Skip to Main Content
Speaking of the Economy
Speaking of the Economy - Mark Bils and Niklas Engbom
Speaking of the Economy
Jan. 12, 2022

The Flow of Human Capital

Audiences: Economists, Policymakers

Mark Bils of the University of Rochester and Niklas Engbom of New York University discuss their research on how easily labor is substituted or shifted between jobs and its implications for policies to bolster human capital. Bils and Engbom presented their most recent work at the Richmond Fed's CORE Week in December 2021.



Jessie Romero: Hi, I'm Jessie Romero, director of research publications at the Richmond Fed. Thanks for listening to "Speaking of the Economy." You can find more episodes on our website,, or subscribe via Apple Podcasts. If you like what you hear, please leave a review and tell your friends.

I'm talking today with Niklas Engbom, an assistant professor of economics at New York University, and Mark Bils, a professor of economics at the University of Rochester. Niklas and Mark are here in Richmond for our second CORE Week, which brings together economists from a range of disciplines for a week of collaboration and ideas exchange. I'm excited to talk with them today about the papers they're presenting this week, both of which address the important role that human capital plays in both individual and aggregate outcomes.

But first, we'll start with some basics. So Niklas, what do economists mean when they talk about human capital?

Niklas Engbom: It's a broad concept, for sure. Broadly, I'd say it's a mesh of a worker's knowledge and skills that you can use in some productive way to produce goods or services, or maybe even ideas.

Mark Bils: I'm trying to stay close to Becker's definition, but it's basically anything where you give up something today for a gain tomorrow.

Anything that gives you a market or welfare benefit is a form of human capital. The difference is kind of embodied in the person. The example I would give [is] I learned how to make a very good cake from just practice. That's my human capital. If I came up with an idea for a better recipe and I wrote it down, that's not human capital. That's not embedded in me and it would be passed on, even though they lead to the same quality of cake.

Romero: I like that. It's the first time I've ever heard cake used as a metaphor in economics.

Let's start with your research specifically. With your co-authors Baris Kaymak and Kai Jie Wu, you're studying labor substitutability. What does that mean?

Bils: We're looking at substitutability, particularly with respect to schooling. Over time, there's been a tremendous increase in schooling. People see this in the U.S., but it's at least as striking in other countries. We're focused on a lot of countries, upwards of 60 to 100 countries over the last 60 years. What you would see is that the amount of years of schooling in the population is tripled in a typical country. In the U.S., we've seen many, many people go to college. It wasn't that long ago … it was more rare.

We're looking at substitutability across the schooling groups. In particular, if you get this increasing in schooling over time, what does that do to the return to that schooling? Does it tend to drive it down because I get more and more of a certain type of education group, and so they're crowding each other out and that lowers the return on schooling? That's what we mean by substitutability.

Romero: Speaking as a non-economist, it seems fairly obvious that a worker with a high school diploma wouldn't be a good substitute for somebody with a master's degree. That seems pretty intuitive. So why is it something that it's important to study more precisely and understand in more detail?

Bils: I'm going to go into a little bit of a different [area] — what economists mean by perfect substitutes.

Romero: Okay.

Bils: Perfect substitutes doesn't mean [trading] somebody with a college degree and somebody with much less schooling. Even if they're perfect substitutes, it doesn't mean they're equal substitutes. It doesn't mean that there's no value to the schooling, that the person with less years of schooling will be said to be comparable. It means that, even though there's some gain to having a more productive worker, if you suddenly shift the economy to having a whole lot of people of the one type with lots of schooling, it doesn't necessarily drive down the relative ability to trade one for the other.

The parallel would be we're going to have an aging workforce and we're getting older and older and older. Maybe age has some benefits, and we see that wages are higher for older workers. So, you wouldn't say necessarily the experienced worker is equal to a new person [being hired]. But if I hired two new people and they spent a lot of time sorting out, making more mistakes and how they should do it, they might be able to be a fairly good substitute for [a] more experienced worker. So there's substitutability and being a perfect replacement. Those are two separate things.

I'll give a couple of reasons why it's important. For instance, if we consider policies that subsidize schooling — and schooling is one of the most subsidized things of our actions — one of the arguments for subsidizing schooling is in terms of inequality of earnings. As a practical matter, though, and when economists look at the effect on inequality of subsidizing schooling, very little of the impact comes from taking people who would have finished high school and then convince them or encourage them to get more years of schooling. In some sense, they're going to get higher wages, but they're going to get the wages later. There's still a trade-off for that person.

Most of the action is that more people go to college and the return on schooling falls and that's what leads to the greater equality. That's the argument which is often made. I mean, otherwise, you look at a lot of the college subsidies and they're really upper middle class transfers, to a large degree. That's often like a justification. But if the schooling is relatively good substitutes, then as more and more people go into schooling it won't really have such a big effect on the route of wages of people with more or less schooling.

Another example, which comes up often when we talk about immigration, [is should] we have more selective immigration like in Canada, where you bring in more of a threshold in terms of skills and assets that come in. The argument is, well, this affects workers in the U.S. who would be less skilled if we bring in less skilled immigrants. That's all completely dictated by how strong the substitutability is. Our particular emphasis is on looking across countries, but I'll just leave it at that.

Romero: Okay, great. Well, that's a great segue to my next question.

We won't get into the technical details of your paper, but we will have a link on the show page for listeners who are interested in them. But as I understood it, the upshot is that you and your co-authors find a much higher degree of substitutability that has been found in previous research. How does that affect our understanding of what's driving these cross-country differences?

Bils: When we look across countries, per-capita incomes are 30, 40 times higher in richest countries as poorest countries. And the logical question is then, could poor countries catch up by investing more? In particular, is there such a high return on schooling? That's led to a lot of work trying to say, how important are differences in human capital in rich and poor countries?

If you do a quick calculation and in rich countries people have six or seven more years of schooling, what's the benefit? When we look at the market in terms of what somebody [gets with] 12 years of schooling versus six years of schooling, it's sizable. Supposedly it doubles your earnings. "Doubles your earnings" is a big number, but relative to a factor of 35 it's very small. This leads to the conclusion that it must be other factors, not schooling or school quality, that matters.

But key in that assumption is that it's just more years of schooling in rich countries. What people have argued, very soundly, is that maybe the quality of schooling is much higher in rich countries, so it's not so much the years of schooling but the quality of the schooling. The problem with that argument is that the market returns don't look higher in rich countries than in poor countries, so to justify why the quality is so much higher than I have to have it, they're very poor substitutes. So schooling is much better in rich countries, but there's already so many people [with better schooling] that the return is low, so I don't realize how much higher the quality is.

Romero: Okay.

What implications do you think your findings might have for policies to address inequality, either across or within countries?

Bils: I would just go back to what I was talking about before. If I think about subsidizing schooling, I think it suggests that there will be some effect. We're not arguing that the schooling groups are perfect substitutes.

I'll just put our numbers in context. If you look at the U.S. over the last 40, 50 years, [there was] a huge increase in schooling. If you use the sort of numbers that people have often used in the literature, if there had been nothing else that had happened and you just had this big increase in schooling, then the return of going to college would have gone from being like 40 or 50 percent to minus 20 percent or minus 30 percent. So it's very extreme in that sense. Our numbers say if there hadn't been these other factors changing which drove up the demand for schooling, just the extra people going to college would have probably driven the return down from 50 percent to 20 percent.

So it's not like we're saying there are no perfect substitutes, but it does matter. It's very key for talking about any policy that affects schooling and what that [does] imply for inequality. If I subsidize schooling, I drive down the return. I indirectly help the people with less schooling that's directly related to this. The higher the substitutability, the less effective would policy be. And then again, it's saying I don't have to worry so much about if we get a set of immigrants say who are less skilled. It's not going to move things so much.

Romero: Great. Thank you.

Niklas, the research that you're presenting this week also looks at cross-country differences. But you're looking at a different driver and that's job-to-job fluidity, how readily workers can move from job to job. How does this fluidity kind of compare [and] vary across countries?

Engbom: Before the study, we didn't know all that much about this particular form of mobility, job-to-job mobility … It's seen across countries. In the U.S. context, we kind of know from the literature that this is a very important source of lifecycle careers. Can workers move over time to better jobs? It's an important driver of wage gains for workers.

In my study, I collect data from a set of developed countries. I find large differences across countries in the rates at which workers switch from one [employer] to another. It turns out that these patterns are also correlated in a way you maybe wouldn't expect, based on the U.S. evidence we have. In a country where people do switch a lot [between] employers, you also see steeper wage growth over the life cycle, consistent with the narrative we have in the U.S.

Romero: Okay. Sounds like the U.S. does tend to be more fluid.

Engbom: Yeah, that's for sure.

There is a lot of dispersion across countries, There are other countries that are similarly fluid as the U.S. — the U.K., for instance, but also some countries in Europe. You tend to think about European in general as a very non-fluid kind of area and the U.S. as a high fluid area. That's actually not entirely true. There's a lot of dispersion within Europe, too, in terms of fluidity and the labor market policies that governments have implemented. Countries like Denmark, for instance, have a fairly flexible system and that shows up as a high rate of worker flows across employers.

Romero: Okay. I didn't realize that there were those differences within Europe. We tend to think of it all as one block.

Engbom: Yeah, there's very much a distinction between, say, southern European countries. Some central European countries generally have a little bit higher fluidity. The Anglo-Saxon countries — the U.K. and the U.S. — that have very high levels of fluidity.

Romero: Is the U.S. job market more fluid than it used to be? I think you hear a lot of people saying, long gone are the days when you could work for the same company your whole career. But how accurate was that characterization, and has it changed?

Engbom: That's a very interesting question. I think it's a debated question. If you ask a layman, they often tend to say what you said that fluidity is higher now, these lifetime jobs are more prevalent in the past.

If you actually look at the data, there's more of a nuanced story. It is true, I think, that if you look at the tenure distribution, the very right tail has become a little less thick, it's thinning a little bit. There is a sense in which these lifetime jobs have become less prevalent today than in the past. I tend to think about that as unionized manufacturing jobs, maybe. But [there are] very few workers that actually hold those type of jobs. Even 30 years ago, that was not that common.

For the average worker, it looks like measures of job-to-job mobility have gone down in the U.S. People switch employers less now than 30 years ago.

Romero: Wow. That's definitely not what I would have expected ...

One of the things you find is that greater fluidity is associated with greater wage increases. What's the mechanism there?

Engbom: Right.

The idea is that by switching an employer, that's a way for you to effectively boost the return on your skills. You move to an employer that maybe values your skills more, in some sense. But in switching more, you drive up the price of your skills over time in the labor market. If you endogenize skill accumulation on the job, then you could see how that expected faster return and growth in the return to your skills could also incentivize people to accumulate more skills in the first place. In my model, higher fluidity is associated with both a steeper growth in the price of skills but also the underlying quality of skills.

Romero: So the idea would be, if I think I have a good chance of going and getting a higher paying job somewhere else, I'm going to go ahead and invest in more skills right now in the hopes of then landing that job.

Engbom: Yeah, exactly. If you're in a job that doesn't really value your skills, there's not much point in accumulating skills. But if you think that, later on, you can at least on some job where you can better use your skills, you might end up responding and accumulating more skills.

Romero: Are there things that policymakers could or should do to enhance labor market fluidity?

Engbom: Even more than targeting fluidity, I would argue that policymakers would maybe try to subsidize on-the-job training in some way. As Mark mentioned, there's a lot of subsidies to formal education prior to entry to the labor market. But we have a lot of evidence now that there's also an important accumulation of skills on the job and, according to my analysis, that might be under-provided in the market.

Romero: I think both [of] your papers are very relevant in the current environment. We're hearing a lot about employers lowering their skill requirements so that they can draw from a larger pool of workers We have an article coming out about that in Econ Focus and we'll have a link on the show page. And the quits rate is at record highs, which implies that a lot of workers are changing jobs.

Do you think that your research can help us think about public investment in education and the value of employer-based training in the current environment?

Mark, I'll ask you that question first.

Bils: What I would speak to is, to the extent people are substitutable, it suggests that this can work. Could it be a successful strategy? I think it would be a very big improvement in the U.S. labor market if there was less emphasis on using schooling as a screen. [With] many jobs, people go to college and they work jobs that aren't really related to what they learned in college. Some of that is some skills they learned and social skills in college. But I think some of it is a signaling that the markets are using.

I think a more fluid labor market where (a) firms will take a chance and be able to hire people and let them prove themselves and (b) be able to fire them easily if they don't pan out [would be] a much better system. I hope that's one of the benefits that comes out of COVID is that we get to a higher fluidity but also more experimentation in the labor market. There's going to be a lot of positions where they're going to say, let's just try to invest more in the interviewing process and find people who don't necessarily have the years of experience [and] don't necessarily have the years of schooling but there could very well be extremely successful. People are reaching and trying to push themselves up and be as productive as possible, given their certified characteristics in terms of schooling, etc.

Engbom: I totally agree with Mark. My work emphasizes a lot of scope for learning on the job. If we get some people that maybe have a little bit lower qualifications on paper into these jobs, there's still a tremendous scope for learning stuff on the job. Maybe you just have to get through that gateway, and then you can accumulate a lot of the skills you need for that particular job.

Romero: That's encouraging to hear.

Last question — how are you enjoying CORE Week?

Engbom: I think it's been fantastic. I'm in the same room with Mark Bils. [Laughter] No, it's been fantastic.

Bils: No, it's fabulous. The Bank has an array of fabulous economists and all the people coming in. And then it's set up beautifully with activities and enough presentations with enough time to interact outside of the presentations.

There's a summer institute that the NBER [National Bureau of Economic Research] would have and it was like that. There were few papers you go [see presented], you'd have time to walk around and go get ice cream and interact with people. There hasn't been anything quite like that where the day's got lots of things going on but not so packed that you can interact with people. It's been great.

Romero: Well, I'll take the comparison to the NBER. It's very high praise, indeed.

It's really a pleasure to have you here. Thank you so much for coming and thank you so much for talking with us. I really enjoyed it.