Make the Patient the Customer

Section 1The question before us

A man gets a small dent in his car and pays for it himself, because he does not want his rates to go up. He would never submit an oil change to his insurer. When a mechanic quotes him a number that sounds wrong, he says so. He knows what his policy costs because he bought it, and he has probably changed carriers at least once because somebody advertised a better price.

That same man cannot tell you what his health insurance costs, did not choose it, has never asked a doctor the price of anything, and submits every routine visit to a third party. He is not behaving inconsistently. He is behaving correctly in two different systems, only one of which was designed.

Car insurance is cheap partly because the person buying it is the person paying for it. Nobody in a medical transaction is doing both.

The doctor bills an entity the patient never speaks to. The patient's employer chose the plan. The employer's cost is hidden inside wages the employee never sees quoted. And the price of the service is not merely unknown, it does not meaningfully exist — the same procedure carries prices varying five-fold within a single state, unrelated to quality.

1.2   Our own share of the failure

The employer-provided health benefit is the largest single tax preference in the American code, and the people who complain most about government distorting markets have spent seventy years defending it.

It is a subsidy. It is worth more to higher earners than to lower ones, because an exclusion is worth your marginal rate. It is unavailable to the self-employed on the same terms, to the unemployed at all, and to anyone whose employer does not offer it. It ties a family's medical coverage to one member's job, which is a strange thing for people who believe in labor mobility to defend. And it is the single largest reason no consumer market in medicine ever developed.

This author has claimed it, both as an employer and as an employee. It is on the list of preferences Issue Paper No. 8 proposes to eliminate, and it belongs there.

Section 2What we know about the waste

In 2019 researchers at Humana and the University of Pittsburgh published an estimate in JAMA of waste in the American health system, drawing on 71 estimates from 54 publications across six categories. The total came to $760 to $935 billion a year, about 25 percent of all health spending.1

Two features of that finding matter more than the headline.

The largest single category is administrative complexity — and the authors could find no studied intervention that reduces it, which is why their estimate of achievable savings excludes it entirely.1 American health administration runs at roughly double the Canadian share of spending.1

The second largest is pricing failure. Not that care is expensive to deliver, but that what is charged bears no stable relationship to what it costs.

Note what both of those are. They are not the cost of medicine. They are the cost of the payment system — of every insurer maintaining a different formulary, network, prior-authorization process and fee schedule, and every provider maintaining staff to negotiate with all of them. A patient paying a posted price generates none of it.

That is the strongest empirical support for the argument in this paper, and it is worth being precise about what it does and does not show. It shows that a quarter of the money is not buying medicine. It does not show that a consumer market would recover all of it.

Section 3Reform I: End the wartime accident

The exclusion for employer-provided health insurance should be eliminated, with the revenue returned through lower rates, including lower payroll rates.

Almost nobody chose this system. During the Second World War, federal wage controls prevented employers from competing for scarce workers on pay, so they competed on benefits. A 1943 ruling made employer-paid premiums exempt from taxation, and Congress wrote the exclusion into the code in 1954.2 An emergency workaround to a price control became the organizing principle of American medicine.

The consequences follow mechanically. Carriers market to employers rather than to patients, so plans come in one size. Patients cannot take coverage with them when they change jobs, so a family's medical security is attached to an employment relationship neither party controls. And because the subsidy is worth your marginal rate, the exclusion delivers more to a partner at a law firm than to a warehouse worker.

The Joint Committee on Taxation puts the exclusion at roughly $226 billion a year — and states that its estimates exclude the payroll tax effect, so the true figure is considerably larger.2 Any repeal must therefore return part of the offset as a payroll rate cut, or a middle-income family with employer coverage comes out behind on its first pay period.

One disclosure, because the same dollar appears in more than one of these papers and a reader is entitled to know it. This exclusion is among the tax preferences Issue Paper No. 8 proposes to eliminate, and Issue Paper No. 11 identifies retained base broadening as the principal source of financing for the Social Security transition. If all three reforms proceed, the revenue from repealing this exclusion cannot be returned in full as rate reduction. Some of it would be retained. We would rather state the conflict than let three papers each spend it.

Section 4Reform II: Two vouchers, on the same curve

Replace Medicare and Medicaid with two vouchers: one to purchase catastrophic insurance, one for routine care and medication. Both phase out on the single schedule described in Issue Paper No. 9 — not on a means test of their own.

4.1   Why this is a smaller change than it sounds

Neither program is what its name suggests.

Medicare Part A is contributory, but pay-as-you-go — today's workers fund today's retirees and nothing was pre-funded. Parts B and D are voluntary programs in which premiums cover about 25 percent and 13 percent of costs respectively, with general revenue covering the rest. Across the whole program, only about 57 percent of Medicare's cost comes from dedicated sources.4 Medicaid has no premiums, no contributions and no trust fund at all, and most states already deliver it by paying private managed-care companies a fixed sum per enrollee.

So the government is already buying insurance on people's behalf with general revenue. This reform changes who holds the purchasing decision, not whether the public is paying.

4.2   Why the phase-out must sit on the common curve

This is the provision we would defend hardest and it is easy to get wrong.

Comprehensive coverage is worth somewhere between eight and twenty-five thousand dollars a year. Means-test it separately and the implicit marginal rate inside the phase-out band is enormous — larger than any cliff in the current system, and created by the reform meant to remove them.

A voucher that vanishes at sixty thousand dollars makes the dollar at fifty-nine thousand nine hundred and ninety-nine the most expensive dollar earned in America.

One schedule, one taper, health included in the sum. The rule from Issue Paper No. 9 applies without exception: all withdrawals plus all taxes must never exceed the top statutory rate.

