SxSW 2008: The Myth of the Rational Voter

by Luke Wroblewski March 10, 2008

At SxSW 2008, Professor of economics Bryan Caplan discussed many of the theories underlying the wisdom of crowds and provided contrary evidence to the theories of intelligent decision-making by groups in his The Myth of the Rational Voter talk.

  • The “miracle of aggregation” is defined as the situation where large numbers of uniformed people balance each other out when voting. Since most people are uniformed, they vote one way or the other and the few people voting that are informed tip the tide.
  • In other words, the result of uniformed crowds is a wash but a few experts can shift the group to a better decision. This only works with very large numbers. In theory, with the miracle of aggregation, you would get the same outcome if everyone were fully informed as you would if a very large number of people were uniformed.
  • It seems too good to be true that this system works. But people believe in it because they want to believe in the tenants of democracy.
  • However, there are many counter examples that prove this system does not work. When you look at any public opinion data, it is very easy to find examples that disapprove the “miracle of aggregation”. Example: the percentage of the budget the US spend on foreign aid is actually 1.2% but surveys show people agree it is 20%.
  • Caplan’s assumption is that when laymen and experts disagree, the expert is generally right. He cities discrepancies between economist expert opinion and public voting to illustrate examples where this is the case in his book The Myth of the Rational Voter.
  • Polling can change answers by changing the way questions are asked.
  • Pessimistic bias: people assume things are going to hell and getting worst. Currently the exact opposite is true.
  • If you are reasonable, you always have to be open to feedback and change your mind. If you want to maintain your beliefs, you should not be rational.
  • Economists expect people to have crazy views about politics because they are free. Private cost for false beliefs is low. Public beliefs could be problematic.
  • People who know the least tend to have very predictable beliefs.
  • Public opinion is often wrong because you are asking everyone to vote and it costs them nothing to do so. In betting markets, people have to pay for being wrong so people who vote wrong will drop out because they are losing money. In a betting market, people who are wrong tend to be silenced.
  • Stress greater alliance on experts. Just because most people believe something, does not make it right.
  • If people don’t know the issue –ask them not to vote. This approach could be interpreted as fascist.
  • Purely changing people’s views of a few facts can change votes.