Bill Kerr: There have always been individuals on the periphery of the workforce, who for one reason or another are unable to gain steady work in line with their capabilities. But in recent decades, automation—intended in part to reduce discrimination by casting a wider net—has excluded many qualified workers. In fact, it's not hard to trigger a robotic rejection. It could be gaps in the resume for caregiving or a spouse's work move, underemployment, a history of incarceration, or a host of other factors. But why, when employers complain of a worker shortage, are these hidden workers still out in the cold? How can employers engage with this varied group estimated to number 27 million in the U.S.?

Welcome to the Managing the Future of Work podcast from Harvard Business School. I'm your host Kerr. I am pleased to welcome as a guest by Managing the Future of Work co-chair and podcast cohost Joe Fuller. Joe led a joint investigation with Accenture into the root causes and potential solutions to this labor market breakdown. In their report, Hidden Workers, Untapped Talent, Joe and his coauthors analyzed the experience of workers and employers in the U.S., U.K., and Germany. They present the business case—as opposed to the philanthropic rationale—for hiring hidden workers. They explain how organizations can expand their talent pools, address their skill shortages, and diversify their workforces. Joe, welcome to our podcast.

Joe Fuller: Thank you, Bill. It’s always fun to be in a position to answer a few questions as opposed to ask them.

Kerr: Well, I have lots of them. The first one I want to ask is, you have a very catchy title. How did you define a “hidden worker”? Then, as you think about their exclusion, do you see it as being due to technology being misapplied, shortcomings of HR, or some other factors?

Fuller: Well, we were trying to understand what caused employers to skip over large numbers of applicants. These workers were often on a cadence of applying once or twice a month for positions but getting employment offers at the rate of one percent of the number of applications they filed. What we came up with was this notion that the way employers were approaching the process—of identifying applicants and then qualifying applicants to be candidates—caused them essentially to screen off, through the filtering they were doing, various subpopulations of applicants, effectively hiding them from the process. The fundamental design paradigm rests on doing things really efficiently. Rather than spend a lot of time—which, of course, time is money when you’ve got an open job position—or spending a lot of professional resources or fees to search firms, they emphasize minimizing the number of candidates that are considered and minimizing the time taken to fill a position. That logic gets embedded in their applicant-tracking system—their online tool for soliciting and then doing an initial assessment of candidates. It dictates how much time—not very much—is spent by recruiters directly reviewing candidates’ profiles. It limits the amount of time that the ultimate supervisor for the worker spends in the process. That quest to maximize efficiency drives the process to seek to minimize the number of people a human being considers and to only consider those applicants that are the tightest fit to the job description.

Kerr: Great. So we’re going to come back to those applicant-tracking systems and potential changes. But let me start with a question that comes from the introduction, which is the eye-catching number of 27 million. We’ve had on this podcast—you’ve cohosted and I’ve cohosted with people from the Bureau of Labor Statistics and other government agencies about official data. You’re developing your own data set and estimates in this report. What was lacking about the official data? And what’s the other background around excluded workers that was present when the study commenced?

Fuller: The data that does present itself through great resources like the Bureau of Labor Statistics don’t do a lot to tell you what the makeup of various populations are. They have very great longitudinal data on definitions like “long-term unemployed,” but it ends there. You don’t know if those people are long-term unemployed because they suffer from some significant disability or whether they’re some other form of what we came to call “hidden workers”—for example, someone with multiple felony convictions or someone who is in a geographic location where there are effectively very, very few private-sector jobs available. We wanted to understand, with more nuance and more texture, who were those populations. The single largest category in the United States of hidden workers are people that are working part-time and would like to get more hours or a full-time job but can’t. Someone who’s working part-time, depending on the number of hours they’re working, is not treated as unemployed. That’s, by the way, very different for the other two countries we studied–the U.K. and Germany—where that phenomenon is much less prevalent, and workers are more likely to be unemployed in the way a man on the street would define it. We wanted to both expand the scope of people we were looking at to include people that are not in the official statistics, but more importantly, try to understand who are the subgroups that make up those various populations.

Kerr: Well, given that these workers are hidden and that they’re not captured in traditional data sources, how did you and our partner on this project, Accenture, organize the research? And how did you go about identifying and surfacing up this group?

