Last week I argued that people who aren’t busy over-interpreting political forecasting models are under-appreciating their meaning and import.
Models are, by definition, simplifications of reality and are necessarily incomplete. If they were complete descriptions of reality, they would be reality itself and too complex for us to understand.
Critiques focused on the fact that models are sometimes wrong and don’t include all possible sources of human behavior are simply critiquing models for, well, being models.
I know as well as anyone that models can be wrong. By my count,
FiveThirtyEight’s usually excellent directional predictions have been “wrong” in six gubernatorial and Senate races since its inception. I’ve been directly involved in four of those six contests.
The question models present is, “Can we learn something about a reality too complex to apprehend by simplifying it and looking at a few factors?” And the occasional mistake doesn’t obviate the fact that the models reveal some important things about the world of politics.
Models based only on fundamentals — presidential popularity, economic changes, midterms vs. presidential year — tell us a lot about political outcomes. They aren’t perfect, they cannot tell us everything, but they do reveal a great deal. And those fundamentals exclude the choices of campaign managers and consultants, party committees and presidents.
Josh Katz of The New York Times’s The Upshot team estimated that models based solely on fundamentals before campaigns were really engaged were only about 5 percentage points worse than polls in projecting Senate outcomes between 2004 and 2012.
Think about the implications of that. In most races, candidates, their big moments, their gaffes, their strategists and their strategies, make precious little difference. To be sure they can have some impact, but not nearly as much as practitioners, pundits, reporters and voyeurs like to believe.
When candidates and their teams make a difference, it’s usually because they are running against the tide of history, not with it. One year some Republican colleagues grumbled about Democrats getting campaign awards in a GOP year. Winning a race you have a 90 percent chance to win, just based on the fundamentals, is no great achievement. Winning one where your probability of victory is 10 percent is a real success.
The models teach us that historical tides count for a lot. I don’t know how this election cycle will turn out, but this newspaper has already carried articles about the blame game that will develop if Democrats lose the Senate.
Blame games are interesting stories, but the stories themselves aren’t necessarily true.
Forecasters tell us that such attributions of responsibility are just wrong-headed. In May, The Washington Post’s Election Lab, using a model based on fundamentals, gave Republicans an 82 percent chance of taking the Senate. A trio of academics using various fundamentals and differing simplifications of reality — that is, different models — projected a Democratic loss of just over six seats.
To be clear, I’m not at all sure those predictions are correct, and I work hard every day to make sure they turn out to be wrong. If they do prove inaccurate, lots of credit will be due to lots of players. But if they prove reliable, there is no one to blame but circumstances. Odds were against us from the start.
Last Saturday, Jews read the book of Ecclesiastes in synagogues around the world. At one poignant moment in the book, King David, a pretty successful guy, writes, “The race is not to the swift, nor the battle to the strong, neither yet bread to the wise, nor yet riches to men of understanding, nor yet favor to men of skill; but time and chance happen to them all.”
Time, chance, circumstances and historical forces matter. We’d all do well to learn that lesson from the forecasters.
Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982. Current clients include the majority leader of the Senate and the Democratic whip in the House.