India’s growth numbers shape how we understand everything from jobs to investment to global standing. But what if those numbers don’t tell the full story? 

New research suggests India may have both underestimated and overestimated growth at different moments over the past two decades. That insight opens the door to a broader conversation about India’s macroeconomic choices, from exchange rate policy to electricity pricing to the quiet persistence of trade barriers.

To discuss these issues and many more, Abhishek Anand joins Milan on the podcast this week. Abhishek is the Founder and Managing Director of Insignia Policy Research and a Visiting Fellow at the Madras Institute of Development Studies. He’s previously worked as an Economist at the World Bank and was a member of the Indian Economic Service, working in key positions throughout the Indian Ministry of Finance.

Together, with Arvind Subramanian and Josh Felman, Abhishek is the author of a new working paper published by the Peterson Institute for International Economics titled “India's 20 Years of GDP Misestimation: New Evidence.” 

Abhishek and Milan discuss the controversy over India’s GDP estimates, important reforms within India’s statistics ministry, and the debate over the Reserve Bank of India’s policies to defend the rupee. Plus, the two discuss Abhishek’s work on power sector reform and the embrace of non-tariff barriers that stymie the spirit of India’s new bilateral trade agreements.

Episode notes: 

1. Abhishek Anand, Josh Felman, and Arvind Subramanian, “India's 20 years of GDP misestimation: New evidence,” Peterson Institute of International Economics Working Paper 26-3, March 2026.
2. Abhishek Anand, Arvind Subramanian, and Josh Felman, “How GDP data misread the economy, complicated policy,” Indian Express, March 14, 2026.
3. Abhishek Anand and Naveen Thomas, “Free Trade on Paper, Protection in Practice: How India’s Policy Interventions Hollow Out Trade Liberalisation,” O.P. Jindal Global University, January 2026.
4. Abhishek Anand, Arvind Subramanian, and Josh Felman, “Going forward, RBI’s rupee policy must not repeat errors of recent history,” Indian Express, December 29, 2025.
5. Abhishek Anand, Praveen Ravi, Navneeraj Sharma, and Arvind Subramanian, “To help India’s economy, unleash the power sector,” Indian Express, August 27, 2025.

Transcript

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

Milan Vaishnav Welcome to Grand Tamasha, a co-production of the Carnegie Endowment for International Peace and the Hindustan Times. I'm your host, Milan Vaishnav. India's growth numbers shape how we understand everything from jobs to investment to global standing. But what if those numbers don't tell the full story? New research suggests India may have both underestimated and overestimated growth at different moments over the past two decades. That insight opens the door to a broader conversation about India's macroeconomic choices, from exchange rate policy to electricity pricing to the quiet persistence of trade barriers. To discuss these issues and many more, Abhishek Anand joins me on the podcast today. Abhishek is the founder and managing director of Insignia Policy Research and a visiting fellow at the Madras Institute of Development Studies. He's previously worked as an economist at the World Bank and was a member of the Indian Economic Service, working in key positions throughout the Indian Ministry of Finance. Together with Arvind Subramaniam and Josh Feldman, Abhishek is the author of a new working paper published by the Peterson Institute titled “India's 20 Years of GDP Misestimation, New Evidence”. I am pleased to welcome him to the podcast for the very first time. Abhishek, so good to see you. Thanks for joining us.

Abhishek Anand Thanks a lot, Milan. Thanks for having me on your podcast.

Milan Vaishnav So I want to start with this paper that you and Arvind and Josh have just put out. We will link to the paper in our show notes. I think many of our listeners may not have read the paper, but they've probably seen the headlines from the media reporting around your paper. In this new paper, you and your co-authors argue that India's GDP growth between 2005 and 2011 may have been underestimated, while growth between 2012 and 2023 may have been overestimated and often by fairly consequential margins. Let me just start perhaps by asking you what motivated you in the first instance to revisit this question of GDP estimation and what convinced you that something wasn't adding up and that this was something where you needed to really dive deep?

