Análisis de datos aplicado al Presupuesto General del Estado para la optimización de la asignación de techos presupuestarios

Abstract

Abstract: This study develops a quantitative methodology for the technical determination of budget ceilings in Ecuador s General State Budget, integrating data analysis, predictive modeling, institutional clustering, and an optimization framework subject to the global fiscal limit. Using historical budget data for the period 2012 2024, the research identifies key stylized facts such as the rigidity of current expenditure, investment volatility, incrementalism, and institutional heterogeneity which shape budget behavior and justify the use of analytical tools to anticipate future execution. The results show that institutional execution can be predicted with high accuracy through supervised learning models, and that behavioral patterns allow institutions to be grouped into profiles with distinct levels of stability, volatility, and spending capacity. These findings provide the basis for an optimization model that assigns ceilings consistent with fiscal constraints and with each institution s actual ability to execute its budget. Rather than simply generating numerical outputs, the methodology offers a practical interpretation of the budget process, demonstrating how systematic use of data can strengthen fiscal discipline, reduce discretion, and enhance transparency in resource allocation. The study concludes that it is feasible to gradually incorporate an integrated system of prediction, classification, and optimization into the national budget process, and that this approach can serve as a valuable tool for the Ministry of Economy and Finance. Its main contribution lies in providing a technical, replicable, and results-oriented framework that complements political decision-making and supports more efficient and sustainable fiscal management.