Pretty sloppy reporting all around. They aren't even looking at the correct statistics to address the hypothesis.
The article cites the percent of new car sales that were from 21 to 34 year olds in 1985 versus 2010 and find that it falls. What they really want to look at is the percent of 21 to 34 year olds that buy cars. You could have an increasing percent of young people buying cars AND a fall in the percent of car sales driven by young people if young people are making up a smaller proportion of the car buying demographic. Indeed, 21 to 34 year olds make up a smaller proportion of the general U.S. population in 2010 than in 1990 (http://www.censusscope.org/us/chart_age.html). It probably isn't enough to reverse their result, but it soaks up some of the claimed behavior.
They make the same type of mistake looking at the percent of teens with a license. Well, 13-15 can't even have a license, so what is happening to the age distribution within "teens"?
Then you have all of the other causal effects that other commenters address. I've come to expect this type of thing from the Atlantic blog posts, so I was surprised to see that this was a magazine article.
The article cites the percent of new car sales that were from 21 to 34 year olds in 1985 versus 2010 and find that it falls. What they really want to look at is the percent of 21 to 34 year olds that buy cars. You could have an increasing percent of young people buying cars AND a fall in the percent of car sales driven by young people if young people are making up a smaller proportion of the car buying demographic. Indeed, 21 to 34 year olds make up a smaller proportion of the general U.S. population in 2010 than in 1990 (http://www.censusscope.org/us/chart_age.html). It probably isn't enough to reverse their result, but it soaks up some of the claimed behavior.
They make the same type of mistake looking at the percent of teens with a license. Well, 13-15 can't even have a license, so what is happening to the age distribution within "teens"?
Then you have all of the other causal effects that other commenters address. I've come to expect this type of thing from the Atlantic blog posts, so I was surprised to see that this was a magazine article.