Univariate and Multivariate Unit Root Tests with and Without Structural Break(s) : An Empirical Examination Using Real Interest Rate Parity (RIRP) Relations

The study examines the methods of testing the presence of unit root and co-integration of time series data using those tests that assumes the presence of breaks and those that assumes no breaks. We employed Eliott et-al(1996) DF-GLS for the latter test and Perron(1997) for the former category. In terms of the co-integration tests, we applied both Johansen(1992) andGregory and Hensen (1996a)procedures that allows to check for co-integration in the presence of no structural breaks and the one that assumed the present of shift under alternative hypothesis. The study applied these tests to examine the Real Interest Rate Parity Relationship between Nigeria and her major trading partners namely: China, India and USA. To do this, monthly time series data was collected on interest rates (call rates and 90 days Treasury Bills rates) and Inflation (Consumer Price Index) from the respective statistical institutions of these countries. The following evidences were found: firstly, there is no difference between breaking and non-breaking univariate and multivariate unit root tests. Secondly, there exist a Real Interest Rate Parity (RIRP) between Nigeria, China, India and USA and the direction runs from Nigeria to these countries. Therefore, the study concludes that it is profitable to borrow money in Nigeria and deposit in China, India and USA at prevailing market interest rate