Bill Kerr: Higher education’s central role in economic development is well established. But will the affordability crisis and a less than optimal relationship with the job market rewrite the equation? Skepticism about U.S. colleges’ and universities’ ability to deliver talent and opportunity is growing. What will it take to improve the situation?

Welcome to the Managing the Future of Work podcast from Harvard Business School. I’m your host, Bill Kerr. My guest today is Professor David Deming of the Harvard Kennedy School. David is a faculty Co-Lead of the Harvard Project on Workforce, along with my HBS Managing the Future of Work Project Co-Chair, Joe Fuller. David joins me to talk about the Project on Workforce’s College-to-Jobs Initiative. The research effort combines a playbook and map-based data-visualization tool. The aim is to shed light on the effectiveness of common approaches to bridging school and career. The College-to-Jobs map combines data on college graduates and job openings at the region level. The tool is intended to spur research and informed decision making by stakeholders in education, business, policymaking, and beyond. Welcome to the podcast, David.

David Deming: It’s great to be with you. Bill. I’m really excited for the conversation.

Kerr: David, it’s always a pleasure to welcome a Harvard colleague with similar research interests. Maybe for our listeners, tell us a bit about your background and also a bit about the Project on Workforce.

Deming: So I am an economist by training. I started at the Graduate School of Education and joint appointed between the School of Education and the Kennedy School, where I am now serving as Academic Dean, in addition to my role as Co-Chair of the Project on Workforce. My research focuses on education, economic growth, economic development, jobs, skills, the future of work, and kind of everything in-between. So did some work early in my career on early-childhood development, Head Start, programs like that. Looked at K–12, school accountability, school choice, things like that. And then, eventually, started working in higher education and have increasingly been focused on how do we build better and smoother pathways between education and training institutions and the labor market—and not just first jobs, but also jobs that have pathways to upward mobility, jobs that pay family-sustaining incomes, and all that.

Kerr: And tell us a bit about the Project on Workforce.

Deming: So the Project on Workforce is a collaboration across multiple Harvard schools. The co-leads of the project are me, your colleague Joe Fuller at the Business School, my colleague Peter Blair at the Ed School, also Raffaella Sadun at the Business School has joined us quite recently, and Bob Schwartz is retired but is kind of like Co-Lead Emeritus. And we think that this is a really exciting cross-school initiative, because we have folks on the team with different perspectives. At the business school, you all tend to focus a lot on employers. Here at the Kennedy School we focus on policy and the connections in-between. And, of course, our colleagues at the School of Education focus on educational institutions. And so we try to catalyze research policy and practice around these areas of job training, workforce development—primarily in the U.S., but with a comparative perspective, trying to learn lessons from other countries. And we do the research, but then we also talk a lot with policymakers. We place students in workforce related internships. And we bring in speakers and just try to be a clearinghouse for all things workforce, try to build a knowledge base, and push us toward a better equilibrium for jobs and careers.

Kerr: That’s great. David, you’d mentioned—and I don’t think any of our listeners will doubt that going from college to work is an important transition—but what prompted this new initiative about that particular progression point?

Deming: Well, you know, Bill, I really got interested in this question when I started to look in a comparative way at what we do in the U.S., versus other countries, in terms of job training, workforce development. The United States spends about 20 percent of the OECD average on so-called “active labor market policies”—so this is things like work subsidies, job training, apprenticeships, a variety of internships, a variety of programs that help smooth this pathway. So we spend $1 for every $5 spent in European nations. We are the richest country on earth, and we spend much, much less on this problem. So if you are a person of means, if you have family connections, you understand how to go from college to career. You use your personal connections, you come to college knowing what you want, or you follow in the footsteps of your parents or whatever it is, and you kind of make it through. If you’re at a four-year college, there are a lot more institutions that help connect people to good jobs that, at either two-year colleges or less-resourced four-year colleges, those don’t exist. And so people just get lost in the cracks. And I thought, “Look, we’re spending all of this money on education and training as a society, and I think we’re wasting a lot of value by not helping people make this really important transition.” So that’s how I got started working on this question.

