I'd consider 3 possibilities:
The data is skewed by factors we don't know about.
The data is accurate and reflects either changes in lifestyle or efficiency gains or both.
The data is accurate and the historical correlations hold. If this is true, the slew of other positive economic data is wrong.
I think the most likely is 1. The least likely is 3.
When an extreme outlier occurs in a data set, you don't dismiss it, but the most common reason is a faulty gauge or collection method.
My hunch is that it's a combination of; refinery closure, weather, changes in driving habits, efficiency, mass transit, sunspot activity, lunar phase, McRib discontinuation, bicycles, motorcycles, tricycles, unicycles...did I leave anything out?
|
