Surviving the Pandemic: Financial Distress Prediction for Slovak SME Manufacturers

Abstract

Research background: In this era of economic uncertainty, small and medium-sized manufacturing enterprises (SMEs) face an increasing risk of financial distress, often with devastating consequences for employment, supply chains, and overall economic stability. The ability to predict corporate financial distress is crucial for financial stability, risk management, and economic resilience, particularly for small and medium-sized enterprises, which are highly vulnerable to financial shocks. Purpose of the article: The purpose of this study is to develop a financial distress prediction model for Slovak SME manufacturing firms by identifying key financial indicators that distinguish bankrupt from non-bankrupt companies over a three-year period before failure. The study also examines how these indicators evolve as financial distress approaches and evaluates the impact of external shocks, particularly the COVID-19 pandemic, on their predictive power. The goal is to enhance early detection of financial instability, supporting more effective risk management in the manufacturing sector. Methods: To develop a robust financial distress prediction model, we employ a three-step methodology. First, we address data imbalance using the SMOTE, which generates synthetic samples to balance the minority class and improve model performance. Next, we implement a Genetic Algorithm for variable selection, optimizing the choice of predictors by minimizing the Schwarz Information Criterion. Finally, we use Logistic Regression to model financial distress, ensuring interpretability and statistical rigor. Findings & Value added: While EBIT/Total Assets, Cash Flow, and Cash Flow/Total Assets consistently reduce financial distress risk, liquidity measures such as Cash and Cash Equivalents/Current Liabilities shift in relevance—acting as protective factors closer to failure but signaling inefficiencies further from financial distress. The model demonstrates strong predictive performance, maintaining high accuracy, recall, and AUC even when tested on real-world unbalanced data, confirming its practical applicability. These findings emphasize the need for adaptive financial distress prediction models that reflect the dynamic nature of financial distress and provide more effective early warning systems in an era of heightened economic uncertainty.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.