Kerr: And your report starts with some rather startling statistics, as well. It says that the college earnings premium appears to be in decline, maybe for the first time, and that a number of college graduates—40 percent—are in jobs after school that don’t have a bachelor’s degree [requirement]. So maybe talk a little about that macro picture.

Deming: A college degree is still a very good investment for people on average, but there are tons of variation in how well it works out. So what that means is that many people go to college and don’t get connected to a good job and don’t necessarily earn the kind of income that could pay back their student loans or help get them on the pathway to mobility. More colleges should be intentionally focused on integrating work-based learning into their curricula. And there are a lot of colleges that are doing this already very successfully, and we talk about these. And that’s actually what students want. If you survey college goers, if you survey parents who are paying tuition bills, they say they want—most of them say they want—colleges to be more work focused. You know, we don’t want to make it just job training. It’s true when you’re in college, you learn a broad set of skills—and I think that will always be true and should be true—but you can do that while also providing more education about how to be a good employee and how to make your way through the workplace.

Kerr: Beyond the differences in expenditures across countries, do you have other factors that you say are behind where we find ourselves and that we now need to intervene in this crisis?

Deming: When our educational institutions were created—so there was a huge wave of college openings in the U.S. in the late 1800s all the way through the mid-1900s—many, many fewer people went to college. And college in the U.S. and certainly in Europe and in many other countries was seen much more as an educating citizen-leaders as part of a kind of upper crust of society. And now college is an expectation for so many more people, and many more people go to college. So in a way, we’re victims of our own success. We have a system that is funded and managed as though 20 percent of the population is going to attend college, when in reality it’s more like 60 percent or even higher. And so we just need to change and adapt. Not everyone’s going to be a citizen-leader of society, and there’s nothing wrong with that. We also need colleges to broaden and expand to serve an array of people who actually have, they want to learn the classics, they want to learn liberal arts, but they also want to get prepared for a job. You can do both, but I think we haven’t adapted our institutions to the reality that two or three times as many people are going to college than they used to. So I think what you see is work-based learning, internships. Those things are actually expensive to do, and they require a lot of institutional effort. And colleges are already kind of serving a lot more students than they used to. So I think it’s going to require a big push to get something like this done.

Kerr: So with this report, what’s the blend of the audience that you’re intending? And what impact are you hoping for?

Deming: So we’re hoping with the College-to-Jobs playbook and the College-to-Jobs map that we can provide actionable information primarily to local folks on the ground, policymakers, mayors’ offices, colleges about what’s happening in their area. So, for example, if you are a new mayor and you want to be the jobs mayor and you want to know what should you invest in, what kind of partnerships should you build between your local community college, let’s say, and local employers, you might be interested in knowing what are the degrees that my local colleges are producing right now. And then if I look at job vacancies, where are employers hiring? And is there a mismatch there? Are there some jobs where employers are posting lots of openings but there are not a lot of degrees being produced and, therefore, the demand for some profession—let’s say nurses—exceeds supply. And so then you say, “Maybe we ought to reprioritize, or if we’re going to make an investment, we know where to look.” And you can look at it just for one year, over multiple years. You can look by degree type. You can look by very detailed occupation code. And so we don’t view the tool, Bill, as an answer to a question, but rather increasing your ability to ask the right questions and to get information to take the next step. Right? Because it’s not like there’s going to be one solution nationwide for creating better pathways to good jobs. In the New York City area, there are lots of finance jobs that may not be true in other places, and here in Boston there are a lot of healthcare jobs. We want to provide people with the information they need to make good decisions. So rather than making an important jobs investment based on an anecdote or based on talking to a few people, you actually can look at the data in a systematic way.

Kerr: So, if we look over time, ideally we’ll see less and less of a gap in terms of this oversupply of one thing compared to the local demand and more alignment.

Deming: I think that’s right. I mean, I think it’s possible, Bill, to decide that both numbers are too low. You know let’s say that we actually want to have demand-induced supply. Just to think of an example, a lot of smaller towns don’t have much in the way of large employers, but every town has a hospital. So maybe that hospital is thinking about a partnership with another hospital or medical center further away that’s larger. Maybe we want to create more jobs and create them in partnership with employers and then also at the same time build the capacity in our local college. It could be saying, “Where do we want to make a big push?” Maybe we look at our area and we say, “Given our population, we actually have a lot fewer healthcare jobs than other cities that we might think of as good comparisons to us, and so we know where to invest.” So I think there are multiple uses for the tool, and I will view it as a success if we hear from our partners around the country that they’ve come up with new uses for it that we didn’t think of.

