In Defence of STV
This morning's DomPost has a somewhat intemperate if not positively spleenful diatribe by Michael Bassett entitled "STV is flawed madness". Leaving aside the vexed question of whether flawless madness would be preferable, it seems to me that the practical problems encountered with STV are not the end of the world. They ought not to have happened but they have been overcome without injury to the democratic process. The people have voted. The votes have been counted and the counts have been audited. The mayors and councillors elected by STV may, indeed, justly claim a more legitimate mandate than that of, for example, John Banks in 2001 under FPP. Experience with FPP voting technologies in the USA show that it could be much worse.
Bassett's article lists a series of advances that had been made over the years
We fought for a standard system of enrolment at the central and local levels instead of maintaining local rolls where many names were duplicated and others were missing altogether. We fought for a ward system so that locals were guaranteed involvement in decision-making that affected them. And when voter turnout to polling booths dipped below 10% in some areas by the early 1980s, we introduced postal voting. Participation rates leapt, then subsided to levels that are still many times higher than in the good old days. Labour's goal was always to make the system as simple and local as possible to encourage people to participate in civic affairs.and then asks
What on earth decided the present Labour Government to play games with these basic principles in the recent election?
This is pretty misleading. The change to STV (where it happened) had no effect whatever on enrolment, wards or postal voting. Perhaps the author only intended the question to apply to the last "principle" (keep the system as simple as possible) and the ambiguity was accidental but then again - perhaps it wasn't.
Michael Bassett's specific objections to STV are that it is too complex, too many spoiled or blank ballots, the particular version used in NZ was the wrong one, second preferences of elected candidates get less weight than those of excluded candidates and that the final results were incomprehensible. I suspect, however, that his real problem is with any proportional system that threatens the two party monopoly.
STV is not, of itself, particularly complex. It has been used successfully in Australia and Ireland (inter alia) for many years. I don't believe that New Zealanders are any less intelligent than Irishmen or Aussies (even if we do earn a little less). Michael Bassett is, however, making a reasonable point when he questions "running different voting systems at the same election" particularly when the STV and FPP ballots looked very similar and often occurred on the same page. It would be interesting to know whether spoiled ballots for the DHBs were higher for those centres where everything else used FPP. Of course, the risk of spoiled ballots for STV depends on how rigorous the rules for valid voting are. If all the measures suggested here were adopted I suspect that the overwhelming majority of those spoilt ballots would have been valid.
Blank ballots are a slightly different matter. A voter might conceivably be put off voting by the perceived complexity of the system but a blank ballot will be generated only when a voter votes for one local body (or referendum) but not another on the same sheet of paper. If they didn't vote for anything they wouldn't (usually) send their vote in. It is hardly surprising that the DHB elections (no wards, no real power and large numbers of candidates the voter has never heard of) would be prime candidates for blank ballots from electors who voted for their Mayor or City Council on the same sheet. this would have happened regardless of the system but, in fact, all DHBs were STV.
In any event the total impact of spoiled ballots, blank ballots and reduced turnout appears minor. The Wellington mayoral election (under STV) had 268 informal votes and 1046 blank ballots out of 53,483 papers returned while the neighbouring Hutt City mayoral election (under FPP) had 83 informal votes and 1282 blank ballots out of 26,018 papers returned. The informal votes are a slightly higher percentage under STV (0.5%) then FPP (0.3%) but it's a pretty small number in any case. The blank ballots were a higher percentage in Hutt City (5%) than Wellington (2%) but this may be due to reasons other than the voting system. The total informal votes for councilors (summed across wards) were about double those for mayor at 563 (1.1%) in Wellington and 171 (0.7%) in Hutt City.
There are major differences in counting procedure for STV electing only one candidate per "electorate" and STV electing several candidates per electorate. The voting procedure is exactly the same in either case but in the single member instance only the lowest polling candidates are excluded. This makes it effectively equivalent to (but quicker than) the exhaustive ballot system used in the French presidential elections or by most political parties to elect their parliamentary leaders. It is also the system used in Australian lower house elections. With multi-member electorates (eg council wards) STV redistributes both the total vote of excluded (lowest polling) candidates and the "surplus" votes of elected candidates and is an (approximately) proportional representation system. This is the system used for the Australian senate.
There are only very minor differences in the way votes can be counted within either of these variants (single member or multi-member) and there is nothing unusual about the specific calculation algorithm adopted for New Zealand. Nor, to the best of my knowledge, was the electoral calculator module responsible for any of the counting problems encountered.
The claim that second preferences of votes for winners counted for less than those of winners is specious. Under STV every vote always counts as a full vote in total. If the quota is 2000 and candidate A receives 3000 votes in the first round then only A's surplus (1000) votes is redistributed. This means that A's 3000 votes will all be distributed but will count as 0.33 votes each. If you voted for A then your vote counts as 0.67 for A (minimum to elect A) and 0.33 for whoever you gave your second preference to. If A had been the lowest scoring candidate with (say) 300 votes then A would be excluded and all 300 votes would be redistributed. A vote for A then counts as 0.0 for A and 1.0 for the second preference. Under FPP a vote for a very popular candidate (who will get in anyhow) or a very unpopular candidate (who has no hope) will both be "wasted".
The last claim - that the presentation of results is confusing - is fair enough. All the required information is available from the calculator output, it's just a matter of presentation. David Farrar gives an excellent example of how this can be done. Hopefully city councils and mainstream media will follow this lead.
However, it is the second last paragraph of the Bassett article that really got to me
The early Labour Party experimented with proportional voting at local elections , then reeled back when they saw the results.
I don't know what the historical incident referred to is but the only legitimate goal of any voting system is to fairly reflect the choice of the voters. Not (not EVER) to produce an outcome acceptable to Labour or any other political party.