« So now what? | Main | RIP, Nicholas Katzenbach »

Wednesday, May 09, 2012

Polarization Means Two Different Things (and is not a purely American phenomenon)

            Howard links to the Norman Ornstein and Thomas Mann book on polarization that is getting a lot of attention.  I haven't read it yet (it's en route) but am pretty excited to.  That said, I wanted to toss off a few quick thoughts about the discussion it has touched off.  Like many previous debates about polarization, reviews have seemed conflate two very different ideas within the term polarization.   The first is the degree to which the parties divide the population -- call it the distinctiveness of the parties.  When people talk about how many (or rather how few) Republicans are more liberal than the most conservative Democrat, it is this concept they are after.  And the second is the difference between the party's idea policy points in policy space -- how left are the Democrats and how right are the Republicans.  As I noted earlier this week, both are at an all time high in Congress, but the normative implications are very different. 

             Distinctiveness is almost surely a good thing for uninformed voters.  Voters with little information need clear party labels in order to allocate responsibility or to translate policy preferences into accurate votes.    Lau and Redlawsk have shown that voters frequently vote for the wrong candidate based on their own preferences, but that incorrect voting has been falling as distinctiveness has risen.  More distinctive parties also are likely to be more able to push changes through their caucus when they have power. 

             Difference is another story.  It is an unexpected result, given the median voter theorem and almost surely a bad thing in a two party system as it likely leads to some policies preferred by a majority failing to be enacted.   And that's ignoring the point raised by Ornstein and Mann and explored by Rick Pildes in this excellent article, that our constitutional system is a rough fit with a Responsible Party Governance (RPG scholars like E.E. Schattschneider and Austin Ranney worried about this issue quite a lot, but came to contrary conclusions, arguing that only coherent parties could get things done in our constitutonal system, but that's a discussion for another day.) 

            But distinctiveness does not necessarily mean difference (although it can).  For instance, James Snyder and Michael Ting have argued that distinctiveness may mean less difference, as candidates in a system of distinct parties have less need to prove themselves as true ideologues.   But perhaps more importantly is that reforms aimed at reducing difference may result in reducing distinctiveness instead of having their intended effect.  Take "jungle" primaries, which allow candidates to self-identify with a party on the ballot, taking away from parties their ability to control who gets on the ballot and who can identify as a member.  This is aimed at reducing difference, but probably only serves to reduce distinctiveness, as candidates eschew unpopular party names (as Dino Rossi did in 2008, preferring G.O.P. to Republican as he tried to avoid George Bush's coattails).  And it can lead to fringe candidates like David Duke getting through to the second round, leading to less pressure on the major party candidate who makes it.  As it turns out, and contrary to popular belief, the type of primary -- closed, semi-open, open or other -- doesn't seem to matter at all in determining the ideology of the candidates.  This is likely because voters in primaries lack on-ballot party heuristics (what is a primary but a non-partisan election inside a party?) and just don't know very much about the candidates.  Voters in legislative primaries (particularly as you go down the ballot) largely follow the lead of organzed groups that can provide information, as Seth Masket showed in his remarkable book No Middle Ground.  But we should be careful what we mean when we say polarization, as difference or distinctiveness are distinguishable concepts that have separate normative implications.  

            Finally, when talking about polarization, we have a tendency to treat it as a purely American phenomenon.   But opinions about politics have gotten weird the world over.  Westminster systems like Britain, Canada, New Zealand and the like are supposed to have only two parties, per Duverger's Law (roughly, that voters don't like wasting their votes and that organizers and candidates don't like wasting their efforts.)   But all Westminster systems in the world are now multi-party.    And things are weirder yet in most Parliamentary systems, with fringe parties of left, right, nationalist, and just plain odd taking increasing shares of the popular vote.  Here and here I discussed why this might be the case -- why our weirdness expresses itself as polarization, while it takes different forms in other countries -- but we should not pretend that our politics is so different from the politics of other countries.  


Posted by David Schleicher on May 9, 2012 at 06:22 PM | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Polarization Means Two Different Things (and is not a purely American phenomenon):


Consider the polarization of "think tanks" over the past several decades. I was born in 1930 and became politically aware early in life of politics with FDR during the Great Depression and WW II. I was not aware of "think tanks" and their political influences back then. I did learn later of AEI and Brookings, but did not think they were that influential; nor was I immediately aware that AEI was considered conservative and Brookings liberal. I recall seeing Norman Ornstein on news and political shows over the years and he always seemed to be sensible, whether critical of Republicans or Democrats. I do not recall Thomas Mann that vividly in the same time frame as Ornstein. (This was before the Internet and blogs.) I recall Robert Novak on Meet the Press and other talk shows. Novak became somewhat the face of AEI as the "Prince of Darkness." Novak was an ideologue, unlike Ornstein. AEI added others some of whom were the neocons that pushed for and supported George W. Bush's Afghan and Iraq wars. Yes, Brookings had the liberal E. J. Dionne, a pussycat compared to Novak.

So it is somewhat heartening to see Ornstein and Mann come together with their book.

By the way, it's clear to me that AEI has been more polarizing than Brookings.

Posted by: Shag from Brookline | May 10, 2012 10:47:49 AM

Charles -- While I agree that Presidentialism does create added Duvergerian pressure, I think you overstate the case a bit. Westminster systems have single member districts and use first past the post vote counting, meaning that in any given race a vote for a third party is wasted. Further, in Congress, mixed coalitions are a theoretic possibility -- independents like Jim Jeffords have swung the balance in the past and I wouldn't be surprised if Angus King of Maine is the swing Senator in the next Senate. Also, while Westminister systems have always had small regional third parties, in 2009 or so something new happened -- there were no Westminster systems with majority governments. (Canada has since swung back.) So there has been change.

If you hit the last links in the post, you can see a discussion of another major difference between Westminster systems and the U.S. -- primaries -- and how that affects third-party creation.

Posted by: D.Schleicher | May 10, 2012 10:34:24 AM

I know it's not the main point of your article, but Westminster systems are far more forgiving of third parties than is the American system. In a parliamentary system, a vote for a third party is not lost, so long as your candidate is elected, because MPs have far more (theoretical) influence on governmental policy than American Representatives. Put another way, in a Westminster-style system, a third party "wins" if it elects enough MPs to deny the government a majority—it can now have a large amount of influence on legislation, policy, etc., even if there is no formal coalition and even if it has only a handful of MPs. In the American system, where most focus is on the presidential elections, a third party is only treated as "winning" if it wins a majority of votes in the electoral college. This is a monumental burden, nearly impossible to achieve. As such, it is unsurprising that most efforts at political reform have focused on the two existing parties rather than on forming third parties.

Also, looking back, third parties have existed in many Westminster-style parliaments since at least the early 20th Century. The UK Parliament circa 1900 contained not only Conservatives and Liberals, but also Liberal Unionists, the nascent Labour Party, and Irish nationalists. Even after the rise of Labour, the old Liberal Party hung onto seats until the present day, occasionally entering into coalitions with or informally supporting one of the other parties. Canada has also had a third party represented in Parliament since the rise of the Progressives in the 1920s and it has not been uncommon for four or five parties to be represented.

All of which is to say that yes, Westminster-style systems do experience some effect of Duverger's law (seen most recently in the dramatic swap between the Canadian Liberals and NDP after the NDP became perceived as "electable"), but it is not nearly as dramatic as in the United States.

Posted by: Charles Paul Hoffman | May 10, 2012 8:52:10 AM

The comments to this entry are closed.