Paul Osterman: The Disposable Worker: How Firms Learned to Stop Investing in People

Ben Zweig: We sometimes use terms of formal employment like W-2, contractor, freelancer, to try to define degrees of job security. But it seems like the target keeps moving. My guest today is the author of the new book, Disposable Workers, which makes the case that American firms have slowly reorganized themselves around workers they don't intend to keep. Paul Osterman is professor emeritus at MIT and one of the most prolific researchers on labor market institutions and job quality. Here's our talk. Welcome to the Economics of Work. I'm excited to welcome Paul Osterman, professor emeritus at MIT and someone who brings a lens of justice and fairness to the field of human resources. Just wrote a very thought-provoking new book, Disposable Workers. I have it right here, which I encourage everyone to check out. Paul, welcome to the show.

Paul Osterman: Thank you, Ben. I'm very happy to be here.

00:46 What is a disposable worker?

Ben Zweig: Yeah, for sure. So let's get started with just defining what is a disposable worker. What do you mean by this new concept of disposability?

Paul Osterman: So I will subsequently go into detail about different subcategories and ways of thinking about it because it's kind of a complex set of different kinds of situations. But the big idea is that, whereas it's true that almost all of us are at-will employees, that is to say a firm can fire us for any reason it wants to, unless we are a tenured faculty member or a judge, basically.

Ben Zweig: Yeah.

Paul Osterman: Normally, when a firm does hire somebody, they expect to keep them. They may not be in the end, but they expect to keep them. They expect them to have a career there. They expect to give them training. They often provide some kind of career ladder. Broadly speaking, a disposable worker is someone who, when a firm puts them to work, they do not expect to keep them. They do not invest much in training. They do not invest anything in career ladders. And so from the very beginning, it's different than at-will because from the very beginning, they're disposable. And the firm designs the jobs or organizes the work, so it is disposable.

02:13 The postwar social contract

Ben Zweig: Well, let me ask you, I mean, just to start with what it's being contrasted to. So when we think of non-disposable workers, workers that have this expectation of growth and investment, where do you think that comes from? What is the origin of the ... Does the firm want that? Is the firm imposing this relationship because it advantages the firm? Or is it coming from the workers that they want this and it's something that's being bargained over? How do you see the emergence of the expectation of more continuity?

Paul Osterman: Right. So we should talk about down the road in this conversation preferences, whether people want it, whether there are generational differences and so on. But I think the big picture is bring yourself back to the end of World War II. At the end of World War II, basically a social contract was agreed to. Now, it was underwritten by a very strong union situation, right? Unions accounted for over 30% of the workforce at that time, and plus firms were trying to avoid unions by imitating union practices. But the deal was that firms would keep people on board. And if they let them go, they would try and recall them after the business cycle evened out. And there was a set of expectations about wages going up parallel to productivity and to profits. It was a deal, not ever written down, though there is this kind of language about the Treaty of Detroit that Walter Reuther agreed to.

Ben Zweig: What's the Treaty of Detroit? I'm not familiar.

Paul Osterman: It's an expression representing the bargain that Walter Reuther, United Automobile Workers, made with the three automobile companies precisely about lifetime employment, recalls after layoffs, wages going up with the cost of living, wages going up with profits. Now, that was the deal. That was the kind of situation in the American labor market, with exceptions, obviously. Agricultural workers were excluded. African Americans often got a much worse deal. But it was a central tendency in the labor market up until the 1970s. And then firms started to systematically chip away at that. The automobile companies moved plants to the South to avoid unions. They opened up their job system, hired people from the outside, not just promoted people from within. They set wages based on performance, not on cost of living step increases. And then Ronald Reagan sent a signal when he fired the air traffic controllers. So I view the disposable worker trend as a continuation of firms acting as pure profit maximizers. And we can talk in more detail about what those motivations are, but acting as pure profit maximizers by squeezing the workforce as much as they can. Now, I want to be a little subtle here. I mean, firms do need flexibility. And so there are kind of benign reasons for having some degree of disposable workers. Firms do create jobs. Firms do innovate new products. All that's good. This is not an argument against profit-maximizing firms. It's an argument that you need to understand what they're about and put kind of guardrails around it. So that's my historical perspective on it.

06:05 Contractors, freelancers, and marginal workers

Ben Zweig: Yeah. Okay. So so much to unpack there. I mean, I really want to get into how we should think about firms and what you mean by guardrails and all that. But before even getting into all that, I mean, you mentioned there's like a breakdown of what goes into this disposable worker definition. How do you think through that?

