Modelling foreign exchange volatility

As exchange rates themselves and their volatilities indicate to economical stability of a country, implemented exchange rate regimes are extremely important for economies. In Turkish case, under the floating exchange rate regime, implemented since 2001, Central Bank of Turkey intervenes foreign exchange markets only in the case of excess volatility. The main reason for this situation is that volatility is considered to be a risk and threat against the economic stability. On the other hand, Central Bank of Turkey does not have a statement describing what level should be considered as excess. In this context, the main aim of this study is to establish a model for exchange rate volatility. Recently volatility models have become one of the most frequently discussed issues. This is mainly due to correct the false assumption of constant volatility which commonly used in previous studies although it is usually not. Among all volatility models, ARCH models are the mostly used ones that can model the variance with the maximum likelihood method and analyse parameters with capturing the changing variance. In this study, volatility of USD sell rate is modelled with the help of highly correlated and lagged currency data. Moreover, while reaching to the best fitting final model (Autoregressive Conditional Heteroscedasticty models), this study tests also many other models including the classical ones assuming constant volatility. After capturing it, study ends with the evaluation of the Central Bank interventions which is performed in the light of the finalized model. Keywords: 1. Foreign Exchange 2. Volatility 3. ARIMA Models 4. ARCH Models 5. Central Bank Interventions