‘For 1st Time In 4 Decades, Young Men Are Withdrawing From Education’

A new report finds that young Indians are more educated than ever, that gender- and caste-based gaps have narrowed, and that young Scheduled Caste (SC) and Scheduled Tribe (ST) workers are breaking away from occupations traditionally associated with their communities.

But it also uncovers troubling signs: young men withdrawing from education, graduate earnings stagnating, and most new jobs being created in agriculture, a sector that contributes least to output and earnings.

Rosa Abraham is associate professor of economics at Azim Premji University and lead author of the State of Working India 2026 report, which traces four decades of data on youth employment and education.

Abraham heads the Centre for Sustainable Employment at the University. Her research focusses on informal work and women’s employment with particular interest in issues at the intersection of labour statistics and women’s work.

In an email interview, Abraham discusses what the data reveal about India's labour market, the structural barriers that persist, and what it means for a country that has less than a decade to make the most of its demographic window.

Edited excerpts:

What were the three findings that surprised you the most, and why?

The first was the closing of the gender gap in entry-level salaried earnings for graduates. By 2023, a 20- to 29-year-old graduate woman was earning as much as her male peers. When we look closely at these trends, there is both cause for celebration as well as for worry.

Between 2017 and 2023, young women’s earnings increased at the rate of 1% per annum (after accounting for inflation). On the other hand, young men’s earnings have slowed down significantly, and in fact, between 2017 and 2023, average annual growth rate was -0.1% in real terms. Therefore the gender gap closing has come from a combination of both these two trends. Graduate women have seen an improvement in their labour market opportunities but not graduate men.

The second surprising trend was the withdrawal of young men from education. This is the first time this is happening in the last 40 years that we have data for. Between 2017 and 2023, the share of young men in education has fallen from 38% to 34%. And, this is not just a poor-household phenomenon. It was seen across all households and more so among the upper-middle income households.

Finally, the large increases in Scheduled Caste (SC) and Scheduled Tribes (ST) youth in ‘modern’ industries has been heartening to see. While other studies have also pointed towards this intergenerational mobility with SC/ST sons far less likely to be in the industries that their fathers were employed in.

Let’s delve into some of these. What explains the real decline in young men's earnings between 2017 and 2023?

There were two possible explanations for this recent decline in graduate men’s earnings. One explanation was that perhaps, entry-level industries have changed for the newer cohort of young graduates. They are entering into industries that don’t pay very well. The other explanation was that industrial composition itself had not changed over time, but rather, the premiums on graduate education have gotten compressed across all industries.

We use a simple econometric technique to try and tease out which of these better explained these recent trends. What we found was that the fall in earnings did not come from a change in the composition of entry-level industries. Rather, average wage premiums to graduate education across all industries seem to have dampened, resulting in the fall in earnings.

Part of this could be a result of the fact that demand for graduates has not kept pace with the supply of graduates. Industry has created too few jobs, and the consequent ‘surplus’ of graduates have resulted in stagnation of earnings.

Secondly, if young men’s withdrawal from education is happening more in upper-middle income households where financial need is less pressing, what else might explain it?

Across all household quartiles (the poorest, and the richest), the share of young men who cite budget constraints as a reason for withdrawal from education has increased between 2017 and 2023. While the poorest household saw one of the largest such increases, it was interesting to see that the third quartile of households also saw a similarly sharp increase. This was also a strange finding for us.

One possibility is that the stagnation of salaried earnings has possibly affected the middle/upper-middle income given that the former are households that are more likely to rely on salaried incomes. Consequently, the salary squeeze may be felt acutely among them. This could explain the increase even among the upper-middle income households.

About half of the 83 million jobs added between 2021-22 and 2023-24 were in agriculture, your report finds. Taken together with the increase in women’s engagement in own-account self-employment, what does this say about the Indian economy, which is now among the world’s fastest-growing major economies?

This points towards the clear decoupling of growth from employment. India’s growth has been associated with ‘jobless’ growth. Perhaps, this ‘jobless’ aspect may have mitigated somewhat given that employment creation, particularly for young women, has picked up in the last half-decade. But if you look closely at where this employment has been created, most of it has been in agriculture—a sector that contributes least to output and investment, and a sector associated with lowest earnings. Therefore, while growth may not necessarily be jobless, it has certainly not created the kinds of jobs that a young, highly educated workforce would want.

What we have seen is a structural regression—workers, especially women, returning to agriculture. So on the one hand, we have a high growth rate, but on the other hand, we have this structural regression. It suggests that growth has not necessarily been very equitable.

The share of young men withdrawing from education citing household income needs has jumped from 58% to 72% in six years, you found. Given gender and caste differentials of access, what does this mean for India’s tertiary education?

Over the last few years, gender-based gaps in tertiary enrolment rates (TER) have more or less closed, which is a phenomenal achievement. Caste-based differences have also narrowed somewhat, although TER among STs still remains relatively low. Against an overall TER of 28, ST enrolment rate is still around 21, as of 2022. Of course, it is an increase from 8, which is what it was a decade ago, but there is still a gap that needs to be closed.

The fact that young men are withdrawing could mean that women’s TER may now exceed men’s—again, a kind of perverse convergence, similar to what we see in the case of graduate earnings.

And this withdrawal, to support household incomes, has not been restricted to only poor households; rather it is seen across all income groups of households. This means that it is likely to affect SC/ST men as much as it does other caste group men too. This is the first time we’re seeing this kind of reversal in TER for men.

While the percentage of unemployment among graduate Indians is more or less in line with the past, you found, a higher population and enrolment are leading to more absolute numbers of unemployed graduates. But graduate salaries are twice that of non-graduates, and the difference widens over their lifetimes. What does this say about the structure of India’s economy?

