Not all macroeconomic problems are recessions
By Scott Sumner
I recently argued that uncertainty over Brexit did not lead to the recession that many forecasters expected. Actually ‘argued’ is the wrong term, “pointed out” is more descriptive. But I continue to believe Brexit was a mistake, and that it will hurt the UK economy.
Tyler Cowen has a reply to my recent post on the Brexit shock:
Scott Sumner has a very good post on that question, noting that UK gdp growth has been robust and suggesting this refutes uncertainty-based theories of the business cycle. I see the matter somewhat differently, however. In standard real business cycle theory, a’la Long and Plosser, a business cycle is defined in terms of comovement and persistence. The real business cycle “victim” can either take a direct hit to wealth, or by various processes of smoothing and substitution, spread the hit out across various sectors and over time, thereby generating comovement and persistence and thus what we call a cycle. Taking the direct, concentrated hit isn’t a “cycle” but it still is very painful, in most simple versions of the model it is more painful than doing the smoothing. . . .
I think Scott is a little too distracted by not seeing cyclical action at the usual intermediate frequency.
A while ago I estimated the costs of this “hit” at 5,625 pounds per capita, though since then the British pound has fallen even further, thereby raising those costs.
Just to be clear, Tyler’s estimate of the cost of Brexit to UK consumers in no way conflicts with my post. I think Tyler misinterpreted me as saying that Brexit didn’t harm the UK (which is understandable, as the post was poorly written on that point, before I revised it.) I do think Brexit (when it happens), will reduce British GDP. So Tyler and I agree on that issue.
Where we may disagree a bit is the business cycle question. The meaning of the term ‘business cycle’ is fairly clear, and nothing Long and Plosser say will change that meaning. At least since Mitchell, the business cycle has referred to co-movements in many key macro variables that play out over a few years. Most economists think of a business cycle as a significant move in macro data such as unemployment and real output, seasonally adjusted, relative to trend. There’s really no question what business cycles look like, all you need to do is look at RGDP growth rates, or the unemployment rate. For instance, here is the unemployment rate for the US:
That’s what economists mean when they talk about business cycles. I encourage people to focus on the 11 recessions shown on the graph. In every case, the unemployment rate rose by at least 230 basis points. They are very easy to see, as during non-recession periods the unemployment rate never rises by more than 80 basis points. Thus macroeconomists have a very well understood concept in mind when they talk about recessions.
Now it’s true that really bad things can happen that which make a country poorer, which do not show up as recessions. One example might be protectionism. Or Ludditism. Or the 2011 earthquake/tsunami in Japan. Those things might result in a large loss in wealth without being recessions, in a technical sense.
I think Tyler missed the point of what I was trying to do with my UK Brexit post. When we have an event that everyone agrees is a recession (with rising unemployment and falling RGDP), there is often debate about the causes. In my view, tight money (falling NGDP) is usually the cause. Others often claim that real shocks were the problem, and a counterfactual where the central bank kept NGDP growing would not have prevented the recession. Instead (in their view) you would have ended up with an inflationary recession—like 1974. I believe, for instance, that this is Robert Barro’s view of the Great Recession.
If someone doesn’t like my definition of recession, I can simply rename periods of sharply higher unemployment as “bananas“. In that case, my claim is that uncertainty shocks do not cause bananas. (Those not around in the 1970s will think I’m losing my marbles.)
The British natural experiment tells us that uncertainty shocks do not cause periods of sharply rising unemployment and falling RGDP. It tells us nothing about the impact of Brexit itself, which hasn’t even happened, and doesn’t even tell us whether the UK residents are already seeing a drop in wealth due to anticipation of Brexit. Thus it’s still possible that Brexit uncertainty causes harm, just not banana-type harm.
I actually think it likely that the uncertainty does not cause much harm, but that Brexit itself will cause modest harm. The fall in the pound is certainly an indication that Brexit is expected to be a significant problem for the UK economy, as Tyler suggests.
My post wasn’t really about Brexit at all. I was using Brexit uncertainty to show that one explanation of recessions, sorry bananas, has been removed from consideration. That makes it even more likely that periods of sharply rising unemployment and falling RGDP are due to falling NGDP growth, caused by tight money. There are still a few competing real business cycles theories of “recessions” (as the term is conventionally defined), but one less theory than a year ago.
PS. Don’t underrate the importance of the very weak pound (soon to be called “ounce sterling”). Without a sharply lower pound, I would consider Tyler’s argument to be very weak. But with the pound at $1.25, I think the argument is pretty strong, at least 70-30 that Brexit will significantly reduce British GDP growth.
PPS. Tyler also said:
As for the uncertainty theories, I’m not sure the Brexit story is a good case study for them. At first I was uncertain as to whether it really would happen, but not very much any more.
In the six months after the Brexit vote, there seemed to be great uncertainty about how it would play out, hard or soft. It’s still not clear what arrangements the UK will be able to negotiate, particularly for industries such as banking. But in the end neither my view nor Tyler’s view matters. The point is that the consensus of economists thought it created great uncertainty, and the consensus of economists expected a dramatic slowdown in UK growth as a result. And they were wrong. Either economists don’t know what uncertainty looks like, or it has no cyclical effect. Either possibility is really bad for uncertainty theories of the cycle.
Here is Tyler back in July, expressing what I believe was the conventional wisdom:
I think there is a pretty simple story here. Brexit increases uncertainty, both in the mean-preserving sense, and in the “very bad outcomes are now more likely” sense, and that lowers investment. That in turns shifts back the aggregate demand and aggregate supply curves, and a recession may result. Less than a year ago, MIT economist Olivier Blanchard published a major paper on capital inflows being expansionary, and now of course we are seeing the reverse. Toss in some negative wealth effects for further transmission. I was at an event in The City a few days ago where the anecdotal data about postponed or cancelled deals seemed pretty overwhelming, and this is consistent with what one reads in the papers as well, not to mention with basic economic theory. It is true of course that we don’t know how large these effects will be, but the more purely British measures of equity value are still down quite a bit.
The uncertainty Tyler refers to did not suddenly go away after he wrote the post. Rather the UK started doing better than expected, and that created the impression that the uncertainty was being resolved. But in a policy sense the Theresa May government was very vague about its plans for quite a number of months.
The term ‘uncertainty’ must have some sort of independent meaning beyond “the economy is spiraling downwards” in order to have an important causal role. (Just as ‘bubble’ must mean more than “prices went up and then down”.)