Oh, I'm sure we can draw some general information from the data ... but I just have a bad feeling that the data, and the relation to GDP, are too self-referential.
I don't see any reason why there's any contamination with GDP stats.  You can argue there's sampling errors in the VMT, but the correlation with GDP is evidence that there is some useful information there, rather than the contrary.