What I've never seen addressed in any of the underlying studies, was whether the rate could be determined to be tied to prevalence of condition or prevalence of diagnosis.
Given the "flavor of the month" status that Autism Spectrum Disorders have attained, I'm very skeptical that the socioeconomic link is anything more than the same helicopter parenting that seemed to lead the explosion of attention disorder diagnoses and prescription rates.
It could be a combination of that effect plus the rise in widespread Internet usage, and the massive increase in "geek out" behavior that has caused in the general population (surfing the web, clicking on every little link, web forums, text messaging, etc. all the kinds of things that a hyper-focused nerdy or Asperger's like person would be more likely to engage in).
Lots of diseases are linked to economic status, but I've never heard of autism being one of them.