Issues in Business Management and Economics
Vol.2 (3), pp. 054-059, March 2014
Article ID BM/014/013. 06 pages
Copyright © 2014 Author(s) retain the copyright of this article. Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 3.0 International License
Original Research Article
An additive seasonal Box-Jenkins model for Nigerian monthly savings deposit rates
Accepted 22 February, 2014
Ette Harrison Etuk1, Igboye Simon Aboko2, Uyodhu Victor-Edema3, Mazi Yellow Dimkpa4
1Department of Mathematics/Computer Science,Rivers State University of Science and Technology,Nigeria.
2Department of Mathematics/Statistics,Rivers State College of Arts and Science,Port Harcourt,Nigeria.
3Department of Mathematics,Ignatius Ajuru University of Education,Port Harcourt,Nigeria.
4Department of Mathematics/Statistics,Rivers State Polytechnic,Bori.Nigeria.
Corresponding Author’s Email: ettetuk(at)yahoo.com
Monthly savings deposit rates of Nigerian banks are herein modelled by seasonal Box-Jenkins methods. The realisation analysed, herein called SDR, spans from January 2006 to November 2013. The time plot of SDR shows an overall negative trend, an indication of a decline in Nigerian bank customers’ savings dividends with time. Seasonality is not obvious. Twelve monthly differencing of the series yields a series SDSDR with a generally horizontal secular trend and no clear seasonality. Both SDR and SDSDR are adjudged to be non-stationary by the Augmented Dickey Fuller Tests. Non-seasonal differencing of SDSDR yields the series DSDSDR. This series has a horizontal trend and still with no clear seasonality. It is adjudged to be stationary. However its autocorrelation plot establishes 12-monthly seasonality. Two seasonal autoregressive moving average models are proposed and fitted: one is multiplicative/subset and the other additive. This latter model is shown to be superior to the former one on all counts. For instance, it better explains the variations in the data than the former model. Moreover, the residuals of the additive model are shown to be uncorrelated. This shows that the model is adequate. Forecasting of the deposit rates might be done on the basis of the model. It is recommended that further studies be done with a view to finding models that could better explain the variation in the time series.
Key words: Savings deposit rates, SARIMA models, additive seasonal models, Nigeria