Economic context
Changemaker Q&A

Repairing Labor Market Institutions to Deliver Opportunity: A Q&A with David Autor

Elisabeth JacobsFebruary 04, 2021


David Autor

David Autor is a member of the WorkRise Leadership Board and the Urban Institute Board of Trustees. He is the Ford Professor in the Department of Economics at the Massachusetts Institute of Technology (MIT), codirector of the National Bureau of Economic Research Labor Studies Program, and coleader of both the MIT Work of the Future task force and the J-PAL Work of the Future experimental initiative. His scholarship explores the labor market impacts of technological change and globalization on job polarization, skill demands, earnings levels and inequality, and electoral outcomes. Autor has received numerous awards for his scholarship, including the National Science Foundation CAREER Award, an Alfred P. Sloan Foundation Fellowship, the Sherwin Rosen Prize for Outstanding Contributions in the Field of Labor Economics, and an Andrew Carnegie Fellowship, and for his teaching, including the MIT MacVicar Faculty Fellowship. In 2017, Autor was recognized by Bloomberg as one of the 50 people who defined global business. In 2019, the Economist labeled Autor the “academic voice of the American worker,” and that same year, he was featured in a Last Week Tonight with John Oliver segment on automation and employment.

In this interview with Elisabeth Jacobs, WorkRise’s deputy director, Autor explains why simply eliminating low-wage jobs isn’t the answer to building ladders of opportunity and how labor market institutions can be repaired to create economic security for workers. This interview has been edited and condensed for clarity.

Elisabeth Jacobs: In an essay you wrote with Elizabeth Reynolds, you describe the major forces shaping postpandemic recovery: an increase in telework among higher-wage workers, urban dedensification of work away from cities, greater concentration of employment into larger firms, and an increased reliance on automation to perform routine work. You say these forces have resulted in eliminating millions of low-wage service jobs, as routine activities in our prepandemic lives—going out to eat, using public transportation, going to the office—have been scaled back or eliminated altogether. Yet you say we shouldn’t celebrate the reallocation of work out of low-wage jobs and that having too few of these jobs is actually worse than having too many. Explain what you mean by that. What function do low-wage jobs play in our economy?

David Autor: What we would like to see are more alternatives to low-wage occupations, but eliminating them doesn’t actually create those alternatives. There are short-, medium-, and long-term trends to consider. In the short term, much of the employment loss has been in-person services, logically, right? That, of course, would happen because most service work can’t be done remotely. But the question we were focused on is the post-COVID labor market and how much the trend away from in-person services will persist, and our hunch is it will persist a lot.

A perfect example is business travel, an industry that is not expected to recover fully. We realized how much we were travelling before the pandemic and that it isn’t necessary. You just don’t need to cross continents for a 90-minute meeting. You can now do it on Zoom. The larger point is that business travel doesn’t just affect the airline industry but also transportation, hospitality, restaurants, and other businesses, like building security and dry cleaning. Because when you leave your home and stay somewhere else, there’s an entire life-support system that has to be activated to keep you going. This, combined with more work done telepresently, is causing contraction for these service-sector jobs, which have been the growth area for workers—both men and women—without college degrees. And yes, these are low-wage jobs that don’t have much of a career trajectory. They don’t provide much economic mobility.

But simply eliminating low-wage jobs isn’t the answer. At a minimum, we would like those jobs to be competing aggressively for workers, which is what we saw in the years prior to the current recession. We saw rising real wages at the bottom of the labor market, the result of extremely low levels of unemployment and a tightening labor market. Ideally, of course, we would like these workers to pursue jobs with more of a career trajectory, even if they don’t require a college degree. But we don’t accomplish that by just shutting down the jobs we don’t like.

We should be simultaneously trying to move people up the ladder while also making sure the bottom rung—since there’s going to be a bottom rung—affords living standards so people don’t live in poverty, can take care of their children, and have a level of economic and health security.

EJ: So, having those jobs at the bottom of the labor market is important, but ensuring there is competition also helps drive up wages. But is there a way of transforming service jobs so they don’t lead to dead ends?

DA: Sure, so two answers. One is that service jobs don’t have to be as bad as they currently are. For the same McDonald’s job done in the US compared with Denmark, there’s a world of difference between what people are paid, whether they have health care, whether they have vacation, whether they have family leave. Those jobs will always be among the lowest paid in the economy because they don’t require a very scarce skillset, but the US makes them much worse than they have to be. Part of it is a lack of competition for these jobs, but also, we could make policy effort to improve them. Such solutions won’t come for free since it will drive up costs for those services. If wages are higher, burgers will cost a little more and so on. But many would argue that it is a reasonable price to pay. That is one way to create redistribution through markets, and it’s not a terrible thing.

Two, there is evidence that increasing minimum wages raises earnings, reduces poverty, and doesn’t seem to have measurably adverse effects on employment. It’s not free, but it seems to help more than harm the workers we are trying to target. In the long run, we would like fewer people doing jobs that don’t have economic mobility, but we also have to recognize that we will need jobs in our economy for people without qualifications or skills and that the work they do has to get done. We should be simultaneously trying to move people up the ladder while also making sure the bottom rung—since there’s going to be a bottom rung—affords living standards so people don’t live in poverty, can take care of their children, and have a level of economic and health security. There are other countries with economies similar to the US that are able to attain this.

The central challenge we face is the real disconnect between the growth in productivity and the growth in median wages and earnings the typical worker receives.

