Reflections on current elections

Last Saturday was the national election in Australia. The country has a system in which all the members of the more powerful house - the parliament - come from electoral districts. It includes a mechanism of transferring the vote for a preferred but losing candidate onto a second, third, etc preference until only two are left, so that every winning candidate will have the support of at least 51% or so of the voters even if they weren't their first preference.

Still, it does lead to the kind of distortions one would expect from a district-based voting system that was inherited from Britain. Look at the estimates made by ABC which, as of my writing of this post, look like this:

The Coalition (Liberals and their immediate allies) gets 45.3% of the primary votes and are predicted to win 59.3% of the seats (89/150).

Labour gets 33.8% of the primary votes and are predicted to win 38.0% of the seats (57/150).

The Greens get 8.4% of the primary votes and have won 0.7% of the seats (1/150).

Palmer United Party gets 5.6% of the primary votes and may have won 0.7% of the seats (1/150) although that still appears to be narrow.

From a German perspective, that all seems terribly unfair. In an extreme case, if 51% of all voters preferred party A and 49% preferred party B, the supporters of the latter party could end up with no representation in parliament provided that the supporters of both parties are distributed perfectly equally across the map. The Bs would be completely unrepresented, just as the Green and Palmer supporters are nearly unrepresented at the moment. (Yes, there is the senate, but that is not the house with the real power and it has its own problems.)

More realistically, people are somewhat clumped: this area here slightly prefers A, that area over there leans towards B. Then we get the problem of Gerrymandering. Even if the As had 55% support across the nation, you could draw the electoral districts so that they would only get comparatively few seats. The trick is to lump as many of their supporters as possible together. You would have to draw some districts to include 80% of As and 20% of Bs and many others to include 47% of As and 53% of Bs. Voila, the Bs win more districts although the As have more votes. Apparently, Gerrymandering is not a bad problem in Australia but it is in some countries.

Germany also has an election in about two weeks. The German system cleverly combines districts with proportional representation, and the latter (the fact that parties get the same proportion of seats as they got votes as long as they clear the 5% hurdle) means that Gerrymandering would be useless. And, of course, with the same results as above under the German system the Greens and Palmers would get a decent number of seats in parliament (looking at party programs I assume the result would be the Coalition and the Palmers governing together).

But of course both have their advantages. Although I much prefer the German electoral system, there are at least two things that I would import from Australia if I could. The first is mandatory voting, the second are the preferences. There have been quite a few cases where a German party got 4.7% or so of the vote and thus no seats in parliament. The votes were thus as wasted as those to a losing candidate under a first past the post system. It would be nice if the voters of a German party missing the 5% hurdle, or perhaps even only the party itself, could decide what other party should get those votes then, as under the Australian system.