Auctions in Political Elections
http://www.pubchoicesoc.org/papers_2011/Macdonell_Mastronardi.pdf
This article highlights the importance of auction theory by modeling elections as auctions and showing that the strategies that determine the amount of campaign money to spend are simply strategies to maximize payoff in an auction. The auction setup is as follows: the candidates are bidding campaign money in hopes of winning the election. However, this auction is an asymmetric all-pay auction where one candidate may have an advantage and both of the parties must pay the amount that they bet regardless of whether or not they win the election. Therefore it is slightly more complicated than the auctions discussed in class. The authors of this paper make a few other assumptions about the auction. There are a certain number of voters (represented as Bd and Br) that will definitely vote for one candidate, regardless of the campaign money spent, while other swing voters(u) are won based on the amount of campaign money spent by the candidates(Ed and Er). Therefore, the democratic candidate’s vote total is Bd + u(Ed/(Ed+Er)) and the republican candidate’s vote total is Br + u(Ed/(Ed + Er)). The goal is to determine the values of Ed or Er to win the election and spend the minimum amount of money as possible. The authors go further to find the optimal strategies for the cases where 1) the number of resolute voters determine the election entirely and case 2) where the swing voters play a substantial role in the election.
The most interesting takeaway from this article is that given that each candidate has a given amount of campaign money and that the district in which they are running for election in has a measurable bias, each candidate has a distinct strategy that they should follow to determine how much money they should spend on the election. However, in practice, this seems a bit too simplistic. If the candidates knew exactly how much campaign money to spend, and knew exactly how much campaign money the other candidate would spend, it would be possible to determine the winner immediately. One candidate would be able to spend just enough campaign money to ensure a majority and it would be useless for the other candidate to contribute anything at all. Auction theory suggests that while the strategies are determined, it is impossible to know the value the other candidate has in mind and how much he/she is willing to spend. However, this explanation is not entirely sufficient. Unlike real auctions where the other person has no idea what the other person is willing to pay and therefore has to play a strategy based on uncertainty, the information about the election is likely attainable to the candidates. For example, in certain elections (ex: presidential elections) it is safe to assume that the value to each candidate is so great that they are willing to spend as much campaign money as possible. Also, with the Federal Election Campaign Act of 1972 candidates are required to disclose their campaign expenditures (to protect against corruption). Therefore, much of the uncertainty around the auction itself is available to the candidates and it is possible to determine what the other candidate’s optimal strategy is. This leads to the same question posed earlier: if the values of the auction are completely determined and each candidate knows exactly what the other candidate should do, why isn’t the outcome of the election determined. I believe that the true answer here demonstrates the flaws of applying auction theory to practical situations. The actual dependence of the number of votes obtained on the amount of campaign money is incredibly complicated, so the goal now becomes trying to best model the political environment rather than outplaying the opponent in an auction. This is what is observed in practice, as candidates spend their time adapting to the political environment rather than only scheming about campaign funds. However, this auction model is very interesting nonetheless because it gives insight to how candidates should react in an idealized situation.