money aside, and whoever gets sick gets the money. That's insurance. My question is how much would each put in. Well, it's obvious they'd all pay the same. They are exactly alike in every way. It's only fair.
The point I'm trying to make is that in John's system, no matter how you slice it, there will be some choke point along the way. What if we switched from our system to John's tomorrow? A lot of healthy people would buy guaranteed renewable insurance, but a lot of sick people would be out of luck and face exorbitant premiums. If we are in a system where everyone is paying a different price for insurance, then we can't possibly be at the complete markets case. If we aren't at the complete markets case, either John's policy is suboptimal or there is some friction that prevents us from getting to the complete markets case. So what is it?
The wrench that I haven't thrown in yet is incentives. If I have guaranteed insurance before I am born, I'm much less incentivized to put effort into maintaining my health. I may drink like a fish or eat too much. I may smoke or do drugs. Maybe I just won't sleep enough. We make decisions that affect our health. This is a "friction." This is a reason that the complete markets case may not work (as if the having to buy insurance before your born wasn't enough). Maybe this is the sort of important friction John has in mind, because for conditioning on pre-existing conditons he gives these examples in the podcast, "Fat people don't pay more than thin people; smokers don't pay more than non-smokers. All sorts of things they don't condition on." Russ corrects him that they can condition on smoking, which in my understanding isn't that big of a bump in premium. I don't have a problem with insurance companies conditioning on smoking or charging a higher premium on people like me that eat and drink too much. It's the all sorts of other things I don't want. What about something simple like being female? Women have to pay higher insurance premiums, because women have babies. In a complete market, women would pay the same premium as everyone else. They'd buy their insurance before they had a gender (or even parents) and pay the same price as everyone else.
Conditioning on effort is fine to the extent that an insurance company can do it. But what they really want to do is condition on how sick we are. People with cancer pay more, and I don't see John distancing from that. I can't imagine by pre-existing condition he really means "big sweet tooth."
I thought I found a lot to like in the essay. I think there are lots of regulations that make health care more expensive and that the health care market should be more competitive. But I was really jarred by this statement, "Romney was pushed on what he was going to do, he said: Well, but of course we can't let insurance companies discriminate on pre-existing conditions. Well, once you've done that, you've swallowed the fly--the old lady that swallowed the fly...And once you do that, you've got most of the ACA." He's really arguing that my disagreement is irreconcilable with his preferred health care solution. Which makes sense, if my story is correct then charging everyone the same price and mandating everyone buy insurance sounds pretty close to the preferred solution. Since in a truly compete market, we'd all pay the same health care premiums. His preferred solution must be a market frictions story, but its a story I can't understand and he spends a good deal of the podcast and paper dismissing frictions in the health care market. If there is a well-formed second best model there, I'd love to know it.
I was hoping Russ would press him on one point I can't comprehend, which is central to Cochrane's argument. First, we agree that insurance should be long-lived and guaranteed renewable. Meaning, if a person gets sick, the insurance company can't drop them or raise their premiums. This makes perfect sense to me. The reason we buy insurance is to insure ourselves from risk. In this case, the risk of getting really sick in a way that will cost a ton of money for care. If I buy insurance that only lasts one year and I get cancer, I'll get the benefit for one year, but the next year when I try to renew my premiums will go sky high. What good is that? The answer is obvious. I need to buy "insurance insurance." I need insurance against the risk that my premiums rise. Of course, if insurance insurance isn't long lived, I'll need to insure the risk my insurance insurance premiums rise and this rabbit hole goes on and on. It's turtles all the way down. This logic leads to the idea that insurance should be long lived. Rather than buying an infinite number of insurance contracts, I can buy on contract that never expires.
What I can't figure out is this statement?
Now, the “adverse selection” phenomenon, that sick people are more likely to buy insurance, and healthy people forego it, is a big problem. But the insurance company charges the same rate, not because it can’t tell who is sick – a fundamental, technological, and intractable information asymmetry. The insurance company charges the same rate because law and regulation force it not to use all the information it has. If anything, we have the opposite information problem: insurers know too much.As best I can tell by following his links, health status insurance is the insurance insurance I just discussed. So John's big idea is the asymmetric information isn't a big deal, the real problem is that insurance companies can't charge sick people higher premiums.
This source of adverse selection is a legal and regulatory problem, not an information problem, and easily solved. If insurance were freely rated, nobody would be denied. Sick people would pay more, but “Health status” insurance shows how to solve that.
