Arvind Panagariya and Shruti Rajagopalan Talking Trade

Episode 3: Arguments for Protection: Beyond Infant Industry (Part I)

SHRUTI RAJAGOPALAN: Welcome to the discussion series on free trade and liberalization as part of the 1991 Project at the Mercatus Center. I’m Shruti Rajagopalan, and in this conversation series I will be talking trade with Professor Arvind Panagariya, who is the director of the Deepak and Neera Raj Center on Indian Economic Policies, and the Jagdish Bhagwati Professor of Indian Political Economy at Columbia University.

In the past, he has served as the first vice chairman of the NITI Aayog in the Government of India, and as chief economist of the Asian Development Bank. He is the author of a number of books, but for today’s conversation, in particular, we will focus on his recent book, “Free Trade and Prosperity.” Welcome, Arvind.

ARVIND PANAGARIYA: Hi, Shruti. Very pleased to be back with you.

RAJAGOPALAN: Thank you for joining us for episode three and talking trade and reforms. I wanted to talk today about a few aspects of protectionism as a continuation of our conversation last time. In the last episode, you detailed the various reasons why the infant industry arguments against free trade are misguided. Today, I want to go over some of the other common arguments against free trade and for protectionism, especially in developing countries.

I would request you to perhaps expand on each of them and tell us if these have any merit, whether it’s a question of information externalities or diversification, or coordination questions, capital market problems, import substitutions, the bias against imports and the bias toward exports. These are just themes that have popped up over and over again, starting in the ’60s and ’70s in India, but these themes and arguments are very much alive today and they are always brought forward in some kind of new packaging to serve up protectionist policies. I think it would be very, very helpful if we could go through each of these arguments, and if we could get your take on them.

Diversification and Information Externalities

PANAGARIYA: Wonderful, Shruti. Let’s begin with the diversification and information externalities. The diversification argument has, itself, come in two different forms. First, there was the early diversification argument that was used, particularly, in the Indian context, where we wanted a diversified structure of industry. As a part of that, we went in and because the constituency for open markets was almost nonexistent, India almost went to the extreme of offering what one might call protection on demand, and sometimes in the earlier literature, particularly Max Corden, I think he used this term, made-to-measure protection.

The point was that if anybody said, “here is a product that the country is importing, but I want to produce it at home,” and we said, “Okay, we will simply prohibit the import of the product.” It’s pretty much anybody who is willing to produce the same product that’s being imported would get protection.

Another way we did it was that, under investment licensing, we would say, “Well, we’ll give you the license.” Suppose these are tube lights. Somebody says, “I’ll produce tube lights, but of course I’ll have to import the parts.” We say, “Well, all right, we’ll allow you for four years to import the parts and issue you the license for X number of tube lights to be produced, but within four years you have to indigenize the product.” The diversification here meant you are not only assembling the tube light, but also, ultimately, sourcing the components that are assembled locally.

That could mean either you yourself manufacture these or seek some local manufacturer for these. That was the diversification plan. That’s one form in which it played out. That eventually fell out of favor, at least even within India after 1991 reforms and so forth. Now there has been a resurrection of that in a very different form. There is some development research that came through.

There’s this paper by Imbs and Wacziarg, and they find that the rising per capita incomes are accompanied by increased diversification within sectors. If industry grows, even if it remains a constant proportion of the GDP, what happens if per capita incomes are rising is that within industry, each product becomes more and more diversified, different varieties of the product come in. There are new products being discovered, which become a part of the basket.

Some sort of innovation seems to accompany either in process or in product. What has happened is that some economists—most notably I think you know Dani Rodrik—have tried to capitalize on that diversification being a part of rising per capita incomes as a part of rising growth rates. He says that product diversification itself requires the discovery of new products, variants of existing products and processes that allow the available products to be produced at a lower cost of production.

He says, then, that these discoveries do not really lend themselves to patenting. If you could patent them, then, of course, there is no externality problem. And this is something we discussed earlier, if you recall, in the context of infant industry argument for protection. He says, “Well, look, because they cannot always be patented”—meaning that there’s a danger that they’ll spill over—“then we’ll have that problem that then nobody would invest in this discovery of new products.” That’s the problem we encountered earlier.

He says there’s a free rider problem here, other firms can copy them without having to incur any of the costs of discovery. Therefore, no firm would want to undertake the discovery and that, of course, then paves the way for government to step in to take some action. Then Rodrik says, “Look, the first-best policy is a subsidy on investment in new nontraditional industries, but that requires close monitoring, which is not practical.”

Then he goes on, and this is something I’m just giving you as a quote, really, that this is how he then builds it up. He says that—in Hausmann and Rodrik, this was a 2003 paper—“We recommend, generically, a carrot-and-stick strategy. Since self-discovery requires rents to be provided to entrepreneurs, one side of the policy has to take the form of a carrot. This can be a subsidy of some kind, trade protection, or the provision of venture capital. Note that the logic of the problem requires that the rents be provided only to the initial investor, not to copycats. To ensure that mistakes are not perpetuated, and bad projects are phased out, these rents must, in turn, be subject either to performance requirements, for example, a requirement to export or to close monitoring of the uses to which they are put.”

