What lies behind election swings

This little essay is for those who find elections fascinating. Not everyone does!

We usually compare each election result with the last one, doing our best to decipher ‘the will of the people’. The politicians themselves do it for us too, setting out their interpretation of what has happened. So, with respect to last Saturday’s ACT elections, Zed Seselja points to the seats his party won, while Katy Gallagher points to the votes that Labor and the Greens won. Election analysts talk about shifts in the vote, and point to this or that event or policy or place in explanation.

Only occasionally does anyone look at the electorate itself. For it changes too, all the time. Sometimes there has been a redistribution of electoral boundaries, and parcels of voters are moved from one electoral division to another. That happened in the ACT, but I’ll pass it by because it was not a major event, and in the multi-member electorates that characterise the ACT’s electoral system it is difficult to pin-point the effects of such changes.

Let’s look at the changing ACT electorate, forgetting boundaries for the moment. In 2008 it consisted of 244,000 voters, of whom 212,000 cast a valid vote. Rather more than 8,000 voted informally, and of the remaining 20,000 all we can say is that we don’t know what happened to them. Perhaps they forgot, or were overseas, or they don’t exist at all. At about 10 per cent of the apparently enrolled electorate, that’s quite a lot of uncertainty.

While we don’t have the final figures for 2012, the ACT electorate has grown to nearly 257,000 — another 13,000 — and the growth wasn’t evenly spread around the ACT, which is why there was a redistribution. Some of the growth represents young voters coming of age, and some of it was new arrivals to the ACT. These additional voters actually represent the net increase in the electorate, because there were departures, too. The ACT has a mobile population, because people come and go because of employment opportunities here and elsewhere. Since Canberra has been growing, the net outcome is an increase, but quite a lot of people who were voted in 2008 had left the ACT by 2012.

About one percent of the population dies each year — that’s about 2,500 per year, or 10,000 over four years. Births and immigration combined, for Australia as a whole, amount to about double the death rate. So call that 20,000 for the ACT. Then there is internal immigration — people coming from elsewhere in Australia. So the gross change in the ACT’s electoral roll from 2008 to 2012 may have been as much as 40,000, which is 16 per cent of the enrolment. That’s quite a lot.

So when we look at the change in party shares in the vote from one election to another, we need to remember that we are not focusing just on the people who decided to vote a different way four years later. Of course, there are such people, but there are also a lot of new voters and the corresponding loss of quite a number, who voted in 2008 but not in 2012. All this creates a kind of ‘churn’.

To find out what happens in the mind, so to speak, we need to talk to the same voters at two different points in time, and I once did that, interviewing voters in 1967 and then again in 1969. There are many possibilities for change: a citizen can decide not to vote, for whatever reason, or cast an informal vote, or shift from one party to another. And the more candidates there are the more possibilities there are. Almost certainly, someone will change their usual vote because he or she knows the candidate. And although we always concentrate on the major parties, and the shift to the more successful, there will have been people who voted against the trend — in the ACT case, for example, those who shifted from having been a Labor or Liberal supporter to voting Green — even though the Green vote went down.

I wrote about all of this in Stability and Change in Australian Politics, now long out of print. But the lessons remain the same. What went on in the 2012 ACT elections, compared to 2008, is a rich blend of different movements: new arrivals, departures of all kinds, differential turnout, major party conversions, the circulation of Green voters, the circulation of voters who supported other groups in 2008 — and its reverse.

In the late 1960s, I could estimate the components of the comparable changes. This was a time when the ALP nationally was doing well, recovering from its thumping at the 1966 elections. And although there was a significant net change in Labor’s position between 1967 and 1969, that was the net outcome, for a very much larger proportion of the people we interviewed had changed their position between those two years, and these changes went in every direction. I would love to see that kind of analysis done again today. But it is hugely expensive, and the kind of extended interviews we were able to do forty years ago are no longer possible.

We still don’t know what the final outcome in the ACT will be, but when you hear of it, you might also think for a moment of the great swirl of decision-making that occurred on election-day. It makes democratic elections even more interesting and important than just the outcome in terms of which party forms a government. The ‘two-party preferred vote’ is such a narrow compression of the real thing.