Hi’.
It has become an habit of mine for the foreigner I am to talk about the studies done on Medicare-for-All. I talked about its pricetag and then briefly felt the need to remind people of the keyword that is cost control.
With Elizabeth Warren being grilled on “how to pay for a plan to reduce health care costs”, CNN relied on the Urban Institute’s second study (2019) to once again say M4A will bankrupt everyone and their grandma. I personally rely on Pollin et al. (2018) — or “the PERI study” — but, as the idiotic centrist that I am, I wanted to check if my preferate study was wrong.
So let’s try to forget about my prejudices towards a study whose bibliography is two pages long (compared to Pollin et alii’s thirteen, let me be petty), we’ll just look at the numbers.
(As a reminder, there are other studies out there, I’m just singling those two out.)
1. Expanded care
Universal health care is the objective. That means the uninsured but also the underinsured. We need to know how many each study takes into account because the more people have access to care, and the better that care, the higher the cost (everything else equal).
The PERI study details 28 millions uninsured (p.24) and 85 millions underinsured (p.27).
The Urban Institute’s 2019 study say 34 millions uninsured (p.35) and… nothing about the underinsured.
The uninsured’s difference might be due to Trump’s sabotage, but also to the undocumented being covered in the Urban Institute’s scenario, which the Peri study seems to exclude. The Urban Institute does consider the extended coverage from M4A (p.8) but I cannot for the life of me find anything about underinsurance. Said otherwise: ‘til the Urban Institute decides to actually provide numbers, the PERI study is assuming a far larger expansion than them (see also Peri et al. 2018: 25 for uninsured spending).
2. Cost of that expansion
Okay, we have the number of people. How much does that expanded coverage cost?
The PERI study estimates a 16% increase for the underinsured, and a 65% increase for the uninsured (pp.35-36, table 6), totalling (for year 2017, total spending $3.3tr) a $133bn and $167bn increase respectively, or a $300bn increase total (9.3%). They then compare that with Thorpe’s study (which find the same increase ratio) and go for the higher-end possible increase, 12%, or (in 2017) $390bn.
The Urban Institute’s 2019 study says total health care spending will rise by $700bn (p.37) for 2020, or a 20.5% increase. Aaaand that’s it.
To be fair, they do give price elasticities due to the lack of cost-sharing (p.61) but uh… yeah. It uhm, it exists. Yay.
3. Cost control
Right now we have a $420bn pricetag versus a $700bn one (both for 2020, 12% vs 20%, every absolute number will be based on 2020 now). But that’s half of the story. How much does Medicare-for-All actually save?
3.1. Administrative costs
The PERI study relies on international and historical comparison (p.45) to derive its 3.5% ratio, and hence a 5% cost saving, or $175bn. That’s only part of the story though: that’s for insurance. There are also savings for hospitals and doctors, whose share is higher (8% and 13% respectively), leading to another 4%, or $140bn, for a total of 9%, or $315bn.
The Urban Institute relies on a 6% ratio (p.38). A 3% ratio reduction would translate into $77bn, so the saving should be ~$65bn. I have to assume. I don’t know.
3.2. Medicare rates
The PERI study does use the base Medicare rate, at 78% of private rates (p.53), leading to a 2.8% cost saving (between hospitals and doctors), or ~$100bn.
The Urban Institute uses a 115% rate (p.38), which I can only guess would equal to a ~$180bn. Without them giving actual numbers it’s just me throwing guesses.
3.3. Prescription drugs
There is a whole discussion in the PERI study about it, so long story short (p.44), it’s a 6% savings, or ~$210bn.
The Urban Institute gives no number. It’s there, it’s mentioned, but no number whatsoever.
3.4. Other savings
There are plenty more cost saving potentials in the PERI study, summed up as (p.67) : unnecessary services, inefficiently delivered services, “preventive care” and fraud. They seem to only apply to the transition phase though, but still, that’s cost reduction of 1.5%, or $52bn.
There is no such mention in the Urban Institute.
4. So who is right?
As a summary, the PERI study estimates a 12% cost increase and a rough 17% cost saving, for a difference of -5% in cost, or $-175bn/year (2020).
The Urban Institute’s 2019 study estimates… a 20% cost increase already taking cost savings into account I guess. Or a $700bn/year (2020).
Looking at cost savings, I would say Pollin et al. (2018) are looking for a high-end. Higher Medicare rates are reasonable (115% sounds about alright, given current market rates are actually roughly ~180% higher). Administrative costs of 5-6% wouldn’t be surprising. When playing with the taxpayer money, let’s not be conservative, ‘cause they are conmen, but let’s be careful. So on cost savings, I would actually align with the Urban Institute.
But here is the thing: it doesn’t make much of a difference. Those two changes take away 2.5% and let’s say 0.8% respectively (that second number is arbitrary), leaving a 13% cost saving ratio, or -1% in cost, $-35bn/year (2020).
But if on cost savings I’m siding with the Urban Institute, to be safe, to be scientific I have to side with the PERI study when it comes to cost increase. Contrary to the Urban Institute and its habit to give near no number (their M4A analysis holds on roughly 8 pages, plus a couple pages of… methodology), Pollin and colleagues have detailed every decision and even offered the calculus. It’s simple, grounded, supported by previous and parallel works and even though the actual increase would be 9%, they go for the high-end 12%. The actual difference could be even bigger, in favor of M4A.
5. The Urban Institute is wrong, plain and simple.
I have quoted Pollin et al. on that (p.162), I’ll do it again:
As we saw in Chapter 2, between the PERI and Thorpe approaches, the estimates for spending increases for the currently insured range between 4.2 percent for PERI and, with Thorpe, 7.0 percent for the non-elderly and 7.5 percent for Medicare recipients. All of these figures are less than half of the Urban Institute’s 15.5 percent unsupported figure. The Blahous assumptions on utilization increases for the currently insured are 11 percent for the non-elderly, and 16 percent for those insured through traditional Medicare. Thus with Blahous as well, his overall assumption on utilization increases is below that assumed by the Urban Institute.
The Urban Institute concludes that, overall, as of 2017, Medicare for All would cost 16.9 percent more than the existing U.S. health care system (p.2). This estimate is based, first, on their assumption that overal utilization would rise by 15.5 percent. But in addition, the Urban Institute report gives no consideration to potential sources of cost saving under Medicare for All. This assumption is inconsistent with a wide range of evidence, as we review in both Chapter 3 and this appendix.
I think the Urban Institute’s second take (this criticizes their 2016 study) does take potential sources of cost saving into account, but their utilization increase ratio is still completely off the chart and they have yet to actually justify it. I would add you don’t consume health care like you consume bread, so using elasticity is brick dumb, but I’m no expert so I’ll refrain.
Once again: Medicare-for-All saves money.
It expands coverage to the whole population, ensures proper coverage for all and makes that coverage cheaper for all (but a small minority), it is revenue-neutral at worst if you are risk-averse like me. It does not cost money. It saves money. Those are the actual numbers. To the best of our knowledge, this is what it does.
I still want Warren to say how the money will be moved around, I’m curious to see if it will be payroll or income… but she doesn’t have to “pay for it”. It’s paid for. It saves money.
It is a plan to reduce health care costs.