What can be expected in credit-risk management from NPL in the western Balkans region in the future?

This paper presents an analysis of the impact of the rate of selected parameters on the banking system performance, specifically to non-performing loans (NPLs) movements. The goal is to investigate which the most influential factors affecting the movement of NPLs in the WB countries are, given that research have pointed out the impact of macroeconomic factors on the formation of NPL rates in banking systems. The authors have added several parameters to their methodology, dividing the indicators into internal and external. On the topic of indicators that affect the performance of the banking system, but also predictions of future trends of NPLs, several hypotheses can be set, but this research will start from the hypothesis that the NPL trend can be predicted by creating predictive models which, as the basis, have a combination of macroeconomic and banking system performance indicators. In addition to scientific literature, publications of international development and those by financial institutions were used as well, and the authors also accessed the international database - CEIC data. The time aspect of the research will cover the period 2010-2019, and the prediction of NPL trends will be performed for the period 2020-2025. To determine the final model and the indicator that will most accurately describe the target variable, the Merton's model in the statistical tool R will be developed and prediction tests will be fey performed. The most important statistical methods: linear regression, R2, ADJ R2, correlation matrix. The results show that in 3 out of 5 observed indicators, the one that most influences the trend of problem loans is the unemployment rate. Based on the modelling, the outputs indicate small deviations between the NPL obtained by the model and the publicly announced NPL-trends are very well presented, and the forecast results indicate a sharpening of the NPL trend curve in the period up to 2025. The contribution of this paper is reflected in the time prediction of NPL trends which can be useful to state authorities for adequate measure implementation.