Revenue forecast of municipal budgets

The subject of my thesis is to analyze the predictions of tax revenues in ten municipalities of comparable size. The main aim of my thesis is to evaluate the accuracy of predictions for selected municipal tax revenues and see if you can not refine their expert estimation using appropriate statistical methods. A sub-goal is to characterize in detail the various components of the budget revenues, and analyze their size and structure in selected municipalities. Another important sub-goal is to compare the communities to highlight their differences and common elements of municipal budget process. These information are based on interviews at municipalities. The result of the analysis is that the optimal use of prediction methods differ not only income from income, but from city to city. For income tax paid by employees appeared in some cases, reliable prediction of the city, in other cities it was exponential smoothing. For the tax on personal income from independent activities is clearly the most accurate regression analysis, which refines the prognosis of this tax by up to several tens of percent. Although the error of prediction of the city in property taxes was none too small, this approach has remained the most accurate. Only three cities have been more accurate by using exponential smoothing.