Gina Kolata recently wrote an article in the NY Times emphasizing the genetic components of overweight and obesity; the headline—Genes Take Charge, and Diets Fall by the Wayside—pretty much sums it up. Kolata covers her behind by saying that behavior and environment do affect your build/weight, but it's quite de-emphasized, and somebody looking to get a quick answer from the article might well miss that toward the end she does cite some researchers' estimation that "70 percent of the variation in peoples’ weights may be accounted for by inheritance," which leaves 30 percent of the variation to other factors.
Now John Horgan posts a response from Ellen Ruppel Shell, a science journalist who writes a lot about food and obesity. She disagrees with Kolata's argument, or at least her emphasis, pointing out examples where cultural practices clearly affect rates of obesity. (Unfortunately, Shell doesn't throw out a number for how much variation in people's weights comes from genetics. I'd also be curious to see how different scientists answer this question.)
One interesting point that quietly emerges from this dispute is that Shell only mentions non-genetic factors that are cultural- or political-based rather than individual-based. Yes, it's true that Americans of the same genetic stock have been fatter because of socio-political changes (sprawl, car culture, etc.), and it stands to reason that socio-political changes could make Americans skinnier again. But that doesn't mean that individual people, in aggregate, have the willpower and control to diet their way to skinniness just because they want to.
So it seems I'm ending with a point in favor of an interventionist, paternalistic "nanny state." People want to lose weight but they don't have the willpower, so the state will step in and show them how it's done!
(Full disclosure: Ellen Ruppel Shell was a professor of mine in J-school, and Horgan used to blog for Discover.)