Abhishek Anand Basically, I think it was not me initially, but Arvind when he was the Chief Economic Advisor who had an issue with the new methodology when it was launched and implemented in 2015. If you just look at his writing, because one of the criticism that he often faced is that he has been speaking about it after leaving the government, but actually when he was the chief economic advisor, his first survey that he wrote and there he had a small box explaining his issues and why he thinks the numbers the new numbers that that came out of the new base 2011-12 was puzzling for him. And he followed it up with another chapter, not a chapter, but again a small box in the following years economic survey. My own of course I was working with him on some of these issues. And again, we followed the work even after he left the government. And in fact, he wrote a paper explaining in detail why he feels it is overestimated and by magnitude of around two, two and a half percentage point. So, it's basically Arvind who had an issue and he thought that there are a lot of puzzles which are not explained by the new numbers that the government released. And for me, I mean, of course, I was working with him and there were one or two indicators which convinced me. Nothing very technical and I don't think there one needs to do something very sophisticated. But for example if you look at the period 2011 onwards, all the key macroeconomic indicators be it exports imports investment bank credit foreign direct investment all of these indicators key macro indicators they have collapsed at by a fairly large magnitude if you compare the period of 2005 to 11 vis-a-vis 2011 to 2023. Whereas the GDP number that you would expect that if there is a collapse in each and every macroeconomic indicator, the growth will fall as well, but you don't see that in data. So that's one smell test. At the second, and this is what Arvind wrote in his survey as well. Look at the year of the demonetization. That was perhaps the largest monetary policy shock that the emerging market economy has faced in recent years. And one would expect if your currency is contracting by such a large magnitude, your nominal, forget about real, but at least your nominal GDP would face a decline. Interestingly, what you see is that there's a huge contraction. 85% of your currency goes out of circulation, but your nominal GDP, it surges from around nine, nine and a half percent the previous year to around 11% a year demonetization happened. I think just these two small data points were enough, at least for me, to get convinced that there is something off and we need to understand GDP data better. Maybe we were wrong and we needed to understand all the technicalities, but suddenly there is something not right and somebody and specifically people in the government, they need to kind of look at these issues and try and understand it better. So it goes back in time, starting from 2015, the very first few months when the data was released that Arvind started writing about it and I have of course later worked with him on some of these issues.

Milan Vaishnav So I want to, Abhishek, get into a little bit of the kind of methodological issues that you raise, and we'll try to do that as best we can so that a broad audience can follow. But there are essentially two core methodological issues that your new paper raises, right? And the first is one of the shortcomings of the way in which India measures GDP is that it uses the formal sector of the economy as a proxy for the informal sector. And the reason that this is a problem, building off what you just said, is the latter. The informal sector was hit by a kind of triple whammy of shocks. First, you had demonetization, then you had the introduction of the goods and services tax, and then finally, of course, you had COVID-19 pandemic, which hit India particularly hard. Help us unpack how this proxying kind of works in practice and why you think it has led to biased estimates.

Abhishek Anand Yeah so let me try and explain it as easily as possible for the audience. So basically, as you rightly pointed out for the informal sector, which is a problem not just in India but many of the emerging economies, I mean it's difficult to track what is happening there. I mean the data system perhaps is not that mature where we can track the informal sector that way. And wherever possible maybe we directly try and observe the informal sector but in vast majority of informal sector we kind of proxy their growth through their counterpart in the formal sector and let me try and explain what that means. Say for example, manufacturing. Now for formal manufacturing the way to estimate it is basically looking at the annual survey of industries that the government does on an annual basis. And you get the data from there. For informal sector, again, the proxy is whatever growth that we observe as per ASI, we apply the same for the informal sector. And similarly, many of these sectors where we do the same, for example, transport and communication, we proxy the growth in the informal sector through the observed growth of motor vehicle sales in the formal sector, for example. Other issue is in some of these sectors where people think that okay methodology is right but there is a problem and I'll explain what the problem is. Some of the sectors, for example education and health, we do not really use a proxy from formal sector, but what we use there is we look at the consumption expenditure survey and what has been the expenditure on education and health and then extrapolate growth from there. Now the problem in India is that we haven't done consumption expenditure survey. I mean, now we have started doing, but after 2011, the first, I mean 17-‘18, it was supposed to be out, but it was not accepted by the government. It was junk and then it was-

Milan Vaishnav Correct. It never got released.

Abhishek Anand Yeah. And it only happened now, I think in 2023 and 2022, 23 and 23-24. So basically the growth in education and health in the informal sector, we were proxying through. Expenditure, household expenditure in these sectors but based on 2005 to 2011 survey because of surveys were not done at the time that it was expected to be done. So there are all these issues. Now just one thing which I often you know as a pushback here from people is that oh but why would you claim that it will lead to an overestimation now and not earlier. I mean the same methodology was being followed even prior to 2011. And as we show in the paper, yes this method is imperfect. Between 2005 to 11 and we show it through data, government's own informal sector survey data and if you compute the growth rate how informal sector was doing and correlated with formal sector they were broadly moving in similar direction. So that proxying, while not ideal, was less of a problem in the prior years. The issue starts as you rightly pointed out post 2015 because it was the informal sector which was hit by three shocks that you mentioned. And the sector was not that the formal was not hit, but it was informal that was hit adversely, much more relative to the formal sector. And you see that there is a divergence in the growth that again we show through government's own informal sector data, that how now the formal and the informal sector, the growth rates have diverse completely, which was not the case earlier. And that's why post 2011, this proxy method is not that good a solution to what otherwise was a good solution in the past.

