Monday, July 13, 2015

If Ordering Dinner Was Like Senate Voting

Another instalment in the Senate reform debate is overdue, following a Richard Denniss op ed published by the Canberra Times a week ago.  I've been extremely busy with work recently and fortunately Ben Raue at the Tally Room has dealt with most aspects of the Denniss article.  (Anyone who hasn't done so already should read Ben's article before reading this one).  There is one main point I want to deal with at some length here, and that is the article's analogy about buying dinner.  I want to repair this analogy and explain what real-life decisions about something as simple as obtaining a takeaway meal would really be like if ordering dinner via a friend was as silly as our current Senate voting system.

In the process, I hope to demonstrate that the way we (and only we) select Senators is so utterly bizarre that if we tried to do anything else at all by such a method we would find it completely ridiculous.  Click the "Senate reform" tab at the bottom for more of my writings on the subject.  (I am sometimes accused of using sledgehammers to crack nuts, and this article is probably a prime example of that, but while unsound claims continue to be given media coverage I will continue doing so.)

The Dinner Analogy

Denniss seeks to demonstrate his case for the current system being maybe not so bad with the following:

"Australia's preferential voting system is as much maligned as it is misunderstood, so before explaining the changes the Liberals and parts of the Greens are inching towards, and the other options on the table, it's important to explain a few concepts.

Imagine you wanted your flatmate to buy you some takeaway dinner on the way home from work.  You could ask them to get you a hamburger if they could find one, or a kebab if there were no burgers, or some Thai food if there was none of either.  You might not realise it, but if you have structured your request in that way, then you have effectively voted preferentially.  That is you made it clear what you wanted to happen if your first or second preference wasn't available.  The Australian voting system allows you to do the same when it comes to candidates.  You put a 1 next to the candidate you like the best and, if few others share your support and that candidate loses, your vote helps elect the candidate you liked second best or third best".

Denniss then goes on to compare this with the US first-past-the-post system, and claims that the proposed changes would move us further in the direction of the US system.

Actually it seems it is most of the Greens (with no verifiable evidence of significant dissent) and parts of Labor as well as the Liberals who support the changes.  But if this was a debate about the merits of preferential voting as we have it in the House of Representatives, and why we have preferential voting in the Reps rather than first-past-the-post, then Denniss's analogy would be a very good one.  The only objection to it then would be that if you told your flatmate that you just wanted a hamburger but if it wasn't available you'd have anything, your flatmate would refuse to buy you anything, and would insist that you place all the available five or ten dinner options in order so he knew which one to pick if there were no hamburgers.  It might seem that the flatmate is being a bit compulsive here (or like me in similar situations, afraid of having to use any initiative), but it's no big deal as there are not that many options.

But as a Senate analogy ...

Let's see how this analogy really works in the Senate, and imagine, The Checkout style (but longer than an episode thereof), what getting your flatmate to get you dinner would be like if it really "worked" like Senate voting.  Imagine you're the hamburger fan from the Denniss article, and this is what happens when, over a series of nights, you try to get your flatmate to get you a hamburger.

On Monday, you ask your flatmate to get you a hamburger.  If there are no hamburgers, you'd like a kebab, if there are neither of those you'd like Thai food, but if all three are unavailable you're happy with anything so long as it isn't a vegemite scroll.  After all, the chance of none of your three choices being available is quite low.  Your flatmate doesn't get you anything and tells you that your choice of a hamburger was not valid since you hadn't indicated a full list of preferences.

On Tuesday, you ask your flatmate to get you a hamburger.  Your flatmate says they can only try to get you a hamburger if you give a ranked list of all the available possible dinner alternatives.  OK, what are they, you ask innocently.  Your flatmate rattles off a list of 110 possible dishes from 42 restaurants in town including dishes you had never heard of at restaurants you'd never heard of (and even some which could be purchased from nameless food vans) and including several dishes you dislike.  Your flatmate gives you a form to fill out listing your preference between these 110 dishes.  Hungry, you omit numbers 62, 77, 83, 86 and 102 from your form, and your flatmate returns some time later with nothing, telling you that these mistakes had rendered your preference for a hamburger invalid.

