Tally Room mail - The tedious Dunkley spin game (Patreon)

I'm releasing this newsletter every fortnight for Patreon donors, and it will be available for non-donors later that week.

Federal by-elections generate tremendous amounts of attention as they come close, and some of that is warranted. After all, a seat in federal parliament is at play, and the person who wins that seat will hold it for a long time. But for so much of the conversation, the Dunkley by-election was played up to be more than it was, supposedly having predictive power for the direction of the federal government and opposition.

We saw a lot of expectations-setting in the lead-up to the election, with Labor and Liberal spinners arguing over what an "average swing" would be - do you just look at first-term by-elections, by-elections caused by the death of an MP, or by-elections involving a government seat. All of these calculations produce different averages, but the sample size is not particularly large.

All of this was about giving the by-election meaning beyond simply the result.

But by-elections are a poor guide for political prognostication. At the most they can confirm trends that we can see in political polling, but there are always a lot of local factors that make them less appropriate for national elections.

Firstly, the whole country is not the same as Dunkley. Even if the by-election told us how the voters of Dunkley would vote at a general election (it doesn't), it wouldn't tell us much about other types of marginal seats.

Secondly, the consideration for a voter is not the same at a by-election. The survival of the government is not in play, and voters are free to vote for someone else. There's plenty of evidence that by-elections tend to favour the opposition, and that doesn't translate into a general election.

There are also issues of lost personal vote that won't play out the same way at the next federal election. Labor lost the personal vote of Peta Murphy. Even if voters understood that this was out of the ALP's control (they didn't expel her or anything), it may still see voters go somewhere else. In contrast, at the federal election Jodie Belyea will be a sitting MP, and in the seats which Labor won in 2022 they will have a first-term MP with a new personal vote.

So what does Dunkley mean? It confirms what the polls already tell us - the Labor government is no longer in the honeymoon period, but they remain in the lead. It's also discouraging for a Liberal Party looking for alternative paths to victory that don't involve winning back the teal seats. Dunkley is the exact kind of seat where the Coalition would need to over-perform if they were to win government without winning back its inner-city core.

What I'm reading

I've been really enjoying reading Why Cities Lose, by Jonathan Rodden. Unlike most American political science, Rodden goes beyond the American experience to also expand his story to Canada, the UK, Australia and mainland Europe. It tells the story of how the industrial revolution flowed into concentration of left-wing voters in cities, and how the change in the economic circumstances of those cities then played into the modern party system.

I'm about halfway through the book, but it has a fascinating chapter telling the history of how mainland European countries switched to PR mainly as a response to the rise of workers' parties and liberal parties seeing a proportional system as a way to survive being sidelined. The second half of the book seems to focus more on how winner-take-all electoral systems then disadvantage parties that predominate in cities.

I also really enjoyed Chris Monnox's recent journal article about the changes in campaigning practice after the introduction of compulsory voting in the early 20th century. I've found Chris' history work about how Australian political campaigning so fascinating. So much of political history is focused on the high-level political events in parliament involving political leaders, when so much went on at the grass roots.