Kerr: Sounds great. So it’s a very complex terrain. Maybe walk us through how the report is structured. What’s the role of the data and the visualization and the other best practices that you’re bringing out?

Deming: So the way the tool works—and I think this is really important to emphasize—is that the data that we use are extremely comprehensive and high quality. When we look at the college side, we’re using a data source called the “Integrated Post-Secondary Education Data System,” or IPEDS for short, which is a survey that’s been collected for more than 30 years continuously by the U.S. Department of Education. And it’s a survey that has almost 100 percent response rates. Why? Because if you are a college, and you want to give out federal financial aid—that is, Pell Grants and Stafford Loans—you are required by law to fill out this survey accurately. And it includes things like how many students do you have enrolled, part-time and full-time, by race, by gender, by age, by degree program? How many certificates of electrical engineering did you give out in 2017? How many nursing degrees—licensed practical nursing versus registered nurses versus everything else in-between—did you award each year? And so it’s extremely detailed and extremely high quality. And then on the demand side or the employer side, we’re partnering with a company called Lightcast who has been collecting, scraping the web for job vacancy data online on job ads and internal company websites for almost 20 years and cleans it up and identifies all kinds of things about the job posting, the education requirements, the experience requirements, the job they’re posting for, something about the employer, and so on. And so you get an actual real-time snapshot of job vacancies in your area. What are employers hiring for, and which jobs, in which locations, et cetera? So we pair those two things together, and we have a really comprehensive view of a local labor market. Our data, we’ve collected it down to the county or metropolitan statistical area. So we’ve got hundreds and hundreds of local areas represented in the data, so it really is a pretty comprehensive map. And it’s, you know, you go to the map, and you can look up, you can start with your own area, you can select, you can kind of customize it, you can look for particular degrees, you can look overall, you can look at the time trend, and it’s really quite intuitive. So for those of you who are listening, I recommend that you stop whatever you’re doing, pause the recording, and go check out the tool. You can go to CollegeToJobs.HKS.Harvard.edu.

Kerr: David, I want to go back to the mayor that wants to have jobs, because every mayor wants to have jobs. But then the next part of that is that there can be great jobs. There can be good jobs. There can be less-than-wonderful jobs. And did you need to take an approach around the quality of the jobs that were being delivered?

Deming: That’s a great question, Bill, and it’s actually a very difficult question to answer, because it’s hard to track people, over time in existing data, to know something like a career trajectory. So I can tell you with some accuracy what are the characteristics of first jobs people get after college. But what’s a lot harder to know is how does that lead to jobs down the line. That’s our next project, is trying to look longitudinally at careers to understand, if I take a particular first job, what kind of outcomes can I expect 5, 10, 15 years down the line? So stay tuned on that. But I do think on job quality we try a bit to look at, for example, jobs that pay the kind of earnings that would allow you to pay off student loans or would allow you to earn some kind of premium, things like that.

Kerr: In your report, you have spent a lot of time thinking about potential interventions to help college to jobs. And you’ve got a matrix that talks about everything from career coaching to experiential learning coursework down to job shadowing and think about, first off, what does research say. Has research even been looking at something, all the way over to implementation feasibility? So talk to me a little bit about how you’re intending this matrix to be used? And what are some of the applications or key things you learned from building it out?

Deming: Think of it as a first resource for anyone who’s thinking about adopting one of these interventions. So what do we know about internships? What do we know about apprenticeships? What do we know about career navigation? We try to just provide a synthesis of the evidence on each of these questions, not like a step-by-step guide for how to do this in your area, but a step-by-step guide for how to get started. The second point is that we try to be very clear about the limitations of our knowledge in certain areas.

Kerr: I’d love examples of things you found surprising. “Hey, we do a lot of that, but I didn’t realize the literature was that thin,” or the opposite side of, “Wow, there’s a lot of evidence about this, but no one seems to actually use it.”

