Forecasting Of Exchange Rate In Regime Switch: Evidence From Non –Linear Time Series Model
No Thumbnail Available
Date
Journal Title
Journal ISSN
Volume Title
Publisher
IJES
Abstract
Description
Most of macro- economic variables exhibit cyclical behaviors, state dependency and regime switching and modeling such series using univariate linear time series models fails. This study test for non –linearity and linearity in the behavior of the Naira/US dollar exchange rate where the exchange rate exhibits a regime switching between the year 1970 and 1994 using a non- linear time series class model to describe the structure of the series. For this purpose, a logistic smooth transition regression was detected to fit a yearly data over the sample period 1970 to 2013. There is evidence of non-linearity in the behavior of the series and also evidence of continuous of regimes or allow of smoothing the series as it was revealed from the large value of the smoothing parameter. The model estimated was powerful and eligible in the sense of conformity to economic reality and sounds usefulness for decision makers, forecasters to forecast macro-economic variables that exhibit cyclical behavior through the use of this type non – linear class time series models
Keywords
QA Mathematics