For more than three decades, India’s growth story has rested on the promise of a large and youthful workforce—but whether that promise is being realized remains an open question.

A new report published by the Centre for Sustainable Employment at Azim Premji University—State of Working India 2026—takes a comprehensive look at how young Indians move from education into the labor market—and asks whether India is successfully converting its demographic dividend into an economic one.

The report documents a striking paradox: even as educational attainment has expanded dramatically, the transition to stable, gainful employment remains uncertain—with high graduate unemployment, limited job creation outside agriculture, and persistent gaps between aspirations and opportunities.

To discuss the report, this week on the show Milan speaks with the report’s lead author Rosa Abraham, who heads the Centre for Sustainable Employment at Azim Premji University. Her research focuses on informal work and women’s employment, with a particular interest in issues at the intersection of labor statistics and women’s work. Prior to joining the university, she worked as a researcher at the Ashoka Trust for Research in Ecology and Environment and as a lecturer at the Madras School of Economics.

Milan and Rosa discuss the state of India’s mythical “demographic dividend,” the quality and quantity of higher education, and India’s stalled structural transformation. Plus, the two discuss the high unemployment rate for college graduates, trends in internal migration, and the loosening of caste-based occupational segregation.

Episode notes:

1. “India’s Middle Class Hits a Breaking Point (with Saurabh Mukherjea and Nandita Rajhansa),” Grand Tamasha, April 15, 2026.
2. Rishita Khanna, “‘We are not overproducing graduates, we are underproducing good jobs,’” Hindu, March 25, 2026.
3. Soutik Biswas, “India's young are more educated than ever. So why are so many jobless?” BBC, March 19, 2026.
4. Karthik Madhavapeddi, “‘For 1st Time In 4 Decades, Young Men Are Withdrawing From Education,’” IndiaSpend, March 27, 2026.

Transcript

Note: this is an AI-generated transcript and may contain errors

Milan Vaishnav Welcome to Grand Tamasha, a co-production of the Carnegie Endowment for International Peace and Hindustan Times. I'm your host, Milan Vaishnav. For more than three decades, India's growth story has rested on the promise of a large and youthful workforce, but whether that promise is being realized remains an open question. A new report published by the Center for Sustainable Employment at Azim Premji University, State of Working India 2026, takes a comprehensive look at how young Indians move from education into the labor market and asks whether India is successfully converting its demographic dividend into an economic one. The report documents a striking paradox. Even as educational attainment has expanded dramatically, the transition to stable, gainful employment remains uncertain, with high graduate unemployment, limited job creation outside of agriculture, and persistent gaps between aspirations and opportunities. To discuss the report, I'm pleased to welcome Rosa Abraham to the show. Rosa heads the Center for Sustainable Employment and is the lead author of State of Working India 2026. Her research focuses on informal work in women's employment with a particular interest in issues at the intersection of labor statistics and women's work. Prior to joining the university, she worked as a researcher at the Ashoka Trust for Research in Ecology and the Environment and as a lecturer at the Madras School of Economics. I'm pleased to welcome Rosa to the show for the very first time. Rosa, congrats on the report and thanks so much for taking the time.

Rosa Abraham Thank you so much for having me, Milan.

Milan Vaishnav I want to start off by maybe just asking you to give our listeners a bit of background about the State of India working report. As I understand it, this is the fifth edition. It's become a kind of flagship report that your center puts out every year. This edition, much like some of the earlier ones, has gotten a lot of attention. Tell us a little bit about the origin story behind doing this annual publication and what this year's focus is.

Rosa Abraham So the State of Working India report, the first edition of this, was released in 2018. And it really started off as an initiative within the university to try to understand the landscape of the labor market in India. So, there were a lot of papers, research, which looked at similar questions. And the intention with the SWIs, the State of Working India report was just to try to take stock of where we are in terms of the growth process and how the labor markets done. And over the years, like you've said, we've had five editions of the SWI. And the first one took stock of where are the jobs, what are the kinds of jobs, who works in these jobs. And after that, we’ve had more thematic focus. So we had the SWI 2021, which looked at the impact of COVID, especially the impact on women in the labor market. And then in 2023, which was the previous one, we were trying to look at the impact of social identity. So, what is the role of gender, caste and religion and how it intersects with labor market outcomes. And then this year we looked specifically at the question of youth in the labor market to understand questions around education and the school-to-work transition. So, this is the fifth edition and hopefully there's a few more to come.

Milan Vaishnav I want to just ask you a bit about the data that goes into this because anyone who opens up this report, and we'll link to this in our show notes, and I hope our listeners will have a read, I mean, it's a pretty exhaustive deep dive into the world of labor statistics. And, you know, as you know better than I, this can be a very fraught world. Labor statistics are politically contested. There have been controversies around their release, around their fidelity, around the distinction between say publicly provided data versus private data like that compiled by CMI or other third-party providers. Tell us a little bit about the kind of data repository you take advantage of to put out these reports.

Rosa Abraham Yeah. So, I think it's interesting, you know, when we started the report, we did face this issue of not having very recent, nationally representative, publicly available labor force statistics. But I can say that since then, especially from 2017 onwards, we now have annual data available on the labor market and that comes from the Periodic Labor Force surveys and that's something that we have systematically been using in these reports and these are nationally representative, they are sort of complying with international standards on how employment is measured broadly. So, in that sense we have now a wealth of data and we've also got, with the next round of these nationally representative surveys, we will also get data that continues to be yearly but will also promises to be district representative. So, there's a huge scope for further analysis now with the data. When we did feel positive data was, you know, when we were doing the COVID analysis in 2021, because we didn't, the labor force surveys had sort of not been released. You don't get it at, at, you, know, there's a lag in the release. What we then did was to rely on the privately available consumer fit image surveys produced by CMI. Now, there are issues around how representative is this. There are issues about the comparability because how we measure employment, how they ask questions is very different. But I think one thing that we've been very consistent about when we've bringing out these reports is really caveating the data. So, for example, even with the periodic labor force surveys, which is the nationally representative surveys, there is a broad understanding that this seems to miss upper end of employment or the high-income households. There's a non-response in urban areas, you have your gated communities. So, there's upper high-income households that these surveys tend to miss out and we're well aware of that. With CMI also, we are aware of shortcomings, especially when it comes to measuring women's income. But knowing these caveats, then we do then provide some kind of benchmarking of these data sets against the nationally representative data sets. So, we do have some idea of what they do measure and what they don't measure. And that's something that we've been doing with the reports. In this year's edition of the report, I think we've kind of pushed ourselves to go beyond the labor force surveys and what we brought in with this 2026 edition is a lot of administrative data. So, these are your MIS portals that's available through different kinds of government schemes or government initiatives. And so, the data that we've used, one of this is coming from eShram, which is a portal of informal workers, a registry of informal workers across the country. It's self-reported, there's no sort of sampling data collection as such, but that gives us a lot of information on migration flows, where do young informal workers move to. We've also relied on information from the industrial training institutes, which [are] the major providers of vocational education in the country, and that comes from the, again, the administrative portals of ITIs, so it gives us information on how many students enrolled, their caste, backgrounds, and so on. And the last kind of big administrative data set that we've used is the All-India Survey of Higher Education. This is essentially a repository where institutes of higher education are expected to submit, you know, basic information on students enrolled, number of vacancies, number of teachers, infrastructure and so on. And I think, you know, we've been able to, again like I said, understand what these things capture, but more importantly, what they don't capture. And that kind of runs throughout the report. But bringing all of this together has kind of given us a very good understanding of, you know, not just labor market outcomes, but things that precede that as well, questions around higher education quality and so on.

Milan Vaishnav I think that's very useful to give people a kind of baseline in terms of understanding, you know, the nuts and bolts that go into this. I want to pivot to talking about this sort of framing of this report, which is very much around this idea of India's demographic dividend. And, you know, I think for people who have been studying, learning, writing about India, the demographic dividend is like this mythical creature that has been talked about for literally decades because it is a process that unfolds over decades. In the report, you define it very simply as an age structure where the working age population increases relative to the dependent population. And you note that India is nearing the final phase of this demographic dividend. So here we are in 2026. Tell us where exactly India sits on its dividend and how narrow is this window as we look ahead.