Paul Osterman: Yeah. Well, the way I can kind of clarify that is by describing some occupations. So the first kind of category I work with are contractors. You're employed by a staffing firm and you go to work somewhere else. So one example of contractors at the relatively low end are people who clean buildings at night. They typically work for a staffing company. They're not employed by the building ownership at all. Landscapers, security guards, often many construction workers. At the high end, contractors travel nurses. They're also contractors. They work for a staffing firm, but those jobs are very well paid. So there's some distribution.

Ben Zweig: Well, what is what is the relevance of that? I mean I think so. So one way to think about it is like I mean, so I used to I used to work at IBM and at IBM only a third of the people that sit around you are technically employed by IBM. Then two thirds are technically employees of Infosys or Cognizant or whatever. They're just like employed through some other some other firm technically legally. But they're your colleagues. They're with you like everyone else. So I imagine it's easier to get rid of people like that. I know you have this case of Google where they were kind of in the same boat. But for all intents and purposes, they're basically like employees.

Paul Osterman: Yeah, well, they're not. So first off, it's easier to get rid of them, right? You can just eliminate your contract with Cognizant and those people are just gone. Whereas letting go you at that time would have been more complex.

Ben Zweig: So do you think when we say at will employment, it's like not really at will like if you're employed as a contractor, you're like fully at will. But if you're an employee, there's like, it's like it's like at will, but like on a spectrum.

Paul Osterman: That's right. There's a process. And not only that, if I let you go and let a lot of you people go, that has an effect on the morale and the performance of the people who remain. And that is not true for the contractor. But but the other piece of it, you ask, well, now remember, we're still on just one of these categories. Yeah, the other piece of it is that by using contractors, you can drive down compensation. So you can force staffing firms to bid against each other to get the contract. And really, the only margin they have is wages. And so you see this all the time, particularly at the low end. I'm a building owner in Boston, and I have a staffing company clean my buildings at night. Those contracts last only for 30 days. I put it out for a bid again, and somebody else wants that job, some other staffing company. How are they going to get it? They're going to offer a lower price. Well, the only margin is wages. So it drives down wages. Now, my data, we can come back to what my data are.

Ben Zweig: Yeah.

Paul Osterman: Support that argument. Contractors make much less money than do standard employees. So those people sitting next to you, some of them may be great. If they knew C++ and no one else did, then they've got power in the labor market. But most contractors are not in that situation. So my second category, again, to give you some examples, is freelancers. Now delivery drivers, ride share drivers are freelancers. Where they should be is a question we can talk about, but they are. So are freelance journalists. They're freelancers, and journalists are becoming more and more freelancers. Now in the book and in my data, I'm able to distinguish people who are freelancers for organizations versus the person who's walking your dog or taking care of your mother. Most data can't make that distinction, but I do. Again, these freelancers are totally disposable. They have none of the protections and benefits of employee status. The person who did the index for my book is a freelancer. And so that's the third, that's the second category. The third category, and this goes to the question of fissuring, are what I call marginal workers. Now a marginal worker is not fissured. A marginal worker is a W-2 employee from inside an organization, but the organization does not treat him or her as a real employee. It does not invest in their training. It does not put them on a career ladder. Who are those people? Well, adjunct faculty are examples. They're never going to get on a tenure track. They're on a short-term contract here today, gone tomorrow, depending on enrollments. Staff attorney.

Ben Zweig: Just a note on adjunct. So I'm an adjunct professor at NYU, and I didn't even know that I wasn't a freelancer until recently. I actually thought I was a freelancer, and I viewed as freelance. I can take on a course here, take on a course there. It's a gig, basically. The fact that I was considered like a W-2, I didn't even register that I was on a W-2. So yeah, I think it does very much feel that way.

Paul Osterman: The reason you are a W-2 and not a freelancer is because if you're around too long, there's court cases that if you have freelancers around too long, they essentially become de facto employees. To avoid that, if you are an employee, it's just that you aren't going to get tenure. Whatever course you were teaching, if the enrollment went down to a low number, sorry, Ben, we don't need you next year. That's not true for me. As Donald Trump would famously say, I could shoot somebody on Fifth Avenue and I wouldn't lose my tenure status. Actually, I might, but you understand the point. Other examples of these marginal workers, many firms create part-time jobs with the expectation that they're high turnover. So Walmart did that, Home Depot did that, United Parcel Service does that. I've got examples in the book of exactly that. So those are also marginal. They're W-2s, but the high turnover is built in. And then many low-wage service jobs. There's just no training, there's no career path, and they're marginal. So that's the third category.

