Alternatif Modellerin Enflasyon Öngörü Performansı: Türkiye Örneği

Turkey has implemented its monetary and exchange rate policy based on pre-commitment which is pre-announced basket peg in the year 2000. This policy leads to provide the additional information to set up the forecast models for Turkish inflation. Inflation forecasts can be improved with the injection of a prior information into the structural (conditional) econometric models. This study focuses on the relation between inflation and exchange rate by comparing the forecasting performances of structural econometric and Vector Autoregressive (VAR) models. We observe that Adaptive Expectations Model gives the better one-step ahead forecasts rather than the other models via Root Mean Square Error(RMSE) criterion. Moreover, forecasts of being obtained from Adaptive Expectations and Cointegration models are combined to reach to the more qualified inflation forecasts.