Izvorni znanstveni članak
https://doi.org/10.17535/crorr.2026.0020
Country-specific financial distress prediction using decision trees: Empirical evidence from Slovakia and Croatia
Dominika Gajdošíková
orcid.org/0000-0001-7705-3264
; University of Žilina, Faculty of Operation and Economics of Transport and Communications, Department of Economics, Žilina, Slovakia
*
Katarína Valášková
orcid.org/0000-0003-4223-7519
; University of Žilina, Faculty of Operation and Economics of Transport and Communications, Department of Economics, Žilina, Slovakia
Tomislava Pavić Kramarić
orcid.org/0000-0002-0974-4423
; University of Split, Department of Forensic Sciences, Split, Croatia
* Dopisni autor.
Sažetak
This paper develops and compares financial distress (FD) prediction models for Slovak and Croatian firms based on a decision tree (DT) approach. The study aims to identify the crucial financial variables that predict firm FD in these two post-transition economies that share historical and institutional similarities but differ in the structure of the economy. A database of Slovak and Croatian enterprises was analysed, focusing on firms being classified as prosperous or non-prosperous based on standardized accounting parameters from financial reports. DTs were employed since they can model complex, nonlinear relationships without diminishing interpretability. The results indicate that high levels of total indebtedness, equity leverage, and debt-to-equity ratios are cross-sectionally associated with an increased likelihood of FD. In the Slovak context, liquidity, particularly the current ratio, showed greater importance than profitability measures, whereas in Croatia profitability ratios, especially return on assets, gained relatively higher relevance compared to liquidity. The findings close a research gap concerning country-specific differences in Central and Eastern European (CEE) FD predictors. The paper provides theoretical contributions by refining knowledge on insolvency drivers in post-transition economies and practical contributions by understanding tailored early warning systems. The study is limited by sample restrictions and the omission of macroeconomic factors. Future research may incorporate wider contextual determinants and alternative machine learning (ML) approaches.
Ključne riječi
financial distress prediction; decision trees; financial distress; firm-level analysis; post- transition economies
Hrčak ID:
347146
URI
Datum izdavanja:
13.5.2026.
Posjeta: 0 *