University of Chicago economics professor Neale Mahoney and Harvard Kennedy School professor Jeffrey Liebman have an idea that they think could prevent wasteful end-of-year spending as agencies scramble to use the last of their expiring dollars: allow their budgets to roll over. I recently spoke to Mahoney about the idea. Our exchange has been edited and condensed.
Can you describe the “use it or lose it” agency spending pattern?
Agencies are cautious. They want to hold back spending early in the year in case there’s a natural disaster or something that requires them to spend a lot of money. Then, as it gets toward the end of the year and this crisis hasn’t occurred, there’s a “use it or lose it” mentality. Agencies also want to spend this money because Congress might say, “You didn’t spend your money, so we’ll give you less.” When agencies spend a lot quickly, they often spend it on lower-quality stuff. What we’re trying to accomplish in our model is to show that if you allow for rollover, you fix both problems. You fix the first problem, of agencies holding back on spending early in the year to build up a rainy-day fund, because they already have one. And you fix the problem of them draining the rainy-day fund because they can just carry it over. They also then theoretically won’t rush to spend on things that are of lower quality.
Do any agencies allow rollover spending?
There are certain types of budget authority that are not on an annual cycle. Nondiscretionary spending, obviously, is not on that cycle. There are some multiyear budgets. On regular appropriations, the Justice Department obtained permission to roll over its IT budget into subsequent years. So at DOJ, you don’t see the spike in year-end spending in IT—it’s basically flat over the course of the year. It’s a little speculative, but it does suggest that, if you allow agencies to roll over funding, they won’t engage in that type of year-end spending.
How would the rollover budgeting process work?
We outline three different thought exercises in our paper on this subject. One we think about is allowing agencies a time-limited grace period. We give them a couple of extra months to spend their money. This allows for some overlap between the new money coming in and the other money going out. We think it solves some share of the problem—not everything, but approximately half of the waste could be avoided by doing this. The second thing we examine with our mathematical model is when there’s some probability that Congress will raid your rainy-day fund—a model with an unlimited rollover period, but with 50 percent possibility Congress comes and takes your money. We show that even if there’s some risk that you might not be able to spend the money in the future, you’d still probably put some money in your rainy-day fund. The third thing we looked at—which isn’t realistic for the federal government, but is more realistic in local settings—is multiyear budgeting. If you’re a state or a municipality and set budgets every two years, you only have this issue half as often.
What issues could arise with rollover budgeting?
The problem with rollover, of course, is that agencies might be skeptical of using it because Senator So-and-So could come in and say, “You have all this money parked in this rainy-day fund; we’re going to give you less money next year.” We did in fact see a senator criticize DOJ for doing this. But it’s still worth having the extra time, even if there is some risk.