20.2 Adverse Selection When Buyers Know More Than Sellers
You’ve now seen what happens when sellers know something that buyers don’t. Now it’s time to turn the tables and explore what happens when buyers know something sellers don’t. It turns out that this information gap creates similar problems, but in this case, it’s buyers who are adversely selected. Instead of too few high-quality goods, the problem in this case is the quality of the customers.
How does customer “quality” vary? Some customers cost the seller less than others. For example, customers who borrow money differ in their likelihood of paying it back. Loaning money to someone who pays it back on time is less costly than loaning money to someone who must be hounded to make a payment. Folks looking to rent on Airbnb differ in how neat and tidy they will be as well as in their likelihood of doing damage to the property. Thus, some renters are lower cost than others. And in the health insurance market, healthy people are lower cost compared to unhealthy folks who are going to require a lot of medical care. The problem for suppliers is that the customers have private information—they know whether they’re likely to be high cost or low cost—while the seller can’t tell. Let’s investigate how this shapes the market.
Hidden Quality and the Risk of Getting High-Cost Customers
Hannah Johnson is the CEO of a small nonprofit health insurance cooperative. Because her business is a nonprofit, her goal is simply to take in enough in payments to cover the medical costs of the people she insures. She hired a team of researchers to find out average annual medical expenditures in her area so that she can sell insurance at a price a smidge above average spending on medical care. For instance, if average annual medical expenses in her town are $5,000 per person, she hopes to sell health insurance for a smidge above $5,000, so that she can also cover her overhead.
People expecting a lot of health care costs will pay more for insurance.
While Hannah only knows the average medical expenditures in her community, it turns out that her potential customers know some things that she doesn’t. Some of her customers know they’re genetically predisposed to cancer, some suspect that they are developing diabetes, and others are planning on getting pregnant. These potential customers have private information that they’re each likely to have higher-than-average medical expenses. If Hannah sells her insurance policies at the cost of average medical care expenses, her policies will be a great deal for these folks! Anyone expecting higher-than-average costs will be excited to sign up for health care coverage from Hannah’s nonprofit.
Giving birth can cost over $10,000—and that’s if everything goes as planned.
People expecting few health care costs will pay less for insurance.
There are other people who know that their health is good. They’re young, they’re fit, and they drive safely. Sure, anyone can get cancer, and accidents happen, but these folks don’t expect to be costly clients. Healthy people value insurance, but if their expected health care costs are only $2,000 per year, they question whether insurance priced at the community average cost of $5,000 per year is really worth it. They’d be willing to pay something for health insurance, but they’re willing to buy insurance only if it’s relatively cheap—something closer to the average medical expenses for healthy people like them. That means that for some of these healthier folks, when insurance is priced at the community average medical expense, it’s too expensive to be worthwhile for them. As a result, they won’t buy health insurance.
Adverse selection of buyers can cause the market to collapse.
If Hannah can’t figure out who is likely to have high or low medical expenses, she’ll charge everyone the same price for health insurance: a smidge above $5,000, the community average cost. But then some healthy people won’t want to buy health insurance, while those who are most likely to need expensive medical procedures will see it as a good deal. This leads to a larger share of customers having high health insurance costs compared to the community average. This is known as adverse selection of buyers—the tendency for the mix of buyers to be skewed toward more high-cost buyers when buyers have private information about their likely costs. This problem arises because of private information—people buying insurance know things about their health that sellers do not. And in this case, Hannah’s customers are adversely selected—the folks who are going to cost her the most in medical bills are the ones who are most likely to buy health insurance from her.
If Hannah could differentiate among people’s likely health care costs, she would just offer a low-price insurance contract to people who are healthy—because they’re likely to have lower medical-care costs on average—and a high-price contract to those who will likely have high medical expenses. But because she doesn’t know, she has to charge everyone the same price. Adverse selection of buyers is a problem for sellers, because it means you won’t get the (low-cost) customers you want.
Figure 2 shows how the problem occurs. The problem is that buyers know something sellers don’t, and they use that information to their advantage. Because some low-cost folks choose not to buy health insurance, the average amount Hannah ends up paying out in health care costs is above the community average. So she needs to charge higher prices to break even. But higher prices lead more people to opt out—particularly those who expect to have moderately low health care costs. This once again means she pays out more than she has charged people on average. And this cycle can be self-reinforcing.
