Vaibhav Rathi on Political Influence on Bureaucratic Priorities & its Impact on Investment in India

SHRUTI RAJAGOPALAN: Welcome to Ideas of India, a podcast where we examine academic ideas that can propel India forward. My name is Shruti Rajagopalan, and this is the 2026 job market series, where I speak with young scholars entering the academic job market about their latest research on India. 

I spoke with Vaibhav Rathi who is a doctoral candidate in economics at Stockholm University. We discussed his job market paper titled, “Political Influence, Bureaucratic Priorities, and Capital Accumulation in India.”

We talked about bureaucratic state capacity in India, the role of bureaucrats in approving large infrastructure projects, whether politicians use their influence to distort administrative priorities, the impact of these processes on private investment 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, Vaibhav. Welcome to the show. It’s such a pleasure to have you here.

VAIBHAV RATHI: Thank you so much, Shruti. It’s a bit surreal and pretty awesome to just talk to you in person and one-to-one in this podcast format. 

RAJAGOPALAN: Well, less surreal because—

RATHI: It’s just crazy. [chuckles]

Bureaucratic Priorities, Political Influence, and “Distortion without Corruption”

RAJAGOPALAN: Less surreal because we’re going to discuss your paper, [laughter] which is actually on a really important issue. You’re talking about some of my favorite areas of work that I’ve been tracking now, it feels like, for decades. 

In this particular paper, you ask whether political influence makes India’s bureaucracy more productive or it just changes whose work gets permitted or allotted. And the setting in which you’re studying this bureaucratic state capacity, but also bureaucratic priorities, is the district administration. In particular, the administration that’s responsible for large capital-investment projects that are owned either by private firms or sometimes spearheaded by central governments and state governments. 

And you basically look at this question of bureaucratic capacity and prioritization from the political lens. You measure the influence of a Member of Parliament and whether the MP belongs to the coalition government that’s governing that particular state, because a Member of Parliament obviously is the Union Parliament for the Union Government. And then, you combine that with the project-level CAPEX records—in this instance, the big database that CMIE has—with election data and also the career histories of IAS officers who are district administrators.

What you’re really looking at is—it’s almost an event study of whether a Member of Parliament who’s currently in office finds themselves aligned with the current state government or not aligned with the state government. Because it’s staggered, because state elections and union elections are not yet “one country, one vote,” OK? So, this is a very useful research design. 

And what you’re really looking at is the total number of approved projects which need to go through this district administration. And the premise is that there are always too many projects and too little personnel. What are the projects that get priority? Does the fact of an MP who’s a union government representative being aligned with the state government have any influence on state-level executive prioritization, which we see through these IAS district administrators? 

What you find is that actually the number of projects getting approved doesn’t change much with political alignment. What you really find is that the composition of what is getting approved changes, right? So, the private projects actually increase and the private share of rather what gets approved increases, but new private announcements themselves don’t increase. 

So, it’s not like there’s this big private-sector boost when an MP finds themselves aligned with the state government or misaligned with the state government. It’s more that now the private sector gets priority. As opposed to the central-government projects and the state-government projects, there’s this whole other thing which we can talk about. 

You argue exactly that—that this kind of political influence can reallocate priority in a state bureaucracy which is already very capacity-constrained. So, an influential MP can move one project which is already in the pipeline to the front of the queue without changing a single rule. There’s not necessarily any major change in policy, any major change in announcements, but it is a change in priorities, right? This is what you call “distortion without corruption.” 

Did I largely understand the setup of your paper and your findings? And feel free to correct me because there are so many moving parts in your paper, which I found fascinating. Did I largely get it right? And what got you motivated into the subject?

RATHI: Yes. I think you gave an excellent summary. I would want to highlight just one thing, which is the directional aspect of things, before we go into detail, and I’ll start with a bit of motivation. But just the directional aspect, that an MP can be aligned or nonaligned. The way we do identification is changes and things. And so, the effect you summarized was when the MP loses influence. In the story terms, it’s like, the district administration knows that an MP, even though he or she’s the same, they have less of an influence within the state government.

