Predicting earthquakes—and presidential elections

Anxiety-inducing horse-race presidential polls are everywhere. So, take a deep breath and remember: polls are a snapshot, not predictive. Gallup, the gold standard in opinion research, had Carter beating Reagan by eight points just weeks before the 1980 election; Dukakis beating Bush by 18 points in the summer of 1988; and Romney beating Obama in their final poll of November 2012.

This is context for the work of Allan Lichtman, a history professor who has spent more than forty years working to understand the structures underpinning election outcomes.

That work started with a chance 1981 dinner party encounter with a Soviet geophysicist who had a theory—untestable in the non-democratic USSR—that the science of earthquake prediction could sharpen our understanding of elections. In the view of Lichtman and Vladimir Keilis Borok, presidential elections are not about a Democrat versus a Republican or Biden versus Trump, they are about stability versus upheaval. Does the American public stick with the incumbent president or party (stability) or do they vote for change (upheaval)?

Lichtman and Borok studied every presidential election dating to 1860, searching for indicators associated with stability or change. They distilled 13 key indicators of electoral success, each either true or false. Taken together, these indicators have correctly predicted every presidential election dating to 1984, save one: the election of 2000.

Click the above and you'll see that four of the indicators are political; two are economic; several focus on policy, social unrest, the presence or absence of a presidential scandal and leadership in foreign affairs. The final two address candidate charisma, a measure of sweeping personal appeal, boxes only Ronald Reagan and Barack Obama have ticked.

In a recent interview with Errol Louis of NY1, Lichtman refrains from making an outright prediction on 2024. But our application of his methodology suggests a clear path to re-election for President Biden this fall. 

More important than his predictions, however, is Lichtman's view that governing—not campaigning—is what counts when it comes to winning presidential elections. That's an unintuitive but encouraging message for our hyper-partisan times: good governance is good politics.