On Wednesday, you ask your flatmate to get you a hamburger.   You are ready after Tuesday's debacle and fill the form out perfectly.  Your flatmate returns with a spamburger.  You are baffled by this as spamburger was not even on the original list of choices, and you don't like the idea much at all.  "Ah yes" your flatmate said.  "You remember that American Hamburger option you ranked sixth?  The one you'd never heard of but you thought it sounds OK? Well, it turns out it's made from tinned Spam, and the five things you listed above it were all unavailable."

On Thursday, you ask your flatmate if there is an easier way.  Well yes there is, he said.  It's called "restuarant preferencing".  Each restaurant has its own ordered list not only of its own meals, but also of all the meals produced by the other restaurants.  You just have to pick a restaurant, and if they have no meals that you like, then they pass your order on to another restaurant. Normally it will be something similar to what you like.  You think that this sounds fabulous.  Tonight you don't feel much like hamburgers so you order a meat pie from Shorty Bill's Meat Pies.  You get a meat pie from Shorty Bill's Meat Pies.  Wow, you think, what a great system this is.  So easy!

On Friday, you decide to try getting a hamburger via this "restaurant preferencing" method, since it worked so well on Thursday.  However, you're in a hurry, and the ordering form is enormous.  You see "Hamburgers Australia" in the first column.  You think that sounds a bit different from the "Australian Hamburger Bar" (AHB) that you would really like to order from, but you can't see the latter on the form at a quick look, and you're feeling a bit hungry and confused.  Maybe they've changed their name or something?  You put a 1 in the Hamburgers Australia box.  Your flatmate brings home a buffalo-burger and points out that your preferred bar was there all along in the 25th column on the right.  You realise that the buffalo actually tastes more like beef than the hamburger you were after, although it does smell a bit off.  You ask Hamburgers Australia why and they tell you the buffalo was shot after straying onto a windfarm, and the post-mortem showed it wasn't vaccinated.

On Saturday, you have another go at getting a hamburger, but hamburgers are not available.  Your flatmate (who now works seven days a week, while you have had to take time off to research restaurant menus) brings you home a bowl of vegan stirfry.  But how is this possible, you ask?  Surely a hamburger restaurant would run through dozens of similarly unhealthy meat dishes in its preference ordering before getting into any of that greenie stuff?  Surely you can trust your favourite burger bar to recommend something similar, just as you would trust your favourite burger bar to cook you a good burger if it could?  No, your flatmate explains.  Your burger bar has done a restaurant-preferencing deal to cross-recommend with Christine's Veggie Hut whenever either of them runs out of meals.  Of course, some customers will complain about this highly unethical deal in the future, but the profits of doing so are so great that neither place cares.

On Sunday, you realise that you cannot get a hamburger without having to either express preferences for 110 different dishes from 42 different restaurants, or else potentially having to eat something chosen in a dodgy deal by your favourite burger bar.  You find out there is a website where you can read the restaurant-preferences of every restaurant in town.  You notice that the local kebab house, your second preference, preferences Australian Hamburger Bar second, and generally their listing of restaurants seems to make sense, with all that veggie rubbish safely down the list.  You pick the kebab house and you get a kebab.  It's not your ideal choice but it's still quite good.  Hey, you think, perhaps I might be getting the hang of this!

Monday rolls around again and it being Monday things are bound to go badly.  You pick the kebab house again but they are out of kebabs, and there are also no hamburgers.  Your flatmate brings you a so-called Aussie Pizza, which contains egg.  Channelling the author of this article, you like pizza but detest egg.  Hang on, you say, wasn't the pizza restaurant's first choice pizza the Volcano?  Ah yes it was, said your flatmate, but so many Volcano pizzas were sold in the first fifteen minutes that by the time your order reached the pizza house, they only had egg pizzas left.

After that disaster, you study the website more carefully before ordering on the second Tuesday.  You decide the Thai place's preference list doesn't look too bad, though you are now giving a first preference for your third option, and they've only ranked hamburgers 17th.  To your horror, your flatmate returns with a vegemite scroll.  But this is impossible, you say, as you were sure that the Thai place had ranked the Vegemite Diner below the lead dishes from every other restaurant.  Ah yes, your flatmate said, this is so.  But after a large number of transfers between various restaurants including a point where your order was randomly transferred between three different choices, a point had been reached where the only choices left were the Vegemite Diner and another place you also couldn't stand.  But the Thai restaurant, while ranking the other place's first option 107th and ahead of both the Vegemite Diner's, had ranked their second option absolutely last on one of their three orderings.  And too bad for you, that was the one that was randomly chosen.

Wow, you think, this is getting really complex!  Trying to pick a meal you can safely select on the second Wednesday, you find that your absolute safest restaurant-preferencing choice is a sardine sandwich, not that you like them all that much, and it's not exactly a filling dinner either.  You therefore choose a sardine sandwich and your flatmate brings back an egg pizza.  Well, it was that or a vegemite scroll after all the other places shut, so at least you got the least disliked of the two remaining options.  Could have been worse ... until your flatmate points out this problem.  The hamburger joint had closed early because it realised it was never going to get your order in time.  Had you instead ordered a hamburger, you would have got one!

Things get even stranger on the second Thursday when you decide it is hamburger or bust.  So you order a hamburger using restuarant preferencing, and your flatmate brings home a stale sausage roll (bad, but better than egg pizza or vegemite scrolls).  Your flatmate explains that if you really wanted a hamburger, you should have ordered with the Vegemite Diner!   An order to the Vegemite Diner would have been pure genius - it would have resulted in Hamburgers Australia shutting its doors early, as a result of which HA would have transferred all its orders to Australian Hamburger Bar, and since the Vegemite Diner luckily ran out of vegemite scrolls and therefore couldn't sell you one, your order would have then been passed on and resulted in a burger.  All so obvious!

On the second Friday it gets odder still.  Three hours after you order a hamburger, your flatmate reports something really strange has happened.  A dispute has sprung up between Squid Rings With Cheese and Porcupine Pasta, two of the deservedly least popular eateries in town.  Some orders were lost and depending on these orders, it is not clear which of these joints should close first.  The dispute is baffling because both of them would have to close later anyway, but each is holding orders from other restaurants.  Which one closes earliest turns out to affect the meals received by thousands of customers in ways no-one could possibly have expected.  Your preferred burger bar tells you you will get a burger whichever way that decision goes, but food authorities announce that because of the confusion all orders for the night at all restaurants are cancelled, even those that would have been filled no matter what.

Come the second Saturday you again order a hamburger.  A strange craze for a kind of yellow ice cream that melts in five minutes has swept the city and your order is again in all kinds of doubt.  But success is yours at last as you see your flatmate carrying home the box marked AUSTRALIAN HAMBURGER BAR.  You are practically drooling with anticipation as you open the box, yep it looks like a burger alright.  You bite into it and find the meat is old and tough and doesn't taste that much like beef at all; in fact it tastes more like a brown paper bag.  You call the restaurant and ask them what has happened, they don't make them like they used to.  "You silly young Pratt" says the burger bar owner, "All our prime burgers are now made from Bullocks."

On the second Sunday you decide that the takeaway scene is awful and the puns are even worse.  You give up, and you make your own dinner at home.

The end.

How Senate Reform Really Works

The really strange thing about the Denniss article is that in fact the main point of the JSCEM-proposed Senate reform is to make Senate voting a great deal more like his analogy than it actually presently is.  Any supporter of the principles underlying Denniss' defence of our system in comparison to first-past-the-post should also support proposed Senate reform in preference to the system we have now.

The reform supported by the Government, the Greens and parts of Labor actually delivers on the principles involved in the dinner analogy to a very large degree.

A voter will be able to vote for their choice of candidates (=meals) in order, provided they specify enough options for a formal vote (normally six).

A voter will be able to vote for their choice of parties (=restaurants) in order, and will be able to stop whenever they reach the point where they no longer care.

If a voter does not know which options to pick, and trusts their party to recommend options for them, they will be able to simply take their party's how-to-vote card and follow it.    If you mostly like your party's card but disagree with one recommendation, you can make your own slight changes, which you can't do without numbering every square below the line under the current system.

