13 August: Additions near the end on parliamentarism and the Virginia Plan.
15 August: More discussion at the OTB post by Joyner that is grafted at the bottom of this planting (and in his comment thread).
At the propagation bench to yesterday’s “blogiversary” planting on two-party politics in the USA, James Joyner (whose Outside the Beltway post prompted mine) suggests that PR might seem good on grounds of “fairness” but that he has often thought PR to be impractical because of the “tail wagging the dog†result.
Fairness to cooky voters like me who favor nutty little parties is all well and good. But the ‘fairness’ argument is the less compelling case for PR. The compelling case is in its aggregate effects on representation, accountability, and governance, not on how nicely it treats society’s cranks of the far left or right or whatever ideological fringe.
But what about the supposed “tail wagging the dog” problem? That’s a systemic issue (the “dog” being the system), and if the problem is real, then we should be very cautious about PR indeed, for we certainly do not want to empower the tail.
So, do small parties have “disproportionate” influence over coalitions that larger parties must form in order to govern? The actual evidence that this happens is, well, nil. I’ve talked about that here (especially in the New Zealand and Germany blocks), as have many of the propagators over the past year.
The best single paper by a political scientist on this question that I can think of is:
McGann and Moran, “The Myth of the Disproportionate Influence of Small Parties in Israel” (the link takes you to the abstract and another link there will get you to a PDF of the paper).
Basically, there just is not much evidence that small parties get more than their weight in votes would entitle them to, nor that they are able to hold “hostage” the bigger parties (which, after all, are also minority parties that get, by definition, disproportionate influence under plurality elections!). And if it does not happen in Israel (where the largest party often has only a third of the votes and seats, the country is a single 120-seat district, and a party can win a seat with just 2% of the vote) it is unlikely to happen almost anywhere.*
Small parties may be needed for coalitions, but they need the larger parties just as much–and often more–in order to exercise any influence and to be able to bring any policy or other rewards back to their voters.
An additional factor in limiting small-party “extortion” ability is a simple fact about multiparty systems: For almost all voters, there are more parties close to the voter in the ideological space than is the case in a two-party system. In other words, small and large parties alike are always under pressure from competitors, and if they overbid, they can be shut out and replaced by some other party that a chunk of their constituency likes almost as much (and perhaps more, if it can actually deliver). Those of us who think competition is an inherently good thing–in the markets for goods and yes also in the market for policy ideas–thus like PR. More competition, more prospects for satisfying the greater number of voters.
In two-party systems, on the other hand, the problem for a voter who wants to punish one party was well summed up by Henry Droop in the passage from a longer and rich paragraph that I am so fond of quoting (so fond that the beginning of it is up there on the left sidebar). Droop notes that the problem with majority or plurality systems and just two major parties is that moderate, nonpartisan voters:
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.
With PR, given that there are almost always multiple possible majorities that can form–the essence of Madison’s famous argument about checking “factions,”–it is much harder for any one party to push its advantage too far. If it does, it risks breaking the coalition. This, coupled with the presence of multiple parties competing to please unafiliated voters, gives such voters a voice between elections, something recent experience in the USA clearly shows is lacking (as I wrote about last November).
James raises a valid point about fragmentation in the USA currently and how it would be reflected in a PR system. Actually, that fragmentation of interests is one of the best arguments for PR: With multiple parties, rather than just two (at best–increasingly districts and states are not even two-party competitive), these interests are reflected in a way that is transparent. Under our current winner-take-all systems, interests are reflected by individual members of congress who are accountable to no one outside their districts, and given safe seats, not even clearly within them. It is much better that voters and politicians alike be able to see–from shifts in the votes for Greens and Libertarians and so on–which direction the electorate is moving in. Two-party competition–and again, in fact, we do not even have that in many places–is an awfully blunt instrument. So, numerous interests are represented, but in a nontransparent and non-accountable way. And many interests aren’t represented at all.
Then there is the question about presidentialism. Of course, multiparty politics works best with a parliamentary system–to keep the coalition partners that run the executive accountable to the people by way of the proportionally elected legislature. We might have had a parliamentary system if Madison’s original Virginia Plan had not been thwarted by the small-state delegates. (The plan called for the House to elect the executive–and also the Senate, with the latter election based on nominations from the state legislatures, each state represented by population.) Alas, we did not get parliamentarism or weak bicameralism–the best systems for multiparty politics. We wound up with presidentialism along with strong bicameralism.