And state governments are the space where the career decision transfers and postings of the state-level IAS cadres are decided. So, they subconsciously, or just by the way of strategic interest and behavior, they internalize these things that not all MPs are the same. Some MPs are more powerful than others. What I am trying to say in this paper is that that power or the power to influence or power to monitor the bureaucracy changes the bureaucratic attention. And so, as soon as you see the MP loses their influence, the bureaucrats immediately shift away from central-government projects, and in the process, private projects end up benefiting. We’ll have more to say on what kind of private projects are benefiting in some time.

The Queue Metaphor

RAJAGOPALAN: Yes.That intervention was very useful. When I was reading your paper, I was thinking of it almost as if there are three lines. It was Ganesh Chaturthi recently. You think of Siddhivinayak Temple or some of these places. There are three different queues. There’s a VIP queue, and there’s a regular queue, and there is a nonpayment queue. It almost feels like the gatekeepers decide which queue opens up when, right? 

And this is exactly what you’re telling us, which is, if the MP is aligned with the incumbent state government, then the central-government queue opens up much faster and much more. And the moment there’s a new state government in place, and the MP finds themselves misaligned, the private-sector queue will now start opening up, right? Is that a good way to think about it?

RATHI: Yes, yes, perfect. Yes, that’s exactly–the queue metaphor was pretty spot-on. 

RAJAGOPALAN: What got you motivated to work on this paper?

RATHI: My motivation going in was to understand the capital formation and the economy. And when we think of capital formation, we have some familiar objects we can go to. One is the regulatory environment. Since 1991, we have had [a] number of deregulations and simplifications of regulatory processes. But [the] other part of it is that, given a regulatory environment, we also need the bureaucracy and regulators and the people to process those regulations also.

So, my motivation was that the one part of it, which is just policy landscape, and when it comes to regulations, is relatively well understood and much more understood compared to just the capacity of the bureaucracy itself to process those regulations.

That was my entry point to see, “OK, what can we say about—does the flow of state capacity—strictly as measured by how well does the bureaucracy process these regulations—A, can we measure it? B, can we see what are the instruments and what are the drivers that vary it along meaningful dimensions?”

India’s Bureaucratic State Capacity

RAJAGOPALAN: On this, I have a couple of things to unpack, Vaibhav. So, I want to first start with, can you walk us through how do you think of capacity or bureaucratic capacity in this sense? And the typical economist question of, relative to what, right? So, how is India on bureaucratic capacity overall, before we get into the specifics of the paper?

RATHI: So, there are two points to it. We can start with just [the] simple size component of it. For whatever reason, historically, we tend to—I think you already are very familiar with what I’m going to say—is the concept and idea of “the flailing state.” It has been well understood and people have this at the back of their mind, that the Indian state tends to produce rules and regulations, and we somehow lack the capacity to just implement the rules and regulations that are already existing.

So, why is that? One is just the simple object of, what is the size of the bureaucracy?

We can go and measure, and, let’s say, we can look at World Bank bureaucracy index. We find that, for some reason, if we just look at all public employees, which includes all the bureaucrats—higher-level bureaucrats as well as the service delivery people who are more at the front line of public service delivery—even if we combine all of them, the number is much lower than many other peer Asian countries and many other developing countries.

For example, Vietnam is at around 90 public employees per 1,000 population, where India is around 28. So, the difference there already is pretty big. But then, not all of those people are involved in regulation. The class of officers that, especially within the context of [the] Indian state, who actually have all the executive powers, then we are basically just talking about [the] IAS cadre. Because [the] IAS cadre employ—starting from district, state, and central levels, these are the people that basically occupy all influential positions. And they’re like managers of the state. They have executive power. They have power to make managerial decisions, and all this.

How do we assess that? OK. Currently, the authorized capacity of [the] IAS cadre is around 7,500 people for the population of 1.5 billion people, around 1.4, 1.5 billion people. 

One way to compare it is just to go back in time. Let’s say if we start from 1947, we inherited [the] IAS system when it was Imperial Civil Services before that. I looked into it a little bit in just a historical, anecdotal way. And I realized that back then, the civil administration systems of India and Great Britain were very similar. 

