A syllabus of errors

by John Quiggin on August 17, 2004

The WashPost runs an Op-Ed piece byPradeep Chhibber and Ken Kollman, claiming that the failure of third parties to do well in the US is due, not to plurality voting or other institutional factors but to excessive political centralisation. The claim is that since third parties

once competed successfully in congressional elections, winning significant portions of the popular vote and often gaining seats in Congress. This was true for most of the 19th century and even the early part of the 20th
the cause of their subsequent failure must be something new – political centralisation[1].

Chhibber and Kollman seem to be well-regarded political scientists. But their argument here is riddled with errors, or at least large logical gaps.

First, they present hardly any data, and don’t answer the obvious empirical objections. Their claim that third parties do less well now than in the past runs into some obvious problems. Two of the last three presidential elections have been decided (or at least greatly affected) by third-party candidates, Perot in 92 and Nader in 2000. The Reform party also elected a governor, Jesse Ventura in 1998, and there’s one socialist member in Congress. That’s not much of challenge to two-party dominance, but did the parties cited by Chhibber and Kollman (Prohibition, Socialist, Populist, Greenback, Farmer-Labor) do notably better? We’re not told, but I’m pretty sure the answer is “No”. There’s also no evidence (beyond the single data point of Canada) that centralism is favorable to two-party systems.

The logical problems are even more striking. Granting, for the sake of argument, that third parties have declined since the 1930s, and that centralisation has increased at the same time, haven’t these guys ever heard that “correlation does not imply causation”. Leaving aside the possibility of purely spurious correlation, there are plenty of possible “joint cause” arguments. For example, it might be that the rise of mass media has both reduced regional diversity (which implies less reason to oppose centralisation of political decisions) and also given advantages to large parties.

But, there’s still a more significant error. Let’s suppose we’re satisfied that third parties were once strong and that their decline was caused by political centralisation. It’s still obviously true that such parties are disadvantaged by plurality voting and other features of the electoral system. If you want to encourage third parties, you can either fundamentally change the relationship between Federal and State governments, reversing 100 years of history, or you can change the voting system. Changing voting systems isn’t easy, but it’s been done in many places, and can be done on a state-by-state basis .

fn1. Looking on the web, I found this book chapter by Chhibber and Kollman, which seems to date the rise of the current two-party system, more plausibly in my view, to 1860. It is certainly true that the Republican Party of that time, devoted as it was to the containment and ultimate destruction of a regional “peculiar institution”, did a great deal to enhance the power of the central government.

{ 30 comments }

1

reuben 08.17.04 at 11:00 am

Speaking of third parties, today’s Guardian features an infuriatingly naive column by George Monbiot in which he urges good lefties to do the good lefty thing and vote for Nader – because Kerry says he would have supported the war (http://www.guardian.co.uk/comment/story/0,3604,1284465,00.html).

Is Monbiot really too daft to realise that had pie-in-the-sky lefties voted for the Dems instead of Nader in 2000, there never would have been a war in the first place?

Makes me want to slap the man in the face with a rolled-up copy of the Socialist Worker.

2

ross 08.17.04 at 11:29 am

Another, and better, alternative explanation for the lack of party system evolution would be party decentralization. Open primaries have weakened parties. Since parties no longer have direct control over who represents them as electoral candidates, candidates can appeal to their local community with a very specific set of policy preferences, often distinct from what their party may appear to support elsewhere. Consequently, expressed party preferences differ significantly across the country. For example, a typical Manhattan Democrat party candidate is quite different than a typical Arizona or South Dakota Democrat. Such intra-party, cross-national heterogenity of preferences crowds out third party formation. Moreover, this tendency is also supported by the fact that parties in the US are decentralized. The national level parties are looser organizations. It is at the state level that party organization matters most. Indeed, it was at the state and city level that political machines developed historically.

3

Jonathan Dresner 08.17.04 at 11:34 am

Well, there’s some logic there: the most highly centralized governments in the world have only ONE party….

