Is it is or Is it isn't a Recession:

And Who Cares?

Economists tend to equate higher output with higher social welfare. People are generally better off when the economy is humming along, compared to when it isn’t. Higher output is associated with more consumption, more jobs and more opportunities to invest. All else the same, these are all good things.

Fine, easy stuff to understand. So why is there such an uproar about “recession” these days. Well it is the first part of that other contentious word “stagflation”. So if we are in a recession, combined with high inflation, then we are truly in a stagflation period. And if the economy is not humming, well then we are not doing so well. Again, easy to understand.

And with stagflation, comes special challenges for our policymakers, most of whom are unfortunately not well equipped for the challenge. We miss you Larry Summers and Ben Bernanke.

I want to talk about something else: the political economy side of the term recession. The issue is how does the state of the economy impact election outcomes

It seems like there is considerable attention in the political discourse about this term, in particular deciding whether or not the US economy is currently in a recession. A recent press conference of Treasury Secretary Yellen was spent avoiding questions on this topic. Does this label matter so much?

It does seem that this attention within political circles is reasonable, given the significance of the state of the economy in determining election outcomes. But the attention given to declaring the US in a recession or not is much less important.

Going way back (actually way way back), a Yale economist Ray Fair has been researching and writing about the connection between the state of the economy and election outcomes. He generally finds that when the economy is not doing well, the predicted vote share of the incumbent in the next presidential election is lower.¹

The point I want to make is simple: the measure that Fair uses in his empirical work is simply the growth rate of real GDP per capita, let’s call it y. So, back to the ongoing discussion of defining a recession, for the purposes of predicting election (both Presidential and Congressional) outcomes, whether or not the economy is in a recession per se is not relevant. The information about output seems to be full captured by y.²

From this perspective, two quarters of negative growth may be called a recession by some, but not by others. And it may lead the NBER to eventually declare that a recession began sometime in early 2022. And some time later we will read papers trying to determine how much of the recession of 2022 was caused by oil price increases. Etc. Etc. Etc.

But if you listen to Ray Fair, this is all a bit off the point. Voters, it seems, just look at the numbers, not the labels attached to them. The effects of low growth on political outcomes are there, regardless of whether someone calls it a “recession”.

1. Looking at presidential outcomes, the estimated coefficient on real GDP growth per capita is 0.673 in the November 2018 paper. This means that if the growth rate was 1 percentage point lower, the vote share of a Democratic incumbent president would fall by 0.673. Here the vote share of Democrats in 2018, for example, was 51.163. So that share would have become 50.49. Of course his results over all the years are much richer and nuanced than this single coefficient on y. It is interesting to see how these relationships change over time, as he shows in his many papers. And of course, the 2024 Presidential election isn’t until, wait, 2024!

2. I looked but could not find among the many papers that Fair wrote on this topic, an analysis that uses the NBER business cycle dates rather than y. Through his many versions, the level of economic activity is captured by real GDP growth per capita, with inflation entering as a separate argument in the regression.