Milan Vaishnav So Abhishek, that was remarkably coherent and I think relatable for any lay person. Let me now challenge you further because I think the second one is even more technical, which has to do with the use of deflators based on commodity prices. And those have moved very differently from broader price trends, right? So essentially, this is coming to the crux of how you factor in inflation and rice rise. Into the data. Can you explain how this distorts GDP measurement?

Abhishek Anand Right let me try and be simple again not sure if I will succeed but I mean as we understand I mean we in case of India we majorly compute GDP or GVA in nominal terms first and for a very small sector for example agriculture which is computed based on production data so you look at actually you know what quantity produced or different commodities and there you basically are computing everything in real terms. But that's not the case with majority of the sector and around 75 to 80 percent of the GVA or GDP are computed first in nominal terms and of course then you have to somehow deflate it to arrive at arrive at growth in real terms. Now the way to do it is generally you kind of look at the sector specific price levels. So, for example if it is manufacturing you would look at what is happening to manufacturing prices by how much manufacturing prices have gone up and then you try and deflate nominal values to real terms. Now the issue in India is that I mean it's fine as long as you are using right set of prices for right sectors but that's not the case. For example, vast majority of services sector and let me give you one example say trade and hotel and restaurant for example right? Where we use wholesale price index and not consumer price index or a producer price index.Now because we are basically using wholesale price Index and wholesale price index have basically a bunch of majorly input. Basic commodities that go into manufacturing, for example, steel, cement, oil, which is being used to arrive at real value of what is happening in say hotels and accommodation. So that's not the right prices that should be used. It should be something similar to consumer price index. That's the price that consumers pay. We shouldn't be looking at the price that wholesalers pay. And for a lot of sectors, that's what we usually do. The other issue is, even for manufacturing, I mean, one argument that we make in the paper is that even wholesale price index is kind of outdated. It has not been updated. If you look at the wholesale prices, basically it should be tracking the key inputs that go into manufacturing, but that's not the case. And you would see that it's basically influenced the way we compute our wholesale price index. It's basically tracking oil price. So everything, what is happening to steel prices, what is happening to cement prices, what is happening to other intermediary input that go to manufacturing. You would not expect it to be, Of course, it will be influenced to some extent by oil. But the correlation is very, very high. So basically, WPI is nothing, but basically, it's tracking oil prices. And that's the problem. So, one, even for manufacturing, wholesale prices, in principle, should have been OK. But just that, in India, it has not been updated. The basket and commodities that go into computing are very, very old and needs updating. And secondly, and that's a major concern, for vast majority of services and even sectors like construction, where we use wholesale price index, which in turn is basically tracking oil prices to deflate everything from nominal to real.

Milan Vaishnav So, Abhishek, let me now just kind of zoom out and try to give this some, what the major takeaways are for the listener, right? I mean, just to reiterate what we've said before is, what you all find is that GDP growth between 2005, 2011, it may have been underestimated and the corresponding growth for the subsequent period, 2012 to 2023, it may've been overestimated. Just to kind of round this out, tell us, why you chose that break point of 2011-2012 to separate the two? And secondly, what is the magnitude of the underestimation in the first period and the magnitude and the overestimation in that latter period?

Abhishek Anand Yeah, so why we chose 2011-12 onwards? Because the new base that India chose, which is 2011-20, although it was released in 2015, but it applies to all the data starting from 2011- 12. So, basically, 2011- 2012 onwards, all our GDP number is based on this new methodology, whereas before that, it was based on the old methodology of 2004-05 base. So that's the simple reason why we kind of look at before and after because of course there are some issues with the new methodology and that's why we wanted to look at what happens after 2011-12. Now in terms of magnitude, so for the period 2005-11, what we understand and we explain in the paper, the magnitude of underestimation is around 1.5% at this point. So, the official growth number is around, if I'm not wrong, around 6.5, but it should be somewhere close to eight percent. For the subsequent period, 2011-’12  to 2023, the growth seems to have been overestimated by one and a half to 2% at this point. So the official growth would be six or 6.1%, but it should be around, say, as per us, four to four and a half percent on an average every year.

Milan Vaishnav So I want to just kind of tie this to kind of where the debate is in India right now. You all wrote an op-ed for the Indian Express, which we'll link to, where you laid out some of your findings, but you also said something I found very interesting, which is that you give a lot of credit to the Ministry of Statistics and Program Implementation known as MOSPE, which oversees this process and is under kind of relatively new leadership in the past couple of years. And you note that your work is both complementary as well as complimentary in its recent revisions. So it's working alongside the work they're doing, but also giving them credit. Tell us what you mean by that.

