The budget planning methods of the regional budget revenues based on econometric forecasting models

The paper deals with one of the budget planning methods of the regional budget revenues based on econometric forecasting models. The relevance of the topic is conditioned by the fact that the planning of regional budget revenues for upcoming fiscal year and the planning period is aimed at the regional sustainable economic development, the rational use of resources and the improvement of living standards. The research objectives are to analyze the dynamics of an individual region’s major budget items, to identify trends in changing the structure of revenues, to build trend forecasting models and to justify their use in the short-term forecasting of the budget revenue items. It’s a well-known fact that various methods of the budget planning and forecasting are used in economic practice. Five econometric forecasting models were used: a linear trend time series model, Brown’s adaptive model, an autoregressive model, an exponential smoothing method, and a method applied in Excel. The structural and dynamic analysis of the main items of the Lipetsk region’s budget revenue part for the period from 2017 to 2023 was carried out. The special emphasis is put on assessing the results deviation of the forecast values by means of the research methods mentioned above in line with the actual indicators of the budget revenues. The possibility of using forecasting models for practical use in the formation of the revenue side of regional budgets is examined by the authors. The structural and dynamic analysis of the main items of the revenue part of the Lipetsk region’s budget showed that almost all income items and total incomes possessed positive trends during the period under review. Some anomalous observations caused by the consequences of the COVID-19 pandemic in 2021 were revealed, the compensation was made by means of the interbudget transfers increase to the constituent subjects’ budgets. These anomalous observations were eliminated in the calculations by replacing them with the arithmetic mean of the neighboring levels of the series (2020 and 2022). The calculations’ results showed that among the models under consideration the linear model has the smallest values of the average approximation error within the time series with an error of 4.4%, whilst Brown’s model possesses an error of 5.4%. Still comparing the forecast deviations presented by the Lipetsk region’s Finance Department, the smallest deviation for 2024 and 2025 and a satisfactory degree of accuracy were shown by the exponential smoothing model (2024 - 6.5%, 2025 - 0.4%). Consequently, the Finance Department of the Lipetsk region can take into account the forecasts’ results based on the econometric models presented above in the budget drafting and planning its revenue, mainly the exponential smoothing model for a two-year planning and the method used in the Forecast Sheet add-on in Excel for a yearly plan.