The proposed new system is giving voters far more flexibility in voting, and far more control over what their vote does, than the current system.  The voter is not actually losing any choice they have under the current system, save that if they want to follow their party's recommendations exactly, they will have to copy several numbers instead of just putting a 1 in a box.  But that effort is nothing compared to the effort required for anyone who wants to vote below the line in the current system.

The only extent to which the comparison to first-past-the-post is apt is that fewer preferences will be transferred, because more voters will exhaust their preferences (especially while people are getting used to the new system).  This, however, has no impact on the ability of the voter to direct their own preferences effectively without having to vote for a huge list of candidates below the line, an ability the voter currently does not have.

It will mean there are cases (fairly rarely) when Party X fails to catch Party Y on preferences, but would have done so had the voters for excluded Party Z been forced to express a preference between X and Y.  However, that is better than what often happens at present, in which the difficulty of voting below the line effectively forces a voter for Party Z to preference Party X although given the choice they would actually choose Party Y.   There is no perfect solution (because forcing chosen preferences all the way through results in a massive informal-vote rate) but it is better that a party with a clear level of support beats one with much less clear support because voters for other parties don't choose to preference the latter, than for the one with less clear support to win because voters for other parties are forced to preference it.  Even when a preference flow might seem ideologically natural (between like-minded parties), there is no way that every voter would follow it.

There is one other thing I will pick up on here.  Denniss argues:

"Given that the whole point of an election is to trust a party to make important decisions on our behalf, it's not surprising that the vast majority of Australian voters trust parties to allocate preferences and opt to vote above the line"

Voters trust politicians?  No they don't, actually; political parties are among the least trusted things in our democracy (and Senate voting farces are part of the reason for this). As Ben pointed out, it's just not about trust; high above-the-line voting rates are about lack of a practical alternative given the increasing number of candidates and the time taken to get voting below the line right.

But there's another point here.  We have indirect democracy (in which we elect politicians to make the vast majority of political decisions on our behalf) not because we positively trust politicians, but because there is no practical alternative.  In the case of Senate voting systems, we know that we cannot trust even the parties we like to distribute our preferences as they should be distributed.  We do have a practical alternative to trusting them, and we should therefore support it.


  1. Kevin, there's another problem with Denniss' article. His proposal of a 2-4% primary vote threshold per candidate would make elections even worse, as major parties would be forced to instruct their supporters to vote below the line, as if not enough major party voters did, major parties would end up with only their top candidate elected.

  2. Thanks, I hadn't picked up on that before and can only hope this is just a matter of imprecise wording on Denniss' part and that he has failed to consider that threshholds per candidate and threshholds per party are different. He indeed writes "And both yours truly, and Professor George Williams have separately argued for a simple threshold test to be added to the current system in which the existing above and below the line system is maintained but only a candidate who receives more than 2 or 4 per cent of the vote can ever be elected."

    Prof Williams at least has clearly advocated a threshhold of 4% per party or per independent candidate, rather than per candidate. Threshholds per candidate would be an awful idea under any system, but especially under the current one.

  3. This comment has been removed by the author.

    1. It's true that the equal representation of states in the Senate leads to disproportional representation of population, and that's an issue. But it's also true that in order to change this, all states need to vote Yes (Article 128 of the Constitution), and I have a funny feeling that Tasmanians won't vote themselves less power in the Senate.

      A 'quota system' is not what is being proposed here. The 4% quota was suggested to the Joint Standing Committee by the Liberal Party, but was rejected, and the bill introduced by the Greens calls for an optional preferential system as used in the NSW upper house.

      Saying that 25% of the electorate voted for minor parties, and therefore minor parties deserve 25% of the seats, is rather odd. Just because someone votes outside the Liberal/Labor/Green group does not mean that they support all minor parties above those three. Do you seriously think that a voter for the Animal Justice party would rather have their vote go to the Shooters and Fishers over the Greens? Or would a Liberal Democrat prefer a Socialist Alliance senator over a Liberal?

      In most cases, voters for these parties are supportive of certain policies, and want a Senator that reflects their opinions.

      Discussion of which parties would be advantaged, if any, by the proposed changes, is here:

    2. Apologies for deleting my comment Henry. It was poorly formatted (as Kevin surmised) so i resubmitted it. I've replied below.

