 My name is Leon Sintiqoumana and thank you for all for coming, especially guests from outside of the department and also guests from inside of the department. So it's my pleasure to welcome you to what is the fifth annual Sambol lecture. The Sambol lecture is made possible in part by a generous anonymous donation. We are very grateful. I would like to thank the staff members of the economics department and Perry who did an amazing job in planning and hosting this event. I also invite you to reception downstairs on the second floor after the lecture. The Sambol lecture series was established to honor the work of our colleague Sambol who is here today. One of the founding members of what is now recognized as the premier heterodox department in the world. The department is proud of what Sambol accomplished while he was here on staff full time, as well as how he continues to be a shining star in economics department and an excellent ambassador for the department. The fifth Sambol lecture today is given by Professor Juliet Shor, who is a professor of sociology at Boston College but most importantly a PhD graduate from here. Welcome Juliet. We are proud of her work and now I'm going to invite Sam to introduce the speaker. Thanks Sam. It seems impossible but it was 40 years ago when Julie was leaving here that she gave to me and my kids a little tree like this tall. And this morning as I was thinking about those 40 years I looked out of my window and there across the field is this tree which is now three times taller than my house. That tells you something about how things have gone on since she left. A lot has changed and she's done wonderful things. I won't go into that because I know we want to get on to hear what she has to say. But after leaving here she taught for 17 years at Harvard in economics and women's studies. She's now an economist teaching in the sociology department. I want to give you some idea of the breadth of her work and how she goes about working on problems. She won the Leontief Prize for advances in economic theory and she also won the American Sociological Associations Award for public understanding in sociology. She is a genuine multidisciplinary person. She gravitates towards questions. The tools are secondary. They are whatever is needed to answer the question. She's also done some stunning research which I think hasn't been recognized like a few of her other blockbuster titles. She wrote what I think is the best research on how labor effort, worker effort, responds to local labor market conditions. I think it's the best study out there except that it's not out there. She was busy doing things which she thought was more important than that and so it never got published. This is by far the best work and some very good pieces of work come out recently by Liz here and others. She was moved to write important works, one of which became a real blockbuster. The overworked American, the overspent American born to buy. If that sounds a little dirty to you, all these horrible things, I didn't know that she had a book with a subtitle. How and why millions of Americans are creating a time rich, ecologically light, small scale, high satisfaction economy. Now who could be against that? She founded South End Press to co-founder with other people from this wonderful department. She was also a co-founder of the Center for Popular Economics also here. She founded the Center for New American Dream, many other things I am not mentioning. She's a great public intellectual. I think some of you know she's also a great teacher. She won the University's Distinguished Teaching Award here in this department. She's one of the very few of us, of our community, whoever has won that. It's a very rare honor. The legendary Steve Resnick, a fantastic teacher he won it. Melissa Osborn won it. A couple of other people, but obviously she was recognized for being a truly outstanding teacher here. She also won, listen to the title of this thing, the George Orwell Award for Distinguished Contributions to, now I get this, Honesty and Clarity in Public Language. Now that's a concept. Couldn't we try that? How about honesty and clarity in public language? I mean, we really need you, Julie. Keep it up. You know, I sometimes see Julie when I'm going through airports, which is the only time I actually see a TV screen. And, you know, people, she's related to what people care about, and people ask about her work. Kenneth Galbraith once told me that he had the impression that his colleagues at Harvard, they only read his books because they were afraid that their cleaning ladies were going to ask their opinion of what the books were. And that actually happened. That is, I think that people who don't consider themselves to be economists read Julie's work. And if you're an economist and you haven't read the book, well, then you're kind of out of it. Just like some Harvard professor, you know, who, well, it happened to me. I was coming through customs once and he said, what do you do? I said, I'm an economist. And he said, oh, what do you think of Galbraith? And, you know, I had read Galbraith, so I could answer the question. But when I think of Julie's work and her contribution over the years, the person I think most about is Thorsten Bedlin. There is a tradition in American economics and sociology which Bedlin began Galbraith was a representative of, and Julie is part of that. And as an intellectual and as a political actor and as a public intellectual, she's very much like that tree in my field, towering. Thank you, Julie. Well, thank you. It's really wonderful to be here on the way I was counting the number of years since I first made that turn up into the campus to get into Thompson Hall and some of you in the room, probably not too many, may remember that when I made that turn, I totaled my car on my very first day of graduate school. But it's a, yeah, it's been an amazing time since I was here. I'm so honored to be giving this lecture. Sam was my advisor and he was an incredible advisor. I certainly have not lived up to the standards that he set for me, but at least I always do know what it is that you do owe your students and the way you are supposed to treat them because of the incredible mentoring that I got from him. And also what an amazing graduate experience I had here. I often feel very sorry for my graduate students because life seems so much harder for them. And I think it's much more difficult today, at least in my university, to have the kind of intellectual excitement that we had here at UMass. It was just an all-consuming and just really kind of utopian graduate experience. And I know that sounds kind of crazy because these days when we hear about graduate students, we hear about depression and pressure and stress and how horrible it is. But it was really fantastic and it was all about the ideas and the work and the reading and the talking and the research. So thank you so much and thank you, Sam, for everything that you gave me. And in the Q&A, somebody should remember to ask me about a cleaning lady's story connected to my books and it was a funny one. And the cleaning lady in question was Barbara Ehrenreich while she was doing research for Nicollin-Dymed. And there was one more thing I wanted to mention from your research or from your intro, but I'm forgetting it now. So can we make sure this is off? That's not on. It is off. Okay, great. Am I hearing a little something? Yes. Is it coming out okay? Yes. Okay. So the funny thing, oh yes, and why it was that I never published that paper because as you said it, I remembered why. So it's a me too story that about somebody that deserves to be talked about. So somebody please ask me about that in the question and answer too. Okay, so just one more prefatory point here, which is I'm going to talk about research that I've been doing for the last, basically starting in 2010, so almost a decade now, on what is called the gig economy or the sharing economy. And the story as it relates to being here, giving the Samuel Bowles lecture, is that about maybe five or six years into the work, I was writing a paper about some findings that we had, which I'm going to talk to you about. And I was well along with the, and in fact we had the finding, and I think I was writing up a grant application to do more research on it. And I've been thinking about and writing about this finding for quite some time, and it wasn't until that moment that it dawned on me that what we actually were seeing was the cost of job loss, which was what I wrote my dissertation about, Samuel and I wrote on it, but enough years had gone by that it wasn't the thing that immediately popped out to me. And I was so pleased that I had come full circle to a finding completely independently, unconsciously that got me back to the cost of job loss. So what you have mostly, unless you're researching and reading in this literature, what you probably have heard most about with the sharing economy is a kind of, it's good, it's bad, pro-con, very normative political public engagement about it. So lots of people saying it's fantastic, and here's Airbnb trying to get New Yorkers to go easy on Airbnb, did a big advertising campaign. And pretty quickly, on the very day that they put up these advertisements all over the city, they were written on. The shared economy is a lie. Joan Rivers was a bigot, not exactly connected, but another great quote from that day for a paper that an article that came out in the mainstream press about what happened when they launched this advertising campaign and all the graffiti and these things were defaced immediately was another one of these, I happen to have been there so I was able to get these pictures, was the dumbest person in your building is handing out keys to