SJR 350

Commission on the Commonwealth's Planning and Budget Process

October 15, 1997, Richmond

The third meeting of the commission featured two perspectives on developments in performance based budgeting, revenue forecasting, and long-range expenditure forecasting in other states.

Performance Based Budgeting

According to a representative of the National Conference of State Legislatures, performance based budgeting "is more than a fad." Two-thirds of the states have adopted legislation calling for the use of performance based budgeting, and 14 other states have initiatives not involving legislation.

Performance based budgeting's central elements are using strategic planning to set the mission, goals and objectives of programs; measuring programs' outcomes; and setting benchmarks to be met by programs. It is intended to hold agencies accountable for performance, provide flexibility in using resources within programs, and encourage management innovation.

Integrating performance based budgeting into traditional budget processes has not been easy. A transition to this method of budgeting requires a focus on activities and outcomes rather than on line-item expenditures. Agency structures may have to change to put responsibility for a given activity in one agency.

North Carolina, Florida and Texas are leading other states in attempting to implement performance based budgeting. When North Carolina's governor introduced both performance based and traditional budgets for the 1997_1999 biennium, the legislature ignored the performance based budget. Florida is phasing in performance budgeting over seven years. In Texas, agencies have developed 11,000 performance measures, and "key" outcome measures are printed in the budget bill.

The experiences of these states reveal difficulties in creating meaningful performance measures. Agencies often lack the experience and historical data needed to provide a basis for comparing performance. In addition, the creation of performance measures must involve both the legislative and executive branches. The NCSL representative also cautioned that performance based budgeting cannot make appropriations decisions. No one has determined how to tie budget allocations to measures and performance. Moreover, it is not clear whether appropriations to an agency that fails to meet its performance goals should be increased or decreased.

Performance based budgeting can be a valuable agency management tool. It provides a mechanism for systematic review of an agency's organization and purpose, helps identify those parts of government that are performing well, and encourages long-term planning. It also gives legislators new, and possibly better, types of information and helps legislators focus on program outcomes.

The former executive director of the National Association of State Budget Officers argued that the concept of performance based budgeting is often oversold. While 48 states claim to be implementing it in some fashion, its main benefit is helping agencies use their resources more efficiently. He praised Virginia's efforts to adopt agency-wide performance measures. While improving efficiency is a laudable goal, it is no substitute for making policy decisions. To the extent performance based budgeting attempts to substitute applying empirical data for exercising policy judgments, it threatens the political system.

Legislative Branch Revenue Forecasting

In 13 states revenue forecasts are prepared by both the legislative and executive branches. In the seven states where the constitution or law requires a consensus revenue forecast, establishing a legislative forecasting capacity is needed to make the legislature a full player in the process. In the other states, the legislature either tends to adopt the legislative forecast or to strike a compromise between the legislative and executive forecasts.

Eleven of the states preparing a legislative revenue forecast have a legislative forecasting unit, while two states contract with outside groups to prepare the forecast. Most states with a legislative revenue forecasting office have a staff of one to three people.

The former executive director advised the commission that "if it ain't broke, don't fix it." Virginia's revenue forecasting process is as impressive as any in the nation. He advised that the best forecasting systems, regardless of which branch administers them, are ones open to public review, which "de-politicizes" the numbers. While revenue forecasting is comparatively easy when the economy is growing steadily, it is difficult to predict an economy's turning points.

To the extent that problems with the legislature's reliance on an executive revenue forecast are attributable to the timing of the release of revenue information, cooperation and communication are critical. In the 1997 Session, updated revenue forecasts based on receipts in January were provided to the General Assembly before the end of that month. Problems arising from delayed receipt of revenue data may be more acute in years with a short legislative session.

Long Range Expenditure Forecasting

Only two states—North Carolina and Florida—are conducting long range expenditure forecasting for all agencies. In these states it is not apparent that expenditure forecasting is helpful in making policy decisions on a long-range basis. When expenditure figures are projected on a long term, figures often show expenditures exceeding revenues, and in North Carolina, long-range expenditure forecasts have made the legislature reluctant to reduce revenue sources.

The appeal of long-range forecasting is knowing what it will cost to keep doing what a state is currently doing. However, it is not apparent that states know how to do such forecasting very well. In North Carolina, computer models have been developed with consulting firms to provide projections that are more reliable than simply extrapolating current expenditures based on assumptions about inflation and population changes.

Most states perform some form of expenditure forecasting for major budget drivers such as transportation. However, states tend to disclose neither how the projections are done nor the results. Preparing accurate long-range expenditure forecasts requires knowing the future, and modeling these forecasts is difficult. Most expenditure forecasting assumes changes in laws and programs over time. Consequently, the rules reflected in the modeling involve policy decisions. Because a legislature or executive cannot bind future ones, assumptions that extend beyond the current period are often questioned.

Multi-year forecasting requires discipline. Citing the back-loading or ignoring of certain program costs in its long range expenditure forecasts, one speaker accused the federal government of fostering skepticism in such forecasts. Rather than trying to model every expenditure, states should focus on major drivers and publicly disclose the results.

Legislative Impact Statements

In response to the commission's charge to examine the mechanisms to evaluate legislation having an effect on the budget and expenditure projections, Department of Planning and Budget (DPB) staff summarized the fiscal impact statement process. In 1997, impact statements were prepared for 1,144 bills, or 60 percent of the bills introduced. Of the 776 bills for which no impact statement was prepared, 211 were not reviewed. Agencies preparing the greatest number of statements were DPB, the Department of Taxation, and the State Corporation Commission.

Most impact statements are prepared pursuant to executive order and are not required by statute. DPB's goal is to prepare statements that are objective, concise, broad in scope, and timely. Agencies attempt to complete statements within three days of receipt if the bill is assigned to a money committee, and within four days if assigned to other committees.

DPB measures its performance by the number of complaints received. For the past couple of years, the agency has not received any complaints. To gauge satisfaction with the impact statement process, the commission authorized staff to survey the members of the General Assembly.