keywords: Univariate, models, exchange rate, forecasting, time-series, mean square error
The exchange rate plays a critical role in an economy like Nigeria because import and export contribute a large part of the economy. This paper considered the forecasting performance of three univariate models (Decomposition, Holt Winter’s and SARIMA models). To achieve this, monthly exchange rate data was collected from Central Bank of Nigeria Statistical bulletin covering from January 1981 to December 2015 for the analysis while data covering January 2016 to December 2016 was used as out of sample forecast. The result from the models revealed an increasing rate in monthly exchange rate in Nigeria, while Holt winter’s model performed best among the competing models for forecasting monthly exchange rate in Nigeria. The CBN should monitor the exchange rate in order to make loan accessible for business men and investors either local or external investors.