large numbers of strangers. So there's been a lot of debate about controversy about Airbnb and then of course if you've followed what's been going on in recent months, focusing on the labor issues, there's been a big fight about ride-hell drivers in particular but gig workers more generally and number one, the treatment of gig workers and a lot of critiques of the increasing degradation of conditions of work and pay for ride-hell workers Uber and Lyft most particularly. On May 8th, there was the first global strike of ride-hell drivers and in September, the California legislature passed and the governor signed into law a bill called AB assembly bill number five which makes them all employees and the companies are now fighting this, they've put together nearly, well, 90 million at the moment to have a ballot initiative. And on the other side you see here the, this is a screenshot from what the ride-hell company, the platforms are sending to the drivers to tell them to go against AB five because it's going to take away flexibility. So a lot of controversy about the classification of these workers and that also has been, you know, very much a pro con kind of thing and then a couple of weeks ago Instacart which is one of the sort of most problematic platforms in the delivery space Instacart is a platform where people go and do shopping for people and then, you know, deliver it to their homes, grocery shopping Instacart workers went on strike for three days and the company immediately reduced their wages as a result of the strike. What I wanted to do today is move beyond the pro and con and there's lots of interesting things to talk about the pro and con I'd be happy to in the Q&A to step back for a second and say, well, what are the analytics of these platforms? How do we understand them as economic and social entities? What are they doing? What's new if anything about platforms? A lot of people think there's nothing new about platforms they're just the latest way to exploit labor in the same old ways that they've always been exploited. What difference does the technologies that they use make? Of course, they use sophisticated algorithms to do a lot of the work that historically has been done by management. Do they have unique features? Is employee status even feasible? The platforms say the whole model is based on the independent contractor status and that shifting to employee status is going to put them out of business and how malleable are their current configurations meaning are there things that could change about the way platforms are operating that might deal with some of the problems that the critics are raising but that might also retain some of the features that the proponents like. So I've done a multi-platform study, one of the few and also the biggest that I know about, which is qualitative. So there are surveys out there that survey people from lots of different platforms but in terms of kind of seeing how these platforms are actually operating in the experience of labor and so forth, very few multi-platform studies. So what a multi-platform study allows you to do is kind of figure out what's unique about the whole form. And what I want to argue today is that these platforms represent a new kind of labor regime that is very different from what's come before and represents something both difficult, troubling and I think very right with possibility. So a word or two about definitions, there are now lots of words which are being used to define what we've been studying. The first terminology that was used was collaborative consumption and oh, you know what, I forgot to take out my notes. Okay, I'm going to do it without my notes. Collaborative consumption and collaborative consumption, really the key idea was an idle resource, an idle asset that could be used by the non-owner while it was idle. That's fine, I think I can. Thanks. So Airbnb, being a classic case, you have a spare room, you rent it out to someone or you loan it out to someone. Sharing economy or collaborative consumption, so it really pertains to those idle assets, that idle assets idea. Labor, something like Rindhale, Uber and Lyft, originally started as an idle asset phenomenon because the idea was that people were in their cars and they were going somewhere and they had a free seat next to them. Of course they pretty quickly devolved into just a labor service with that collaborative consumption or what came to be known as sharing economy dimension, really not particularly relevant, particularly once people started to acquire cars for the purpose of driving them. So it came to look a lot like taxis with more sophisticated technology. The on-demand economy also refers to those sort of labor services that people get on apps that come here and basically ride them and deliver them being the key ones. Circular economy speaks to the growth of second-hand markets on platforms. I'm not going to talk about those today, but that's another part of what's been thought of. The peer economy meaning it's person-to-person exchange and I called it originally the connected economy, connected consumption of connected economy. But I'm going to use the term sharing economy in part because where I do use the term sharing economy in part because there was also a very important non-profit dimension of the sharing economy which I think is really relevant, has been key in the success of this sector and is important I think for thinking about how to change it. Today I'm going to focus on the labor side and the gig. So gig economy which is what I put in the title is really the part of it that I'm going to focus on. Okay, so what's really different and important about this sector? Number one is the technology. So the technological innovations that make possible peer-to-peer exchange meaning person-to-person exchange and basically undermine the need for corporations or companies who do various things in economic transactions. So what are the two key things? Number one, companies develop brand reputations and reduce risk for consumers. And my study has focused on consumer services but a lot of these issues are also relevant for the business-oriented services, gig work on things like mechanical turf and so forth and then also business-to-business. So when you have a peer-to-peer situation where you have incomplete information, asymmetries of information among the peers. So how do I know that David Cotts is a reliable person to trade with? If David Cotts is part of a brand like Ramada Inn or Yellow Taxi, I know that the brand has done the work of vetting and so forth. But in a peer-to-peer economy, you don't have that. And what the technology allows you to do is crowdsource ratings and reputation and that actually came from other places like eBay is where a lot of that started so that by seeing the five stars and reading the reviews, I can feel comfortable making a trade. And many of these trades are very intimate trades because it might be somebody in your home or in your car. So they carry high possibilities of risk. So number one, the technology reduced the risk associated and even more than reducing the actual risk, it reduced the perceived risk quite a lot. And number two, it also facilitates search because one of the things about peer-to-peer economies with lots and lots of small sellers is for the consumer you have to go and come through the yard sales and the flea markets and so on and so forth. But the digital technology allows you to have nearly instantaneous search. So the technology enables a whole, sort of really the growth of a new sector and sort of moving away from centralized provision, corporate provision in lots of areas. What else was on my notes? One was cashless payments which doesn't figure too big in the story. Another key thing is ease of entry and exit. So the platforms are very open access, very easy to get on them, very easy to get off them. And that turns out to be really important in my story. I had some other things I was going to tell you about here and I now forget them and I'm sure they'll come to me. Okay, so this is the name of the project. It was funded by the MacArthur Foundation. Officially ran from 2011 to 2018. I now have an NSF for the next three years with the Future of Work program at the NSF to study algorithmic workplaces. So this work is continuing. And these are the graduate students who worked with me on the project. They were all PhD students in sociology and most of them now have their PhDs. So we took a case study approach. Is that funny? We took a case study approach and I'm going to show you all the cases that we did. In most cases we interviewed users, we've interviewed consumers, but also earners on these platforms. In the case of Airbnb, we also scraped and have a national database and have done a lot of other things which are not sort of qualitative. But the material I'm going to talk about today, your stuff is from qualitative data. And we surveyed everybody so we have some numbers too, but I'm not going to really talk too much about that. So we did Airbnb Touro which may show up in the slides as relay rise it rebranded. It's basically Airbnb for cars where you rent out your car when you're not using it to somebody. So not with a driver, the people who rent it do the driving. TaskRabbit which is an errand site subsequently acquired by IKEA who still runs it, still runs independently. And a task that used to be called Rabbits which may give you some sense of how the company felt about the people working there. But you could just post a task, you could post anything that was sort of vaguely legal and people would bid on the task to add an auction model. Eventually they moved to something different from both the auction model phase and also once they had sort of posted wage rates for