Abhishek Anand Complimentary with I because I think genuinely in last one, one and a half years, the MOSPE is doing good work and there is a lot of activity all of a sudden happening and you would see a lot new surveys being launched and also the old ones that were irregular happening on time. For example, the survey for informal sector now happens every year and it wasn't happening earlier. I mean it was supposed to happen once in five years earlier but it wasn't happening even once in 5 years but now you would see that happening every year. Similarly, they have launched a pilot for surveying the informal construction sector and as of now we don't track construction sector which is big in India. Similarly CPI has been updated. GDP revision itself the process itself was very consultative this time around as opposed to 2011-12 when that happened because this time there were a lot of seminars, workshop, outsider experts were invited, lot of engagement that happened. So in that sense, I mean full credit to the department and of course we do see a clear break from how it used to happen in the past vis-a-vis how it is happening now. Complementary with E, the reason being I think what we have done it should inform the MOSPE in its own work because while they have updated the base and they have released numbers starting from say 2022-23. They have to also produce a PAX series. They have to go back in time and maybe produce numbers for last seven, eight, 10 years. I don't know for how many years they would so that we can compare based on new methodology, what the growth rates would have been in the past. And what we are doing basically is using broadly the same methodology that MOSPE follows with maybe one or two changes where we differ. For example, now as per new base, they are now not proxying, it would be wrong to not to, that they are not proxxing. I mean, but the level of, you know, the degree to which it used to get proxied earlier, it has come down considerably in the informal sector. And that's what we have done in our work as well, because, you know we had, because you have this survey data available, not every year, but you know over a period of time and one can look at if you don't proxy, but just look at informal survey what happens and also in some of the cases where we use a different deflator. And some of these deflaters that we use, they are using in their new methodology. In some, of course, we don't agree and they continue to use what they used to do in the past, but there is some level of convergence in our own work and the new methodology, and I feel that our work perhaps can inform them when they start working on the back series and they produce data for last 8-10 years. Some of the work that we have done and we have written about in the paper could be useful for them, is my hope.

Milan Vaishnav So Abhishek, let me now just pivot, if I could, this conversation to talking about a few other areas where you've done research and you've worked. And maybe let me move now from the kind of Ministry of Finance, MOSPE, to the Reserve Bank of India, the Central Bank, and exchange rate policy. You and Arvind and many of your co-authors have also written about the RBI's management of the exchange rate of the rupee under the current RBI governor, Sanjay Malhotra. And you give him a lot of credit saying, look, the RBI has done a much better job, and perhaps previous leaders, in maintaining exchange rate flexibility, allowing the rupee to sort of float, but that there's still concerns. For listeners who don't track this closely, what's the downside of the RBI directly intervening to kind of defend the rupee to make sure that it doesn't depreciate too much, say, vis-a-vis the U.S. Dollar?

Abhishek Anand Right. So that article that you're referring to, so we were basically critical of the rupee management starting from late 2000 or mid-2023 to 2024, where, I mean, if you look at the currency and the exchange rate, it was basically flat. And the downside risk to that is that, of course, what people generally understand is that okay, if your currency is not undervalued but fairly valued and for India, that would have meant allowing depreciation to happen, it would have boosted our exports. That's of course partly the reason that's one. Second issue when you try to defend currency and don't let it depreciate and find the right market value is that how the RBI would defend the rupee. It will have to sell dollar and buy rupee from the market, which basically means the rupee liquidity tightens. And the rupee is not as easily available as it should and that leads to increase in interest rate and which is not good for an economy like India where you want to encourage private investment. I mean we keep worrying about why private investment is not picking up. What are the issues? So many things have been solved and in that scenario the interest rate is higher than what it should be. It becomes difficult for private sector investment to pick up. So that's one reason. Second reason is why I mean, basically what you're doing is when you are trying to defend rupee and keeping it artificially high. So, you are keeping imports in a way more attractive than what it should have been. I guess letting the currency depreciate is much better a subsidy policy than doing all these PLIs and stuff and trying to help domestic manufacturing. Because here you don't have to pick winners. So basically if you let the currency depreciate, it makes imported input some of the raw materials more expensive. And that's a uniform subsidy to any industry in India who is in this business of producing some of these raw materials. And you don't run into this problem of big groups lobbying, getting something specific for them. So basically, they are few winners, but a lot of losers. Rather than that, just have this uniform subsidy for every industry in India that is trying to compete with imported inputs. That's much better policy than having PLI and selective interest rate subvention and some other benefits that the government does. And that can lead to rent seeking as well, so that's broadly the issue.

[…]

Milan Vaishnav I mean, so what you're saying, in other words, Abhishek, is that if you decide as a policy or political matter that make in India is going to be your objective and that you want to increase the manufacturing share of GDP, which has been relatively flat over the last 10, 15 years, actually allowing the rupee to depreciate would help your case and do so in a less distorted manner, because essentially what you are doing is ... Is making India's exports cheaper across the board without giving special favors to one business house or one state or one sector.

Abhishek Anand Absolutely, absolutely. And this is specifically important because I mean we keep talking about China plus one opportunity but that window is narrow. It still exists but it is narrow and the best way to exploit it is to basically give this confidence to the domestic manufacturer that they can compete in the global market and one of the key ways to do it is to let the currency depreciate. That incentivizes domestic production and helps them compete globally. So, and also, I mean, specifically for labor-intensive sectors such as leather, footwear, apparel, where margins are very narrow, exchange rate is very, very crucial. And that's the sector where also a lot of jobs would be created. It will not be created in services sector or high-skilled sectors like electronics. The jobs are there to be created in apparel, leather, footwear, textile, food and processing. And for them the exchange rate really, really matters because the margins are so low and how competitive globally these sectors are. So that's why I personally feel that it's of utmost importance that we let the currency find its natural value and not keep it artificially high.

Milan Vaishnav But Abhishek, I guess I'm wondering, you know, how do you think about the trade-offs in these things, right? So if you just take the case, I'm just picking one up top of my head, like India wants to now be the iPhone manufacturer for the future. And we see statistics coming out every month saying, you know, India now accounts for X share of global iPhone production and so on and so forth. And so a lot of the component parts that are going to making a cellular device are imports, right, which would be made more expensive, no, if the rupee depreciated. So how do you think about that trade-off in terms you want to lower the input cost, but you also want to make your exports more competitive?

Abhishek Anand Right, so just coming to your Apple example, and of course, we are doing very well when it comes to exports, but it is debatable how much domestic value addition happens in India and whether it has improved over time or not, because of course when a new industry is setting up here, it will take some time for local or domestic value edition to happened and at least my sense is that it has improved over time. Now, of course, coming to this question of what happens if a rupee is artificially high, now if rupee is artificially high, so basically then imports are cheaper, right? And then some of these components that go into manufacturing mobile phone, Apple will have less of an intensive for example to try and get its partners in India and produce it here. It can very well import everything from outside because the imported prices are low. And in this case, actually, if the rupee depreciates, there is more incentive for Apple to try and nudge its partners to relocate to India and produce some of these things here rather than importing all of it. And in that case, my own sense is that the domestic value would improve. Not that it is not improving but the rate at which it will improve would be higher than what we currently see in India.

Milan Vaishnav Abhishek, one of the things that's so interesting about your work is that it's very eclectic. You work on many different subjects, and all of them are extremely relevant to the here and now. I want to turn now to another place where you've done some work, and that's the power sector. This is, again, something that you've written recently about in a couple of columns. In one recent column, you've talked that He basically said, look, you know, everybody knows that there are all kinds of problems with the power sector in India. These have been well documented. There's a lot of data on this. People have written about this, but one of the things which has been maybe underappreciated is the macroeconomic costs of the failures in the power sector. And I wonder if maybe we could just kind of start there. If you, if you kind of step back and look at it from a kind of 30,000 foot level, how do you characterize the core weaknesses of say India's electricity distribution system, the kind of core piece of this puzzle? How would you, what are the kind of stylized facts that you would point to?

Abhishek Anand Right, so one thing before I start answering what you were saying, and that's right, many people commented once we wrote that article that all of this is known, what is new here. What I just want to say is that that's not right. I think our understanding of power sector is very, very wrong. One basic fact, we all look at how much power subsidy government is giving. I mean, we look at the budgeted numbers. And that's a very accounting method of looking at the subsidy that the government gives. And let me give you one example. So basically, say, for example, a household, it consumes electricity 100 units. And the prices are set by the regulator and the price that the regulator sets is say four rupee per unit but the government says that okay no we are going to give it for free and the numbers that everyone has in their mind is that okay per unit subsidy is four rupees because that was the price the regulator set and based on that they compute overall subsidy number and even best of the experts in India they always refer to that number. And that's completely wrong and you would appreciate it if you just think of it from an economist's perspective because what is important is what is the cost of production of electricity and the cost of the production of the electricity is much higher per unit in India will cost roughly around say 9 or 10 rupees. Yes, for whatever reason the regulator sets it 4 rupee and then they will pass on the burden to some other sector but the subsidy that we are giving to either agriculture or domestic households when they are consuming one unit of electricity is not 4 rupees but 10 rupees because that's the price and that's a very high price and then we try to recoup some of those subsidies by passing on the burden. In a good scenario even say manufacturing and it links to your other question on manufacturing. They should also they should actually be paying say 10 rupee per unit, which is the cost of production But they don't they pay much higher because the government has to get subsidy burden shared with either commercial entities or industrial entities and they end up paying much more and I mean, this is one overall taxation number that we compute and Arvind has it in his book that what is the overall taxation that manufacturing has to pay. So it turns out to be around 80-85 percent that for one unit of electricity a manufacturing or commercial entity has to pay and that has consequences for a country like India where we want to push manufacturing and it could be a sizable cost for some of the sectors, for example textile, where it's very, very energy intensive. We want to attract a lot of data centers. Their energy, I mean, they consume a massive amount of energy, and if you want to give them energy at such high prices, they will not be willing to come in. And that's why you see that government will give a lot subsidies to them, and that's why we are able to attract. So, manufacturing in India really needs that support from the government, where not that they should be subsidized, but at least they should be paying the fair price, and not. And not sharing the burden that should have been actually paid by agriculture and domestic households. Now what is one way, I think over time, I mean, this is a consensus that we have reached, you know, all people that, so many of us, I'm not the only one, but my other co-authors who work on this sector is that it's very difficult to break this policy regime where you price it fairly for each sector. And the other issue is that, of course, when it comes to distribution discords, they are basically monopoly, right? And there is no competition. And then they are inefficient. And you have absolutely no incentive to become efficient, to think about, oh, how can I manage my company better? How can I reduce my transmission losses? I mean, they just have no and I think the way is either privatization and we have seen it happen. Odisha, I think, two years back, they completely privatized. There is no public sector discount. And I think it's doing better now. It will take time, but it's going much better. Second, allowing the industry to exit. And you would see now the news items coming out every day that, OK, in Andhra Pradesh, first, Google, and I think now Reliance as well, they have been given parallel licensing by the government. So they're completely off the grid. They will have their own procurement, transmission, generation, everything. Because that allows them to produce and consume electricity at a much cheaper rate, for which otherwise they will have to pay almost double the rate. And these are data centers. Again, as I said, for data centers, it's very, very crucial because they use a massive amount of energy. And if you want to attract them, you have to give them this flexibility. So that's what I feel. So, at a macro level, it's basically a management and efficiency issue. Ownership Problem? And the only way to deal with it is that you will let privatization happen or at least competition that where you also have a private discom and a public discom maybe geographical separation but more importantly while that will happen after maybe few years it may take some time allowing industry and commercial entity to exit and manage things on their so that they can manage their costs better.

Milan Vaishnav You know Abhishek, you talked a little bit about the fact that the way electricity is priced imposes, you know, almost a kind of 100% tax on manufacturing, right? It's pretty significant The two other things that you talk about in this piece, which I was wondering if maybe could elaborate on a little bit which is first households are now receiving subsidies comparable to agriculture, which is a pretty striking finding. And the second is that a large share of those subsidies are actually going to middle- and upper-income households as opposed to the poorest. And I wonder if you could just elaborate a bit on those other two findings.

Abhishek Anand Yeah, so the first one, I mean, this, your question I think was 100% taxation to manufacturing, right?

Milan Vaishnav Yeah, so there's kind of three things. You talked a little bit about the taxation on manufacturing, but I was also wondering about the kind of, you know, how do you calculate how much subsidies households receive? Because the fact that those subsidies are comparable to agriculture now is pretty striking. And the second is, you know, how much of this is actually going to houses that don't really need these subsidies who we might classify as middle- or upper-income households?

