PREDICTING CURRENCYCRISES WITH SIGNALAPPROACH: THE CASE OFTURKEY

Currency crises creates extremely high costs in economies. The depletion of foreign exchange reserves, a severe recession and negative GDPgrowth rates are observe in the countries where the currency crises are occurred. There seems to be great benefits from identifying reliable early warning indicators helping to predict currency crisis. Thus, policy makers would enable to forestall impending pressure on their countries currency system by using early warning indicators. The “signal approach†to proposed by Kaminsky ve Reinhart (1998) to predict currency crisis is commonly employed in empirical studies as an early warning system. It involves monitoring a number of variables that tend to behave differently or to exhibit an unusual behavior in the periods preceding a crisis. Acertain threshold is determined for each indicator, which, when exceeded, issues a warning signal that a crisis may occur within a period of time. This study aims to predict the 1994 and 2001 currency crises in Turkey by using signal approach. Empirical analysis reveals that almost 67 percent of the indicators examined signal the occurrence of currency crises. It is concluded that 9 of 15 indicators are determined as leading indicators and are found to be significant in signal approach based on their noise to signal ratio and six of indicators lacked predictive power from list of potential indicators.