  4. In fairness, the Australian Motoring Enthusiast Party represents nearly 67,000 voters across Australia, whereas the average Tasmanian Senator represents a little over 20,000. If you're looking for unrepresentative swill the AMEP probably isn't the place to start. That isn't a problem for people like Denniss', who is in effect a paid cheerleader of the Greens. The Greens, of course, receive no disadvantage from the current state based senate voting system. They stand to gain if a quota system is introduced.

    In regards to the proposed reforms - a quota system is a ridiculous idea. This assumes the purpose of the reform is to retain the pretence of a senate voting system based on proportional representation, and not a cynical attempt to lock out independents. Nearly 25% of electorate voted for third parties (outside the LNP/ALP/GRN troika) at the last election and they elected 7 of 40 members, or 17.5%. Even without the LDP's huge donkey/confusion vote in NSW, Independents and Others would still be underrepresented in a Gallagher Index.

    In principle a quota isn’t a terrible idea, but 4% is surely too high and almost certainly arbitrary. A more reasonable quota would be 2.5% in half senate election (1/40), and 1.3% (1/76) in full senate election. At least this system pretends to support PR, without blatantly rent seeking for the existing three players.

    In terms of potential future advantage for the Greens…. It’s hard not to look past the argument David Leyonhelm made on this front. Getting rid of Group Ticket Voting? What are the Libs smoking!? The parties that are elected under GTV are also the ones that poll under 4%, and low and behold, they’re almost entirely to the right… Are they deliberately trying to entrench a permanent left wing blocking majority in the senate? Half senate voting already means the Greens come close to winning 1/6th of the seats with 1/12th of the vote. We already know it’s impossible to win 4/6 senate seats as an individual party. It certainly rules out a future L/NP senate majorities. A combination of a quota and the abolition of GTV will be like signing their own suicide note.

    Further, it's always amused me how it is always the minor parties who are singularly accused of abusing GTV, while the L/NP and Greens get away with it unscathed. The Libs preferencing the Greens above the LDP, Sarah Hanson-Young being elected off the back of preferences from PUP, a party controlled by a coal mining billionaire. They are all as bad as each other.

    1. For anyone confused above, Henry's post was in reply to Ross's post but Ross resubmitted it (presumably to improve the formatting).

      As with Henry I agree that the state based malapportionment is a big problem, but it's also one the small states won't agree to solving so I don't devote much energy to trying to get it changed. Also at least those Senators elected off moderate vote shares in small states are accountable to their populations, in terms of what they do tangibly affecting their re-election prospects. A crossbench Senator elected on a tiny statewide vote knows that their fate at the next election probably won't be decided by their popularity and will be decided by backroom deals and luck, so there is little accountability.

      The Greens have said that they do deals with other parties (including non-Green ones like PUP) because while they are opposed to the current system, they will still use it to avoid being disadvantaged. I think this is acceptable except that the Greens have failed to do their research on exactly who they are preferencing at times - there was a very narrow escape last election from an extremely anti-gay candidate winning in Tasmania on Greens preferences.

      The big advantage of reform for the Greens in my view is that they won't have to make these decisions anymore. More than most other parties the Greens have a support base that is very sensitive to dodgy deals with non-likeminded parties, and it is brand-damaging for the Greens to be seen to be just like the other parties.

    2. I appreciate the point about it being unfeasible to change the constitution. Having a non legal background I naively assumed it could be achieved through a simple (in theory!) referendum (four states and an absolute majority) but that seems to not be the case.

      In reply to Henry's point about minor party support. It's obviously true not all minor parties would support other micro parties. The more nuanced point I was trying to illustrate is the potentially undemocratic nature the abolition of GTV could have due to the arbitrary 14.3% quota. Blaming GTV is putting the cart before the horse. Or more precisely, addressing one problem without fixing another. Minor parties make these preference trades precisely because the 14.3% quota is too high and against the principles of PR. They do it because trade off of seeing a party elected with completely opposing views is greater than never having the voices of their constituents heard. Optional Preferential worked in the NSW upper house because of the lower 4.5% quota. Parties that receive 2 or 3% of the vote aren't locked out.

