Jordan Ragusa has an intriguing post up investigating the tenures of national party chairmen since 1856. There are a few nice findings in there, but what particularly caught my attention was the relationship between a party's success in congressional election and a party chair's likelihood of retaining his position. The relationship is a big nothing -- a chair is no more likely to keep his job if his party just saw epic wins in the House or got decimated.
The finding suggests that party insiders recognize that the choice of chair is relatively unimportant to the fortunes of the party. Yes, generally it's nice to have a unifying figurehead in office and to avoid appearances of incompetence and malfeasance, if for no other reason than these things affect the value of the party brand. And while I'm quite confident that the Democrats would have done well in 2006 even without Howard Dean at the helm, his 50-state strategy could possibility be credited for unexpected Democratic pickups in places like Virginia and Montana. But overall, this job doesn't matter that much, and party elites seem to know it.