Friday, March 31, 2017

The cost of single payer: roughing it out

With sweeping ACA repeal apparently off the table for the moment, the Democratic imagination is turning once again toward some version of single payer, or Medicare for all.

I thought I'd try an overview from more than 30,000 feet -- say, from space -- of the tax revenue that might be required to move those now covered by commercial insurance into some kind of public insurance program.  That is, converting about a third of current total healthcare spending from private to public.  I hope this is not too general, or too vaguely informed, to be useful, but here goes...

1. According to the National Health Estimate, in 2015:
  • Medicare spending was $646.2 billion, or 20 percent of total NHE.
  • Medicaid spending was $545.1 billion, or 17 percent of total NHE.
  • Private health insurance spending was $1,072.1 billion, or 33 percent of total NHE.
  • Out of pocket spending was $338.1 billion, or 11 percent of total NHE.
2. Private health insurance covers more people than Medicare and Medicaid -- about 170 million (combining the employer and individual markets) vs. 120 million.* But it costs less in total -- about $1.1 billion vs. $1.2 billion.

3. At the same time, private insurance pays far higher rates to healthcare providers -- no one knows exactly how much more, but I've heard estimates in the 160% of Medicare range. Medicaid, in contrast, pays about two thirds of Medicare rates on average. [Update, 4/1: It's been pointed out to me that MEDPAC reports commercial rates for physician payments to be about 128% of Medicare -- that is, Medicare pays 78% of commercial rates on average. See p. 98 here. OTOH (added 4/4), a brand new CBO report estimates that commercial rates for hospital inpatient services average about 188% of Medicare, with wide variation. Hospital services accounted for a bit over $1 trillion in spending in 2015, compared to $635 billion for physician and clinical services.]

4, Total government expenditures are higher because the old and the disabled are expensive (though 33 million children covered by Medicaid and CHIP are relatively cheap).

5. If the federal government picked up the 170 million people currently enrolled in private insurance, the new or adapted program would pay considerably lower rates to providers than does the private sector. Plus there's the little matter of about 20 million remaining uninsured who are not undocumented to add to that total. If the uninsured cost as much to insure as those insured through the Medicaid expansion and marketplace subsidies, that would add about $100 billion per year to the tab.

6. At the same time, the government (or governments -- states pay nearly half of Medicaid costs) would probably have to pay somewhat higher rates in, say, Medicaid, to avoid system collapse. Or else, there would be a gradual phase-down from commercial rates in the new program covering half the population. Medical providers can't sustain a sudden double-digit overall rate cut.

7. If government-supplied insurance for those now covered by commercial netted out, say, 20% cheaper, that would mean about $900 billion per year in new government spending if the system could be projected back to 2015 - call it $1 trillion with the currently uninsured added.  According to NHE projections, private spending will be about 46% higher in 2022, while GDP rises more slowly, 36%, over the same period, and the consumer price index rises 18%. Savings from converting private spending to public might grow with time if current tight controls on Medicare spending hold up.

8. The new spending would be partly offset by saving the bulk of $234 billion in annual revenue (estimated, 2016) lost to the employer tax exclusion and other related tax breaks (including ACA marketplace tax credits) for health insurance benefits [or maybe not - see Bob Hertz comment below].

9. A new program, however blended with existing programs, might cost $800 billion in 2015 dollars, boosting total federal spending by about 20% (or somewhat more by 2022) and increasing federal health spending by about two thirds.

10. New taxes would have to raise a bit more than triple the amount raised by the Medicare payroll tax, which raised $234 billion in 2015.  So, another 9-10% in payroll taxes, however split between employer and employee, would be needed. [Update, 4/4 -- probably more like 10-11%, given update correction re likely savings from replacing tax-protected employer spending on healthcare with taxable increases in wages.]

11. At the same time, that federal spending would replace about $700 billion spent on premiums annually by employers (including government employers) and $310 billion spent by employees -- plus another $29 billion spent by the federal government on ACA marketplace subsidies ( $35 billion with Cost Sharing Reduction subsidies included) and $60 billion spent by individuals in the individual market (all in 2015).  .

Easy-peasy, right? Not. But ultimately worth it, if we could bell the cat. And savings would grow over time. The real trip is to get all healthcare payment rates under (gasp) government control.

P.S. The spending above - From the spending shares above, which represent almost all the actual spending on medical care, if Medicaid payment rates are 66% of Medicare and all others are 150%, I get a blended or weighted average payment rate of about 113% Medicare.


  1. Good effort, thanks.
    One point I have made numerous times to Gerald Friedman and others is this:

    - the $234 billion in tax breaks does magically flow into the Treasury to offset the cost of single payer.
    At the present time, the few firms which do drop health insurance rarely raise taxable wages at all.
    And even in your plan, employers are going to face some new payroll tax, and so will be reluctant to raise wages.

    Also, the single payer plan's new taxes of 9 to 10% can have NO exceptions or the plan will be desperately underfunded. There is no national will right now to force every little restaurant or dry cleaner to pay a new payroll tax, or to chase down everyone who works for cash and get them to pay new taxes also.

    1. That's true, isn't it. To the extent that the private spending is replaced by the tax, counting it as a source of added tax revenue is double counting.

  2. Pro tip to people who think that a marijuana tax would fund single payer: People do not spend 10% of their income on marijuana, on average. Even a 100% tax would be a drop in the bucket.

    Seriously, I've seen several people suggest this. We need to do a better job teaching critical thinking and estimation.

  3. I think there may be is a good reason employer provided insurance is more efficient than government provided. When large companies get insurance for their employees they are large enough to bargain with the insurance companies. And many times they are really self insured and the insurance company only administers the payments. Also Medicare part D forbids bargain ing for lower drug costs where private companies are not so constrained. This bargaining power is only available to large corporations or small business who manage to be part of some large group.