Lane Miles per Capita vs Congestion

[Mdcastle]

I've been plotting freeway lane miles per capita vs hours spent in congestion for major cities. One thing I've noticed is
1) Through the early 1990s, both lane miles per capita and hours spent in congestion both increased.
2) From then on, in some cities (Kansas City, Minneapolis) there appear to be an inverse relationship. In other cities (Portland, Chicago) it's not so clear.

I've also plotted this relationship for a number of cities for 2011. Congestion decreases sharply for the three most built up cities (KC, Cleveland and St. Louis) and increases sharply for the three least built up (Miaimi, New York, and Chicago). In between there are some wild outliers (with San Franciso, LA, and DC on the high side relative to their lane miles), and Phoenix on the low side), but otherwise is more or less flat.

Is this a valid way of doing this, and is there some explanation for these trends?