Fuller: We did about 125 face-to-face interviews across the three geographies we looked at, just to give ourselves a better sense of the life of these people and how they viewed it to help structure our questionnaire. Then we went out and interviewed three groups of people. We interviewed employers as to their employment practices as applied to hidden workers. But more importantly, we interviewed—we surveyed, I should say rather—two groups of hidden workers: current people who fall in these categories of hidden workers who are either trapped in a position where they’re working part-time but would like to work more hours or are actually unemployed; and those who were formally hidden workers but now have successfully reentered the workforce. We wanted to do that so we could not just understand the magnitude of the phenomenon, but what seems to drive it. What causes someone to get on the margins of the workforce and then just stay there? Are there practices through government policy or employer practices that effectively wall those people off for consideration? By comparing the experience of currently hidden workers and former hidden workers, we were looking for what the differences were, what seemed to enable that formerly hidden worker to get in the workforce. What are the implications for this large pool of people who remain hidden?

Kerr: As to the difference between those two groups, what allows some to be able to get back into the workforce?

Fuller: Certainly, persistence is one thing, and that you do have asymmetries in terms of how often people are presenting themselves. We did find that workers expressed a high level of self-agency about trying to get ready for jobs of the future, trying to enhance their attractiveness as candidates, for example, by learning a new skill or adding a credential, although there are a lot of barriers to people doing that. There are issues also of people’s life circumstances changing. A significant percentage of hidden workers are active caregivers for kids, for seniors. If you’re suddenly an empty nester or if a parent or other senior that’s been living with you goes into some type of care facility so you’re no longer that primary caregiver, your circumstances can change. But I think an overarching impression we had is that, once one is in one of these hidden worker categories, there’s a strong gravitational pull. That’s compounded by the way employers assess candidates, who view someone who’s not been employed or doesn’t have the most recent skillset as not qualified.

Kerr: Well, this brings us, I think, into something of a trillion-dollar question, which is, you actually began this research a couple of years ago, before the pandemic. But the pandemic has pushed a lot of caregivers into more marginalized employment. How has the perspective around hidden workers adjusted from when you began the project to now, 18 months or so into the pandemic? Are we going to have more hidden workers as a consequence of the caregiving responsibilities during the pandemic?

Fuller: It’s clear that, as both caregiving responsibilities became more significant for a lot of people under Covid—for example, maybe having their parents move in with them, or very importantly, having school out of session and having virtual learning—that completely changed the mosaic of who was either stepping away from the workforce or changing their hours as a function of the caregiving requirements of Covid. As we go forward, of course, those impacts are going to be relaxed asymmetrically by region and by caregiving category. Also, Covid had a significant impact on the caregiving industry, with a lot of capacity, particularly for kids, effectively being shuttered. The one thing, though, that was really interesting in our research, Bill, was that, whether you are a caregiver or a veteran or someone with a criminal conviction, these categories of hidden workers—when we asked them what was the impact of Covid going to be on their capacity to get work—almost half of them said it’s going to be equally hard as it’s always been. That, despite the fact that employers were frantic to find both medium- and high-skills workers to fill positions in essential roles that were still operating under close-to-normal business rules. I think that really illustrates how daunting it is once you fall into one of these pools to get out of it. These workers were pretty convinced that it didn’t matter if there was buoyant demand for workers; they were still going to be largely on the sidelines because of the way companies go about looking for talent.

Kerr: Earlier, Joe, you mentioned the difference between the U.S. and the European countries on the role of part-time hidden work. Were there any other key differences or points of similarity between the three countries or across the three countries in terms of hidden workers?

Fuller: I would say there are a couple of important things. The first is that, across all three markets, employers were regularly saying it was very difficult to find talent with the relevant skills. American employers were the most satisfied of the three countries in terms of the quality of the candidate supply. They weren’t thrilled with the quality, but their feeling about the quality is higher. I think relative to hidden workers, their experience was quite uniform across the countries. They found things like the fact that the experience of applying for a job was rather complicated, the requirements weren’t very clear. The fact that they never got any feedback about why they had not been considered was really very varied—not merely made it hard for them to understand what they had to do differently, but very demoralizing. You can imagine going to the trouble of applying for jobs regularly and just getting silence back. It’s really psychologically distressing. We also found that they all shared the same dilemma, which is, they knew that they needed to do something to break the impasse, to add a credential or two to get closer to the state-of-the-art and the jobs they were seeking. But they really lacked the insight as to what they could do that would make a difference. They didn’t really have the resources, the slack time, and the financial resources to make the investments that they think would allow them to get out of this isolated position. The final thing I would say is, across all three countries, the personal impact—in terms of everything from sense of self-worth and optimism to level of energy people were bringing to the process—it was very common that people really feel personal distress about being regularly on the outside of the workforce looking in. And that’s a human condition that isn’t associated with any rational labor market.

Kerr: Yeah, it’s tough. As you think about the industry space—you have geographies then industries—are the hidden workers particular to an industry or the segments of hidden workers most relevant to industries differ, or do they feed into a common hidden worker pool across industries?

Fuller: I think it’s more the latter, that there are life circumstances or other attributes that tend to identify you as likely to end up in one of these hidden worker categories. There are, across all markets, impact of what I would describe as “stranded skillsets.” If you look at certain industries—let’s just take the United States as an example—foundries, computer manufacturing, aerospace manufacturing. The employment level today is half of what it was 20 years ago. Part of that is automation, part of that is globalization. But if you spent 15 or 20 years in an industry—let’s just take foundries as the first one I mentioned—and that’s what you’re really knowledgeable about, and that’s what your resume describes, or the skills that you have high confidence you can bring to somebody, so when you get looking for a job, you don’t see job positions out there that really speaks to what you think you can do, you end up out of the labor market. Your specific skills related to that industry are not terribly relevant to the jobs that are available. But more importantly, the longer you stay out of employment, the further away you are from having the skills profile that employers are going to want. This becomes a bit of self-fulfilling prophecy. For whatever reason you’re walled off from the main sequence of the labor market, once you’re on the outside, the problem compounds, which makes it very, very important for both workers to take the steps they can take to get re-engaged in the market quickly, and also ultimately for policy makers to think about, what can we do to actually get people who are on the short- or intermediate-term unemployed reemployed quickly, as opposed to let them drift in that longer-term unemployed candidate, where the slope of the hill they have to climb to get back into a good job is getting steeper with every passing month.

Kerr: Another part of that is, in that foundry example, “I’m reading job descriptions, and they don’t seem to line up with me, but I actually have the skills to do that type of work.” That could be another barrier. In the report you talk a lot about job description. Tell us, who generally writes a job description, and what do they seem to be getting wrong?

Fuller: What we found about job descriptions is, particularly for middle-skills jobs—jobs that require applicants who have something less than a four-year college degree, they can have an associate’s degree, or a certificate, or a professional license that doesn’t require a college degree—but even, also, for higher-skill jobs, that job descriptions are not revisited very often. When they are rewritten, it’s generally done by someone in an HR function, human resources function, that actually runs the recruiting process. They generally are not going to have the type of nuanced understanding that a supervisor or someone who’s a success in the role today really will have for what it takes to be good at this job. Job descriptions, also, they tend to be like rings in a tree stump. The attributes that were sought 10 years ago and 15 years ago and 20 years ago still remain embedded in the job description. Just new paragraphs of the additional requirements are stapled on at the end. So what does that have the effect of doing in the hiring process? It means a lot of the requirements may now be superfluous, certainly secondary or tertiary considerations relative to someone’s likely success in the job. When you get one of these very complicated job descriptions with lots of requirements, we know that a lot of candidates, particularly those who are a little bit further down the path to being hidden, essentially abandon the process of applying. The consequence of this is that it’s actually an expression of that overarching design paradigm I talked about earlier. The process does succeed to the extent that it starts to winnow down the number of people who are going to get considered very quickly. But in so doing, because these job descriptions are overly complex, not updated, and not really skills-focused—What do you have to be able to do to be a success here but rather experience and credential focused? What jobs have you had previously? Did you get this degree or this certificate? How long have you been employed in this job? How many years of experience?—it excludes lots of people that would actually have a good chance to be a success in the job. What’s perverse about this is that employers know it. We asked employers, “Do you think your process excludes qualified candidates?” For both middle-skills and higher-skills jobs, about 90 percent of employers said, “Yes, we know that happens.” That was a head-scratcher for me, because these exact same employers are the ones saying, “I can’t find the right talent.”

