Saturday, October 11, 2008

Risk Segmentation In The Health Care Market

The Jaded Consumer hopes to outline a few salient health care finance issues in a series of issue-focused posts. There may be many solutions to some issues, and the Jaded Consumer does not pretend to offer ideal solutions. However, a policy piece finding fault with the status quo should offer proposed interventions, or it is a mere gripe session. Other nations, with different national predilections, would doubtless steer a different course likely to be recommended here. However, The Jaded Consumer intends offering practical solutions that might be accepted in the United States.

This first article addresses risk segmentation as a barrier to affordable universal coverage. Universal coverage is not addressed as a right, though there are states in which legislatures sought to make it a right before encountering federal barriers to achieving the goals of their constituents. Universal coverage for health risks is addressed here as a national aspiration.

The barrier raised by risk segmentation is escalating marginal cost. The idea is something like this:
0) The risk of health care expenses is distributed unequally across the population
(This point is a huge deal and is numbered zero as a sort of emphasis.)
1) Most easy-to-cover people get covered easily through private funding. (This is in part a result of the tax-advantaged status of employee benefit plan contributions under statutes designed to enshrine compensation practices dating from the time of World War II wage freezes, when non-cash benefits seemed to employers a plausible way to compete for scarce labor. Today, anyone sufficiently healthy and socially stable enough to hold down a permanent full-time job is advantaged in accessing externally subsidized health care. The subsidy isn't, as often thought, the subsidy of employers -- they will never pay more than they can afford for labor -- but the subsidy of the federal government, which has chosen to push people into employee benefit plans by electing to tax this form of compensation at zero percent.)
2) People who argue most forcefully that they need coverage, and those whom advocacy groups most urge should be covered, are drawn overwhelmingly from a subset of the population far from the left tail of the health care expense curve; these people are, on the whole, less likely to be cheap risks to cover. (Think about impoverished, permanently disabled Medicaid enrollees and the elderly, who are already in high-cost federally-supported safety net systems precisely because the risks posed by these persons has led to government intervention.)
3) Uncovered and safety-net populations experience increasing concentrations of high-risk individuals. This one takes a little explanation, and is critical to the problem of incremental establishment of universal coverage. The cherry-picking of "good" risks by for-profit coverage schemes (like mandatory coverage participation at some level for young full-time employees who've never been ill a day in their working lives) cause the population of already-covered people to have a lower cost of coverage than the groups described by (2) above, making the "high expense" of covering those easy-to-cover people seem to be chump change by comparison to the cost to cover the concentrated risks in the remaining population. As the cost of coverage increases, the benefit of the tax breaks in (1) will push more people into available private schemes because of the amplified after-tax cost advantage, despite legal uncertainty regarding enforceability of rights under employee benefit plans (compared to non-employment-related insurance coverage).

How Insurance Works

You don't hope to crash your car.

However, for only a few hundred bucks, you can get months of protection (by a licensed insurer, regulated by the state for its solvency in the face of its insurance obligations) against the risk that you cause a wreck that crumples someone else's fender -- protection that includes sending someone to inspect the claimed damage, haggling on the phone or by mail with the other driver about the cost of his repairs, and even defending you against his law suit -- hiring lawyers, conducting depositions, running a trial -- if he doesn't like what your insurance company is telling him about the value of his rusting jalopy.

All things considered, this coverage is a deal.

As it happens, you don't have a choice about this coverage; every state in the Union has a vehicular financial responsibility law, and when last I checked forty-eight of them specifically required insurance as the way this financial responsibility must be demonstrated. In order to evade this coverage, there are people who work all kinds of scams -- paying one month on a six-month policy and then ceasing payment, printing false insurance cards, you name it -- but if you want to renew your driver's license and keep your car registered so you're not pulled over and ticketed every week, you either have the coverage or you work very hard to evade the requirement. The reason this coverage is so cheap is part of the secret of how insurance works, and why it works so well when the covered population is (almost) the whole population at risk. Only scam artists escape the pool; all the responsible people like yourself are in the covered pool, and are sharing the cost of the risk. The legal minimum insurance is cheap not because law suits aren't costly and not because people don't get into wrecks in this country and because there aren't tens of thousands killed every year on U.S. roads and many more injured -- the insurance is cheap instead because the burden of all this cost is allocated across an enormous pool of covered risks. (Okay; it's also cheap because the legal minimum caps out at a cost well below the cost of a human life. Most fender-benders aren't million-dollar claims. The outlier risks are left where they started, in the hands of individual motorists, because the many state legislatures creating the mandatory coverage laws have reached a compromise between the premium-paying public and the premium-hungry insurers, and have drawn a line where the mandate ceases, and private choice to obtain more coverage steps in. Tweaking local coverage requirements isn't rare.) In short, since the whole population contributes to mandatory coverage, the per-member burden is comparatively modest.

