WA 2021 – the not-so-level playing field for the upper house


Australia has a long history of electoral systems that slant the playing field in favour of particular ideological positions, favouring certain voters over others. Limiting the vote to people with property was common in the mid-19th century, and survived in some upper houses well into the 20th century.

It was also very common for rural electorates to be drawn with smaller populations than urban electorates, or for boundaries to be left unchanged as urban populations boomed.

Western Australia is the last relic of this “malapportionment”, with power in the upper house favouring rural voters and the parties who do better in those areas.

Malapportionment in the lower house was a common tactic for skewing the popular will in the middle of the 20th century, propping up conservative governments in South Australia, ensuring the Country Party were a viable party of government in Victoria, and inflating parliamentary majorities for Joh Bjelke-Petersen in Queensland.

Western Australia finally abolished its malapportionment for the lower house before the 2008 election, with a redistribution moving a number of electorates from rural areas to urban areas. But that same process was unable to fix the malapportionment in the upper house, with a Greens upper house MP refusing to go along with the reform plans. Instead the malapportionment in the upper house was made even worse.

The Legislative Council is elected from six regions. Three of those regions cover metropolitan Perth, which as of 2021 includes 75.8% of all enrolled voters. The other three regions cover the remainder of the state. Each region elects six members.

This means that less than one quarter of all voters get to choose half the members of the upper house. The malapportionment is even worse, since the South West region covers more electors than the Agricultural and Mining and Pastoral regions combined.

I’m going to run through the voting numbers between these regions, which shows the impact of the malapportionment.

This first table splits the vote between the three metro regions and the three rural regions.


There is a big gap in the Labor and Liberal vote between these areas, but the Nationals are the most dramatic, polling 18% in the rural areas, but only 4.4% statewide. The Nationals vote is even more concentrated in the two most malapportioned regions, polling 30% in Agricultural, 19% in Mining and Pastoral and just 12% in the South West.

There are also similar skews for the right-wing minor parties, with One Nation and the Shooters both polling substantially better in the rural regions. The Greens are the reverse, doing much better in the metro regions, and doing much better in the South West than the northern rural regions.

It’s then useful to look at the seats won by each party in these different areas.


Labor holds two more seats in the metro regions than in rural WA. Overall the “left” holds ten out of eighteen seats in the metro area, while the “right” holds ten out of eighteen in the rural areas, resulting in an even split of eighteen seats each.

The actual result would be quite different if the upper house seats were distributed evenly. The Nationals would struggle, and I’d expect Labor to do better.

The 2021 election may well be decisive enough that Labor and the Greens will have a decisive majority, and that will likely see reforms proposed that will end this last population imbalance in Australia’s voting systems.

Liked it? Take a second to support the Tally Room on Patreon!


  1. The logical first step would be to roll Agricultural and Mining and Pastoral into one region – that would fix the grossest of the malapportionment.

    This would still leave non-metro WA with 40% of seats with 25% of the population, but it’s a start.

    If you then added the missing six members back to the three metro regions (making 8 per region), then you’d get Perth having 66% of the seats with 75% of the population. I think that would not be an unreasonable split (a slight regional imbalance, but not unreasonably so). An alternative would be to recreate the Metro regions as four six-member regions – but how that would work I don’t know; perhaps Central Metro, North Metro, East Metro and South Metro.

    M+P being merged with Agricultural is a bit of a mishmash of communities, but given nearly every suggestions for the Federal seat of Durack has a worse abomination (the Kimberley in a seat with outer suburban Perth), it’s not the worst setup in WA politics.

  2. My understanding of the original 2005-8 Legislative Council reform plan is that it was only a dilution of the rural and regional weighting from equal with the Perth Metro area to a 15:21 split (through adjusting the number of member per region) in the Metro area`s favour, still leaving significant rural and regional weighting and with the same increase in non-metro malapportionment through equalising the number of members in each non-metro region without equalising voter numbers. It was, on balance, fairer than what actually happened but not one vote, one value. There is not real way of having 1 vote, 1 value without reforming the balance of regions between metro and non-metro (other than ridiculously large metro regions, e.g. 15-18 members each).

    I suspect that work from home expanding the residence rage of Perth based “office jobs” in Agricultural and South West might help with malapportionment (except for Mining and Pastoral).

  3. I’d like to see a 4*9 model. Keep the metro regions the same, merge the three rural ones. Easy to do, keeps the size of the upper house the same, and it’s roughly proportional. Regions with an odd number of seats are generally better (less chance of deadlock), and with a 10% quota, the Nats can still get a couple of seats, as can the Shooters or whoever.

    A 50-50 election would probably have a 5-4 split to the left in East and South Metro, 5-4 to the right in the other two, and any majority Labor govt should be able to get things through the upper house with the Greens. 2017 would’ve had a fifth left seat in North Metro and a sixth in East Metro – something like 16 ALP 4 Grn, a clear majority.

