Steven Brownstone on Agricultural Subsidies, Mechanization, and Historical Land and Labor Institutions in India

Brownstone and Rajagopalan explore the interplay between technology adoption and rural labor markets.

SHRUTI RAJAGOPALAN: Welcome to Ideas of India, a podcast where we examine the academic ideas that can propel India forward. My name is Shruti Rajagopalan and this is the 2024 job market series where I speak with young scholars entering the academic job market about the latest research on India. 

I spoke with Steven Brownstone, a PhD candidate in economics at the University of California, San Diego. His research focus is on the fields of development economics, agricultural economics, and political economy. We discussed his job market paper, Labor Market Effects of Agricultural Mechanization: Experimental Evidence from India. We talked about the reason there isn't a natural mechanization in rice plantation in Telangana, the role of the state in the uptake of mechanization, the labor market in a developing country that is undergoing a structural transformation and much more. 

For a full transcript of this conversation, including helpful links of all the references mentioned, click the link in the show notes or visit mercatus.org/podcasts.

Hi, Steven. Welcome to the show. It’s a pleasure to have you here.

STEVEN BROWNSTONE: Hi. It’s a pleasure to be on the show.

The Wage and Labor Effects of Mechanized Drum Seeders in Telangana 

RAJAGOPALAN: You’re doing something quite interesting with your job market paper. Your intervention is to provide farmers mechanized drum seeders; basically, it’s capital substituting for labor. You also give them training to increase the uptake of the capital, arrange for rentals of these drum seeders, and so on. 

You find that this intervention, as one would expect, tripled the uptake and replaced manual rice transplantation, which is a very labor-intensive, tough job. Consequently, again, as one would expect, the wage bill decreases for the top farmer, or the entrepreneur in this case, both in terms of labor days of who they are hiring out, also in terms of the wages that they need to pay, and obviously, profits thereby increase.

The second part of what you find is also very interesting, that the labor displaced as a consequence of this mechanization that you do through the intervention doesn’t fully get absorbed in other parts of the labor market outside of rice transplantation in agriculture or outside of agriculture. 

Is this a good summary of what you’ve tried to do?

BROWNSTONE: Yes, I think that’s a quite good summary of what I’ve tried to do. Maybe I’ll steal some of it for my introduction. I think one thing to clarify is that this was a government of Telangana intervention. How I came to this project actually was not as an econ PhD student knowing a lot about drum seeders and thinking they were interesting, but I was in Telangana thinking about state capacity and was working with the agriculture department to see how they could improve extension. 

Going around in the villages, everyone was talking about this rice labor shortage. The number one issue for panchayat leaders was that there weren’t enough transplanters to transplant the rice, and the wages had gone up to these unreasonable levels in their eyes. There was this clamor for something to do about this. 

The government was actually promoting these drum seeders. Even in the control villages in this program, there was a government program to promote, through extension, drum seeders. What the experiment really identifies is what happens when you add these capital rental markets on top of this government extension push, because as a grad student, I wouldn’t be able to create a complicated extension system that the government of Telangana has invested in. Drum seeders are a relatively cheap technology. We’re able to create these rental markets in a few places.

RAJAGOPALAN: What is the mechanism at work? Is this just standard structural transformation, and you’re trying to see the labor-capital substitution, or is there something more going on here that’s not obvious to an economist who doesn’t study agriculture and rice plantation in Telangana?

BROWNSTONE: I think broadly, yes, this is the standard. In the presentation, I have a labor supply curve and a labor demand curve, and the labor demand curve shifts. The broad story is the classical economic story. The nuance which you get into is the nature of the transplanting labor market being this very congested market within maybe a broader story.

The Indian agriculture labor market in general, as many studies have shown, is defined by a lot of excess labor. There’s a lot more people that work in agriculture in India than probably need to produce the food that is produced. On the flip side, there’s many different tasks within agricultural production and within different crops. 

One particular feature is the timeliness of certain tasks. Particularly, planting needs to happen at a very specific time during the year so that farmers can get their crops in the ground. Rice transplanting is a type of planting that’s actually the most labor-intensive.