things. Postmates which is a delivery service will again deliver anything legal or vaguely legal. It's the equivalent of Uber Eats. We did not do Uber Eats but Postmates and Favour which is another, one of these delivery sites started as mostly bicycle couriers. Now increasingly people are using vehicles. Uber, eventually I gave in and said yeah, I guess we have to study Uber which is what everybody was studying. Lyft and then the final case that we did is a platform which is owned by the workers. It's a stock photography platform so the workers are photographers or they're artists. They don't call themselves workers and I'll talk a little bit about that at the end. It has a different kind of analytics than some of the others and some similarities. So we've researched a lot of different things. We started also looking at nonprofits. I didn't include those in the cases but alternative economic forms like a maker space, a food swap, a time bank, etc. We looked at moral, sort of the moral and cultural dimensions of participation. We looked at the culture of Airbnb hosting with our national database. We looked at racial discrimination and gentrification and also ratings, how the ratings function on Airbnb and racial dynamics there. We looked at platform labor and its relationship to income inequality, new kinds of vulnerability among Uber and Lyft drivers, systems of labor control which is what I'm going to talk about today and the dynamics of platform cooperatives. So we've identified and I should say back on the other slide was our project website. Most of our papers are there. I have a book coming out on this in September called After the Gig. How the sharing economy got hijacked and how we can win it back. And the next little bit of the talk comes from a review article that I've done with a labor sociologist called Steve Ballas which is coming out in the annual review of sociology this summer. So there are four major approaches in literature to understanding the analytics of these platforms and I want to tell you what they are and then tell you what my alternative is. So the first comes from economists and economists tend to focus on efficiencies and entrepreneurialism. So the point that I made earlier about these technologies enabling peer-to-peer exchange. So one of the things that economists have argued is that you can get many much more self-employment in very small firms as a result of these technologies and I think that's right. They also focus on the algorithms and what the algorithms can do. They can reduce management costs. They can get you increased efficiency in logistics. In RideHel for example instead of taxis driving around in what are called wild goose chases the algorithms can because they allocate the passengers to the cars they can get much more efficient matching so they're really important matching technologies. They can lower search costs for consumers as I already mentioned. They foster entrepreneurialism and self-employment. They can solve information deficits to reduce risks. Again I mentioned that. Efficient payment systems. So economists tend to think this is a really great thing. It's going to yield all these benefits and it's going to yield benefits to individuals and it should reduce the average size of enterprises because you'll have more and more self-employment as a result of it. Some have gone so far as to say it's going to end employment because everybody can just be a self-employed platform entrepreneur and this is also the sort of rhetoric of the companies. I'm going to go through them and tell you what I think is wrong with each of them. The second one is what we might call the algorithmic manager or my boss is an algorithm which is a famous kind of popular meme that goes along. So people in the camp of algorithmic management who tend to be either sociologists or lawyers or legal scholars basically argue what's new about these entities is the use of algorithms and these algorithms are incredibly powerful and they can control all aspects of the labor process and they're also powerful in terms of power dynamics because the firms can basically set the algorithms in ways that create long sort of structural asymmetries of information between the company, the platform and the driver or the worker that gives the companies a structural power over the drivers. So for example some years into ride hail the companies started blinding destinations to the drivers. So when drivers have to take a certain number of rides but they no longer knew where the rides were going and that is the kind of asymmetric information we're talking about. Another asymmetry that they're doing now is that they're charging the customer one price and they're calculating the pay of the driver on another price. This is especially true when you have long, long drives and so the workers complain, the drivers complain, they the customers might be getting twice being charged twice the base that the driver. So these kinds of things so this yields systemic power over the workers also a lot of criticisms of the lack of human content arbitrary decisions and so forth made by the algorithms and of course if you've followed any of the debate about AI and algorithms you know there are lots and lots of problems with algorithms associated with racial bias and gender bias and so forth and that actually shows up in some of the studies of gig labor also. So for example the task-ragment algorithm is more likely to recommend white workers and black workers in some cities and so forth. The third view is it comes up sociology there's also a trend within economics and that is basically arguing this is there's really nothing new here. This is just a continuation of this drive to precarity that we've seen. So whether it's guys standing in the precariat work that you know did Eileen Applebaum's you know working on this back in the 1980s a lot of economists working on the growth of contingent temporary part time alternative work arrangements and so forth. So this this school of thought basically says look what's really important here is the precarity basically symbolized by the lack of W2 employment and the independent contractor so-called 1099 status of virtually all gay workers and that this makes them ineligible for all kinds of benefits they have no job security they have no rights when bad things happen the platform says it's not our fault it's you know this is a peer-to-peer exchange we're merely facilitating an exchange and so this has created a highly precarious situation for workers but what's key in this theme is that this is something that has been going on with you know beginning in the 80s with the rise of neoliberalism and the decline of unions and this is more of same kind of neoliberalism on steroids so this view really focuses on the legal dimensions of the business model the legal dimensions of employment the technological aspects of the platforms really don't play much role here because the story is the same it's the Fisher workplace it's the you know precarity etc so it's a very un-technological very much focusing on labor institutions and the regime of labor and the idea of a change in the regime of labor and it also has you know obviously a very strong normative dimension because it's very critical of these changes and then there's a very small literature we could call this the view is sort of the idea that platforms are that they're really not anything on their own and it really depends the way they are depends on where they are and this is a comparative literature particularly with Europe versus Europe in the US comparisons done mostly both Kathleen Thelan who's a political scientist some of these others are economists but basically they're looking at the ways in which European states forced platforms to operate differently so the Germans who pretty much kicked Uber out because it didn't conform to their laws or the French who had been you know fighting the Swedes who forced Uber to change what it was doing and to conform to its taxi laws and so forth so this idea says nothing really unique about platforms all that matters is the regulatory context in which they operate and that's really powerful okay so what's let me just say something about what the weaknesses of these four so I think the economists were these economists it's not all economists these economists who focus on the efficiencies and the technology I think are really this is a really important point to view because the technologies do I think the technologies really all can be very efficient and they can do a lot what they miss is the political economy so they miss the ways in which the technologies affect power relations so between workers and the platforms and then also between the platforms and the state and there's a whole sort of it's not so large but there's a whole literature on the ways so the platforms have been able to avoid regulation for the most part and that's beginning to change a little bit but they've been very successful and you know there are a number of reasons for that but there's a lot of power that's being exercised in this sector and this approach is absolutely oblivious to it the algorithmic manager it's interesting it brought me also back to another another faculty member who I worked with here at UMass and that's Rick Edwards because his work is actually very it's both still very present in this literature but also really relevant for thinking about algorithmic management because one of Rick's three Rick has the three types of control you know direct or simple control and bureaucratic control I think I don't know if I have the words or I think bureaucratic but the idea that there's something new in the technology controlling the workplaces absolutely wrong I mean you just have to think back to the