In Search of a Theory of Three Party Systems
As I understand it, party competition in mature democracies is supposed to converge around two major parties, one on the center-right and one on the center-left. The reason is quite straightforwardly strategic: if either party strays too far from the center, it gives the other party the opportunity to seize the center and with it a majority; and, if either party splits or faced serious competition for votes from a new upstart party, vote splitting would doom both parties on whichever side of the ledger they happen to be on to perpetual minority status.
Obviously, you can have periods where there are multiple parties competing to become one of the two stable alternatives, but these should be relatively short transitional periods, not the normal course of affairs. And obviously as well this analysis works better for some democratic systems than for others. It makes more sense for American and British-style systems than for proportional-representation systems, for example. And it works better when you don’t have a regional or sectarian or other identity-driven party throwing a wrench into the works. But it’s a sufficiently thoroughgoing assumption that whenever a political system doesn’t conform to this assumption, it’s treated as an interesting exception.
But the exceptions are starting to devour the rule.
Canada just had an election whose results might be described as “what was supposed to happen in Britain’s last election.” That is to say: the third party came in second, the NDP substantially overtaking the Liberals for the first time in history. If the assumption that two-party competition is “natural” is to hold, the Liberals should fold, leaving two major parties (Conservatives and NDP) competing for Canadian votes. But for reasons of history if nothing else, this is unlikely to happen any time soon.
But apart from history, it’s worth pointing out that if the standard assumption of two-party normalcy were true, then the NDP would never have gotten off the ground in the first place. Ditto for the Liberal Democrats in Britain, who have been affecting the outcome of British elections for over a generation. Further afield, France, in spite of having a Presidential system, has nothing resembling two-party competition. Israel has a proportional representation system, so it’s a different case, but for a long time it was functionally a two-party system, with a right-wing bloc of parties competing with a left-wing bloc for the majority of voters, and with identity-driven parties for the ultra-Orthodox and the Arab vote that were open to any coalition (in the case of the ultra-Orthodox) or systematically excluded (in the case of the Arab parties). But not anymore. For over a decade, Israel’s political system has been rocked by the emergence of centrist parties aiming to be alternatives to the right and left. These parties have never won an election, but they haven’t gone away, and neither have Likud (even when it shrunk to a tiny rump after the first Netanyahu government) nor Labor (the former center-left powerhouse now dwindled to near-irrelevance and sitting in a right-wing coalition). Should I go on and talk about Mexico? Belgium? Switzerland?
What’s the current political science take on this? Why are third parties much more durable, under a variety of political systems (though not the American), than theory would suggest?
I think you’ve got the theory a bit wrong. <a href=“http://en.wikipedia.org/wiki/Median_voter_theorem”>Median voter theory</a> is respected and useful. However, I think most of the talk about the two party system refers to studies of American politics rather than some sort of universal rule. In the U.S. we’ve got both winner take all voting and a strong President, so median voter theory is more likely to hold.
I wonder if some of the two party systems world wide were more a result of cultural influence from American and the UK rather than necessarily the natural result of their electoral systems. If that’s the case, it might help explain why you’re seeing such a breakdown that can’t be traced to changes in constitutions.
— Gregory Sanders · May 3, 02:54 PM · #
France has a strong Presidential system, but they don’t even have a clear two-party competition at the Presidential level. And Britain and Canada both have winner-take-all constituency-based voting. And the United States even has a region with a unique culture and history of separatism. On some level, it’s kind of peculiar that the Dixiecrats didn’t become a stable regional party that threw their support in the Electoral College and for Speaker of the House to whichever major party bid the most for their support. Right?
— Noah Millman · May 3, 03:20 PM · #
I don’t think there is anything wrong with the theory as far as it goes. But there are a lot of tweaks that can make third parties more viable. In the U.S. the party that wins the legislature is not necessarily the one that wins the administrative branch. So it’s not so easy to have a minority president. If the Prez was chosen from the party that got a plurality in the legislature it might be different.
When one of my sons was working with the Workers World Party, I was made aware of hurdles that have to be overcome just to get on the ballot. Those can keep third parties down and out of sight. Lowering those hurdles can give third parties more influence.
I have sometimes voted for 3rd party candidates who didn’t have a chance of winning, mainly because I didn’t want any of the candidates to win, and didn’t want any connection with those who would. Yet I want the vote count to be as high as possible, so the people in office will be aware of the need to watch their backs. I am not aware that very many people use that rationale when voting. You’d need a LOT more people who think that way to give 3rd parties much of a chance.
As to the situation in Canada, I don’t know. (I went to Drudge to find the results. The left-wing version of Drudge (Google News) seemed to have been ignoring it on its front page, at least when I looked.)
— The Reticulator · May 3, 04:40 PM · #
“As I understand it, party competition in mature democracies is supposed to converge around two major parties, one on the center-right and one on the center-left.”
Not really, that ALSO presupposes that you have a unimodal distribution of voters on the ideological dimension.
Of course, Duverger’s law suggests that even if you don’t have such a distribution, you might still have a two party system, provided you have the first-past-the-post system. (Since it favors formation of two large political voting coalitions due to the possibility of “wasting” your vote on a minor party that will not stand a chance. Think Nader in 2000. Also, note Nader’s vote total in 2004, after the whole Florida business four year before.)
In countries with multimodal distributions (or where you can’t neatly collapse the multidimensionality of political issues) AND with proportional representation, you’re going to have political system with more than two relevant parties.
