You've heard — perhaps on this very blog! — that our long-term deficits are almost entirely driven by health-care costs. That's true over the next 50, 60, 70 years, which is, absurdly, the time frame people often talk in. But over the next 20 years, it's not quite right.

A more accurate way to put it would be that in the coming decades, new spending is almost entirely driven by health-care programs. But what's really driving the spending in those programs is the aging of the population, not the rise in health-care costs. Over at the Concord Coalition's blog, Joshua Gordon makes this point in an unusually clear way — by which I mean, of course, with graphs.

Here's a breakdown of new spending by program between 2012 and 2037. As you can see, Medicare and Medicaid far outpace Social Security, and all programs that are not Medicare, Medicaid or Social Security are expected to shrink as a percentage of the economy:

There are two reasons health-care programs could be spending more. One is that health-care costs are going up. The other is that more people are using them. We typically talk about the problem as if the problem is rising costs. But over the next 20 years, the cost increases are driven by more people.

Much of the growth in Medicaid spending is coming because the Affordable Care Act expands Medicaid to cover many more people. But the growth in Medicare is coming because the population is aging into it rapidly. This graph tells the tale:

This doesn't fully account for the recent slowdown in Medicare costs, so if current trends continue, those red lines are too large. But even at this higher estimate, the increase in spending is really coming from the increase in old people. As Gordon writes, "The total increase in Medicare costs from [health care] inflation is only 0.7 percent of GDP. Alternatively, growth due to aging accounts for 1.9 percent of GDP."

There's a reason that policymakers prefer to talk about health-care costs than old people. If the problem is just rising costs, then perhaps there's some cost control "silver buller" —  maybe premium support, or paying for quality rather than service — that will cut costs without hurting anyone. But if the problem is more people, then the answer, really, is higher taxes, lower benefits, more debt  or some combination of the three.

Which isn't to say that rising health-care costs aren't a problem. The 75-year predictions, which I don't put much stock in, really are about health-care cost inflation. Consider this chart from the Council of Economic Advisers' most recent report:

If current cost trends hold, then after the demographic bulge ends, Medicare will actually shrink as a share of the economy, consuming little more in 2085 than it is today. As I've tried to emphasize throughout this post, I think 75-year predictions are a bit ridiculous — here's a fuller argument on that point — but since they're in wide use around Washington, it's worth being clear about what's driving them. And that is, over the next 20 years, more people using government health care, particularly Medicare, and over the next 75 years, health-care costs growing a lot faster than they have in recent years.