TESTING FOR STATIONARITY IN TIME SERIES DATA (FOREIGN EXCHANGE RATES IN NIGERIA) USING SOFTWARE

ABSTRACT

This study focuses on testing for stationarity in a time series dataset of the Nigeria's foreign exchange rates, comprising the USD, GBP, EUR, and CFA currencies against the Nigerian Naira. The data which was collected monthly from the Central Bank of Nigeria (CBN) statistical database and was documented on a monthly basis from the year 2004 to 2021, is analysed using the autocorrelation function (ACF) in R. The ACF is computed to assess the correlation between each currency's exchange rate and its lagged values up to 25 lags. The t-statistics of the SACF are then used to test for stationarity in the original data, the results showed a strong evidence to reject the null hypothesis of stationarity, which states that |trk| < 1.645 (at a 90% Confidence Interval).

In order to make the data stationary, differencing was carried out. The t-statistics of the SACF of the first order differenced series was computed and it showed that the series met the conditions of the null hypothesis i.e the first-differenced data is stationary for the four currencies for at least 22 lags.

This study demonstrates the importance of stationarity in time series analysis and the effectiveness of the ACF in identifying and addressing non-stationarity in a time series data.