Two parties are not enough

That subject line will not surprise anyone who has spent time on this blog over the years. The need for a multiparty system (and electoral institutions to support it) in the USA has been a consistent theme here since the very earliest green shoots of the virtual orchard.

Two recent essays by Seth Masket demonstrate the point very well, even if it was not the author’s intention (and I see no indication that it was).

What broke the Republican Party? Masket notes just how different the GOP of Romney and Trump are, and concludes that “Those in the new GOP no longer see the Reagan-Bush Republicans as members of the same party.”

The Party of Self-Doubt. Masket reviews the recriminations between progressive and moderate wings of the Democratic Party over an election win that felt more like a defeat in some ways, due to more limited success in non-presidential offices than expected.

These essays, taken together, are excellent demonstrations of the tension in trying to cram different tendencies into one of two large parties. In most democracies, these intra-party groups would be separate parties. And if they were, voters could actually choose among them in general elections, and thus shape the direction of whichever left or right (or other) camp wins leadership of the executive branch, and of the majority coalition in control of Congress, at a given election.

Perhaps even more importantly, voters who dislike the currently ascendant tendency on their side of the wider left–right divide would not have to cross over to the other side in order to punish the incumbent. This is not a new idea, of course. I will leave you with a quote from Henry Droop who noted this problem in his critique of two-party politics and the electoral systems that support it.

With majority voting they can only intervene at general elections, and even then cannot punish one party for excessive partisanship, without giving a lease of uncontrolled power to their rivals.

A longer excerpt and context is offered at this blog’s Droop page. Oh, and Droop observed this a century and a half ago.

27 thoughts on “Two parties are not enough

  1. In a similar vein, the Washington Post’s Jennifer Rubin divides the Republican Party into a “right-wing nationalist” caucus and a “reality” caucus and writes, “The longer the Reality Caucus members fool themselves and remain in an alliance with Right-wing Nationalists, the more they enable and empower the latter, further fraying the fabric of constitutional government. It is time for the Reality Caucus to make a decision and choose its strategy. The status quo is unsustainable.” (See ) Like almost all journalists she doesn’t acknowledge the role that institutions play in keeping the two “big tent” parties sustainable.


    • Good that she said this. On the institutions point, I don’t know how to feel about its exclusion. I mean the typical op-ed statement would be something like “it is against Duverger’s LAW, so we can’t have a party split.” (I am exaggerating only a little.) So I’d almost rather see the idea that you ACTUALLY CAN FORM A NEW PARTY be put out there as a legitimate course of action.

      Then we can have a deeper conversation about institutions–and reform!


      • I see what you mean. Your point is consistent with the observation that changes in the party system more often precede electoral reform than follow from it.


  2. Open party list, single transferable vote, or MMP? Perhaps a US state should experiment with one of the above. Populists can win no matter what system is used.


    • Any of the above. I am fine with states trying out different things for their own legislatures, although I’d think that it would be preferable to have some considerable uniformity for the House or US Senate.


  3. Sure, two parties are not enough – but I really don’t care about different groups being crammed into the same parties, as such. It happens even when moderates have the upper hand in both parties (which I think used to be the case in the US). My problem is polarisation and the high stakes and lack of accountability they create. Droop’s comment relates more to this concern.

    As a sidenote, Droop was talking about efficient parties, not about the unhierarchical parties of the US. I think a big take-away from the current situation in American politics needs to be not just the value of third parties as such, but also how primaries do nothing (at best) to help parties stay moderate.


    • I can’t disagree with that. I also take Droop’s point to be about polarization. And further to be that either third parties or competing factions within parties (i.e., at general elections) would provide moderates with a voice. It is true he did not address primaries, and conditional on having PR, I’d prefer to get rid of them. But PR in a world with primaries still addresses Droop’s (and my) concern, right?


  4. Alaska passed Measure 2. It provides, among other things, for preferential voting. It will be interesting to see how it works out.


  5. 2 party domination worked better in the US when it was harder for it to tilt to one-party domination. IL’s use of 3-seat PR with a droop quota for its state representative elections made one-party domination nigh impossible in the economically important state of IL for over a hundred years.

