SJR 350

Commission on the Commonwealth's Planning and Budget Process

August 25, 1997, Richmond

The Commonwealth's performance budgeting system is being touted as an improvement over state government's traditional incremental budgeting. The performance budgeting system and agency expenditure forecasting efforts were reviewed at the second meeting of the Commission on the Commonwealth's Planning and Budgeting Process.

Maximum Size of the Rainy Day Fund

The revenue stabilization (or "rainy day") fund is intended to ease budgetary stress caused by revenue shortfalls resulting from unforeseen economic downturns. Neither the Constitution nor the Virginia Code places a dollar cap on the fund's maximum size. The fund is expected to grow to $338.9 million by July 1999.

In response to concerns about the absence of a dollar limit on the size of the fund, a Joint Legislative Audit and Review Commission (JLARC) official recounted the factors that led the General Assembly to base the maximum fund size for each upcoming fiscal year on a formula. The formula bases the maximum fund size on 10 percent of the average corporate and individual income tax and retail sales tax revenues for the previous three years. Basing the maximum fund size on a formula allows the fund to grow with the state's economy. Tying the maximum fund size to a percentage of income and sales tax revenues intentionally provides an automatic basis for accruing a substantial fund.

Performance Budgeting in Virginia

According to the director of the Department of Planning and Budget (DPB), Virginia's performance budgeting system provides substantial benefits over the previous method of incremental budgeting. JLARC has noted that the program budgeting approach instituted at the recommendation of the Hopkins Commission had evolved by the early 1990s into incremental budgeting. Under incremental budgeting, agency allocations are determined based on the current appropriation, adjusted by funding for new program initiatives, and amended by requests for funds above the current base. Incremental budgeting has been criticized as lacking systematic means for reviewing programs, setting priorities, and focusing on long-term plans.

Performance budgeting attempts to address these deficiencies by linking long-range strategic planning, performance measurement, and budget development. The Commonwealth's performance budgeting system has been recognized by the National Performance Review as "best in class" for its strategic planning and performance measurement components.

Executive branch agencies are required to prepare six-year strategic plans as part of their budget submissions for the 1998-2000 biennium. In these plans, agencies are expected to reexamine programs, consider financial goals, reassess priorities, and base funding on program performance. This approach was contrasted with six-year expenditure projections that merely extrapolate current programs into the future, assume priorities do not change, and disregard program performance. The Commonwealth required agencies to prepare six-year expenditure estimates starting in 1976, but discontinued the requirement in 1978. These expenditure estimates were characterized as an academic exercise. The estimates were highly inaccurate and not widely used. However, long-term client projections were acknowledged to be useful in identifying budget policy options.

Since 1993, the Code of Virginia has required agencies to submit to DPB, with their biennial budget estimate, an estimate of the amount that will be needed for the two succeeding biennial periods. The DPB director acknowledged that agencies have had latitude in complying with this requirement and noted that no requests have been made for the information submitted by agencies. The five-year plans prepared by agencies in 1995 for the 1996-98 budget cycle were submitted as part of the Governor's confidential working papers. In contrast, the agency strategic plans being prepared for the 1998-2000 biennium will be available to the public.

Six recommendations were offered to the commission. The General Assembly should (i) review agency strategic plans as part of programmatic and budget deliberations; (ii) take advantage of the statewide strategic planning process to obtain meaningful analysis of policy options and scenarios; (iii) prescribe additional performance measures and investigate targets and results; (iv) expand cooperation with the executive branch in developing six-year capital outlay plans; (v) take advantage of existing opportunities to participate in the revenue forecasting process; and (vi) implement an automated activity-based accounting and budgeting system.

Forecasting Major Budget Drivers

The manager of DPB's Economic and Regulatory Analysis section reminded the commission that forecasting is the first step in planning. While short-term (biennial) forecasts have been very accurate, long-term forecasts face insurmountable uncertainty. Five- and six-year forecasts for revenues and for Medicaid, education, and prison populations have generally been off by between 7 and 8 percent. Investing in sound forecasting pays big dividends by improving the efficiency with which scarce resources are allocated.

DPB's forecasting responsibilities include coordinating expenditure forecasting, evaluating agency forecasts, assisting agencies in preparing forecasts, and forecasting budget drivers. Budget items forecast by DPB include TANF caseloads, Medicaid costs and enrollments, and correctional system populations. The Medicaid forecasting process, which required development of an econometric model, and the public safety forecasting process, which uses a consensus process, were praised by the Department of Planning and Budget.