But, since 1947, the system of civil service in Great Britain has undergone significant change, where now they don’t have this high level of career protectionism where you hire elite cadre officers, and they occupy all breadth and depth of all civil services and public life. They have moved away from the system, where they actually have [an] internal labor market of civil servants, where if there is a posting, some civil servant can actually apply. There is not a centralized posting system that can be influenced, and . . . There are a lot of strategic influences and power centers that can arise from having a centralized decision-making system.

And the second thing is that, if you just compare the count also, in the UK right now, A, it’s a bit tricky to compare because the setting is a bit different. But even if you just cut down the numbers by a lot and just try to distill officers by the role, which have similar roles as IAS officers in India, the number you’re going to end up is around 25,000 to 30,000 people for the population of around 70 million people in the UK. In India, we have around—the authorized strength is 7,500 people. Even that is not fixed. Recruited in full capacity, the actual number is somewhere between 5,500 to 6,000 people for a population of 1.5 billion people.

So, I’m trying to give you a sense of the personnel constraints and capacity constraints that exists within the elite bureaucracy as the country is growing, as government welfare schemes are exploding. It’s just—an IAS officer at the district level, I assume, is just drowning in work. And there, you get a sense of when you have such insane levels of tasks to complete—which, by the way, can be easily improved upon by just increasing the cadre size. But for some reason, we don’t, and there are going to be consequences arising. Even if we hire highly skilled, meritocratic people, they are human beings at the end of the day, and there are only 24 hours in the day. So, yes. So, that’s the setup which we’ll build upon.

The Procedural Context for Large Investment Projects

RAJAGOPALAN: Awesome. So, before we get into your paper, I want you to build out one additional context, especially for those who are unfamiliar with the process. What actually happens between the announcement of a large investment project until it becomes, say, an operating power plant or railroad or factory or something? So, what does that procedure really look like?

RATHI: Yes, so, that’s a very good point to start. This part is not strictly unique to India. In any regulatory environment in most countries, when you have [a] large capital expenditure, there is a project announcement, and then there is a long chain of events that needs to happen when that announced investment or promised investment can graduate to productive capital. And this long chain contains events which are both bureaucratic in nature and nonbureaucratic in nature.

I specifically focus on events that are bureaucratic in nature - “I need land access approval,” “I need environmental clearances,” “I need forest clearances.” And all of these clearances—I can submit an application and I can put together a proposal. But that proposal will go into the bureaucracy, and then it becomes a question of how efficient and how productive the bureaucracy can be to process these inflows of potential capital investment and graduate them into productive capital.

And the way it works is that it’s basically an exercise in gathering information and assessment. Suppose, even if we abstract away from, let’s say, district, state, and center levels, there are different levels of bureaucracy. If an application comes in, even if the decision is made at the state level, or if, let’s say, there’s a big investment project, the state ministry of power or something is involved, the way the Indian state is set up is that they will have to request information from the district because the district administration is both their information-seeking arm as well as the enforcement arm.  

If somebody sitting at the state level has to enforce some ruling, they will pass on the information to the district. And even if they have to seek some information, they will seek that information from the district administration. So, in this way, the district administration is a singular place where the state can extend its reach to collect any information. That’s how the regulatory process works, which, in very simple words, is just an exercise in collecting information and assessing information that, “OK, does this project qualify all the regulatory criteria as specified in the legislature?”

Political Alignment and Project Prioritization

RAJAGOPALAN: Yes. With this context, both on how a large project gets approved, and now also the IAS bureaucracy, how overworked they are, and what their incentives or disincentives are depending on how we recruit and promote them, I want to get to the real meat of your paper, which is about political alignment and what is it that the political alignment or nonalignment actually does. 

How should we think about alignment between a Member of Parliament at the union level, a state government, and a district-level IAS officer? If you can explain it to us conceptually, then we can get into what happens when the state government and the MP are from the same coalition or not.

RATHI: Yes. There is a feature in the Indian state design, which has been previously explored by other papers also, which is that because of this capacity constraint, there is a relationship between bureaucrats—and especially bureaucrats at the district level—and the local politician.