Deming: So there’s a lot of evidence about internships, and I had always thought—maybe kind of casually—that the benefits of internships were mostly about making connections between an employee and employer. So I get an internship with a particular firm, they learn something about me, and if they like me, they decide to make me a job offer. So there is a lot of high-quality evidence suggesting that, if I do an internship with a company, it increases not just my chances of getting a job, but my chances of getting a job with another company that doesn’t really have any relationship to the previous one. And it increases my chances of earning more down the line. So, actually, there was this great—and a lot of the evidence comes from other countries—programs where the people who are getting these internships were actually prohibited by law from working with a company that gave them their original internship. So you’re not allowed to stay with the same employer. And yet you see people who won a lottery, basically randomly were offered internships while they were in school, have higher earnings 3, 5, 7 years down the line as a result of that work-based experience with a completely different firm. And so, to me, that tells us something really important, which is not just about making the connection—which is really hard to scale—but it’s actually that work experience when you’re in school has a true causal effect on your employability and your earnings down the line. And I think that’s an extremely important finding that I was not really aware of before doing this work.

Kerr: I know you want to shy away from saying a specific, “Here’s a paint-by-the-numbers exercise for an employer.” But imagine this does land in the employer’s hands, which is the goal of publishing it. What role should employers do next with it? How can they think about taking the matrix and acting upon it?

Deming: So I would say that, well two things: One is, employers should understand that a lot of information about what makes someone a good employee is actually not very well captured by their college record. And so I would say, “Screen less on college credentials and screen more on actual experience.” So give kind of lower-dose opportunities for potential employees to show what they can do—so co-ops, internships, different opportunities—and many employers are already doing these things—to get a look at someone and not judge them too much based on their transcript. And even the college you went to is actually not that predictive of early-career performance. So you get people on the job even for a little while, and you learn a lot. So create more opportunities. I think that’s one thing.

Kerr: Great. So we can almost think this matrix is a way to make the boundary a little bit more porous and to experiment with people and try out a match of a worker and a firm.

Deming: Yeah. That’s a great way of saying it, Bill. I would say the other point is that, the more that employers can think positive-sum rather than zero-sum about this—create a kind of consortium—there can be huge benefits to everyone from that kind of cooperation. So I would say bring in even your competitors, in some cases, together with a group of colleges, and do something together, and you’ll all benefit much more than if you try to go it alone.

Kerr: You’ve mentioned you’ve gotten feedback. I’d love to hear any other angles people have taken on some of the reports and some of the visualizations.

Deming: So I think folks have found the visualizations really interesting on a couple of dimensions. So, one is looking at demographic characteristics. So there is some really interesting work being done right now at the pipeline in STEM-type jobs for employees of color, people from disadvantaged communities, and really huge differences across areas in the success or failure of colleges to produce a diverse pipeline of candidates in high-needs jobs. We’ve looked at that a little bit, and I think it sort of begs the question, “Why are some places succeeding and some not?” And so I think, again, the tool is supposed to be a way to ask questions, not an answer. So we don’t know the answer, but it certainly tells you where to look. I think another thing is being able to work backwards from anecdotes about success and to frame it in terms of the tool. What do I mean by that? So in this space of workforce development research, there were an awful lot of case studies of success. So a particular college—let’s say in Florida—develops an advanced manufacturing credential in partnership with a handful of employers and builds a pipeline. And everyone’s happy, and it seems to work really well. Is that scalable? Would it work in Iowa? Would it work in California? And we don’t have a lot of generalized knowledge. We have a lot of case studies of success. So what this tool, I think, allows you to do is say, “Okay, well, we know that there was this partnership in this particular year. Let’s look and see if we can see evidence of it working in the data. Do we see that employment outcomes are better? And then, can we look for other places where that, maybe, this was already going on or maybe it wasn’t, but they have similar characteristics, and kind of reverse engineer things that work and things that don’t based on patterns in the data.”

Kerr: You earlier noted, “Hey, colleges have been getting some things right.” So I’d like you to, let’s spend a little bit time on that positive part of it. And then, also, just what would be the best way of using this toolkit as a college going forward?

