Modelos de predicción de quiebra empresarial para las micro y pequeñas empresas del Ecuador (2018-2023)

Abstract

Abstract: This study evaluates the predictive capacity of the Altman Z2-Score, Springate, and CA-Score insolvency models in Ecuadorian micro and small enterprises (MSMEs) during the 2018 2023 period. Through a quantitative, descriptive, and comparative approach, using a pseudo-panel design based on 213,411 records from the Superintendency of Companies, the validity of these instruments was analyzed against the legal reality of the entities. The methodology included data engineering, winsorization, the removal of outliers, and the implementation of the McNemar Test to contrast binary classifications. Significant findings reveal a critical asymmetry among the models: while the Altman and Springate models exhibit structural optimism, the CA-Score displays superior sensitivity by detecting vulnerabilities overlooked by the other models. Therefore, it is concluded that the CA-Score is the most effective prudential supervision tool for the local environment, as its mathematical rigor prevents false diagnostics of financial health. This work provides substantial evidence for the design of early warning policies and strategic management in Ecuador.