Abstract
Using a simple linear regression model for estimation could give misleading results
about the relationship between
Y,
and X,. Possible problems involve (1) feedback fromthe output series to the inputs, (2) omitted time-lagged input terms, (3) an autocorrelated
disturbance series and, (4) common autocorrelation patterns shared by Y and X that
can produce spurious correlations. The pri~ary aim of this study was therefore to use
the Box-Jenkins Transfer Function analysis to fit a model that related petroleum
consumption to disposable income> The final Transfer Function Model
z = C(1-w, B) B' z <x> + (1-8 B)a significantly described the data. Forecasts
t (1-o,B) I I I
generated from this model show that petroleum consumption will hit a record of up to 4.8636 in 2014 if disposable income is augmented. There is 95% confidence that the
forecasted value of petroleum consumption will lie between 4.5276 and 5.1997 in 2014.