13:07 Employers of record and H-1B workers

Ben Zweig: Where do you think of employer of records? When you're trying to hire someone in another country, sometimes you have to technically employ them through some entity that exists in that country. So you technically need to classify them as a contractor, even though you want them as a formal employee. I think these EOR structures have been set up, have been exploding a lot in the past few years since COVID. And I think many companies hire people across borders and they're classified as contractors, but they're treated as employees. Where do you think that fits into this framework?

Paul Osterman: So you're asking me about Europe now as opposed to the system here in the United States?

Ben Zweig: Yeah, I mean, for a firm that's based in the United States but wants to hire someone in Europe or wherever else in the world. And they want to hire someone through Deel or one of these EOR platforms. How does that fit into this framework?

Paul Osterman: Well, I don't really talk about that. And I don't want to pretend I'm an expert on that in any way. I do talk in the book about H-1B people. H-1B is another example of this because I hire somebody to be a programmer through Infosys or some staffing firm, but they're located in the United States. They're not working in India, but they're completely temporary. I can't keep them around for more than a couple of years. And in the book, I show large numbers of firms that use these things, this category, in lieu of hiring somebody as an employee. So that's actually another example, but that's somebody maybe using an international staffing firm to get people to work in the States. I don't really know enough to tell you about if I'm a firm in America and I want to hire someone in France, what I have to jump through, what hoops I have to jump through.

Ben Zweig: Yeah. The H-1B example is such a fascinating example because in my mind, that to me feels like the firms would... That's an example of firms wanting to have more permanence than is legally allowed. I think there's a lot of firms that get very upset when if their employees on an H-1B don't win the lottery and have to leave, that feels very destructive.

Paul Osterman: That's right, but they're not employees and they can be let go at a drop of a hat. It builds in exactly the kind of flexibility that they're seeking via these other ways of having disposable workers. It's just another example of trying to get out from under... There's research on this. Many people who are H-1B workers have skill levels that American workers could match. Those are not C++ experts. It gives the firm more flexibility. It's just another example of a disposable worker.

Ben Zweig: Yeah, interesting.

16:23 Doesn't economics say this evens out?

Ben Zweig: I think... Yeah, let me give you my understanding of what standard economics theory would say and maybe you can help us understand where this is maybe going wrong. From a textbook economics perspective, I would think there's a lot of determinants to someone's compensation. Someone gets paid some amount. There's real compensation. Then there's a variety of other things that employees want. Maybe one of those is career paths training. Maybe some of those are formal status and benefits. Those are things that I guess I would think of as part of the total rewards package and to the extent that those are costly to deliver. Some firms will deliver those non-salary benefits. Some firms won't. The workers that are more attracted to those types of things will sort into the companies that offer them and the ones who are less attracted to it will not. That's the hedonic wage function. You can think about it as just a compensating differential where you have some benefits that are given to the workers. If the firm makes the determination that they want to give that benefit and take on the additional cost, it will attract people who are more attracted to the benefit. It's nice to have things that are not just cash and things that can reconfigure. You get better matching. That all sounds positive. What do you think is missing from that?

Paul Osterman: That's a provocative and interesting argument. I'd say it's more complicated than on two dimensions. One is part of the answer about compensating differentials as you just laid it out is that the burden, this argument, the burden of these benefits or the burden of these non-pecuniary benefits of one kind or another are shifted to the workforce so that the aggregate cost to the employer is the same one way or another. It offers me a good wage or it offers me health and safety or it offers me a lot of training or whatever and the cost is shifted to the workforce.

Ben Zweig: So at the end-- Is that an argument about the mechanism or the incidence?

Paul Osterman: That's an argument about incidence of who bears the cost for these non-pecuniary benefits. An identical worker, one might receive a high-- you just said this, an identical worker on all observables, one would get a higher wage and the other would get a lower wage but a gym membership or whatever. There's two problems with that. The research on that is completely negative. It doesn't find the burden of shifting. It's a little like the minimum wage argument that logically the minimum wage should reduce employment, but within a reasonable range it just doesn't. The world is more complex than that simple supply and demand argument. But the other, and I think the killer response to your argument is that identical workers in the compensating differential world that you describe should have equal satisfaction. They just get their satisfaction from different sources, one from wages, the other from the gym membership, a third from whatever. I've got data on job satisfaction. In a survey, I need to describe to you where my data come from, but in my survey of all these people, I have data on job satisfaction. Observationally equivalent people who are in these different statuses, so someone who looks the same on the dimensions of age, education, experience in the labor market, whatever. If you're a contractor or a marginal worker, you're less satisfied than if you're a standard worker with the same characteristics. So they're not getting, it's not evening out the way textbook theory would suggest it should.

