Predicting Business Continuity in Italian Luxury Firms: A Dual Logistic Regression Approach

Abstract

Business continuity prediction is a topic of great relevance, as it affects the economic well-being of all countries and the stability of production systems. In this context, the study develops two logistic regression models to predict business continuity in the Italian luxury sector. The analysis is conducted on a sample of 114 companies operating in the luxury sub-sectors of fashion, jewellery and watchmaking, and automotive, observed over the 2021–2023 period (342 observations). The dependent variable is business continuity, while the independent variables include six financial indicators: current ratio, current assets/total assets, equity ratio, natural log of total assets, return on assets, and retained earnings/total assets. The two logistic regression models differ by validation criterion: the first employs a time-based split (training set covering 2021–2022 and test set for 2023), while the second applies a firm-based split (training set with 84 companies and test set with the remaining 30, randomly selected). Both models exhibit high predictive performance, with overall classification accuracies of 96.49% and 94.44%, respectively. The results confirm the effectiveness and robustness of logistic regression in predicting business continuity in the Italian luxury sector, a field characterized by a distinctive combination of resilience and vulnerability.