Conventional accounts of retrospective voting assume that voters can use changes in their own welfare to discern whether incumbent leaders have governed well or badly. We challenge this assumption, arguing that the connection between incumbents' performance and voters' subjective well-being is likely to be highly random. As a result, retrospective voting is harder than it seems, and incumbents often get rewarded or punished for events beyond their control. We illustrate this point by tracing the electoral impact of a clearly random event-a dramatic series of shark attacks in New Jersey in 1916. We show that voters in the affected communities significantly punished the incumbent president, Woodrow Wilson, at the polls. We also develop a simple formal model demonstrating that the randomness of "blind retrospection" significantly impairs the effectiveness of elections as mechanisms for political accountability.
The 1916 New Jersey shark attacks resulted in the deaths of four people and one injury, and partially inspired the book and movie Jaws. According to the paper Wilson -- himself a New Jersey governor and Asbury Park resident -- held cabinet level meetings to discuss the crisis, though there wasn't much the Federal government could do beyond sending in the Coast Guard and the president's time was beginning to be dominated by the deteriorating situation in Europe.
Of course, Wilson won the election, even without New Jersey, by a razor thin margin, so in the end the sharks didn't have a huge impact on history, though it's tempting to wonder how President Hughes might have handled U.S. entry into World War I if the maneaters had made it down to, say, Virginia Beach.
My favorite detail from the paper is that the incredulous New York Times, skeptical that a shark would ever travel as far north as New Jersey, put the story on page 18 with the headline, "Dies After Attack By Fish."