    I don’t see why it’d be bad to still tend to have two big major parties if neither can dominate and campaign finance regulations that were well enforced and adaptable made them both tend to serve closer to their respective median party-voter. Also, if there were minor parties trying to bring about a DFL-like merger with one of them, and local third parties(LTPs) either raising up neglected issues or new frames for wedge issues or being vents for extremists/racists from the major parties…

    There tends to be 1 or 2 big parties in any democracy if LTPs successfully made the key issues more dynamic / unpredictable then it’d be hard for a faction within either of them to claim dominance with the promise of enabling the party to dominate.


  6. Let’s imagine that the people of California decide on a bold experiment. They pass a constitutional amendment providing that for 10 years:

    The assembly would have 80 members elected by MMP.

    The senate would have 42 members elected by statewide STV with half the senators elected at every general election.

    After 10 years there would be a preferential ballot with the options of:

    MMP for both houses
    STV for both houses
    Reverting to FPTP for both houses.

    If nothing else we would finally have an answer, or at least a suggested answer, to the perennial MMP/STV debate. That answer would be given by an electorate that had lived experience of MMP, STV and FPTP.


    • Pretending for a moment that you are serious, I hope your MMP model would have open lists (with or without party option, also known as flexible lists). Perhaps with four regions like 15, 15, 35 and 15. Otherwise I wouldn’t vote for it, and I doubt the lived experience would be a good one.


      • I know it is not my model being asked about, but I maintain that the evidence is that closed lists work fine with MMP, and you did not need the additional complication of open lists. I also fully concede that there are jurisdictions where closed lists, even as part of MMP, are not politically feasible. I have no principled objection to open-list MMP (and have written about that on this blog before).

        Note that as most of us political scientists use these terms, lists are “open” only if there is no party-given default list ranking. What makes lists “flexible” is not whether there is an option for the voter to give a list vote instead of a candidate-preference vote. What makes lists flexible is the existence of a default pre-election order of the list, which prevails in case there are not enough candidates with whatever quota or threshold of personal votes the law specifies as necessary to give them a priority in claiming the party’s seats.

        Finally, Wilf, seriously? You would vote against an MMP proposal if the lists were closed, or if compensation were statewide rather than regional?


        • MSS asks “You would vote against an MMP proposal if the lists were closed, or if compensation were statewide rather than regional?”

          Well, at the system design commission I would. Fair Vote Canada has, since 2013, not supported any proposal with closed lists. Bavaria has seven regions because state-wide open lists are not feasibly accountable. And closed regional lists were not recommended to be continued in Scotland by the Arbuthnott Commission. Would I vote against MMP with closed state-wide lists? They work in New Zealand with under 5 million people, but not in California with 40 million, and were unsaleable in Ontario with 14 million. Having been there and done that, I wouldn’t donate a cent to that Vote Yes campaign.


      • The point of the bold experiment would to get an electorate with actual experience of both systems and then see which system they preferred.


  7. If Australians were made to choose either AV or STV for both the House and the Senate would that tell us which was the better system?


    • Most Australian electoral systems are path-dependent. They are an artefact of history rather than design. The exception is the ACT. When federal parliament created the ACT legislative assembly some form of proportional representation was inevitable, because single-member districts would have resulted, at most elections, in an assembly monopolised by the ALP.

      They experimented with the disastrous modified d’Hondt system which emerged from a series of compromises between the major parties. A series of public consultations and referendums resulted in a decisive majority for STV. The final referendum, between single member districts and the Hare-Clark version of STV showed that 65.30% of the electors voted in favour of a proportional representation (Hare-Clark) electoral system. That is reflected in the Proportional Representation ( Hare-Clark) Entrenchment Act 1994.

      The ACT is peculiar. It’s the most educated electorate in the country. It is the most progressive electorate. They had recent experience of FPTP in federal actions, the modified d’Hondt system, and non-Hare-Clark STV in the senate. The major parties were relatively quiet in the referendum campaign because of embarrassment over the failure of the modified d’Hondt system. It’s an open question whether the rest of the country would follow their example, but there’s at least a reasonable chance they would.