It’s backbreaking, terrible work. This is a task where you get these moments of congestion where the farmers want more labor than is available in the village. If you look at a histogram of rice transplanting wages, there’s this mass of peak congested wages, and it’s that mass that the treatment removes. It’s actually not changing the average wage, but it’s removing this congestion in the labor force. 

Maybe one way of reframing is not wages going down, but wages failing to rise. That probably explains why there was this political support, to begin with, in this project.

Wages Failing to Rise: NREGA and a Profitability Ceiling

RAJAGOPALAN: Why are wages failing to rise? What’s going on in that political economy that the farmers refuse to throw some money at the problem and hire daily wage labor or whoever can be attracted? Is it because NREGA is pulling them in another direction? What is going on in that market?

BROWNSTONE: Actually, another interesting aspect about the context of this study, is that there is very little NREGA during these peak seasons. Not to go way ahead, but one of the results that I do find is that there’s more unmet demand for NREGA. There’s not more NREGA actually happening because really, NREGA isn’t happening during the peak season.

I think part of the issue was just profitability. At a certain point, rice is not a particularly profitable crop. There’s a certain kind of ceiling where if you pay above that ceiling, it just doesn’t make sense economically, given where rice prices are. I think that’s probably a big reason why they’re not just letting themselves get bid up to higher levels. To some degree, they are. That’s why, in the treatment villages, there are these higher wages. That’s why there is this kind of clamor for paying people. They are paying the higher wages, they’re just not happy about it, and they’re looking for mechanization. 

I think that’s been the story around the world. If you look at tractors in the United States, it’s the same story of wages have to go up first and then you get the mechanization. I think the thing that’s different about this context, and maybe about women in rural India that don’t have many outside options, is that wages go up, which inspires the mechanization, but then because there’s nowhere for the labor to go, the mechanization can drive the wages back down again. I think that dynamic is different than we see in other contexts and really matters for the pace of structural transformation overall.

Women’s Changing Role in the Farming Labor Market 

RAJAGOPALAN: One follow-up on that. Typically, this backbreaking work is done by women. The women in the household who are mostly working in the household, when it comes to rice transplanting season, they come out of the household, and they really chip in. In fact, it’s mostly a woman’s job this part of the season. 

Why isn’t that pool available, given that female labor force participation in India is so low and, in many places, declining? I would have thought that would be a ready-made pool of labor.

BROWNSTONE: I think part of it, and this is part of the story of why these wages are going up, is that in general, Telangana is getting much richer. This is the downward sloping part of the female labor supply curve that households are getting richer, and women can afford to no longer do this work.

Another result that I have is the overall treatment effect on women in these villages reporting that their men will let them work outside the village is actually negative. That’s probably driven by the farmer profit results, that farm profits are going up, not just among people that adopted drum seeders, but among people that didn’t adopt drum seeders. I think a lot of it is just driven by this area getting richer, and as you get richer, you can afford not to do this backbreaking work. It’s not just ‘job versus other job’. It’s ‘job versus no job at all.’

The Puzzle of Mechanization: Is Government Intervention Necessary?

RAJAGOPALAN: This is the point of your paper where I started seriously getting puzzled, because, until this point, it made perfect sense. I was like, “This is a classic story of structural transformation and economic development, wages are going up, which means you can’t quite find labor, which means you need to substitute labor with the capital and some kind of mechanization.”

However, at this point, why isn’t the mechanization automatic? Why does there need to be government subsidy, government intervention, government training? This has happened everywhere else in the world in agriculture. One would think it’s quite obvious that this would also happen in India. Why doesn’t it just happen without an intervention? I guess what I’m asking is, what is the market failure here that the government needs to intervene and provide a subsidy?

BROWNSTONE: I think the market failure is this capital-skill complementarity. I was also puzzled. This was my initial puzzle, “Why is there this?” I think it’s the first question that anyone tells you to ask if you’re designing an experiment. “Why hasn’t this happened already?”

I think the reason is that the farmers weren’t interested in being trained on something they didn’t have access to, but at the same time, it didn’t make sense to rent out a device that farmers didn’t know about. There’s this issue where you need to combine this training on growing rice through drum seeding—direct seeding is a completely different process than growing rice through transplanting. It’s not like a tractor that you can just slot in, and this is the harvest. The tractor does the harvest. Nothing else has to change. You can grow your rice exactly the same.

When you switch from transplanting to direct seeding, you have to change up your irrigation, you have to change up your weeding. There’s a lot of things that change along the way. For farmers having that training combined with the access to the device was very important. 

Unlike with some of these seed companies that have these big private sector extension services, it doesn’t really make sense for a drum seeder manufacturer that’s going to sell a device for 5,000, 6,000 rupees because the farmers can share them, as they do in my study. One village maybe needs 20 of them maximum. It doesn’t make sense to invest in that training given the amount of profit that you could get from one individual village. I think this training is the big issue.

RAJAGOPALAN: That ought to be a one-time thing. Is that what the intervention is finally studying? Does one-time training work? Because when I read the paper, it seemed like this is a more systematic subsidy that’s now going to be given for mechanization. Did I get that wrong?

BROWNSTONE: I think the future direction is unclear. I think the idea is it’s not even really a subsidy. The rental price that was set matches what was observed in other villages that have already adopted drum seeding in Telangana, and it’s 250 INR per day. Over the course of even two seasons, the devices would pay for themselves. It is a subsidy in terms of providing access early, but you could theoretically structure this all as a Panchayat using its own Panchayat funds to buy one of these and then making it back through rentals. It’s not a really aggressive monetary subsidy. It’s more about accelerating the pace of these things being adopted.

It is possibly true that over the natural course of more seasons, these devices would be adopted even in these villages. It would just take more time. I don’t think this is a stark case where without government intervention, farmers would never use drum seeders, because that’s not true. There’s other districts, there’s other areas where the drum seeder adoption rates are higher than you see in my study.

I do think this training capital complementarity is really important. I think in these other districts and villages, the farmers are able to train each other or talk to each other to understand how to use this device and so they just buy it naturally. I think leveraging this massive government agricultural training infrastructure that the government’s already invested a lot of money in makes a lot of sense for increasing the uptake of some of these capital goods.

Mechanization or Migrant Labor

RAJAGOPALAN: This is what happens to those of us studying India, just curiouser and curiouser as each step goes along when the government intervenes. 

India’s got this massive unemployment–underemployment problem. Lots of youth labor who are just not able to find jobs. They’re underemployed and lurking in agriculture in rural areas when they can, otherwise they’re just studying for the government exams or something like that—different demographics. 

I understand that Telangana is a place which is getting richer. The young people of Telangana, especially those who are from poorer families, don’t wish to do this backbreaking work. 

Why isn’t there an influx of migrant labor, which does happen in other rural areas? In Punjab, you do see an influx of migrant labor because they do multiple crops in a year. They also grow a rice crop. Why isn’t that an option? Why are we forced to mechanize in a labor-surplus country where the farmers themselves are not that excited about even relatively cheap mechanization? The whole thing doesn’t make sense to me when I zoom out.

BROWNSTONE: First of all, there is some migration to Telangana for this work. Not as much in the districts that we chose for the study, because I think those networks take a while to be developed. They’re just starting to figure out this guy will run a bus to rural Bihar and get you your paddy transplanting workers. 

I also think to some degree, India is running out of this endless labor pool in rural Bihar that will work for very, very low wages. That group now has more options. I think that imported labor is still fairly expensive. It’s cheaper than the local labor but it’s still fairly expensive. 

Part of it is that, yes, India does have a lot of unemployment and underemployment, but it almost makes sense—it’s even true in the US—we want to mechanize the worst jobs first. Paddy transplanting is probably one of the worst jobs out there in terms of marginal product per drudgery to human.

RAJAGOPALAN: The way you phrase that is super interesting. The way you phrased it is, “We want to mechanize the worst jobs first.” The way I would phrase it is, “Through the market process and structural transformation, the worst jobs automatically get mechanized first because that’s where the drudgery-to-benefit ratio is the best.” The fact that it needs an intervention is baffling to me.

Two things could be going on. One is just, for political economy reasons, the government of Telangana can’t exactly load up cheap labor from other poorer states in a bus, bring them during harvest season, and send them back, because that’s not a feasible solution. There’s a lot of NIMBYism. We don’t want people from outside. There’s a lot of linguistic sub-nationalism and those frictions. That’s possibility one.