assembly line and the idea that the assembly line controlled the process of work the pace of work etc and using technology to control workers is a very old story as everyone in this room knows number one, number two the idea that the technology is somehow all powerful and takes away all ability of workers to resist or shape their own environments etc is also something we know from a lot of work that's been done in heterodox economics and sociology is wrong there's a lot of technological control and they will be more or less successful depending on whatever so these folks over really overstate the novelty and they overstate the power of the algorithm and it's both because as we get more and more studies of these workers we find there are lots of ways that workers resist the algorithm including through technology and they trick the algorithm into thinking they're doing one thing versus another they learn how to for drivers they learn how to drive their car in a particular way which then makes the algorithm think something's happening for the delivery guys we learn that they can tell the algorithm they're on one kind of vehicle but they're actually on another and that affects what they get so people are fighting back against these algorithms they share information you have great computer scientists who reverse engineering the algorithms to figure out how they work so there's a lot of resistance to the algorithms and that's one of the things that's wrong with them and there's another thing that's wrong with this story which I'm going to get back to the cost of job loss which I'll get to in a minute so the precarity story is that it misses the technology and I think any analysis of what's going on without recognizing that technological innovations and dimensions here is very flawed the technology does matter it matters both because it creates possibility but also it has really changed the way these entities are operating in comparison to conventional firms so yes there's precarity but it turns out there's not precarity for everybody and that's again, cost of job loss I'm going to get to that in a minute but our findings really pass down on the idea that all that's happening is more and more precarity and the institutional chameleons I think understates the extent to which there are common features of platforms the story is not always as different we might be getting more convergence here which is a kind of first first round of things would suggest and I may have forgotten one or two things that was on my notes that's not turning out very good sorry what did we find and what is my argument about how to understand these firms and the approach that I'm arguing the first thing is the firms are engaged in what I call a retreat from control so there are really important aspects of the labor process which conventional management attempts to control but which these firms do not they have given up or they have ceded control over lots of things number one they're open access pretty much anybody can join if you have certain kinds of convictions or things on your record you can't but for the most part they take all comers very different than conventional employers who screen people and hire for things number two one of the things that means much more heterogeneous than conventional firms in a conventional firm what happens is that you come into an HR department and you get steered into a place where the other workers are more like you so across the platform you get lots of heterogeneity in a variety of ways well I'll talk to you about some of the ways but highly heterogeneous, heterogeneous third thing is they cede control over hours of work you can work as much or as little as you want this again brings me back to that first book I wrote which was all about the ways in which employers controlled hours of work this is something that is really just so absent from the debate about what's going on and it's really really important because it's a dimension of this super this high levels of heterogeneity in this labor force and also the way the work gets done so Uber and Lyft may try to do more scripting than some other than some other platforms they'll say give water or have some mince or offer people a phone charger be polite etc but for the most part gig workers are on their own to interact as they like and particularly outside of RideHail in posting for Airbnb or TaskRabbit and so forth these people are performing the labor process in the way that they want to very much in contrast to a lot of service work which is highly highly scripted we want to think of the canonical example of McDonald's and the McDonaldization of society these are consumer services we're talking about so many many jobs in which particularly lower paid jobs in which individuals have are providing services to other whether it's call centers or retail or fast foods and so on and so forth okay so they've given up controlling all of these things how are they going to make a profit well we know first of all that they're not all making a profit I'll come back to that basically they're using technology and market discipline to extract labor out of these workers so the technology is through the crowdsourcing of ratings and reputation and the algorithmic direction the algorithmic management so that is there but they're also using the discipline of the market so they are relying on the market to get work out of workers and I'll show you a little bit more so what we find is that what we call platform dependence the extent to which a worker relies on platform earnings to pay his or her basic expenses that's platform dependence it varies a lot across the platform labor force for many people it's just an extra for others they're wholly dependent and what we find is that all of these outcomes a wide range of outcomes depend on the degree of platform dependence for, as the light went off in my brain a couple of years ago, the cost of job loss for people it's not exactly the cost of job loss but it's a very close cousin and the second point here what I call homo-variance is the finding that people have very different behavioral strategies for they're not all strategies behavioral actions on the platforms there are people who are real homo-economy they're very rational they're calculating, they're maximizing they're optimizing, they've got spreadsheets they know exactly how much money they're making they're trying to figure out how to make more there are people who have I call them homo-socialists very strong social dimensions to what they're doing they won't charge the max they are in it for other things etc they will not do things that they can make a lot of money from because they would involve status insults or a whole range of things that are sort of much more social and not income maximizing and then a third category much debated in the economics literature I call homo-instrumentalists which is that people are just they're just trying to get a certain amount of income that they need so that's the income targeting story and we can talk more about that I think we only got these findings because we did a multi-platform study and one of the problems with a lot of the literature is it studies Uber Uber and Lyft and Uber and Lyft, RideHail although it is very large within this sector it's also very particular in certain ways and just looking at that I think gives you a distorted view of what's going on in lots of other platforms but the key a key point here is that these differences exist not just across platforms but very importantly within the same platform so within the same platform you have people with very different platform dependence very different behavioral strategies very different kinds of outcomes so let me get to that next okay so we did in-depth interviews of 60 to 90 minutes plus surveys today I'm going to talk about 111 earners on 7 platforms I showed them all to you before Airbnb task rabbit that tour relay rides which is the car rental Postmates in favor which are delivery Uber and Lyft which are RideHail we collect data from 2013 to 2016 you had to have done at least 5 trades many of these people have done many many more we had started originally with this younger age range 18 to 34 because they were almost everybody on platforms when we started we were one of the first people to study these things and it started it was very youthful participation on both sides of the market and we mostly kept with that the older Lyft drivers tend to be a little bit older just because we wanted to be a little bit more representative with them and we recruited sometimes through the platform sometimes we sent researchers to orientations we did advertising snowball a lot of different ways it's a whole complicated thing to recruit we got kicked off some of the platforms I'm going to go really fast just through some of the dimensions of the sample our sample is about two-thirds men and one-third women that's pretty normal in the sector as a whole although caveat, one of the things in the notes that I didn't have a slide on is there's a lot of question about exactly who this population is we don't yet have a really good understanding of who they are studies that have been done are pretty problematic but I think this is this is not too off we looked at the care work platforms which are a little bit different of course there would be more women on them the care worker I can get into those in the Q&A on race we have about 60% white 10.5% Latinx 14% black just under 10% Asian not that different than the city as a whole these are mostly people in Boston we have a few others that we did remote interviews the really important demographic thing is the very high levels of education of people in the gig economy they're kind of off the charts now we're a