STATISTICAL MODELLING OF MONTHLY DOMESTIC CONSUMPTION OF NIGERIAN PETROLEUM PRODUCTS ETTE HARRISON ETUK ABSTRACT: This work is a time series analysis of the monthly domestic consumption of Nigerian Petroleum products. The realization analyzed is a 110-point series which spans from July 2006 to August 2015. A time plot of the data shows a generally downward trend. Even though the Augmented Dickey Fuller Unit Root Test is significant, the series could not be said to be stationary because its correlogram reveals an underlying trend and a 12 monthly seasonality. A seasonal difference of the series shows no trend. Its correlogram is suggestive of a seasonal autoregressive integrated moving average (SARIMA) model of order (0,0,1)x(1,1,1) . Diagnostic checking 12 confirms its adequacy. For instance its residuals are largely uncorrelated and are normally distributed. Forecasting and simulations of the monthly domestic consumption quantities could therefore be done using the model. KEY WORDS: monthly domestic consumption, crude oil, Sarima model, Nigeria INTRODUCTION: Crude oil is the mainstay of the Nigerian economy. Until early 2016 Nigeria has been the leading African country in crude oil production. It is now overtaken by Angola. To all stakeholders the trend of petroleum related activities is of interest. Petroleum reserves in Nigeria appear to be dwindling. This situation is exacerbated by some factors, one of which is the activities of militant groups, notably the Niger Delta Avengers. The Daily Trust newspaper of 23 August 2016 reports that the country's oil production has dropped to a 30-year low in May 2016 (Daily Trust, 2016). There is a general call to diversify the economy to avert possible depletion or extinction of the reserves. In the sequel emphasis is on the modelling of the monthly consumption of the products with a view to obtaining an objective basis for forecast, control and monitoring of the pattern. Observed from an inspection of the sampled data is a secular trend as well some seasonal tendencies. It is therefore thought that a seasonal Box-Jenkins approach or a seasonal integrated moving average (SARIMA) modelling approach is appropriate. The SARIMA modelling technique proposed by Box and Jenkins (1976) has gained a lot of popularity as it has been hugely successful in the modelling of time series especially those which have exhibited seasonal natures. For instance, Akpanta et al. (2015) and Afrifa-Yamoah et al. (2016) fit SARIMA(0,0,0)x(1,1,1) models to monthly rainfall in Umuahia, Abia State of 12 Nigeria and Brong Ahafo Region of Ghana, respectively. Williams and Hoel (2003), Mombeni et al. (2013) and Farhan and Org (2016) fitted SARIMA models to vehicular traffic flow, water demand in Iran and container throughput at international airports, respectively. Khorasani et al. (2016) have modelled groundwater nitrate contamination by the SARIMA approach. Etuk (2013) modelled Nigerian monthly crude oil prices as a SARIMA(0,1,1)x(1,1,1) and Etuk and 12 Amadi (2013) fitted a SARIMA(0,1,1)x(0,1,1) to monthly domestic production of crude oil in 12 Nigeria. This is just to mention but a few. 32