Representation, Governing, Elections, and Deceit

I'm a big Andrew Sprung fan, and of course I'm always flattered when anyone bothers to critique what I've said, so I owe a response to Sprung's weekend post concluding that I'm a bit inconsistent on representation, and generally continuing his critique of how I deal with the subject.

This all goes back to my view of representation (which owes a lot in turn to Hanna Pitkin and Richard Fenno, but of course I take responsibility for my interpretation of their work and my synthesis of it). The gist is that I want to say that good representation involves politicians making and keeping promises to their constituents, and that those promises can be not only matters of public policy but also all sorts of questions of behavior and style. As Sprung says, I also follow Richard Neustadt in suggesting that presidents who look out for their own interests in building influence (Neustadt's "power") will tend to have the side effect of producing viable public policy.

So: last week I attacked Luntzism, although I didn't use that name at the time...Luntzism is the fallacy that if you find the words that poll well and perform well in focus groups that you can slap them on any policy and then mistakenly believe that the underlying policy will be successful. Sprung asks why Luntzism is wrong:

Wait a minute. Suppose a pol's "contract" with her voters consists of 'promising' always to push their hot buttons -- pandering to their prejudices, regardless of policy outcomes? Suppose voters repeatedly reward such rhetoric irrespective of policy...In other words, suppose poll-tested terminology works, either by asserting a distinction between the attacker's policies and his opponents where none exists, or by effectively lying about the opponent's policies? In some cases, it may not matter whether the underlying policies are popular. Is Bernstein suggesting that such manipulation of language -- say, excoriating "regulations" to create a space for gutting regulations that voters would support if they knew their content -- will not work in the long run, that voters will sniff out the truth? Or that poll-tested language deployed to mislead will work in the sense of helping to get the smearer elected, but not work in the sense of serving voters' interests?

I'll say a few things in response. One is that as a practical matter, I'm convinced that Luntzism just doesn't work. Since neither side has a monopoly on political information, the other side is going to undermine whatever happy talk you come up with to describe your policy proposals. Of course, in many cases it just doesn't matter, since in many cases people aren't voting on the basis of specific policy proposals. But if the policy in question is one that people do focus on, I just don't think it will work. See, for example, the Iraq War after the insurgency started and the Bush Social Security proposal. The second, again as a practical matter, is that my stance against Luntzism isn't so much about attempting to fool constituents as it is about mistakenly fooling politicians. During the Contract for America period, we have stories of GOP Members who were assured that various Republican proposals were wildly popular and supported them on that basis, when in fact all that was popular was the question wording.

More broadly, and this holds whether I'm right or wrong about all of that: we shouldn't, in my view, assume too close a link between representation and election outcomes. Both are really complicated, and they overlap, but they're not the same thing. A politician, in my view, can have an excellent representative relationship with his constituency and still be defeated, for example if a national partisan tide has a massive effect on a local constituency. Or the other way around: a politician who has a weak representative relationship can still hang on for years. And then there's the additional problem that we know very well that much (most? nearly all?) of the constituency is not actively engaged in their side of the representational relationship. Which I've written about, but I don't think blogged about...it's a very difficult problem.

And then...well, I do think Sprung is pushing it a bit. I just watched the one where Buffy tells Giles to lie to her...it's a wonderful scene, and Giles is indeed doing what Buffy requests when he lies to her and tells her how easy the world is. We have no evidence, however, in the case of Luntzism that either side is actually building a relationship based on "lie to me." I'm open to saying that a "lie to me" representative style could be a healthy one, but I really doubt that such a thing is very common. Now, could a politician "promise" (as Sprung says) to always push constituents hot buttons, regardless of policy? Sure. But that sure sounds like a very partisan representation style -- which is fine (as far as my view of representation is concerned), but it implies weak connections outside the partisan core, and at any rate it's not usually a winning formula for a district featuring a partisan balance.

Again: part of this is that I think Sprung is trying to link three separate things I've said, things that I would agree can go together, but which I don't believe are quite as tightly connected as he takes them to be. Good representation is about making and keeping promises and about explaining actions in office. For presidents, and perhaps for other elected officials, a positive side-effect of seeking to increase influence is that it tends to create viable public policy. And viable public policy and good representation help in re-election. I do think all of these things are true, and together they do a lot of the work of real democracy...but the latter two are tendencies, not absolutes, and the links between all of them are tricky indeed.

Let me put it another way...one can, for argument, entirely separate elections and representation, and think separately about how they constrain politicians. Elections give pols a general incentive to keep constituents happy -- to produce policy that works (yes, I know about mixed incentives from divided government in the US, but put that aside). Produce a recession, and voters will vote against you, end of story, end of career. Good times? Everyone gets re-elected.

Representation works by constraining politicians to explain what they're up to and to keep their promises. As long as pols want to have strong representational relationships, and we have good evidence from Fenno and others that they do (perhaps for electoral reasons, perhaps not), then they stay within those constraints.

Both of these are factors in making representative democracy work, even with an electorate that's only minimally engaged. Which is, and has been since at least Madison, a terribly difficult problem.