The coverage you elect, the one that protects you against the risk of someone totaling your car and fleeing into the night, is more expensive. It's above the point the mandatory coverage, above the point risks are spread population-wide, that risks are concentrated among a subset of willing buyers -- and this is the point at which it starts to become expensive. If you don't owe a bank money secured by a car (banks require more-than-minimum insurance to protect their loans' collateral), and you aren't frightened by these risks (fearful people often overinsure, to the delight of insurers), then you need not pay for them. You can opt out of the extra coverage.

Adverse Selection and The Problem Of Personal Choice
Insurers like to talk about moral hazard -- the tendency of the insured to conduct themselves so as to experience losses at a greater rate or severity than the non-insured -- and they like to allege that the world is full of insurance fraud. The world is as equally full of insurers trying to shirk their contract obligations in the face of loss. It's probably a wash, each side diligently working to cheat the other.

The real reason insurers hate the individual coverage market is adverse selection.

Adverse selection is the reason that when insurers price a lifetime annuity (a bet the covered life will last forever) and life insurance (a bet the covered life will end tomorrow), they don't use the same mortality tables. When you shop for an annuity, insurers use a mortality table that assumes you are healthy and going to live an awfully long time. When you shop for life insurance, they use a table that assumes you are hiding anything that can't be picked up on their screening tests, and will drop dead without a great deal of provocation. In short, the insurers assume when you shop for insurance that you know you need the coverage.

While some folks can be convinced by an insurance agent to buy anything, there are others who are particularly keen to get life insurance because they know all their near relatives died young, or they just learned they have a lethal illness and have been approached by a viatical settlement buyer, or the like. Prospective insureds who fear losses can be in a much better position than an insurer to appreciate the risks facing them. In this world -- where paper medical records can lay unnoticed by searching insurers, where billing codes are manipulated to ensure payment rather than accuracy -- insurers fear the information gap in underwriting important risks like health (which can quickly run into six or seven figures on a bad risk) is stacked against them. In a market in which members can opt out of sharing the population's risk, insurers fear the worst and charge accordingly.

Mandatory Coverage Is Coverage For Insurers
This is why full-time employees "get" health coverage in most companies whether they like it or not. Without the choice to opt out, employees are powerless to exercise adverse selection, and insurers can rest easier, unworried that new applicants are seeking coverage to run a scam; they are seeking coverage because they want a job. People with genuine work qualifications are probably less likely to be running a scam than the general population, besides. Moreover, insuring a whole pool of covered risks is a more attractive enterprise; there's little wonder insurers who deal with employee benefit plans structure their offerings so as to yield this result.

With very few employers failing to insure health risks at some level, and none to my knowledge processing claims directly, it's clear that the preference of insurers in structuring their offerings directly drives benefit plan coverage decisions. Insurers don't want to have to insure loners who turn up begging for coverage (what does he know that we don't?), but they are happy to write policies on whole populations who can't say "no" when their employers are looking for a way to pay valuable compensation that isn't subject to employer taxes.

Covering The Uninsured
So let's look at a common solution: we'll take the people who we're worried about -- the uncovered folks we want to stop choking emergency departments to death by allowing neglect of primary care to mature into a succession of expensive minor emergencies, the chronically ill who don't seem to qualify for stingy government programs, and so on -- and we'll cover them all in a special program to stem the tide of uninsurance. Because this population is selected for its high costs, coverage of it will be through the roof.

Think about it: if these people's claims weren't breaking local hospitals' ER budgets and didn't swamp county hospitals' funding and weren't killing everyone with an inclination to provide charity care, nobody would be clamoring for their coverage. If the claims weren't significant, the whole problem would be written off as not-a-big-deal. The county would cover it, or teaching hospitals, or charities. But that's the problem: the cost of this care dwarfs the resources people want to use to address it.

Covering these folks will cost a fortune.