Looking on the bright side of the not-so-Great Compromise at Philadelphia, an advantage of presidentialism in a multiparty system is that the chief executive does not “fall” when coalitions shift. However, that is also a disadvantage, inasmuch as it limits the range of possible alternative coalitions that can form and rules out the possiblity of early elections to refer interparty disputes back to the electorate. Besides, as I alluded to above, the role of small parties in forcing such crises in parliamentary systems is much exaggerated anyway.
Basically, presidentialism would be almost certain to keep Democrats and Republicans as the major parties even under a House (and ideally, Senate, but that’s more complicated) elected by PR. Smaller parties ideally should be given a role in forging coalitions for presidential elections, which is hard to do with the electoral college. But even if the electoral college were not abolished, the National Popular Vote concept would be highly likely to generate pre-election coalitions for presidential competition–and especially if PR were also adopted.
One concern I have with the NPV idea (which, basically, would have states collectively having 270+ electoral votes agree to give their electors to the national popular-vote winner) under the current congressional electoral system is that third parties would expand their influence in presidential elections without a corresponding influence in congress. For multiparty politics to work, it has to be in both branches, consistent with the Madisonian incentives for interbranch cooperation upon which the entire edifice of separation of powers is premised. (Already, third parties are more active and receive more votes for president than for congress, though the gap is not as great as it once was, and in 2004–unusually for the USA–third party voting was higher for House and Senate than for President.)
Would US Presidents appoint coalition cabinets if neither their nor the main opposition had a majority in Congress? I do not know. Behavior in other presidential systems is mixed on this point. It would be advantageous to them to do so, but they would retain the right not to do so. But just eliminating the single-party Leviathan in control of the House of Representatives (and, to a lesser degree, the Senate) would be in itself a powerful advantage of PR: No more minoritarian governance of the legislative bodies.** It would become genuinely majoritarian, in that parties collectively representing a majority of the electorate (and less regionally biased, too) would be in control of the production of legislation.
With a direct vote for President and PR in Congress, you come pretty close to a best-of-both-worlds scenario (or at least as close as one can get under presidentialism): Pre-election coalitions among parties to elect the president (and these coalitions could shift from election to election), and inter-party bargaining in Congress between elections to keep the partners accountable to their respective constituencies.
Footnotes
* What about Italy? I think the evidence is also limited there, though I am aware of no study of precisely this phenomenon. However, there were various factors–more or less unique to Italy–that might have made Italian small parties more capable of making high demands than elsewhere: the absence of a feasible coalition centered around any large party other than the Christian Democrats (because of the large Communist party), the high internal factionalization of the largest party, and the presence of secret voting by MPs on most matters in parliament. Of course, since 1993, Italy has abandoned PR (and, no, it did not return to it in April, 2006). The peculiar form of a mix between overall majoritarianism and intra-alliance PR that both Italian electoral systems since 1993 have consisted of probably–and ironically–gives small parties far more blackmail potential than is the case under almost all PR systems.
** Whereby the leadership of the single party in control–which can even be the second most popular party nationwide–often refuses to bring to a vote a bill on which there is majority support in the public and even in the chamber itself, but which divides the party internally.
Does the Tail Wag the Dog in Proportional Representation Systems
In response to my TCS article and blog post yesterday on why the US has maintained a Democrat-Republican two party system since 1860 and likely would for the foreseeable future, UCSD political scientist Mathew Shugart, co-author (with Rein Taagepera) o…
Scion grafted by Outside The Beltway | OTB — 11 August 2006 @ 12:03
What do you think about minimum thresholds as a way to reduce the total number of parties to a manageable level? My thinking about electoral systems has always been a bit vague, but I always thought that MMP with a 5% or 10% threshold might work. Giving a party parliamentary representation as long as it gets a portion of the popular vote equal to the reciprocal of the number of seats maximizes fairness to those with odd political preferences, but to my inexpert mind seems like a recipie for trouble.
Seed planted by Pithlord — 13 August 2006 @ 08:50
[revised] A 10% threshold?? Way too high! Actually, I think 5% may be too high. Why have ‘PR’, and then make the threshold so high as to guarantee a lot of wasted votes, and hence substantial deviation from proportionality?
The propagators and I have had some discussion about this matter. For example, the thread on the Ukrainian election (3% threshold with over a fifth of all votes cast below it and thus wasted); and on the Czech election (where the wasted votes, although onlly 6%, may have contributed to the tied seat result between the two blocs).
I don’t know if there is such thing as an “optimal” threshold, but 3% or 4% seems reasonable to me; however, there are examples even of such “low” thresholds resulting in very high levels of wasted votes (such as the Ukraine case mentioned above). I’d prefer even lower, though for practical political reasons, PR adopted in a jurisdiction with a history of plurality systems will generally have thresholds of 4% or 5%, I would expect.