That’s the argument. What he is saying is that well, look, you can’t patent them, and so you have to incentivize these innovations. How do I incentivize innovation? Then there is a carrot-and-stick strategy. What you need to do is ensure that anybody who makes a discovery gets the rent, at least for a while, that accrue to it. That can be some sort of subsidy, or then it could also be trade protection or the venture capital.

First of all, if it’s venture capital, the government doesn’t need to intervene, and venture capital is a phenomenon that should come endogenously. Why is that intervention even required, really, if you allow room for venture capital to come through, if the market sees that there are some entrepreneurs out there who can innovate, right? I mean, that’s the whole idea of the startups, et cetera, being financed by venture capitalists.

Protection now, he also throws in there, trade protection could be the carrot. We saw, if you recall, in the context of infant industry argument, that trade protection really does not solve that problem. I don’t see how if it fails in the infant industry context, then how will it not fail here when it comes to product innovation, right? Any kind of innovation that costs money and then can be free ridden by somebody else—firms are not going to innovate. They will continue to produce the existing products so that they can take advantage of the protection, basically. Yes, subsidy can do something, but I come to that issue also in a minute. Then he goes on to mention the stick and says that to ensure the mistakes are not perpetuated and bad projects are phased out, these rents must, in turn, be subject to some sort of performance requirements, exporting or close monitoring of the users to which these things are put.

First he says that, well, the first-best policy is subsidy on investment in new nontraditional industries, but this requires close monitoring, which is not practical. If you can’t monitor that, how can you then say, simultaneously, that you’ll do close monitoring to which you are putting the subsidies and so forth. It’s arguing both ways first saying that, “Oh, first-best cannot be achieved because there is a monitoring problem, but then I can do all sorts of other monitoring,” which is the stick part of his argument.

There is an inherent contradiction. Also, going further, even leaving these facts aside, there are at least three other problems with this argument. Well, one I’ve already noted, which is that state protection doesn’t solve the externality problem. So throwing in trade protection as one of the kinds of carrots is patently wrong, and he doesn’t elaborate how it will solve it, as far as my analysis is concerned because the problem arises in the infant industry case also, it is very much the infant industry case.

In the way, this whole diversification business is a sideshow for his argument. Anytime there’s an externality, of course, there’s a case for intervention. I don’t see why he has to bring in this diversification issue here. If there are products to be innovated, processes to be innovated, regardless of whether or not there is a diversification issue, there is a case for intervention. Question is: Is there a case for trade intervention? And that answer is a resounding no.

Now, also, the success stories that we know from Rodrik and Hausmann, if you go back to the work that they refer to, to bolster the argument, these are cut flowers in Columbia, soccer balls in Pakistan and hats in Bangladesh. The reality is that none of these were actually successes made by the government. These were all market successes. The governments can come in later on after they have actually succeeded.

It really reminds me of something that my friend, the famous Japanese economist Tatsuo Hatta used to tell me. We were together at the World Bank at the time and he used to say that in Japan, the METI, the Ministry of [Economy] Trade and Industry, in those days, in the ’60s and so forth, they would simply see what sectors are succeeding, and then they would give them the subsidy. It would look like they were creating that miracle but this was happening anyway. This is a very similar story—cut flowers in Columbia, soccer balls in Pakistan and hats in Bangladesh. These are not successes created, actually, by the government. These happened much more generically in the countries themselves.

RAJAGOPALAN: You just said that every externality may need some intervention, but it does not need trade intervention. I would actually go one step further and say every externality doesn’t necessarily merit an intervention. There are some externalities which are inframarginal, right? They don’t affect the optimal production or allocation of the good in question. Sometimes, externalities get resolved through vertical integration, the kinds of arguments that Demsetz has made, right? Sometimes there’s a Coasian solution to externalities, especially in innovation markets if there are tradable property rights.

The inframarginal externality, of course, Buchanan and Stubblebine have laid that out. Every time people see externality and therefore it needs to have a Pigouvian resolution, which now spills over as trade protectionism and interacts with the other side of international political economy, it’s not so clear to me that the argument is convincing. I think the standard has to be higher.

PANAGARIYA: Absolutely. In fact, now you could actually go even yet another step and argue that to think that the governments in these countries are capable enough to actually pinpoint the externalities and actually implement the solutions, even very simple solutions, and to think that this will not create perverse incentives on the part of the lobbyists is to assume a lot.

RAJAGOPALAN: Absolutely.

PANAGARIYA: The governments have actually created, often, more problems in supposedly trying to correct the externalities than they have solved. I remember both my friend and mentor Larry Westphal, who was one of the very early economists to study South Korea, he ultimately, eventually, wrote a more retrospective piece in the Journal of Economic Perspectives.

Larry does believe in, and certainly in those days—I don’t know his current thinking—but he certainly at the time believed in the infant industry protection and all. But he was very careful to say that, “Look, this requires a very, very capable government, and that is not the case in most developing countries. It’s only in very rare cases, such as those of South Korea and Taiwan, that you had capable governments. To draw lessons from their experience and to think that you can replicate it in many of the African countries, or even in South Asia, is to assume a lot.”

RAJAGOPALAN: Even when you have capable governments, there is a question of what is one’s model of innovation, right? Is it a process of trial and error and a sort of bottom-up feedback mechanism through which innovation arises—this is through consumer feedback, this is because of competition, other margins through which the innovation is being propelled? Or is the innovation the result of a small group of experts and their ability to identify winners and losers. Very often, they conflate venture capitalists with governments. Venture capitalists are a small group of elite people who have the ability to pick winners and losers, but they’re also the residual claimants of their choices.

That model can’t automatically be superimposed on the government, and now we say there’s an externality, so we’re willing to let go of the bottom-up, trial-and-error process of innovation, and now the government can pick winners and losers, as well as a venture capitalist or someone else can. I think that’s another really big conflation, but that’s a conflation of frameworks. It’s not about a particular case or a particular country. It’s more a mental model of how one thinks innovation takes place in an economy.

PANAGARIYA: True, and on that, we have very little, limited understanding. Generally, what seems to work is broader policy on this and the kind of model that the United States has pursued generally, that very good higher education, incentivizing research at the universities and then interactions between the universities and the industry. At the end of the day, that seems to work a lot better than any specific kind of intervention that you do by incentivizing specific entrepreneurs and so forth.

You can sometimes pick up certain areas where R&D is required—3D printing kind of thing. Even there, there are actual successes, but when you pinpoint these things, it is much harder. If you take the broader framework, you get better successes.

RAJAGOPALAN: Yes. At the end of the day, the fewer frictions for factors of production to get together in different combinations, the greater innovation one will get in the economy. Even broader principles like free movement of goods, free movement of people, free movement of capital, at the framework level, these are the necessary ingredients so that then you can launch a thousand combinations of different things and some kind of innovation takes place.

This kind of pinpointed intervention here and there to solve one externality, that problem, diversification, capital market imperfection, actually, it undermines the larger framework which will allow more combinations. It’s increasing frictions in the economy. In one sense, this kind of intervention, if you take a very big picture view of growth and innovation, something like what Paul Romer talks about in the endogenous growth theories, this is actually a problem for innovation in the larger scheme of things.

PANAGARIYA: Yes. What you’re saying reminds me of the story that Shanta Devarajan used to tell us in the very early days. These were the days that laptops were just beginning to come in the United States. He traveled to South Korea, and the custom guys stopped him, then they opened the laptop, started taking pictures and everything. So he thought, “What was wrong?” Later he understood that this was a new product.

RAJAGOPALAN: Yes, and they had no idea.

PANAGARIYA: Therefore, they wanted to do some sort of reverse engineering or something. There’s some interest because of that.

RAJAGOPALAN: Absolutely.

PANAGARIYA: What I’m hinting toward is that trade, if it moves freely, diversely, there’s a lot of diffusion of technology through trade because technology is embodied in the machines. There are products, also, which have technology which you can then reverse engineer if trade is happening.

I remember, in India, we were not even allowing these imports to happen. The innovators or entrepreneurs never even had a chance to see the product and then to reverse engineer. It was on the other hand things like medicines did move and those got reverse engineered because of the very weak patent law, and so a fantastic generic industry emerged. It was not so much any specific intervention that the government had done, but simply a law which allowed you to copy the patented medicines, in those days, that created the generics. It was through trade that the first medicines would come in.

RAJAGOPALAN: This happens even at a personal level. Because I have traveled and I have lived outside India—I’ve lived in New York City and Washington, D.C. which are food capitals within the United States—I am now more likely to cook Italian food or Korean food or Thai food at home because I have tasted it in a restaurant. I know what it tastes like. I have been to stores where I can get the ingredients.

My grandmother who never set foot outside the country, nor did she eat at restaurants, was a fantastic cook, but she has never made spaghetti or she has never made kimchi or anything. Even at a personal level, we experience this on a daily basis, but we forget about it when we think about it in economic models or anything else. The idea that exposure and diffusion of ideas and technology takes place when there’s free movement of goods and people. That’s the most basic insight that each one of us has.

PANAGARIYA: Absolutely. I think those are all fantastic examples and the food one is great, actually, because I’m a generation slightly before yours. My mother on traditional Mewari food was an absolutely fantastic cook. Even if it was Punjabi, Bengali or any other foods, she wasn’t, because that was a generation that didn’t travel very much at all. Even within India, because of the lack of movement or very restricted moment, people did not. Within our lifetimes, those barriers have broken because of the movement of people, information and so forth.

Capital Market Imperfections

That’s one of the reasons I want to go through the nuts and bolts of these arguments. Now, moving on to the next one, which is the question of capital market imperfections. This is, basically, the argument that interventions through protective tariffs, or subsidies or some kinds of benefits to domestic producers, they’re justified in order to correct the distortions in the capital market.

If there were absolutely perfect and free and open capital markets, we wouldn’t need this, but given the imperfection of capital markets, we need to solve that distortion with an additional intervention, in this case, protectionism. Can you walk us through this argument? To my knowledge, virtually, no economy has perfect capital markets, but many economies have high degrees of free trade. Also, some sense of how different countries solve the problem of imperfect capital markets without protectionism.

PANAGARIYA: Good way to phrase it. First of all, I want to go back to the infant industry argument. You remember that one of the things we said there is that if learning really is internal and the industry is socially desirable, then the market will ensure that it will happen. The way it will happen is that I incur losses today, but my profits tomorrow more than offset the losses. I can cover my losses by borrowing from the banks. I go to the capital markets, borrow to cover my losses today. Tomorrow, when I make profits that more than offset today’s losses, I can pay back the loan to the creditors. That’s how it was.

First of all, the capital market imperfections argument is brought in right in the context of trying to resurrect the infant industry argument. It was the thought that the capital markets may be imperfect, and so the bankers might not see it correctly. They might think that there is no learning or tomorrow your costs will not fall enough. Therefore, there is an argument for infant industry protection.

First, that clarification is essential as a starting point. Imperfection in capital markets is not an argument for infant industry protection because, even if there is no learning involved, if capital markets are imperfect, then the problem can arise almost anywhere. Anything that, let’s say, requires a heavy investment, very capital-intensive industry, and has a long gestation period—you will make an investment, it’ll take three years before output begins to happen.

For three years, no output is going to happen. For three years I’m only making losses. Even though there is no learning involved here, it’s simply the fact that it takes that long to set up the entire factory. It’s not a learning issue at all. The capital market imperfection issue is quite independent of infant industry. One should not conflate infant industry argument with capital market imperfections. What we assumed there when we discussed the infant industry argument saying that you can go and borrow from the capital markets, is just the right way to approach the infant industry argument.

Now come to the capital market imperfections themselves. Certainly, in trade, there is quite a body of literature where this is cited as a possible argument for protection. If somehow the bankers are kind of underestimating the return to industry, then when anybody goes to borrow from the bankers, they would under-lend to the industry and that would leave the industrial production suboptimal.

Now, forget even that, actually. Even if for any other reason, let’s say, the industrial output happens to be below optimal—there may be other market failures and so forth—even then, you could say that the first-best will be to subsidize the output itself. As a second-best, you could subsidize capital. That is how a lot of the trade literature made the argument, that the first-best intervention is credit subsidy to the extent of underestimation.

Even if we actually locate the distortion in the capital markets, then, of course, the first-best thing is some sort of credit subsidy. The second-best intervention would be a production subsidy, where the distortion is on the production side so that the industrial output is suboptimal then you subsidize output.

As a second-best, you can subsidize capital—the use of capital in industry. If, on the contrary, the distortion is directly in the capital markets, then the first-best is really capital markets subsidies, is what we would say, and the subsidy on output is the second-best way to solve it. Now, the problem is this. In principle, let me grant it that, conceptually, you can make that argument. When it comes to practicing that kind of argument, what is the problem?

The first one was very nicely pointed out by Anne Krueger. Again, you’re thinking in terms of this two-product model: one is agriculture, the other is industry. Somehow within this framework, I see that because of this whole export pessimism that at other times we were talking about, if I rely on agricultural exports, it’s a nonstarter for me because demand is relatively inelastic. Demand is subject to low pricing, price elasticity as well as low-income elasticity.

In any case, because of low-income elasticity, as industrial economies become richer, demand will continuously shift in favor of manufacturers and against primary products. Therefore, my terms of trade as the exporter of primary products will continue to deteriorate. It’s not a winning strategy. Likewise, even if I increase productivity in my primary products and want to export more, then my own exports will lead to a massive decline in the price because the demand for agricultural and primary products is also price inelastic.

The price inelasticity would mean my attempts to sell more will result in a massive decline in price and therefore a fall in the revenues that I am able to get. If primary products are a nonstarter, I’ve got to work on industry. How do I do the industry? I find some ways so we can have capital. We won’t say that the reason industry is suboptimal is that my capital markets are not working and this industry is not getting a fair share of credit, so let’s subsidize the industry.

Let’s recast the problem in three goods. You’ve got two industrial goods, one is an export good and one is an import good. Now, if you start subsidizing the use of capital in the import good, it’s entirely possible that your comparative advantage was actually there in the export good and that was under-produced good. Once you do this, you will actually pull capital out of the export good into the import good or import-competing good.

That will be the wrong thing to do because, by assumption, I’m saying, my underproduced good may have been actually the export good. It’s not so straightforward. The biggest problem, I think, in my view on this capital market intervention thing, was pointed out by Ronald McKinnon, and very rightly so. He came from the finance angle. He was an expert on financial markets, one of the very early experts on that subject.

McKinnon’s point was that when you’re in a developing country—in the end, this is sectoral intervention that you work for. Even if you’re trying to work, let’s say, on the export sector, and you want to say that not enough capital is going to the export sector, you are talking about subsidizing the use of capital in the entire export sector. He’s saying that the developing country problems are actually very different.

You cannot go with these kinds of subsidies at the entire sector level. What happens is that every sector, including even agriculture, but other industries, even services, because these are very fragmented, they all have entrepreneurs with very high returns and entrepreneurs with very, very low productivity. Your high-return activities are not concentrated in a particular sector so that you could give the subsidy to the sector.

You really have these high and low productivity activities going on simultaneously in the same sector. There is no easy shortcut actually to developing the financial markets. He says, ultimately, this is all about development of the financial markets, and work on that, otherwise, you are going to make a mistake. Because if you say, I’m going to give capital to industry as a whole—by the way, a good example of all this is there in the Indian context, which is the priority sector lending.

RAJAGOPALAN: Yes, which is terrible.

PANAGARIYA: What are we doing? We are doing exactly what Ron McKinnon      says. The MSMEs are not getting enough. Agriculture is not getting enough. Exporters are not getting enough. Farmers are not getting enough. Allocate a significant part of your credit to these sectors. These sectors, as you know, have high-return producers and low-return producers.

These are priority sectors, there are nonpriority sectors, but many nonpriority sectors actually do have a lot of very high-productivity entrepreneurs. Your job is actually to really identify the high-return entrepreneurs or high-return activities, which are spread throughout the economy, and give credit to them. That requires development of the financial markets. If you think that you can really solve the problem by giving credit at the sectoral level, then that’s what you would think you’re not developing your financial markets properly. That has happened in India.

RAJAGOPALAN: It’s a giant rent-seeking exercise with additional problems. You give loans to priority sectors and candidates that are low productivity, because of political bargains and political connections. Before you know it, you have massive NPAs, and you have a banking crisis that is on the brink because of this kind of priority sector lending.

PANAGARIYA: Well, suddenly there’s a contributor. It is also this reckless lending that happened in the contracts.

RAJAGOPALAN: Absolutely.

PANAGARIYA: I don’t pin the entire problem of NPAs on the priority sector lending, but it certainly has contributed. Even if it doesn’t contribute, even if things are normal and all, and banks are able to cross-subsidize and manage, the point is that you are neglecting certain very high-productivity activities because you have to allocate a significant part of the credit to these priority sector lending activities, and you don’t even know. At what level do you want to maintain MSME sectors? That should be an endogenous matter. You don’t want to, through artificial credit subsidies, keep alive small little firms. The other big problem is actually these very small enterprises. If you then also try to reinforce that through the credit subsidies and all, then you’re only perpetuating a structure which is the wrong structure.

Industrial structure itself actually becomes hostage to that kind of lending, because if you say that these are my priority sectors, with no clear justification of why these are the priority sectors, then you’re perpetuating a particular structure of the economy. The whole development process is actually about the structural change. At the end of the day, you want the share of agriculture, particularly in employment, to fall, and the share of industry and services to rise.

RAJAGOPALAN: Yes. This does the opposite, right?

PANAGARIYA: Yes. It perpetuates it.

RAJAGOPALAN: Absolutely. There’s an additional pattern I’m observing now that you’re walking us through multiple arguments. It seems to me like the problem is usually not a trade-related problem. They’re using trade interventions and protectionism to always solve a problem which is somewhere else in the economy, and they’re not able to solve it, or it’s a much harder fix.

It’s much easier to enact a credit subsidy for this sector or that exporter, as opposed to the decades it takes to have financial sector penetration ease, frictionless, or very low friction lending, good identification. That kind of infrastructure takes a really long time to build.

All the models basically assume that that can’t be done, and try to put a Band-Aid on it using a trade protection as a model, when actually the problem has nothing to do with trade. That’s what it seems to me like. Is that an ungenerous interpretation of all the models you’re walking us through?

PANAGARIYA: Yes. Well, a lot of truth to that. A lot of truth to that it’s an easy fix, because even subsidy which would do less harm—meaning production subsidy suddenly will do less harm than trade protection because trade protection is a combination of output subsidy and consumption tax.

Protection is not only giving subsidies to production, but it’s also taxing the consumer of the product which is subject to the tariff, so it’s worse. It will do less harm to do output subsidy, but output subsidies are hard to maintain, because politically people see through it that you’re taking out of my pocket and giving it to somebody else.

