Calculating Obama’s statistics

Big league baseball recognizes an informal standard known as the “Mendoza Line,” a minimum batting average needed to stay on a Major League roster. This threshold is named after Mario Mendoza, a slick-fielding Mexican shortstop who played for three clubs in the ’70s and ’80s. Mendoza could catch and throw, but struggled to hit the ball. Though his career batting average was .215, he spent lots of time below that, giving rise to the saying that a weak hitter now risks “falling below the Mendoza Line” when his average dips below .200. No matter how good your fielding, you can’t stay in The Show below that number.

We need consensus on a political Mendoza Line for incumbent presidents like Barack Obama. What is the surrogate for batting average and what is the minimum threshold of performance to stick around for a second term? Taking cues from the established Mendoza Line, we need a nice round number, like .200. And we want something that’s easy to compute with trusted data, like batting average. The standard should also be an absolute number, not a measure of change. Mendoza had relatively better years, but his legacy is the solid .200.

A lot of attention is given to the unemployment rate, of course. But there are several statistical problems inherent in using it to define a political Mendoza Line, not the least of which is the small number of “at-bats” for presidents seeking reelection. Mendoza had hundreds of plate appearances to set his standard. In presidential politics, there are only 11 cases that apply. Nate Silver, a sophisticated number cruncher and blogger for The New York Times, has described the relationship between the unemployment rate and reelection as “maddeningly inexact,” but acknowledges his own newspaper’s reporting that “no American president since Franklin Delano Roosevelt has won a second term in office when the unemployment rate on Election Day topped 7.2 percent.” He estimates that Reagan might have been reelected in 1984 with 7.5 percent unemployment, given the Gipper’s fat margin in that election. Neither 7.2 nor 7.5, however, seems like a round enough figure for a proper line in the sand, so I guess we’d have to round up to 8 percent to achieve consensus around an incumbent’s drop-dead line for unemployment rate.

Another standard sometimes used to define a Mendoza Line for incumbents is garnering 50 percent of the vote in pre-election polls. Savvy pollsters never look at the gap between the candidates or the challenger’s percentage. Keep your eyes on the incumbent. So a poll that puts the incumbent at 49 percent and the challenger at 44 percent, with 7 percent undecided, would be seen as a probable loss for the incumbent, despite the incumbent’s 5-point advantage. The logic is that everyone knows the incumbent, so if they are not voting for him now, they will eventually be voting against him, even if seemingly undecided at the moment.

Yet another possibility is approval rating in polls like those conducted by Gallup. But, again, there are statistical problems and stylistic ones. In additional to the small sample challenge, changes in polling methodologies across time makes this tricky business. Furthermore, it seems like a non-even number, like an approval below 47 percent, might be the number. No good for Mendoza purposes.

A nifty Mendoza Line might be fashioned from foreclosure rates. I estimate that a president will struggle winning a state if more than one household in 50 is in default. That’s a nice round figure. Results from the last two elections suggest that the top 10 foreclosure states will weigh heavily in deciding who wins. Keep an eye on it. Maybe one day we’ll speak of an Obama Line that was defined by defaults in the 2012 election cycle.

David Hill is a pollster that has worked for Republican candidates and causes since 1984.

---