TPM Reader BP on the redistricting article …
I think you only need to read a few pages of the Friedman-Holden study to conclude it’s bogus. The phrase “regression discontinuity approach” translates to a finding that in the first election after redistricting, more incumbents lose, not less. I think anyone who follows politics would respond to that finding with, “Well, DUH!” Incumbents lose more often after redistricting because the redistricting is being done by the opposite political party of some of the incumbents – they are drawn out of their districts, or forced to compete with other incumbents of the same party (see Pennsylvannia, or the second Texas redistricting). The Abromowitz et al. study is better – it looks at the partisan balance of districts before and after redistricting and concludes the redistricting is not what is skewing the partisan balance toward extremes – that has happened mainly in between redistricting years, as red districts get more red and blue ones more blue. (This Austin American-Statesman analysis documents the same trend by county.)
The results of both studies highlight the imprecision of the term “gerrymandering”. Some gerrymandering is to protect incumbents (e.g. California, allegedly). Some is to get rid of incumbents (e.g. Texas, or Connie Morella’s old district in Montomery County Maryland). And some is for general long-term partisan advantage, as described by reader DB * but I’d also note that this goal does not necessarily imply greater reelection rates for incumbents, as it often involves taking reliable voters away from safe incumbents; Delay did this to himself in the Texas redistricting (which combined getting rid of incumbents with striving for long-term partisan advantage). I don’t think there’s any reason to think the balance between those different goals would imply that better technology for gerrymandering would increase incumbent reelection rates, so the outcome of the studies is not surprising.
