I just got finished teaching Congress this semester, and for my money, the most policy-relevant idea that I came across comes out of David King’s work on political polarization. Briefly put, there is a large literature in political science that argues that political polarization in Congress can be explained by the decreasing competitiveness of Congressional districts. In a perfectly competitive district, there would be very strong pressures for party nominees for the House to move to the location of the median voter in their district, because if they did not, it woud leave open an opportunity for their competitor to swoop in, seize that median voter and win office. This “median voter theory” was made famous by Anthony Downs in his Economic Theory of Democracy.
For a variety of reasons, we now have very few competitive districts, which means that Congressional elections almost everywhere now look a lot like 1940s Alabama, with the primary tantamount to election. Where this is the case, it is argued, ambitious office-seekers will get the message and appeal more to their primary electorate, and less to the median voter in their district. Add this up over 400-plus non-competitive seats in the House, and you’ve got a prescription for very few moderate members of Congress.
David King has looked into this and found that the old median voter theory doesn’t work so well. He found here that some of the most extreme members of Congress come from some of the most competitive districts. It turns out that when a district is competitive, the most extreme voters get highly mobilized and turn out for primary elections, driving the median primary voter even further to the poles. Were candidates chosen by fat, cigar-chewing men in back rooms, then this would create a great opporunity for Downsian action–one party would look at who the other was nominating and say, “wow! what good luck! those guys are slitting their throats by choosing a candidate at the 95% of district preferences” and pick someone closer to the median voter. But we now have primaries instead of having the party as an organization choose the candidate. And the important point is that primary voters are: a) uncoordinated and; b) self-selected on the basis of ideology–these are generally people who put a greater priority on ideology than victory.
Given a competitive district, this suggests that elections will tend to pit NOT two center-seeking candidates against one another, but two pole-huggers. The consequence will be that districts such as this will have a great deal of turnover, as voters elect one wingnut by 51%, and then throw him out the next time (for the opposite kind of wingnut) by the same spread.
Now, notice the logic of this. If King’s argument is true, then polarization is primarily a function of who shows up for the primary, NOT what the conditions of competition are in the general election. The further implication is that the more the primary becomes self-selecting on the basis of ideology, the more it will produce ideologically extreme candidates, REGARDLESS of the character of the district. So what can explain why this tendency to produce extreme candidates in competitive districts seems to have increased? David suggests two theories. First, turnout in Congressional off-year primaries has fallen through the floor–now it is only the most motivated (and thus ideological) activists who bother to show up. Second, the character of those party activists has changed–as the parties’ base have sorted out, the most active party members are those motivated by ideology, and not solidary benefits.
The implications here are quite stark. I’m not quite sure I buy David’s argument that competitive districts produce MORE polarized members. But I certainly think he’s right that it is the conditions that now prevail in the primary that really matter. The problem with this theory is that it suggests almost nothing can be done about the absence of moderates in Congress. It is relatively easy to hand districting over to a bunch of retired judges and tell them to draw straight lines that don’t split counties. That is a TECHNICAL fix. But how do we draw more moderates into primary elections? That is a cultural and behavioral problem.
On behalf of the whole markarkleiman.com family, who (at least some of us!) when we are not attacking the Bush administration are actually pretty moderate folks, hereby invite suggestions on how to drive up primary turnout.