Estimation the exchang rate with ARCH-GARCH models in the forex market

Foreign exchange market or FOREX is the biggest and the most liquid financial market in the world. According to the latest report of BIS (Bank for International Settlements) announced in April 2013, the daily trading size in the FOREX is about 5.3 trillion dollars. The dominating volatility and fluctuation in this market made it a complicated business for policy makers and consequently caused investors and speculators to be very cautious in their trading strategies. In financial series, instead of using linear time series because of its features, using nonlinear conditional variance models are becoming widespread because of its characteristics. With this, the aim of this study is to determine the best model within the exchange rate estimation modelling and alternative models in FOREX marketplace with using the non-linear conditioned variables of variance ARCH and GARCH modelings. In the study between January 2, 2013, and December 27, 2014 using USD / TRY selling prices per day, primarily exchange rate was investigated by the ADF test series is not stationary and is found to be so stable that the inclusion of a unit root. Then, the best mean equation model inside the alternative autoregresive models has been chosen as the ARIMA (3,3) model. In addition, it has been seen that the residual of mean equation model has the effect of ARCH, which has been determined by making ARCH LM test. Because of the serie has an ARCH effect, conditional variable variance models has been predicted with ARCH(4,0), GARCH(2,1) and the best forecast model between them has been appointed as GARCH (2,1). Keywords: Forex Market, Foreign Exchange Rates Forecast, ARIMA, ARCH,GARCH