The way that relationship works is that—in a simple political economy framework, if you go ask an economist, you’re going to find a split that some economists will take the side that politicians should be able to influence bureaucrats because they carry the will of the people and mandate of the people. Bureaucrats are not really answerable to people. But on the other side, some other economists would say that there is a real information asymmetry. By and large, bureaucrats are the ones that have better information because they are the ones who are tasked with doing whatever needs to be done. Politicians generally have a much poorer information set than bureaucrats. 

The question is of power-sharing. The way it ends up happening in the Indian context is that, even if we have designed a bureaucracy which is on paper supposed to be free from political interference, because the politicians cannot really hire and fire people—in many countries, the politicians get to appoint their own bureaucrats, but in India, it’s not the case.

The other component which is relevant for the paper, which is just characterizing the political economy at the district level, at the roots level, and when it comes to political economy, there are two main players. One is the politician and one is the bureaucrat. Then there is a game, as we political economists tend to talk about things. There’s a game that is being played between the politician and the bureaucrat, and that sets the state—at the district level, the state of governance can be either in, let’s say, a good state or a bad state.

But the headline finding is that when the politician is able to control or monitor the bureaucrat in a much better way, it gets reflected in the task allocation and the effort allocation of bureaucrats.

Measuring Bureaucratic Activity

RAJAGOPALAN: Yes, yes. So, here, it’s very useful to start thinking, because so much of your paper is about measuring these concepts that we intuitively understand, right? One part of it is, as you mentioned, the political alignment, which we can measure through elections and who is in office and who is not in office, and you’ve constructed a dataset around both the IAS officers and the political alignment. 

Now, the other aspect of what you need to measure to get this headline finding is bureaucratic activity itself, right? In this case, we’re talking about very specific bureaucratic activity when it relates to these large investment projects. Can you walk us through how do you observe and measure this kind of bureaucratic activity in your paper?

RATHI: Yes. Just one quick point before that, for people interested in political economy and the space who are well aware that there are few objects that people tend to think of when it comes to state design. One is selection. We have already taken care of selection in Indian conext; there is highly meritocratic selection. And incentives—there are well-defined career concerns, incentives. But the third important point is the point of monitoring, which is a good entry point [to] the measurement thing we’re going to talk about.

The problem with monitoring is basically that when we come to these managerial types of tasks, which are inherently a multiple-task framework, it’s really very hard to even define what is the task list. Let’s say there’s a big literature on what is the value add of managers. I’m not too familiar, but my understanding is that in these kinds of roles, a lot of measurement ends up being residual, TFP-type managerial quality, which is precisely because this is a setting where oftentimes even the task list is not defined.

An IAS officer might be doing some other set of tasks in District A, and in District B they might have completely different concerns. And it might vary across time also. Let’s say, in election years, there are just different things that you have to do. In order to monitor or properly measure their bureaucratic productivity or efficiency, it’s not an easy thing to do. It’s a very hard empirical challenge, which is the reason why even at the state-design aspect, we have not been able to have a formalized monitoring mechanism at a higher bureaucratic level.

That’s how there’s this soft signal that gets passed around that some politicians can pass their feedback, their soft signal, how cooperative . . . If you go look up all these 360-degree [assessments] of IAS officers—they have these annual assessment surveys—a lot of weight is being put on how cooperative the officer is, how big of a team player they are. There’s a lot of value in the soft signal that gets passed from the local politicians to the state committee, the state government that decides their transfers and postings and so forth.

And bureaucrats also know this. So, all of this is like the setting is an as-good-as-perfect, complete-information setting because in bureaucracy, things tend to be highly specified. So, then it becomes a scenario where bureaucrats know that politicians presumably have a good amount of sway over their future career trajectory, the kind of postings they have, and the kind of departments they get posted in. 

But then, we come to the point that not all politicians are the same. This is where the key identification assumption, and then, now we are getting into the main econometric setup of the paper, which is that not all politicians are the same. In this setting, the main identification assumption I’m making is that by matching an MP to a district administrator, which is the IAS officer, I can make a claim that an MP who is aligned with the state ruling party—and presumably they have a better communication channel with, because of their shared party affiliation and party alignment—they would have more of an influence over the IAS officer compared to another MP who is not in the same party.

