Özet:
There is a vast literature on financial stability issue after the onset of the recent global financial crises. Stress-testing, which can be defined as the techniques used by policy makers to assess the vulnerability of the financial system to extreme events is one of the key issues in financial stability literature. The basic idea behind stress-testing is finding a risk proxy for financial stability and performing impulse response and sensitivity analysis on this proxy. There are different risk measures used in the literature such as; credit risk, market risk (FX risk etc.), liquidity risk, and contagion risk. In this study, we choose to analyze credit risk and use Non-performing loan (NPL) ratio as a proxy for the credit risk. The reason we focus on the credit risk is the financial structure of the Turkish economy. The main players in giving loans are conventional banks. Thus we think that NPL ratio will be a good proxy for assessing banks fragility and consequently financial sector stability. There is no published study on macro stress testing for Turkish case. Therefore, this study as a first attempt for macro-stress testing applied on Turkish economy opens a window for further studies. Also this thesis is one of the first studies, revealing the determinant of Non-Performing Loans after the 2000-2001 period. Using a 7-lag VAR model, we performed dynamic out of forecasts which reveals quite accurate results compared with the actual data. Since forecasting is a very crucial tool for both policy makers and market players, these results are one of the main strengths and contributions of this study, Also policy recommendations are provided in accordance with the robust findings. Firstly, industry production used as a proxy for GDP has a significant effect on NPL ratio, which can give prediction insights to the policy makers depending on the real state of the economy. Secondly, capacity usage has a significant positive effect on NPL ratio which conflicts with the prior expectations. Therefore policy maker should be concerned about the discrepancy between real stimulus in the economy and people's expectations. Lastly, spread used as a proxy for the expectation of banks reveals quite useful insights about the expectations o the market.
