Skip to main content
Faculty Viewpoints

What’s the Real Jobs Picture?

The top-line employment numbers seem encouraging, but campaign rhetoric makes it clear that many people feel the jobs market is deeply troubled. Yale Insights spoke with labor economist Lisa Kahn about her research. She argues that long-term trends driven by technology and trade have been amplified by the Great Recession to create new challenges for job seekers.

The economy has added 15 million jobs since 2010 and the unemployment rate has dropped back to where it was before the Great Recession—and well below where it was in 1984, when Ronald Reagan declared it “morning in America.” That kind of economic news is usually a campaign bonanza for the party in power. But the prevailing mood remains uncertain, with many, including young people and blue-collar workers, feeling that the prosperity that previous generations enjoyed has passed them by.

Yale SOM’s Lisa Kahn discusses what’s behind the profound and, in some cases, unsettling changes in labor markets over the last two decades, including the shift to more service sector jobs, the effects of increasing automation, the declining rate of male participation in the workforce, and the lingering effects of the Great Recession.

Q: How would you describe the labor market in the U.S. today?

I think we’ve seen a lot of improvement since the Great Recession. Unemployment is back where it was. Employment and jobs are growing. Wages are now growing for the first time, and, in the last year, they’ve grown across the board, from the low end to the high end of the income distribution. There’s a lot to be pleased about with the labor market, especially relative to 2008.

But there are longer-run changes to the labor market which existed long before the Great Recession that are troubling. In particular, some workers have lost their jobs to technology and cheaper overseas labor. A large swath of these workers have yet to find exactly where they can be productive elsewhere when, say, their factory job is replaced by a machine or someone in a country with cheaper labor. Those problems existed before the Great Recession, but in some recent research I’ve found evidence that the recession may have exacerbated them. So, I would say the economy is still struggling with these issues, just as we were before the Great Recession.

Q: What is happening with manufacturing output and manufacturing jobs?

We have certainly seen gains in manufacturing output in the U.S. as well as some small gains in manufacturing employment. When I was working at the Council of Economic Advisors for President Obama, we were able to highlight those successes even as early as 2010 and 2011.

But you see a different manufacturing in the U.S. now than you did before. Our comparative advantage is in high-tech manufacturing, bringing in machines to replace assembly line workers. That’s why we still have very high productivity in manufacturing even though employment has been on the decline for a long time. On top of that, jobs in manufacturing tend to be lost in recessions and then not come back.

This pattern is mimicked in most of the developed world. Even in developing countries, manufacturing is changing. In China, the wages of manufacturers have grown substantially. It’s no longer the place to do your cheap manufacturing. Now, you need to go to the more recently developing countries.

Q: Where do you see manufacturing in the U.S. going?

I think in the U.S. we have incentive to specialize in high-tech, high-skilled, cleaner manufacturing because a richer country can tackle those problems better. That’s our comparative advantage, and I think we’ll continue to do that.

There is huge affection for manufacturing in the U.S. That’s because manufacturing has been a huge part of our success in the past. So now we hear political rhetoric that we need manufacturing to be a successful country. That’s just not true. Actually, we could have a service economy that is very productive.

In 1900, we didn’t have a manufacturing economy; we had a booming agricultural economy. As technological changes rendered most of the employment in agriculture obsolete, people moved into cities and shifted to work on assembly lines and other manufacturing jobs.

Now machines and overseas labor have replaced those jobs. The people who were doing them need to find employment elsewhere, but unfortunately, for quite a few, we haven’t yet figured out where. That doesn’t mean we won’t. That doesn’t mean there isn’t a productive role for those workers in the labor market.

Q: This election cycle has been heated and polarized. Part of that has been giving voice to those who, despite an extended period of job growth, are frustrated by their job options.

As you can tell by our political climate these days, there’s certainly a sense that some people think things have not gotten better. We’ve seen declining participation of men in the work force for a long period, as well as stagnating wages.

Jobs that you can get with a high school diploma or some college are hard to find. But we still have demand for very skilled jobs at this education level, like plumbers, welders, and craftsmen. We also need personal service workers: nurses, medical technicians, house cleaners, janitors, and protection service workers. These jobs have seen demand grow substantially at the same time as technology and outsourcing have eroded manufacturing jobs. But former factory workers or young men who would be taking factory jobs if they were available do not seem to be transitioning into jobs where we clearly have booming demand.

Q: You mentioned that recessions can amplify the impact of automation and outsourcing. Could you explain that?

I have some research that showed these effects are exacerbated by recessions. This goes back to Schumpeter, who hypothesized that recessions are a cleansing period where the least productive firms die off. This lets new firms come in and produce in a different way.

Even inside stronger firms, there can be a tendency to delay adopting new technologies. A pharmacy might like to bring in self-scanner checkout machines that save them on labor costs, but it’s costly and will create inconvenience for customers during the transition, so they might wait.

It turns out that recessions might be a very good time to make this kind of change because if you have to stop production for a while, fewer people are buying. If you have to lay off workers, everybody understands that you have to do that in a recession. And as things pick up, you can hire workers with the skills to use the new technology. For all these reasons, adopting new technologies and deciding to move production overseas might actually be concentrated in recessions.

When these changes happen gradually, it may be easier to help small groups of people acquire the new skills they need to move into another sector. But if the impacts of a recession are being amplified by a large number of firms changing how they produce, a large swath of the workforce is losing jobs all at once and seeing their skills depreciated, that’s going to generate some really big problems. It probably contributes to jobless recovery.

For government policy, this means being cognizant of structural change around recessions and potentially expanding need for re-training, in addition to subsidizing unemployment.

Q: Could you explain the phenomenon of job polarization?

In the U.S. and a number of other countries, we’ve seen a decline in employment and wages in the middle of the skill distribution at the same time as there has been increases in employment and wages in high-skilled and low-skilled jobs. Research has been pretty successful at tying these changes to automation and, to some degree, offshoring.

The jobs that are codifiable, that is, made of tasks that can be turned into a repeatable roadmap, tend to be in the middle of the skill distribution and the middle of the earnings distribution. Those are the jobs that are done more efficiently by machines, and that’s where automation has had the biggest impact: manufacturing, or the bank tellers who were replaced by ATMs, or the clerks who did payroll, word processing, or data entry.

Machines haven’t yet developed the versatility, judgment, analytical thinking, and tacit understanding associated with jobs at the higher end of the skill distribution. And machines have not been able to do the manual work that janitors, house cleaners, or line cooks do. So, we’ve seen employment growth and wage growth at the ends of the skill distribution where the work can’t be replaced by machines.

At the same time, we’re seeing increasing skills required for the jobs that are available. In the research I referred to earlier, I looked at job vacancy postings to study the skills that employers want and how that has changed over time. We have seen increases in requirements for education, experience level, and computer skills, as well as increases in key words associated with critical thinking and analytical skills. That’s been across the board in the job vacancy postings but especially in cities hit hard by the Great Recession.

Is it just because firms decided, a lot of people are unemployed, so we can get a better worker? Or is that because they actually need workers to produce differently? My research shows that there is some of both, but given that the demand for higher skills continued even after labor markets have essentially recovered, what remains is upskilling due to changes in how work is done.

Q: We think of automation impacting blue-collar workers the most. How much is it reshaping white-collar jobs?

I think that’s a really important question. We have seen the influences of automation in almost every pocket of the labor market, from the very low end of the skill spectrum to the very high end.

Automation is becoming more and more sophisticated. It’s doing more and more tasks that we previously thought maybe were not codifiable. The future of labor essentially comes down to, where are computers going to replace us and where are computers going to augment us?

A computer can be as good as or better than a human at doing taxes. Accountants have, as a result, been pressured to shift to aspects of their work that demand interpersonal interaction with clients, something like strategizing about portfolios, which requires give and take to establish priorities and tacit knowledge about the whole system. They haven’t lost their job to automation, but the job now involves a different set of tasks, which makes them more productive.

One area where machines are weak is in pretending to be people. Interpersonal tasks have been shown to be growing the most in terms of employment and wages, especially when they’re combined with some skill.

Q: Is there a role for education and ongoing skill training in addressing these challenges?

Education and training have a vital role. We should double down on policies that help ensure that people have the right skills for the right job.

There are very important policies that are already on the table. Investment in early childhood education has a really high return on dollars invested in terms of getting people on track for future jobs. We should also increase access to college and universities through scholarship programs and something as straightforward as simplification of FAFSA, which is the incredibly complicated form required to be eligible for financial aid.

In addition, we need vocational programs in high school and community colleges aimed at both young workers who are not entering the labor market at the rate we would like to see and older workers who have lost their jobs and could very much benefit from training to develop a specific skill that can use to weather the outsourcing and machine revolution.

However, it’s a hard problem because training programs have a mixed record in the U.S., and it’s surprisingly difficult to pinpoint which ones were actually successful and which ones weren’t. Some things that seem promising include programs that work very closely with employers to give workers the skills those employers actually want rather than skills that we abstractly think they’re going to want. A variation on that could be matching workers to employers for on-the-job training opportunities. Community colleges also tend to be very flexible in terms of changing the size of their enrollment and their curriculum in response to immediate demand. We probably have some of the pieces of the puzzle, but we need to work harder to get it right.

Q: What are the limitations to retraining? Does it make sense in a community where there aren’t jobs?

There are two approaches to the problem facing places where there used to be a lot of jobs like mining or manufacturing towns that are now struggling. One is a place-based approach where investments and tax breaks aim to help the local economy move to a new phase. Pittsburgh is an amazing example of this; it suffered when steel died, but it has transformed into a hospital and research town with great amenities. It’s doing wonderfully.

In some cases, a second approach may be better. That approach helps people move to where there are currently opportunities. That’s less popular, especially in the political sphere, where politicians look to help the voters who are currently in their district. But I think there’s a lot to be said for increased mobility.

Q: How much is social mobility linked to these issues?

There’s really fascinating work on social mobility in the U.S. by Raj Chetty and others. His team looked across the country and asked, “If you’re born in Arizona, how likely is it that you’re going to be earning more money than your father earned? And how does that compare to someone born in Connecticut?”

It turns out that there are vast differences. There’s very little mobility in the South and the Midwest, but quite a bit of mobility in the Northeast and the West. Why do we see these big differences? It can’t be the case that somebody born in Louisiana is any less capable of achievement than somebody born in New York. An interesting correlation is that places where there’s less social mobility are also places that have generally more inequality—that is, how much money someone at the top earns is very different from how much money someone at the bottom earns.

According to economic theory, there’s not necessarily a link between social mobility and inequality, but the empirical link is present. You might wonder if the extreme poverty or crime or educational institutions of diverging quality that can come along with increased inequality might be why people can’t get beyond where their parents are to reach their full potential.

Inequality is not bad in and of itself. The inequality that says, “You’re more productive so you should get paid more.” That’s good. It’s motivating. It’s incentivizing. The inequality that says, “You should get paid more because you came from a rich family.” That’s demotivating and inefficient.

If I ask, why does a college graduate gets paid more than a high school dropout? Is it because the college graduate came from a rich family? Is it because the college graduate is much more productive than the high school dropout? I think it’s both, and we want to separate out the good inequality from the bad inequality.

This interview was edited for length and clarity.

Department: Faculty Viewpoints