Figure 2 | Adverse Selection of Buyers
In fact, insurance companies can face an adverse selection death spiral, where they charge more in an effort to break even, but that causes even more people who are likely to be relatively low-cost to stop buying health insurance, and that in turn raises their average payouts. Sometimes the spiral doesn’t end until only the very highest-cost buyers are left.
Adverse selection makes it hard for some people to buy the products they want.
As a buyer, if you have private information that you’re likely to be a low-cost customer—you’re healthy and unlikely to need expensive medical care—you may not be able to find insurance that is fairly priced for you. The problem is that insurance companies can’t tell the difference between you and unhealthy higher-cost customers, and therefore must charge you a higher price because they’re worried that you’re a high-cost customer. This higher price may lead you to decide not to buy insurance.
The result is a market failure. The fact that some people don’t buy insurance doesn’t mean that they wouldn’t benefit from insurance. Even relatively healthy people want to be insured. And insurance companies would love to sell more insurance policies, especially to low-cost customers. The problem is that the insurance company can’t offer low-cost customers a policy that would be attractive to them without also attracting so many high-cost customers as to render those policies unprofitable. Adverse selection in insurance markets pushes prices up, which causes many low-cost buyers to be unable to find insurance that makes sense for them.
Interpreting the DATA
Adverse selection in Harvard’s health insurance plans
Think you understand adverse selection? If so, perhaps you should explain it to the smarties at Harvard. The university used to offer two types of health insurance to its workers. However, it offered a bigger subsidy to those choosing the more generous plan, and so a lot of people chose that policy. When Harvard decided to equalize the subsidies across the two plans—effectively raising the price of the more generous plan—it quickly learned a lesson about adverse selection. Soon Harvard’s younger and healthier employees—who didn’t really need the more generous plan—left it. Only a relatively unhealthy pool of workers remained in the generous plan, which caused that plan’s costs to soar. Consequently, Harvard was forced to raise the price of this generous plan further. But this caused even more healthy people to leave the generous plan. And because the remaining workers were even less healthy, costs per worker rose yet again, leading the generous plan to lose even more money. Within three years, this cycle—of rising prices leading healthy people to opt out, which raised average costs—became obvious even to the folks at Harvard, and it abandoned this health insurance plan.
Harvard isn’t immune to adverse selection either.
Shared ignorance is bliss in insurance markets.
The problems of adverse selection depend on how large the information gap is between buyers and sellers. For instance, you don’t have much private information about the likelihood of a robbery or an electrical fire at your house. The likelihood that you will file a claim is largely determined by factors such as how dangerous your neighborhood is, how old your house is, and what materials were used in the construction. Since insurers are able to learn this information, they can tailor the price of your homeowners insurance to your specific circumstances. You don’t have much private information, so your best guess as to the how likely you are to have an electrical fire is similar to the insurance company’s best guess. As a result, adverse selection isn’t as much of a problem in the market for homeowners’ insurance. But note that insurance works best when neither side knows how likely it is that the insured event will happen.
Letting people opt back into insurance when they need it worsens adverse selection.
This insight that insurance markets work best when there is very little private information explains why there are often limits on when you can opt into or out of insurance. For instance, if you could purchase fire insurance as soon as you smelled smoke, then only people whose houses were already on fire would buy insurance. That’s the sort of risk no insurer will take on, leading the market to collapse. To prevent people from exploiting their informational advantage, there are often rules that limit the ability of buyers to opt into insurance once they have more information.
This is a big issue in the health insurance market. If customers could opt out of health insurance when they’re healthy, and then opt back in as soon as they got sick, insurance companies would end up with only customers who are sick and require expensive treatment. This worsens adverse selection problems. And it explains why health insurance companies often refuse to cover medical expenses resulting from preexisting conditions (unless required to do so by government regulations). They’ll let you opt back in to insurance, but won’t cover conditions you developed before buying insurance. Effectively this eliminates the incentive to opt in only after you get sick, and so it limits adverse selection. But it also means that folks who develop medical conditions while they’re uninsured—perhaps they’re between jobs or temporarily can’t afford insurance—face a lot of risk. That’s why the question about how to deal with preexisting conditions has been so controversial. This logic also explains why insurance companies that are forced to cover preexisting conditions by government regulations typically want these policies to be matched by a mandate requiring that everyone buy insurance. If everyone must buy insurance, then insurers no longer need to worry about buyers opting in to insurance only when they’re sick.
Risk aversion may help undo some of the problems of adverse selection.