Enter the High Risk Pool
The idea of creating government-established high-risk pools and offering group coverage to everyone in the pool surely sounds like a great idea at first: it's group coverage and not individual coverage; everyone in the group must paricipate; there should be a group discount -- right?

Ha, ha.

Americans like to think you get a deal buying products in bulk, but this is insurance, not a sack of yams. You get a discount by spreading risk, sure. But a high-risk pool isn't a place risk is spread: it's a place risk is concentrated. Insurers don't want any part of it at typical prices. Unless the premium is so high that there's a safe bet on the large concentration of risks, all bets are off. Insurers required to play this game would sooner leave the market. (Mind you, you don't need zillions of insurers to have a competitive market; two or three are enough, if there is no price collusion.)

The usual fix for this is to have government subsidize the high-risk pool. This, in effect, is the very thing the pool was to avoid: soaking the public fisc with the cost of the uncovered care. Taxpayers don't want to do this, since they are themselves (ask them) already paying for their own care. The only folks they see covered by a scheme like this are folks who don't have a job and don't pay taxes. Not exactly an easy sell, is it?

The Newest High-Risk Pool Proposal
President Obama -- I say President Obama because it's sure he will win, so why bother dreaming up silly terms like "hopeful" and "prospective" to term what's already a fait accompli? -- will urge a payroll tax to cover the cost of covering everyone not already covered by an employer plan. He spells it out right here. (In it he also spells out some funny things like how he'll save money for the health care system with tort reform, which is ironic considering the biggest estimates ever calculated for the annual financial cost attributable to malpractice claims and so-called "defensive medicine" don't amount to a tenth of a percent of the national health care budget. This, from a man who has claimed earmarks aren't a material expense!)

President Obama's plan expressly exempts from the payroll tax any employer already providing "adequate" coverage. This means that every employee will have a choice: whether it is cheaper to pay the payroll tax, or to maintain the existing plan. In other words, covered populations that are cheaper than the payroll tax will be able to avoid shouldering any of the risk-shifting associated with the incremental coverage. Once the price of the new plan is known -- that is, once the employer tax is spelled out in a statute -- employers will have even greater incentive to game their benefit plans for tax purposes. Employers with young, healthy populations will pay insurers to keep them out of the employer tax, and employers with older or less healthy populations will cheer, throw their health plan out the window, tell employees they have joined the new federal program created by Congress, and save all those extra health care dollars. After all, the payroll tax will be a tax-free benefit and fully deductible, it'll just be cheaper.

This market behavior will ensure that only populations that cost more than the payroll tax are in the federal risk pool. By spelling out the alternative in a statute, the payroll tax will enable insurers to conduct their ordinary cherry-picking operations even more effectively. Employers will be beseiged by cut-rate plans offering to "cover" for less than the cost of the payroll tax mostly-healthy populations with plans designed to scrape through the payroll-exemption language in the post-Obama tax code.

Spreading Risk For Real
To spread the risk of the whole population -- including the population of covered lives cherry-picked by insurers from an applicable risk pool -- it is necessary to tax not payroll but premiums. A premium tax of the sort states ordinarily collect on insurance premiums can be used to conduct risk allocation by funding the overpayments caused by the risk pool. A premium tax ensures that insurers price all their products -- to any market segment -- so as to cover the demands of the high-risk pool. The premium tax structure can thus be designed to make cherry-picking futile, ensuring more coverage and less coverage-avoidance.

Federal Health Services Funding Monopsony? Not The Cure!
In the law of the sale of goods, every state in the Union has the same substantive law. (Louisiana does not have common law, so the Uniform Commercial Code Article 2 language that depends on the existence of the common law cannot be used there; however, a local analog containing the same substance has been enacted to achieve the same result with different words.) Congress didn't need to create federal law on the sale of goods; the utility of uniform law was itself enough to inspire coordination and uniformity. State-law insurance insolvency statutes are the same way: federal law does not govern them (due to the McCarran-Ferguson Act), but there is nevertheless uniform law nationwide. Automotice drivers' responsibility has worked the same way. The argument that federal law is needed to create uniformity is a canard.

Federal law is in fact the major obstacle to universal care in this country, and has been so since the late 1970s when employers used federal law to enjoin a payroll tax funding Hawaii's health plan. Oregon had a radically different plan -- one with distinct and innovative cost-containment measures from which the nation could have learned valuable lessons for health plan design -- but the lessons brave Oregon would have taught us have been lost to federal preemption. Tentative movements of other states toward universal coverage have been slowed by the lack of freedom to experiment in this regulatory arena.