EJ: I want talk about a few key points you make in the MIT Work of the Future report released in November, which is that an economy based on dynamism and innovation didn’t have to result in a labor market that has failed to deliver economic security and prosperity to so many workers. Labor market policies and institutions have failed to blunt the effects of automation, technological change, and trade. How can policies and institutions be repaired so the market delivers returns and opportunity to a broader swath of the workforce?

DA: The central challenge we face is the real disconnect between the growth in productivity and the growth in median wages and earnings the typical worker receives. So our challenge is how we close that gap or at least get these two trends in sync with one another. Most advanced economies have seen similar divergence between productivity and wages, but like many things, the US is an outlier in how extreme the gap is between productivity and wages.  

So, how have other countries done better? One thing they’ve done is have much more effective mechanisms for worker voice and labor market institutions that mediate between workers collectively and employers. The notion of “worker voice” or “worker power” has a terrible reputation in the US. Part of the reason is that our form of worker collective action is uniquely adversarial and, at least in the recent past, uniquely ineffective. Organized labor has been losing membership for decades, and, in addition, it doesn’t cover a lot of the people we would like to help. For example, agricultural and domestic workers were explicitly forbidden (PDF) from unionizing in the 1935 National Labor Relations Act. This was of course overtly racist, since Black Americans were overrepresented in agricultural and domestic work.

That provision is still on the books today. So we need new forms of collective action. I’m not saying we need to eliminate existing unions, but we need a new path for restaurant workers, agricultural workers, domestic workers, people engaged in app-based gig work. They need to have better forms of negotiating power and collective bargaining, particularly in the pandemic, as so many face potential job loss and/or exposure to COVID-19. We have seen new forms of collective organizing for these sectors, such as Restaurant Opportunities Centers United and the National Domestic Workers Alliance, but these organizations don’t call themselves unions for a reason. But they are trying to raise labor standards and practices within their industries.

Another flaw of our system is that unionization in the US happens one firm or establishment at a time. The bottom line is that no employer wants to be unionized when its competitors are not. Imagine if you were passing minimum wages one restaurant at a time. That would be a terrible idea, as it wouldn’t be viable for the single restaurant who chose to raise wages. Most countries have forms of sectoral bargaining that represent all workers in an individual sector or occupation. The bargaining seems to be less aggressive but also met with less hostility since all firms in that sector have to deal with a unionized workforce simultaneously.

It’s important to think about economic security as not something that just happens when you are working but also what happens when you are not working.

EJ: You also discuss modernizing the unemployment insurance (UI) system and having meaningful minimum wage reform. Which, to you, seems like the most promising pathway or the most urgent?

DA: Passing minimum wage legislation is, in some ways, easier than other things because it doesn’t have a direct fiscal cost to states. Many states are pursuing this path, including states that went for Trump, including Florida, which passed a $15/hour minimum wage in November. President Biden has spoken a lot about raising the federal minimum wage, and now, with Democrats having a slim majority, there may be momentum to do this.

I think reforming the UI system is much harder because it is in such terrible shape. On average, only about one in three people who is unemployed actually receives UI benefits. In many places, you could work full time for half the year at minimum wage and not qualify for UI because qualifying is set to wages and not the number of hours you worked. So a high-wage worker would qualify after one week’s worth of work, while it could take a low-wage worker eight months. That’s crazy. That means that some of the least paid and least secure workers are being de facto disqualified from unemployment insurance. And many states view UI as a cost rather than an insurance program, and they are trying to minimize people getting on it. (Read recommendations for making UI more equitable in the Work of the Future report.)

Raising the minimum wage and reforming the UI system are complementary strategies since they are both concerned with strengthening economic security for workers whose economic security has been in decline. It’s important to think about economic security as not something that just happens when you are working but also what happens when you are not working. A well-functioning labor market needs to provide support for these types of transitions. In our report, we cite Denmark’s example as one model to consider, where there is a well-established state mechanism for providing unemployment insurance, retraining, and pushing people back into the labor market. If the US provided comprehensive training and support for unemployed workers, then we would worry a lot less about what happens when workers lose their jobs.

EJ: You also note that the workforce system designed to help unemployed workers—and really all workers in need of skills and training—is one that even workforce practitioners will say is too fragmented and disconnected from other systems that support workers. How does workforce training in the US need to change to get us to where we need to go?

DA: As you say, the system is highly fragmented, but its strength is that there are many points of entry. In many countries where you have a highly centralized training system, it’s hard to change tracks once you’ve made your choice. Another problem is the quality of training is heterogeneous and often disconnected from what employers want. And even if the training is good, how would someone know that it’s good? Several states have made admirable efforts to coordinate specifically and work with local and regional businesses. But there are a few areas where we could do better. One is formalizing credentials and standardizing training programs so employers have something to react to and targeting certain sectors, such as trade, health care, or IT.

Another part of it may involve giving people some resources or financial credits or assistance to get training, but then we have to be even more concerned about the quality of training they buy. We know that when the federal government just pays for things and doesn’t certify the quality, we get a lot of abuses of the system. One final thing, and this isn’t really the answer, but very much a hope, is we are in an era of innovation in education. We know adults don’t always do well in a traditional classroom setting. Training can be more hands-on, more interactive and personal. Education has gotten so much more accessible with the rise of online learning, virtual and augmented reality, and, of course, YouTube. My hope is that the Biden administration will tackle workforce systems and education and training—particularly for dislocated workers who are eager to be retrained—in a bold, comprehensive, and transformative way.

Share your ideas for research, topics, or events to be featured on Working Knowledge by emailing