Interrupted Time Series Analysis of Nigerian Monthly Distribution of Petroleum Products
- Author Ette Etuk,
- Co-Author Azubuike Agbam, Imo Moffat
- DOI
- Country : Nigeria
- Subject : Mathematics
Keywords : Petroleum, distribution, intervention, Nigeria, Arima modelling
Abstract
This work is about an intervention model for monthly distribution of petroleum products in Nigeria. A look at the time plot of the distribution from 2009 to 2015 reveals an abrupt jump in the distribution in January 2013 which continued to 2015. The intervention is believed to be the deregulation of the downstream sector of the Nigerian oil industry which took place January 1, 2012. This means that the impact was not immediate but came twelve months afterwards. It has been noticed that this jump is statistically significant. The pre-intervention distribution is adjudged as non-stationary by the Augmented Dickey Fuller test. Therefore it has been differenced and the differences are stationary. An AR(2) model is found the best for modeling them. Post-intervention forecasts have been obtained on the basis of the ARIMA(2,1,0) model. The difference between post-intervention observations and forecasts is modeled for the intervention transfer function. A close agreement is observed between post-intervention observations and intervention forecasts. Intervention measures may therefore be based on the intervention model so computed.
Comments
No have any comment !