But there are plenty of reasonably unified nation-states (some with considerably stronger central governments than our Federal system: Japan, Israel and France come to mind) in which smaller parties (I hate the term ‘third’ parties, by the way, as it legitimates the idea that there are only two ‘real’ parties) are numerous, influential, and which do not have the regionalistic character they describe for 19c US parties (Though in the case of Israel, religion and ethnicity largely replaces regionalism as the core of smaller parties, the others have a real ideological range).

Moreover, if regionalism was going to be a strong base for a new independent party, a Southern party (Dixiecrats, for example) would have been viable. No, the issues are fiscal, rhetorical, habitual, and legal (who plans the debates? a supposedly independent commission entirely staffed by evenly matched big party functionaries), but you can’t blame the federal system.

4

Patrick Nielsen Hayden 08.17.04 at 11:46 am

“Open primaries have weakened parties. Since parties no longer have direct control over who represents them as electoral candidates, candidates can appeal to their local community with a very specific set of policy preferences, often distinct from what their party may appear to support elsewhere.”

But back before the rise of open primaries, there was just as much–arguably even more–ideological difference between (for instance) Democrats from different parts of the US. If anything, in the era of open primaries, the parties have become more ideologically consistent and less hospitable to regional variation.

5

abb1 08.17.04 at 1:04 pm

Seems really obvious that the single most important reason for the existing two-party model is our winner-takes-all electorial system. Plurality voting would be an improvement, but proportional representation appears to be the best solution.

6

Idiot/Savant 08.17.04 at 1:16 pm

I’d have thought that the real limit was restrictions on ballot access – the Democrats and Republicans have tended to raise threshholds in a deliberate attempt to protect their political duopoly.

7

ross 08.17.04 at 1:36 pm

“But back before the rise of open primaries, there was just as much—arguably even more—ideological difference between (for instance) Democrats from different parts of the US. If anything, in the era of open primaries, the parties have become more ideologically consistent and less hospitable to regional variation.”

Yes, but open primaries are only one factor. The other is that even before open primaries, endorsements were controlled by state parties, not the national apparatus. So regional variation was sustained.

Also, that party policy preferences have become more “ideologically consistent” is largely an artifact of post-civil war southern Democrats having finally merged into the party system in an ideological way rather than as simply being anti-Republican. This variation is a function of the distribution of preferences rather than institutional incentives. One can’t think about these things in a bivariate way. I was only commenting on institutional effects above and should have mentioned that I was not considering preference variation, though that is an important factor.

8

digamma 08.17.04 at 2:08 pm

Is Monbiot really too daft to realise that had pie-in-the-sky lefties voted for the Dems instead of Nader in 2000, there never would have been a war in the first place?

That doesn’t seem obvious to me. For whom should such “lefties” have voted in 1996 to prevent the bombings of Iraq, Kosovo, and the medicine factory in Sudan?

9

zaoem 08.17.04 at 2:40 pm

It is pretty unfair to come with the standard set of methodological objections to an Op-Ed. As you could have seen at the bottom of the page, the piece is based on a larger comparative study of political systems. Moreover, the study is clearly about the ability of third-party candidates to hold elective office, not their ability to influence the outcome of what is essentially a two-candidate race (we know about that).

10

reuben 08.17.04 at 3:06 pm

Fair points, Digamma, but Monbiot is crying out for a revolution that just ain’t gonna happen, and he and other Nader supporters are actually pushing us much further towards the bad. He couldn’t have known in 2000 what the results of voting for Nader would be, but he does now: four more years of the most rightwing president in living memory. For someone with Monbiot’s environmentalist beliefs, just to take one example, four more years of Bush is the single worst thing that could happen to the world, yet he seems to want to give us that in order to punish Kerry for what I view as politically expedient talk aimed at the domestic political market. And as for foreign policy and war: by calling for people to vote for Nader, Monbiot is effectively calling for them to not vote for the Dems. Because enough people – including many reasonable ones – did that in 2000, we’re in the mess(es) we are now. And we all know that.

And after four more years of Bush, is there going to be a revolution? No. There might be a Democrat in the White House, but that Democrat will have to fight like mad to rectify even one-twentieth of the damage Bush has already done.

