Long Memory in Exchange Rate Volatility
Considering rapidly evolving technology and effective markets, wherein information and news are quickly and effectively reflected in financial asset prices, the positions of investors trading in financial markets regarding financial asset prices vary according to the continuous stream of information coming to the market. However, markets are not fully efficient in terms of maintaining a long memory that enables future pricing estimates based on the past market price of the financial asset. Revealing the existence of a long memory structure is essential to the development of monetary policies since exchange rates that tend to return to average exert high resistance. In this study, the exchange rate’s long-range dependence is determined in the scope of the log-periodogram estimator and using a fractional model structure, the average model, and the variance model structure related to the exchange rate between February 22, 2001–March 16, 2020 are examined. In this context, the parameters in the model allow an examination of the long memory process. According to the fractionally integrated exponential generalized autoregressive conditional heteroskedasticity model, it is determined that the effects of shocks in the exchange rate market continue and persist for a long period. Policy suggestions within the scope of exchange rates are evaluated within model outputs.