Rosa Abraham Yeah. So I think, you know, there's a lot of different definitions of demographic dividends. So if you start looking at this literature, especially in the Indian context on what do you, what is the phase where you're in a dividend phase and when does that dividend shift and become a liability? Like you said in the report, we adopted the simple number of working age people per dependent, that ratio, and as long as that ratio is increasing, we say, well, we're at a demographic dividend, and we find that it's peaking at 2030, so that's four years from now, so we're very much at the end of that dividend phase. But if you look at…there is different definitions, and we stayed away in the report from engaging with the different definitions mostly because the broad takeaway is that we are in the far end of this dividend phase. So if you know there's another definition which says well the dependency ratio at least in terms of the young dependents should be less than 30 percent. And the old dividends should be less than 15% and when this shifts, that's the other point at which you know you're getting into the demographic dividend shifting. Now if you use that measure, you will get the number at 2035, which is another 10 years from now. Different measures but the latest that pushes it till is 2040. So, we have at best another 15 years of you know a working-age population that is probably growing fast or is accounting for a larger share than the dependent population. So, in that sense, irrespective of what definition you're looking at, we are at a place where our youth share is declining in absolute numbers. In fact, the zero to 14-year-olds, which is the youngest of what, has already started a decline. That set in from 2020 onwards, and the 15 to 29-year-olds, that is now recently in 2025, that started to decline. So, the absolute number of the youth is falling, but population has this inbuilt momentum. If you already have a young workforce, they'll start growing and that working age will still be growing in a sense. But it's only a matter of 10-15 years when we're going to have more dependents, support, you know, relying on every working age individual. Which is why it's all the more imperative that it's not just about the immediate need to productively engage a young workforce, but it is also thinking about we're getting into an aging society. And so every young person in that society is going to be supporting more older people. And those older people who are, you know, today's young, if they haven't built up the kind of social security provisions or the pension systems to support them when they are old, then they're just going to have an additional financial burden on this already burdened young person in the future.

Milan Vaishnav So, you know, the report emphasizes that in order to convert this demographic dividend into an economic growth dividend, you need to have a series of steps taken both on the supply and demand sides. And a lot of the report is really structured around those different supply and demands side constraints, and we're going to talk about some of them. But I wonder if you could just at a kind of macro level, tell us: What are the most significant binding constraints on each dimension today, and where do you see policy falling short?

Rosa Abraham I think, so there's both a supply side constraint and there's a demand side constraint. So, on the supply side, what we've seen is that there's this massive influx of graduates into the job market. And that's come because we've had this huge expansion of colleges, especially post-liberalization, a lot of private universities and private colleges coming up. So, if you think of when... Skills in courts defined as just having a degree, we've kind of ticked that box. But the constraint that we have is transitioning that skilled workforce into employment. And there's multiple kinds of constraints. And I'm sticking with the supply side. One is that a lot of the school-to-work transition is not happening very smoothly. So what we find is that if you track a young worker that has just graduated and you track them for up to a year, you're finding that only about 7% of these young graduates are finding salaried employment within a year. So, half of them do find employment within a year, but it's only about 7% who eventually find salaried employment. And the other side of this is where we're confronted with the demand side, that we just haven't created enough jobs that's going to employ this highly educated workforce. So, because of education, there's also this aspiration on what kinds of jobs they want to work in. If you look at the growth story that India has, most of our jobs have been largely in construction, to some extent in manufacturing, but following construction, the next major employer is low productivity services. And by low productivity services, I mean trade, so retail and wholesale trade, transport, and to some extend communication as well. Now, for a graduate, there's a sense that there are certain kinds of jobs which they think are suitable for them. And when those jobs are not available, you will find that they kind of stay back and may not really engage with the labor market immediately. And so that's one kind of constraint that and that's I think the sort of overarching constraint which is that we just haven't created enough jobs and not just that, it's even if we have created jobs a lot of the job has work has been informal, which on its own is possibly may be okay because there's also this literature that argues that, okay, you enter the labor market as an informal worker, you know, maybe as a gig worker, or in kind of some kind of part-time contracts, and then over time, then you transition into a full-time, paid, secure employment. But you're finding that in the labor-market in India, that kind of transition from informality to formality also doesn't happen, there's no job ladder, so to speak. So, you have a supply of graduates with certain kinds of aspirations and then on the other hand you have a private sector which has failed to create that kinds of jobs that they aspire for and a public sector which would have traditionally or at least historically created that jobs but that's also shrinking over time. So together a lot of these are kind of creating these kinds of surplus labor.

Milan Vaishnav So there's a lot there to unpack. Let me just focus on one element of it, which is the kind of provision of higher education. There is a remarkable set of statistics around this in the report, which documents this massive expansion in the provision of higher education. So you write that basically around the time of liberalization, early 1990s, there were around one thousand six hundred forty-four institutes of higher education. At last count, there were nearly 70,000, right? So it's just been like explosive growth. And this has totally upended and remade the education landscape in the country. And I wonder, what kind of data is available to tell us about quantity versus quality and the kinds of skills and education that people are actually getting from attending these colleges, which have blown up over the past couple of decades?

