Predikce příjmů obecních rozpočtů

The subject of my bachelor thesis is the elaboration of a case study that deals with the preparation of the revenue side of the budget of the municipality Hodslavice. The main aim of my work is to determine the accuracy of the actual estimates of the municipality and to compare the accuracy of predictions of several statistical methods. I use these statistical methods: simple moving average, exponential smoothing, transformation moving average, regression against time with an exponential trend. Drawn from the analysis, I found that for the prediction of each type of revenue, there is a more accurate method than expert's assessment of the budget administrator in Hodslavice, but there can not be applies a single method of prediction for all types of revenues. The most accurate method for predicting taxes of individual revenues is exponential smoothing. However the differences between the predictions by this method and the expert's estimation of the administrator of the budget are minimal. Regression against time was the best in forecasting tax corporation income and value added tax. For real estate tax and non-tax revenue, which the municipality has at least some competence in influencing the future level of revenue appears to be the most accurate method of simple moving averages.