The Washington PostDemocracy Dies in Darkness

Health care is crowding out everything. Social Security isn’t.

Third Way has generously responded to my critique of its report alleging that spending on Social Security, Medicare and Medicaid is crowding out spending on investments like education, research and development, and infrastructure. Given what good deals some of those investments are, this would be a serious problem indeed. But while it's clear that health care spending, including Medicare and Medicaid, is growing at an alarming rate that threatens other priorities, I don't think the evidence supports the view that Social Security is doing the same.

Jim Kessler of Third Way disagrees:

[Matthews] is correct, and we showed in our paper, that over the past 50 years, all of Social Security’s growth relative to GDP has occurred in the first 20 years and has stayed roughly static since. But five percent of the economy is a lot. It’s roughly equal to Medicare, Medicaid and CHIP combined. And it’s not going to stay static – that is a certainly. In less than two decades, Social Security is poised to jump from 5.0% to 6.0% of GDP, according to CBO. One point may not seem like a big deal, but it represents a 20% rise in the cost of Social Security relative to the size of the economy.

No one disagrees that the retirement of the baby boomers is going to increase the cost of Social Security as a percentage of GDP. But it's worth noting that the increased cost does not deplete the Social Security trust fund until 2040, according to the CBO. And it got that trust fund because other programs kept borrowing from Social Security's revenues and are required by law to pay that debt back. So more and more of our Social Security obligations will take the form of debt repayment financed through general revenues, principally the personal income tax.

One way to view this is as Kessler does, as Social Security "crowding out" other priorities in years to come. The other way to view it is that Social Security has been, in effect, subsidizing the rest of the government for decades and now it's their turn to pay it back. In that view, Social Security hasn't been crowding out other spending at all. It's been enabling it! The net effect on spending by the rest of the government, over the course of the loans and their repayment, is zero. So until 2040, talking about Social Security crowding out spending by the rest of the government is conceptually sort of weird.

This becomes clearer if you compare changes in Social Security spending to changes in investment spending during the period studied by Third Way. If the "crowding out" story is true, there should be a substantial correlation between changes in Social Security spending, as a percentage of GDP, and changes in investment spending. But  there is no such correlation. Changes in Social Security spending only accounted for 2.4 percent of the change in total investment spending, 2.8 percent of the change in defense investment spending, and 2 percent of the change in non-defense investment spending. Not only that, but the correlation coefficients are positive - that is, increases in Social Security spending are actually mildly associated with increases in investment spending! This isn't to say that the two actually promote each other, just that there isn't really a relationship at all, historically.

Kessler writes that "These two areas of the budget – both of which are critical priorities for America, born of Democratic Presidents and Congresses – are locked in a zero sum game." This is the bigger problem with his narrative. The way the Third Way paper talks about this, you'd get the impression that entitlements and investment spending together make up most of the federal budget. But they don't: