Seeing this post on Fivethirtyeight made me chuckle, since I had made exactly the same basic claim a few days before the election: where the economy is weaker, the stronger the support for Trump. Now that the official results are in, one can run a better correlation than I did, comparing apples to apples, not the estimated support with some caveats, drawn from polling data. If one were to plot the changes in Democratic voteshare (of the major party votes) between 2012 and 2016 against the job creation by states since 2008 relative to the national average, we get the following graph:
The Y-Axis is the change in Democratic voteshare and the X-axis is the job creation since 2008 in each state, relative to the national average–i.e. how far a state raced ahead or lagged behind the rest of the country in creating jobs, which, as pundits remind us, has not done badly on average, but, in truth, has been pretty variable by geography. This is not a clean graph: while there is an apparent positive correlation between job creation (or the lack thereof) and the changes in votes, it is not so obvious enough to make the correlation very large (corr = about .25–which is roughly what shows up in the model with the polling data as well.) The numbers are mangled by some states where the economy did well (North Dakota) or at least not bad (Oregon) but still saw huge drops in support for Clinton, relatively speaking, but I could not find systematic means of accounting for them, even though it is not too difficult to conceive reasons why individual outliers should be, well, outliers. Still, a large majority of states lie in the appropriate quadrants, consistent with a positive correlation–more jobs, more votes for the Democrats. Of course, the caveat is that, since the numbers are as messy as it is, this looks apparent only in retrospect–I suppose if I did post the earlier version of this plot, there will have been many claiming how I’m seeing Jesus in a piece of cheese, and they’d have been right, of course. I suspect the 538 folks had access to better data than I, but this does add another datum to the maxim, “it’s the economy, stupid,” and the economy is not just the average.