Ben Zweig: Yeah. I guess what is still a remaining question at that point is two opposite questions. One is that some benefits are offered at the cost of employers. If companies offer gym membership or whatever it is, presumably they're doing that for some sort of reason. I mean, they want to attract people. There's this example of, at one point I heard this story from Hal Varian, who was the chief economist at Google, where he basically said that it was because Steve Jobs was so anti-competitive in labor markets with Eric Schmidt at the time. He said, "We have to cap our salary for entry-level engineers." So they couldn't compete on salary. It was textbook case of anti-competitive behavior. There was just this explosion in amenities. Google took on all these benefits at very high cost because they thought it would be cool or attractive. Now that's obviously been exaggerated and pulled back on. Yeah, that's one thought.

Paul Osterman: Yeah, but let me tell you about that, I think. The Affordable Care Act, Obamacare, exempted firms, exempted the requirement to offer health insurance or pay a penalty if people were part-time. Right? Now, in the textbook world you're describing, it shouldn't make any difference to the firm if the people were part-time because the cost would be shifted to them in just the way you described of health care insurance. They get their health care insurance and the firm would cut back on their wages. But the research shows that after the passage of that act, there was a growth in part-time employment beyond trend expectations. Why? Because they wanted to avoid having to pay that benefit. The firms did not believe that they could simply shift the burden to the employee. All the kind of evidence of your eyes, plus the evidence of the data, suggests that there's... I'm not saying there's no truth to the compensating differential. I'm saying it's not strong enough, strongly enough true, to obviate the fact that firms are being strategic about using disposable workers to avoid benefits in career ladders and training.

Ben Zweig: Yeah. Yeah.

23:37 What stops every firm from doing this?

Ben Zweig: Let's say, I mean, yeah, I think that sounds kind of consistent in a way where if the cost to a firm is going to be going up, they'll try to avoid that in any way. Let me ask you another question. Why are more firms not doing this? What are the limitations on this? Is it norms? Why don't all firms just squeeze as much as they can? What is the limiting factor on this?

Paul Osterman: Well, so first off, more and more firms are doing it. So the trend is up, right? Clearly there are some costs to the firm. So I do have data, again, from the survey. I asked disposable workers of each category, freelancers, contractors, marginal. How committed are you to the success of your organization? And freelancers are on average, are committed, more committed than standard workers. Contractors and marginal workers are not. They report themselves substantially less committed. This is called organizational citizenship in the literature. I ask a parallel question. How willing are you to put in extra effort for your firm's success? Again, I get the same pattern where freelancers are willing to put in extra effort, but contractors and marginal workers are not. So there is a cost to this. It's not a free good. And now I don't want to pretend, and the real literature doesn't suggest that firms are really completely rational profit maximizers. I mean, there's a lot of organizational glitch in the world, as I'm sure you saw at IBM. You'd see inside of any organization. You look at what's going on and say, this isn't efficient, but they're doing it nonetheless. But firms are, they are facing a cost for doing this. The other cost they face is teamwork. So everyone, work is increasingly organized in teams. And if your person in your team is one of these disposable workers here today, gone tomorrow, that team is not going to be as efficient. So there are some limiting factors to it. That said, it's growing.

26:06 Is there a sweet spot of disposability?

Ben Zweig: Yeah. So, I mean, yeah, I want to ask about like drivers of that more, but also, I guess one more thought that I want to gauge your reaction on that I think is kind of, I don't know if there's like real evidence on this, but I get the sense that like within the field of economics, there's a belief that when we look at national measures of disposableness, it's not clearly bad. Like if we think about this example in France where it's like basically impossible to fire someone, like that's the extreme where like workers are not employed at will. It's like the opposite of disposableness. So that seems to be really bad. Like that creates like lack of dynamism in the labor market. Like it's it hurts workers, it hurts productivity. So do you think there is like a sweet spot of disposability?

Paul Osterman: I would be completely making it up. Here's what I'm willing to say. Firms do need flexibility. I mean, there is such a thing as a business cycle, right? And skill needs change over time. So there is a need for flexibility. The trick is, as a policy discussion, the trick is to think about how you provide firms that flexibility without being overly damaging to the people who are being flexibleized. And that's the word, right? So I don't argue in the book that we should become France, right? And I don't know, you're too young to remember the Coneheads. Did you ever see the movie The Coneheads with Dan Aykroyd? Dan Aykroyd and Jane Curtin, these are actors from Saturday Night Live at the very beginning, and they're somewhere in the middle of the country, and they have an upside down ice cream cone on their head. And that's what they look like, right? They're from some alien planet somewhere, and people have ice cream cones upside down on their head. So the people in the middle of the country look at this, "Who are you guys?" And they say, "Well, we're from France." And everyone says, "Oh, yeah, you're from France. The people are different in France." I am not arguing that we should move to a completely rigid labor market in which you can never lay people off, you have no options for freelancers. I'm absolutely not arguing that. I'm arguing, A, that there's kind of too much of it, but B, there's consequences for people. And if we're going to do it, we need to alleviate those consequences.

