Tuesday, January 25, 2005

Opposition Strategies 

[NOTE: I've been meaning to write about this for some time, and a longer version will be posted soon]

Here's a question to get you thinking: what has been, historically, the most successful electoral strategy for US political parties who are out of office? In other words, if the other party just ate your lunch at the polls, and you want to win in the next election, should you:

a) Adjust your policies to be more like the other (more successful) party's?
b) Stick to your guns and maintain the same policy positions (or move farther away from the other party)?
c) Innovate: find a new set of policies to champion?
d) Go out of business?

If you chose "d", you're not paying attention, I guess. The surprising answer, however, is that none of the other options is clearly superior, historically.

Lots of smart people have made the case for the other three options. The cartoon version of the DLC is that it always shouts out "Go right! Take option 'a'!" Option "a" is also endorsed (or at least predicted) by a huge literature in political science, drawing on the work of Anthony Downs, which suggests that in a two party system candidates will tend toward the political center. [Those of you with access to a good library can find an excellent review of this literature, by Bernard Grofman, in the 2004 issue of Annual Review of Political Science. It's worth noting that the newest research on the issue is driven by the empirical observation that even in the US, the two parties and their candidates tend to support and enact very different policies, pace Ralph Nader and most college sophomores.]

The left half of the blogosphere regularly endorses option "b". And the DLC's own version of its history suggests that option "c" has worked in the past (Clinton 1992) and could work again.

So what does history tell us? The most rigorous attempt to answer this question that I'm aware of is a 2001 paper by Kenneth Finegold and Elaine Swift (you can find it here if you have access to the British Journal of Political Science). Their result? Despite the impressive arguments mustered by proponents of each option, none of the strategies stands out as superior, at least in the US presidential context. Each strategy has produced a statistically similar won-lost record. A more sophisticated test (examining the size of deviations from predicted vote totals) returns the same result: even when controlling for the non-policy factors (national economic performance, personal popularity of the incumbent) that typically go into vote-prediction formulas, the out-party hasn't done any better or worse when it has chosen to move toward, away, or in a different direction from its incumbent opponent.

There's an important caveat to be made here: the results presented above don't take into account the policy strategy chosen by the incumbent. I'd suspect this is an important omission.

While we're waiting for further research, though, where does this leave us? I'd draw at least one pretty strong conclusion: reasoning from historical analogy here (e.g. "President X gained election and established a legacy of control for his party by adopting strategy Y") is pretty worthless. Rather than seizing on a narrative based on a single past election, Democratic policy strategists should focus on the here and now.

UPDATE: Mark Schmitt has a terrific post up about the use and misuse of George Lakoff's "framing" ideas, which makes a similar point about extrapolating from individual cases of electoral success.

This page is powered by Blogger. Isn't yours?

Weblog Commenting and Trackback by HaloScan.com Referrers: