Flag election data: Electoral Commission bleg

I wonder whether the Electoral Commission will keep and publish the data on vote preference orderings in the preferential flag ballot. There are 120 potential permutations on the 5 orderings, but ...rather more where there's ballot drop-off. Still, it shouldn't be unmanageable.

It would be really interesting to check whether there were major or minor cycles that went unrevealed, whether there was a Condorcet winner, whether the Condorcet winner was picked by the preference elimination routine, and whether different run-off options would have yielded different winners. Recall that a Condorcet winner would be the flag that beats each and every other flag in a head-to-head vote. If one exists (and it isn't necessary that one does exist), it is desirable that a selection mechanism select it. If there isn't one, all choices are kinda arbitrary anyway so no particular reason to prefer STV and top-of-order elimination over voting by veto or other mechanisms, but no particular reason to prefer the others either.

I've sent a note in to the Electoral Commission via their website asking that they keep the data and publish it. We'll see.

While we're at it, though, it does seem ... ridiculous ... that we'd be collecting all the preference data and then running an elimination routine without first checking for a Condorcet winner. STV doesn't necessarily pick the Condorcet winner, if there is one. It took me all of five minutes to come up with a preference ordering that failed to reveal one. Shouldn't that be the first stage check before running the preference ballot? The preference data needed for running STV is all you need to check for a Condorcet winner. The algorithm for checking wouldn't be all that hard.

Since we seem to be running all kinds of process changes under extreme urgency anyway, why not have one more process change and ask the Electoral Commission to start by looking for a Condorcet winner?

Malpass on MMP

Luke Malpass and Oliver Hartwich argue for a rather substantial change in New Zealand's Parliamentary structure: a lower house elected by FPP and an upper house elected proportionately by region. They rightly argue that the aim should be for a "least worst" system rather than a best one.

There are some big problems in squaring bicameralism with Westminsterian parliaments. First and foremost, from whom must the executive seek confidence? If an important bill fails in a Parliamentary system, there must typically be a confidence vote. Would a Kiwi senate striking down a piece of legislation trigger a no-confidence vote? If so, then all of MMP's problems with coalition formation return. If not, and the Senate can strike down money bills, what happens in case of deadlock? I've no particular problem with deadlocks, but they don't sit well with the notion of Parliamentary supremacy.

Malpass and Hartwich suggest adopting the Australian system where the Prime Minister can dissolve both houses and hold a full re-election should a government bill be twice rejected by the Senate. This solves one problem, but eliminates a potential benefit of bicameralism: namely, having an upper house elected on a much longer term than the lower house to insulate policy from the transient whims of the electorate.

Brook-Cowen, Cowen and Tabarrok's 1992 primer on constitutional change in New Zealand suggests that, relative to the unicameral first past the post system then operant, bicameralism offered little.

In New Zealand, the introduction of a strong second chamber would fundamentally alter the nature of accountability in government, and in a manner which would in our view be unsustainable. Westminster systems of government revolve around the accountability of the executive to parliament. With two equally powerful but differently composed chambers, the executive will face continuing conflicts in defining the interests to which it is accountable. As a result of these conflicts, we might expect a general weakening of accountability to the electorate. In particular, we would expect:

• a reduced incentive on the part of politicians to mirror the preferences of the median voter;
• an increased incentive to serve the interest of strong special interest groups;
• an increased incentive to maximise revenue and redistribute resources from citizens to the government;
• an increased incentive to favour particular regions and districts at the expense of other regions and districts; [fn: This outcome will hold where the second chamber is elected under a federal system or by means of regionally-based proportional representation.]
• an increased incentive for politicians to indulge their own policy preferences or ideology;
• a reduced incentive to respond favourably to international constraints.

Accountability conflicts of the kind described here are not necessary features of bicameral systems. Rather, they are a product of the particular combination of strong bicameralism and a Westminster parliamentary system. Strong bicameralism and accountable government could, by contrast, be combined if New Zealand were to adopt a more consensual system of government (for example, with the first chamber being elected on a proportional representation rule), or a presidential system of government (with executive power distanced from the legislature).

The status quo has of course changed since then. MMP means that Parliamentary parties must form coalitions to govern; forming coalitions across houses ought not be particularly more difficult than forming them within houses. So the costs of adding a second chamber are now much lower than they were from the 1992 status quo. But it's not clear to me that the main problems of the current system - too short an electoral time horizon, difficulty in ascribing responsibility in coalitions, generally excessive power for minor players - would much be solved by adding a second chamber.

The best argument I've seen for a second chamber is that it slows the implementation of reform such that it's more likely to be embedded: Australia having taken the slower route to economic reform but having brought more voters along. But would a bicameral system as here proposed really slow things down that much?

Every three years we'd vote for a lower house by FPP and for an upper house by regionally-based PR. It's the rare case in which the party with the smaller fraction of the vote gets a Parliamentary majority under FPP. It happens, but not all that often. So the dominant lower house party would also be the dominant upper house party. In the lower house, it would likely have a majority; in the upper house, it would have to form coalitions with minor parties. I'm not sure I see much difference between a large centrist party having to form coalitions in a unicameral parliament and that same large centrist party having to form coalitions in one house of a bicameral system. Moreover, as the party lists developed for Senate elections would be controlled by the party apparatus, the odds of there being substantial within-party differences across the two houses are slight.

I'd modify Malpass's proposal as follows:

• House elections (FPP) every four years; Senate terms of eight years with staggered terms - half elected each time there's an election in the House
• Senate elected by a national PR vote rather than regional PR to reduce geographical coalitions across houses: enough incentives for geographical redistribution in the lower house; demographically-based coalitions in the upper house should check the worst geographic pork-barreling (though massively attenuated as upper house parties are clients of lower house via party list structure)
• The Australian double-dissolution trigger seems a good one. The Executive is beholden to the House for confidence, but a bill twice rejected by the Senate lets the Prime Minister dissolve both houses for election; in that case, some Senatorial seats would be up for three-year and others up for six-year terms.

I'm still rather unsure of the benefits of a bicameral system with an upper proportional house over a unicameral first past the post system in a small country. Increasingly, the main criteria for me is whether voters can adequately apportion blame for bad outcomes. Westminsterian FPP fares best on that metric.

Other reactions:

• Brad likes the proposal.
• NBR reports here.