Presidential Elections: There’s a Theorem for That
To follow up on my colleague Aaron Nathans’ roundup of AIER blogs on the election, I want to highlight some economic perspectives on elections as they may relate to this one.
One classic insight is Arrow’s impossibility theorem, originally published in a 1950 paper by Nobel Prize winning economist Kenneth Arrow. It shows that there is no way to ensure any voting system, will always simultaneously satisfy a list of seemingly desirable conditions. More generally, it makes the same point about the impossibility of aggregating individual preferences to societal preferences.
The biggest sticking point is usually the “independence of irrelevant alternatives” condition, which states that societal preferences between options A and B should only depend on individuals’ relative rankings of A and B – in other words, no spoiler third options.
In our general elections, this usually comes up in the possibility of a third party candidate taking votes primarily from one side and thus allowing the opposing candidate to win. It’s undesirable because the presence of the third candidate changes the voters’ rankings of the two leading candidates. Some people propose a preferential voting structure, such as “instant-runoff” voting, that would allow people to rank their choices and eliminate this spoiler effect.
But there are some “rock-paper-scissors” situations this can’t resolve. As a simple example, suppose there were only three types of Republican voters, with different preferences over the final three candidates. Type 1 prefers Trump, Cruz and Kasich in that order; Type 2 prefers Kasich, Trump, Cruz; and Type 3 prefers Cruz, Kasich, Trump. (These were the final primary vote tallies in North Carolina, Ohio, and Utah, respectively.)
No matter who is nominated, two of the types prefer someone else. Suppose, as in reality, Trump is nominated. Two of the three types, Types 2 and 3, would rather have Kasich than Trump. But then Types 1 and 3 prefer Cruz to Kasich, and Types 1 and 2 prefer Trump to Cruz. Any of them could win or lose a two-man race. There is no perfect choice.
Other studies look not only at who ends up winning, but how we got there. One much more recent paper than Arrow’s classic, measures the effects of political uncertainty, in particular uncertainty about presidential elections, on implied stock market volatility in the United States. They find “the presidential election process engenders market anxiety as investors form and revise their expectations regarding future macroeconomic policy.” Not only the policy but its changes can be important; as a 2013 cross-national study shows, policy volatility can lead to negative impacts on economic growth.
If instead of reading economists’ thoughts on politics, you want to cut out the middleman and simply vote for one, that’s an option too. FiveThirtyEight recently ran an interesting feature on Laurence Kotlikoff, a Boston University professor who will appear on the ballot in two states, and a man who I originally came across in his work on risk sharing within U.S. extended families.
Click here to sign up for the Daily Economy weekly digest!