Revenue forecasting at the local government level always has been of critical importance in annual budget development and execution. The projections that underlie a budget have a significant impact not only on how the locality allocates limited resources and responds to service-level demands by residents but also on how the locality plans for the future.
The variety of forecasting methods typically used by local governments--judgmental methods, trend analysis, deterministic approaches and econometric modeling--have varying degrees of precision, depending on the level and accuracy of the components. Judgmental methods have the benefit of being easy to use; these projections, however, relying on the use of past information and knowledge of each particular locality, can be jeopardized when the in-house "expert" leaves the locality or the information is used incorrectly. Trend analysis, also a fairly simple projection method based on the assumption that past revenue trends will predict the future, can result in significant variation from actual performance of major indicators when external factors change, as in a recession or during a major demographic shift.
The deterministic approach involves the use of one independent variable to determine the projection of a revenue. While this method incorporates outside factors into the process and acknowledges their effect on the projection, problems occur when the relationship of the one variable to the revenue changes. Econometric modeling has the advantage of using multiple variables to predict future revenue streams, but it also requires the ability to determine which variables accurately represent a relationship to a particular revenue and to predict the future course of these variables. In addition, econometric modeling calls for advanced technical expertise, extensive data collection, and significant software and hardware acquisitions.
The use of these projection processes, in many cases, has failed to predict key revenue streams with any great accuracy. The need for enhanced projection methods has become particularly important, especially in light of the recent recession and expected continued slow growth in the immediate future.
An additional dimension to the revenue projection process comes from the desire of elected officials to respond favorably to the needs of their constituents. The result can be a forecasting process that becomes politically charged and that often results in revenue projections that are unrealistic. Senior-level project managers, aware of these conflicts, tend to view revenue forecasts prepared by finance staff with suspicion and fail to develop a sense of confidence in the projections. This environment, coupled with a local economy in recession, produces an even more complex revenue projection process.
A Solution Explored
With similar problems in revenue forecasting prevailing in Chesterfield County, Virginia, county officials realized that different methods and innovative approaches were critical. They sought ways to augment and enhance their revenue projections and at the same time provide a degree of confidence regarding staff revenue projections for senior managers and board members.
Chesterfield County is a suburban locality with a population of 220,000 located just south of Richmond, the state capital. During the 1980s it experienced rapid growth in revenues; in the early 1990s, however, these trends changed, and the county began to experience significant shifts in three of its largest revenue sources--real estate assessments, personal property tax and local sales tax--as shown in Exhibit 1. The percentage increase in real estate assessments fell for tax year 1992 to 4.1 percent from an annual average of 12 percent since 1984, and it was anticipated that assessed values would decline further in future years and not keep pace with inflation. Personal property revenue increases in FY91 were projected to be consistent with the . …