Open lists sound great in principle, but I've yet to come across a concrete proposal that seems workable.

Allowing voters to rank their preferred party's list candidates in order has a big problem: most won't bother. (There is a lot of experience to this effect, from preferential-voting systems around the world.) If those non-votes are just ignored, you end up with a party list decided by a tiny minority of the party's voters. If the non-voters are assumed to be acquiescing to the party's preferred list order, then that order is pretty much the one that prevails. Either way, there's no real improvement over closed lists.

Alternatively, the list order can be based in some way on the electorate vote (e.g. the best-runners-up rule). The problem here is that it's far too easy for the parties to manipulate individual electorates to achieve their desired result. ("Vote for my party, but please, please don't vote for me" - sound familiar?)

Political parties know very well which electorates (and sub-areas within electorates) are likely to yield them the most votes; they put a lot of effort into finding this information out, so as to plan their advertising spend efficiently. Under a best-runners-up system, they would simply place their candidates accordingly. The only real difference from the current system would be that the voters (but not the politicians) would have a much hazier idea of what was going on.

The notion that electorates can't possibly be larger than 1/70th of the population isn't one I subscribe to myself.

It's interesting to note that MMP could work just fine with a 99-seat Parliament - if we were willing to reduce the number of electorates from 70 to 60 (leaving 39 list seats to maintain proportionality with). That doesn't seem so different from the first MMP election in 1996, which had 65 electorates.

how do you get from 70 to 120 rather than 140? Maori seats?

MMP uses the list seats to achieve proportionality. For this to work nicely, the list seats need to be roughly 40% of the total; having 70 electorates thus implies 120 (-ish) seats in all.

It's not a hard-and-fast requirement - the system would still work with fewer list seats, but there would be a risk of less-proportional outcomes. Since proportionality is one of the main things MMP has going for it, that would be a serious flaw.

The Maori seats (currently 7) are included in the total of 70 electorates.

The argument that "MMP needs 120 MPs" is based on the assertion that the maximum practicable size for an electorate is 1/70th of the population. Thus, the current 70 electorates are already as large as they could possibly be. To be kind, the evidence for this notion is less than overwhelming, especially when one considers that (a) NZ has never actually tried having larger electorates than that; (b) other countries have, with results that don't seem unworkable.

If we accept this "1/70th rule", then FPP and PV could both work with a Parliament as small as 70. If you don't really care how politicians are chosen, but are just holding out for the slim chance of reducing their numbers, then these are the options to go for.

For STV (where every electorate has at least 3 MPs) a Parliament of at least 210 is required, although perhaps we could argue that larger electorates are more feasible when they have multiple MPs to service them.

For SM, Parliament needs to number at least 94 (if we are to maintain the ratio of 1 list seat for every 3 electorate seats).

And for MMP, we need 120 MPs.

If I just vote in favour of MMP and not tick an alternative, does that make MMP stronger, weaker, the same?
If I did give a second tick to say PV, would that help to undermine say SM?

Casting a vote on the second question does not affect MMP's chances of winning on the first question.

But: if MMP doesn't win this time, its chances of surviving long-term may depend very much on which of the other 4 systems it goes up against in 2014.

So yes, if you want to "undermine SM" (for strategic reasons, or just because you don't like it), you should give a second tick to one of the other three alternative systems.

In most respects, the alternative that is closest to MMP is STV. So that might be a reasonable choice for those who are voting in favour of MMP on the first question.

An online bookshop - a good idea, if not exactly novel.

But, why involve the authors? Your own analysis confirms that the authoring part of the book trade suffers from very low returns. You could do almost as well - with a lot less effort - by focusing on the retailing part only, and acquiring the rights to titles in the usual way.

Matthew's comment is far too long, but dead right. The projections used by Keith and Treasury are just the midpoints of some sprawling probability distributions. But over this kind of time horizon, uncertainty dominates: it's the rest of the distribution that is most of the story.

The world could indeed catch the Japanese disease, or any of a number of other ailments. Keith seems to want to ignore those possibilities, on the grounds that they aren't in the most likely scenario, and we can't stop them happening anyway. But they matter: NZ will have its best chance of surviving those kinds of external shocks if we haven't put borrowed money into the Super Fund.

The big thing missing from this post is any consideration of risk. Borrowing to invest in sharemarkets might be a winner, or it might not - and the risk of a poor outcome is high enough to stop prudent people from doing it.

The Super fund has the advantage of a long time horizon, which makes poor outcomes unlikely (at least if the experience of the 20th century is any guide to the 21st). But there are still bad scenarios. The next 20 years might see the Chinese economic miracle collapse, or a major war, or the peak-oilers proved right. And there's the problem: those bad scenarios for the Super fund are the very ones where it will be most desperately needed, because they are also bad for the NZ government's fiscal position. That's really what makes the risk intolerable.