Statistical Modeling of the Impact of Program Budget Expenditures on the Economic Dynamics of the Kaliningrad Region

Issue № 1, 2026, article № 14, pages 145-155

Section: Competition of young authors

Language: Russian

Original language title: СТАТИСТИЧЕСКОЕ МОДЕЛИРОВАНИЕ ВЛИЯНИЯ ПРОГРАММНЫХ БЮДЖЕТНЫХ РАСХОДОВ НА ЭКОНОМИЧЕСКУЮ ДИНАМИКУ КАЛИНИНГРАДСКОЙ ОБЛАСТИ

Keywords: PROGRAM-TARGETED FINANCING, REGRESSION ANALYSIS, CORRELATION ANALYSIS, BUDGET, GROSS REGIONAL PRODUCT, KALININGRAD REGION, EFFICIENCY OF BUDGET, MACHINE LEARNING

Abstract: The article presents a study of the effectiveness of program-based budget financing at the regional level using correlation and regression analysis methods. Based on data from the Kaliningrad region for 2014–2023, the impact of six key budget spending areas on Gross Regional Product (GRP) dynamics is assessed. The analysis incorporates both classical statistical methods and modern machine learning algorithms, including polynomial ridge regression, Random Forest, Support Vector Regression (SVR), and k-Nearest Neighbors (KNN). In addition to analyzing the relationship between funding volumes and GRP, the study includes an in-depth analysis of data properties, including testing the normality of distributions, identifying multicollinearity, and assessing the robustness of models to sample changes. This approach not only allows determining the impact of individual programs on regional economic performance but also improves the accuracy of predictive models; by complementing traditional expert assessments, it allows detailing the understanding of the multiplier effects of budget investments and optimizing resource allocation based on their nonlinear contribution to the economy. Additionally, a comparison of the corrective capabilities of various machine learning methods was conducted, enabling the identification of algorithms that most accurately describe the nonlinear structure of budget funding data. The study identifies priority areas of funding that have the greatest influence on the region’s macroeconomic indicators. Special attention is given to the agro-industrial complex, which demonstrated the lowest performance relative to other programs. The results obtained can support decision-making and contribute to improving program-based budgeting mechanisms at the regional level in the Russian Federation.

Authors: Mnatsakanian David Albertovich