State and National VMT Estimates: It Ain't Necessarily So
The enormous jump in vehicle miles traveled (VMT) reported by the 1990 U.S. Nationwide Personal Transportation Survey (NPTS) caused a great deal of concern among planners and policy analysts. Such a jump seemed to portend an era of ever increasing travel, pollution and energy consumption.
This paper re-analyses the NPTS data and shows that the VMT jump was a statistical error. The 1990 NPTS oversampled new vehicles and undersampled old ones. Since new vehicles are driven two to three times as much as old one, the sampling bias will overestimate VMT. And the result may have been intensified by an underestimate of VMT in the 1983 NPTS, thus increasing the apparent jump form 1983 to 1990.
I also calculate alternative VMT estimates using data from two other national surveys and a massive odometer-based California study. The three new estimates are in close agreement with each other. I conclude that VMT per vehicle actually grew at only half the rate estimated by the NPTS.