little bit higher because we're in Boston which is a high educated place but it has started to change more lately but particularly until recently I would say the lower wage parts of this sector so even in the driving and ride hell which are the two low wage parts here you have a lot of highly educated people either people who are current students or people of college educations and many of the immigrant drivers for example are people of college educations in their home countries and have come here so to some extent I think it started some technological facility like smart phone and so forth to get started on this it just tends to be very highly educated Airbnb of course is a higher even higher educated than this varies across the platforms so Airbnb and task rabbit and relay rides tend to be very high educated the first two have higher capital requirements than delivery and ride hail and task rabbit has higher human capital requirements something like 70% of task rabbits have which is what we find also have college degrees so I've mentioned this platform dependence thing and this is a key variable that we find is driving outcomes wages satisfaction labor process how they do the work levels of algorithmic control other things too so dependent workers are people who are wholly or primarily dependent on the platform for their livelihood they rely on the earnings to pay for monthly expenses roughly equivalent to full time workers partially dependent pretty much just what it sounds like rely somewhat but either work on multiple platforms and a lot of these they work on multiple platforms especially ride hail and delivery or they have part time jobs small businesses or other sources of income supplemental earners platform earnings are not part of their regular income source and are considered extra many have full time employment or activity and on these platforms the dependence are the smallest group that differs in some cities with ride hail workers but even even there you have one of the things that happens is you'll have a lot of workers who work a small number of hours and then you'll have a smaller group they do the vast majority of the work because they work very long hours so you might have 20% doing 80% of the work but what we find is pretty consistent with survey data both in the US and Europe and it's interesting how consistent this is across lots of platforms and places the bottom is all the platforms together that dependent workers are about 25% partials are about a third and supplementals are the largest group the plurality over 40% it differs a lot in our cases for example we don't have any Airbnb hosts who are full independent and we have no relay rate or relay ride renters who are fully dependent on the other hand are Uber and Lyft people are and that's because if you sample by getting rides since they do most of the rides you're more likely to get done okay so I didn't show you the demographics and this because I was trying to tell you that we have a representative sample I just want you to know who our people are it varies, it will vary a lot by the platforms that you include in your sample and also how you sample so let me just give you some sense of how this platform dependence thing works out so supplemental earners on task rabbit, they get really good wages somewhere between 25 and 150 an hour so really high wages they have high non-pecuniary benefits they do it because they're bored outside of work so they may have really good full-time jobs we have an MIT graduate who works in a lab and she just doesn't work with mice all day I want to talk to people afterwards I hate sitting around being unproductive earning money to go to concerts and so forth because they can be selective with the jobs that they choose they can demand very high wages they can also avoid unsafe or problematic jobs they can see if something looks a little sketchy they don't need to take it and they're able to reduce precarity earnings work is a safety net we have people who are supplemental earners who are using their earnings to build savings or retirement so the idea that this is just creating precarity is clearly wrong from the point of view of supplemental workers it allows them to avoid low end exploitative work so we have a number of people who talked about working in bad low wage jobs and how much more they like this because they have so much more autonomy and some manage a portfolio of earnings and that flexibility and autonomy is really important to a lot of these people both scheduling flexibility so they can control so many people who work on these platforms they do it because they've got something else going on and that needs to take priority whether it's children, family responsibilities a career that they're building and they need to go on auditions or all kinds of things that ability to choose their own hours and their schedules is really key and also to not have a boss is really important and even for the dependent workers the ability to not have a boss is something that's highly valued in the labor force today particularly for people who don't have really high status jobs where you have a lot of autonomy on the other hand the dependent workers lose a lot of these dimensions they do get high hourly wages but there's not enough business to give them full time incomes so they're living the dependent earners in this sample where task rabbits have below poverty or annual earnings so good hourly wages not enough demand and task rabbit has that's been one of the things with task rabbit and its wages got really high but that really limits its demand you see the same thing going on in Rytale the low wages the low prices create lots of demand but then also poor earnings for the workers they lose flexibility and autonomy because in order to make those expenses they have to make their monthly expenses they have to take jobs and this deals wage jeopardy where they take jobs where in the end things they're going to get so they have to take more scammy jobs and those are out there they have to take more unsafe jobs and so forth and another thing that's been happening on these platforms because we've watched them over time is the kind of downward trajectory in terms of degradation of the labor conditions supplemental earnings for postmates in favor so people think it's reasonable extra money they love the non-pecuniary benefits so there are people who do it so they're going to earn money while they're exercising if you're a supplemental earner you can avoid unsafe conditions so one of our examples is a woman she does a lot of different gigs she's a married sort of middle class woman she has a kid she does postmates with a car she goes in the evenings she keeps her daughter in the car in the car seat she insists that the person come down to her car total violation of the rules of the platform no no no no I will not get out of my car so the ability of the algorithm to manage the workers is much more limited with the supplementals or another supplemental who says oh I never do the things they want us to I don't put them in the bags I don't use the stickers and more and more well what if the platform didn't like your ratings and said you needed to come for an orientation oh well I wouldn't do it because I don't need to and they don't worry about their ratings this is one of the things that's been very big in the literature the idea that all the workers live in fear of the ratings and being deactivated so they have a lot of autonomy they do the job the way they want in contrast to the dependence this is a kind of job of last resort it has the lowest earnings in our sample demand is very erratic so you have the afternoon and the evening and then maybe some very late evening you've got a lot of dead time in the middle of the day that also happens with the ride hail vulnerability to weather and traffic a lot of these full time dependent couriers particularly in Europe where there's a lot more of this are getting into accidents they're having lots of problems they've also the dependent workers are more likely to have lost their scheduling autonomy because the firms are putting more pressure on them and so there's this wage autonomy trade off it's also built into the way the algorithms go if you want a wage guarantee you have to sign up for a certain block and take everything in that block so the companies are in some cases they're moving away from that total what I call retreat from control finally at the last case I'll tell you about this uber lift so the supplementals the earnings not as good as they were but the earnings have been good people really like the flexibility and autonomy they use their spare time productively they reduce costs associated with full time work so somebody who has a job in the city but lives in the suburb turns on the uber app gets somebody to drive in with him gets the toll gets paid for gets expenses for the car it works very well so they use it to supplement inadequate compensation in their full time jobs to finance leisure spending and we have some interesting cases of people who went from part time to full time one really sad case of a guy who felt his job was dead end he loved the part time uber thing decided to go full time on uber and it was a nightmare because one of the things about these probably the lower wage jobs is people many people have no idea how much money they are making or aren't making we had one interesting interview where we asked the person about their hourly wage he actually had never calculated he sat there in the interview and calculated and found out he had been making $6 an hour and that was for a delivery and if you've read the literature you know that once you have a if you put in a sort of hefty charge for depreciation and maintenance and so