Kerr: I’m going to throw in another statistic, Joe, just while we’re on there. You have in the report that 90 percent of the employers are using an applicant-tracking system. If we start multiplying 90 percent by 90 percent, it ends up being a large percentage here. Is this, then, cause for optimism, like, “Hey, if we think of it in a different way, we’re going to be able to make significant traction toward hidden workers fast”?

Fuller: Well, I think there’s optimism. As our report indicates, employers that have launched a specific program to access certain subsets of the hidden worker population really report encouraging results. If they say—for illustration, would be CVS—the retail pharmacy and health insurance company has a program for people who have cognitive impairments, are not often even high school graduates. To train them to work in their stores in things like stock shelving, putting goods on the shelves… Or a Microsoft doing that with cognitive diversity individuals, for example, autistics… Or more commonly, companies that develop customized programs for onboarding veterans… When they focus on a subset of hidden workers, understand what kind of accommodation, upskilling, onboarding they’re going to need, partner with representatives of that subpopulation, they really report good results on multiple business dimensions in terms of productivity relative to what I’ll call “open-cycle-hired candidates ”: loyalty to the company, rates of voluntary turnover are lower, and a general softer benefit of enthusiasm of the other workers, the incumbent workers, who welcome these new colleagues into the workplace. That’s one cause for optimism. Obviously, Covid has caused a lot of employers to revisit what they actually need out of workers. That’s going to, I think, relax some of these constraints that crept into the system over time, accreted in the system, but shouldn’t really be disqualifiers for consideration for people. Finally, the simple demographics in all these markets are such that the absolute size of the workforce is essentially stagnant. Unless you think that—if we take the American example—the U.S. is going to have a very well designed, well honed, and quickly implement and change its immigration laws—of course, something that you’re an expert in, Bill—employers are going to be having to deal with a pool of workers that are accessible now. That means they’re going to have to confront the fact that we’ve got many more job openings than we have unemployed, at least as classically defined. They’re going have to find ways to access pools of workers that are not reported as unemployed currently or underemployed currently to fill their growing needs.

Kerr: I think important in the examples you were just giving, Joe, is that this isn’t staying in corporate social responsibility, this is the business case for CVS or for Microsoft. How are most of these companies defining success? Are they able to put numbers behind it or does it stay typically directional, like you were describing?

Fuller: They define success, those that do it, I think importantly, as a normal business process would be evaluated. Are we getting talent we need? Employers regularly report, employers with these types of programs, that this has helped alleviate their problem finding skilled talent. Are people able to become productive quickly? When you know what that hidden worker is going to need in terms of accommodation—it could be a workplace layout accommodation, it could be some remedial training, it could be certain type of more specialized and skilled supervision—but once you’re able to define that, then you can build a management process around it and cause that person to be more productive. I think a skeptic might say, “Well, that kind of accommodation causes the employee to be perhaps captured, if you will, by the employer, because you’ve made this special accommodation, and some other employer may not be able to do that, so where else do these workers have to go?” Well, my response to that is, first of all, let’s remember that they were isolated from work, and the vast majority of them badly wanted to change their work circumstances. But just like many of our parents told us over time, the easiest way for any person to get the next job they want is to have an adjacent job. Even if you’re getting some special accommodations, you’re developing those state-of-the-art skills, you’re no longer showing a gap in your resume, you have a work experience that other employers, maybe who aren’t so sensitive to the needs of hidden workers and the opportunity vested in that population to find talent, are going to view you as a legitimate candidate for the next job, even though you are once a hidden worker who benefited from this type of intervention. I do want to say that the people involved in corporate social responsibility or ESG efforts to engage in work populations have had, really in many ways, pioneered this and give us confidence that these programs can work. But what’s most important, as you suggested, is that the company—the leadership of the company, the co-workers in the company—don’t view this simply as doing good. It’s doing well by doing good, because hidden workers can become highly effective and productive employees that are going to advance the prospects and performance of the company just like those who have never fallen into the hidden-worker segment.

