A Comparative Analysis of Budget Forecasting Methods: A Systematic Literature Review Covering the 1983–2024 Period

ABSTRACT This study systematically analyzes 69 peer‐reviewed works comparing budget forecasting methods. It explores methodological evolution, geographic distribution, and evaluation trends. Four phases of development are identified: statistical methods, diversification, machine learning, and deep learning. A division emerges between traditional and next‐generation techniques. Geographically, 43% of studies focus on the United States, while developing economies remain underrepresented. In evaluation, MAPE, RMSE, and MAE dominate, with directional errors largely neglected. Findings show that optimal method choice depends on context, supporting a pluralistic, context‐sensitive approach rather than universal reliance on a single forecasting method.