Experimenting on a state-by-state basis to provide data points on coverage features that are effective -- and which are ineffective -- is not a strength a single federal plan can possibly offer America. Once in place, the plan will be pulled about not by data or the lessons of neighbors' differing lessons, but by pure politics. Health care providers who have been terrorized by their powerlessness before federal payors and their ruthless watchdog agents' hungry zeal to threaten fee recoupment will understand immediately why a federal monopsony in health care finance is a danger to the field. Hospitals and physicians familiar with federal funding games will understand why giving the federal government an even larger fraction of the covered population will work against health care workers and their desire to provide quality care.

The worst obstacle to the development and innovation in health care policy in the last thirty years has been federal law preventing state experimentation in funding policy, plan enforcement, and quality assurance. The last thing America needs for the improvement of health care policy is more involvement by Congress. A new federal uninsurance pool funded by an opt-out payroll tax will accelerate risk segmentation, escalate the cost-shifting of the cost of health coverage onto government, and will prevent local efforts to innovate -- as Oregon heroically tried to do before it was stabbed in the throat for its efforts -- in ways that will result in genuine efficiency gains.


Anonymous said...

I don't understand how this premium tax would work and how it "ensures that insurers price all their products -- to any market segment -- so as to cover the demands of the high-risk pool". Please elaborate.

Jaded Consumer said...

Premium taxes are levied by states on insurers for a variety of reasons -- to fund enforcement, to raise funds for funds designed to cover claims against insolvent insurers, you name it. It's just a tax.

A premium tax isn't a tax on insurers' incomes. A premium tax is a levy on whatever insurance premiums are collected by an insurer and are subject to the tax. For example, surplus lines insurance is typically subject to a surplus lines tax levied against any premiums collected for surplus lines insurance. If a policy isn't a surplus lines policy, its premiums aren't subject to the tax.

The idea here is that by cherry-picking good risks from the risk pool the government needs to find a way to get covered, insurers have created a cost-shift to the public. The premium tax attempts to shift back to the insurers -- and thus to their premiums payors -- the cost of the risks that sound policy would have more broadly spread. Insurers can get back some of this tax by covering people from the government high-risk pool (people who can't afford the cost of insuring members of the risk-concentrated pool), and getting a government subsidy (funded at least in part by the premium tax) to cover the difference between the affordable premium rate and the cost to provide the coverage.

In theory, HMOs with expertise in high-risk groups like diabetics could make a living serving exclusively high-risk people, and profit doing it, because the premium tax funding their risk-pool premium-match (or however it's structured) has shifted the cost of these high-risk people back to the insurers who were working so hard to make sure they didn't cover too many people who were unemployably sick, or had highly comorbid medical conditions.

There are other ways to cover everyone -- Oregon's plan would have done it in part by saving money in claims handling, which is a huge black hole of insurance overhead -- but these innovations will probably remain regarded as "unproven" because federal law prevented a population-wide implementation, meaning that we will never know how successful Oregon's plan might have been.

Back to your question: Insurers who know they face a premium tax will (logically) price their products so as to leave them with enough funds to operate after paying the tax. Assuming that the government agency setting the rate of the premium tax is roughly correct about the amount of cost attributable to cherry-picking (and this would likely be wrong to some degree in either direction from time to time, as coverage trends and insurer marketing changes), insurers would have to charge their low-risk clients an amount suited to subsidizing the demands of the previously-avoided high-risk groups whose cost of care insurers have historically been able to shift to hospitals (uncompensated care), local governments (government-run safety-net hospitals), and states and the federal government (Medicaid costs, etc.). With cherry-picking rendered ineffective due to cost-shifting premium taxes, the "true" cost of broad coverage would become known, and measures to address those costs could be considered and implemented.

At present, uncompensated medical costs are simply a hot potato tossed from party to party in the hopes someone else will get stuck with it. Very little has been done to effect control of the problems created by non-coverage, which is blamed for driving up aggregate costs by causing avoidance of presumably cheap preventative care and leading to use of free-to-the-emergently-ill services like emergency departments, which are both not designed for primary care and are by their very mission unable to accommodate problems until they are severe and expensive.

Kent said...

