I looked at the numbers: the decline of which you speak is pretty small. In 1980, 94% of the population did not bike or walk. In 2012, that number is 96%. Half as many people walked to work, but they certainly were not in any sense make a damn difference in the overall driving rate.
Plus, your stats show biking rates are up, too.
My argument is that the average person was probably just as lazy in the 80s. This study is arguing that given similar rates of reported activity, there should be similar reported weight levels.
It doesn’t matter who is driving or biking if you aren’t studying average activity rates (which they weren’t).
Personally, I am betting that there have been definitional changes that skew the results. Just look at how activity guidelines have sunk from “30 minutes of vigorous activity” to one I recently saw that asked people to just “be in motion for 30 minutes”.
If those represent the end points on a spectrum charting a single definition of physical activity, then the quality of that self-reported activity is naturally called into question.