forth for these cars it erodes the hourly the net hourly wage quite a bit so people may think they're making a lot of money and in fact they're not and there's a lot of worry about what happens in April when the tax bills come due so the dependent owners have pretty much lost control of their hours they have to work to the market they work long hours so that guy I just talked about was working 12 hour days and they get back problems they have the tax bills the platforms are constantly changing things so the asymmetric information is important and they worry a lot more about deactivation so that's the story the market discipline matters a lot and the technology matters and I would argue has created a new kind of firm which has seated control over a whole range of things is using the market to control and the technology let me just say I said a word or two I pretty much probably went through this already homo variants by platform we found that about 35% of the sample were um excuse me I'm missing that one anyway so homo economicus about 43% on air bnb a little over a third on relay rides 20% on task rabbit homo sociologists is the big one on air bnb a lot of them are there for social reasons and then you have that homo instrumentalist air bnb it's not it's pretty simple and mostly people have a social orientation but relay rides and task gravel we have people that just say oh I'm just here to earn some money they don't think about strategy they don't really have a social orientation they just need to make a certain amount of money to pay their bills or as one of them said I needed cash she was a musician how to get some cash and relay rides came out I love that quote ok there's also a hierarchy I think I've mentioned it air bnb is at the top to task rabbit postmates in favor and that matters also the hierarchy matches the conventional labor force in terms of race education and other things that sort of structure conventional labor markets ok I'm going to finish up in a second I just want to say one thing about our last case which was a platform cooperative as I mentioned and there it's no longer technology plus the market maybe it's technology plus solidarity but this is a platform that is owned and governed by the artists themselves it's been very successful it went a bit up market there's a lot of photography world there's a getty images is sort of the McDonald's of stock it's a boutique so the earnings are better because it's a cooperative it returns a much larger fraction of the revenue to the earners the artists are very happy the governance is working well highly satisfied members one little asterisk about the governance there's not a lot of literature about cooperative governance they get somewhere between 20 and 30 percent of people participating which I think for co-ops is actually not too bad it has challenges because it's global there's people from 65 countries so meetings you can't have synchronous meetings you also have problems of language but there are other challenges for cooperatives on platforms one of the reasons these are really interesting because I got interested in them really early on because I thought this technology can solve a lot of issues for service providers because it can solve the reputational issues so you can cut out all those middle people agents and brokers and so forth particularly if they can home health aides and so forth who may take up half of the really high fees and so forth if you can solve the reputational problems you could actually give a lot more back to the providers and so it seemed like a really interesting thing and also we saw how rapidly these things scaled typical worker cooperatives may take years to get off the ground and have a couple of employees but these platforms were just growing like wildfires so if you get the economics right and people can just hop onto the platform you could actually cover lots and lots of workers in a short period of time so there's a freelancers co-op in Europe that has over 35,000 members and it just you know just grows really fast there are challenges one as I talked to earlier the heterogeneity of member orientations so just like all the other platforms there are people out here for whom this is a full time job they're devoting lots and lots of money attention effort and then there are what they disparagingly call mom-tographers women who take their iPhone out at the breakfast table and take pictures of their cute kids so those are the low hours the equivalent of the part-time Uber drivers so that creates really tensions among the membership in terms of what they're looking for the kinds of policies they want and the other thing is then on a platform the unlike a plywood cooperative or any of the traditional worker, Mondragon the traditional worker co-ops these are individual producers and in where you have artistry unlike say taxi driving or you know some of the more things where the work itself will be more homogeneous you have you know very heterogeneous levels of talent also the fact that it was a cooperative attracted some really highly talented photographers who would never do stock photography otherwise so that was a positive thing but they ended up having a hugely unequal revenue distribution and more unequal than a US wealth distribution so they've got 9 people who get a quarter of the annual revenue and so it's but these people are putting a lot of money and effort into it so there's that and then there's one other thing which is that it's not an open platform so they capped membership which is another a challenge for these cooperatives are they going to be open and if they are will they be able to get enough demand to get decent earnings for everybody or you know I think this was a smart choice for these guys but once you close it you lose important dimensions of the platform and then there are the generic challenges some of which are for all cooperatives but these are financing attracting customers and what we call the tyranny of the market one of the things that the artists told us about was that the buyers wanted pictures of affluent white people and so that was the market you know they have to succeed they've got to work to that market and in fact where they are the images are of people from global south countries they are listed under travel so there's a kind of what we call a neo imperialist a neo imperialist market there which the co-op you know to succeed it's we all co-op space this they are just firms in a market where if they you know have to meet customers okay so key findings I would argue we have a new kind of labor regime here high heterogeneity across numerous dimensions technology is key but of course the political economic factors are also central these platforms can be highly desirable for workers because they can offer high levels of autonomy and freedom but first of all the platforms have to allow that and one of the questions really important question to ask is is this just a transitional stage are the platforms that you know they're not all making money some of them are ride hell isn't I do think ride hell is kind of different and special but can they make money in a world in which they allow people to work as much or as little as they want and so and so forth can they extract profit without more labor control we know a platform like Airbnb can do it but it is a platform with a lot of capital relative to some of these others and so this is this is a question I mean Airbnb is making money and it has been from a pretty early stage because it really is just taking a transaction fee on something that is much more a peer-to-peer exchange employment status should it happen is likely to transform the whole model and a lot of the people who have been pushing for employment status are arguing that it doesn't need to take away flexibility and they are right technically that the firms can continue to maintain flexibility for the workers even if they make them employees but I am almost certain they will not because if they have all those additional costs of employment they are going to want to get more work out of these workers and they are going to there is one platform that I mean confidentially we have found out that what they switched some they switched their California workers to employees in anticipation of the change in employment status so I think that employment status is going to make a huge difference and if it ends up happening it may or may not but I will just leave it there and open up for questions thank you very much so questions comments thank you Jerry saving me yet again so picking up the last thing you said about employment status is not going to likely be a way to improve these platforms and help these platforms change the labor process and so forth so how would you recommend regulations norms anything else to try to make this a more viable model for everybody well I mean you can go the employment status way but it basically means they are going to revert to being conventional firms so I think that is the first thing I think the co-op model is the way to go because the co-op the workers can decide how much flexibility and autonomy they want they might be able to set minimum hours but since autonomy and flexibility seems to be highly desired by many people in this workforce they can also give more and give up a little bit on the profits because they are going to be getting a lot more in profits through the technology and the other efficiencies that they are getting on the platform so I think that is some of the there are now some small co-ops I picked this one to study because it was the only large size platform co-op that I could study in the United States it is actually in Canada but that I could study at the time now there are some cleaning co-ops and