Possibility two is they don’t know what the hell they’re doing. A lot of money got allotted as a budget to agricultural training and mechanization, and now they have to spend it, and this is the way to spend it. 

If there are all these serious problems, and it is really at crunch time, the worst congestion—the fact that farmers aren’t willing to put in the effort to learn how to use a new process seems a little bit bizarre. Is it a cluster of market frictions but no failure, and this will just ease those separate frictions? Is it state-created market failure? Do you know where I’m going with this? I just don’t understand it.

The Role of Government in Mechanization Adoption

BROWNSTONE: Yes. This has been a puzzle that has puzzled economists for generations. The non-adoption of productive agricultural technologies is something that has baffled development economists even before the ‘19. The money for this project literally came from the Agriculture Technology Adoption Initiative, which is this huge Gates thing to study why farmers don’t adopt productive technologies.

RAJAGOPALAN: Why do we need it? Why is that even a question to be studied? Farmers adopt cell phones when it’s convenient to adopt cell phones. My point is, when it’s convenient to adopt, they’ll adopt it. Why is that not the baseline assumption? Why does the Gates Foundation need to tell us what is the appropriate time to adopt a technology?

BROWNSTONE: No, it’s not the appropriate time for it.

RAJAGOPALAN: Either I’m missing something or they are missing something. One of us is totally off in the way that we look at Indian political economy.

BROWNSTONE: There are real training and knowledge gaps that really do matter. Farming is a very complicated process that has a lot of different complementary inputs. I agree that, yes, in the long run, farmers have cell phones now, but they’re bombarded with information. There’s a lot of information that’s out there.

I do think that there’s some role for the government to play in terms of helping farmers understand the universe of technologies out there, and also the training that they want—from talking to a lot of farmers as part of this project—is relatively hands-on. They want someone that’s going to go out there and see the field. If something comes up, they need to know that this person is in their village, they can go and complain to them. Trust is such a huge issue. I think, and maybe that’s also why it’s different than other technology diffusion cases where you just see something on YouTube and you’ll say, "Oh, I’ll try this."

When you choose a farming technology, you don’t see the result of the technology for an entire agricultural season. You’re making this big bet, and there’s always some risk in the back of people’s heads that the entire crop will fail, which is devastating for these rural families. It’s harder for information to spread from trusted messengers in the agricultural sector than maybe other sectors. These information frictions, combined with the capital market frictions, are real.

The other thing that I think is important to note is, for the smallholder farmers in India that have very small plot sizes, it doesn’t make sense for them to buy most of the capital technologies. All the tractors, all the combines, everything is rented.

RAJAGOPALAN: That was my hunch.

BROWNSTONE: It all comes back to land markets. That these farmers are way too small. That’s the argument, that the farms are too small. I looked this up before the call, in India, the average agricultural holding has gone down.

That’s actually decreased over the years. It’s decreasing from 2021. As it’s been decreasing, the penetration of tractors has almost doubled. It’s gone from 30 percent to about 60 percent, which is still too low. 

The point is, because of these active rental markets and tractors, that we actually are able to get some of this capital mechanization that standard economists would think is impossible with such small land holdings because of the rental markets. I think for rental markets to occur, you don’t just need an individual farmer to make a decision. You need coordination. You need a critical mass of farmers in a village to decide that they want to do something, that they want to take up a new technology.

That again points to this idea that you need some outside support to help overcome this coordination problem, or for certain technologies that are obvious, the rental markets will observe themselves. I do think there’s an argument to be made that governments can improve the pace of mechanization, can improve the pace of structural transformation. I do have hope that India, even if the government does nothing, will eventually—just the forces of rising wages and economic growth—can adopt better technologies in agriculture. I think the question is how fast any of this will happen.

Should Telangana farmers grow rice?

RAJAGOPALAN: This is where I start getting quite worried. Now in Telangana, rice is already enormously subsidized by the government, other factor markets. You have irrigation, you have free or highly subsidized electricity so that you can drill down for groundwater because rice is very water intensive.

Given that [in]Telangana, the wages are high, it is not exactly a water-rich state—you need to give all these additional subsidies. India spends, I think 0.75 percent of its GDP on fertilizer subsidy, which is pretty crazy. Should Telangana farmers be in the rice farming business to start with, asks the economist who believes resources must go to their highest valued use?

BROWNSTONE: I think that’s an interesting aspect of policies. At the same time the government was supporting this study, they were also supporting subsidies for crop diversification. I think this is an issue that at the same time, people are looking at this, and it’s like, should farmers be growing rice? 

Politically there’s this huge push to push out irrigation because it’s so politically salient, but at the same time, everyone knows that the groundwater is not unlimited, and that farmers should be switching to less water-intensive crops. It’s a really challenging issue because once you have irrigated land, patty is still this safe bet crop that earns. The whole push for millets is challenging because millets don’t really make sense on irrigated land. They’re really great, relatively dry-land crops, which is why that is a traditional crop.

RAJAGOPALAN: They are dry lands.

BROWNSTONE: They are fundamentally dry lands. Farmers are maximizing within the land that they have.

RAJAGOPALAN: No, I totally understand. Again, I’m neither blaming the farmers for getting something horribly wrong, nor am I blaming you. Not at all. I’m just saying this entire government approach, or the Gates Foundation approach, makes no sense because it’s like, oh, we’re going to take all these crazy market distortions as given, and now we are going to create this intervention as opposed to why not remove these crazy market distortions, and the problem may go away in the first place.

Market Distortions and Maximizing Food Production

BROWNSTONE: I think the crazy market distortions do serve a very important purpose, which is around maximizing the production of staple food grains. Something I learned through the course of this study—that the agriculture department’s yearbook, their annual, “we did a great job this year,” was “we produced this many metric tons of patty, and this many metric tons of wheat”—that the state produced. “And the food stores are full, and there was no risk of famine.” It is a very 1960s version of India’s development problems.

I was at a discussion where people were discussing reducing the fertilizer subsidies. What came up was they’d done some modeling on how much would this reduce total food output of India. I think it was 10 percent based on other countries doing this. The politicians basically said, “This is not something that we can play with, reducing the total food output by 10 percent.” I think there’s this really big concern that if you remove some of these distortionary subsidies, the total food prices would rise slightly. There are ways to obviously cushion this, but I think that’s the real concern. These subsidies are fundamentally in place not really to maximize farmer welfare, but to maximize aggregate food production.

RAJAGOPALAN: I’ll tell you why I don’t buy this, because when there is that much more production of rice and wheat, the prices go down, as we know, and then the government needs to intervene with minimum-support prices to keep the prices up. This whole thing seems like it is designed so that politicians can make transfers and win votes in rural areas where there’s a captive audience.

The whole thing doesn’t make sense. If 10 percent of total production was reduced, farmer prices would actually go up. They should be happy about this. Then we don’t need an additional government distortionary transfer, which currently is minimum support prices. In sum total, like I said, fertilizer subsidy is about 0.75 percent of GDP. All agricultural subsidies put together is about 2.25 percent of GDP. We’re not talking about these tiny marginal things. When you say this is concerning, is it really concerning for anybody other than the politician?

BROWNSTONE: I do think it is concerning for both the rural and urban core. The big driver of poverty is food prices. I think if you were to big bang, get rid of PSB, get rid of MSP all overnight, you would need to think about something to cushion the people that are on the precipice of food insecurity.

RAJAGOPALAN: We all have that already. 800 million people in India already get a food subsidy in addition to all the separate agricultural subsidies. 

BROWNSTONE: These are all interlinked.

RAJAGOPALAN: 800 million people need the subsidy because we’ve kept the prices too high artificially. Do you know what I mean? Again, I’m not even saying the markets will solve everything magically. That’s not the case I’m making. I’m saying this is like layer upon layer of distortions, which all seem to be solving opposing and conflicting problems. Did you find this puzzling at all?