The only assumption that I need to defend, which I think I’m very comfortable in defending, is that party alignment matters. If you share the same party umbrella or coalition, the communication within that coalition is better than communication across coalition. Even at the state government, if they have to prioritize, they would put more value on signals from MPs who are from the same party compared to signals from MPs who are from the different party. 

Understanding Changes in Politicians’ Influence

RAJAGOPALAN: This is very helpful context, Vaibhav. Now, can you tell me, within your state election design, how do you isolate a change in an MP’s influence? And since we’re talking about what happens when an MP loses political alignment, can you actually help us measure that and unpack your headline finding?

RATHI: Yes. I’m going to claim that political alignment of the local MP is a useful proxy [for the] influence the MP has, over the bureaucrat. There is a three-step causal chain that I’m laying down. I’m going to claim that the MP has some effect on bureaucrats that’s going to change bureaucrats’ behavior and that will generate some outcome in the real economy, which is our y variable, which is capital formation in the local economy. Now, I’ll need to be very careful at each step because three steps is already long for a causal chain.

The first stage, which is that, can I confirm in the data that political influence matters when it comes to bureaucratic career decisions? And then I’m going to claim that, given that bureaucrats know this, that the political influence can have a measurable impact on their career, that will have [an] effect on their behavior, and that behavior I’m going to measure in just one specific dimension, which is regulatory process. There are other kinds of things that bureaucrats might be doing. It’s not possible in an empirical setting, or especially in a standardized way, to measure all [the] things that [they] are doing. So, for the sake of simplicity, I just take one thing, which is standard across all districts in India. That allows me to make the point in a much clearer way. 

Going back, the first stage is my claim that somehow politicians are able to have a reasonable, measurable influence over bureaucratic career trajectories. There, we already have some bit of evidence. There’s Iyer and Mani’s paper from 2012, which says that bureaucrats in India—and bureaucrats here in the sample there are IAS officers in India treat political connection as a substitute for their own performance as well as investment in their own human capital. The point of the paper is that more-connected bureaucrats end up performing worse, but they get the same career trajectory, as good as somebody who is not well-connected but they are putting in more effort. There’s this complementary nature in the production function of career trajectory of IAS. The officers know this, and they are well aware of the value of political connection, so to speak.

Then I have to say that, “OK, that’s true across the board, but I’m going to pick one measure, which is political alignment.” There, in a descriptive way, in a fixed-effect framework, I show that bureaucrats that are connected to aligned MPs, immediately in the next three or five postings, they get much better postings compared to bureaucrats that happen to be connected to nonaligned or nonsenior MPs. There are a number of dimensions along which MPs differ. Senior MPs might have more influence than junior MPs. Aligned MPs might have more influence than nonaligned MPs. The alignment thing is two-sided. 

Now we get into the idea that, “OK, we see that alignment does matter for bureaucratic career trajectory. How do we measure the causal effect of it?” The alignment is like a relationship object; it can change from two sides. Either the state government can change, or the MPs themselves might change because of national Parliament elections, and somehow the new MP might have a different alignment state because there’s a government turnover at the central level.

I can use both elections, basically. Given that, in India, we have staggered state elections, it allows me to construct a staggered, difference-in-difference event study-type stacked panel where each cohort group is basically one state election treatment. Each treatment spawns a bunch of control units and a bunch of treated units. Treated units are the ones where the MPs are the same, but, because the state government changed to a different party, initially they were, let’s say, aligned, and then they become nonaligned because the state government changed.

Then, we can put some restrictions over control units, where the control units are in the same initial level of alignment, both in pre and post period. And then, that allows you to compare treatment and control in a causal way. That’s one part of the empirical design.

The other design is that this is aligned to nonaligned, but it can go [the] other way also. It just so happened that in this other way, I don’t have many treatment events. That’s how I can use this event study framework for that also, and I do that, but the effects there are slightly more noisy because just low treatment events, low unit counts. There I use a different empirical study, which is closed election regression discontinuity, where I just specifically ask, “OK, what happens when there is an alignment status change from nonaligned to aligned because of a close election of the MP themselves?” The good thing is that these two designs are independent in nature in the sense that they use completely different treated units. They use completely different times. They complete in a different set of districts, but they’re shared by just one mechanism, which is, how does local district bureaucracy behave when there are these political changes? 