      • “recent experience of FPTP in federal actions” = “recent experience of IRO-AV in federal actions,”
        Canberrans had also used some form of non-Hare-Clark STV their advisory House of Assembly in the Seventies (9 members from each of the ACT’s two federal House divisions). I have had a very hard time finding any details about this other than that independent MHAs were replaced via direct party nomination (this was before statutory registration of parties was even introduced… they were treated as private clubs with limited legal recognition, as in company law) whereas independent MHAs were replaced by countback.
        When the ACT H of A was established in the 1950s it consisted of a grand total of 3 members, chosen by territory-wide Multiple Alternative Vote, like the then recently-replaced 1918-48 Senate system (elect one by AV majority: restore all the defeated candidates; elect a second by AV, rinse and repeat) still at the time being used for the SA upper house.
        Canberran residents are not just better educated formally but have a stronger culture of critical thinking. Having lived there and also Queensland I was struck by the difference in epistemic processes. Queenslanders, over coffee: “Someone loaned me a tap that a friend of a friend gave them, warning how the Hawke Socialist Government is planing to set up separate Aboriginal States by 1988. So we’d better vote no in the referendum to recognise local government.” Canberrans, over coffee: “The Minister made these claims in Parliament yesterday, and it’s immediately striking what she DIDN’T mention.”
        That said, Canberrans have (puzzlingly) no idea how to merge traffic when entering a motorway.


      • I’ve lived about half my life in Queensland and half in New South Wales. A couple of years after my escape from the Deep North I spent Christmas with my parents. They had a view of the Brisbane River at Auchenflower. On Christmas Eve a parade of pleasure boats with coloured lights went up and down playing seasonal music. a friend of my parents waved at this pleasing spectacle and said that of course we had nothing like that in Sydney.

        The Royal Australian Navy had recently celebrated an anniversary with a grand fleet review of RAN and allied ships. I was still young and brash enough to answer that in Sydney we did it with aircraft carriers.


      • I haven’t yet worked out a consistent metric for quantifying the Alan/ Tom Score of any given F&V thread, but one that starts with “extremists in the GOP are crowding out moderates” and ends with Canberra drivers and Auchenflower river boats must, if only in ordinal terms, be one of the highest scoring.


      • There’s always been a strong tendency in bicameral systems to elect the second chamber on a different basis. The constitution mandates that all states except Queensland vote as a single electorate. The Queensland exception is probably now spent, but there is some dispute. It’s never been used.

        The senate was elected by something called block preferential before 1948 which almost always ensured a government landslide. The 1946 senate election result was 33 Labour, 2 Liberal and 1 Country 1948 Labor was facing defeat in a general election and legislated for STV as an act of high principle or to avoid a senate wipeout, depending on why you talk to. The idea of an upper house elected by PR is now part of the country’s DNA. The only exceptions are Queensland, the ACT and NT, which lack a second chamber, and Tasmania, where the legislative assembly is elected by STV.

        The proportional upper houses generally enjoy much higher standing than the majoritarian lower houses.


      • Though how it actually came to be involved more history than design, as Alan noted, I recently learned that the AV-lower house + STV-upper house was raised as a possibility during the negotiations over Federation and actually put before the first Federation Parliament, by the government, no less (surprisingly not even AV was adopted despite the presence of three major parties, the unsustainability of which under FPTP not being something people were oblivious to).

        Does someone know more about these somewhat prophetic early proposals (specifically, the PR for the upper house in combination with a majoritarian lower) and what arguments were made for them?


      • The 1891 draft came very close to providing for popular election of the governor-general. The idea was enthusiastically promoted by Sir George Grey, a New Zealand delegate who had been twice governor of New Zealand, governor of the Cape Colony, and then prime minister of New Zealand. Some unkind delegates suggested Sir George may have had some idea of who would make a great governor-general.


  8. Perceived political advantage for the party proposing the change. (The same reason Winston Churchill, once back in the bosom of the Conservative party, was against AV but OK with STV provided it was only in LIb/Lab strongholds.)


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.