Last night I happened upon this article at the TNR website that advances the unorthodox claim that partisan gerrymandering has nothing to do with the late-20th century decline in the number of contested races in the House of Representatives.
Here’s the article in TNR; and here’s a much lengthier and (to some of us) impossible to understand academic paper that sets forth the results of regression analyses and other fancy math that, we’re told, prove the point.
As you know, though gerrymandering is literally as old as the USA, the conventional argument is that the practice has become increasingly effective in recent decades because computer programs and modeling have allowed the folks who design congressional districts to make a qualitative leap forward in the specificity or perhaps granularity of their knowledge. And that, along with computer programs that can spit out endless permutations of possible district maps, allow the line-drawers to come up with maps perfectly designed to ensure one or the other party’s control of the district. So, same agenda behind the gerrymanders but just a qualitatively leap forward in the technology used to implement it.
But the authors of this piece say this isn’t so.
They don’t say these highly precise districts don’t stack the deck. What they argue is that the legal strictures of the voting rights revolution (‘one man, one vote’, the civil rights act, etc.) have cancelled out the greater precision. In fact, they say it’s slightly more than cancelled it out, making partisan gerrymanders slightly less effective than in the old days.
They suggest that the more likely culprits for decreasing competitiveness are media penetration (probably an overlooked suspect), money in politics and political polarization itself.
(ed.note: Totally obscure reference. I’m curious whether some analogy could be drawn between the effects of media penetration and the role of religious confessionalization on the hardening of confessional lines in the 17th century. Perhaps a strained comparison; but maybe not as strained as we’d like.)
Now, whenever I see a poli-sci study with regression analyses and number-crunching I immediately know the best thing for me to do is not even read it. Because I just don’t have any serious grasp of the methodology that would allow me to independently evaluate the quality of the argument and evidence. Either that or I’m too lazy. But I’d be curious to hear what others make of this. And I have to say that I find myself inclined to believe the dumbed-down overview of the argument presented in the TNR article. The way they interpret the evidence of the last couple decades — particularly, what happened after the 1990 and 2000 census/redistricting — seems questionable to me. But that may simply be a matter of the sample set we’re dealing with being too small.
In any case, I’m curious to hear what others think of this.
