The Market for Health Insurance Is the insurance market stable? Buyers hold the information advantage: -Likely to know more than insurers about their health status and habits -Insurers worry that if they offer a particular plan at a particular price, enrollees might be an adverse selection of risk types in the population
Uncontrolled Adverse Selection - If insurer underestimates risk, claims experience will be worse than anticipated - Premiums rise in the next period - Even more good risks drop out - Adverse selection “death spiral” possible: only the high risk people remain insured (at a premium consistent with their high expected costs)
Adverse Selection Example All consumers have the same U(I) and incomes of $25,000 - Half are low risk (25% probability of $20,000 loss) and half are high risk (75% probability) 50% average risk -Insurer cannot identify risk types -No loading or moral hazard -AFP = $10,000 -Insurance guarantees income = $15,000 (utility = U*)
Example (cont.) - For Hs, U* > EUH, so they buy the policy - For Ls, U* < EUL, so they do not buy even though they’re risk averse and we’ve assumed no loading
Adverse selection in choice of plan Suppose: -1 insurance plan (full coverage & community rating) -Low risks prefer this plan to being uninsured -No moral hazard & no loading Compare the equilibrium in a competitive insurance market under full information vs. asymmetric information
Full information (insurers can identify risk types) Community-rated plan can’t be an equilibrium: -Another insurer has an incentive to charge a lower rate but only accept Ls -Ls have an incentive to enroll in this plan to avoid cross-subsidizing Hs -In equilibrium, there will be 2 types of plans: each risk type gets full coverage for a premium = their own AFP, eliminating cross subsidy from Ls to Hs -Full coverage is optimal (efficient) because of risk aversion and lack of moral hazard; EU maximized by having the insurer bear all of the financial risk
Asymmetric information (insurers can’t identify risk types) -Again, CR won’t be an equilibrium -2nd insurer will try to attract Ls. However, if they offer full coverage at Ls’ AFP, Hs will also enroll and the plan will lose money. -2nd insurer tries to design a plan preferred to the CR plan only by Ls -Lower premium plus coinsurance: Ls opt for the new plan to avoid cross-subsidizing Hs, but Hs may prefer high coverage plan even at the higher premium (coinsurance is less desirable to Hs)
Welfare Implications of Asymmetric Information Equilibrium -Compared to full information equilibrium, Hs are no worse off -Ls worse off as they bear substantial risk while full insurance would be optimal -Like FFS plan with coinsurance, managed care may provide alternative plan designs that separate the market by risk type: restrictions on access to specialists and other care management tools may be less bothersome to Ls than to Hs
Evidence of Adverse Selection Cutler and Zeckhauser (Handbook, 2000) review the evidence: -8 of 12 studies of FFS vs. Managed Care find AS into FFS (2 show AS into Managed Care) -3 of 4 studies of the decision to be insured vs. not find AS into insurance -14 of 14 studies of choice between high option and low option plans find AS
AS in Medicare managed care (Mello et al., HSR, 2003) -HMOs receive favorable selection of newcomers -Favorable selection persists over time for some, but not all, health status measures -Risk-adjusted payments to health plans are appropriate
Other ways insurers avoid adverse selection -Medical underwriting -Basing premiums on observable characteristics (age, smoking status) -Pre-existing conditions clauses -Public programs face difficult issue of risk-adjustment -Clinton plan: mandatory CR, but plans enrolling a disproportionate share of persons with high cost characteristics would receive higher payments -HMOs in Medicare and Medicaid managed care -Ability to forecast expenditures from readily available data (age, sex, etc.) is poor
Ex ante vs. ex post risk Private insurance market equilibrium has premiums that reflect each risk type’s expected loss -Not controversial for auto or life insurance -Conversely, many people would like health insurance to provide cross-subsidies from the young and healthy to the old and sickly
Ex ante vs. ex post risk (cont.) -Desire for health insurance to make transfers not only ex post, but also ex ante -Insurance markets tend towards avoiding the ex ante transfers -Even in large employer groups where the ex ante transfers are apparently made, compensating wage differentials may erase the transfer -Thus, accomplishing ex ante transfers likely requires regulation. Because health insurance is voluntary, low risks can avoid paying the cross-subsidy by dropping out or by selecting plans unattractive to high risks.
