I have following definition of structural unemployment (taken from "Modern principles of economics" by Tyler Cowen and Alex Tabarrok):

Long-term unemployment caused by shocks or permanent features of economy making it harder for certain workers to find a job

So the answer to the question in the title would seem to be "No, because quarantine won't last so long that prices of wages and other factors of production would fully adjust to inflation. So unemployment wouldn't be long-term, thus it wouldn't be structrual unemployment".

But on the other hand, other three types of unemployment (seasonal, frictional, cyclical) don't seem to fit the bill. An example: Suppose prostituion is legal and I'm a prostitute. Due to pandemic the government decides that prostitution isn't an essential business and because it's not something that can be done online brothels are forced to be closed, including my brothel. My brothel is also a small business with very thin profit margins (consequence of very competitive market caused by low price of entering the market). In short, without its workers working it's unable to pay its rent and is kicked out by its landlord, spelling end to it. I try to find another job, but due to many other businesses closing down competition for very few vacant jobs is crazy and they disappear with speed of light. Is my unemployment seasonal? No, we don't have seasonal nationwide quarantines. Is it cyclical? No, because even if economy was booming nationwide quarantine and ban on non-essential businesses would still leave me and many other people without job. Is it frictional? No, I'm fairly certain that there is no a vacant job that is waiting for me to find it. Structrual unemployment seems to be closest to that I experience by its sprit. So, maybe the definition is too narrow and it's possible for unemployment be both short-run and structrual?

  • 1
    $\begingroup$ Having businesses shut down by the government for a limited period is an action that is extremely rare. All we can say is that it does not fit those categories, and should be a new category. However, some job losses are due to a lack of demand, and would be considered cyclical. $\endgroup$ – Brian Romanchuk Mar 29 '20 at 21:07
  • $\begingroup$ @BrianRomanchuk Is it really cyclical if lack of demand is caused by the government closing businesses? $\endgroup$ – user161005 Mar 30 '20 at 2:50
  • $\begingroup$ I said that that some job losses are due to a lack of demand. That definitely counts as cyclical. $\endgroup$ – Brian Romanchuk Mar 30 '20 at 12:41
  • $\begingroup$ @BrianRomanchuk Saying "definitely" is neither an argument nor an explanation for why jobs lost due decreased demand caused by state-imposed bans should be counted as cyclical unemployment $\endgroup$ – user161005 Mar 30 '20 at 12:57
  • 1
    $\begingroup$ “Cyclical” refers to the business cycle. A collapse in demand - e.g., as seen in the oil market - is part of the business cycle. If we wanted to decompose observed unemployment into categories, I believe that one would estimate structural and frictional, leaving cyclical as the residual. You want to add a new category - unemployment caused by government health decree. That’s fine, but it’s a new breakdown. However, cyclical will still absorb everything not in that new category. And depending on the country, that’s non-zero. $\endgroup$ – Brian Romanchuk Mar 30 '20 at 13:47

There are lots of different people unemployed, and the answer will be different depending on the person. In your prostitute example, if you left your medical assistant job because prostitution paid more, then it's frictional, because there are lots of medical assistant jobs, but it might take you time to go through the interview process and be onboarded. If you have no marketable skills other than sex work, then it's structural unemployment. If there is another job that you're qualified, but you don't get the job because there are so many other people competing for it, then it's cyclical (which is a bit of a misleading term, since there it's not like there's a "quarantine/no quarantine" cycle, at least not yet, fingers crossed).


The issue is how we can decompose changes in unemployment into frictional, structural, cyclical, and seasonal.

The question has been updated to include the source for the definition of “structural” unemployment. However, I do not have access to that text. There is nothing in the definition that allows us to calculate “structural unemployment” (nor the other components).

Nevertheless, we can step back and attempt to answer this question based on conventional definitions.

“Seasonal” is very easy to deal with: the raw data is non-adjusted, and then statistical agencies do a statistical adjustment. We can work with the seasonally adjusted data, and throw out the notion of seasonality.

We are then facing “structural” and “frictional” unemployment. I assume that researchers have proposed methodologies, which are likely to be controversial. In any event, we estimate those, and remove them from the total. Since “cyclical” is the only remaining category, “cyclical” gets all the remaining unemployment.

I am not expert in labour market economics, but have read recent literature. I have seen no recent use of the concept of “structural unemployment.” My opinion is that it is not meaningful. However, I will attempt to discuss the definitions that I seem to recall seeing.

Frictional unemployment is more straightforward: workers cannot instantly be matched to new jobs, so a certain percentage of the workforce will always be in transit between jobs. Studies exist that attempt to estimate this, but I believe the consensus is that this is a small and steady percentage. It will not move much in a crisis.

Structural unemployment is highly problematic. It is supposed to be a steady state value. One typical way to estimate “structural” variables is to pass them through a smoothing filter. Under this definition, the smoothed series is “structural”, and the residual is “cyclical.” Under this definition, the construction ensures that almost any spike is “cyclical.” However, I have never seen anyone proposing a decomposition like this in at least a decade.

Until the methodology for estimation of “structural” and “frictional” changes in response to the pandemic, any rapid movement in seasonally-adjusted employment is “cyclical” by definition.

  • $\begingroup$ It would be useful to define more precisely here the notions of "structural" and "frictional" unemployment, as well as "cyclical" unemployment, as you have done in your comments! $\endgroup$ – emeryville Mar 30 '20 at 22:02
  • $\begingroup$ Like I wrote, there’s no fixed definitions, and any techniques to estimate them are likely to be controversial. You are working from some source for your definition, and I have no idea what it is. From what I have seen of labour market economics, the concept of structural unemployment makes no sense, so I cannot offer a definition. $\endgroup$ – Brian Romanchuk Mar 31 '20 at 0:33
  • $\begingroup$ @BrianRomanchuk Why does it make no sense? $\endgroup$ – user161005 Mar 31 '20 at 3:16
  • $\begingroup$ @BrianRomanchuk has a point. A source for user161005's definition of structural unmployment is required. Otherwise, to say that the concept of structural unemployment makes no sense is a bit too definitive. If you Google "structural uemployment" in gogle scholar, you will find 37,300 results and many very good labor economists. $\endgroup$ – emeryville Mar 31 '20 at 7:35
  • 1
    $\begingroup$ @emeryville I updated my post to include the source of the definition $\endgroup$ – user161005 Mar 31 '20 at 15:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.