Suppose that two parents have a kid that is covered on their insurance, until he's twenty-five. Then he has to go to the market and buy his own insurance. Now suppose when he's twenty-four, he gets cancer. In John's world his premiums should skyrocket. I think he would argue that the parents should buy "insurance insurance" for the kid. The parents need to buy insurance against the kid's future high premiums, before he gets sick. Obviously, they can't do that after he's diagnosed. The "insurance insurance" would be too expensive. They need to do it earlier. Perhaps, when he is born. Though what if we find he has complications? Perhaps, when he's conceived? But what if the parents have genetic predispositions to certain diseases? The kid is more likely to become sick, thus the insurance on his future premiums must be higher. Maybe they'd be high enough that the two parents wouldn't get married, because of the possible complications. Oh no, two people who love each other can't get married. What if the grandparents stepped in and helped out? They could buy "insurance insurance insurance," they could hedge the risk that their son or daughter grew up wanting to marry someone who wasn't as genetically compatible.
Maybe this sounds convoluted, but I'm trying to get to the notion of complete markets. Complete markets means that all risk that can be insured is insured. Since we are risk averse, complete markets make us all better off. If one of our houses burns down, we'll all chip in $100 bucks to get us back on your feet. The Amish did it literally, but nowadays we do it through insurance companies. The premium on home insurance is just the $100 bucks that goes to some unlucky person that just watched their home go up in flames.
But in a truly complete market, ALL risk is insurable. I could insure against a risk, before I was ever born, before I ever had a medical history. In a truly complete market, we insure all risks at "time zero," before we even exist. I picture a random soul in heaven, just like all other souls. One of the risks those souls want to insure against is sickness. Maybe being born sick. Maybe being sick later in life. Basically, they don't know when it will or if it will happen, but there's some risk they'll get sick. So they get with all the other souls and agree, we'll all put money aside, and whoever gets sick gets the money. That's insurance. My question is how much would each put in. Well, it's obvious they'd all pay the same. They are exactly alike in every way. It's only fair.
The point I'm trying to make is that in John's system, no matter how you slice it, there will be some choke point along the way. What if we switched from our system to John's tomorrow? A lot of healthy people would buy guaranteed renewable insurance, but a lot of sick people would be out of luck and face exorbitant premiums. If we are in a system where everyone is paying a different price for insurance, then we can't possibly be at the complete markets case. If we aren't at the complete markets case, either John's policy is suboptimal or there is some friction that prevents us from getting to the complete markets case. So what is it?
The wrench that I haven't thrown in yet is incentives. If I have guaranteed insurance before I am born, I'm much less incentivized to put effort into maintaining my health. I may drink like a fish or eat too much. I may smoke or do drugs. Maybe I just won't sleep enough. We make decisions that affect our health. This is a "friction." This is a reason that the complete markets case may not work (as if the having to buy insurance before your born wasn't enough). Maybe this is the sort of important friction John has in mind, because for conditioning on pre-existing conditons he gives these examples in the podcast, "Fat people don't pay more than thin people; smokers don't pay more than non-smokers. All sorts of things they don't condition on." Russ corrects him that they can condition on smoking, which in my understanding isn't that big of a bump in premium. I don't have a problem with insurance companies conditioning on smoking or charging a higher premium on people like me that eat and drink too much. It's the all sorts of other things I don't want. What about something simple like being female? Women have to pay higher insurance premiums, because women have babies. In a complete market, women would pay the same premium as everyone else. They'd buy their insurance before they had a gender (or even parents) and pay the same price as everyone else.
Conditioning on effort is fine to the extent that an insurance company can do it. But what they really want to do is condition on how sick we are. People with cancer pay more, and I don't see John distancing from that. I can't imagine by pre-existing condition he really means "big sweet tooth."
I thought I found a lot to like in the essay. I think there are lots of regulations that make health care more expensive and that the health care market should be more competitive. But I was really jarred by this statement, "Romney was pushed on what he was going to do, he said: Well, but of course we can't let insurance companies discriminate on pre-existing conditions. Well, once you've done that, you've swallowed the fly--the old lady that swallowed the fly...And once you do that, you've got most of the ACA." He's really arguing that my disagreement is irreconcilable with his preferred health care solution. Which makes sense, if my story is correct then charging everyone the same price and mandating everyone buy insurance sounds pretty close to the preferred solution. Since in a truly compete market, we'd all pay the same health care premiums. His preferred solution must be a market frictions story, but its a story I can't understand and he spends a good deal of the podcast and paper dismissing frictions in the health care market. If there is a well-formed second best model there, I'd love to know it.
No comments:
Post a Comment