— Marko · May 3, 08:17 PM · #
There are different models that you can apply that will actually, in an efficient system, create a certain number of parties (or permanent coalitions). Without proportional representation, the number is generally two. Yet there are situations as in Canada and the UK where you have more. That’s because the theory implies an efficient system. The Canadian system is inefficient, as evidenced by a Conservative majority even though the two more liberal parties garnered considerably more of the vote. Of course, the previous Liberal majority benefitted from the other: divisions among the right, between the Progressive Conservatives and Canadian Alliance Parties. When those two merged, you had an efficient system again until the Liberal Party destroyed itself. The UK system is also somewhat inefficient, with the two center-left parties winning a plurality for the center-right party (and the curious expectation that it would be improper for the two center-left parties to form a government on their own). So yeah, when voters on one side or the other simply cannot unite, and when they act inefficiently, third parties are born.
The vast majority of France’s national assembly belongs to one of two parties. The main difference in France is that they simply don’t have a strong party system, so parties come and go with greater frequency. Different names, but pretty similar allegiances. Their minor parties are tyupically in regular coalitions with the main party. It’ll be interesting to see what happens in Mexico, with two left parties benefiting a united right one. Will this act out in perpetuity?
Marko also makes a good point, which is that it presupposes a continuum. In Canada’s case, even setting aside the NDP, you’ve always had the Bloc. That’s because you have a cluster of voters with interests (perceived to be) independent of the rest of the nation. The Two Party System assumes, in addition to efficiency, the population not having a degree of tribalism above a certain threshold.
Will the Canadian system iron itself out? I think it probably will. If not by the dissolution of the Liberal Party, then by a permanent coalition between the NDP and Liberal Parties, analogous to the National/Liberal Parties of Australia and CDU/CSU in Germany, or a defacto coalition as with the FDP and CDU/CSU. Either way, if they keep throwing elections to the Conservatives, something is likely to give.
Notably, constitutional theory does not hold much of a difference between the number of parties and number of permanent or implied coalitions. So in Australia, the fact that you have the National Party and the Liberal Party does not contradict the theory. The biggest problem with the theory is the assumption of a lack of tribalism and regionalism.
So can the US ever have a third party? It’s difficult. It’s easier for a party to get started in the legislature and we have a strong executive. And unlike France, we have a very well established party system with a huge advantage to members of those parties. If we got rid of those instutional advantages, the likelihood is that while new parties might come and go, we’d (a) still be looking at two main coalitions and (b) these coalitions would shift similarly to how they do now.
— Trumwill · May 4, 12:08 AM · #
In First-Past-the-Post Parliamentary systems, the median voter theorem should work — provided that ideological cleavages dominate regional ones. That has only sometimes been true in Canada.
Historically, the Liberal/Conservative split was not ideological, but linguistic/religious. Catholics voted Liberal; Protestants voted Conservative. That gave the CCF (the predecessor of the NDP) and the Credistes (too complicated to explain) some entry point, but the reality is that the party system was essentially Liberals and Conservatives into the late eighties/early nineties. At that time, regional and linguistic cleavages became enormously more salient than ideological ones, giving birth to the party system that just died yesterday.
This new party system will probably be stable, and the Liberals will probably decline further, so long as ideological cleavages continue to be more salient than linguistic or regional ones. That’s a big “so long as” in Canada.
— Pithlord · May 4, 04:49 AM · #
Noah: Good points. I don’t know the French system that well, but I’d guess that the U.S. electoral college makes our Presidential system especially inclined to two parties, although I stand by thinking that Presidential systems in general are more conducive to two party systems than the alternative.
As to the Dixiecrats, that is an interesting point. Obviously the Republican southern strategy was intended to capture them and it succeeded. However, I’d suspect that major parties plotting to capture regional parties is a common worldwide phenomenon. I’d suspect that they may no longer have qualified as a nationwide power broker after the electoral portion of the Jim Crow system was broken. The South is obviously still a huge power broker, but without the ability to disenfranchise African Americans, the dixiecrats portion of Southern white votes became a smaller portion of the overall pie.
— Gregory Sanders · May 4, 06:08 PM · #
Just to rephrase the observation made by ‘Marko” – a multiparty system would be expected where you have cross-cutting cleavages in the political society and a political party arises to represent different combinations from the Chinese menu.
In the example of Israel, the cleavages would be Zionist v. non-Zionist, secular v. religious, syndicalist v. bourgeois, &c. The Labor Party was Zionist-secular-syndicalist; the Herut Movement was Zionist-secular-bourgeois; the National Religious Party was Zionist-religious-bourgeois; Agudah Israel was non-Zionist-religious-bourgeois; and so forth. In France, you had Catholic v. secular, authoritarian v. constitutionalist, socialistic v. bourgeois, &c. You also had a great many local patron-client assemblages.
You have cross-cutting cleavages in the U.S. as well: welfarist v. classic liberal, traditionalist v. libertine, internationalist v. isolationist. However, the cleavages co-vary to such a degree that some of the permutations make small platoons. Add to that first-past-the-post and a pair of antique political parties of which each have been a vehicle for an evolving jumble of interests (and practiced at co-opting and assimilating new interest) and you have a bipolar system even though a multi-polar system might represent the balance of opinion better.
— Art Deco · May 5, 01:54 AM · #
After buying a CCTV security System, the more we worried about is the after sale service. However, if you buy a Security Camera System in D CCTV Security, you do not need to worries about it. Because D CCTV Security offer a 100% Satisfaction Guarantee on all of our products. Most of our products come with a 2 YEARS WARRANTY, and FREE LIFETIME TECHNICAL SUPPORT.
— DCCTV · May 6, 07:57 AM · #
How about this as a third party idea: http://thenewthirdparty.blogspot.com/
— LJP · May 7, 03:07 AM · #