    Obviously there’s other things that need to be fixed. GVTs need to get gone – there seems to be a grim inevitability that there’ll be at least a couple of microparties elected with 2% of the vote on Saturday.

  4. Given the size of the state, the diversity of the regions and the sacrifices that many make to live in regional WA, I’m not too bothered by the existence of malapportionment. Perhaps add in an extra metro seat in each region if need be, but like Tom I too hope that we start to see a return to the regions as people’s work conditions change.

    Group ticket voting is much worse, although I don’t think getting 2% of the vote really counts as a micro party. A micro party is something that basically only exists on a sheet of paper, and never gets more than 1%. There are some actual minor parties (SFF, AC, One Nation) who generally preference each other, but their voters won’t take much issue with one of them being elected, and collectively would be not too far removed from the Greens total proportion of the vote. Get rid of GTV, and a bunch of the actual micro parties (Daylight Saving, anyone?) will disappear.

  5. There are lots of ways in which some people are disadvantaged relative to other ones, why is living in the regions the only one that gets to have the weight of their vote tripled in value (on average, for some it’s more than that)?

    I think GVTs should be got rid of but if I was Labor or the Greens I would want both done simultaneously.

  6. While the system as it is may be unfair, it is perhaps more workable. Absolutely the Lower House as the chamber of government should have been and was reformed to basically one vote, one value but that is a very Perth based chamber with 41/59 seats.

    Where a government can win on one city alone, as was the case in the previous election, I think it can be beneficial for the stability of the state overall to have a malapportioned upper house. If you allow the sheer numbers of people in Perth to dictate you run the risk of the regions becoming increasingly disaffected.

  7. Tom, here’s a funny one for you: there generally isn’t a tie in the 36-seat LC. Reason being: the govt has to pick a speaker, who doesn’t get to vote unless breaking a tie… but without him, there’s only 35 (an odd number), so there can never be a tie to break. Look at the current one: Labor+Green-1 is 17 out of 35, an effective minority.

    Unless, ironically enough, votes on things like upper house reform – then it has to be an absolute majority. That’s why Labor needed the ex-Lib Alan Cadby as well as the Greens back in 2005

  8. A tie with the total numbers is fine. I think what Tom is referring to is the numbers in each electorate. Odd numbers are generally fairer than even numbers for STV elections. It means in theory a party with a majority of votes wins a majority of seats.

    I think the number of lower house seats should be boosted to 60, then 5*7 with each region consisting of exactly 12 lower house seats. This keeps the total number of MPs the same.

    7*5 with 63 seats (making 9 seat regions) would also work nicely. That would likely still result in a merger of M&P and Agricultural, and create 2 new regions in Perth, but it would leave “South west” largely intact.

    One thing to point out is outer suburban seats in rural regions is normal in Victoria (e.g Monbulk, Melton). I don’t think it will be too big a deal if seats like Darling Range and Swan Hills end up in rural upper house regions.

  9. @Malcolm, every state on the mainland is more than half it’s capital city (Brisbane is line ball, but “South east Queensland” is easily more than half of QLD). They all theoretically shares that ability. WA is about 80% greater Perth but that’s comparable to SA and Vic.

    In every single state both the party of government and opposition hold both urban and regional seats.

    Victoria, which is over 75% greater Melbourne, manages to have 3 of its 8 regions be predominantly non-metro/rural with no malapportionment. This is done with a slight mixing in of outer suburban seats.

    Even without that, neither major party can afford to forget the upper house seats up for grabs in the regions.

  10. A couple of comments have suggested things like : “I too hope that we start to see a return to the regions as people’s work conditions change”. Let’s be realistic about this – to balance this malapportionment, you’d be talking half a million people at least abandoning Perth for the regions. That’s simply not going to happen. If you got 50,000 that would be huge.

    Either way, I can’t see why my LC vote should be worth twice as much if I moved to Mandurah, and three times as much if I moved to Lancelin. That does not make sense. Some *minor* extra benefit for the regions is acceptable, but not a 3:1 ratio. That’s where I came up earlier with the regions getting 33% of the vote off 25% of the population.

  11. Even all these years later i am still astonished that the Greens agreed to – or actually insisted on – such an horrendous anti-democratic outcome in the WA Upper House. 🤯

    Hopefully that will be fixed after this election. (along with scrapping GVTs – horrifically anti-democratic in a different way).

  12. @Andrew Bartlett – The Greens position in 2001-2005, as described in a paper by former Democrats MLC Norm Kelly (https://www.parliament.wa.gov.au/parliament/library/MPHistoricalData.nsf/0/3a1d179df0311c7b48257744001739fb/$FILE/MP271%20Kelly%20N.pdf.PDF – sorry no idea how to link) is certainly a somewhat different take on democracy. I somewhat cynically refer to it as “votes for trees”.