29:01 What's driving the trend?

Ben Zweig: Yeah, makes sense. So now, how do you make sense of... I guess I'm curious about what you think of as the drivers, but also as the policy levers, and maybe they're related. Like maybe if we understand the drivers, we can think of better recommendations for various policy levers. How do you make sense of what's driving this and what's actionable?

Paul Osterman: So I talk about that. So one set of drivers we've already discussed about, which is flat out cost. There's the wage cost. You can drive by having staffing firms bid against each other.

Ben Zweig: Would you think of that as something that drives this cross-sectionally or over time? I guess I mean, what's driving the increase over time?

Paul Osterman: Okay. But what's driving the increase... Let me come back to that. I want to mention one other driver that we haven't talked about. We've talked about what you could call cross-sectional drivers. There's a cultural driver. In the book, I quote a McKinsey report. Now McKinsey, as if there's any establishment voice of the business community, it would be McKinsey. McKinsey had a report called The State of Organizations in 2023. And they write that 95% of a firm's value is driven by 5% of its employees. Now if that's not a cultural attitude about the workforce, I don't know what it is. Right? And you see that... I mean, that's why I call that the quiet part out loud. So there is that in addition to kind of the direct kind of cost drivers. The other thing that explains the trend is if the social constraints around behaving right have loosened radically. They just loosened. And that's why I kind of engaged in that conversation about the decline of the post-World War II contract. What was seen as best practice... Another kind of clever way of describing it is it used... The best seller X years ago was In Search of Excellence. It was about firms as families. Then along came Reengineering the Corporation. It was about firms as firing everybody. Yeah. I mean, that was a transition, right? About what's seen as best practice.

Ben Zweig: Yeah. And you think this is related to general just like politeness and norms and behavior in like, in society.

Paul Osterman: I actually think, and I don't mean this as a left-wing critique of organizations, but at the end of the day, they're about maximizing profits. And I mean, think about DEI. After George Floyd, every large firm in America told us that DEI was a core value. They really cared a lot about it. They don't even mention it today, right? I mean, it's about profit maximizing. And that's not a bad thing. I'm absolutely not arguing that's a bad thing. But I am arguing it's something that needs to be managed, an impulse that needs to be managed so we get the benefits with fewer of the costs. So this kind of attitudinal stuff exemplified by the McKinsey report and just a loosening of constraints about what's best practice plays a role. But I think also clearly cost and savings costs and getting flexibility are important.

32:37 Monopoly power and rent sharing

Ben Zweig: So in a previous conversation on this show, I was talking to a colleague of yours at MIT, Anna Stansbury. And she was making-- I mean, you may be familiar with her. She was making this case that this-- I mean, she was really talking about the fissuring aspect just about how employees get classified. But basically making this case that there's a rent sharing between firms that are making monopolistic profit. There's some distribution to their employees, but less so because there's been declines in worker power. And when there is a fissuring, then the contractors who are part of some janitorial agency or security agency or whatever, they're not part of that negotiation over rents. And I wonder if this-- I wonder how you think this concept may be related to kind of anti-competitiveness on the product side, but also on the worker side. Do you think there is an anti-competitiveness angle to this argument?

Paul Osterman: Well, first of all, I would never argue with Anna because she talks with a British accent. And if you talk with a--

Ben Zweig: Yeah, yeah.

Paul Osterman: You just assume people who talk with a British accent are smarter than you. It's true.

Ben Zweig: In this case, it might be true.

Paul Osterman: Well, I'm not going to admit to that. But she definitely talks with a British accent. I know Anna quite well. She is a colleague. And she's great. There's two sources. First off, let me just say the importance of this marginal worker category is that it is not consistent at all with that argument because those people are inside the organization. They're not fissured. And if there are social norms in the organization that are driving stuff, you would think they would share in that, and they don't. And in my data, marginal workers are a big percentage of the workforce. And so that's one of the significant things about that kind of category. Yeah, the anti-competitiveness on the product market side-- I haven't worked on this, but other of my colleagues have-- clearly, Walmart has so much power on the product market side that it can go to Procter & Gamble and say, if you want us to carry whatever shampoo Procter & Gamble makes, you're going to have to give us a better price. And then that works its way down to lower wages. So the degree of monopoly on a product market side, the purchasing side-- not the selling or the product, but the purchasing side-- makes a difference in wage structure. And there is research on it.