In MMP, as opposed to pure list, you tend not to have a lot of little parties anyway, unless parties that fail to meet the threshold are allowed in by winning only one SSD (as in New Zealand). I would advise no such “loophole,” though I recogize that the politicians who designed the details of the NZ-MMP system would never have gone for any system without that provision (except for the Greens, all the small parties in post-reform NZ have been parties formed around an individual MP).
Seed planted by MSS — 13 August 2006 @ 15:19
As you probably know, Turkey does indeed have a 10% hurdle, which in the last election resulted in the largest party getting 66% of the seats with 34% of the vote. Which sort of makes a mockery out of PR.
A decision on the “proper” level for a hurdle is of course a political question, and in Turkey the hurdle is presumably set so high to prevent the Kurdish minority party from getting in (independent candidates are exempt from the national hurdle, and a handful of Kurds got elected this way, as well as, one would think, through the lists of the two parliamentary parties). I assume the hurdle could be altered in connection with future EU membership talks.
In the US a 10% hurdle, even on a local level, would be a tool for keeping the two-party system solidly intact.
Seed planted by Espen Bjerke — 14 August 2006 @ 03:17
“In the US a 10% hurdle, even on a local level, would be a tool for keeping the two-party system solidly intact.”
Exactly right, Espen.
Seed planted by MSS — 14 August 2006 @ 09:29
If Ontario’s Citizens’ Assembly on Electoral Reform designs an MMP system next March and April, they will have to decide on a threshold, unless they use the Scottish finesse — a uniform district magnitude with unlinked districts so as to have an effective threshold without saying so. Uniform district magnitude in Ontario will not be as easy as in Scotland.
Thresholds have always been set by self-interested politicians, such as Turkey’s very undemocratic 10%, and the lengthy failed negotiations with the FDP in Germany to drop the 5% to the Scandinavian norm of 4%.
What will independent citizens make of this debate?
In Canada we have two precedents: BC and Quebec.
The BC CA, before it gave up on designing an MMP model, voted on the threshold for its stillborn MMP model. Some wanted no threshold at all. Some wanted the classic 5% of New Zealand and Germany. The moderator asked them to vote on 5%, 3%, or no threshold. Predictably, 3% carried the day on the second ballot, 4% and 2% not being on the table.
The Quebec Citizens Committee may have had a similar debate, but their report was squarely for a 5% threshold, necessary since they wanted province-wide proportionality. Context is important here. They were trying to improve the hideous “draft bill” tentatively proposed by the Liberal government, which had Scottish-style unlinked regions on five MNAs each, an effective threshold of 16% or so, the highest in the MMP universe. So they had tactical reasons not to propose anything too radical.
No one can tell what Ontario’s citizens will say. However, the Greens got 4.5% of the vote nationally in January, 4.7% in Ontario. They mostly say they are confident they can live with 5%, since PR would surely result in more Green votes. The peak of the Family Coalition Party (anti-abortion) support in Ontario was 2.7% in 1990. Most people don’t seem to want a “pizza parliament.” I can see a choice of 4% or 5% being a plausible threshold.
Seed planted by Wilf Day — 14 August 2006 @ 11:21
Scotland’s MMP system’s threshold is indeed implicit in the use of regions for electing list members, but it is relatively low – in 2003 MSPs were elected from the lists with just over 5% support in that region (and less than 2% nationally). In Wales the regions are much smaller and the effective threshold higher – the Greens were nowhere near winning a seat in 2003 with 4.8% in their best region. In London’s 25-seat Assembly there is a threshold of 5% (which should make little difference, but in 2004 did what the government hoped and prevented both the far-right BNP and the left wing/ Islamic Respect from winning seats with 4-5% of the vote).
I wonder if the ‘tail wags dog’ phenomenon is most acute with Independents accountable to a local district rather than parties with a wider mission? A simple trade can be done giving the Independent’s constituents favourable treatment while permitting most of the government’s agenda. The clearest case of this was the ‘Gregory deal’ in Ireland in 1981/2 but it can happen under FPTP when a party is just short of an overall majority (as in 1976-79 and 1994-97).
Seed planted by Lewis Baston — 16 August 2006 @ 07:18
Could “…the ‘tail wags dog’ phenomenon [be] most acute with independents accountable to a local district rather than parties with a wider mission…”
Yes, Lewis, I think that is precisely correct. More generally, the narrower the constituency, the more an agent (party or individual legislator) is likely to be a high-demander. And it is hard to get narrower than one MP who was elected in one district (especially if with well under 50% of that district’s vote).
