South Australia and Tasmania both go to the polls tomorrow. I don’t have a great deal to say about either; I share the general expectation that the Liberal Party will comfortably win both, thus removing Australia’s last two Labor state governments. But there’s one aspect of the Tasmanian election that’s worth a look, because it ties in nicely with yesterday’s post about manipulation of electoral systems.
There’s no real doubt about the Liberals winning a majority in Tasmania – today’s Newspoll gives them 53% of the vote – but it’s not entirely certain who will emerge as their official opposition. Although it’s not likely, there’s a small chance that the Greens will win more seats than the ALP, even though the ALP will certainly have more votes (Newspoll says 23% to 16%).
How could this happen? In particular, how could it happen in Tasmania, whose electoral system is acknowledged to be the most democratic of any of the Australian lower houses?
The problem is that there’s a big difference in the distribution of the parties’ votes, as is shown by Kevin Bonham’s polling aggregates. Labor’s vote is quite evenly distributed across the state; Bonham has it somewhere between 21% and 26% in each of the five electorates. The Greens’ vote, however, varies much more: from a low of 8.7% in Braddon to a high of 24.3% in Denison.
A quota for election is 16.7%, but with a bit of luck a party can scrape together enough preferences to win from a primary vote of 13% or 14%. The Greens got up in Braddon last time with 13.8%, and the Liberals won two seats in Franklin with 30.5%.
So imagine that in four of the five electorates, Labor has about 24% and the Greens have about 15%, with most of the rest going to the Liberals. Each of those electorates will therefore return three Liberals, one Labor and one Green. In the fifth electorate, Denison, where the Green vote is concentrated, imagine it’s something like 40% Liberal, 27% Greens, 23% Labor and 10% others. That would probably result in two Liberals, two Greens and one Labor.
The total would be 14 liberals, six Greens and five Labor, even though Labor in aggregate would have outvoted the Greens by about 6.5%.
As I said, I don’t think a scenario like this will actually be realised. But it’s a possibility, and the fact that it is amounts to an important lesson for the ALP – or anyone else who thinks they can fix the electoral system to their advantage.
Although we refer to Tasmania’s Hare-Clark system as “proportional representation”, its proportionality is limited by the size of the electorates. If the whole state voted as one electorate, we know the result would very closely match the totals of votes. With 5-member electorates, that’s not necessarily so.
In 2010, as it happens, it was pretty good. Labor and the Liberals each won 40% of the seats with a bit under 40% of the vote, and the Greens won half as many seats with their 21.6%. But there’s no guarantee that votes and seats will be such a close fit.
And, of course, the reason that this is so capricious is that in 1998 the two major parties did a deal to try to shaft the Greens by reducing the number of seats per electorate from seven to five. With seven seats each, and a quota of 12.5%, proportionality would be much more assured. A scenario of Greens outperforming Labor would be much harder to construct; on the numbers given above, Labor would probably win nine or ten seats to the Greens’ six.
In the short term, the deal worked well: the Greens were reduced to just one seat at the 1998 election. But by 2002 they were back to four, and in 2010 they won the balance of power again.
If Labor were to be reduced to third-party status after tomorrow, it would be the ultimate lesson on how partisan changes to electoral law can backfire very badly.