Graduate education still carries with it a high premium. For that alone, investing in a graduate education is worth its while, leaving aside the social prestige that comes with it.

In the Report, we try to estimate the returns on investment in education—if someone were to enroll in education, that entails some foregone income in terms of lost years working. At the same time, graduate education subsequently allows for entry into better, higher paying, usually salaried, employment.

Taking these two counteracting trends into account, we estimate the net returns on investment in education and find these returns to be very high. This explains the continued investment in education, aside from the other social factors that push for graduate education.

However, the pathway into good jobs, after graduation, has become more tenuous in the last few years. In a separate study, with my co-author Surbhi Kesar, we find that there is no ladder into formal/salaried employment from informal employment. The only way to secure permanent salaried employment is to enter into the labour market as a salaried worker, i.e. your entry level job.

And as a result, you’ll find graduates waiting longer to get their ‘aspirational’ employment, which with age, mellows and perhaps recalibrates, and they eventually enter, but not necessarily into their aspired job.

This is the result of constraints in job creation alongside a highly dual economy—a salaried, formal, but small segment coexisting alongside a large and highly informal segment.

A related finding is that the overall capacity to spend time searching for a job has increased, with existing divides showing up on gender and caste lines. Fifty percent of young graduates find employment within a year, but salaried employment is rarer. Young men either find employment within a year or do not find, even up to three years. And sons of less educated fathers find any employment earlier, but are less likely to find permanent salaried jobs. Do you see this affecting choices at the tertiary level?

The mechanical expectation is that a slowdown in graduate earnings and the difficulty in acquiring salaried employment may result in households investing less and less in tertiary education. But, the reality is that firstly, graduates still earn double what non-graduates do, and as I mentioned their returns on investment in education is very high.

And secondly, households do not invest in education solely for its labour market value. There is social prestige too, not just for men but for women too. Education has an important role to play in the marriage market too.

For young girls in India, education has become a vehicle for upward social mobility not through educational homogamy but rather through educational hypogamy—women marrying men who are less educated than them—but belong to a socially/economically higher group. Therefore, it is too early to see these dampening trends affecting tertiary enrolment rates.

Younger cohorts are less likely to be concentrated in traditional occupations associated with their social background, you found. What does this say about education or skilling being social levelers?

The Report finds that younger generations of SC/STs are less and less likely to be in industries that are traditionally associated with these caste groups—leather, waste, mining etc, compared to the youth in the older generations. This points towards a weakening of caste-based occupational segregation.

In the Report, we have not looked at how these patterns are different for graduates versus non-graduates. Separate studies have established that education has enabled a faster mobility away from these traditional occupations. Therefore, it is likely that higher education has indeed facilitated this intergenerational mobility. What was also good to see was what the new employers were.

The last 40 years have seen a sharp rise especially in manufacturing of motor vehicles, paper manufacturing, and in the modern, higher earnings services industry such as business support and financial services.

When SC/ST youth enter manufacturing or business support services, are they entering as salaried employees with some security, or predominantly as contract and casual workers? Has the nature of employment—not just the sector—also improved?

In the report, we do not look specifically at the youth cohort of SC/ST within each industry to see how their pattern of employment has changed. The reason we avoid this is that the datasets that we work with (the national labour force surveys), being a sample survey, do not have a large enough sample for us to make a concrete/definitive analysis. The SC/ST youth sample within specific industries when disaggregated by employment arrangements become a very small sample, in terms of absolute numbers, and it would not be reasonable to infer from such a thin sample.

However, in the State of Working India 2023, we do find substantial intergenerational mobility when comparing employment arrangements of father-son pairs. This is seen across all groups including the SC/STs, although less so for SC/STs. So, sons of fathers who were in casual wage work were less likely to be in casual work themselves. Extending this finding, one can infer that the industrial mobility that we see may also be accompanied by mobility in terms of employment types too.

In many cases, employers now increasingly rely on independent testing of credentials rather than relying on degree certifications, you write. What does that mean going forward for India’s education systems?

We have cited papers by Fuller (2022) and Copestake (2023) that many employers are moving away from using degree certifications alone. We do not have studies that concretely establish this but hiring patterns among employers do point towards independent testing becoming an important screening device, over and above degree credentials.

There are two ways this can go. Either our existing educational curriculums are upgraded to better address industry needs, or students will increasingly have to invest in other kinds of certifications beyond the degree to ‘signal’ their quality.

The report mentions the potential fall in numbers of entry-level jobs due to AI. How do you see this affecting employment across groups?

It is too early to say what the AI impact is going to be. Will it substitute for workers or will it complement workers? It is likely that AI will affect entry level jobs, especially tasks that are repetitive and non-cognitive. To the extent that these are jobs most occupied by young workers, it is likely that they are going to be most immediately impacted. We are already seeing a slowdown in entry level jobs, according to different sources.

Given that you have studied 40 years of data, are there historical parallels to such slowdowns, where technological change disrupted entry-level absorption, that offer any guidance?

Unlike earlier disruptions, I think the difference with AI is we don’t really have an idea of what the kind of disruption is going to be. For instance, when computers were introduced, or say, digital booking platforms for cabs, there was an idea of what kind of work was getting displaced, even if the extent of displacement itself was unknown. With AI and the labour market, I think the big difference is we don’t really have any idea how AI is going to progress, what’s going to take off, what is going to have a wider scaled-up expansion.

AI can substitute for certain tasks in which it can displace labour; at the same time it can also augment workers in their tasks, in which case it can increase productivity. What we do know comes from initial trends and not all in the Indian context. This is what makes it difficult to really prepare for the kind of disruption that might unfold in the coming years.

We welcome feedback. Please write to respond@indiaspend.org. We reserve the right to edit responses for language and grammar.