So, why do we have managed care? -Lowering OOP price causes overuse, creating a welfare loss that offsets part of the welfare gain from risk-spreading -FFS dilemma: better protection from financial risk implies more distorted post-illness incentives. Patient C-sharing, the only FFS tool to control overuse, defeats the purpose of insurance. -Managed care’s fundamental tools are: -Selective contracting/price negotiation -Steering enrollees to selected providers -Utilization review/management/monitoring -Non-traditional provider payment mechanisms (e.g., capitation, salary, holdbacks)
So, why do we have managed care? (cont.) MCOs’ tools have the potential to control overuse without placing the consumer at greater financial risk: -For most goods, the price the consumer pays is exactly what the producer receives -OOP price paid by consumers can be different than the price paid to providers -Overconsumption can be controlled using traditional D-side cost-sharing and incentives to influence provider behavior
Example: Conventional FFS Insurance Suppose objective is to provide incentives to consume m* W/o cost-sharing, mH will be demanded; imposing cost-sharing such that consumer pays p* reduces demand to m* but consumer bears substantial risk
Example: Managed care -Charge nothing out-of-pocket demand mH -If providers are paid pH, they supply all units demanded -Suppose MCO negotiates lower rate with providers (the important part is a lower marginal payment rate; e.g., salaried physician’s marginal payment = 0) -If price = pL, providers supply only m* and moral hazard is controlled without imposing risk on patients -Optimal payment is not “fully prospective”; if providers were paid 0, they’d undertreat
Patients and Providers: Allies or Enemies? Imposing financial (or non-financial) disincentives on providers puts patients and providers in conflict: -If patients pay 0 out-of-pocket, they’d like more care than providers are willing to deliver -We’ll consider supplier induced demand later
Fundamental economic difference between FFS insurance and managed care Conventional FFS Insurance: Patients unconstrained at point of care, but have to pay for the expected costs up-front with a higher premium Managed Care: Patients receive less care than desiree given their cost-sharing, but pay lower premium since moral hazard is reduced; may get better protection from financial risk since S-side incentives help limit care
So, which is better? Managed care may raise consumer welfare by limiting inefficient consumption without imposing additional D-side cost-sharing Risks of “overtreatment” replaced by risks of “undertreatment” “Management” isn’t free: Negotiating contracts, monitoring providers, running UR, implementing practice guidelines, etc. Some financial arrangements (e.g., capitation or holdbacks) place providers at risk; due to limited practice sizes they can’t diversify risk as effectively as insurers HIE showed that D-side cost-sharing discourages care to a greater extent in lower income groups; thus, S-side cost-sharing may be preferable on equity grounds - By changing incentives, S-side cost-sharing can affect development and diffusion of technologies: cost-reducing techs become attractive
Product differentiation in health insurance Nearly all plans now use some “management” techniques: -By 1989, 65% of conventional insurers (those that pay FFS and do not restrict choice of provider) required preadmission certification for at least some hospitalizations, 54% had UR and large minorities had second opinion or case management programs
MCOs shift the locus of “shopping” from the consumer to the health plan - Individuals have little incentive to price shop, little bargaining power and often are unable to shop around even if they would want to (e.g., heart attack) - MCOs have incentive to negotiate lower prices, and ability to steer large #s of patients to providers; some control costs by providing services directly - Success in negotiating discounts depends on competition among providers (i.e., competition in physician and hospital markets is needed to make MCOs effective competitors in the insurance market) - Before MC, providers had little reason to compete on price. With Medicare’s move to DRGs and MC’s growth, “excess capacity” developed in the hospital market and in some physician specialties, helping MCOs win discounts
Competition and managed care discounts -Melnick et al. (JHE, 1992): -Greater PPO discounts in more competitive hospital markets (more hospitals and more empty beds) -Smaller discounts when the PPO was more dependent on 1 hospital -Kaiser owns its hospitals -Allowed Kaiser to bypass the high prices of other hospitals, providing competitive advantage in the ‘70s and ‘80s when there was little price competition -Recently, Kaiser enrollment has been flat while IPAs and PPOs, which contract for services, have grown. With low hospital occupancy, these plans can negotiate substantial discounts.
Health plan choice: responsiveness to premiums - Many employees face a “menu” of plan options - Increasingly, employees are charged different co-premiums for different plans (employer contribution is a fixed dollar amount, so employees pay the marginal cost of more expensive plans) - Willingness to switch is crucial to the success of competitive strategies for controlling costs - If enrollees became “locked into” their plans, price competition between insurers would be reduced
Evidence on plan choice Short and Taylor (JHE, 1989): -$100 decrease in the HMO premium relative to traditional coverage led to 2.6 percentage point rise in HMO enrollment -$100 decrease in the price of FFS with high cost-sharing relative to a more generous traditional plan led to 5.3 percentage point rise in enrollment in high cost-sharing plan Buchmueller and Feldstein (HA, 1996): -“Natural experiment” in which the U Cal changed the way it priced health plans to employees -Employees facing small premium increases ($1-10 per month) 5 times as likely to switch than those facing constant premiums (25.2% vs. 4.8%); 42% of facing $50- 60 monthly increases switched
Evidence on plan choice (cont.) Buchmueller (HA, 1998) -Shift of enrollment to cheaper plans substantially reduced UC outlay -AS death spiral for FFS plan Wholey, Feldman and Christianson (JHE, 1995): -Increased competition in HMO market lowered group/staff/network model HMO premiums and IPA premiums -Provides some hope that competition between insurance plans will help reduce premiums -Helps explain why the large profits enjoyed by many HMOs in the mid-1990s were transient
Uninsurance and the Labor Market Measures of the uninsured vary, but 45M is a reasonable figure (15% of the US population) 63% have at least one full time worker in their household 27% have household incomes > 3 x FPL
Uninsurance and the Labor Market (cont.) One quarter of working uninsured were offered coverage but declined -Reflects low WTP for insurance; even substantial subsidies may not have much effect on take-up If workers with low demand for HI sort into firms that do not offer insurance, non-offering firms are not likely to begin offering insurance in response to a subsidy -Several studies are consistent with a weak response to employer subsidies State coverage mandates make insurance more expensive, encouraging firms and workers to drop coverage -Particularly true for small firms, because large firms often escape state mandates by self-insuring