Section 5Reform III: Keep the change on price — never on care

Where a patient chooses a provider charging less than the reference amount, he keeps the difference. Where he declines care altogether, he keeps nothing.

That distinction is the entire design and it is not a technicality.

5.1   It has been done, and it worked better than expected

In 2011 the California Public Employees' Retirement System set a reference price of $30,000 for hip and knee replacement, after finding prices varying five-fold across hospitals with no relationship to quality. Members choosing a facility at or below the reference price paid ordinary cost sharing. Members choosing a more expensive one paid the difference.

The share of members using low-price facilities rose from 48 to 63 percent. Volumes rose 21.2 percent at low-price hospitals and fell 34.3 percent at high-price ones. Prices charged fell 5.6 percent at low-price facilities and 34.3 percent at high-price ones. Average price paid fell about 26 percent — more than $9,000 per procedure — for roughly $5.5 million in savings over two years.3

And the finding that matters most: Berkeley researchers attributed roughly 85 percent of the savings to hospitals cutting their prices, not to patients switching. Some renegotiated rates across their entire book of business.3

The patient does not have to become a skilled shopper. The credible possibility that he might is enough to discipline the supplier. That is a considerably lower bar than the argument usually requires, and it is the reason we think this mechanism travels.

We should report the disagreement. One analysis found members saved roughly $300,000 in reduced cost sharing; another found members paid roughly $700,000 more in the first year, because some stayed with an existing surgeon and absorbed the difference.3 Both cannot be right, and any published version should resolve it.

5.2   Why the patient must never profit from going without

A voucher that can be pocketed by skipping the appointment is a cost-sharing scheme wearing a friendlier name, and there is definitive evidence on what those produce.

The RAND Health Insurance Experiment randomized thousands of families across levels of cost sharing through the 1970s. Higher cost sharing reduced spending substantially. But participants reduced effective and ineffective care in roughly equal proportion — they were not able to tell which was which — and among poorer and sicker participants, health outcomes were worse.6

CalPERS never paid anyone to skip an operation. It paid them to have it somewhere cheaper. That is why prices fell without outcomes falling, and it is the line this reform must not cross.

Section 6Reform IV: Preventive care free at the point of use

Screening, immunization, prenatal care, chronic-disease management and the medications that prevent expensive events should carry no patient cost at all — not a copay, not a deductible, and no keep-the-change incentive.

The incentive runs the wrong way everywhere else. Every dollar a patient saves by not taking a statin is a dollar the system spends on the heart attack, and the patient captures the saving now while somebody else pays the cost in a decade. Keep-the-change is a good instrument for choosing where to have a colonoscopy and a terrible one for deciding whether to have one.

Note that this is not a concession to the argument — it is the argument applied correctly. A market works when prices reflect costs. The cost of skipped prevention is real, it is large, and it arrives on a delay that no individual consumer has any reason to price.

Section 7Reform V: A price must exist before anyone can shop

Providers should be required to publish a binding cash price for every service they offer, and to quote it in advance on request. A price that appears only on the bill is not a price.

Everything else in this paper is decoration without this. Consumers cannot discipline suppliers they cannot compare, and the current chargemaster system — list prices bearing no relation to what anyone pays, negotiated rates concealed as trade secrets — exists precisely to prevent comparison.

This is also the least radical proposal in the paper. Every other industry in the country manages to tell customers what things cost before they buy them.

Section 9What we are not claiming

We are not claiming this reduces anyone's care. The proposal converts existing spending into vouchers; it does not cut it. Nothing here requires anyone to receive less, and if a version emerged that did, we would oppose it.

We are not claiming consumers are good judges of clinical value. The RAND experiment says plainly that they are not — which is why prevention is free at the point of use and why keep-the-change applies only to where care is purchased, never to whether it is.

We are not claiming to know how much costs would fall. Nobody does. A quarter of spending is waste; how much of it a consumer market recovers is unknown, and any number offered is a guess wearing a decimal point.

And we are not claiming this is proven. Four of the five reforms here rest on demonstrated results. The system as a whole does not exist anywhere in the form proposed, and by the standard this organization applies to everyone else, that is a real objection rather than a quibble.

Section 10The argument you can carry

The paper compressed to what a person can remember and repeat.

And the one that is ours rather than theirs. The employer health exclusion is the largest tax preference in the American code, it is worth most to the people who need it least, and those of us who complain about government distorting markets have spent seventy years defending it. This author claimed it too.

Section 11Conclusion

The expensive thing about American medicine may not be the medicine.

A quarter of the spending is waste. The largest identified category is administrative complexity — the machinery of billing, negotiating, authorizing and denying, which exists because the person receiving care is never the person paying for it. The second largest is pricing failure, which is what happens when nobody in the transaction knows or cares what anything costs.

Neither of those is the cost of a doctor's time or a hospital's equipment. They are the cost of an arrangement that began as a workaround to a wartime wage control and was never examined again.

What we propose is not a smaller commitment to anyone's health. It is the same money, handed to the patient instead of to an intermediary he never chose, with the right to keep what he saves by choosing well — and with prevention free, because we want more of it, and with a floor beneath everyone, because the alternative is indefensible.

We have said clearly what this cannot do. It does not reach the half of spending concentrated in the sickest patients. It has never been run at national scale. And it requires real regulation to stop insurers from selecting the healthy, which is an awkward thing for a market proposal to need.

But a system in which the same operation costs five times as much across town, and nobody involved can tell you why, is not an expensive market. It is the absence of one, and we have been paying for that absence for eighty years.