Larger Economic Questions About Agricultural Subsidies

BROWNSTONE: No, no, no. I definitely find—I think there’s some interesting papers that try to link all of the different subsidies in very large models. I think it’s a really puzzling system, but I also think that really big agricultural subsidies are not just a developing-India or a developing-country specific issue. If you look at the EU, if you look at the US, across the world, even places where farmers are now a tiny percentage of the electorate, the political economy story that, “Oh, this is just vote-bank politics,” doesn’t really make sense in the US or the EU. There’s still these big agricultural subsidies.

RAJAGOPALAN: There’s no vote-bank politics, but it’s the different version. It’s like concentrated benefits and dispersed costs, which is why you have a tiny group of people who are extremely well organized.

BROWNSTONE: I think it’s a mix of concentrated benefits and dispersed costs and also real concerns about food price stability and food security. I do think food prices are a little bit separate from other markets. Governments fall when there are famines and when there are food price banks. 

Onions and tomatoes, which are the unregulated foods in India, when those prices go up and down, it’s a huge political issue. You could only imagine if we let rice and wheat and these things that are crucial to households’ well-being be as volatile as the tomato and onion markets. That would start creating a lot of political issues for politicians. 

I think there’s a medium. You can have a more sensible system than we do have. I do think that these are slightly different than standard products.

RAJAGOPALAN: For governments, I understand that there are these very specific incentives. For folks like the Gates Foundation, and I understand that you can’t speak for them, why isn’t the solution, “Hey, why don’t we intervene and do some experiments on cold storage warehousing, which will help decrease volatility of prices? Why don’t we float an experiment on futures and options markets, which will, again, help farmers plan better with prices? Why don’t we start looking into how land markets would work if we started doing land pooling and made the size of the land holding bigger so mechanization has a greater uptick?”

Even the interventions, I’m just wondering, why are they all in one direction, which is: Let’s add another distortion in an already massively distorted labor market? Why is there no push in the other direction? Because if we take the developed country example, I agree with you that they also give huge farmer subsidies, but they don’t have such land markets, and they don’t literally subsidize every single input that a farmer requires and then also subsidize the output. There is some learning. If you’re just going to borrow best practices, why isn’t that the one we are borrowing?

BROWNSTONE: They definitely do subsidize the output at times. There is research on some of these other topics. It’s not that people don’t try to study insurance for farmers. I’m trying to start a project looking at how to better structure contract-farming prices, which is another thing that people have explored. And this other paper that I have is about the long run consequences of land reforms, trying to look at how land size is affecting the village-labor market. I do think this is a thing that people study.

This question of how can you increase the pace of mechanization, I still think is important, and I think you can do it in a way that is far less expensive. The actual kind of government outlay for this project is eventually almost zero. It’s the discounted cost of tying up capital for two years, in some cases, the subsidies that you would need to kickstart a rental market. One way that I think about this project is it’s about kick-starting the private sector. It’s not about displacing the private sector.

Actually, I did a survey: 70 percent of farmers were interested in buying their own drum seeders at this point, or getting private access to drum seeders after doing this for two seasons. I do think there’s this aspect where the government can come in for a little bit, introduce someone to a new technology and then leave it and then step away. The Indian government is terrible at stepping away, like I said.

RAJAGOPALAN: That is my question. What’s your hunch on how this is going to play out? Is this going to be one more subsidy or do you think the private market there is pretty robust, the rental market is going to be pretty robust because we’ve seen this play out with tractors and other kinds of technology-rental markets in agriculture. What’s your hunch on what’s going to happen if we have this chat five years from now in Telangana?

Future of Mechanization and Agricultural Policy in Telangana

BROWNSTONE: I think that the nice thing about this kind of subsidy of a one-off purchase is that just the way the Indian state works makes it relatively one off. It’s very hard to procure things, and it’s kind of a one-off line item. Because of that, I think the risk of just endless purchases of drum seeders going everywhere is relatively low.

That being said, Andhra Pradesh actually has a souped-up version of what’s going on in my paper. Their kind of farmer training centers now offer rentals of all sorts of things, including tractors.

RAJAGOPALAN: Oh, boy.

BROWNSTONE: Oh, the private sector does provide that. It’s not to say that other states won’t do crazier versions of this, but I think this idea of combining extension and capital access is a powerful idea. One Acre Fund, which is the most successful social enterprise in Sub-Saharan Africa around agriculture, that’s basically their model. Their model is combining this capital and training. I see the appeal, but I do somewhat share your worries of once a government program starts, it’s hard for it to stop.

