A number of diaries here recently have pointed out that several Democratic incumbents are sitting on very large piles of cash that could be better used in close races. Recently released research by political scientists Jonathan Kastellec, Andrew Gelman, and Jamie Chandler [KGC] suggests that the critics have a good point:
http://polmeth.wustl.edu/...
KGC are interested in understanding the relationships among the proportion of the vote, the geographic distribution of that vote, and the final result in congressional seats gained or lost. The optimal strategy for allocating campaign funds is to give just enough to those in safe seats to make sure that they stay safe, but no more than that. (Winning with 60% is just as good as winning with 75%.) Viewed in this way, large margins for incumbents are a symptom of an inefficient allocation of funds across all campaigns.
KGC find that in the 2004 congressional elections, the Democrats won their victories with an average of 69% of the vote, while the Republican winners averaged only 65%. This suggests that spending by Democrats in safe seats is too high, and that some of their funds ought to be diverted into more competitive races. If the party has formal rules or informal understandings that induce incumbents to share their bounty to help the entire party, they don't seem to be working as well as they could.
(Some of you no doubt have noticed that I posted a very similar diary entry on this same paper the other day. I deleted it and replaced it with this one, which emphasizes different points than the previous one.)