Democritus University of Thrace (DUTH)
Πρόβλεψη αποτυχίας επιχειρήσεων με την χρήση μη παραμετρικής μεθόδου (modified super-SBM-DEA) σε συνδυασμό με μεθοδολογία επιλογής των χρησιμοποιούμενων λογιστικών μεταβλητών (LASSO)
Abstract
dc:descriptionOver the last six decades, corporate failure prediction has been a major topic in the fields of accounting and finance. This dissertation aims to develop a corporate failure prediction model in combination with the application of an appropriate empirical methodology for the selection of the variables. In addition, emphasis is also given to the study of the theoretical framework of corporate failure, the overview of existing prediction models in the relevant literature, the basic categories of explanatory variables that have been used, as well as the methodologies for their selection. In the frame of the empirical analysis, it was deemed necessary to create three separate samples. The first one consists of 270 listed U.S. firms, 135 failed and 135 non-failed, operating mainly in commercial and industrial sectors. The second sample consists of 90 listed U.S. firms, 45 failed and 45 non-failed, from the energy sector. Finally, the third sample also includes listed U.S. firms from the energy sector, however, the number of failed and non-failed firms is not equal. Specifically, it includes 149 non-failed firms, while the number of failed ones varies from year to year, with each firm being excluded from the sample during the year of its failure. For the examination of the research hypotheses, both econometric models and methodological tools from the field of operational research were utilized. In more details, the Least Absolute Shrinkage and Selection Operator (LASSO) has been selected as a method for corporate failure prediction variables selection. Data Envelopment Analysis (DEA) has been selected as the primary corporate failure prediction model and more specifically, an advanced version of DEA, the modified super-SBM-DEA model that is used for the first time in the establishment of a corporate failure prediction model. Logistic regression has been selected as an alternative corporate failure prediction methodology for the performance assessment of the modified super-SBM-DEA model as a corporate failure prediction tool. For the dynamic assessment of failure, Malmquist DEA has been implemented. According to the findings, the proposed methodology that is based on the combination of LASSO and DEA offers superior results in terms of corporate failure prediction accuracy and is a valuable tool that contributes to the decision-making process.
Degree
thesis:*- Grantor dc:publisher
- Democritus University of Thrace (DUTH)
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Katsimardou, Sofia
- Κατσιμάρδου, Σοφία
Subjects
dc:subject × 18- Πρόβλεψη αποτυχίας επιχειρήσεων
- Αριθμοδείκτες
- Logistic LASSO
- Modified super-SBM-DEA
- Λογαριθµικό υπόδειγµα πιθανότητας
- Malmquist DEA
- Corporate failure prediction
- Financial ratios
- Modified super- SBM-DEA
- Logistic regression
- Οικονομικά και Επιχειρήσεις
- Κοινωνικές Επιστήμες
- Λογιστική
- Διοικητική Επιστήμη και Επιχειρησιακή Έρευνα
- Economics and Business
- Social Sciences
- Accounting
- Management Science and Operations Research
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/59625