      Many of the examples used against GTV, like Steve Fielding's election in Vic in '04, are actually reasons why the system worked before systematic gaming. FF got 1.7% of the national vote and 1.3% of the senate seats. In almost all cases, save Ricky Muirs election, and the potentially ridiculous ASP in WA, the minor parties won a seat would have been entitled to anyway had there have been proper PR!!

      I would also add that reasons why people vote for minor parties (including the Greens) may not be as ideologically driven as many assume. Certainly from a Queenslanders perspective, the dramatic drop in Greens support in both the 2012 state election and 2013 federal election, seemed, on face value, to decline at elections that saw unusually large votes for 'other' third parties, KAP and PUP respectively.Without being able to prove that the Greens decline in these elections were a protest vote, i'd point to the fact that 10-20% of Greens preferences at most elections flow to the Liberals, which would suggest at least a portion of the Green vote is a protest rather than ideological one.

      To sum up, I agree the system is being gamed. But the 'solution' is in effect enriching larger, established parties at the behest of smaller ones.. If changing the size and proportionality of the senate is unfeasible, then we need to weigh up whether it is worth the price of consistently disenfranchising 25% of the electorate. Lets remember here, while only a small fraction of people vote below the line, it is an option. Who is the wise person assuming that voters who place their trust in minor parties don't also endorse their preference swapping decisions, on the basis of the trade off previously discussed? It would be interesting to see polling of various minor party voters to gauge whether this is indeed the case.

  5. The 14.3% quota is a consequence of state-based Senate elections, which are, once again, constitutionally entrenched. Any attempt to change this would meet with similar protests to abolishing equal representation of the States.

    I too would be most interested to see polling on the motives of minor parties, and I think that in the cases of the less obviously ideological parties (Sports Party, Australian Voice, Building Australia), you would see a willingness to support other minors. But, under the proposed changes, they will be able to make that choice for themselves, rather than having it made for them.

    It is true that voting below the line is an option, but at the moment, it is an unnecessarily cumbersome one. I maintain that most voters don't endorse the preference swapping decisions of the parties they vote, simply because you have to be a fortune teller to be able to understand exactly where your vote will go and what its effects will be. How would a Western Australian voter know that the placing of the Australian Christians ahead of the Shooters and Fishers on their party's ticket would determine whether ALP and PUP or Sports Party and Green senators would be elected?

    1. My issue here is whether a disinterested Western Australian would actually care (or even endorse) where their vote went, leaving that decision in the hands of the party they chose to trust and support. If a party makes a decision that they are better off swapping preferences with the devil, then so be it. After the last federal election and the accompanied media scrutiny the electorate is more aware of the problem.

      More substantively, if we all agree it will be neigh impossible to change the constitution, then consideration has to be given to the consequences. Are we truly better off with an electoral system where only parties who receive 8%+ of the vote are in a position to contest a seat? I understand the constitutional constraints, but GTV offers a way around this issue.

      Any changes should be outcome focused. To stop tiny parties winning seats off a minuscule primary vote, introduce a low and reasonable quota. 1.3/2.5% seems like a logical and democratic starting point. I just don't agree that the mere fact people assume voters don't know where their vote is going, is a reason in itself to disenfranchise that vote and implement a system which is weighted in favor of the larger parties.

    2. I think people do care about where their votes go. The idea behind the reform is that if you do care where your vote goes, you can make that choice without having to number 110 boxes. Forcing those who do like a party, but don't like its preference ticket to number so many boxes is unfair, and parties shouldn't have so much control over their votes.The new system doesn't disenfranchise these people, it makes it easier for them to choose who they actually want.

      Senators are elected on a statewide basis. No other proportional representation system would give 16.6% of the seats in a state to a party with 2.5% of the vote. We have six senators per state. Personally, I think there should be seven, but the fact remains that parties with very low shares of the vote don't deserve such a large share of the seats on a statewide basis. Parties below 8% simply don't have the support to win one seat out of six, and they shouldn't be allowed one simply because of a deal voters don't understand.