Ben Zweig: Wait, why does that make a difference in wage structure? Wouldn't you think that, OK, so on the product side, Walmart has this monopoly pricing. They can get a bunch of economic rents. They can get some excess profits. If-- yeah, like maybe--

Paul Osterman: No, no. It's not about them having monopoly power when they're selling to you and me. It's about them having-- you could call it monopsony power or monopoly power when they're talking to one of their suppliers. They can squeeze their suppliers. Just like a firm using a staffing agency for building cleaning can squeeze the contracting agency.

Ben Zweig: Yeah. Well, so let's say they have monopsony power on their suppliers. But is that like-- is there a relationship between also having monopsony power on the worker side or they're just-- they happen to have both in their--

Paul Osterman: No, no. Well, Walmart is now all over the country. Maybe when they were located only in Arkansas, they had monopsony power. But they're everywhere now. So that's not Walmart's story. So much. And by the way, Walmart, which for many years was kind of the poster child of bad behavior towards their employees, they've gotten better. And when we come to policy, I'll say that public pressure and public policy pressure and just norms and campaigns can make a difference. And Walmart's an example because they were responsive to Fight for $15 around compensation. I've lost the thread though. Where are we in the conversation?

Ben Zweig: Yeah. No, I guess just like whether monopoly power on the product side is potentially explaining some of this--

Paul Osterman: Well, I mean, the automobile industry is, I guess, is an example. The big three totally lost their monopoly power to Japanese, Korean imports. And that's led to the diminished role of not only the union workers and interestingly enough, a large fraction of automobile assembly workers now work for staffing firms, not for the big three. And totally in automobile supply chains, heavy, heavy use of contractors, not regular employees. And that's reflected in the loss of monopoly power on the product side.

38:11 Remote work, AI, and uncertainty

Ben Zweig: Yeah, it's a really interesting example. So if we're thinking about drivers over time, why has this increased so much recently? Yeah, what are some theories there? Does this have anything to do with remote or-- I mean, one theory I heard recently, I was just talking to a neighbor of mine that went to that quit her job a couple of years ago as a W-2 and then came back as a contractor. And in her case, she actually liked it because she was excluded from the return to office mandate because of that. And I thought that was such a funny example because it's just no one really cares about this return to office mandate. They're just like trying to circumvent these rules. Yeah, I mean, I wonder if remote is-- I know that's not about the marginal workers. In particular, that's about these other categories of disposable workers. But yeah, I guess I'm curious--

Paul Osterman: In the survey there, I do ask freelancers if they work on the site or remote. Now, keep in mind that the survey I'm working with was done in 19-- 2022. So it was executed at the tail end of the pandemic when there was more remote work because of the pandemic. But clearly, freelancers are much more prone to working offsite or remote than contractors or marginal workers are. So I'm willing to admit that's a piece of the trend. I mean, thank you very much for not asking about AI. But I think AI is another explanation in the following sense. I think the issue of whether AI is going to change the amount of employment in the economy is unclear. It's not clear what it's going to do to the number of jobs. But there's so much uncertainty around AI right now on everybody's part, but including employees' part, that if you're an employer and you don't know what your staffing need is going to be because of AI, you're going to be much more inclined to use disposable workers because they're easier to get rid of. And to the extent that there's any good research around the impact of AI, it shows that AI is affecting young workers more than established workers. And that's the margin on which employers can use disposable workers, entry level. So I think that is a reasonable expectation about trend.

Ben Zweig: Yeah, I think that sounds right. My read on this younger worker phenomenon is that, like you're saying, it kind of represents a higher discount rate in that younger workers are always a more uncertain bet than someone who's done it before. And when the path ahead of you is so unclear, all you can really do is optimize for the short term and that advantages older workers. So that's part of it. And also firms, when they have so much uncertainty, there's more of a premium for flexibility. I mean, both from AI, but also policy, tariffs, just like all the madness in the world.

41:39 Can policy reverse the trend?

Ben Zweig: Yeah, so do you think if the world becomes more calm, do you think we'd see a reversal?

Paul Osterman: I think it would take policy tools to really reverse it. You might be able to slow down the trend, but it would take policy tools. And so here again, this is where these categories make a difference. We've been talking about disposable workers. But the issue around, say, Uber drivers or food delivery drivers who are freelancers, who are treated as freelancers, is a different than the issues around contract workers or marginal workers. The first in cases of drivers, you're talking about defining employee status. And whereas in the other people, they are employees, but there you're talking about just the floor of job quality, improving the quality of the jobs. So there's a different set of policy tools around this. The other thing I would say, if we're moving into where we're going on this, this is a broad-based issue. So it's not just about poor people, right? So it's about journalists who have been pushed out of the newsroom and are now freelancers against their will, which I interviewed quite a few of those. It's about people who are contractors at the high end but would like much more stable employment. So it's up and down the kind of hierarchy. And so you would think it's an issue that politicians could latch on to. There's some evidence that they have, but not much evidence.

Ben Zweig: Yeah. So what are some of your favorite policy levers for this?

Paul Osterman: Well, on the freelance defining employee status question, first off, Uber. I mean, it's not just Uber. It's Lyft and so on. And by the way, it's being pushed at the margin. So Amazon is moving its drivers into freelance status. Walmart has something called Walmart Spark, which is just like Amazon moving its drivers into freelance status. So they're pushing at the margin of making more and more people independent contractors or freelancers. What's happened with that is that the federal law is ambiguous. So depending on who's the president of the United States and in the Department of Labor, the definition of employee status is revised. So Obama pushed to broaden it. Trump one narrowed it. Biden broadened it. Trump two narrowed it. About 32 states have a much more simplified law called the ABC criteria for who's an employee. It shifts the burden of proof onto the firm to prove that someone is not an employee. These states can't use this for collective bargaining because that's preempted by the federal government, but they can use it for state benefits and regulations. In principle, the adoption of that standard federally would make a huge difference. Uber claims that workers want flexibility and they lose it, but that's just not true. There's nothing in the law about employee status that prevents people from being part-time or prevents them from choosing their own hours. They could give workers flexibility. That's not a legal barrier. So for freelancers, the issue is sharpening, sharpening who's an employee and who's not. For marginal and contract workers, there's a range of tools. Unions when they're there improve people's work situations quite a bit. But as well as I do, unions are rarely there these days. They represent 6% of the private workforce. Public campaigns like Fight for $15 actually did make a difference. I like the example of Activision, which had a huge contractor workforce that needed, I guess it was FTC approval to do something or other. I can't recall just what.

Ben Zweig: Their merger with Blizzard?

Paul Osterman: Yeah. So they were hauled in front of some Senate committee and asked about how they're treating their contractors and lo and behold, suddenly those people were made into regular employees. So there's a range of tools. And Google is another example in which the regular employees pushed hard against the company around the quality of the situation of the contractors. And then Google issued a standard of how we're going to treat our contractors. So there are a range of ways of thinking about how to improve job quality. And that applies for contractors, that applies for marginal workers. But again, at the end of the day, if it's not a politically salient issue, it's hard to imagine a lot of progress.

47:04 Who is responsible for job quality?

Ben Zweig: Yeah. It's interesting. I mean, I'm in some ways sympathetic to the idea that like, who are we to say what's good job quality? We have wages, we have career paths, we have good work environment, we have good management. It's such a multi-dimensional concept that it seems kind of hard to think of policies. Then do we have to kind of pick what we view as good parts of a job? Is it about learning from others? Is it about mentorship? Is it about low commute times? There's so much to get at. I think it's such an important thing that you're documenting on that job quality seems to be eroding in very identifiable and important ways. And yeah, at the same time, I don't know who is responsible for that.

Paul Osterman: Well, we do know who's responsible for it. The employers are doing it. I agree that people care about a list of things and it would be foolish for you and the public policy to kind of check off every box. I don't think there'd be much dispute that a living wage would be a component of job quality. We have a history of setting a living wage. The minimum wage is a joke at the federal level, but the precedent of wage standards is quite clear. Some level of job security and fair treatment on the job, I don't think people would argue about that. I'm not saying you have to have a rule. We're going to require gym memberships.

49:04 Contractors and the headcount problem

Ben Zweig: I think this concept of the marginal worker really makes it very stark that there's a whole bunch of things that people expect from employers that they're no longer getting. And that is a very interesting thing to at least to note and probably do a variety of things about. Here's another point that I think you didn't really highlight in the book, but I want to kind of make it because it's something that I see and struggle with. I wonder if this messiness and classification is just making it very hard to monitor the state of the economy. When companies disclose their employee headcount, they only disclose their W-2 employees. That makes it very difficult to understand their total workforce. I struggle with that as a company who tries to understand what's happening in companies. That data is not reliable and we have to find all sorts of workarounds, which makes it very difficult. It would be very nice to know the workforce of a company, maybe split by different contractors, but companies don't need to disclose contractors. And also just in...

