IJAMML 1:2 (2014) 165-180 December 2014 ISSN: 2394-2258 Available at http://scientificadvances.co.in * Corresponding author. E-mail address: ettetuk@yahoo.com (Ette Harrison Etuk). Copyright  2014 Scientific Advances Publishers 2010 Mathematics Subject Classification: 11D04, 62D05, 62M10, 62M20, 65K99. Submitted by Jianqiang Gao. Received November 3, 2014; Revised December 5, 2014 A MODEL FOR THE FORECASTING OF MONTHLY NIGERIAN BANK PRIME LENDING RATES: A SEASONAL BOX-JENKINS APPROACH Ette Harrison Etuk a and Uyodhu Amekauma Victor-Edema b a Department of Mathematics/Computer Science, Rivers State University of Science and Technology, Port Harcourt, Nigeria b Department of Mathematics/Statistics, Ignatius Ajuru University of Education, Port Harcourt, Nigeria ___________________________________________________________________ Abstract This work involves the fitting of a SARIMA model to the monthly prime lending rates of Nigerian banks from January 2006 to September 2014. Its time-plot shows a generally horizontal trend with a peak between 2009 and 2010. Evidence abounds for 12-monthly seasonality: The correlogram is sinusoidal-patterned with period 12 months and an inspection confirms the seasonality hypothesis. Hence it was necessary to difference the series seasonally once. The resultant series bears a lot of resemblance with the original series. The Augmented Dickey Fuller (ADF) test considers both series as non-stationary. A further but non-seasonal differencing of the series yields a series that is adjudged as stationary by the ADF test. Its time-plot shows a horizontal trend and no clear seasonality. However, its correlogram shows an evidence of stationarity and seasonality of period 12 months. Applying a new algorithm for subset SARIMA modelling, the ( ) ( ) 12 0 , 1 , 1 0 , 1 , 1 SARIMA × model was fitted. It is shown to be more adequate that the corresponding additive model. It is also demonstrated to be multiplicative and not subset. Forecasting of the rates may therefore be based on it.