Cashing Out

Our household is just like a state: good at forecasting revenue, but when it comes to expenditures, things get dicier.
May 2005
Barrett and Greene
By Katherine Barrett & Richard Greene  |  Columnists
Government management experts. Their website is

Every April as we fill out our tax forms and reflect on our financial situation, the same question crops up: "What happened?" More specifically: "What happened to all that money we predicted we'd save this year?" And almost every year, the answer is about the same. Our income was what we expected it to be, but expenses for our family of four went up more than we anticipated.

If we had a few million more people living in our house, we'd be behaving just like many of the states. Absent unpredictable changes in the economy as a whole, we and they are not bad at forecasting revenues. But when it comes to the expenditure side of the equation, things get a little dicier.

Maine, for example, generally does a good job at estimating revenues, thanks to its consensus revenue forecasting committee, which includes members of both the executive and legislative branches. Meanwhile, when the governor develops and presents a budget, the Medicaid piece of that is based on prior history, trendline, cost per case and new eligibles. When that's presented to the appropriations committee, all those factors are on the table. But, says Christopher St. John, executive director of the Maine Center for Economic Policy, "They're not recorded in a particularly coherent way. Then, if either the legislative or executive branch needs a few more dollars, they'll go in and say 'Well, Medicaid's not going to cost so much.'" This results, of course, in mid-fiscal-year shortfalls when Medicaid exceeds the revised expectations.

Michigan earns reasonably high marks for its estimating of program costs, according to Dale Fickle, senior fiscal analyst at the Michigan League for Human Services. The exception is Medicaid, where the administration has low-balled both caseloads and costs for a few years. "I don't know if I'd say they've done that knowingly," he says, "but they've gone to the lower edge of any reasonable range, and it's come back to haunt them year after year."

The big problems in expenditure estimation most often occur in Medicaid and in other caseload-driven programs such as corrections. But there are some smaller expenses that aren't trouble free. In Massachusetts, according to Noah Berger, executive director of the Budget and Policy Center, the cost of snow and ice removal is always underestimated. "This is a way you can pretend to save money," he says. "They project that the following year, it won't snow. They know that if it snows, they will plow the streets. A supplemental budget is an annual rite of winter here."

States experiencing rapid growth have the most difficult job in estimating expenses--that's no surprise. But, it's most important for them to make accurate predictions, because the potential to be overwhelmed by a miscalculation is great.

In fast-growing Nevada, the Department of Human Resources, which handles health and human services, is conscientious about developing forecasting tools for expenditures. External partners are invited in to help--tax consultants as well as county government and legislative fiscal staff. "We try to make sure we're as smart and consistent as we can be," says director Michael Willden.

A handful of other states also are doing well on expenditure estimating--among them: Delaware, North Carolina, Utah, Virginia and Washington. North Carolina, for instance, formally models both revenues and expenditures and takes both legislative and administrative numbers into account. Jim Johnson, director of the legislative fiscal research division, says that "both sides of the street in this state have very experienced lead economists. Both sides employ economic firms to do outside econometric forecasting." Although there may be policy questions about determinate sentencing systems, when North Carolina undertook that effort about 10 years ago, it instituted a formal system of modeling prison population growth, which has stood up to the test of time.

The truth is that one of the most significant elements to good cost estimates is the resolve and desire to get them right in the first place. And when a state is in the middle of miserable budget debates, with no easy out in sight, it takes almost saintlike resolve to resist the temptation to fiddle around a bit with expenditure estimates. The temptation is even harder to resist since there generally isn't as much notice taken of mistakes on the expenditure side as on the revenue side.

It's like that in our house. Our son is going to college this fall, and we think we have a handle on costs, assuming gobs of financial aid. If we're wrong? We'll make college happen, even if we don't have the money. Just like the states.