As a general rule, predicting the future is not something historians are typically comfortable with. Most of us believe that human behavior is far too complex to reduce to a deterministic formula, and we generally maintain that the value of studying the past is not that it helps us to predict the future, but that it equips us to face it more wisely.
But not all historians would agree. An interesting exception is Professor Allan Lichtman of American University, a specialist in American political history. I first figuratively encountered Professor Lichtman as an undergraduate when I was assigned his book Historians and the Living Past, a primer on the theory and practice of history. Then in graduate school I did my best to wade through a second book of his with the scintillating title Ecological Inference (an introduction to multivariate regression analysis that was mostly over my head).
Professor Lichtman is in the news these days because he has developed a model that has successfully predicted the outcome of every presidential election since 1984. Lichtman’s basic premise is that presidential elections are referenda primarily on popular attitudes toward the party controlling the White House, and he has devised a series of thirteen key variables that shape the level of public satisfaction or dissatisfaction with the incumbent and his party.
You can follow this link to a brief five-minute interview with Professor Lichtman. Lichtman’s model predicts that the GOP should win this November but he himself is skeptical. Lichtman suggests that the idiosyncratic nature of Donald Trump’s candidacy may mean that his explanatory model will no longer work. If Lichtman is correct, his analysis offers a sophisticated statistical defense for a claim widely made this fall, namely that almost any Republican presidential candidate other than Donald Trump would be heavily favored to claim the White House.