A Difficult Decision
Whether you are happy with its outcome or not, 2016’s election is a political scientist’s dream. Love them or hate them, Donald Trump and Hillary Clinton shined a bright spotlight on the economics of politics, especially in the area of political endorsements. As an American, I must note that it is objectively surprising just how many scandals each candidate managed to survive while still managing to secure the endorsement of a major party, and in the case of Donald Trump, win the general election. In the article linked below, Andrew McGill interviews Professor Kevin Zollman of Carnegie Mellon University about a very specific facet of the 2016 election: how individual GOP members, even after a cascade of revelations about the personal conduct of their own candidate for president, never quite managed to “dump Trump.”
For context, it is important to note the personal character of candidates for public office is an important factor in many voters’ choice of who to vote for, often forcing candidates to present themselves as the future conscience of the nation while simultaneously painting their opponent as a base person of little credibility. The behavior of politicians in this circumstance is simply rational behavior: if you want votes, give the people what they want. So how then, in an election which placed such a heavy emphasis on the character flaws of each candidate, did rational politicians who themselves either have to or will have to answer to voters manage to stomach endorsing someone whose public image clearly did not fit within their moral framework? The answer is surprising and reveals an interesting truth about information cascades: they can be stopped, but how?
In the article linked above, Professor Zollman’s answers this question though an analysis of Trump’s strategy in response to the release of the Access Hollywood tape. It boils down to a simple threat: unendorse him and there will be consequences. The consequence, in this case, was the agitation of Trump supporters by calling those who unendorsed him “hypocrites,” a label which can put the brakes on any politician’s aspirations in any liberal democracy. Trump, in this case, managed to change the game politicians were playing by changing what we in class labeled the vg value when calculating expected payoffs in the Accept/Reject scenarios we analyzed (In this case, GOP members had the choice between accepting or rejecting their candidate). For GOP members, the value of not selecting to reject the candidate was magnified because an endorsement now reduced the threat of a particular politician being viciously attacked by the most animated block of voters within the GOP. Politicians in the Republican Party were therefore forced to, however weakly, accept the candidate put in front of them by not unendorsing him. From a game theoretic perspective, this stopped a cascade of people realizing that, as Zollman states, “nobody was happy with the state of affairs” and started a cascade of people in the other direction of people either reendorsing or at least stating their intention to vote for Trump. This was accomplished by the simple threat of being branded a traitor to the party by its most enthusiastic voting block. As Professor Zollman notes, politicians do indeed not like to be “left out in the cold.” Therefore, in the end, retraction became too unpalatable of an option for incumbents.
Professor Zollman’s analysis in this interview is indeed spot on, and provides a new way to view the 2016 election in terms of information cascades. Further, it shows that stopping an information cascade is possible, and that a crude, yet effective, way of doing this is a well placed threat which changes the payoff equation for individuals. His view of the behavior of individual GOP members in endorsing Donald Trump as an information cascade is also exactly in line with what we learned in class. Though individuals held private reservations about their candidate, they nevertheless swallowed these, as for them the most rational choice was to follow the crowd.