Most people don’t like to take gambles, particularly with their health. They’re risk averse, which means that they dislike uncertainty and will often pay to avoid it. Risk-averse people will pay more than the actuarially fair price for insurance, meaning that they will pay more than they expect to get back on average. They help insurance markets function because, by being willing to pay more, insurance companies can sell policies to some risk averse, healthy customers even though they have private information that they’ll likely be lower cost on average. This helps limit adverse selection.
Adverse selection problems can arise in any market where buyers have private information.
The adverse selection problem extends well beyond insurance markets. Indeed, the central insight here is that anytime buyers have private information about their likely costs, your business will find it hard to get the customers you want. Low-cost customers are more likely to decide what you’re selling is too expensive, while high-cost customers are more likely to see it as a great deal. Here are a few real-world examples:
- Life insurance companies—which offer your loved ones a payout if you were to die—may find that people who don’t take good care of their health are more likely to purchase life insurance.
- Car insurance companies may find that people who drive the most recklessly are the most interested in buying a lot of collision protection.
- An entrepreneur trying to offer divorce insurance may discover that only people whose marriages are on the rocks are interested in purchasing it.
- Landlords may find that people who are financially irresponsible and not interested in taking good care of their living space are the most interested in renting rather than buying a place to live.
- Restaurant managers may find that people with really large appetites are the most interested in the all-you-can-eat buffet.
These are all situations in which you might not get the customers that you want. As a buyer, these are also instances where you might not be able to buy what you want. You may prefer the variety of a buffet, but if the enormous appetite of your college’s football team has pushed the price up, paying that much might not make sense given your small appetite.
So what can you do about it? Let’s now turn to considering solutions to adverse selection among buyers.
Solutions to Adverse Selection of Buyers
Thankfully, there are ways to address the adverse selection problems that arise when buyers have private information. These solutions can help you get more information about your customers—helping you tell high- and low-cost customers apart.
Solution one: Sellers can use information that is related to buyers’ likely costs.
Insurance companies are always on the lookout for buyers exploiting their private information. That’s why they ask you so many questions before telling you how much your policy will cost or whether they will even offer you one at all. Auto insurers use your age, your gender, your marital status, your credit rating, your grades, your car’s safety rating, where you live, and of course, your driving record. Similarly, health insurance companies will ask whether you smoke or exercise. But the law limits the characteristics they can use. For instance, it is illegal to charge men and women different prices for health insurance even though they can have very different medical expenses.
This explains why car rental agencies won’t rent to you if you’re under a certain age—they’re worried you’ll be a high-cost customer, perhaps more likely to crash the car. Vacation rentals may specify “families only,” effectively rejecting unrelated groups of people—that is, groups of friends—because they figure that friends are more likely throw a party that trashes the place. And health insurance companies charge a lot more to smokers, figuring that they’re the folks most likely to face high medical costs.
Solution two: Sellers can offer different contracts so that buyers separate themselves.
One way to get the customers you want is to offer a product that only low-cost customers will want. For instance, insurers are sometimes able to differentiate between high- and low-risk customers by varying the size of the deductible. A deductible is the amount you pay out of pocket before insurance kicks in. For example, your health insurance might require you pay for the first $2,000 of childbirth-related expenses—that’s the deductible—and it will pay any expenses beyond that.
An insurance policy with a large deductible will only be attractive to people who think it’s unlikely they’ll need to file a claim. As a result, these policies attract low-cost customers. And that in turn means that insurance companies can offer these high-deductible plans at particularly low prices. By contrast, the high-cost people who believe they’re likely to file a lot of claims will want a low deductible. Thus, when people choose the low-deductible plan, they’re effectively revealing their private information—their choice to opt for a low deductible reveals to the insurance company that they expect to be high-cost customers. In turn, this lets the insurance company know that they’ll need to charge a lot more for those plans. This explains why you typically pay a lot more for insurance plans with low deductibles.
This leads to some advice: If you think you’re a low-cost customer who is unlikely to need to file a claim, you’ll often get a much better deal if you choose the insurance plan with a high deductible. But realize that this doesn’t completely solve the problem of adverse selection. In order to get the good deal, you have to risk some of your own money—if something goes wrong, you’ll have to pay a high deductible—which means that you’re only partly insured.