Deming: Colleges are under a ton of pressure right now. I think, both in the short run—the kind of fallout from Covid and the slow progress and regaining kind of the enrollment trend before then, getting students to come back. The fact that we have very low unemployment means, particularly community colleges, enrollment is countercyclical. And so enrollment is low, and budgets are strained—so I think the problem a college has right now, especially a kind of regional college, is very acute. And so having said that, I think a lot of them are starting to understand that this workforce mission is the way to go, that this is what potential students want, people are willing to borrow money, people are willing to forgo earnings, leave their jobs to enroll in a program that has a good chance of getting them into a better job than the one they’re in. And so you do see a lot of evidence of that happening, particularly at the community-college level. There’s a lot of money—coming in part from the American Recovery and Reinvestment Act—into some of these workforce programs into community colleges. The CHIPS Act has a potential, at least for the semiconductor industry to be a big boost. And so I think that there are a lot of green shoots, Bill, of potentially promising workforce-development programs across the country right now.

Kerr: And, David, recognizing that colleges have a lot of heterogeneity to them—from Ivy League schools, community colleges, state schools, and you mentioned more rural agricultural settings, center city, and similar—is it the kind of thought that economic mobility should be placed at the center mission kind of point for all colleges? Or are there going to be differences across?

Deming: I think it should basically be at the center of every college’s mission in some way. I think that looks different in different colleges. I think colleges have a liberal arts and sciences mission—and I don’t want to say that’s unimportant—but I don’t think it needs to be a tradeoff between that and a workforce mission. I think that a lot of learning happens on the job and there’s no reason why some of that can’t start in school. And so I do think, actually, every college, including Harvard, should have more work-based learning opportunities. I think that would serve the students better, and I think it would serve the taxpayers better. So my answer is, basically, I know there’s a lot of heterogeneity, and so it’s going to look different in different places, but I would say all colleges should do it.

Kerr: David, we span back out to your broader research agenda—which you mentioned has gone everything from early-childhood development all the way up through in the workplace—how has this particular research changed your approach, your teaching, your advising, what you want to do next?

Deming: So it’s an interesting question, Bill. There are 1,000 studies showing the causal impact of education on earnings and increasing life expectancy and making you happier, and there are all kinds of evidence of the benefits of education, but very little evidence on the why. And if you look just at earnings or wage growth over the course of one’s life, there is a lot of it that happens later in life. So people’s earnings accelerate rapidly in the first decade of their career or so, but there are huge variations in earnings growth over people’s lives. And that, to the extent that’s skills, too, it’s something to try to understand that I don’t think we really understand. So, if there are some jobs that provide good opportunities for building skills and learning and earning more money and becoming more productive, and there are some jobs that don’t, that’s a big source of lifetime inequality, and it’s kind of understudied. And, of course, it has lots of implications for policy and practice, so...

Kerr: Yeah. Maybe a final question about the Project on Workforce. What are kind of the next research dreams? You’ve mentioned earlier the upward mobility that some jobs will provide versus others, but I know you have a lot of student teams working on stuff so tell us a bit about what lies ahead.

Deming: So we have a couple of things coming out. One is, we’re trying to comprehensively catalog workforce development programs across the country to try to understand what’s out there. We think of ourselves, Bill, as trying to be a clearinghouse for good information in this space, a kind of neutral arbiter. We’re trying to get the best evidence out there to help aid decision making. And then the other thing I mentioned—I’ll just say it again because I’m really excited about it—is an attempt to say something about what are the good first jobs that lead to good careers? What are the first jobs that create more upward mobility? What are the kinds of pathways that people should be looking at out of college or out of high school? And what are the pathways that lead to family-sustaining incomes and economic mobility? So stay tuned for that, and maybe you’ll have me back to talk about that.

Kerr: You’re super productive, so I’m sure we will in the not-too-distant future. David Deming is at the Harvard Kennedy School. He’s also the faculty Co-Lead on the Harvard Project on Workforce. David, thanks so much for joining us today.

Deming: Thanks for having me, Bill. It was a real pleasure.

Kerr: We hope you enjoy the Managing the Future of Work podcast. If you haven’t already, please subscribe and rate the show wherever you get your podcasts. You can find out more about the Managing the Future of Work Project at our website hbs.edu/managingthefutureofwork. While you’re there, sign up for our newsletter.