Rosa Abraham Yeah, so the quantity we do have. So all of that quantity numbers that the report has, which is just the number of educations, comes from the All India Survey of Higher Education Institutions. Now on quality, that's where we actually have very little information, especially the quality of higher education. So in India, we have information on the quality of our primary and secondary schooling and this comes from the annual status of education report which is a survey done by an NGO called Fatam and they kind of look at learning scores, they look at, you know, how well can students read, how well can they do basic maths. So, there is some understanding on quality of primary and secondary education, but we really have no such understanding when it comes to higher education. There’s two ways in which you can think of quality at the higher education level. One is just quality in terms of, well, how good are these colleges, how good are they meeting the needs of the students? And the other way to look at quality is if you think of education as something that allows youngsters to smoothly transition into the labor market, where they meet the needs of industry and employers, you know, and that's kind of also another mark of quality, that higher education is doing what it's supposed to do, which is providing a skilled workforce to industry. It could be manufacturing, it could be services. Now on the first kind of measure, which is really how good our college is in providing education, we try and get at this somewhat. I don't think we've done a thorough job of it and partly because we are constrained in terms of the data available. One way we try to do this is we have information on the number of youngsters at the district level, so number of youth in the district. We also have information of the number of colleges in the districts. So we say, well, are colleges available? What is the kind of provisioning that college, access to college education? And that has grown. So it's grown from about 29 in early 2000, so 29 colleges per lakh youth to about 45. So, there's been a huge increase. The other kind of way that we look at this in terms of how good our educational institutions are meeting these student needs is through teacher availability. I will say that there's not been any systematic and it's not completely unambiguous relationship between number of teachers per student and quality of education, but I do think it's a good measure of resource constraints being faced by higher education institutions. And if you look at that, you can see that even that number has not really kept pace with the huge number of students that has gone up. So, the prescribed norm in India is, and this is norms prescribed by the ministry and sort of formal national organizations that mandate this, which is that it's around 15 to 25 students per teacher. And even when we were at our best, that was in early 2010, we have never met that number. So public colleges, the number of students per teacher has always exceeded 25. And in the last few years, that number has steadily increased. So, what you're seeing is that colleges, in terms of at least the infrastructure of colleges and the physical provisioning of colleges has increased, but resources that can make these colleges function effectively, one of just being teachers, has not kept pace. So that's the closest that we can get in terms of at least quality of higher education from one end. The other one, which is the, you know, how good are they in doing what industry requires them to do? Again, in this report, we don't have that. That's something that we hope to look into in the future. But there's a lot of other studies that we rely on. So, there's these surveys of employers that have been done which find their measures of employability. And I do think the concept of employability is a very amorphous thing. It's very hard to measure. But they've done these kinds of tests of young graduates based on industry requirements and found that only about 50 percent of young graduates are employable. And this is not employable, not just in terms of technical skills, but also in terms just speaking and cognition and communicating in English. So even on those dimensions, there seems to be a huge shortcoming. So, but I think there's still a lot more to be understood because I think it's also a slippery slope to think about employability, you know, because it also then just becomes an individual problem that students have to work on themselves and improve cognitive skills and so I think it has to be taken partly with a pinch of salt, but I do think that over the last few years especially with increased privatization of education, you have education being, you know…it's seen more as a profit oriented business. There has been some compromise on learning outcomes which needs to be looked into.

[…]

Milan Vaishnav So I want to ask you specifically about the case of young men, because this was, I think, one of the most interesting data points in your report, which is you find that the share of young man in education has fallen from around 38% in 2017 to 34% by the end of 2024. And I think later you say that the share of young men in education is the lowest it's been since 2017. So I want to ask you, you know, to what extent is this a kind of distress signal about household finances versus changing labor market incentives and how does the state of young men's employment, or excuse me, education compared to young women?

Rosa Abraham Yeah, so I'll answer the second part first, which is the status of young men's education compared to young women. So since 1983, you've seen in the last 40 years, a huge increase in tertiary enrollment rates for men. For women also, you see that although the catch-up comes much later, so they catch up eventually. And there are different estimates, but broadly the story is that tertiary enrollment rates, so young men in colleges and young women in colleges, are now almost similar. There's some studies, if you look at certain data sets, it will indicate that women have actually surpassed men's educational enrollment. But that's the broad story over the last 40 years where, you know, you've seen this huge convergence and it's really quite, I think it's a remarkable feat that we've gotten this to where it is. If you look at where we are internationally, given our level of per capita income, our tertiary enrollment rates, both for men and women, are on par with where we should be. So we are having levels of tertiary enrolment as you would expect given our per capita. That's where we've stood up until now. Now, when I started looking at the educational enrollment numbers and that's when I saw this from 2017, you see this kind of shift where consistently there's been an increase in education, but from 2017 onwards, the share of young men, like you said, starts to fall. My first instinct was to then say, well, maybe this is a COVID impact that you've seen men maybe withdrawn from education because of COVID and then that of persisted. But this is something that precedes COVID. It has started from 2017. COVID doesn't make a huge dent in this. It just continues even after COVID. Now, it's interesting because, well, if they're withdrawing from education, then what is happening? A lot more what is happening is that they're almost entirely going into employment. Now contrast this with what's happening with women. With women, in the same period, you don't see a decline in education. Instead what you are seeing is actually an increase in the share of women in employment. Now, just to give some context, the share of women and employment in India is, historically it's been very low. So the last few years has seen a reversal in this trend. So there's an increase in employment for both young men and young women, but the story is that the young men increase in employment is coming from them withdrawing from education and whereas for young women it's a withdrawal of women who were previously out of the labor force are now entering the employment scenario. Now, again, when I first saw this falling share of men in education, my explanation was that perhaps you overlay with the fact that in the last few years, we've also seen a stagnation in graduate earnings. So, especially entry-level salaries for young graduate men has actually fallen in the last 10 years. So I thought, okay, maybe this is a response to that, you’re seeing where it doesn't seem to be doing you any good to be investing in education, so let's just enter the labor market straight away. But actually, what the survey allows you to do is that it asks a question that if you're a young person, if you are below 29 years of age and you're not enrolled in education, it actually asks these young individuals why are you not studying and it gives a set of reasons. It could be that you felt that education is not necessary. The other reason that they say is that you needed to support household incomes or the school or college was too far. If you look at what are the most commonly cited reasons, in 2017, about 58% of these young men who had withdrawn from education, they said that they needed to support household incomes and that's why they withdrew from college. Whereas by 2023, which is a period when you see this huge withdrawal from education, the share who say that, well, we need to support household incomes and that's why we withdrawn, that number has increased to about 72 percent. And so, that's where I felt well maybe this is not a tempering of expectations from graduate education. This could be potentially a sign of distress. I do think that it's still too early to see this kind of withdrawal because even though graduate salaries have mellowed over time, they still earn about twice what a non-graduate would earn. There's of course a lot of social prestige attached to graduate education. It has a lot of sway in the marriage market as well. So I think it's too early to think that there's this disillusionment with graduate education; I don't think we’re there yet.

Milan Vaishnav So can I just ask you about this, the kind of graduate unemployment woes? I thought this was something again, that was quite surprising. You know, unemployment tends to be higher at higher levels of education and lower amongst older cohorts. And so, I think you say in the report that 40% of graduates in the labor force report open unemployment. And if you kind of look over, say the last couple of decades, between 2004 and 2023. There were five million graduates who were added every year. And during that same period, the number of employed graduates rose by only around 2.8 million. You know, that's a huge gap between the two. You know again, if you could help us unpack, you know, to what extent is this a demand problem? Is this a skills mismatch? Is there something else going on? How do you make sense of that chasm between these two numbers?

Rosa Abraham Yeah. So, I want to first spend a little time on this graduate unemployment number. So the numbers say is that it's for someone below the age of between 25 to 29 or below the age of 25 who's a graduate, the unemployment rate is around 40%. Now, I do want to say that this is not something new. This is something that has been a structural feature of the Indian labor market. If you go back to where we have, you know, the same data available, which is the earliest we have is 1983, the number is around 35%. We are now at 40%. Now, I also went and tried to look at some more earlier research and I stumbled upon a book by Mark Blore and others, which was written in 1969, and the title of the book is called The Causes of Graduate Unemployment in India. And he talks a lot about even pre-independence, the number of commissions that were set up to understand why graduate unemployment is so high in India, So, I think that this is not something new, it's also not something completely unusual to India. So, if you look at graduate unemployment rates across the world, for youth, unemployment rates will typically be about two times more than adult unemployment rates. And that's because this is a period when you're transitioning, this is the period that maybe you have certain aspirations on what you want to do and you're still looking for that perfect job and so you choose to stay out. So that's the one, that's one thing that, you know, it's not completely new and it's not completely unexpected, it's completely unusual to India. But I do think that the problem is what I mentioned earlier on, which is that if you look at this number over as people age, as they grow older, you'll find that this unemployment rate starts to fall, which means that there is eventually a match happening. But the problem now, and that's where, you know, the number that you cited, which is that you have 5 million graduates entering and then off that only 2.8 [million] get any kind of employment. But the most striking is that the number who gets salaried employment is even lower, it's about 1.7 [million]. So I think that's where, that there's a long wait and ideally you would want that long wait to be met with kind of employment that you aspire for, but that match isn't happening. And that's where I come back to the thing that there is a demand side problem here, which is that we have just not created the kind of aspirational jobs. But I also think that, you know, there's something that we often miss when we are talking about this is the skills mismatch question and the role of caste in the equation right. So, what also happens is that when you sign up for a graduate education, then there's an aspiration around what kind of job that you want to do. Given the hierarchy of caste and the kind of disparagement that manual labor has because of this association with marginalized caste, there is this understanding that graduate education is your pathway out of manual work. It's your entry into your test-based white-collar jobs. And because of that, there's a resistance to really take up highly labor-intensive manual kind of work that may be available. So that's also kind of exacerbated the graduate unemployment problem. Alongside that, and one more layer that comes in, is that more and more households, it looks to me, are able to afford the wait. So, if you look at the, you know, pool. Who's able to wait? Between 1983 and 2023, what we found is that the graduate unemployment has become more democratic. So even poorer households are seeing a larger share of their graduates staying out of the labor market. On the one hand, yes, the transition is hard, but now the transition is also a lot more affordable. So, you will see that a lot more households are having graduates that choose to stay out for some time. So, you know, all of this compounded with the fact that you have a large youth population that's growing. You have a large number of graduates. So the size of the pie itself has increased. The incidence of graduate unemployment has remained the same. But it's just that the scale of the problem has just become much, much more.

Milan Vaishnav I mean, I'm sure it's always very interesting, right, whenever you're researching something and then you realize that somebody's written a book about this in 1969, like, actually, this is not a new phenomenon, even though it often gets cast that way in the kind of media discourse.

Rosa Abraham It was interesting because I also listened to a lecture where somebody had quoted a paper by Cain Raj in his Cairo lectures, which I think was in 1980s, where he also sets up this framework where he talks that we have this idea of disguised unemployment, that a lot of workers are really just in very low productivity work and they are actually unemployed but not really unemployed in the open sense that we understand it. And then what he says is that as countries, as economies develop, a lot of this disguised unemployment will manifest then as open unemployment because incomes would increase and you don't need to maintain this labor force in low productivity work and they might actually report open unemployment and potentially that's what we're also seeing. So it's also as I'm reading more and going into more of the historical debates around it, it's really fascinating that some of these things have not really changed. In the last, I don't know, 70, 80 years.

Milan Vaishnav So, you know, another thing that I wanted to kind of talk to you about is this question of women, and there's a bunch of stuff to talk about. But one of the things that came out in this report is that manufacturing has emerged as a major employer for young women, particularly in recent years. And you know I think when you listen to what a lot of economists say about the about the kind of Indian economy, about what some people in government say. There is this kind of yearning, I think, to replicate the success that many of the East Asian tigers had, whereby they were able to get large numbers of women onto the factory floor and find productive, gainful employment for them. I'm wondering, when you talk about women in manufacturing, to what extent are we talking about that kind of transition? Or are we looking at a fundamentally different kind of quote unquote manufacturing model setting up in India?

Rosa Abraham Yeah, I think where we are in India is we're very, very far away from the East Asian type of manufacturing model, especially when it comes to women's implants. What we've seen is that younger women are more and more likely to be in manufacturing industries, so leather, textiles and apparel, manufacturing of motor vehicles, electrical components. These have become major hirers. But if you really look at the numbers, that's still only about 6% of the non-agricultural workforce, women workforce that it hires, So, we are very far away when you think of the scale of employment and specifically women's employment that manufacturing is currently catering to. And I do think, and this is where we've kind of fallen short in terms of thinking around a cohesive industrial policy strategy. So I think that where we do see success, especially where we see success in terms of manufacturing and women's employment, has been this kind of phone assembling, for example, and manufacturing of motor vehicles as well. And these are also sectors that have seen some kind of favorable policy and some kind of cohesive thinking around promoting these kinds of enterprises. So, from the women's perspective, and a lot of this is really states like Tamil Nadu and Karnataka that has kind of led the ground in terms of employing more women. And it's come from setting up hostels for young people, making those work jobs far more accessible. We are not, and this is just, I think, right now, we are sort of outliers in that sense. We're nowhere near the scale that we have, that we should have given that, you know, we still have a largely unutilized female workforce and I think we need to be thinking about that and need to think about this more in terms of the labor-intensive industries as well, which is still not happening yet.