The Dems might be too militaristic for you, and that’s a far complaint, but they are better than the only realistic alternative, and they are so far better on so many substantive issues that I just don’t see how voting for Nader is an option. IMO, a vote for Nader won’t have one iota of positive effect re teaching the Dems a lesson (if anything, it will only show them that they can’t trust the left end of their possible voter base, thus pushing the party further to the right) but it will contribute to making the world a worse place for millions and millions of people.

In some elections, voting third party is the right thing to do, I think. But not this one.

11

Nicholas Weininger 08.17.04 at 3:29 pm

Your points are fair, but the authors nevertheless have part of the picture right, I think. Decline in regional diversity, certainly correlated with and possibly caused by political centralization, has been a significant factor in the decline of third-party success and power. Populist parties used to do very well among Midwestern voters of a certain stripe; the Socialists and Farmer-Laborites enjoyed considerable success in Wisconsin and Minnesota respectively for some decades (Milwaukee, IIRC, had a Socialist mayor for much of the early twentieth century, and the Farmer-Labor party retains enough sentimental prestige in Minnesota that the state’s Democrats still call themselves Democratic-Farmer-Labor). And back then influence in a single state did mean a lot more.

Now even a state’s governorship means relatively little, and the distinct sensibility of that pocket of voters is gone. This is not entirely a bad thing– the midwestern populist parties’ actual policies were mostly demagogic idiocy– but it’s a real and important change.

12

harry 08.17.04 at 4:03 pm

Nicholas w is right that the data are much more in their court than you suggest, John, though I agree that you make many fair points.

However, the data there, and the UK experience of multiple parties in a first past the post system of voting count against the claim that its that which maintains the two-party duopoly (as abb1 asserts). Sure, if we got a better electoral system we’d see the emergence of more parties, but its possible to have more parties within a winner-take-all system. State control of parties, the financing system, ballot access restrictions (the two Parties have protected their duopoly by making it much harder for third parties to get onto ballots), and the vast size of US Congressional and most Senatorial districts probably have a lot to do with it too.

13

harry 08.17.04 at 4:41 pm

Not, I hasten to add, that I disagree with abb1’s prescription of proportion representation — completely agree with that on numerous grounds.

14

Silent E 08.17.04 at 5:05 pm

There must be a strong regional bias in UK politics that allows smaller parties like the Liberals to succeed in a first-past-the-post system in some districts, while the Conservatives win elsewhere.

In the US, were social conservatives to rally around the Constitution Party (for example) and desert the GOP en masse, you would not see right-leaning politicians still controlling the House and Senate. Many of those incumbents, when faced with a loss of half of their base, would simply lose to the Democrat. The current 220 R – 215 D Senate would not become 110 R, 110 C – 215 D; it would be more like 55 R, 55 C – 325 D!

It makes sense for a rational voter to choose a smaller party over a lesser-of-two-evils major party only if there are good odds that the smaller party has a chance of winning, or at least that a vote for a smaller party is not a vote for the voter’s most-hated polar opposite. This applies both to the particular election, and to the composition of the legislative chamber itself.

I suspect the real reason can be summed up as part ‘path dependence’ (there have only ever been two major national parties at any one time), and partly because a two-party system is a stable equilibrium solution in an electorate that is well centralized or that varies in viewpoint primarily along a well-agreed right-left axis – no player has an incentive to pursue an alternative to the big-tent national 50%+1 coalition strategy.

15

Dave 08.17.04 at 5:41 pm

Proportional representation would be a problem in the U.S., because there is still a huge regional variation in both the politics and the needs of the population. It is much easier for, say, ranchers in Texas, farmers in the Midwest, and Native Americans in the Dakotas to get representation when they have individual representative districts to vote in.

The whole structure of the U.S. legislature is designed to concentrate power in the states and in local areas. Using an at-large voting scheme would effectively disenfranchise a lot of people.

