The measure of transportation BTUs per chained 1996 Dollars of GDP is used as a measure of dependence on oil. This time series can serve as a national security indicator, since it shows the level of energy use for transportation with respect to the U. S. economy.
Figure 1 provides a graph of the quarterly data since 1990. This time series exhibits strong seasonality, as well as a strong downward trend. At issue is whether this strong trend has experienced any changes over time. So the trend needs to be extracted from the data.
Figure 1 - Transportation Energy Use per GDP
In the process of analyzing the data, a short-term intervention, or pulse, was found around the first quarter of 1996, where the winter seasonal impact was not as strong as expected. Figure 2 illustrates the energy use per GDP along with its underlying trend; the intervention at 1Q1996 is noted on the graph.
As can be seen on the trend values in the graph, an upward shift in the underlying behavior occurs around the second quarter of 2001. What now needs to be determined is the cause - is the shift due to the change in GDP or the change in the transportation energy use?
[It should be noted that this shift might be exaggerated because it occurs at the very end of the time series. As more quarterly data are added to the analysis, the increase in the number of data points beyond the point of change should allow the behavior to be more easily measured and analyzed.]
Figure 2 - Transportation Energy Use Per GDP
Figure 3 provides a graph of the quarterly GDP over the same time period. The graph does highlight the fact that the shift in underlying trend is due to the shift in GDP. The last two quarters of GDP show a decided drop, which is to be expected since that period corresponds to the beginning of the present recession.
Figure 3 - GDP: Billion Chained 1996 $
To confirm that GDP is the reason for the change, a plot of transportation energy use (in monthly values) is provided in Figure 4.
Figure 4 - Transportation Energy Consumption
An analysis on the seasonal components of the transportation energy use per dollar of GDP does indicate that the quarterly seasonal behavior is very stable. On average, the first quarter of the year results in approximately a 128 unit drop below the underlying trend; the second quarter experiences a 22 unit increase over the trend; the third quarter reached about 94 units above the underlying trend; and the fourth quarter holds at approximately 13 units above the trend. Figure 5 is a graph of these averages.
Figure 5 - Seasonal Deviation