Paul Osterman: Well, do you know who Susan Houseman is? She's an economist at the W.E. Upjohn Institute for Employment Research. She has written work precisely on this, that by not including contractors and freelancers in employment numbers, we've misestimated productivity change, we've misestimated employment change. She wrote a couple of papers precisely on your point, documenting that you're absolutely right, the exclusion of these people from what companies report. And by the way, there's an inside bureaucratic reason for this, of course, because you're a manager and let's just pick IBM. You're a manager in some company and this chief financial officer says, "Ben, you have to keep the head count in your group down to X." And you say, "Okay, I'm going to do that." And then you sign a contract with a staffing firm to get more people. And that doesn't show up on the personnel records because they're not in your... It shows up as some kind of cost, but not tracked to bodies, to head counts. So yes, that's a very fair point.

Ben Zweig: Yeah. Yeah, it's tricky and seems like there are better solutions for it. Although, I mean, as I'm thinking through this a little bit more, I don't know if there's like a perfect solution because even if we were to include contractors and freelancers, at Revelio Labs, we're a data vendor. We sell data, but we also sometimes provide support. And sometimes there are some cases where someone who we're selling data to says, "Oh, my team has been cut. My head count restrictions," blah, blah, blah. And then we say, "Well, we could just provide support." And basically, as part of the data license, as part of the amount that people pay for any sort of vendor, there's levels of support you can include or not include. So even just vendors that companies work with are ways to kind of get more without it being classified as employees. And should that be included as employee expense? Probably not.

Paul Osterman: No, I think that's right. That's right. What I try and do in the book to avoid the supply chain argument, which is essentially what you're making, I only look at people who are working on site. So if you send your IT operation up to Amazon Cloud, I don't count those cloud employees as contractors. It's only if instead of having somebody on site that otherwise exactly you go down the slippery slope that you're describing.

Ben Zweig: Totally. Do you think that... I mean, you mentioned Susan Houseman's research on measuring productivity. I wonder if there's an effect on just the numbers that come out of the BLS. They're surveying establishments, like how many employees do you have? In the establishment survey, it's often unclear who's an employee and what even is an establishment right now. The nature of how establishments are. In JOLTS, there's this measure of inflows and outflow rates, hiring and separations. And those are just extremely high, just implausibly high. And I wonder if it's because of this recategorization.

Paul Osterman: So if you want to... Now, I'm playing the role of the professor. If you want to pursue this question, which I think you're actually right, that's an important question that we don't have good answers to. The two names you should talk to are Susan Houseman and Katherine Abraham. Katherine Abraham was the commissioner of the Bureau of Labor Statistics and Susan Houseman, who has done a lot of work jointly with her, is also an expert on this stuff. And they could give you a much more interesting, coherent discussion of this topic than I can.

Ben Zweig: All right. I'll have them on the pod and we'll cite disposable workers.

54:56 What labor economists should do more of

Ben Zweig: Yeah. So that's great. One final question for you. As a lot of listeners of this podcast are in the labor economics field, labor economists. What do you wish that labor economists were doing and thinking about from your perspective?

Paul Osterman: More generally, you mean?

Ben Zweig: Generally, or related to this topic.

Paul Osterman: Well, not related to this topic, but I mean, a generalization is that labor economics has been very... Like other fields in economics, but particularly labor economics has been caught up in analyzing large data sets and using very sophisticated econometrics to study that, which is all fine. I'm not complaining about that. The number of young labor economists who have been out there in the world talking to people, getting inside of firms, seeing what's really going on is diminishing. And I think that weakens the econometric research and the survey research because it's hard to put a context around what you're measuring and seeing. So I wish that there was kind of, I mean, a crude way of saying, I wish there was more field work. I think people would be out there looking at reality more.

Ben Zweig: Yeah. I feel like organizational psychology does that really well. They're very related to the field of HR in a very admirable way. Yeah. Do you think there are other fields that do that well?

Paul Osterman: Well, organizational sociologists sometimes do too. Sociology has been a little caught up in, I want to be like an economist syndrome. So there's probably in the top sociology journals, there's too much fancy econometrics still. But they still are a little more out there. Now, I'm not arguing that you should substitute being out there for careful data analysis. Not at all arguing that, but I'm saying that the lack of being out there and seeing reality diminishes what we can say.

Ben Zweig: Yeah. I mean, it sounds like they're complements. Yeah. Yeah. That's great advice. Thank you so much for doing this, for being part of the show. And again, like, check out Disposable Workers. It just came out and it's a really excellent read. Thank you so much.

Paul Osterman: It was a great conversation. I really appreciate it.

Ben Zweig: The Economics of Work is brought to you by Revelio Labs, workforce data for research and benchmarking. If you enjoyed, please remember to rate, review and subscribe. It'll help us reach other curious listeners. I'm your host, Ben Zweig, our producer is Cole Wagner. Thanks for listening.