    3. In my simulations of what actual Senate results would have looked like under OPV I did sometimes find that parties could have won with something around 6-7% rather than 8%; it all depended on the size of the remainders of the various other parties. Sometimes there might be four seats filled on whole quotas and then only one party with a remainder much over half a quota, for instance.

      Regarding threshholds, the problem with the present system is not confined to micros winning off very small shares of the vote. Even with a threshhold of 2.5% we could still have unfair cases like SA in which Family First (3.8%). the Greens (7.1%) and the NXG (24.9%) each ended up with exactly one seat. Threshholds would probably encourage a lot of the micro-parties to merge into larger units capable of beating the threshhold, thus reducing candidate numbers, but this wouldn't stop those slightly larger micros from all cross-preferencing each other, nor greatly increase the below the line voting rate. Actually I suspect many uncompetitive micros would continue to run as siphoners for other parties.

      A threshhold would be some improvement but would probably shift the problem from random tiddler micros getting seats they don't deserve to mostly religious-right parties getting seats they don't deserve.

    4. You may just get the opportunity to argue against disproportional state representation if this idea gains traction.... I really don't think they considered the potential ramifications.

    5. While I agree with Henry that voting below the line is cumbersome, allowing optional preferential voting below the line would encourage further uptake. This could be a nice half way proposal. Give a legitimate option to voters to decide how their preferences are distributed, without taking away a parties ability to be elected under the effective 6-8% threshold.

      I briefly entertained the argument that minor left wing and left libertarian parties would merge into the Greens, and minor right wing parties into each other. But on reflection, this is utopian thinking, and will most likely lead to more voter disenfranchisement. Excluding the front and non-ideological parties, the parties that poll above 0.5% (70,000+ votes) mostly exist for a reason (Animal Rights and Sex party supporters may not be attracted to the harder left elements of the Greens, libertarians loath religious parties).

      A lot of the evidence against GTV is based on individual unfair state results. Individual cases where parties lose a seat they were otherwise entitled to win. Had these results gone to the party that would have won under full OPV, minor parties would be dramatically underrepresented nationally.

      I'd be interested to see some of Kevin's simulations under full OPV, particularly in regards to how it will effect senate proportionality when compared to national vote.

      To take one example. NXG is a novel case. It is already benefiting from highly concentrated regional support in a small state (of course it didn't run in other states, but its potential national support is a hypothetical). When you compare national results to seats won, the problem is not nearly as pronounced.

      Sure, it's unfair if FF wins 3.8% of the vote and wins the same number of seats as a party with 24.9% in a particular state. But sharing the spoils with 1.1% and 1.9% of the national vote isn't an issue. Given how close to two full quotas NXG polled in SA, below the line OPV may fix this problem in any case.

      It's also true that no other proportional representation system would give 16.6% of the seats in a state to a party with 2.5% of the vote. I'm not arguing this is right at state level, but if it redresses the injustice of an unreformable national system it may be the best of bad options.

      I'm also not adverse to the risks of consistently giving the balance of power to religious or other minor right wing parties. But disagreeing with regressive views doesn't mean they should be shut out of the political process altogether because of the convenient argument that real PR reform is unachievable.

    6. My estimates for seats won under the proposed JSCEM model if the exact same votes had been cast can be found here: Of course, a different system produces different behaviour. Antony Green did a far more thorough job of the same thing at where he also went into a number of the cases where the outcome might have been different to the actual result, but might not have. I have a different view about some of the One Nation cases because I think there would have been especially strong efforts to ensure preferences were mobilised against them.

      Generally I think it is more important to ensure clearly unworthy parties do not win than to ensure that marginally worthy parties do. If that means making it extremely difficult for parties to get up at half-Senate elections off below 7% then I'm fine with that, so long as the possibility is still there should the voters actively want it.

      On party mergers, I'd expect some of the religious-right parties to merge with each other and likewise for the right-ish usergroup parties.


The comment system is unreliable. If you cannot submit comments you can email me a comment (via email link in profile) - email must be entitled: Comment for publication, followed by the name of the article you wish to comment on. Comments are accepted in full or not at all. Comments will be published under the name the email is sent from unless an alias is clearly requested and stated. If you submit a comment which is not accepted within a few days you can also email me and I will check if it has been received.