By James Kwak
Single-payer health care has emerged as the primary symbol of the differences between Bernie Sanders and Hillary Clinton. For his supporters, Sanders’s plan is the centerpiece of his vision of a European-style, social-democratic society in which health care is a right that is fully funded by largely progressive taxes. For Clinton supporters, it is the best example of Sanders’s misguided utopianism—a naive plan that has no chance of passage and that would be less effective than incremental improvements to Obamacare.
The Clinton camp has been ecstatic about a critical analysis of Sanders’s plan by Kenneth Thorpe, a health care expert who has supported single payer in the past. Thorpe argues that the plan is underfunded by more than $1 trillion and, if fully funded, would cause most people to pay more for health care than they do today. Thorpe’s analysis was publicized by Dylan Matthews and highlighted by Paul Krugman as further evidence that all the serious people support Clinton. Clinton herself likes to say, “The numbers just don’t add up.”
I have a few points to make about Thorpe’s analysis and what we should take from it. The first, simplest, and most obvious is this: Even if Sanders’s plan has problems, if the goal is single payer, then Sanders is the only choice. We know a Clinton nomination will bring us no closer to single payer. A Sanders nomination will bring it within the spectrum of future possibilities, even if it won’t be passed in the next two years. (Remember Newt Gingrich’s crazy plan to voucherize Medicare in 1995? Well, now it’s not that far from passage.) A Sanders presidency would make it possible to develop a better, more realistic plan; that’s what advocates of single payer should be hoping for.
OK, now to the specifics. Thorpe’s most important conclusion is that Sanders’s plan will cost an average of $2.5 trillion per year, not $1.4 trillion, and that financing it will require significantly higher taxes. As a result, most Americans would end up paying more for health care (in additional taxes) than they currently do (in premiums they pay, premiums paid by their employers, and deductibles, co-payments, and coinsurance).
But there’s something funny about this conclusion. We’re not talking about smart bombs and cruise missiles that will explode somewhere in the Middle East. We’re talking about health care services being provided to Americans. So if we (in the aggregate) have to pay more in taxes, we’re getting it back in the form of more health care. On its face, that’s not necessarily a bad thing.
As the country that already spends more money per person on health care than any other in the world, however, perhaps increasing spending is not a good idea. So let’s look at why total spending goes up. For the total to go up, the price per unit of health care must go up, the number of units delivered must go up, or both. I’m pretty sure that Thorpe doesn’t think that the price per unit would go up under Sanders’s plan. He projects an overall payment rate of 105% of treatment costs. Medicare currently pays less, so it will pay more; but providers make up for that by overcharging private payers, and Thorpe’s model says those costs will go down by 20%. In addition, Thorpe projects 4.7% savings on administrative costs relative to the private system.
So the problem must be that the number of units is going up. And that is basically what Thorpe says. He assumes a 10% increase in utilization by people who already have insurance because the Sanders plan eliminates cost sharing, plus additional spending on people who are currently uninsured. Again, this might be a good thing: not only is it good to provide coverage to the 11% of people who are still uninsured under Obamacare, but there are also plenty of insured families who are forgoing treatment because of high cost sharing.
But let’s assume for the sake of argument that higher utilization is bad. Then the obvious solution is not to toss out the baby with the bathwater, but to put some cost sharing back in. For example, we could have income-based, sliding-scale cost sharing, where the idea is that deductibles and co-payments should be roughly equally significant at different income levels. (Obamacare includes cost-sharing subsidies for low-income households.) So people on minimum wage face a $5 copayment, and people like me face a $100 copayment. I’m fine with that.
The point is, some of the projected increase in utilization is good, because it means people who currently are being underserved are getting the health care they need. And if we think some higher utilization is bad, we can limit it through smarter cost sharing.
What really matters as far as aggregate, long-term health care costs and outcomes are concerned are two things: the price per unit and the value delivered per service. And on both dimensions, we have every reason to believe that single payer will be more effective than Obamacare (and far more effective than not-Obamacare, which is all the Republicans are offering). Obamacare includes a number of tools to try to increase efficiency among providers and bend the cost curve: the Independent Payment Advisory Board, the incentives for Accountable Care Organizations, experimentation with different payment structures, tying hospital payments to readmissions, bundled payments, and so on. Do you know what all of those have in common? They’re all part of Medicare.
Even with Obamacare, the federal government’s influence over the private insurance market is limited. So in order to steer the health care sector toward more effective procedures and more accountability in payment practices, it works through Medicare and, to a lesser extent, through regulations governing the exchanges and employer-sponsored plans. The problem is that just because Medicare pioneers some value-based reimbursement practice doesn’t mean the private sector has to follow. If we really want to bend the cost curve in an intelligent way—and most Democrats do—then it would be much more straightforward under a single payer system than under Obamacare.
As I’m fond of saying, the government is just a tool that we use to arrange some of our affairs. Whether we buy our health care through private insurers or through a single payer, what matters are the outcomes we get for the money we spend. It’s OK to spend more money if we get truly universal coverage and better outcomes; the elderly get a lot more medical care today than they did before Medicare was created in 1965, and most of us don’t think that’s a bad thing. But if we think that we’re buying too much health care, or we’re getting the wrong kinds of health care, a single payer system is a more effective way of redirecting spending priorities—in a democratic manner—than the unwieldy public-private mash-up we have today. Single payer has its downsides, such as the bureaucracy (not that we don’t have one already). But whether we can afford it is not one of them.
 I agree with Thorpe that it makes sense to evaluate the plan on a fully financed basis. Contrast this with Republican tax cut plans, which talk about how much money people in each income quintile will save—but not how much they will lose in government transfers and services if the tax cut is fully financed by spending cuts.
 Technically speaking, it’s possible that total spending will not go up. As far as I can tell, Thorpe’s tables only say that most people will see an increase in total costs, assuming full financing. So maybe a minority of people will see their costs decline so much that they will balance out everyone else. That would be a good thing, both because total spending would go down and because risk-sharing would increase (basically, this would be a transfer from the healthy to the expensively sick).
 I still don’t quite see how total spending goes up with these assumptions. Spending on people who currently have private insurance—most of the population—clearly goes down by about 15%. (“We take private insurance and out of pocket spending … and make several adjustments to develop a single payer funding estimate. First we increase it by 10 percent to reflect the increased total health care spending that results from a reduction in out of pocket payments. Second we adjust downward by 20% to reflect the lower blended payment and finally we reduced the total by 4.7 percent to reflect potential administrative costs savings.”) Spending on the uninsured and Medicaid and Medicare beneficiaries goes up. Thorpe’s report doesn’t specifically compare spending under the Sanders plan to spending currently; the table only compare the number of households that would pay more or less under the (fully financed) Sanders plan than they do today. Matthews’s article compares Thorpe’s estimates to those of the Sanders team, but also does not compare his estimates to current aggregate spending levels. But again, if total spending does not go up, then there’s no problem to begin with (see previous footnote).
 One quibble: Thorpe notes that Sanders assumes that the states continue spending the amount that they are currently spending on Medicaid, even though there is now way the federal government can compel them to. But if we’re going to cross out the revenue provided by state contributions, then we also have to cross out the state taxes that raise that revenue, so in aggregate its a wash (and those state taxes are often pretty regressive).