But I think that this technology, if it’s framed as a technology demo, which I think [is] how this was framed internally, even though it’s not a new technology, then I think you reduce some of that risk. Maybe that’s something that we can try to message or we can hopefully message going forward.

There is a move this way. I think the farmers were complaining that they used to get all these technology demos and all these other kinds of interesting subsidies, but when Telangana put in Rythu Bandhu, which was this cash payment to farmers, they cut the budget for a lot of this other stuff. There is this kind of move that, “Oh, we give the farmers cash. We shouldn’t give them free seeds or free technologies.” Some libertarian voice within me says this is great. All the farmers are getting cash. They can do with it what they want. There was some logic, to some degree, to the old system, at least in terms of technology adoption of giving farmers a chance to try technologies that they wouldn’t otherwise try as an effort to speed their adoption, which has these broader productivity spillovers than just giving cash, which ends up being used for consumption support, doesn’t provide.

The Long Shadow of Feudalism: Concentration of Land and Labor Market Power in India 

RAJAGOPALAN: I’m curious about the other paper that you have going on, which is super interesting. You’re looking at the long shadow of history, this is not a 5-year, 10-year thing. This is like a 200, 250-year thing. You study something very interesting. In the same region in the erstwhile princely State of Hyderabad, you’re looking at land concentration because of a particular set of programs which were land grants given to the feudal power structure that was in place when there was the Nizam of Hyderabad versus the lands that the Nizam owned under the Nizamat. Now, 200 years later, the places that were more feudal look quite different from the places that were first under the Nizamat.

Second, the places that were feudal also had a very different response to the post-colonial government of India’s push of land reforms than the places that were under the Nizamat. 

Can you just walk me through the Cliff Notes version of what’s happening in this paper, and then we can talk about why you think you find this?

BROWNSTONE: You did a good job of doing the Cliff Notes version of this paper. This is all within this broader Hyderabad polity. It’s not these other papers that are British colonialism versus not colonialism. It’s different shades of the same kind of institutions. In one set of places, the Nizam was more at an arm’s length, and so basically, when the land record system started to get modernized, these places weren’t part of that.

RAJAGOPALAN: Can you tell us the years and time periods when this is happening?

BROWNSTONE: Yes, so the grants are happening from the 1700s up through the mid-1800s. These were mostly grants to feudal vassals that were helping the Nizam in his wars. Eventually, when the British took over, everything in the war stopped. You stopped needing to grant new land to your feudal vassals because you weren’t fighting any more wars.

RAJAGOPALAN: Which is early 19th century. That’s when all this started.

How the Feudal Structure Shaped Current Farm Ownership and Women’s Labor

BROWNSTONE: Yes, all this is early 19th century. Then in the early 1900s is when the Nizam is starting to think carefully about, “How do I actually modernize my revenue collection system?” But that’s only really happening in the areas that they control directly. And then in the 1950s—and that’s where the shock that the paper ends up leveraging is—[are] these big land reforms in Andhra Pradesh, There was this really interesting this guy, Khusrowho was on the policy planning committee, did a spatial difference-in-difference design study in the 1950s, which is wild.

RAJAGOPALAN: I saw the map and the colors. I was like, “Wow, this looks very modern relative to what I would have expected.”

BROWNSTONE: The map is something that we found, but the idea to actually pick villages an opposite 20 kilometers away and compare them in the 1950s was, I think, way ahead of its time in terms of what people were doing in econometrics. What they show, and what we see in our data as well, is basically this registration of small farmers’ lands only happened in the areas that were not controlled or that were controlled directly by the Nizam, where the Land Records Administration had been properly functioning for the past 50 years, and then the areas where it had not, the bigger landlords were better able to bully the farmers into not getting their little piece of land paper.

What you see is there’s this missing mass of small farms, even in the land records data today. That’s the first stage for our experiment. This is this big difference in the landholding distribution between these villages that are only 10 kilometers away from each other and in the same district. We’re holding constant this Iyer story of institutions and district level—we’re holding constant all of that. These are literally villages that are 10 kilometers apart.

