Tuesday, October 29, 2002

Mickey Kaus argues that we're in for a long-term 50-50 partisan stalemate, not because of who we are (the red state-blue state cultural divide story), but rather because of what the parties are-- vote-maximizing machines with the knowledge and the polls to constantly adjust to our center of gravity.

An exercise for everyone who took an introduction to political science course at some point: What's that called?

That's right; it's Anthony Downs' 1957 Median Voter Theorem! In a competitive two-party system with a unidimensional political spectrum (and some other technical constraints) the parties will tend to converge around the preferences of the 50%+1th voter. They sometimes make mistakes in their estimation of those preferences (though increasingly-refined polling makes that less likely). They are sometimes beholden to a base for volunteer effort, money, and so on in a way that keeps them from migrating to the center. And sometimes the threat of a third-party split at the left or right margin can exercise a short-term pull on the appropriate party that needs to head off the threat. But, in general and over time, we'll just see fluctuation around the preferences of the median voter.

(A political scientist friend of mine has another friend who's an independent in one of those Reagan-Democrat Michigan counties that splits 50-50 most of the time. He uses this friend as a very reliable weathervane, and proudly tells his colleagues that he knows the median voter. It might seem that in 2000 the decisive voter was more likely to live in Florida, New Mexico, or New Hampshire; but appealing to the median voter in those places requires some local idiosyncrasies. The median voter in Pennsylvania, Ohio, Missouri, Illinois, or Michigan is something more like the median voter in terms of ideological position. This has unfortunate consequences for trade policy. But I digress.)

Most of the conditions Kaus identifies as going on in politics right now could be understood as making this competitive political market (with voters as consumers and parties as firms) work more efficiently and elastically. The absurd heights to which polling, focus-group-testing, and micro-targeted vote appeals have risen are central. As important is his claim that unions and the Christian right have been weakened enough to allow each party to move to the center. I'm not sure he's right about that; remember the teacher's unions and vouchers. Also remember that the Greens are lurking out there on the Democrats' leftward flank, and the Libertarians-- while not right-wing in any general way-- often grab enough of the GOP's anti-tax and pro-gun base to tip House or Senate seats. And we've seen some evidence in the past few elections that one's base can simply get alienated and stay home on election day. So even now, neither party can simply lunge for the center, with no regard for what's happening among their most dedicated supporters.

Note that there are complications in our electoral system. The electoral college, the House, and the Senate might each have a pretty different median voter to appeal to (the Floridian Panhandler-- or do they not like that term?-- the Michigan Reagan Democrat union man and the Missouri soccer mom, respectively). And primaries are a very serious complication. Primaries with lots of candidates have bizarre scrambling effects on who the nomiee will be. Primaries with two or three often pull the nominating process toward the base in a way that makes the lunge for the center more difficult.

But as a general tendency over time, there's a great deal to be said for the Median Voter Theorem-- and Kaus is right that this political market has gotten a lot more efficient lately...

One more complication in mapping Downs' model onto the U.S. is that we don't live in a party-list system. We vote for candidates, not parties-- at least directly. That means that the race for the center will often happen at the level of the individual district-- hence, northeastern pro-choice Republicans and southern pro-gun Democrats. But actual governing happens, in important ways, by party-- most significantly, in control of the two houses of Congress. We lack a strong caucus-and-whip system, but virtually all members elected from one party can be counted on to support their caucus' choices for the chamber leadership (e.g. Speaker), and overall party makeup determines party makeup of the all-important committees. A schematic version of the politics of the mid-90s: the Republicans did a much better job than the Democrats of estimating where the median voters in the south and the Mountain states were. But the Republican caucus elected by appealing to those median voters had an overall position significantly to the right of the median voter nationally-- even though the median voter nationally preferred his or her hometown Republican to his or her hometown Democrat. The Republican House caucus thus became a juicy target for demonization by Clinton-Carville-Morris. District-by-district the Republicans faithfully represented the views of their constituents; and in the nation as a whole more people preferred their district's Republican. But a significant share of northerners and Californians who liked their own Republicans grew to detest the House Republican caucus as a whole, and so the Republicans kept losing PR battles even though they kept holding the House. Call this the Connie Morella problem. Remember that it's happened to the Democrats in other election cycles; indeed, it has something to do with the Democratic House wipeout of '94. And read my earlier post on what how complicated this makes things from the perspective of the individual voter: should I vote for the candidate who most accurately reflects my views? For the candidate whose party most accurately reflects my views, since my Representative or Senator will end up facilitating his or her party's control of the chamber? Or, most interestingly-- assuming local candidate convergence around the local median voter-- should I vote for the candidate whose party does not reflect my views, on the grounds that it's desirable to have a pro-choice segment of the GOP, a pro-gun segment of the Democratic Party, a caucus of pro-trade Democrats, a caucus of pro-gay Republicans, or whatever? (I think the answer is sometimes the last-- but it can involve serious short-term costs.) And this dilemma becomes most serious under conditions of a 50-50 split nationally-- because control of either chamber really could turn on my local election-- and local convergence around the local median voter.

UPDATE: See more above about Kaus' (and Downs') prediction, leading up to an analysis of it vs. John Judis' prediction of a new Democratic majority.

ANOTHER THOUGHT: This is an uncharacteristically goo-goo thing of me to say, but the Florida 2000/ Missouri 2000/ Florida 2002/ Torricelli-Lautenberg/ South Dakota vote fraud/ Minnesota absentee ballot/ etc/ etc problems make it tempting. A time when the parties are in near-perfect balance is as close as we get to a veil-of-ignorance situation for electoral laws. In other words, this-- or rather next January, once all of this year's election lawsuits end-- would be the perfect time to try to generate a model uniform election statute and get it adopted by the states. Election statutes are the quintessential case of rules that should be clearly enacted in advance, without knowing whom they will benefit, because interpretation of ambiguities in particular cases cannot plausibly be separated from the identity of the parties (in both senses). We don't know in advance which party is going to want to yank losing candidates off the ballot. We don't know in advance which party is going to have a Senator die a week before the election. Once the case arises, there are rampant doubts about those who interpret the rules-- the Florida, New Jersey, U.S. Supreme Courts. We've lately discovered all sorts of issues that arise in elections. What's really required to meet a residency requirement (or, in the case of Cheney-Texas-2000, a non-residency requirement)? On what date before an election does a ballot become irrevocably fixed? Should parties be able to yank their living nominees off the ballot? How does the need to get absentee ballots out in a timely fashion balance against the interest in having the Election Day election proceed? What is the right thing to do if polling places open late, or if they still have lines of voters come closing time?

These aren't questions about technology, though one could also add "what standards should apply in triggering recounts, and what standards should apply to the counting of ballots during those recounts?" Deciding them and fixing them isn't a matter of adopting touch-screen ballot booths, the goo-goos' favorite recommendation in early 2001; it's not a matter of passing this law. It's a matter of settling legal questions in advance. A few matters remain pretty clearly linked to partisan advantage: the Republicans have an interest in military ballots, the Democrats in felons' enfranchisement and in same-day registration. But a whole lot of disputed questions of electoral law could right now be treated as indeterminate with respect to who would be benefitted by which answer. The system would be benefitted by clarity, though. It's in everyone's interest that state and federal courts not have to step into a half-dozen elections per year. Let's get that bi- or multi-partisan commission rolling-- not another Ford-Carter or even Brookings-AEI group, but one made up of election-law and procedure specialists who know all the ways in which elections have gone wrong in the U.S., and that has the specific charge of reducing election-law ambiguities and suggesting a model uniform code that could keep elections out of the courts.

No comments: