As the Presidential races gather steam, there is the familiar topic of the timing and location of the primaries. Early primaries gather "too much" attention, though there was once some charm, or even value, to testing the candidates in American living rooms in New Hampshire and Iowa. Super Tuesday sought to increase the importance of Southern states. And California is always in the news because it is not sure where it wants to be on the election calendar. One idea has been to create a national primary, another is surely to rotate a set of primaries, regional or otherwise. But of course a problem with deemphasizing any one state is that the cost of entry increases for candidates, and the pressure of campaign finance is already great.
Missing in this calculus is the role of all-or-nothing primaries. In recent years the Democrats have insisted on proportional primaries, so that second-place finishers can still garner substantial delegates. Republicans have left the choice of how to allocate delegates to the states, some of which have chosen proportional schemes (60% of the vote gets 60% of the available delegates from that state, more or less) and some winner-take-all schemes. Plenty of big states are winner-take-all. Winner-take-all makes some sense if the idea is to reward in a manner that might be correlated with how the candidate will do in the general election. If the (party's) strongest candidate in California will also win in California in November, then that is an enormous prize, and that candidate should get many points or delegates in the primary/convention calculus. That is a big if, however. On the other hand, proportionality does a good job of distinguishing between fairly successful candidates and hopeless ones; a winner-take-all rule gives the fifth place finisher the same reward as the one who finishes a close second. This method would seem more likely to produce a candidate with appeal in many states, and therefore November appeal as well.
It is difficult to know beforehand which method will give the state the most attention. The winner-take-all method can cause some candidates to give up on a state, and then even that state's front-runner will spend less time on that particular prize.
It is surprising and perhaps unfortunate that we have not paid attention to linking the method of delegate allocation to the timing of the primaries and caucuses. It is plausible that proportional allocation is better early in the primary season, and that winner-take-all is superior when the field had thinned. So there is one idea: allow states to choose whether to participate in early primaries (before March 6 perhaps) on a proportional basis, or in later primaries with winner-take-all. It might also be useful to think about linking campaign finance with primary schedules. Early primaries might have stricter spending caps, hard as these would be to enforce on an in-state basis.