Findings of the Study

RAJAGOPALAN: And what results do you find?

RATHI: So, it’s interesting that the results that I find are consistent across both independent empirical strategy. I thought that, assuming that political alignment matters and these MPs have more influence over bureaucrats—and we have these things that people assume that there’s a value in political connection—I thought the private approvals, and because private players can collude with MPs to get their influence, I’m going to see that private players benefit. But what I find is exactly the opposite, which is that political influence ends up being a hindrance in this bureaucratic workspace.

MPs actually use their influence more for their electoral gains and more of their own public projects, and that crowds out, through the bureaucratic attention span, the approval of private space. 

And so, the results are basically that when the alignment state changed from aligned to nonaligned, and MPs lose their influence over the bureaucracy, bureaucrats immediately shift their attention, and they start processing more private projects. The rate of approval of private projects increases by around 30 percent, which is mostly coming from central-government projects and state-government projects, which is because state government is irrelevant when it comes to relationships between MPs and bureaucrats. You don’t find much effect for state-government projects, which is nice because, theoretically, they also got good placebos because there is no change in the relationship between state government and the central bureaucrat at the local level.

On the other side, using the regression discontinuity design—in the previous election, let’s suppose the MP was nonaligned, and there is a political turnover, the MP themselves change, but the new MP, who is aligned, but they won the election by a close margin in the RD logic type. Immediately, you see that, because of this, the exogenous increase in political influence of the MPs, you see immediately the share of effort that bureaucrats put towards central projects increases by around 25 percent. The effort share that goes towards private projects falls because the total-effort budgeted constant in this side of movement, when the central projects get more priority, the private projects suffer. That’s how the district-level regulatory environment just worsens for private players without any legislative change in [the] regulatory environment.

This is just purely a political economy thing where there are these career concerns and the desire to control bureaucracy from the politicians’ side. Combined with the overarching capacity constraint that we have in the bureaucracy, [it] all ends up together to produce these behaviors and patterns where we find this cyclical movement in the rate of approval itself and how the local capital is being formed at the district level.

Why Central-Government Projects Get Bumped in the Queue

RAJAGOPALAN: No, that’s a very interesting set of results, and the magnitude is quite large given what you find. Here, Vaibhav, I had one question about why is it only central-government projects that lose priority to private-sector projects, going back to that queue analogy? What is it about state-government projects that are fundamentally different or respond differently within this political design?

RATHI: This is a good puzzle. I have some answers, but the thing is that, a priori, I was actually surprised that I did find some effects for central projects, because the way the data is structured, and the things that I know about how institutions work, I did not expect that, given the central projects, they have too many touch points with the bureaucracy, and not all of them are at the district level. Central-government projects already tend to be very big, both in terms of scope and cost. Think of, let’s say, [a] big highway project, or, let’s say, a big piece of infrastructure which is funded through central-government funds.

A priori, you can imagine that it’s not obvious that they would be so bottlenecked by district bureaucratic capacity itself. It’s surprising that I do end up finding effects for them, which gives me a signal that the district-level bureaucracy is actually quite instrumental, even for big infrastructure projects, where a lot of other bureaucratic agents at the different levels are also involved.

The other part of your question is, why don’t we see effects for state-level projects? When we started, the whole thing is basically operationalized by career concerns of the bureaucrats. Nothing has changed between [the] relationship of the state government and the local district bureaucrats. So, a priori, I’m measuring the effects of a very finely tuned treatment, where I’m not claiming that I’m measuring a total effect of political influence or something. I’m measuring political influence as driven by just one thing, which is party alignment. So, there is no a priori reason that bureaucrats should even change their behavior when it comes to state-government projects. Does that make sense? Because nothing has changed as far as their relationship [with the] state government. The state government has carried the same level of influence over their career, which is independent of the treatment that I’m measuring the treatment effect of.

Implications for India’s Political Economy

RAJAGOPALAN: That’s super helpful. I think I do agree with the way you’ve set up this explanation. The broader question I have, Vaibhav, before I let you go is, how should one think about different aspects of political economy given your paper? I can disaggregate three, four different parts. One is, of course, how do we think about the civil service or the bureaucracy: their incentives, their recruitment, their career progression, so on? The second part of it is, who controls approvals of a project, which is also a big policy question? The third is, should we start getting into questions like automatic approvals for the most part, given that there is this big political economy knot which gets tangled up for no reason other than a state election? 