there are some small taxi co-ops and so forth but the thing is that the technology is pretty replicable it takes over a huge fraction of management functions and the question you have to really ask and I talk about this in my book is have these firms innovated themselves out of a role because the more that the technology does yeah you need some of the platform designers but what are the capitalists doing what are the owners contributing here not very much so I mean proponents of worker co-ops have argued forever that you don't need capitalists and workers can do the management typically you get a set of managers but they basically don't even exist in this world there are very very few employees at these platforms you develop your model your technology and now there's so many of them and they're so replicable it just I think that's the way to go it just makes a lot more sense and then the workers get so much more of what they want or should be able to that's one follow up which is so one of your key findings is this thing about dependence how the workers benefit or not depending on whether they're dependent so as a compliment to this to make them be able to benefit more would you propose universal basic income as a floor so they become less dependent on this kind of employment for their standard of living well you could do that there are other ways to do it you could be on multiple platforms the worker the worker platforms also give them more incomes so they can if they're in a worker owned so if they're in worker owned platforms they could they can meet their expenses a lot more easily and also have autonomy the thing about being all in on one co-op is the governance but that's also the fact of that question of size like is it a small platform where the workers actually have a lot of say over what happens or you're talking about a global platform where each member is really small so they probably wouldn't have that much say they would just have to shop around for the co-op that does things the way they want it Lee really interesting project we have here I'm curious about the European regulation you alluded to that as at least being one of the so what are those forms of regulation so in Sweden when Uber came in it basically just had to operate like a regular taxi so they had to do minimal wage I think when the affairs were set the government made a little bit of accommodation because they had a lot of regulations around the taxi meters themselves and Uber was able to get them to change that because why have to buy those meters when you have everything on the smartphone so that's one thing and actually some of that has already come to the US so New York now has a minimum income it has a cap on drivers and also the minimum wage both with and without expenses so the calculate expenses and so forth in Germany Uber just didn't want to follow their rules why they did it in Sweden and not in Germany I'm not sure in the Netherlands they had to close down one kind of Uber driving and they opened up a different platform they had some called UberPOP but now the UberX they work more like higher in livery services so they adjust their pricing they adjust their employment practices pretty much a lot of things and then there's work that's just being done in global south countries where you have a lot of other stuff happening for example quite unfortunate some of it contract labor so kind of bosses subcontracting and so forth they might provide the cars but then they're extracting a lot of it so you're losing a lot of what you could get from the platforms in China which has a lot of it's not Uber, it's a DD Uber didn't work out in China but it's all through these subcontracting relations which is I mean that speaks to that institutional chameleon view because they adapted to in these places they're adapting to those existing hiring and stuff yeah a retreat from control was very interesting which is sort of as I understood it opposed to the algorithmic manager theory and I was just wondering whether there are still elements there that you do think you know you sort of haven't disappeared for instance in the technology inheritance that workers aren't less likely to be in the same locations of that more generally you know isn't different from traditional workplaces but you know where we also workers also find ways of getting around managerial control algorithmic control yeah so I don't want to go all the way in the direction of saying the algorithm doesn't determine anything obviously it does there are things that it can do but you know what we found is that the extent to which people pay attention to the algorithm does vary by their dependency status so that's one thing the organizing thing is really interesting so until recently everybody kept writing these people are going to be impossible to organize because they don't know each other and they're you know they're not in a centralized workforce and it turns out it's not true we were seeing connection we were seeing that people were making not everybody the people were making connections so for example actually one of the people on my new NSF team studied Uber drivers at Logan airport because they conquered the airport they congregated there in the waiting parking lots and they organized a strike some years ago and it didn't succeed but there are also networks, ethnic networks and other kinds of networks of recruiting so people know through their prior social networks know each other before they get on to the platforms but what we saw in May and what we saw increasingly now with the Instacart strike and so forth is that the social media, digital communications is actually can actually work to organize people so there is a union in Southern California the rideshare workers union they are doing face-to-face work but they're able to find each other there are also a lot of driver forms and so forth so the communication technologies are helping them and I don't have the view I never really did have the view that this is an impossible to organize workforce people who are hard to organize are the very part timers because they don't have much of a stake but where you have those groups of very long hour platform dependent workers it seems that they are organizable you also see it on what they call the crowd work or the digital labor like mechanical Turk and so forth that they have forums and they find each other I mean that's all virtual they don't do face-to-face David I thought you are finding that there is such a dramatic difference between the experience of the supplementals and the dependence seems to be an important result which assuming that these platforms are not going to be mainly caused but owned by someone and operated for profit it suggests that there has to be a power relation to have a chance of getting a profit for the owner and if the workers are not fully dependent on their earnings from it then it's hard to insert power over them and their experience is going to be pretty good but if they are fully dependent then you always have a lot of power and probably will be able to extract profits they do, the owners do control the technology and the branding so this raises the question of whether this can become a major alternative source of full time earnings in the economy yeah well that was the question that I asked on the right there so I think if you look at some of these platforms they are able to be profitable by just taking a fairly small fraction of the transaction because there are so many transactions so Airbnb Etsy a lot of them they are making money and I think the delivery also are they are more profitable the ride hail is a more complicated case I mean I think in ride hail they could first of all in ride hail right now the reason a key part of the reason they are not making money is that they are subsidizing the rides at a very high level they are subsidizing about 40% subsidy for every ride so it is a predatory pricing strategy because they want to wipe out all the competition and they think once they are the only game in town then they can raise the prices to become profitable and I don't think that is right I mean unless there is absolutely no public transportation and nobody can have access to cars people go on these ride hails in large part because they are so cheap and we know also that they have increased the number of rides instead of walking, cycling or not going anywhere a large fraction of their rides are people because they are so cheap I mean you can get ride hail at the price of public transportation so you get incredible convenience at no higher cost but when what was now a $5 ride or a $3.5 Uber pool is suddenly going to be the price of what a taxi used to be that market is going to shrink dramatically so they knocked off the taxis easy because it was a a big rent high barriers to entry and now they have a model that I just don't think is profitable at the level I don't think it is so much the drivers I think they actually could take a reasonable fee from drivers but they are just way too big right now to be profitable that is my view of it because the extract labor from labor power model is not the only model that works in the economy other models are take a fee on a transaction a peer to peer transaction and that is what these could be yeah, Nancy so I heard a kind of a researcher reporting here .com that kind of fits into what you and David were talking about there because it is in some sense a peer to peer service and that it is matching potential home care providers or customers but it is so oriented towards attracting customers to use the interface that is a fundamental asymmetry which is it allows employers to post reviews of workers but it doesn't allow workers to post any reviews of employers so they don't