I’ve gotten to the point where I would like to see if I can get some articles published at a higher level—places like Washington Monthly. Nothing too big, but a lot bigger than Frankly Curious. And so I’m working on some longer articles with original research. As part of that, I’ve been reading a lot of political science on the effect that the economy has on presidential elections. I’ve read various estimate that about 40% of the results of these races can be explained by the economic trend leading up to the election. This is an idea that Ray Fair did a lot of work on in the early 1970s. (You can get an overview of his work in a pdf, Reﬂections on Macroeconometric Modeling.) But that is rather complicated. I’m not interested in predicting the size of the wins. Just looking at it from a binary standpoint (who won), I’ve noticed something quite simple.
I looked at the unemployment data from the Bureau of Labor Statistics from 1948 onward. At this point, I haven’t done a detailed statistical analysis, but I’ve done enough work to note a startling result: if the unemployment rate is going up, the incumbent party loses and if the rate is going down, it wins. I’m not going to go into exactly how I did the calculation because it is rough and I’ve already improved upon it. But I think the results are pretty clear.
In that period, there were 17 presidential races. If the unemployment rate changed by less than 10% (Not percentage points!) I called it a tossup. There were four of those races: 1948, Truman-Dewey (Dewey had a slight edge but lost); 1956, Eisenhower-Stevenson (Eisenhower had a slight edge and won); 1976, Ford-Carter (Ford had a tiny edge but lost); and 2000, Gore-Bush (Gore had a slight edge).
The 2000 election is a special case that I should be be more clear about. When I get to a slightly more coarse discussion later, I put this in the win category for Gore. This isn’t because Gore won the popular vote. I don’t think the popular vote matters. Politicians know the popular vote doesn’t matter, so they don’t campaign to win it. The electoral college is what matters. But in the case of Gore, I believe he did win the electoral college. Just because the Supreme Court stepped in and nullified an election result does not mean that the election result didn’t occur. I fully accept the fact that this model is useless if the Supreme Court steps into a presidential race and just proclaims one of the candidates the winner. (I can come up with a model for that too.) So keep that in mind. But remember, it was a tossup election regardless.
The one example of a presidential candidate losing when the tend in unemployment said otherwise was 1968, Humphrey-Nixon. Humphrey had a good, but not overwhelming economic advantage. He also had some disadvantages. The biggest I think was the civil rights legislation of the Johnson administration. The race was close, but if George Wallace had not run, it might have been a blowout. It is hard to imagine that many Wallace voters would have gone for Humphrey, but I really don’t know enough about the election to say. But let’s not forget that this was the election when the Republicans invented the Southern strategy that they are locked into to this day. The Democrats did not yet have an answer for it. It was the political Rumble in the Jungle when Ali beat the stronger Foreman by inventing the “rope-a-dope” strategy.
One thing I can say about that race was that this was a time when the American economy was booming. The month before the general election, the unemployment rate was 3.4% and it had not be as high as 4.0% for two and a half years; it hadn’t be consistently at 5% since Johnson ran in 1964. So it’s possible people didn’t care so much about the unemployment because it was already about as low as it could go. Regardless, my little model is wrong about it.
Removing the tossup contests, the model predicts 12 out of 13 races, or 92%. If we use the tossups, the model predicts 14 out of 17 races, or 82%. These numbers seem pretty good to me, and hardly surprising. Political scientists have long claimed that the economic trend is the biggest single effect on presidential elections. But the metrics they use like GNP and inflation rates are not things that directly effect most voters. But unemployment affects everyone, even they aren’t unemployed.
What’s especially interesting about these results is that from the 1980 election onward, the model is perfect. This is why I’m studying these data. Both of the major political parties are convinced that they win elections for reasons that have little positive, perhaps no, and very likely negative effects on their political fortunes. For example, Republican elites are certain that raising taxes is the worst thing in the world; people hate taxes and so they must never raise them. Democratic elites are certain that the only way to win the White House is to be a moderate.
Since I don’t want to get into the other article I’m writing, let’s look at the 2008 election. Did Obama win because he was a moderate? Did McCain lose because he was too moderate? Not at all! The unemployment rate at the beginning of 2008 was 5.0%. By October it was 6.5% and it was clear to everyone that it wasn’t leveling off. McCain could have promised to eliminate all taxes and made a credible case for having a secret plan to cure cancer and the Republicans still would have lost. On the other hand, if the Democrats had nominated Karl Marx himself, the Democrats still would have won.
What I find frustrating is that I think all the political elites (at least on the Democratic side) know this. But they want the conservative policies they advocate for their own reasons that have nothing to do with winning elections. But they have the voters cowed. We are told it would be terrible to nominate Elizabeth Warren or Bernie Sanders. And the Republicans are told it would be terrible to nominate Ted Cruz. But the truth is that if the economy improves throughout 2016, the Democrats will win pretty much regardless of who they nominate. And if the economy tanks, the Republicans will win; it won’t matter whether they nominate firebrand Ted Cruz or “moderate” Chris Christie.