Presidential Party is a Bad Predictor of Economic Success

There is, you might have heard, a meek debate happening around the 2016 election. Some people think we should put more emphasis on economic policies. Others think we should stay the course. Before I offer my take on that, I would like to submit that economic policy is not an especially obvious source of economic growth.

The graph below shows the per capita GDP growth of all the quarters since July 1948. I start presidential terms in the third quarter of their term, the idea being that is the first full annual budget they administer. The horizontal axis is the percent of quarters that are below the quarter in question. The vertical axis is the percent growth.


The graph is a surprisingly good test of your willingness to read a graph in a partisan way. It shows that there is not a large differences, on whole, between Democrats and Republicans. It also shows there is a measurable advantage on average to voting democratic, especially around a recession. If you, for example, so a chasm between them, you might want to consider the ways you twist information to support a Democratic worldview. Likewise, if you cannot see the clear advantage of voting Democratic, you should consider that you are very partisan.

To grasp the data here, imagine that electing a president in 2016 was merely having the new president draw, at random, a quarter from a previous president in their party to have that growth again. This is obviously a gross misrepresentation of Presidential power, but it gives a sense of what this kind of backwards-looking analysis says. For virtually every draw our Republican president would make, the Democrat would have a better chance of doing at least as well. (There are some exceptions around the 95th percentile.)

So, if the Democrats drew their 50th percentile quarter, the GOP would only have a 37% chance of beating them on a draw, which makes them the superior choice on average. If the GOP drew there 50th percentile as well, growth would only be fifth of a percent lower. So while the Democrats do nominally better, the difference over most of the distribution is quite small, averaging to a third of percent. For the median earner, this is about $8.25 a month

One place that this simple exercise suggests a stronger Democratic advantage is around recessions—the very left of the graph. The Democrats have fewer quarters with very deep decreases in per capita GDP. At the widest gap, there is nearly a percent difference in income contraction. This is approximately $25 a month difference. Still not an incredible sum, but more noticeable. This seems to validate expansionist policies, a Keynesian glut of spending, in a recession. I suspect that this accounts for a fairly large part of the finding that Democrats do better than republicans. Make of that what you will.

This is not an unexpected finding for a few minutes with Excel. Let’s take a widely cited 2004 study by Achen and Bartels (PDF). (Bartels is the author of Unequal Democracy, which makes use of these findings.) Their study implies* a higher difference between Republicans and Democrats, at about half a percent. For the median earner, that is $11.75 a month.

Tomorrow’s post will make use of this observation that the difference in growth between parties is small, but for now let us draw a simple conclusion. The Republicans and Democrats have surprisingly little effect on overall growth on average. One way to explain this difference by pointing out that treating Reagan and Bush or Clinton and Carter as cut from the same cloth is a bit silly; I am inclined to agree. I think it is reasonable to say that if all you care about is average income growth, you can detect a slight advantage to voting Democratic. But it is easily overwhelmed by policy details and simple economic luck.

*Table 5 gives values for election and non election year income growth. (This is a slightly different metric than the one I chose, but comparable enough.) Using a weighted average, it appears there is just shy of half of a percent difference between democrats and republicans. The authors do not report this number in their study. You might note that they draw the conclusion that voters are very bad at factoring this small difference and credits the ways both parties tend to have higher growth in election years. I am inclined to say that voters are very bad about caring about 12 dollars a month when it arrives so abstractly, but at any rate, I’m not really here to disagree with the original study.


2 thoughts on “Presidential Party is a Bad Predictor of Economic Success

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s