Finansal krizlerin tahminlenmesi: Sinyal yaklaşımı ile Türkiye uygulaması

In this study, the crises in the Turkish economy are analyzed using the KLR signal approach between 2002-2018. The primary purpose of the study is to determine macroeconomic and financial indicators, predicting any possible economic crisis for the Turkish economy. Besides the KLR signal approach, a variation of this model, the 1.5 standard deviation threshold value approach, is applied, and the empirical results are reported. For the Turkish economy, four crisis dates between 2002 and 2018 are determined by exploiting a financial pressure index; then, abnormal movements of indicators before these crisis dates are monitored. It is observed that real sector indicators react much later than financial indicators and react during or after the crisis point(s). In this respect, real sector indicators are evaluated as worst-performed indicators. These are the industrial production index, unemployment rate, the ratio of exports to imports. The indicators signaling before all crisis periods and resulted in a low error signal ratio below (1) are considered the best estimators. The producer prices index, foreign exchange deposit accounts position, Republican Gold price, and the current account to foreign exchange reserves ratio are among the successful indicators. Nevertheless, the best three indicators with the lowest level of error signal rate and high signaling success are the CPI-based effective exchange rate, short-term external debt to foreign exchange reserve ratio, and domestic debt expansion. Lastly, according to the analysis results, it is observed that the KLR signal approach has a higher and more consistent prediction potential than the 1.5 standard deviation threshold value approach. v Keywords : signal methods, KLR signal approach, financial crises, economic crises, logit-probit