get any information about whether they will be treated well or not and it seems like an interesting mix of kind of a control over workers within a peer to peer yeah and you definitely have you've got that variance across the sectors on the runway I don't know if they're still known but they used to rate the customers too and we just don't know we don't see the ratings that they give so for one of the things that I've heard I think is true is that if you're a low rated customer you will get matched with a low rated driver yeah so whereas Airbnb has fully bilateral there are issues around when you reveal so do both sides have to good practices that both sides have to put their reviews in before they can see the other that's a problematic feature and I think one of the things is because care.com is trying to be like the uber of its sector which is to grow really rapidly and there's a little bit of research on it that I've seen and interesting things about care.com is I mean I absolutely agree with you about the asymmetry it should be symmetric you've got people moving from a problematic but you know has its pros and cons very informal labor market to a more formal one which should give them some benefits and you know some I don't mean literally benefits I mean like you know they're going to get their pay instead of the employer not giving it to them or there are supposed to be a set of guidelines that the buyers are supposed to versus others who are going to be losing autonomy so you know there are independent contractors who have lots of autonomy who go on to platforms and then they're more subject to platform rules and that's a kind of interesting thing that's happening in some of these markets how platforms are changing the relationships. Can I just talk a little bit about the the I mean it seems to me that care.com is a really good example of eliminating a very inefficient middleman that you know there's small care contracting companies that really do extract huge rents from both customers and rooms about it and so there's been a lot of discussion of cooperative home care there is a really good you know really big cooperative home care firm in New York City but it seems like the problem with the cooperative model that you're talking about is it's a co-op for kind of the management of a really large national or international company and in a way that preempts it preempts the potential for local co-ops you know like local care providers to get together to provide their own kind of interface and model because you know it just seems like a lot of the power of these platforms comes from the huge economies of scale that have to do with standardizing across local communities. Yeah well some I mean the I think I said something about you know size of co-op and how though I was talking about the governance so you've got so yeah so there's the cooperative workers I think it's I know the one you're talking about in New York which is not a platform as I understand it they may be going moving some stuff. There's a very small nurses one that's starting somewhere in the Midwest I mean there are a lot of these there's a whole movement now for these I agree that that localness brings you a lot of very positive things but partly it depends on what you're talking about so for stock photography where there's really no interaction between the buyer and the seller the big or sort of generic freelancing or digital labor those things the bigger more global even platforms make more sense I think for the some of these care services and other personal services smaller makes sense brightly which is a house cleaning co-op is franchising itself now so you're getting the benefits of the small but they're going to get the economies of scale because they're going to use the technology and marketing stuff okay let's take I'll speak faster I think there's these digital platforms we thought there was something else going on they're highly profitable is it the same kind of labor is it the way we use these categories to fit all that yeah it's a great question so I didn't get into it in the definitions when the sharing economy what came up all the discussion of it was how novel it is and there's never been anything like this before in all of which is mostly propaganda John Zeisman and Martin Kenny wrote a really important paper in which they said do not think about these sharing services as something new or different they're just they're just another part of the ecosystem of the platform economy and the big platforms of course being the ones that you talked about so they are trying to think of them as just sort of minor you know add-ons to these big platforms and of course these platforms were important investors in some of these I mean Google is a major early investor in Uber for example so I do think there are dimensions of these consumer service platforms that are very similar but the economics the business the services etc are different on Facebook it's all that free labor right they're subsisting on all the free labor of our posting and our eyeballs and all of that which is different here this is more conventional in the sense of service provision I mean these are people I'm talking about offline service provision if we think about mechanical Turk what it's doing I mean it's doing all that AI work for those other platforms so they are related and I do think that that's you might want to look at their work Martin Kenny and John Seisman to sort of think about that ecosystem platform ecosystem really quick the last two very short so I think on something I mean for the potential for the cooperation of this I mean it seems to me that the biggest non-technical barrier of entry is advertising right I mean you've got to get a ride share it's basically there's one or two Uber and Lyft but how can you have a cooperative model in which you need this enormous amount of money to make it work and I'm wondering what are the owners doing here I'm assuming it's advertising and also is it regulatory fixing I mean are they lobbying in some way and how would that interfere with having a cooperative work well I mean the cooperative probably wouldn't need as much regulation because they wouldn't be doing a lot of the things that people are although that's not completely the case I mean if you look at issues of traffic or Airbnb I mean there's a FairBnB is a cooperative alternative to Airbnb that's starting up I mean they may if it got really big it would create the same problems Airbnb is like a whole other store because it's an insider-outsider issue rather than a labor exploitation but the thing about most of these services and this gets back to Nancy's point they are local services so Uber never should have been or needed to be a global platform it wanted to do that because it thought that was the way to make huge amounts of money but you need interoperability across the apps so like the easy pass and so forth that it works in lots of different places we could all just have local apps and then when we travel they could work so you don't actually have to market to huge numbers of people you have to market in a local area because these are all face-to-face services Airbnb is a bit different because it deals with travel but the rest of them are the errands the delivery and the ride-hail which are the big there's a big segments are all local what's up? So I'm interested to hear what you think about the size of this marketing so the Bureau of Statistics survey backwards 0.17 and they found that about 1% of those employed were actually operating these electrically-mediated platforms there's plenty of reason to believe that that's an understatement that it might be biased downwards there's other data from the gallery the the survey of household economic and city-making from the Fed if you look at the data now it might look like it's around 10 to 12% so I'm curious what you think of the size of this market and seem to speak to to say the point you draw is whether or not this is a transitional phase is this something that's actually going to change to make sure So the contingent worker survey by the BLS was kind of a disaster because they asked about full-time and you know what they needed to do was actually it's an old friend of mine who ran that thing I didn't talk to him beforehand but anybody who was studying the sector knows that the full-timers are a small fraction so that one's way too low and they know it in there the shed is I think it's probably around 5% the bank data the JPMorgan Chase data gets to about a 5% of the labor force participating when you go to the shed you're adding in one of other things like selling stuff on eBay and so forth so if you're just talking about gig work versus resell because the biggest thing that people do is sell stuff on eBay in terms of online earnings so it's not that big yet I do think that more and more of the way we interact economically is going to move on to platforms and I think there is a potential for a significant increase here so I mean you think about care work I mean there's a lot more care work that can go other domestic service just so I think there's potential I mean it's never going to be 100% like some of these end-of-employment people but where you also have a lot of these freelancers and so forth moving on platform too I mean I didn't talk about some up work and you know they're higher end platforms you've got lawyers on platforms you have doctors on platforms now so I think you're going to see more and more of it because it offers a lot of efficiencies but what economic model is going to be a different story they may not you know they may use a very different economic model they may just use a conventional model but they use scheduling or something like that thank you very much