With the midterm elections rapidly approaching, professional forecasters are focusing much of their attention on traditionally red states in middle America, such as Iowa, Arkansas, Louisiana, and West Virginia. Many of these races are tighter than they should be, given the unpopularity of the president and his policies. At the same time, polling companies are admitting that they are expecting greater polling error this year. Part of this, they attribute to the difficulties in predicting turnout, but one cannot help but wonder to what extent rapidly shifting demographics are playing a role in this election’s high degree of uncertainty.
There’s also an inherent problem with simply asking people what they prefer. What people say they want and what they actually want are two different things. Economists recognize that we cannot know people’s true preferences ex ante, but that we can get a clue to these preferences by observing past behavior. In this way, consumers are said to reveal their preferences. When a consumer purchases an apple for a dollar, he reveals that, at that particular place and time, he values the apple more than the dollar. Of course, this can always change. The reality of diminishing marginal returns will mean that the next apple he buys will be worth less to him than the first, if they are purchased in quick succession. But we as economists can generalize and assume that most of the time, the apple will be preferred to the dollar.