So... those of us in low risk pool will see rate increases so that insurance companies can pay tax to government. Government will give portion of tax money back to insurers to subsidize insurance for high risk clients. Rest of tax would go to support government programs for unemployed, uninsurable and aged.

I don't see the difference between indirectly taxing us via the insurers vs a direct tax. In either case, some government agency sets a tax rate that hopefully offsets the cost of subsidizing the high risk pool.

Jaded Consumer said...

Kent: There are two schools of thought. The first -- the every-man-for-himself school of thought -- is certainly one idea, and people who are part of the cherry-picked low-risk pool will love that while it lasts. Moreover, adherents of this idea are not in the least offended by cherry-picking and other risk segmentation strategies, which they would regard as logical results of market behaviors to avoid losses.

The other idea is that the cost of health concerns are one of those social costs that should be spread across the population. Whether the idea is rooted in theories of social justice or in economic ideas about finance policy, this second idea requires that cherry-picking to avoid risk be addressed with cost allocation that unwinds its effect on the remaining risks, which otherwise would be concentrated in an unaffordable pool.

Income taxes, per capita taxes, payroll taxes, fuel taxes -- any widely-applicable tax could theoretically be used to fund a population risk-allocation program. The reason the premium tax is attractive is that it directly enables allocating risks to insurers that have avoided accepting risk through cherry-picking activities, and can fund programs to subsidize remaining risks. This is only of interest if one expects insurers to act as payors. The reason that premium taxes are preferable to income taxes (or no-opt-out payroll taxes) is that the latter alternatives serve to charge twice those who have incomes (or are part of a payroll) but have paid for costly coverage despite not being in a cherry-picked group. The premium tax simply offers easier aim where the problem is: the cherry-picking insurers who cause the risk concentration -- and it offers the very solution desired by those intending to spread the cost of coverage, by sharing coverage costs to those who've opted out (for financial reasons, naturally) of the pool intended to put an end to uncovered risks.

One could as easily outlaw health insurance to avoid cherry-picking, then offer government programs to cover everyone. This isn't a particularly American solution, however, and it leaves some unhappy consequences -- among them, a lack of opportunity to see the effect of changes in the system, to learn how to improve it over time. A single monolithic federal system is likely to start looking like other government systems over time, with government-specified resource allocations, queues, professional practice structures, standards of treatment -- all slowly ossifying in a system that is not easily subject to improvement.

In a system that enables state-by-state regulation and halts federal prohibition on state regulation of health merely because of ostensible impacts on health benefit plans, we would see the successes and failures in plans like Oregon's and Hawaii's, and other states could work out how a real solution should be made. Innovations that produce clear results could be replicated. Innovations with adverse effects could be recognized in comparison to different systems, and those effects unwound.

It's perfectly fair to say that government-mandated health risk allocation is a bad idea. It's fair to disagree that health coverage should be addressed as a policy issue. However, so long as we have privately-funded emergency rooms required by law to stabilize anyone who arrives with an emergency, someone will be picking up these costs regardless whether injured or severely ill people can pay -- and we will have a risk transfer as a result. The question is whether the risk transfer will be by lottery to whomever hospitals can contractually shift risk (e.g., by overcharging paying customers to subsidize the care), or will simply serve over time to make hospitals unprofitable if they operate an emergency department.

In the case of government hospitals, the result is even clearer: the risk really is allocated to the public through taxes -- even to members of the public who've done the responsible thing and paid money to ensure their health risk isn't borne by some third party.

It's these folks who've already paid to make sure their risks aren't borne by third parties that I think are most offended that there's an expensive pool of cost-producers who are looking to tap their wallets again, for strangers' health.

I simply advocate that the risk transfer be made simpler by creating both a minimum coverage requirement (which in my view should be done at a level with some expertise in insurance, and not at the federal level where they have no clue at all because they've never regulated insurance) and a premium tax designed to counter cherry-picking. Allocating the premium tax among the covered risks' insurers can be done on the basis of insuring risks from the high-risk pool, or it can be done on the basis of a severity-adjusted insured risk comparison between insurers' coverage pools, or some other manner -- preferably, several different manners by different states so we can see what works -- but without coverage to ensure an equitable risk distribution we will get a non-equitable distribution as a result of the wrangling of everyone seeking to avoid uncovered risks' costs like they are tossing a hot potato.

I didn't say it was pretty, I just said it was an answer :-)