    I’d have liked to have been a fly on the wall as they arrived at their “consensus”. Kelly’s paper has a long quote ascribed to “one Greens MLC” but isn’t clear whether that is Dee Margetts, who is cited for the general principle in the previous sentence.

  13. @Andrew Bartlett, totally agree. It was a horrible model that backfired quite spectacularly. However, there now appears to be some appetite to reform the Upper House, considering the malapportionment gets worse every election.

    In terms of reform models, Jim McGinty’s proposed model in 2005 (as referred to above) was 3 x 7 member regions in the metro area, and 3 x 5 member regions in the country – a 60:40 split to Perth, which still involves some generous country weighting, albeit not as much as is currently the case.

    The model with 4 regions of 9 members each has some appeal and would bring the numbers close to one vote, one value in the Upper House (assuming the 3 country regions mostly become one), however it would create a significant backlash in the country.

    I personally lean towards a model of electing the Upper House as a whole, along with the abolition of the archaic GVTs. This could be done by either electing all 36 members every 4 years (which may require a minimum quota to be established to win seats) or electing half the Legislative Council (18 members) every election, with each member having 8 year terms (as is done in SA and NSW). The WA Parliament has previously expressed a reluctance to have 8 year terms, on the basis that they are too long.

    Any model that reduces the size of the Legislative Council requires a referendum to be passed, which would be exceptionally difficult achieve without widespread bipartisan support.

    Any Bill that alters the six existing Legislative Council regions (although not the number of members in each region) may also require an absolute majority in both houses at the second and third reading stages (this area is extremely complex, and Parliament was far from clear on this during debate in 2005). If Labor and the Greens obtain a majority in the Upper House on Saturday it is likely to be by the barest of margins (most likely 19 out of 36 seats), and with Labor needing to supply the President this may also affect the models and options available.

  14. https://www.tallyroom.com.au/41191#comment-751996

    If there is a tie over who should be President of the Legislative Council, that could get messy (although not as messy as a tied Legislative Assembly in a fixed term system).


    I was actually referring to overall numbers within a chamber. A tie in overall numbers has potential not to be fine (such as I describe above).

    My understanding is that reducing the size of either house of the WA Parliament would involve a referendum, so a 35 seat Legislative Council is more difficult than expansion.

    Victoria does have suburban fringe electorates in the regional regions, however, I suspect the Vic Nats would like to expand the Victorian Parliament to significantly reduce this. Expanding Agricultural into Perth`s outer suburbs has merit as an idea.


    I agree, work from home is no substitute for reform, however every little bit helps.

  15. If shrinking requires a referendum, then I think a 40 seat upper house 8×5 is ideal, and expand the lower house to 64 seats.

  16. Odd numbers are better, but if 6×6 is easier to retain.
    Region 1: “Mining and Agricultural”, made of Mining&Pastoral + Agricultural + Swan Hills + Kalamunda
    Region 2: “South West”, made of the current South West + Darling Range + Warnbro I guess
    Region 3-6: Perth (not sure which compass directions best suit a 4 seat configuration)

  17. I think keep it simple. Either do away with regions like SA or NSW, or go to a Vic-style configuration.

    Whatever happens, just get rid of the malapportionment and the craziness of the GTV.

  18. A fairer system would have eight regions instead of the current six. Six regions would each elect five Members and the two large remote regions with low populations, three each. The ratio between the number of electors in each region and the numbers of Members to be elected would be steady for all. It would restore odd numbers to be elected which ensures a majority of votes will win a majority of seats. It’s basic democracy.
    There is a web page on this issue;
    which TallyRoom correspondents may wish to view.
    Best from Graham

  19. I make ‘the middle of the twentieth century’ to be 1950, when, far from ‘inflating’ Bjeke-Peterson’s representation, the Labour Party was able to win government despite being ourpolled, by the opposition. This continued through to 1956 when, with 51.2% of statewide vote, Labour were able to snag 49(65%) of 75 seats. This, incidentally marked 53 years of near-consecutive incumbency (punctuated by Great Depression only) since federation, for Labour.
    Rural depopulation started the shift, & I guess after the change of government, the unicameral nature of the Qld., parliament favours incumbents. Over to you, Andrew Bartlett.

  20. A century is a long time. I’d define the middle of the twentieth century to cover numerous decades. Roughly you could include 1925-1975. It was a common tactic in 1950, even if it was still helping Labor in some places at that time, and it did all of the things I listed before, say, 1980 (and a little bit after that), which I would define as the middle of the century.


Please enter your comment!
Please enter your name here