The new Tasmanian premier, Jeremy Rockliff, announced on Wednesday that the government will proceed with long-discussed plans to expand the size of the House of Assembly from 25 to 35.
This is taking place a quarter century after the Assembly shrunk in size as an attempt to reduce the influence of the Greens in the parliament, something that was effective in the short term, but left the Assembly too small for its job, and out of line with general trends of parliaments growing as the population expands.
Tasmania’s House of Assembly is elected according to the boundaries of federal electorates in Tasmania. Since Tasmania has a small population, it has consistently elected the minimum five federal MPs since the first election in 1901, and those boundaries have been used as multi-member electorates to elect state lower house MPs since 1906.
The House of Assembly had been expanded from 30 to 35 seats in 1959, when the population was about 340,000. The population had grown to 472,000 (39%) in 1998, when the House was cut to 25. The population is now about 541,000, 59% greater than the population when the 35-seat House was first created.
The reason for the change in 1998 was blatantly partisan. The Tasmanian Greens had started winning seats in 1983, but since 1989 had held numerous seats in the House, and two out of three elections over that decade had resulted in a hung parliament with the Greens in the balance of power.
There was a theory that a reduction in the magnitude of each district, thus increasing the quota necessary to win a seat, would reduce or eliminate the Greens. This mostly came true, with the Greens reduced from four seats to just one, with the quota increased from 12.5% to 16.7%. The Liberal Party lost six seats and lost power, while Labor managed to maintain their numbers despite the seat reduction, pushing them into majority government.
But this anti-Greens effect was short-lived. The Greens bounced back to win four seats again in 2002. They have held at least two seats ever since, although it has definitely been harder to win seats than it had been before the seat reduction, with the Greens reduced to three seats in 2014 and two seats in 2018 and 2021.
Meanwhile the impact on the major party benches has been serious. With the exception of the landslide 2014 election, when the Liberal Party won 15 seats, no government has won more than 13 seats, compared to 16-19 seats for the largest party during the pre-reduction era. This has often caused problems with finding talent for a ministry, particularly with some members being brand new and others on the way out.
The government has not clarified whether the expansion would be achieved by increasing the number of electorates to seven, or increasing the magnitude of the existing electorates to seven, but the latter seems far more likely. Tasmania currently doesn’t have to worry about conducting redistributions for its lower house, or for maintaining that part of the electoral roll. If they went to seven districts, that would need to change, and they would need to ensure the redistribution is complete in time for the next election, due in 2025.
On the contrary, increasing the magnitude from five to seven would be easy to accomplish and easy to understand. Indeed Kevin Bonham has already estimated how the extra seats would have likely been distributed with the 2021 election results: the ten extra seats would have been split between 4-5 Liberals, 3 Labor, one Green and 1-2 independents.
I’m sure the Greens would prefer a higher magnitude, although even an increased assembly size would probably help them even without a higher magnitude, since it would concentrate their vote in higher areas and provide more opportunities to win seats, even if they didn’t win them all.
Indeed there is strong evidence in political science to back up the idea that increasing assembly size and increasing district magnitude both enhance political diversity within the parliament.
The book Votes From Seats by Matthew Shugart and Rein Taagepera runs through the explanatory power of the “seat product” which is created by multiplying the total number of seats in the assembly by the average magnitude per district. This statistic then tells you a lot about the shape of the party system – not just who gets elected, but who wins votes. I believe this paper explains the theory more succinctly.
In particular I wanted to look at one metric predicted by the seat product which is the effective number of seat-winning parties. The effective number (ENP) counts the number of parties and weights the count to their relative strength. If you take the seat product to the power of 1/6, you get something which is quite predictive of the ENP.
This next table shows the results of the last ten elections and the parliamentary ENP. Most of the elections in the 25-seat era have had lower ENPs than those before, although the 2010 election, when the major parties were tied and the Greens had their proportionally best ever result, had the highest ENP during this era.
The next table shows what ENP you would expect from various seat products. I also included a more extreme version, increasing the number of districts to seven, each electing seven.
You can see that the seat product did a good job of predicting the ENP during the 1989-1996 era, and also during the 2014-2018 era, but at other times it has usually underestimate the diversity of parties, probably due to the Greens being relatively strong in Tasmania.
One of the fascinating things about seat product is that it indicates that the same trend of greater political diversity can happen whether the product is increased through a greater assembly size, or through a higher magnitude. This explains why the 650-member UK House of Commons or the 543-seat Indian Lok Sabha have so many political parties and don’t just split into the two parties traditionally expected of first-past-the-post systems. But it is easier to achieve diversity by increasing magnitude instead. A 130-seat assembly with 5-member districts should achieve a similar effective number of parties to the House of Commons.
Global evidence does suggest that increasing the size of the assembly, whether or not magnitude increases, does result in an increasing diversity of who gets elected. It’s not just more of the same. And doing so with increased magnitude speeds up that trend.