Kerr: Great. Let’s also loop back to the government, and I’m going to take off the table for now a well-functioning immigration system and instead ask you, what other elements of a policy agenda would you be putting forward here?

Fuller: Well, there are several. I mean, first is that many of the workers fall into a category called “neither in employment or education”—an acronym for that is “NEETs”—employment, education, or training. It’s very, very hard once you have a big gap in your work history. Let’s say you are in your late 20s and you’ve had a collection of part-time jobs and then periods of unemployment. It’s very hard to put together a strategy for getting back into the workplace. Often by that stage in your life, life has begun to happen. You have children or have used up, for example, all the grant capacity or loan capacity that you’re entitled to or that you know how to access. One is for governments—and I think this is particularly relevant of the U.S. government at state and federal level—to really think about creating paid work-based learning experiences for categories of hidden workers, especially these so-called NEETs, but also others. So that essentially the employer has an enticement to go into what I describe as a rent-to-own cycle. But it’s important, first of all, these positions be paid, that companies have skin in the game, but also that they be really learning engagements. I’m not using the word “apprenticeship” because that defines a very specific version of paid work-based learning, which is a terrific vehicle for enhancing a workforce. I’m talking about new, innovative vehicles for that. I think a second is for the government to help fill in the big data gaps in the market. We don’t have data that’s easily accessible to ordinary workers about, for example, what programs at what schools or skills providers actually have a good record of putting people in employment. We don’t have good data about what skills are actually associated with success in jobs. Whether it’s employers participating in initiatives like the Open Skills Network, where companies are spending more and more time trading thoughts as to, “What are the actual skills we’re looking about?”—government encouraging that—or government stipulating that, for example, educational institutions that rely heavily on federal funding at least begin the process of making results for their programs and the achievements of their program completers visible to others, to give them more confidence that, if I do sign up for that course at a community college or that certificate program, I’m not just making a bet based on my best instincts. I’m making that investment based on a clear sense that there’s going to be a return at the end of the race.

Kerr: Joe, tell us a little bit about the reception that the work has received as it’s come out now for about a month. Then, with that in the backdrop, what do you see is happening on the horizon in the medium to near term, on business side, on policy side, on advocacy groups, and so forth?

Fuller: A lot of employers, and very importantly companies like the technology companies that provide the recruitment management systems and applicant-tracking systems that we discuss in our report have been really rather remarkable in terms of their willingness to acknowledge that there are real opportunities here to improve what they offer their customers and for their customers to work with them to create a system for attracting and assessing applicants that is a win-win—that gets more applicants considered and helps relieve the really tough constraint in finding workers. We were very lucky in our timing, because it’s pretty apparent for a lot of companies, particularly those offering middle-skills jobs, that the model they’ve relied on historically just is no longer fit for purpose. It’s just not creating enough people with the requisite skills to fill roles that they need filled to continue to be a success commercially. I think that there are two or three things that may get triggered by this, or this might add to the debate to advance a shift that a lot of people believe is necessary. The first is moving further and further away from credential-based hiring to skills-based hiring. That may not be a distinction that immediately speaks to our listeners, but historically companies have relied on proxies—like years of experience or a college degree—or a very important one that ends up disrupting a lot of people’s prospects for employment—continuity of employment. If someone’s had any significant break of more than six months in their employment history, that’s viewed as a significant decrement in their candidacy. Fifty percent of companies actually exclude someone from consideration on that variable alone. Rather than relying on those proxies from which we infer someone has skills, someone has self-efficacy, someone can work effectively in a group, someone can be productive and understand what’s required to be productive in workplace. Let’s really try to focus on what are those skills that someone has to have, and how do we start finding ways to measure that or to assess someone’s background for that? Let’s take a significant group of hidden workers, veterans. Almost any company you talk to in the United States, and I imagine Germany and the U.K. as well, if you say, “Would you like to hire veterans, or how do you feel about hiring veterans?” they all express enthusiasm for it. But when they go into the labor market, they are using that applicant-tracking system to define skills and experiences in words that are completely alien to the actual background of the uniform branches. Unless they have a very specific program for hiring veterans, they exclude lots of veterans with no human being ever knowing that’s what they’re doing, because they got excluded because they didn’t have five years’ experience in sales management, or they hadn’t ever worked before in a job that had 15 key words that are sprinkled around the job description which don’t match to military jargon. I think technology companies and human resource areas ought to be working hard to really audit the way their process works. I hope also that considerably more effort is put in to understanding the user experience of a job applicant in responding to a request over long descriptions that are very intimidating in their use of language, that don’t make clear what the deadline is for applying, or timing of a decision that don’t provide any feedback to someone. They reinforce this cycle by causing people not to apply, to abandon an application mid-process, and to give up hope that they’re ever going to get a position. I think all of those things can get revisited. I also hope that some leading company in the areas of reskilling will start taking the logic they’re applying with their incumbent workers, stroking their chins thoughtfully and saying, “Well, if we can upskill our 48-year-old who’s in our organization, why can’t we take that same type of skills enhancement and apply it to a hidden worker that doesn’t currently work here?” We’re hopeful that those discussions will be joined and that more and more companies will be able to point to their success in here in using hidden workers, as we pointed out, with companies like Microsoft, like CVS, and even small employers like our friends in Columbus Hot Chicken Takeover and start extending that logic. One other thing I’m very encouraged by: real discussion now about what’s generally described as a second chance initiative, where we are looking at people with criminal records and saying, for a large percentage of that population, the nature of the crime they committed, what it suggests about their potential to workforce shouldn’t be an employment death sentence. We ought to be revisiting the degree to which companies either have an absolute bar on hiring people with criminal convictions, or, even though they don’t have a bar, they use it as a filter and it excludes potentially qualified candidates once again without a human being ever knowing that’s what happened.

