232 | Amy Finkelstein on Adverse Selection and Hidden Information

Episode Summary

Executive Summary: Sean Carroll and Amy Finkelstein explain why insurance markets are intellectually rich and often fail in practice. The conversation covers adverse selection, moral hazard, government mandates and subsidies, regulatory trade-offs, reinsurance, and the value of empirical evidence in policy design, using vivid examples from health, life, auto, pet, and homeowners insurance.

Main Topics: Why insurance is economically interesting: Insurance is framed as a trade-off: paying now for future security and peace of mind. Finkelstein argues that insurance provides real value because most people are risk-averse and willing to pay for certainty. Adverse selection and market collapse: The core failure mechanism is that high-risk buyers are more likely to purchase insurance than low-risk buyers, causing prices to rise and further drive out healthier customers, potentially creating a death spiral. Private information and hidden risk: Buyers often know more about their own risk than insurers do, even when insurers use actuarial tables, medical exams, and data. This asymmetry appears in divorce, life, nursing home, and auto insurance. Moral hazard and behavior change: Insurance can change behavior after purchase, making people more likely to use healthcare or take risks. The transcript distinguishes this from adverse selection and shows how both effects coexist. Government intervention: mandates, subsidies, and regulation: Mandates, guaranteed issue, community rating, and subsidies can reduce selection, but enforcement and design matter. Well-intended regulations can also collapse markets if pricing constraints are too strict. Evidence from real-world policy and experiments: The episode emphasizes randomized trials and empirical work in economics, including Oregon Medicaid and Massachusetts health insurance reforms, to show that intuition alone is often misleading. Practical consequences for consumers and insurers: Examples from pet insurance, homeowners insurance, Medicare Advantage, and end-of-life care show how pricing, deductibles, and regulatory gamesmanship affect real people and markets.

Key Arguments: Insurance is not a scam; it is valuable because people prefer certainty and peace of mind over small expected gains with large downside risk. Insurance markets are structurally prone to adverse selection because high-risk buyers are likelier to purchase coverage, which can cause prices to rise and the market to unravel. Insurers cannot perfectly infer risk, even with sophisticated analytics, because buyers retain private information about their health, behavior, or intentions. Moral hazard is real: giving people insurance changes their behavior, so researchers must separate the effect of selection from the effect of coverage itself. Government mandates can help by broadening the risk pool, but they only work if enforced and can fail if penalties are weak or rules too rigid. Community rating and guaranteed issue improve fairness but can backfire by driving healthy people out of the market and increasing premiums. Subsidizing healthy or younger people may be economically efficient for solving selection problems, even if it feels politically counterintuitive. Empirical causal evidence is essential in economics because many policy intuitions are wrong; randomized trials reveal actual effects better than theory alone. End-of-life spending is high, but the data do not show that most spending is on people who are clearly known to be imminently dying at the time care is delivered.

Data Points: Pet insurance premium vs payout: $4,300 annual premium; $5,000 maximum payout - Example of a very expensive pet health insurance policy for a 12-year-old bulldog. Health-loss example: 10% chance of losing $100; expected loss $10 - Used to explain why a person might rationally pay more than expected value for insurance. Example willingness to pay: $11 for insurance against a $10 expected loss - Illustrates that paying above actuarial expected cost can still be rational due to risk aversion. Divorce/layoff insurance: No market exists - Used as examples of adverse selection where buyers know more than sellers. Huntington’s risk: 50-50 chance for someone with an affected parent - Used to show how private genetic information affects nursing home insurance demand. Medicaid randomized trial: Randomly assigned coverage in Oregon - Evidence that insurance increased doctor visits and emergency room use. End-of-life spending: About 25% of medical spending on the elderly occurs in the last 12 months of life - True statistic discussed, but interpreted carefully. Homeowners deductible: $10,000 deductible - Finkelstein’s personal choice to reduce premiums on homeowners insurance. Age-related insurance rule: Guaranteed issue at age 65 in Medicare-related plans - Used to explain why older people can enroll in private Medicare Advantage plans, then face restrictions later.

Pivotal Quotes: "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." — Amy Finkelstein: Her opening sales pitch for why insurance is intellectually and practically important. "There is a race to failure built into the very nature of an insurance market." — Sean Carroll: Explaining adverse selection and why offering insurance can attract only high-risk buyers. "What’s amazing about insurance markets is there can be that surplus on the table and yet the market doesn’t function." — Amy Finkelstein: Summarizing the central puzzle of gains from trade that fail to materialize.

Implications: Insurance policy design requires balancing fairness, access, incentives, and market viability. For listeners, the lesson is that better insurance often depends on careful mandates, subsidies, and regulation—plus evidence, not intuition.