Trade protection is very easy because you think that you are imposing it on the foreigners. Nobody sees that my own consumers actually are paying most of this tariff revenue that you are collecting. You just give the impression that it is the foreigners, we are collecting it from them. They are the target, so politically they’re a very easy target actually. The thing about trade taxes or tariffs is that they generate revenues.

The finance ministry generally is very happy that they’re getting these revenues, and so they see this as a good fix. The bigger damage that happens, of course, is that in the end, that delays the development of financial markets. If you take the argument, financial markets will always remain imperfect, and therefore, I have to solve these problems through these alternative measures, you are delaying the development of the financial markets. Fair enough that financial markets which maybe even in America or Europe, they’re imperfect, but they are far, far less imperfect. They deliver a hell of a lot better outcomes than in most other countries, ultimately the financial markets.

There is an argument that you shouldn’t delay development of the financial markets by substituting all of these other interventions, especially the trade intervention, which really is ultimately not fixing the problem at all. It’s neither fixing the problem, and at the same time, it is delaying the development of your financial markets.

RAJAGOPALAN: To take an example which has nothing to do with foreign trade or financial markets, but will illustrate the point: think about farmers in India. For a very long time, a very big problem with farmers is that the markets are very local. And because we don’t have good cold storage, good infrastructure, large agricultural markets, transportation, infrastructure access to large agricultural markets, farmers don’t get a great price for their produce.

They get a very low price. Therefore, they need all kinds of subsidies. We need to give them free electricity, we need to give them fertilizers, we need to give them all these other loan subsidies or loan waivers, worse still. What really needs to happen for the farmer is you need to build the roads. You need to build the infrastructure for the agricultural market.

You need to build the dispute resolution system. You need to create a proper system so that spot contracts and futures contracts can actually be enforced. Those are the things that are going to improve the prices that farmers get. It’s very difficult to build roads. It’s difficult to bring in a new dispute resolution system, or create a large marketplace.

What we instead do is we give them loan waivers, or subsidized fertilizer, or free electricity, or MSP or something like that. What they really need is the road. In fact, there’s some great literature showing that the moment there is increase in road penetration and railway penetration, basically access to transport, agricultural productivity starts increasing.

This is an age-old insight. We see good evidence for this in India, but to me the trade intervention argument because of imperfect capital markets, though that’s a completely different part of the economy, it seems very similar to this kind of logic. Except now we’re talking about foreigners. It’s really a matter of infrastructure development.

On this, my question is, how do countries that have open and free trade, how do they deal with imperfect capital markets? They must also have this problem to some extent, right?

PANAGARIYA: Yes, but look, certainly in the advanced countries, still the capital markets are much better developed.

RAJAGOPALAN: Something like South Korea, Arvind. This was not an advanced country when it opened up its trade.

PANAGARIYA: Yes.

RAJAGOPALAN: It must have had imperfect financial sector and capital markets. How did a country like Korea deal with this problem without protectionism? I guess that’s my question. Is there a way or a model or examples of developing countries that might have dealt with this?

PANAGARIYA: There were certainly some credit subsidies in Korea, but that I don’t think was the dominant instrument. Some interventions happen, and this is where a whole business on targeting and industrial targeting, et cetera, comes in, but that’s not the dominant story at the end of the day. To me, it seems, the South Korean story is really very much taking advantage of the global markets.

Very early on, by the way, 1960s onward, because Korea really devalued big time the domestic currency, which very quickly generated response of the manufacturing exports. 1960 to 1964, manufacturing exports are expanding at the rate of something like 50, 60% a year. It’s a very slow base.

I admit it’s coming from a very low base, so it looks a little exaggerated, but the point is that their structure of the exports dramatically changes from agricultural to manufacturing. Manufacturing initially being something—20, 30% in 1960, by 1964, they’ve crossed more than 50—more than half of the exports come from manufacturing. That, I think, convinced the leadership that this export business could deliver.

By 1965, Park Chung-hee was very much onto the ball game on this. He very much started having monthly meetings, or bimonthly, once in two months or something. He’d meet all the export interests, figure out what the bottlenecks are. Korea is number one small, but at the time, income levels being what they were economically, it was even smaller, so it was not very hard to find out who the big exporters were.

They didn’t have the small-scale industry business, so it’s the larger firms which export. It was easy to collect and then find out from them what the bottlenecks were and all. That is how I understand the Korean story. Certain credit subsidies existed at that time, but I don’t think that was a big part of their story.

RAJAGOPALAN: Their side of the story is they are growing their engagement with the global economy, and simultaneously strengthening their financial sector.

PANAGARIYA: Exactly.

RAJAGOPALAN: That’s the story.

PANAGARIYA: This happens, yes. They didn’t scuttle it by nationalizing the banking sector, for example. Also, there is the capability of government issue. These were capable governments consulting the experts, and all, as well, but largely making good commonsensical decisions.

