Constructing a financial stress index for Turkey: A multivariate GARCH approach

Various economic crises have been experienced due to the dynamics of the country as well as the global crises that have lived in the world from past to present. Many of the important studies in economics aimed to estimate these fluctuations in the markets and to examine their relations with each other. In this context, since the early 2000s, financial stress indexes have been established by economists or central banks for various countries. The financial stress indexes created are specific to countries or regions and have some differences. These differences are the methods as well as the data types that are used. In this study, the financial stress indexes created for Turkey. Also, different countries are examined in the literature review. There is the necessity of constructing a new financial stress index because of the economic fluctuations in Turkey in recent years. The financial stress index of Turkey demonstrates the economic activity after important financial events that are called as financial crisis. The financial stress index which is a continuous variable as a time serie and its extreme values are called financial crises in the literature. The financial stress index can be expected to increase in the case of expected financial loss, risk potential or uncertainty financial conditions. In this study, the financial stress index of Turkey is generated with using the banking sector, equity market, money market and exchange market variables. The financial stress index of Turkey is constructed on daily basis for the period from 2 January 2007 to 29 November 2017. The data diversity has been expanded in this study when compared to previous studies in Turkey that is why this study has importance for Turkey. Firstly, the variables of each sector were standardized. Then, the financial stress index for each sector was composed with the equal weighted average method. FSI of Turkey was created again using the four-sector financial stress index data with the same method. The financial stress index of Turkey is analyzed with banking sector, equity market, money market and exchange market indexes. Modelling volatility is very important for investors and researchers The Autoregressive conditional heteroscedasticity models (ARCH) are used especially to model volatility clustering. In the market, there are periods of small volatility followed by the periods of high volatility. This phenomenon is called volatility clustering. For volatility clustering, historical volatility models can be used mostly. Most popular model is generalized autoregressive conditional heteroscedasticity (GARCH) model which is generalization of ARCH model. In this thesis, multivariate GARCH model was used because of the presence of more than one variable when analyzing correlations. A special type of multivariate GARCH model which is dynamic conditional correlation GARCH (DCC GARCH) model was used in this study. DCC GARCH model can be used to estimate conditional correlations between two series. DCC GARCH model was used to examine the correlation between FSI of Turkey and the markets in Turkey. Conditional correlations give information about volatility spillover. This phenomenon is very important for financial markets. What is observed in financial market is that correlation increase or decrease after financial crisis between two financial markets. If there is a volatility shock in one market, then it affects the other market. If the correlations between two markets are high, it means that there is less risk diversification opportunity between these markets. In other words, hedging between these markets seems not possible. On the contrary, if the correlations between two-time series of financial markets are very low, in that case, it is possible to hedge this portfolio because of the potential of diversification risk. After generating the FSI of Turkey for the last ten years, the peaks of it are interpreted in terms of important financial events in this study. In 2008, the giant Lehman Brothers declared its bankruptcy and the whole world paid the price of Lehman's collapse. Billions of dollars had evaporated in stock markets, millions of people were unemployed, many employers had gone bankrupt etc. The Turkish economy was also negatively affected by this fact. As a result of that, FSI of Turkey which is created in this thesis seems to have a peak in 2009. In 2013, the Federal Reserve System (FED) made a declaration about its tapering decision which is to reduce asset purchases in the beginning of 2014. It means that United States withdraw money from the market and the interest rates will rise for United States. This decision also implied the depreciation of the Turkish Lira because investors wanted to invest in the US. As a consequence of this event, there is another peak in FSI of Turkey for this time interval. At the end of 2016, after the US presidential election, the perception of uncertainty in international finance world had increased significantly. After election, FED's monetary policy was expected to change with new administration in US. Moreover, the slowdown in global economies and the geopolitical risks had the negative effect on economic activity over foreign trade. In that time, the United Kingdom's decision on leaving the European Union provided also negative effect on global market. On the other hand, the main reason of activity in foreign exchange market is that the uncertainties for future interest rate policies of developed countries. The coup attempt had especially big impact on Turkish economy. Thus, there is another peak in FSI of Turkey at the beginning of 2017. Moreover, the relationships between the markets which are used to generate the FSI of Turkey are analyzed according to the correlation coefficients gained from DCC GARCH model. At the last part, the financial stress index of Turkey is interpreted whether it can be used as a sign of financial crisis for Turkey or not. Its extreme values show the financial crisis that the events that suddenly and unexpectedly occur in the economy will produce macroeconomic consequences for the country's economy. Furthermore, firms have serious consequences in that time from the micro perspective.