And finally, is none of that relevant? Is it just a matter of lack of capacity, and if we just triple the number of civil services officers, this problem would largely go away? What is a good way to think about the broader policy and political-institutions implications of your paper?

RATHI: Yes. I quite like this question because this is a good point to think about, “OK, we have some sets of effects, but what is the main instrumental latent variable which is driving the whole thing?” To me, in terms of priority, I think just the effort budget—it’s striking how tight the effort budget binds, because how wild the movements are in terms of what bureaucrats put their effort to. It indicates to me that the first- important thing, it’s basically just the number of offices that we make available to process all this paperwork that we have to do.

At the district level, there is no reason that one single officer should be responsible for all public service delivery, maintaining political incentives, or just catering to political demands for local politicians, as well as processing all regulation when it comes to gathering information and passing that information to upper bureaucratic levels and all of these things. There is just one single office, which is district DMCO office, which sits at the middle of a lot of things. It’s not obvious why we just need one officer, who is in charge of so many things.

For example, one research paper which sheds more light here is this paper by Jeffrey Weaver, where he looks into district-splitting —his point of the paper is more political entities. But I think a lot of effects in that paper are also driven by just how [the] administrative state in India is designed. Where you just increase the districts, you also increase administrative capacity because each district has their own administration. He finds that when you split districts, you just increase the number of public goods that are being offered. Just good things that happen. There’s a lot of increased service delivery.

In my framework, you can imagine that if one district is split into two, immediately the effort budget has twice [the] amount of endowment that they can—each administrator has, let’s say, only half of these things to worry about, and so they will have much more capacity—just doubling in capacity when it doesn’t . . . I think maybe that’s where the biggest gains that we have lie when it comes to easing the burden on the regulatory state.

RAJAGOPALAN: Yes.No, actually, I don’t know if I like this so much because it matches with my intuitions perfectly. I had written a paper many years ago with Alex Tabarrok on premature imitation in India given its poor state capacity.

RATHI: I know that paper.

RAJAGOPALAN: We make a similar argument intuitively, which is, the Indian state does too much too soon. I think this is a great example of what happens when the Indian state does too much too soon. “Are certain kinds of environmental clearances and approvals and so on important?” Absolutely. “But do we have the capacity to enforce regulation at this granularity and at this burden?” Not really. We need to actually think of our regulation in line with the capacity of our bureaucracy or our implementation. In that paper, that was purely conceptual; we never got into what are the implications of something like that.

I think you very nicely highlighted to us what are the implications of that. That’s why that whole queue metaphor, where who gets to be in the VIP line or which line converts into the VIP line because of an outside event, like a state election being won or lost, suddenly starts changing so many things on the ground, and it changes which road gets approved and where it passes. Does a factory or a new university get built? So, these are very big and important questions. I really enjoyed reading your paper. Thank you so much for doing this.

RATHI: Thank you. Thank you so much, Shruti. It was really, really nice talking to you. I hope you enjoyed reading the paper. I think a lot of people already know that there are challenges in state design and a lot of things that could be improved upon. But my goal was to highlight that there are real growth effects that are just costs to growth that we are already incurring, and which is not born from any bad policy or bad regulations. They are different. They are independent. But even with good policy, we need a capacity to execute that policy. And somehow, we have paid too little attention and emphasis on just making sure that the state is capable of doing what it needs to do.

RAJAGOPALAN: Absolutely. Thank you so much again for joining us on this show. I’m excited to see the rest of your work in future years.

RATHI: Thank you. Thank you so much. It was nice talking to you.

Ideas of India is produced by the Mercatus Center at George Mason University. If you enjoy this podcast, please subscribe on Apple Podcasts, Spotify, or your favorite podcast app. Help us grow by giving us a rating and leaving a review. Follow us on X @srajagopalan and @Ideasofindia. Also, check out our initiative commemorating 30 years or India’s market reforms at the1991project.com.