What's Fair?

That's what Democrats are asking now with the increased possibility of the Clinton-Obama contest going all the way to the convention in August. On Chris Matthews yesterday Ron Allen said this was most important factor in picking the nominee, especially if Obama wins more pledged delegates with the popular vote while Clinton wins more superdelegates and insider party support. Then Matthews teased, "You mean, the Democratic party should be democratic?!?"

This is a bizarre box the party has put itself in again and again. E.J. Dionne wrote another article last week laying this out, "Dems' Agonizing Obsession with Fairness." Interestingly, the article's html title of the web page is: "The Dems' Losing Obsession."

The real problem is that our political elites, especially Democrats, have never explained how every social choice voting mechanism is flawed in some way, even majoritarian direct democracy. This has been demonstrated by Arrow's impossibility theorem and Condorcet's voting paradox.

So, what's fair? That depends. In democracy models the outcome usually depends on procedural rules. Unfortunately, the myth persists that merely tallying up all the individual votes across the nation somehow yields the most 'fair' result. The Democratic party has propagandized this view to its own detriment.

A voting mechanism should be chosen according to what the desired result is and then reverse engineered. In democracy the best result is one that arrives most efficiently at a single choice that gains the acceptance of a majority of the participants. The desired result is NOT the maximum happiness of the greatest number of people and there's an important difference. There are winners and losers in elections and everybody's not going to be happy. And in a country divided up so that no candidate gets a majority of the popular vote, such a result is virtually impossible.

We should understand from the difficulties of both Obama and Clinton that identity coalition politics is a disaster for a majoritarian voting system based on such a foggy notion of fairness. The Republicans don't seem to have this procedural problem, even though they seemed more fractured when the nomination process began. In short order they settled on John McCain as the nominee. How? Because they designed a nominating system of winner-take-all, closed primaries yielding a clear winner to most Republican voters. Are all the Republicans happy? Hardly, but they have a nominee the majority can accept. This is going to be difficult for the Democratic party if the split between Obama and Clinton is roughly 50-50. Half the voters are going to feel cheated and nobody knows how they will react. So much for unity.

Another clear case of how Democrats misinterpret voting is the 2000 election result and the subsequent backlash against the Electoral College. The objective of presidential voting is not making the maximum number of US voters happy - it's to pick a president that the broadest representation of American interests will willingly accept. The ultimate objective again is not voter "fairness," but the enduring strength of the union of states. This is why the geographic distribution is just as important as the magnitude of the popular vote. And when the popular vote is split 50-50 (i.e., indeterminant), the distribution becomes the most important factor. This is as it should be, but why that's so is a topic for another discussion.