Some managers use a related approach, bundling their products with something that only low-cost customers will want. For instance, if you think that parents eat less at a buffet than college students, then it might be worth providing Lego sets for kids to play with. Those Lego sets will attract more families to your restaurant, which is profitable because they’re your low-cost customers.
EVERYDAY Economics
Will you soon pay more if you are a bad driver?
For decades, insurers have charged young people much higher rates for auto insurance. Their logic is that some young people take more risks behind the wheel and get in more accidents. But not all young people are bad drivers, and if you’re a good driver you might feel that you deserve a better deal. The problem is private information. Insurers don’t know if you’re a good or bad driver, so they have to charge you a price that accounts for the risk you’re a bad driver. But what if your auto insurer could use technology to close this information gap?
Auto insurers such as Progressive, Allstate, and State Farm now have programs where you can allow them to track your driving—in many cases, using a simple phone app. Why would you want an insurance company tracking your movements? It’s simple: If the data they collect show that you’re a safe driver, they’ll offer you cheaper insurance. This also helps insurance companies figure out who the bad drivers are—they’re the folks who won’t sign up to use the app.
It’s worth thinking about where this will lead. As fewer good drivers buy unmonitored auto insurance, insurers will need to raise prices on unmonitored insurance to cover the higher average costs of covering mainly bad drivers. Those higher prices will push more good drivers into using the monitoring app. Eventually only bad drivers will be left buying unmonitored auto insurance. At that point, insurance companies have app-related data to identify the good drivers, and they can also identify the bad drivers, because they’re the ones refusing to use the app. The insurance company can use this information to offer prices tailored to each person’s driving skill. As a result, technology that reduces the adverse selection of buyers will lead to lower prices for good drivers and higher prices for bad drivers.
Solution three: Government can increase information or directly reduce adverse selection.
The government can help solve adverse selection problems in insurance markets in four ways: By providing incentives for buyers to reveal their private information, by subsidizing insurance, by mandating that everyone buys insurance, and by providing insurance directly.
Government creates incentives for buyers to reveal private information.
The government can give people an incentive to tell the truth, thereby revealing private information. If you lie to an insurance company, that’s called insurance fraud, and it’s a crime. The government helps the insurance market work by ensuring that the insurance company can count on buyers to tell the truth. You won’t be asked if you expect to die soon, since it would be hard for anyone to ever prove if you’re lying. But insurers often ask if you’re a smoker, if you use drugs, or if you have dangerous hobbies like skydiving. The threat of punishment for lying helps ensure that you tell the truth. And your truthful answers to these questions reduces the extent of private information, thereby limiting adverse selection.
Government can subsidize insurance.
The government can also subsidize insurance so that it’s a good deal for more people. The idea is that if you’re only paying part of the cost of health insurance, you’re more likely to buy it. That inducement might lead more healthy or low-cost customers to buy health insurance, reducing adverse selection. There are many forms these subsidies might take, and in the United States, the government subsidizes health insurance both directly (if you buy it on your own) and through tax breaks (if you buy it through your employer).
Government can require everyone to buy insurance.
The government can eliminate adverse selection by requiring everyone to buy insurance. When the government forces everyone to buy insurance, sellers no longer have to worry that only the high-cost customers will buy it, and so there’s no pressure to set higher prices. The information gap between buyers and sellers still exists, but buyers can’t exploit it to their advantage because they don’t get to choose whether to buy insurance. This is why most states require that all drivers have auto insurance, and it helps keep the price down. It’s also why health insurance used to be required. When the government mandates that everyone buy health insurance, whether sick or healthy, there’s no adverse selection.
Government can provide insurance.
Finally, the government can ensure that everyone has insurance by providing the insurance itself, making sure to provide it to everyone who qualifies. When the government offers insurance, your taxes usually cover the cost. As long as everyone in a specific group is covered by the government-provided insurance, there’s no adverse selection problem because no one can opt out. The government directly provides health insurance to around two-fifths of all Americans—including anyone over age 65 (who qualify for a government-provided health insurance program known as Medicare), and many lower-income people (who receive a health insurance plan known as Medicaid). Beyond health insurance, the government also provides unemployment insurance to protect you against job loss, disability insurance that insures you against a disabling injury or disease, and Social Security that insures you against outliving your savings.
So far, we’ve analyzed two types of private information, exploring what happens when sellers know something that buyers don’t (that was the lemons problem), and in this section, what happens when buyers know something sellers don’t. Our final task is to explore what happens when your actions can’t be observed, meaning they are private information.