In the China story, where there is much more presence of the public sector, it’s similar when at least in the sense of capability. Deng Xiaoping really was a capable leader. Again, this leadership issue, of course, does play a very critical role. At least, in my thinking it does.

RAJAGOPALAN: Basically, the state also making a credible commitment that they’ll pursue free trade and focus on removing infrastructural bottlenecks as opposed to pandering to specific interest groups and lobbies, and things like that.

PANAGARIYA: Definitely.

RAJAGOPALAN: That seems to be a part of the story somewhere, right?

It’s easy to give in to the interests of exporters. Not all exporters ask for better capital markets. Usually, they just ask for a very specific subsidy for their particular good. I found that quite fascinating.

PANAGARIYA: For them, the ability to the import the inputs is very important. Those are the bottlenecks that Park Chung-hee, the President of Korea in those days, tried to remove. Are they getting their inputs in time? Are they getting these duty-free?

Even though they had tariffs, they did ensure that the exporters really didn’t have to pay those tariffs, or they were reimbursed rather quickly. Often, they would actually reimburse them a bit more than what they paid, so there may have been some element of subsidy. Since they also had some import protection, that small subsidy was only neutralization of the import tariffs that existed in the first place.

Let the financial sector, financial markets evolve and develop alongside. In India, what we have done is we nationalize the banks. Innovation has been very limited. Much of innovation only began happening after ’91 when we began to allow private banks again in a big way.

RAJAGOPALAN: And foreign collaborations.

PANAGARIYA: Yes, foreign banks to come in as well. That’s when some innovation began. Then, of course, information technology came through also. Even then, you can see that the private banks are way ahead of the public sector banks.

RAJAGOPALAN: We are treating public sector banks like any other public sector enterprise, as basically a jobs program for the middle class that is formally unionized. That’s what the Indian government banking sector has become. It’s become a jobs program for the elite, as opposed to its actual job, which is to make greater financial resources and penetration available. This is one of the original sins of Indian socialism, and of course, Mrs. Gandhi nationalizing banks in a really big way in ’69.

PANAGARIYA: Right. Also, you see, the problem is that the leadership of the public sector banks is always under the fear of vigilance.

RAJAGOPALAN: Yes. They never innovate.

PANAGARIYA: They will never innovate, exactly. It impedes the innovation.

RAJAGOPALAN: Yes. You don’t get promoted and get big bonuses for the good bets you make, and you get pulled up and punished and taken to the CVC and      CBI for even one bad bet. 0% bad loans or error is bad for any bank. Right?

PANAGARIYA: Yes. Absolutely.

RAJAGOPALAN: It means it’s not making appropriate number of bets. Now, we err on the other side, which is we have very, very high number of nonperforming assets.

Your point is well taken that this is not the capital markets and a weak financial sector. It’s not clear that the obvious solution is protectionism. It could be some other intervention or state capacity development in those areas, but a protectionist approach is clearly not the way forward.

PANAGARIYA: Right. The big thing you need is the expertise to evaluate projects. The bankers have to have the ability to evaluate the projects properly. That’s what you need, because ultimately, you want to invest in high productivity projects. If you don’t have that ability, then otherwise, you impede that ability even if you have the ability and big decisions are political through this crony lending.

RAJAGOPALAN: But Arvind, that’s also endogenous to the ownership, right?

PANAGARIYA: Yes, it is.

RAJAGOPALAN: The ability to pick good projects.

PANAGARIYA: Yes, that comes out right. In our case, it comes out very clearly because the private sector banks don’t have that kind of NPA problem. They did not take a rupee of the taxpayer money to clean up their NPAs. They did their own clean up. Whereas in the public sector, NPAs have been massive, and they had to be cleaned up through taxpayer money.

RAJAGOPALAN: When all the big scams like Vijay Mallya and Nirav Modi and these sorts of things happen, I don’t remember which private sector bank it was, I think it was HDFC. They were asked, “Did they approach you for a loan?” They said, “Yes. And we offered them a cup of tea and said, ‘Thank you.’”

That’s basically it. It’s not like there’s an additional, some kind of special superpower that the private banks have which the public sector employees don’t have. You put the same employees in a different bank, and they will perform just so much better in picking winners and losers. I think the ownership is what is driving—

PANAGARIYA: Very much, very much so. Absolutely. It’s not that these scandals don’t happen in private sector banks. We did have a bit of it in ICICI and all, but you see far less.

RAJAGOPALAN: Far fewer bad loans.

PANAGARIYA: Ultimately, it goes back to the commercial pressure. The private banks have the commercial pressure and public sector banks don’t.

RAJAGOPALAN: They’re fully subsidized, basically.

PANAGARIYA: That’s one big factor, of course, but then there are also these other issues of constraints of vigilance, which constrain the ability of the leadership. Even good leadership really can go only so far in these banks because of the vigilance problems.

RAJAGOPALAN: Also a lot of the fiscal spending is now done through the banking system. This is, of course Viral Acharya, Urjit Patel, they’ve talked about how very high fiscal spending, and what they call the theory of fiscal dominance, is now entering the banking sector, creating these spillover problems and entering monetary policy.