16

harry 08.17.04 at 5:48 pm

bq. There must be a strong regional bias in UK politics that allows smaller parties like the Liberals to succeed in a first-past-the-post system in some districts, while the Conservatives win elsewhere.

I’m not sure whether I agree or disagree with this. Yes, the Liberals retained (tiny) local strongholds throughout the 1924-1992 period in which they were basically out of the picture nationally (you can end the period in 62 or, more plausibly, 83 if you like). But the regionality doesn’t play into their post-92 resurgence much. On the other hand, I guess my (opaque and hurried) point about constitutency size was that it is easier to make an impact in much smaller constitutencies, both because there are fewqer voters and because it is easier for the parties to make an impresion by leading local campaigns, etc. Finally, the Liberals were, until about 1999 perceived as a *middle* party (for most of that time rightly), so could win votes from their left and their right (still do, in fact, even though the view that they are a middle party is almost completely exploded); whereas there has ben no space at all for a middle party in the US for generations.

17

harry 08.17.04 at 5:54 pm

Dave — sure, a completely at-large system would be wrong (and absurd, and unmanageable, and….) But, what about a list system combined with a geographic top-up (or vice versa), or State-wide PR, or PR for State assemblies and senates only, or… there are lots of ways of doing it. For President, obviously, we’d need something different, like run offs, etc.

18

Sebastian Holsclaw 08.17.04 at 5:55 pm

I’m not sure an analysis of smaller parties can make sense unless you try to analyze how the larger parties change. We don’t today have the Republican party of Lincoln or the Democratic party of Martin Van Buren. Each of the bigger parties has been transformed from time to time. The most recent set of changes are probably the rise and relative fall of the ‘Moral Majority’ within the Republican Party. Inside-Party movements have many of the characterisitcs analyzed above as ‘third party’, but when they gain influence they get the power of being one of the larger parties.

19

Jeff Licquia 08.17.04 at 7:50 pm

From a historical perspective, Quiggin is on better ground than Monbiot.

I count the following parties as “major” in American politics: Federalist, Democratic-Republican, Democratic, Whig, Republican.

All other parties that have arisen in American politics have been pretty much the same: influential during a short ascention, followed by collapse.

I will note three exceptions:

– The minor parties participating in the 1860 election. All serious students of history have to agree that this election was a major anomaly.

– The 1864 “Union” party, a Republican party in all but name.

– The National Republican party, which had the unique position of having only one major party in opposition. Its anemia led to the formation of the Whig party.

Some might suggest that the whole debate is circular, in that we assert “all third parties fail except for those which don’t”. But it is instructive that our system seems to be unable to support more than two parties at a time; every instance of a third party ascending to major-party status in our history has been accompanied by the demise of a major party.

It’s also worth noting the periods in our history that were characterized by virtual one-party rule: roughly, the Washington era and the period immediately preceding Jackson.

20

abb1 08.17.04 at 8:34 pm

Poportional Representation vs. Winner-Takes-All.

PR: a number of small parties with definite platforms form alliances after elections. What you get, eventually, is a ruling coalition and opposition coalition.

WTA: a number of small parties with definite platforms form alliances before the elections. What you get is two parties with amorphous platforms. One will be the ruling party and the other one the opposition party.

Clearly the PR model is much more flexible, which is not necessarily a good thing, but probably is. It’s also much more transparent: if you want to vote for your regional party or, say, for the anti-war party – you just do it and let your party find the best alliance; you don’t have to guess who is going to be better for your pet cause – republicans or the democrats.

21

John Quiggin 08.17.04 at 9:01 pm

Zaoem, I write Op-Ed pieces regularly, and an important component of the art is writing in a way that indicates that you are aware of obvious objections, and have an answer, even if it’s too complex to be spelt out. To write a piece on third parties and not even mention the Reform party fails this test. On your second point, the authors mention “winning significant portions of the popular vote” as a criterion.

Nicholas, it’s begging the question to say that, since the US is now more centralised, winning governorships does not count as much as it used to. On this formulation, it’s tautologically true that centralisation has caused the decline of regional parties.