Milan Vaishnav You know, one of the things which you talk about in the report as a kind of safety valve, I guess, or a shock absorber in a way, because of this gap between aspirations and what's actually available on the job market, is this question of migration, right, and specifically this question internal migration, which is, as you put it, one channel through which the demographic dividend gets spatially rebalanced, because what happens is you have younger states that are exporting all of the surplus labor to older, more economically dynamic states. So, you know, this is a kind of traditional, you have migrants going from UP, Bihar, places like that to, say, Gujarat, Maharashtra, so on and so forth. But what I found interesting is that migrants from different origin states are moving through very different, distinct migration channels. Tell us a little bit about the heterogeneity in what those channels kind of look like in practice.

Rosa Abraham Yeah, yeah, this was also, you know, completely new for us when we started looking at this. And this comes from the Eshram Registry of Informal Workers. So I will say that there's caveats around, you know, it's not a nationally representative study. But I think it does still give very useful insights. Now, there's different patterns of migration, like you said. So one thing that we see is there are migrants from Bihar, Jharkhand, who have this very wide distribution across the country. So, if you look at the share of those migrants in the total migrant population in districts, you see that they are spread all across. Workers from Bihar moving up north, but also to the south as well. And this is just partly, there's this huge volume that is moving, but also that there is clear distress and they're not necessarily moving only to industrial clusters, they're moving across the country. The other kind of migration that we saw was this kind of very self-contained migration. And this was in the case of, we saw this from Adi [sic] Pradesh, but also in the case of Assam where you have these informal migrants and most of these are young workers who kind of stay within the boundaries of the state. They move to the other state, but in and around the boundaries and they kind of cluster around that. Similar to that is again migration, for example, where we see in the case of West Bengal, where a lot of the migration is around the eastern corridor, but also you have a high influx of migrants into Kerala. So this points towards, you know, there's some pre-existing social networks, there are some preexisting kinds of channels that exist and so migration has kind built on those channels. Now, I think it's interesting also because, you know, when you think and how you, you kind of set it up, which is that you have states which are at very different stages of demographic transition. So, states like Kerala, where the average age of the population is around 40, and you have states like Bihar and UP, where average age the population's 20-25. So, you would think that okay, it's a migration is kind of evening out these disparities, but also evening out economic disparities. And ideally, you should see some kind of, one would like to see some kind of policy framework that emerges that supports this migration and kind of makes this more mainstream and not something that is mediated around channels of social networks, channels of distress, and that unfortunately has not happened, and it's ironic because we see a lot of this kind of migration support emerging for sending our young workers into Gulf countries into Israel, you know, so this kind of mediation of migration to international networks. To international regions, but we don't see that happening across states. And it seems like the most obvious thing to do, that why don't we really address two things with this one kind of policy in a sense. And I think that's where it's important to recognize that these channels are not necessarily only in the direction of economic opportunities, but it's also hugely mediated by pre-existing networks, which may or may not be a good thing. But can we think of these as ways to address both the fact that certain states are not economically developed, but also certain states have a large share of the youth population? And at the same time, there are states which are old, aging, and now need a young workforce. So, can we make this work across states?

Milan Vaishnav I mean, you know, one of those networks that connects migrants and lots of people have written about this is caste, right? And people tend to use caste networks as a way of bridging between their home state and their eventual destination or their place of work. One of the, I think, kind of positive things in this report is that you note that at least going back to the 1980s or so there has been a substantial kind of weakening or loosening of caste-based occupational segregation, right? Because we know that one person's entry into the labor market is going to be shaped by a whole variety of things. There's education, there's opportunity, but social identity is also another factor. How significant do you think the shift is in terms of breaking the bond between one's caste identity and one's kind of, you know, daily labor, their daily wage.

