Purpose of the study. The purpose was to determine the applicability of time-series models to revenue forecasting in Florida cities. / Procedure. Revenue data for were collected for five fiscal years (60 months) in twelve cities. Participating budget officials were interviewed to obtain information regarding forecasting expertise. / Data for Fiscal Years 1982 through 1985 were fit with seven time-series models ranging in complexity from the moving average to Box-Jenkins, in order to obtain forecasts for Fiscal Year 1986. / Mean Absolute Percentage Error and Mean Percentage Error of the techniques tested were compared using one and two way analysis of variance and nonparametric analogs. / Findings. At the annual horizon, the judgmental-naive forecasts of the local officials were generally more accurate than those arrived at with the models utilized in the study. Consistent with findings in the 1982 M-Competition of time-series models, no one model consistently outperformed any other in terms of accuracy or bias. However, curve fitting was significantly less accurate than the other models tested. / Interviews indicated a paucity of knowledge of forecasting methods on the part of local officials. Revenue forecasts were made primarily on an annual basis. Local officials engaged in intentional underforecasting of revenue to protect themselves against potential revenue shortfalls. / Conclusions. Data analysis and interviews produced the following conclusions: (1) At the annual forecast horizon, the currently utilized naive-judgmental forecasting methods generally outperformed the methods tested in terms of bias and accuracy. (2) Aggregation of sources improved the forecast accuracy of the models tested. (3) Large general fund balances in the communities studied minimized the need for improved forecast accuracy. (4) Lack of methodological expertise on the part of local officials may preclude implementation of the models tested. / Source: Dissertation Abstracts International, Volume: 49-12, Section: A, page: 3865. / Major Professor: Gloria A. Grizzle. / Thesis (Ph.D.)--The Florida State University, 1988.
| Identifer | oai:union.ndltd.org:fsu.edu/oai:fsu.digital.flvc.org:fsu_77879 |
| Contributors | Frank, Howard Alan., Florida State University |
| Source Sets | Florida State University |
| Language | English |
| Detected Language | English |
| Type | Text |
| Format | 343 p. |
| Rights | On campus use only. |
| Relation | Dissertation Abstracts International |
Page generated in 0.0016 seconds