My friend and colleague Daniel Ziblatt, a political scientist at Harvard, writes in response to my post:
bq. From a comparative and historical perspective, the interesting thing about this issue is that while we may be correct to assert that “theories of parties” predict establishment candidates win primaries, this is really a theory about the U.S, particularly apt for the U.S. Republican Party. But what is so fascinating is that this itself is of course a variable that looks different in different countries and different time periods, especially with great consequences in “parties of the right.” This is precisely what my book is about: in 19th century Britain, the party leadership nearly always won these kinds of internal battles, giving rise, I would argue to a relatively moderate Tory Party that made its peace with democracy; its leaders were moderate.
bq. By contrast, in other parties of the “right,” insurgents often beat traditional party leaders in these same kinds of battles in late 19th century. In Germany and Weimar Germany (a perverse high “internal party” democracy if you like), traditional parties of the right were usurped by “grassroots” and usually right-wing radicals, making German conservative parties persistent opponents of democracy. The source of the difference is how these parties were structured internally, how the parties were financed, how party congresses were run, who set the agenda, etc. The consequence was monumental. When “insurgents” won in parties of the right in Europe’s past, democracy suffered.