Medicaid and Corrections

The Department of Medical Assistance Services prepares its own six-year Medicaid expenditure forecasts. After projecting baseline expenditures, the fiscal impact of proposed policy changes and new mandates are estimated. DMAS staff bases its primary forecasting process on exponential smoothing, a time series forecasting technique. Current forecasts extend through fiscal year 2010, though the accuracy of time series forecasts decreases as they are extended that far into the future. Alternative forecasts involving regression methods are used for several of the larger Medicaid categories, and the DMAS forecasts are compared with DPB's independent forecasts to arrive at a consensus forecast.

JLARC reviewed the Medicaid forecasting process in 1992 and 1997 and found that the forecasting modeling appears to be sound. JLARC found that Virginia's expenditure estimates do not appear to be less accurate than those of nearby states, states in the South, or across the nation.

Among the lessons relayed by DMAS to the commission were that forecasts require constant maintenance and updating, cannot be produced in isolation from program operations, and benefit from the use of multiple forecasting techniques. The 1997 expenditure forecast was cited as an example that short-term forecasts are more accurate than long-term forecasts. The July 1993 forecast exceeded the actual 1997 sum by 9.24 percent. The variance between the 1994 forecast and the actual number declined to 5.98 percent, and it fell to 1.79 percent for the 1995 forecast and to 0.62 percent for the 1996 forecast. While no one expects projections made six years in advance to be on target, long-range forecasting provides a measuring point that can be constantly adjusted based on changing circumstances and additional data. One measure of a long-range forecast's success is its usefulness as a starting point for planning purposes.

A representative of the Washington-based National Council on Crime and Delinquency (NCCD) described the consensus process for preparation of the corrections population forecasts. NCCD prepares prison population forecasts under contract with the Department of Corrections. The forecasting process uses a simulation model to prepare a baseline forecast based on one year of data. Under Virginia's consensus process for projecting inmate populations, the baseline projections undergo scrutiny by a technical review committee and a policy review committee. Using its simulation model software, NCCD can provide monthly forecasts for various categories of offenders.

Education and Transportation

K-12 spending is a major driver of increases in the Commonwealth's budget. As the average daily membership is used to distribute the majority of state funds to localities, much effort is invested in predicting this number. A representative of the Department of Education noted that instead of producing one statewide forecast, the department must prepare separate enrollment forecasts for each of the 137 school divisions in the Commonwealth for each year of the biennium.

Local information, in conjunction with the department's methodology, has produced highly accurate statewide projections in the short term. As with other expenditure-related projections presented to the commission, the accuracy increases as the forecasting horizon decreases because the later projections incorporate more actual data.

The State Council for Higher Education inVirginia (SCHEV) is responsible for reviewing and approving all enrollment projections proposed by public colleges and universities. Rather than generating a single statewide estimate, at least three projections are prepared for each institution. SCHEV's interim director observed that while the forecasting methodology has improved over the past two years, the process lacks a policy focus. An example of a policy issue not addressed in the forecasting process is whether enough Virginians are receiving tertiary education. While higher education can be characterized as a retail market driven by choice and demographic factors, high rates of participation are essential to the economic competitiveness of the Commonwealth and the nation.

The director identified serious problems obtaining basic data. Prior to 1995, demographic data was provided by the Weldon Cooper Center at the University of Virginia. Since the contract was terminated in 1995, SCHEV no longer has access to forecasts of the number of high school graduates beyond a three-year period. The importance of accurately forecasting enrollment growth is underscored when an institution's appropriation includes funding for projected changes.

Highway and other transportation expenditures comprise nearly one quarter of the Commonwealth's non-general fund operating budget. A Department of Transportation official noted that in addition to preparing annual and biennial budgets and a six-year financial plan, the department integrates agency budget requirements with state and federal legislative processes to provide annual updated financial plans. The 1996-98 biennial budget total allocation is $4.5 billion.

The six-year financial plan is based on the Department of Motor Vehicles' official state revenue estimate for Commonwealth Transportation Funds and utilizes current federal transportation funding estimates. The six-year financial plan is the basis for figures contained in annual updates to the six-year improvement program. The current forecast for 1997-2003 is $14 billion. Depending on the results of Congressional reauthorization of ISTEA, the Commonwealth could receive between $100 million and $300 million in additional federal funds.

Next Meeting

At its next meeting, the commission will review the efforts of other states to improve planning and budgeting processes. The commission will also review the experiences of those states that have established a legislative revenue forecasting capability.

Joseph V. Gartlan, Jr., Chairman
Legislative Services contact: Franklin D. Munyan