Bank loans are the most important component of bank’ assets and income. Therefore, deterioration of bank’ loan quality, measured by the increase of non-performing exposures, impacts banks performance. Level of non-performing loans can be impacted by a macroeconomic factors, such as a change of gross domestic product, unemployment or inflation rate, but also by a bank specific factors, such as a level of profitability, amount of approved loans or capital level. After the financial crisis of 2008/2009, European union was exposed to the severe increase of the banks non-performing loan exposures. Certain countries, which were exposed to the high increase of non-performing loans ratio, were successful in quick decrease of these ratios, while certain countries were exposed to the high non-performing loans in the long run. Taking into account importance of banks in the financial system and loans as an external financing source in the European union, deterioration of banking assets quality impacted financial stability and economic recovery. This doctoral thesis researches determinants of the non-performing loans and economic impact of non-performing loans on financial stability in the European union. Sample consisted of all 27 member states for the period 2002 to 2019 based on aggregated annual data. To test determinants of non-performing loans, both macroeconomic and bank specific factors were modelled by the dynamic panel model, specifically generalized method of moments. Research results indicate that non-performing loans ratio is influenced by macroeconomic and bank specific variables, where increase of the GDP growth ratio and return on assets decreases non-performing loan ratio, while the increase of the inflation rate increases level of such loans. Higher bank cost to income ratio decreases non-performing loans, indicating that higher efficiency of the banking system does not decrease non-performing loans ratio. Economic impact of non-performing loans on financial stability was tested by the impulse response function and variance decomposition, methods of the panel vector autoregression. Variables which were used as a proxy of the financial stability include domestic credit to the private sector as s percentage of GDP, cost to income ratio, return on assets, GDP growth rate and inflation rate. To test the difference between member states with higher and lower non-performing loan ratio, countries were divided into two groups, with the threshold ratio of 5%. Analysis proved that increase of the non-performing loan ratio had a severe negative impact on the financial stability in countries with higher level of such loans. Research implications could be used by bank regulators and management. Rate of the inflation should be taken as the early warning predictor of non-performing loans, while increase of the bank cost to income ratio contributes to the decrease of non-performing ratio. Banks should therefore be very cautious in the period of crisis in reducing quality or number of employees in the loan approvement procedure. Taking into account stronger impact of non-performing loans on the financial stability in countries with higher non-performing loan ratio, which mostly did not introduce the countercyclical capital buffer, different approach to the calculation of this capital should be taken into account, as well as possible introduction of special loan countercyclical reservations. In the case of strong banking assets quality deterioration, regulators should impose quick and resolute measures, directed mostly to the removal (extractions) of non-performing loans from banks’ balance sheet and change of insolvency regulation. Future researches should focus on variables specific to the each individual loan, such as an interest rate, loan duration and collateral, and research determinants of different loan types (loans to the households, business loans,...), while the impact of the non-performing loans should be tested on individual banks’ level of efficiency and performance.