The really exciting result that’s actually not in what you read is that we’re doing a wage survey right now. We’re calling up these villages and asking about agricultural wages. That what you see is for women, there is about a 10 percent difference. There is this very arbitrage-able 10 percent difference in wage for women, but it doesn’t exist for men, which matches to some degree what we find in my study.

RAJAGOPALAN: Because there’s mobility for men.

BROWNSTONE: There’s mobility for men. The market power is less binding for men. They have more real outside options than for women doing agricultural tasks. They don’t.

RAJAGOPALAN: They’re much more tied to the land, and then whatever dysfunction comes with the land, they’re tied to that.

BROWNSTONE: The structure of the labor market that’s induced by these big farmers controlling everything does actually play out even in modern wages for female workers. 

New Research in Relation to Existing Stories of State Capacity

RAJAGOPALAN: I love this paper. I found the whole thing fascinating, especially because it’s a new area. We have the famous Abhijit Banerjee and Lakshmi Iyer, your paper. We have Alex Lee. We have Ajay Verghese, who’s done this qualitatively for the Travancore side, and now you’re looking at the Hyderabad side.

What is your intuition for what is going on? Is this a typical Zamindari-feudalism versus non-feudalism story, as is pitched by Banerjee, and Iyeror is this the Alex Lee story of state capacity, which is when the governance system functions better, and they directly collect revenue, they also have revenue offices. Those people become accountable to the people. They start issuing titles. Whereas in the feudal areas, you’re never going to have that development of state capacity? The state is the feudal Lord. What is your hunch on what is driving the difference?

BROWNSTONE: I think it’s more likely the latter bit, that there was some state capacity that existed, but also the absence of the very empowered local elites to hold the farmers away from getting their title. The Nizam was off in Hyderabad, and his agents maybe didn’t care as much or were more like bureaucrats. When the land reforms came, they acted like bureaucrats and granted the title. Whereas in these former feudal areas, because there was an absence of these bureaucrats, and the bureaucrat equivalent literally was the local elite, literally was the large landowner, they had very little incentive to give that title.

What I think is interesting, especially in the Hyderabad case, though, is all of these elites have now left, mostly because these are mostly Muslim elites. A lot of them have left to Turkey or left to other [places], especially in rural Telangana. I don’t think it is a story of persistence in terms of the actual people, but I think it’s much more a story of how land markets shape political economy persistently. Because they’re so static in India, that this land concentration really matters for wages and service delivery, even to this day.

RAJAGOPALAN: Yes. When there is a vacuum in feudalism, what happens is even if the original land grants were given to a Muslim feudal Lord, all the sub-zamindar, patidars they all tend to be Reddys or Kammas depending on which area that you’re talking about. Then they become the de facto feudal Lord of that area or the local political elite, which you did see happen in the early years. This whole Reddy versus Kamma conflict is largely around political power and feudal power and land markets.

I do see your point, but I thought the paper was also very much confirming the Alex Lee side of the story in that older debate. This is new interesting evidence in a new context, and it’s quite fascinating.

BROWNSTONE: Yes, for sure. We do see some of the elite capture in Alex Lee’s story as well. Where we contribute is trying to take it also in the 
“what does this mean for rural labor markets” direction, which I think is my focus across both of these projects. What do rural labor markets mean for these big questions of structural transformation? Because I think rural labor markets matter in themselves because of poverty, and this is where the poorest Indians work and earn their income.

Especially in my first paper, I also show that these labor markets also matter for technology adoption. If these aren’t functioning well, if people don’t have outside options, the pace of how quickly farms mechanize and the rural sector transforms will be depressed. Overall, that’s bad for this goal of lifting Indians out of working as rice transplanters. 

RAJAGOPALAN: No, absolutely. I think the way I would phrase it is you’re not just looking at structural transformation, you’re also looking at what are the impacts when the structural transformation is not quite as smooth as is normally promised, as taking place over centuries. This is a really fascinating line of research. Thank you so much for taking time to speak with us.

BROWNSTONE: Thanks so much for having me.

About Ideas of India

Host Shruti Rajagopalan examines the academic ideas that can propel India forward. Subscribe in your favorite podcast app