Odhad nákladové funkce podniku a analýza fixních a variabilních nákladů

Abstract

This bachelor thesis addresses the estimation of cost functions and the decomposition of total costs into fixed and variable components for a selected company in the automotive industry. The primary objective is to apply selected mathematical and statistical methods to empirical corporate data from 2011–2025 and to assess their ability to capture the relationship between production volume and total costs amidst the dynamic transformation of the sector. The theoretical framework defines cost concepts, costing systems, and methods for estimating cost behavior. The practical section utilizes the high-low method, linear regression analysis, and second-degree polynomial regression, with the evolution of cost parameters tracked through the rolling windows technique. Model accuracy is validated using MAE, RMSE, and MAPE statistical metrics. The analysis demonstrates that the polynomial model is the most suitable method in terms of economic interpretability, as it reflects the progressive cost growth associated with high investment intensity. A significant finding is the identification of a "cost paradox" in 2024 and 2025, where strategic investments in electromobility and portfolio renewal led to an increase in total costs despite a decline in production volume. The thesis confirms that during periods of profound structural change, traditional volume-based models lose their predictive reliability and must be supplemented by qualitative managerial analysis.