(‘Hypothesis’ is just an overly fancy Greek word for English ‘hunch’!)

2. I do not know the extent to which party orders prevailed over preference vote-driven reorderings. I am trying to find more detailed election returns to determine, at the least, the extent of preference voting.

3. My claim about the undesirability of CLPR is based on the intuition that, in the presence of popular party leaders apt to make ethnoreligious campaign appeals, electoral rules reinforcing those leaders’ control of campaign dynamics create security dilemmas for voters. Even in divided societies, not every voter’s first choice is the party most representative of his ethnoreligious identity. Call the performance-first voter Voter X. When Voter X has reasonable expectations that an ethnoreligious party representing another group may capture state power and use it to the detriment of people from his own group, he has an incentive to abandon his first choice for the party most likely to protect his own group.

I realize this is a working hypothesis that needs testing. I am thinking about that.

Was that conclusion wrong?

Did (some) Iraqis take to the streets over the choice of open vs. flexible lists?

Also, I do not consider OLPR to be a “candidate-based” (or what I term “nominal”) electoral system. And I think the discussion we had earlier implied SNTV, which is a purely candidate-based system (party votes being literally non-existent). In an open list, on the other hand, the list is the first criterion in allocating seats, and only afterward do the candidate preference votes come in to play. That is, of course, an enormously important distinction relative to closed (or flexible) lists. However, the basis of allocation under OLPR is still that candidates pool votes and a preference vote for one candidate may help the election of another of the same list. Thus it is meaningfully a “list-based” system far more than a “candidate-based” one.

I am not aware of any evidence that suggests that list type affects the extent to which the party system reflects ethnic cleavages. SNTV would tend to allow the representation (whether in a larger party or a micro-party) of very small (for a given district magnitude) ethnic divisions. But I see no reason to expect that the distinction between open, closed, and flexible lists affects the overall relationship of the party system to the cleavage structure. It might do so, but it is not immediately clear how, and I certainly do not think the political science research on this question is clear (yet).

I would also be very skeptical of claims that open list PR allows voters to select “performance oriented” candidates. Typically the margins between winning and losing candidates on a list are so small–at least if M is greater than about 6–that any connection between performance and winning is likely somewhat nebulous.

Finally, security of candidates was certainly a reason given back in 2005 for closed lists. And closed lists are clearly simpler to vote and count, though I actually find that one of the least compelling arguments in their favor.