The Role of Total Assets in Financial Performance: Evaluating Econometrics and Machine Learning Approaches

ABSTRACT This study examines the financial performance of North American firms, accentuating the significant role of total assets in determining profitability, measured through EBITDA. By incorporating traditional econometric models—Fixed Effects and Random Effects—with machine learning techniques, exclusively Support Vector Machines (SVM), the research assesses both explanatory and predictive capabilities. The findings reveal that total assets significantly influence EBITDA, with the Fixed Effects model offering robust explanatory power (adjusted R 2 = 77.86%) and SVM exhibiting superior predictive accuracy (89.4%). Additionally, the nuanced effects of financial ratios, such as liquidity and leverage, emphasize their indirect contributions to financial performance. These results offer actionable insights for stakeholders to optimize asset utilization and financial strategies. This study's implications extend to advancing predictive methodologies in financial analysis, providing a scalable framework for analyzing profitability across diverse firms and contexts. This research specifies constructive insights for stakeholders, investors, and policymakers, enhancing their understanding of the financial dynamics of North American firms.