I was fortunate to be a guest Saturday morning on the show Up With Chris Hayes, where I discussed the latest jobs numbers, and the policies that might make those numbers improve.During one segment, we discussed the varying theories for why unemployment is so bad, and a clip from Bill Clinton’s speech was played, during which he said this:
We do need more new jobs, lots of them, but there are already more than three million jobs open and unfilled in America today, mostly because the applicants don’t have the required skills. We have to prepare more Americans for the new jobs that are being created in a world fuelled by new technology. That’s why investments in our people are more important than ever.
He didn’t use the term, but what Clinton was saying was that a large part of the unemployment problem was “structural” (the unemployed don’t have the proper skills) rather than cyclical (the result of of a lack of demand).
On this point, Bill Clinton was incorrect. All of the evidence shows that lack of demand is the problem, not the structure of the economy.
There are a few different lines of attack against Clinton’s argument.
The first is simply to point out that 3 million job openings is not very high.
Years into this recovery, the number of job openings across all non-farm industries (on a population adjusted basis) is still incredibly low.
So mostly we can say that lack of job openings, not unfilled jobs, is the big story.
But there’s much more.
At Jackson Hole, economist Edward Lazear of the conservative Hoover Institute published a paper titled: The United States labour Market: Status Quo or A New Normal?
It looked at this question of cyclical vs. structural unemployment specifically looking at which industries have lost and gained jobs during the bust and the comeback.
Here’s a key paragraph from Lazear’s paper regarding construction workers, a key block thought to be in the crosshairs of the structural shift in the economy (away from housing, etc.):
There has been discussion that the most recent recession has been characterised by a large increase in the unemployment rate of construction workers and the unemployment rate will not decline to previously seen levels until the housing market recovers and the unemployed construction workers find work again.
The data in Table 1 show that the construction sector had a large impact on the increase in the unemployment rate between November 2007 and October 2009. During the recession, 19.4 per cent of the increase in the national unemployment rate was accounted for by higher unemployment among construction workers, even though construction workers make up less than 10 per cent of the workforce (8.4% in 2007). During the 2001 recession, 9.9 per cent of the increase in the national unemployment rate was accounted for by higher unemployment among construction workers. Construction is, not surprisingly, more important in explaining the rise in unemployment in the 2007-09 recession than in the 2001 recession. But also note that 21.5 per cent of the moderate decline in the unemployment rate since October 2009 is attributable to the decline of the unemployment rate in the construction sector. In this sense, the symmetry associated with the onset of and recovery from the 2007-2009 recession resembles the symmetry associated with the onset and recovery of the 2001 recession.
That doesn’t mean that some industries don’t bleed labour into other industries. The mix of the economy obviously changes over time, as workers leave fading industries (manufacturing) towards booming industries (healthcare). But the point is that there’s nothing unique to this moment or this crisis that would suggest that this has gotten particularly bad or that there’s an acute mismatch between the jobs available and the skills out there to fill them.
In the paper, Lazear even takes on the concept of skills “missmatch” in a truly fascinating way. In his paper, he quantifies missmatch across sectors.
These below tables are a little complex, but worth figuring out. What they show is the relationship between job openings and unemployment across various sectors between 2006 and 2011. The vertical line is job openings in an industry. The horizontal line is unemployment. Not surprisingly, the general pattern is that in most industries, a greater number of vacancies is associated with a lower level of unemployment, and that lower vacancies are associated with a higher level of unemployment.
From prior to the recession to 2011, you can see that the slope of the diagonal line las gone down, meaning that the general trend has been for there to be more unemployment, and lower vacancies (obviously) but for the most part, the same industries are in the same places relative to each other. Education and health services have always had a high number of vacancies and low unemployment. Construction and manufacturing have consistently had a high level of unemployment, and low vacancies. Other industries are right around the average with respect to both unemployment and vacancies.
If there had been some profound economic shift, you could expect to see dramatic changes somewhere, with some industries suddenly having a lot more job openings (relative to unemployment) than they did before. You don’t see that. The relative tightness of the labour market stays fairly consistent within each industry, accounting of course for aggregate labour market trends.
So Lazear’s paper definitely dispenses with a lot of the missmatch error. He goes into other areas as well, for example debunking the idea that “housing lock-in” (People unable to move due to the fact that they are underwater on their homes) is a major problem for the job market.
Prior to Lazear’s paper, the most prominent work on the cyclical vs. structural unemployment subject came from economist Amir Sufi at the University of Chicago and Atif Mian of Cal Berkely.
What Sufi Mian did was look at unemployment by county, and he found that it was highest in areas where there was the highest household leverage going into the bust.
Here’s the abstract to his paper titled What Explains High Unemployment? The Aggregate Demand Channel:
A drop in aggregate demand driven by shocks to household balance sheets is responsible for a large fraction of the decline in U.S. employment from 2007 to 2009. The aggregate demand channel for unemployment predicts that employment losses in the non-tradable sector are higher in high leverage U.S. counties that were most severely impacted by the balance sheet shock, while losses in the tradable sector are distributed uniformly across all counties. We find exactly this pattern from 2007 to 2009. Alternative hypotheses for job losses based on uncertainty shocks or structural unemployment related to construction do not explain our results. Using the relation between non-tradable sector job losses and demand shocks and assuming Cobb-Douglas preferences over tradable and non-tradable goods, we quantify the effect of aggregate demand channel on total employment. Our estimates suggest that the decline in aggregate demand driven by household balance sheet shocks accounts for almost 4 million of the lost jobs from 2007 to 2009, or 65% of the lost jobs in our data.
Note that the non-tradable sector refers to that part of the economy that produces goods or services that must be done locally, and thus has max exposure to the economic situation in its vicinity.
It may seem obvious that unemployment would be highest in the most leveraged sectors going into the bust, but actually it’s profound: It means that it’s the household debt overhang that has kept employment subdued. In pace with less overhang: More employment.
Here are two key tables from their paper.
The first shows that generally, high debt to income counties have had worse employment prospects.
The second pair of charts: emphasise that it’s the non-tradable (by necessity local) industries that have suffered the hardest.
Photo: Sufi and Mian
Bottom line: The real unemployment pain is associated with high household debt levels.
Between this and Lazear, the idea that unemployment is about massive industrial shift and people not having the right skills looks completely wrong.
There’s other evidence for the structural case as well. One fact is that unemployment among college graduates is unusually high (by historical standards) which is not something you would expect to see if this were a skills or education crisis.
Here’s a look at the ratio of unemployment among people with just a high-school education vs. those with at least a bachelor’s degree. The number bounces around a little, but generally we’re right around the average level.
One chart that people constantly bring up when talking about structural unemployment is this one, showing that there’s this freakishly large block of people whoa re “long-term unemployed” and thought possibly to be permanently out of the game.
But this doesn’t hold up either. As Mike Konczal has pointed out, a given job opening is more likely to be filled by a long-term unemployed person a fact that’s at odds with a huge mass of unskilled workers.
This chart from Konczal shows that the odds of getting a job at any given moment have moved down significantly for people in all different durations of unemployment, but that actually it’s the short-term unemployed who are the most worse-off relative to where they were pre-crisis.
Bottom line: Big thinker types like to talk about things like structural unemployment, because it sounds smart, and long-termy, whereas just saying it’s about demand sounds crude and short term. In America, we don’t like to say that problems can be solved just by throwing money at them.
But the unemployment problem is about demand, and when demand comes back, so too will the jobs.
Business Insider Emails & Alerts
Site highlights each day to your inbox.