Abhishek Anand Yeah okay yeah so the first question that how did we come up with this number that the subsidy is now fairly evenly distributed because everyone thinks that okay it's basically the agri sector where we have free power and not domestic sector but as you would notice it has changed in last few years and you would see now increasingly every election one of the I mean, most loved schemes that any government, either in power or in opposition, they will announce is that, okay, if we come to power, we will give X units of electricity free to domestic household. And it has changed things quite a bit in last, say, five to seven years. And in many of the states, majority of the states, what you would see is that domestic households, they either get subsidized electricity and that links to my previous point because people feel that okay there is no subsidy. But even if you are not explicitly saying that 100 units free or 200 units free, the rate or the tariff that the regulator sets usually are very low. It would be say 2, 2.5, 3 rupee per unit of electricity on an average whereas the cost of production is much higher. So, that's one, the gap between what the households pay vis-a-vis what the cost and even if you consider efficient because of course part of the cost of producing electricity is because of inefficiency and in our paper we take out the inefficient part and then just look at what the efficient cost would have been and even taking that into account there's a sizable amount of subsidy that goes to household and second because now many of the governments have explicitly committed that they would give 100, 50, 200 depending of the state. Units of electricity free to households every year that adds up to around 50, 45 to 50 percent of overall power subsidy that goes to agri and households combined. Now why do we say that most of it goes to households? Very simple, so you know the states wherever we worked and we could do this analysis. For example, consider a state. I don't want to name any state, but you pick any state. Suppose 200 units free. The government says that, OK, I'm going to give you 200 units free. If you look at the consumption pattern of the households, that basically means that close to 85% to 90% of the household, they don't pay any electricity bill, because their consumption is usually lower than 200 units. And then you have some states where, in fact, 300 units are free, where virtually nobody pays any bill. So just imagine like 80-85% of households or 90% of households do not pay electricity bill or highly subsidized electricity bill. Now I mean it's hard to argue that 90% of the households are poor in India when the government claims that perhaps poverty rate has declined to single digits in recent times but even if you are very conservative and you believe that maybe 25%, 30% households are poor and rest are not. And based on that liberal assumption, we do the calculation that, OK, 25%, 30% households, maybe they are poor, they need protection, but rest of the households are either middle class or rich. What is now based on their overall electricity consumption? What proportion of subsidy is going to reach vis-a-vis poor? And then you would see. And that same goes for electricity as well. For example, in Punjab, I mean, you have many, many rich farmers and their share in overall electricity consumption would be very high. So maybe 15% will actually go to poor households and 85% is actually consumed by very rich farmers and then you do the calculation you would see that most of the subsidy is actually going to middle class and rich households in India.

Milan Vaishnav Wow. I mean, it's kind of a remarkable finding and again, kind of links back to something that you all had done in one of the surveys when Arvind was CEA. Remember you had this great table, right? Looking at different subsidy schemes and actually trying to quantify how much of each one went to the people who really deserved them. I want to make sure that we spend a little bit of time talking about your work on trade policy because this is something that's very much in news these days India has become. A very active player in terms of trying to strike trade deals, particularly bilateral trade deals. So, we saw one with New Zealand announced, there was a mini deal with Australia, one with the UK, one with the EU, one the US, we don't exactly know where it stands, but both sides are still talking. You have this interesting paper with your colleague Naveen Thomas, again we'll link to this, called, “Free trade on paper protection and practice.” And one of the things that you point out is that, you know, while FTA's free trade agreements are often framed as kind of tools of liberalizing, of liberalization, Indian governments have often used domestic policies to offset some of those gains. And one the FTAs you look at specifically in this paper is the free trade agreement between ASEAN and India. And I'm wondering if you could just say a little bit about what your analysis revealed in terms of the gap between what the intent of the free trade agreement was, at least on paper, and what the implementation actually ended up being in practice.

Abhishek Anand Right, so in that paper we looked specifically the free trade agreement that we have with ASEAN and basically looked at the apparel sector because the expectation was that once we had this deal with ASEAN, the domestic industries, the apparel producers in India, they should have had access to zero tariff raw material that goes into producing apparel. And in this case, it was very important because Indonesia is world's biggest exporter of viscose, which is one of the, apart from polyester, the second most important raw material that is required for modern apparel because now we are moving away from cotton based to synthetic apparel, right? And viscose is one the key input and India should have benefited. But that's not what happened. Of course, even before, it's remarkable that even before on paper, the deal was signed and it became effective. Of course, there was lobbying, things happened, whatever inquiry. And then the largest exporter of viscose from Indonesia, and of course, China and other countries, but for us, Indonesia matters much more. They were slapped with anti-dumping duties that you are resorting to unethical practices. And that's why anti-dumping duties of X dollars per kg would be imposed on you. And it was supposed to be for 6 months initially and it kept getting extended for many, many years and it was only in may be 2000 so it was imposed in 2013-14 I am not exactly sure now when the deal became effective but before the deal become effective just few months before that it was in post and it continued for years. Despite this thing of having just for six months initially, but it kept getting extended. Then what happens of course, on paper you have 0% tariff, but then even if you have an FTA, it doesn't matter you don't have to pay, you still have to anti-dumping duties. And that basically meant it nullified whatever the gains were expected to be, and it made imported input. Cheaper sorry expensive as expensive as it was earlier so basically there was no benefit of the free trade agreement and it's not the case I can understand that you know if you have a domestic that's what we see in the paper that you have domestic industry of viscose there where are the lot of producers lot of domestic competition and you are trying to you know I give some benefit of doubt that okay there is an infant industry argument you are to protect but that's not the case I mean, basically, we just have one dominant producer, basically it's virtually a monopoly in India and this one big firm that produces the product that we are talking about. And then what happened is basically, of course, they were insulated from the foreign competition and they could continue to charge higher tariff, in fact, more than what they were earlier. Just matching the anti-dumping level prices to domestic producer and interestingly of course because they are also one of the big exporters and we show in our work that how the same commodity is being sold in India at a higher rate whereas in domestic of course global market they have to be competitive and they're selling it at the global or the world price level and we just use that gap that what it meant in terms of rent being extracted from Apparel industry in India and who's getting benefited is basically this one dominant producer and There is no other producer here. Whereas your downstream sector is a parallel industry which is highly competitive lot of producers a lot of job creation that happens and India could have benefited from this sector, but virtually nothing happened and when anti-dumping and we can discuss it later But of course with anti-dupping duty got replaced with QCOS which are which are even I mean more brutal major then even anti dumping duty because virtually I mean you just cannot import anything and that was also imposed on the same in Indonesia. So that's how it has happened in practice on paper looks every everything looks nice. And that's one thing I just want to mention is that many a times I mean I have heard this argument that all free trade agreements are bad. Look at the free trade agreement. What has India gained? Our export hasn't gone up. Import has gone up. All of that, but it's just that we have never really let free trade agreement to work the way it should. And just to conclude, that's why FTAs are bad is not a fair assessment. And that's what I'm now hopeful that now because of whatever constraints and reason we are signing a flurry of FTAs hopefully it will be given a chance to work better then what has happened in the past?