Kerr: Yeah. I’ll put a plug in. Joe DeLoss, we did two podcasts with him from Hot Chicken Takeover, and we don’t have to wait for policy reform, he’s showing how to do it in his own business. Joe, final question here. What do you see this report as raising for future research opportunities?

Fuller: Well, I think there are several. One is that, as we say, to really understand hidden worker population you have to invest in one or more of the constituent sub-segments. I think there are great opportunities to research what would be vehicles, for example, for taking military skills and articulating them into syntax, a taxonomy, that employers could recognize. The obligation there not only falls on employers, but also is a call to, in the U.S., the Pentagon or Defense Ministry in the U.K. or Germany to lead that and to start thinking about, how one both inculcates more skills in people when they’re in the service but also starts making that skills transcript, if you will, more comprehensible to employers. I think that’ll be a great avenue. I think also there are really important areas to study: How do we impart what are often called “social skills” to people? Very, very often, what the employers are crediting someone who has a long work history is they’ve got those skills, like the capacity to work with others, to communicate effectively, to build relationships that are more and more important in people’s success and work. How do we start thinking about enhancing people’s social skills through learning and training interventions? I think it’ll be very, very important to understand, how do we think about displaced workers in a more thoughtful way? There’s been some good work done here by people like the World Economic Forum with Burning Glass Technologies. But it’s only the initial phases of, when you look at someone who had an employment history but now has those stranded skills I described earlier, how do we get them into the workforce? Historically, our answer has been a very ineffective one, which is, “Well, they did work on an airplane assembly line, but we need a lot of more cybersecurity technicians, so let’s train them for the jobs that are in short supply.” That hasn’t worked. But if we think about that job as a working on a manufacturing line as having lots of shared DNA—not 100 percent, but lots of shared DNA with jobs that are growing—how do we build those bridges? Finally, how do we... I would very much like to have people explain to me why the market for labor has such lousy information flows and inaccurate information flows. What would be the highest impact improvements, whether it’s skills providers publishing more results data or employers providing more feedback to applicants as to what they were missing. If we could just get some more insight into that we might find some plausible, affordable levers we could put in place to try to make this more efficient market.

Kerr: Joe, thank you so much. I love advertising free stuff, and so let me just point out that on our website, as well as in other places, the recent report, Hidden Workers, conducted by Joe and the team at HBS along with Accenture, is available. Thanks so much for joining us.

Fuller: Thanks, Bill.

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.