I think very few parties can be said to have constituencies that narrow.
On the other hand, governments might sometimes prefer to make the kind of constituency-specific deals that you mention, rather than compromise their overall program.
Seed planted by MSS — 16 August 2006 @ 09:51
“Governments might sometimes prefer to make . . constituency-specific deals . . rather than compromise their overall program.” If the “Gregory deal” is the example, it turned out to have wide enough repercussions to compromise a lot. First, by raising expectations it quickly led to the election of 1982 election of the Fine Gael/Labour coalition. Second, it started the auction of promises that then led to such a counter-reaction that, after the 1987 election, both major parties agreed to massive budget cuts and tax cuts. Third, it whetted the public appetite for constituencies to trade for favourable local deals by electing independents. Until the Gregory deal very few independents had been elected, but in 1987 four won, and by 1997 it was seven, four of whom replicated the “Gregory deal” for their four local ridings by forming the “Gang of Four” which met as a quasi-caucus, held the balance of power, and negotiated such good deals that, in 2002, 13 independents were elected, all from constituencies hoping to ask for more, only to be disappointed when the Fianna Fail/PD coalition won a clear majority.
So now STV has the reputation of being a system which elects lots of independents, which it never did in Northern Ireland or Tasmania, or indeed in the Republic of Ireland until the Gregory deal. I would argue it compromised their whole political culture, which was already strongly oriented to local clientelism within parties, towards non-party activism.
We’ll see at the next election how lasting this impact is. I’m expecting a large drop in the number of independents, but how large, I don’t know.
Seed planted by Wilf Day — 19 August 2006 @ 20:49
Leaving aside for a moment the issue of transparency, perhaps, in the way reality almost always trumps theory, we have already achieved something of proportional representation. Meet with any Democrat about tort reform, and see how long it takes for someone to ask someone else “What will the trial lawyers say?” Do we have a “trial lawyer” party? Do we need one? If we had one, wouldn’t the other parties be competing for its votes on tort reform? What would be different?
When significant legislation is passed, it’s because the majority party has persuaded a few folks from the other side of the aisle to come with them. But what does it mean, in today’s Congress, for the Democrats to “pick up a few Republican votes”? Aren’t those votes typically of a bloc, like the farm states or the rust belt? Do we need an agrarian party in order for the interests of that “virtual party” to be served by its power to enter into a “virtual coalition”?
Even the apparent counterexample of “party-line” voting is deceiving. Party discipline is not automatic; it is engineered by good leaders (and, yes, enforced by brute politics – ain’t nothing perfect). Why do all of the Democrats support S301X? Because those who don’t support it can count on the support of other Democrats on H301Y. Maybe a hold-out gets a subcommittee chairmanship to close the deal. Now his virtual party will get a thing it wants through the skillful wielding of the associated gavel. Leaders build coalitions every day. It may not be as efficient as the prepackaged ones in PR systems, but we’re rich enough to afford some inefficiency.
Prof. Shugart speaks of the better matching of party ideology to voter ideology in a PR system, but what “ideological” issue is not just the first or second derivative of substantive interests? If I want high corn prices, I can argue for subsidies, or I can argue for disproportional representation of people who want high corn prices, or I can argue for a system that happens to assure the disproportional representation of people who want higher corn prices. When we mind the pennies, the dollars only seem to be taking care of themselves. In reality, we only mind the pennies so that we will have the dollars.
At the end of the day, I want a system that protects my life, my liberty, and my ability to pursue happiness. Those are all substantive, political goals, and they are not shared by everyone. Some would disadvantage me with respect to any of these things if that would help them, for example, to hold on to happiness that they have already acquired. It is much easier to hold onto happiness than pursue it, and the holders are often at odds with the pursuers, at times attacking the latter’s life and liberty to dampen the pursuit. But it is wrong to claim an “ideological” basis for these stuggles – they are material struggles, distilled into ideology for the useful purpose of suggesting to the haves, for example, that they are better served by a minimum wage that creates customers or a political system in which former leaders are not ipso facto dead.
So we come back to the plasticity of politics in a free society. Maybe our “non-proportional” system survives because it is existentially proportional. Caucuses replace minority parties, temporary voting majorities as granular as one bill at a time replace coalitions. But interests get adjusted. And the interests that are not effectively represented fill the role of those minority parties in parliamentary systems, parties whose impotence Professor Shugart is at pains to demonstrate anyway.
Seed planted by Lawrence Kramer — 31 January 2008 @ 11:41