Milan Vaishnav I mean, just incredibly interesting. Abhishek, you're right. I want to just maybe end by asking you about this issue of kind of non-tariff barriers, and you've written a lot about these, what you mentioned just now, QCOs, which are quality control orders and they're, they have been used extensively by the Indian state to restrict imports. Um, the government has recently scaled these back and it has actually claimed that move as a major reform. Something that hopefully will benefit India in terms of the amount of trade it's able to do. I wonder if you could just tell me, you know, to what extent does the government's moves kind of address some of the concerns that you and your co-authors have had, or do you think it's just a drop in the bucket and there's still deeper structural issues when it comes to the non-tariff barrier side that they need to address?

Abhishek Anand Right. So let me answer it in two parts. So one thing is it really a reform because now we do see India signing FTAs, a lot of QCOs being rolled back. Even tariffs when the government came to power starting 2016-17, there was a huge hike in tariff rates. And Arvind and Shoumitro has written extensively about it and they have a great paper on how the tariffs increased post 2017. Now, there have been corrections and governments usually they claim that major reform and I have written it about in the past and my thing, what I say is that these are all self-inflicted pain. I mean, we have done it in the past and it's not that we experimented with protectionism and then learnt it is bad and now we are trying to correct it. I mean in 1991, we saw, I mean we learnt and we course corrected. I mean there is nothing new to be learned and gained from again. Going back to that era and being protectionist. So it's certainly not a reform and so that's one my worries. So while things have been improving and on paper it looks fine and some of these things are happening because there is an external pressure globally what is happening. Also, there is a push from the US administration and that has certainly contributed to some of changes that we see. But, and this is what Josh and Arvind, they have written about in the past, is that the state of the government, I mean, if we just look at some of their policies, a majority of their trade policy, it doesn't give me confidence that, you know, they believe in being open and not being protectionist. So maybe slightly political, but I don't want to get there. But my hunch is that, or maybe I would just be skeptical that: For how long will we allow it to continue? Or will things go back to what it used to be once the global scenario changes, things settle down, the war is over, relationship with US changes for better? Do we again, will we again resort back to some of these trade policies that we have experimented with in the past and we have miserably failed? So, I would be cautious. I mean, I'm happy what they're doing, but I would still be cautious, and I would just wait how it plays out for the next couple of years.

Milan Vaishnav My guest on the show this week is Abhishek Anand. He is the founder and managing director of Insignia Policy Research and a visiting fellow at the Madras Institute of Development Studies. He's also a co-author with Arvind Subramaniam and Josh Feldman of a brand new working paper published by the Peterson Institute titled “India's 20 Years of GDP Misestimation, New Evidence.” Abhishek, this was a pretty wide ranging conversation, exchange rates, trade, QCOs, power sector, GDP. Thank you so much for really educating us and I appreciate you taking the time

Abhishek Anand Thanks a lot. I mean, it was wonderful to interact with you. And, you know, these discussions, of course, I also learn a lot and just to think through all these things with you on this podcast. Hopefully I did justice and the audience, your audience will not regret if they listen to the podcast.

Milan Vaishnav I'm sure they won't. Thanks again.

Abhishek Anand Thanks. Thanks a lot.