22

John Quiggin 08.17.04 at 9:19 pm

Adoption of instant-runoff voting, as in Ireland or Australia would eliminate the “wasted vote” problem for minor parties while maintaining local representation.

We currently have independent and minor party representatives in most of the state and national lower houses, elected under this system (upper houses are usually PR).

23

Dave 08.17.04 at 9:51 pm

[W]hat about a list system combined with a geographic top-up (or vice versa), or State-wide PR, or PR for State assemblies and senates only, or… there are lots of ways of doing it. For President, obviously, we’d need something different, like run offs, etc.

I prefer approval voting for president. I think PR by state *might* work (and would avoid gerrymandering issues). Anything that does away with the primary system, however, is bad – I hope this isn’t what you’re suggesting by going to a “list” system.

Clearly the PR model is much more flexible, which is not necessarily a good thing, but probably is.

There is a counter-argument here. In a PR system, people are still locked into specific parties with narrow platforms. Often, there is only one candidate available from each party.

In a two-party system with amorphous, coalition parties (as in the U.S.) the voters get to decide not only which party to vote for, but also who is the representative of that party in each race. There may be any number of different viewpoints expressed (compare Dean, Gephardt, Kerry, and Kucinich).

Especially in regional and local elections, people can come up with their own platforms, which might even overlap between the two “major parties” (see Northeast and West-Coast Republicans vs. Southern Democrats). They can have candidates who directly reflect their views and are not necessarily beholden to a particular party ideology.

My Republican Representative, Tim Johnson, is socially and fiscally conservative, but sides with Democrats on a lot of farm issues, since I live in a mostly rural area. If voters had to choose between a Tory, a LibDem, a Labour, and a Green candidate, it would be hard for voters in my area to find one that really reflected their politics.

I’m not saying that one system works better or worse than another, just that the system we use in the U.S. is just fine for us. Now, that’s not to say I wouldn’t prefer going to PR for local elections, or to approval voting for winner-take-all races. But I don’t think either of those changes would or should lead to a system with a large number of smaller, tightly-focused parties.

24

Walt Pohl 08.17.04 at 10:26 pm

How do the LibDems do it? Who votes for them?

25

abb1 08.17.04 at 11:43 pm

I think PR systems are much better in all situations and for for all levels, except, of course, the presidential election.

I’m not saying that one system works better or worse than another, just that the system we use in the U.S. is just fine for us.

With our WTA system if you are, for example, a republican in Massachusetts, your political views are not represented on the federal level at all. Your vote is wasted, you have exactly zero representation. Sure, your region is represented, yet it doesn’t sound like a good deal at all. And republicans are a significant minority in Massachusetts.

Adoption of instant-runoff voting, as in Ireland or Australia would eliminate the “wasted vote” problem for minor parties while maintaining local representation.

Your interests on the federal level have a regional component and the rest – let’s call it “political” component.

Under WTA (even with instant-runoff) your regional interests are guaranteed to be represented – 100%; your political interests, however, will be represented only with 50% probability.

Under PR you have nearly 100% chance that your political interests will be represented, and, if there is a strong regional uniqueness where you live, then you will most likely have a regional party to vote for, if you are so inclined (like Basque Nationalist Party in Spain for example).

26

cure 08.18.04 at 12:40 am

abb1…We’ve put Republicans in the Governership of Massachusetts for well over a decade. In a PR system, individual Republicans would have a worse time of winning than they do now. Party affiliation is loose enough in the United States that, as many have noted, the variation within parties allows the wishes of local voters to be reflected quite well with WTA.

That said, is the lack of third party representation that important anyway? I don’t think so. Coalition building isn’t very important when a PM isn’t selected. The Senate and Congress both contain members that could just as easily be members of Green, Libertarian, Constitution, “potato-throwing farmer” (as the Economist puts it), etc.. In many cases, party affiliation is useful merely as a fundraising tool.

Further, the importance of major parties lies in their ability to vet candidates. Look at the Filipino elections, with their weak post-Marcos party system. Because the party system is weak, celebrities dominate policy wonks in elections.

Lastly, it must be remembered that the US is essentially a contractually linked set of states, who entered a union with the promise of local representation. Despite the fact that post-Warren we are much more centralized, it’s simply a non-started that smaller states would give up considerable political power and agree to a PR style system.

27

Tom Morris 08.18.04 at 1:05 am

The reason why the Liberal Democrats have lost the middle ground is because they have sold out their liberal principles for namby-pamby social democracy. They now treat the individual as an idiot rather than as a rational being. There is no proper liberal alternative, since they have become social democrat. I vote for them only in the hope that they might become liberal again.

28

Brennan Griffin 08.18.04 at 4:35 am

I think part of the answer is in structural ballot access patterns. Third parties in the US used to be much more common and more powerful until early in the 20th century. Republicans at the time were consistently getting pluralities, but were beaten by politicians running on joint Farm-Labor and Democratic fusion tickets (one politician running on the ballot line for both parties).

Republicans used their political might and with the help of Democrats to get rid of fusion in all but a few states, vastly complicating things for third parties. With fusion, a party can use ballot access to form coalitions on the big tickets, without having to play the spoiler a la Ralph Nader.

Its no accident that New York, which still has fusion, has some thriving third parties (Conservatives, Working Families Party, even a Right to Life Party). These parties wield significant statewide influence, command patronage jobs, and get candidates to sign off on anything from higher minimum wages (WFP) to freezes on taxes (Conservatives).

These parties can wield significant influence in tight elections – throwing the election to Giuliani for example, when the Liberal Party (misnamed these days) endorsed him and gave him the margin of victory in his first race. They can always play the spoiler if they need to, but they also have the carrot of an endorsement that carries real, provable votes as well.

29

Jeff Bogdan 08.18.04 at 5:09 am

First, why do the comments on this site display so weirdly (at least, on IE? Some lines are completely obliterated, as if a child had got hold of a bottle of whiteout and applied it in streaks to a manuscript.

Anyhow, though I’m not voting for Nader this year (tho’ I remain unrepentant about having voted for him in Democrat-safe NJ in 2000), his analysis of of the U.S. duopoly is still relevant and so far absent from this discussion.

The nutshell version goes like this: National campaigns are fabulously expensive–mainly becauase modern campaigns are primarily conducted through TV ads and public election funding just can’t compete with all the private cash sloshing around. Only someone who is personally wealthy or who can raise really big bucks can even dream of running for President. Big money tends to come, duh, from very rich people and big corporations whose polital interests, while not completely determined by their economic interests, are certainly heavily weighted toward them.
This naturally has an effect on what parties (and candidates within each party) they are willing to contribute some of their wealth to.

So that while the Republicans and Democrats are not identical, in many respects they are quite similar in their views about economic policy, social insurance, distribution of wealth free (so-called) trade and large swathes of foriegn policy.

Which is why in 2000 we had a Democratic President who tossed thousands of families on on welfare out of a 10th storey window, while them assuring on the way down that he felt their pain and promising them the good jobs that somehow never materialized; who pushed through NAFTA and wanted to extend that mad-scientist scheme to the entire hemisphere; who couldn’t be bothered paying attention to the outright theft of African American votes in Florida–many more votes lost to the Democrats that way than through Nader’s incursion and dangling chads combined–and a Democratic Party that couldn’t be bothered doing anything about it even after the fact. (See the footage, as shocking in its own way as that of Bush reading My Pet Goat or something or other while Manhattan burned, of the Black Congressional Caucus pleading before the U.S. Senate for one, just one Senator to co-sign their protest of this 1980s Salvadoran-style election, and coming away empty-handed–not a single Democratic Senator would lend a hand.)

So vote for Kerry, United Fronters, but don’t get too relaxed when he wins. La lucha continua, as they say in Central America and , nowadays, in Florida, too.

30

Pay per click Search Engine Advertising 08.25.04 at 12:13 am

Good point guys! You can search for more information at http://www.giveramp.com
as it’s better then Google for this type of topic.

Comments on this entry are closed.