Could Federal Spending be Capped at 20 Percent of GDP? Should it Be?

As the budget debate heats up, we will hear much about capping U.S. federal government spending at 20 percent of GDP, roughly its level for several decades leading up to the global financial crisis. Likely presidential candidate Rep. Mike Pence (R-Ind.) has been among the most vocal backers of this idea. Together with colleagues, he has packaged it in the form of a proposed Spending Limit Amendment. Would it really be possible to impose such a spending limit? Yes. Would we like the results if we tried it? Not all of us would. Here is why.

Part of its attraction is that the 20 percent solution appears to require no real sacrifice. If we were content with the level of government services enjoyed in past, pre-crisis decades, why would there be any hardship in holding to that level in the future? Unfortunately, the pretense that it would be possible to maintain the same level of real government services as in the past without future increases in spending is an illusion. The reality is that holding government spending to past levels would require a significant reduction in real public services. To see why, we need to look in more detail at what has happened to individual components of government spending over past decades. The following chart gives the basic data.

First, the chart shows that the discretionary component of federal spending was not constant over the decades leading up to the crisis. Instead, discretionary spending, which includes defense, law enforcement, education, highways, and all the daily business of government, followed a long downward trend, from more than 12 percent of GDP in the 1970s to less than 8 percent in the mid-2000s. A 20 percent cap on total federal outlays would lock in not current levels of discretionary spending, but the downward trend.

Second, we notice that there was a moderate recovery in discretionary spending in the mid-2000s. The recovery was facilitated by a decrease in net interest payments on the federal debt, which fell from their usual level of around 3 percent of GDP to about half that. The decline occurred partly because of unusually low interest rates and partly because of a decrease in total debt due to the the budget surpluses that appeared at the end of the Clinton administration. Unfortunately, federal interest expense is now on its way back up, with a vengeance. By the end of this decade, it will be well above its historical average. If total spending is capped at 20 percent of GDP, interest payments will soon be adding to the squeeze on discretionary programs, rather than relieving it, as in the mid-2000s.

A third budget trend evident from the chart is the growth of entitlement spending, from 6 percent of GDP in 1970 to over 10 percent by the mid-2000s. Perhaps something could be done here. If the growth in the share of entitlement spending could at least be stopped, even if not reversed, that might preserve at least an austere sliver of funding for defense, schools, and highways.

Unfortunately, the idea of freezing the growth of entitlement spending runs up against some hard demographic realities. The growth of entitlements, mostly in the form of social security and Medicare, does not primarily reflect increasing generosity of these programs. True, in some respects benefits have grown, for example, through the addition of drug benefits to Medicare and overly generous inflation adjustment of  social security benefits. However, future growth in entitlements will be driven less by growth in benefits per recipient than by demographics, especially a dramatic increase in the country's elderly population, as shown in this chart:

The implication of these demographic trends is that capping total entitlement spending as a share of GDP would do more than simply deny benefit increases to future generations of retirees. Instead, it would mean sharply cutting both social security and medicare benefits per recipient. Real cuts would be needed even if stringent cost-cutting measures held medical inflation to the rate of increase of average consumer prices.

So, could federal spending be capped at 20 percent of GDP? Yes, it could. Some set of measures like this would do the job:

• Cram discretionary spending down to its pre-9/11 level of 6.3 percent of GDP, compared with 8 percent in the last year of the George W. Bush administration. That would make room in the budget for a doubling of net interest payments, which, given the size to which the total debt has already grown, will inevitably be necessary when interest rates recover to normal.
• Within the discretionary spending category, make a choice between drastic cuts to defense spending, including withdrawal of troops from foreign deployments, or cuts deep into the muscle, not just the fat, of every domestic discretionary program, including law enforcement, education, infrastructure and all the rest.
• Cap entitlements at their current level of GDP. Convert social security from a pension program for all into a means-tested safety net for the neediest. Implement cost-cutting measures for Medicare that have been rejected in the past, such as malpractice reform, elimination of tax preferences for employer-provided plans, a phase-out of fee-for-service medicine, and increased co-pays for all but the lowest-income elderly.

Should it be done? That depends on what kind of government you want. If you want approximately the real level of government services per capita as in the past, and the same level of real social security and medical benefits per retiree, then forget about the 20 percent solution. On the other hand, if you are content with a pared down government that provides a pre-1930s social safety net along with third-world levels of defense, education and infrastructure spending, then by all means, slap on the cap. But please, don't pretend you can have it both ways.

Follow this link to view or download a brief slideshow containing additional data related to the federal budget and the 20 percent solution.