Home > News > Rational irrationality
107 views 6 min 0 Comment

Rational irrationality

- February 6, 2008

“Matt Yglesias”:http://matthewyglesias.theatlantic.com/archives/2008/02/spoilage.php is bemused by a John O’Sullivan “post”:http://corner.nationalreview.com/post/?q=YWJlNDY4NDgxNzFjNDU5NDNlODVkMTJlMTlhZWE3NGQ= saying that it would be rational for conservatives not to support McCain when he wins the nomination.

O’Sullivan:

bq. Many conservatives believe that the key question in this election is: Are there to be two multiculturalist open-borders parties or one? If McCain’s election were to make the GOP fundamentally similar to the Democrats on immigration, bilingualism, racial preferences, and all the National Question issues, that would be a resounding historical defeat for conservatives. … The willingness of a President McCain to cooperate with the Democrats would give such issues as an immigration amnesty a better chance of passage than under a President Hillary or Obama even against strong GOP resistance in Congress. Opponents of such policies, despite enjoying majority support among the voters, would find themselves politically marginalized.

Yglesias:

bq. I think it’s probably true that, in practice, a comprehensive immigration reform is more likely to come in a McCain administration than it would in an Obama or a Clinton administration. So in a narrow sense, O’Sullivan’s making sense. … But this analysis seems to entirely lack context. If electing a pro-amnesty Republican whose administration fails to ban affirmative action programs would be the end of the conservative movement, then Ronald Reagan’s eight years in office were, just like George W. Bush’s, a “resounding historical defeat for conservatives.” Conservatives can be purists if they like, but the reality is that these are issues on which people who agree with O’Sullivan have never held the whip-hand, and it’s unlikely that they ever really will as long as the GOP remains the party of business first and foremost.

_Contra_ Matt, there’s a good case to be made that it _would_ be rational under many circumstances for conservatives to oppose McCain. George Tsebelis, in his book _Nested Games_[1], makes just this argument about the internal dynamics of the UK Labour party in the 1980s. The relevant chapter is entitled “Why Do British Labour Party Activists Commit Political Suicide?” As Tsebelis discusses, left-wingers within the Labour party often opposed more moderate candidates, even when there was a real risk that this would lead to defeat for the party in the general election. This is because they were playing a nested game, in which they were concerned not only about a one shot electoral victory, but also in getting others to take them seriously over the longer term. In Tsebelis’ words (pp.155-156):

bq. The activists were choosing what would have been a suicidal strategy in a one-shot game because they were involved in an iterated and nested game. By choosing to replace their moderate MPs even when it had catastrophic consequences for the party … they were sending a signal to other candidates and to the leadership of the party: they could not count on their loyalty to the party and the structure of the reselection game to impose on them what they considered to be unacceptable solutions. … they were sending an additional signal and creating a reputation of “toughness” or “irrationality” or … “fanaticism.” … An actor builds her reputation by choosing actions that seem suboptimal and can only be explained as the result of some particular characteristic.

Thus, it may indeed be rational for conservatives to show themselves willing to scotch a Republican candidate’s presidential chances to demonstrate their seriousness on this set of issues. This isn’t necessarily intended to gain the whip hand (although they presumably would like it if they could get it), but to make sure that others don’t take them for granted, and feel obliged to offer them concessions to ensure their support. If game theorists like Tsebelis are onto something, then opposing your party’s candidate for apparently quixotic reasons can be a valuable way to build your future reputation The question of whether Reagan was, or wasn’t, a good president for the conservative movement seems to me to be a bit beside the point.

Even so, to say that this behaviour may be rational over the longer term doesn’t necessarily tell you very much. As the “Davies Corollary”:http://crookedtimber.org/2006/11/29/reputations-are-made-of/ to the Folk Theorem suggests, game theoretic arguments about reputation do have a habit of disappearing up their own fundaments if they’re given half a chance. Yet it’s also worth noting that anti-immigration conservatives have a much better chance than elected leaders in democracies (Davies’ example) of committing credibly to apparently irrational and/or fanatical long term goals and getting others to provide them with various concessions to buy them off, as they’ve _already_ developed some of that highly valuable reputational capital.

fn1. George Tsebelis (1990), _Nested Games: Rational Choice in Comparative Politics_ (University of California Press).