The Laws of Politics

            Mark Schmitt of the New America Foundation argued last week that the laws of politics as we know them have broken down.  The biggest casualty, he says, is the “median voter theorem” which holds that politicians will calibrate their actions to win the vote of the median voter.  He observes that politicians are no longer doing constituent service or seeking earmarked spending projects in their districts they way they did in the past.

            The median voter theorem is not dead.  It’s just that most of us in the poli-sci biz learned it before our professors understood that we were living in special times.

            Consider Figure 1 that describes a society where opinion is normally distributed along an ideological continuum that runs from extremely liberal to

Figure 1

extremely conservative.  The fattest part of the curve, where most of the voters are, is at about 5.  Suppose one politician portrays his or her own ideology and portfolio of policies as being somewhere close to 6 while the other sets up shop at 3.  Voters who are at positions 0 through 4.5 will vote for the candidate at 3, but everyone else will vote for the candidate at 6 because they are ideologically closer to 6 from every point greater than 4.5.

            In a world where public opinion is normally distributed as in Figure 1, it is rational for politicians to try to position themselves as close to the median—which also happens to be the midpoint of the ideological continuum--as they can.  When they do that, we end up electing a lot of moderates.

            But this is not the world we have.

            Nothing requires the distribution of opinion to be centered on 5, the midpoint between liberalism and conservatism.  Suppose the curve shifts to the right so that the median is 6.  Candidates from the liberal party can still compete, but their views will have to be more conservative than the average national liberal party candidate to capture the vote of the median voter.  

            What happens if the curve shifts to right so that the median is 8? Or, suppose instead that we have a community with a bimodal distribution of opinion as in Figure 2 or Figure 3?  The median voter theorem still works, but it doesn’t predict the election of centrists.

Figure 2

Figure 3

            If we have public opinion distributed normally, but centered on position 8, a rational candidate will aim to be strongly conservative.  There is no benefit to any candidate to try to moderate his or her positions so as to win some left-leaning voters.  There just aren’t that many lefties to win, and moderating one’s position could set up a challenge from the right.  All of the meaningful competition is going to take place on the right side of the scale.  Conservatives are going to win the election and it barely makes sense for a liberal to compete.

            If we have a community with public opinion distributed as in Figure 2, a rational conservative politician should try to take up a position close to 8; a liberal will want to be close to 2.  Both of these politicians will look like extremists to people on the opposite side of the scale.  There aren’t any voters at the midpoint of the scale, and so it’s irrational to come across as a centrist. The winner of an election in this kind of a community will be the politician who best mobilizes his or her base.

            In communities characterized by Figure 3, the winner is going to be the politician closest to 8. The rational politician here is going to ignore anyone to the left of position 5.

            The America of 2016 is not the America of the 40s, 50s 60s or 70s.  Over the last 40 years, Americans have been moving to communities that are much more politically homogenous than ever before.  It’s now far more likely for communities to resemble the communities described by Figures 2 or 3 or to be shifted versions of Figure 1 than it is for them to have normally distributed opinion centered on the midpoint of the liberal/conservative continuum.

            This is why Schmitt thinks the old rules don’t work.  He’s mistaken.  Politicians are still trying to position themselves close to a median voter, but because our communities are more ideologically homogenous, the median voter no longer occupies a place near the center of the ideological continuum.

            And that’s why politicians can get away with saying and doing things they couldn’t get away with before.  It’s also why politicians don’t have to compromise, reach across the aisle, pursue pork barrel projects or even seek benefits for constituents who may make up a substantial portion of the electorate but aren’t part of the politician’s electoral coalition.  Even unpopular politicians such as Sam Brownback of Kansas and Paul LePage of Maine, two wildly unpopular governors, could win re-election: they simply stood closer to the median voter in an electoral coalition than did their opponents.  When politicians don’t face serious electoral challenges from across the ideological divide, there is no reason to make overtures to the other side or even acknowledge it at all.

            Politics is not physics, and the laws of politics are not as immutable as they might be in the hard sciences.  But even Isaac Newton’s universal laws of gravitation had to give way when Einstein realized that they were but a special case of general relativity applicable to the world in particular ways and under particular circumstances.
            That's why the median voter theorem still lives.