Rosa Abraham Yeah, so I think, I think this and it's not just our report, there's also been other studies which have looked at, you know, father-son mobility, what did fathers do occupationally and our sons doing the same thing, which also points towards…there's this kind of intergenerational mobility that's happened. So, for example, we are finding that you don't find young schedule caste and schedule tribe workers in their caste-associated industries. So typically these industries would be waste, leather, mining, and you're finding that the younger generations of SCST workers are less and less likely to be using this as entry points into the labor market, which is good to see. But I think here's where there is some kind of caution around, you know, there's still a lot more to be done. And there's two things that I'll talk about. One is that another industry that was also a major employer of marginalized caste was the government sector. So public administration and defense services was also an industry where you see a higher share of these caste groups compared to their share in the total workforce. So, they were sort of over-represented. But you're seeing the same trend in this industry as well, which is that the younger generation of SCSTs are less and less likely to be using this as an entry point. And that points towards the shrinking role of the public sector. They're just not hiring at all. And because the public sector employment itself has contracted, one of the groups that has taken possibly the hardest hit from this contraction is marginalized or caste groups. So, affirmative action reservation policy in public sector was a good way to allow for intergenerational mobility, but that role itself is shrinking over time. So that's one kind of flag here when you think when you're talking about mobility across generations. The other thing that you know—and this doesn't come necessarily from the employment outcomes—when we were looking at education and the choice of streams of education. So what we were looking at as well, where do graduates come from in terms of the household income quartiles? So, we can divide households into the poorest 25 percent and the richest 25 percent. You have quartile one and quartile four and between that quartile three and quartile two. Now you'll find that most graduates, 50 percent of graduates, came from the richest households and these are also likely to be results from the non-marginalized or general category households and only about 7% of graduates came from the poorest households. So that's one that your access to higher education is still being mediated and that number has shifted somewhat but it's still highly unequal. So access to higher education still being meditated by income and also thereby possibly caste as well. Now, the other thing as well, once you are in graduate education, what kind of graduate education do you sign up for? And you hardly see lower income households, which are most likely your SCST households, signing up for engineering and medicine. And this, as you know, is in India is, you know, the occupations that’s associated with highest labor market returns, highest labor market mobility, and you're not seeing that. So, I think you might see that well, your young SCST worker is less and less likely to be in the leather industry or waste management. They might be entering into health or education, but they may not be entering as a teacher, they may not be entering as a doctor, they maybe entering still in lower rungs in terms of the occupational hierarchy. So even though industries of entry have become a little less caste-oriented, it's likely that their point of entry into these industries is still being stratified by their caste identity, partly because of the kind of education that they've also had access to.

Milan Vaishnav So Rosa, let me maybe just ask you one final thing, which is a bit unfair because it wasn't an explicit topic in the report, but I think it's something that's on everyone's mind. I mean, in India, outside, really everywhere, which is this question of automation, AI, and what it's going to do to labor markets, right? And wherever you are, there's just a ton of anxiety about this. Based on what you've seen in your report, to what extent do you think this is a near-term threat, or do you that India's labor market challenges are still kind of primarily structural, they're pre-existing, and that this may be a problem in the years to come, but it's not necessarily a problem of the kind of here and now, because India's needs are much more basic in some way than that.

Rosa Abraham Yeah, I think we are towards that, which is that I think AI is not our most immediate threat. I think, we still don't really know what it is that AI is going to displace. There is this understanding where AI is going to displace entry-level jobs and possibly of young workers then. I think where we are in terms of the Indian labor market is that, you know, we are not even thinking about the entry-level jobs that the kind of displacement that AI has is probably not sectors that are major employers in India. So, if you look at say computer and information services or even say high-end manufacturing services, they employ less than 10-15 percent of our workers. If there is some displacement, sure, there's going to be displacement there. But I do think the challenge for us really is not going to come from AI-related displacement. I think it's really about the fact that we just are not creating, and I feel like I sound like a stuck record here, but I think the thing is that we've got this supposedly high-skilled labor force, or at least highly educated labor force. And we are not able to even engage them in basic factory floor level jobs, for various reasons, both demand and supply side. Can AI create huge further disruption in that spectrum and for that group of workers? Not immediately. And I think it is warranted to be alarmed around AI, but I don't think it's a near immediate term threat. I think the immediate term challenge is really around the high graduate unemployment, at least the high on graduate unemployment numbers, not as much the rate, but also how do we make sure that we are able to engage this workforce in high earnings and secure jobs, which we just haven't been able to do.

Milan Vaishnav My guest on the show this week is Rosa Abraham. She heads the Center for Sustainable Employment at Azim Premji University. She's also the lead author of the State of Working India 2026 Report. It is a gold mine for people who are interested in data and statistics on education, on the education to employment transition, where India's labor market is heading. Rosa, thank you so much for taking the time. Congrats on the report and thanks for coming on the show.

Rosa Abraham Thank you so much.