A (sort of) argument for partisan gerrymandering
Scott Keyes writes on Poltical Insider: Congressmen who hail from competitive districts are forced to run challenging reelection campaigns every two years, which in turn leaves them better prepared for the difficulties associated with a bid for higher office, right? That is, the closer a district's PVI is to zero, the better chance that congressman has of waging a successful bid for governor or senator because they must constantly win difficult elections and hone their campaigning skills in the process. Practice makes perfect.
However, looking at data from all the election since the 2000 redistricting, this does not appear to be the case. Congressmen from competitive districts do not appear to have any more success in their bids for higher office than those from non-competitive districts. Since 2002, 36 sitting congressmen have ran for higher office, half of whom won. However, representatives from competitive districts - those with a PVI of between D 5 and R 5 - were considerably less successful, winning just five of their thirteen races. Even in those states that swing states that Kerry or Bush carried with less than 55% of the vote, representatives from competitive districts won just three races and lost six. ...
Two possible reasons come to mind that would explain why representatives from competitive districts don't win their campaigns for higher office at a greater rate than other congressmen. First, if a representative must run a difficult race every two years, he or she has little opportunity to amass a large amount of money that would help in his or her bid for higher office. -- Political Insider: Competitive Districts Not a Stepping Stone for Higher Office