That’s one side, but the other is also just plain, simple political pressure. This is, to some extent, missing or reduced in private banks. You don’t have political leaders putting pressure on bureaucrats or the boards of Punjab National Bank or SBI or something like that, and saying that, “You need to give so and so a loan for this project.”

Typically, it’s because that person lent them money for the election campaign or some other quid pro quo that has taken place on some other side of the economy. You just see less of this. It’s not to say that the private banks are perfect, and they can completely develop the financial capacity of a nation like India with no other inputs necessary. It’s just that we have a massive public sector system which has just failed to deliver.

PANAGARIYA: Absolutely.

Unemployment

RAJAGOPALAN: To go back to the arguments for protectionism. One argument, it would keep coming up for developing countries again and again. It started with the Asian countries. Today, people are saying the same thing about many African countries. This idea that open trade is bad for employment domestically. This is, again, an argument that doesn’t go away.

In India, of course, we have a particular demographic structure. We have had what many people have called jobless growth. We haven’t seen that massive employment in manufacturing that you saw in Taiwan or Korea or China and other places like that. This argument has come back in a new way in this century. It was slightly differently packaged in the previous century.

What is the relationship between free trade and unemployment, especially domestic employment or unemployment? Is there any merit that protectionism will increase employment within the country?

PANAGARIYA: There is not a whole lot to say on this one because I think the evidence is very clear. If you’re talking of aggregate employment, show me any country you know where there is a relationship between protection and aggregate employment. Surely, if you’re talking about any sectoral employment, then, of course, you give protection to auto industry and you will get higher employment in auto industry.

What about the jobs in the rest of the economy? There is just no evidence actually that the trade policy has an impact on the aggregate employment ultimately, which is what employment is about. India itself is a good example. 1991, starting with a massively opened up the economy. In fact, trade also expanded, quite dramatically. Exports were 7%, went up to 20%. Imports were at about 10%, went up to some 30% of the GDP. During this period, our own employment and unemployment surveys actually show hardly any impact. The unemployment rates, whether you look at either two or three different criteria that get used by any criterion, it doesn’t shift much at all.

It’s only the PLFS produced a slightly higher unemployment rate, which has been discussed and all. But during that period, if anything, we had increased protection, not reduced protection. If one were to draw at least some basic correlation, it’s the opposite. The correlation is moving in the opposite direction. I don’t want to draw that because I don’t think that a number that rises from about 2, 3% to 6% had anything to do with trade policy. I don’t think so.

In fact, if you look at some of the other countries—South Korea, Taiwan, China, successful countries—quite the contrary. What you see is massive employment opportunities created by this expansion of trade. Therefore, the industry and services are drawing workers out of agriculture in a big way and in the face of very, very rapidly rising real wages. In Korea, in 1965 onward, you see these wages are rising at the rate of some 9 to 10% a year through the 1980s. Every one of these decades, you see this 9 to 10% growth in the real wages. At the same time, it is absorbing a very large volume of the workforce.

Clearly, you know that it’s a labor scarce economy. It’s turning into labor scarce economy, which is why the real wages are rising. Korean case is not like Arthur Lewis type of disguised unemployment model even. The workers have to be paid higher and higher wages actually as they’re being drawn out of agriculture. Evidence on the aggregate is simply not supportive.

Now, in fact, if we then go back and see the Indian problem, which is the problem of underemployment—I have always said this, underemployment in agriculture and underemployment in industry and services, meaning, that two workers or three workers are doing a job which can be easily done by one worker. That, of course, means that output per worker is incredibly low. Low productivity employment, that’s under employment.

There, actually, trade will help to go the other way. It will reduce this underemployment and raise the productivity. I think politically it’s easy to make that kind of argument perhaps. Everybody who is looking for a job that identifies with it, but it really doesn’t have anything to do with—an even more stark, truly stark example is our days of ’50s, ’60s and ’70s.

RAJAGOPALAN: Closed economy.

PANAGARIYA: When we were a completely closed economy. Look how many jobs we created. Hardly any jobs got created. For my generation, even movies that would come out were all about—

RAJAGOPALAN: —Berozgar, Naujawan.

PANAGARIYA: Yes. These young men seeking jobs and all. You know that often for any single job, large number of people arriving, and then the hero making the sacrifice that “mein ye job le leta to tum berozgar reh jate” and all. Let go off the opportunity. Progressively, in fact, in the clerical jobs, there was a generation when high school was enough for clerical jobs. Then bachelor’s degree and then master’s degree became the norm.

RAJAGOPALAN: Now you have Ph.D.s applying for there’s a peon job in the railways and two lakh [200,000] people applying, and things like that.

PANAGARIYA: Exactly. That’s a reflection of it’s still very limited jobs that we have created in the private sector. The best jobs are you get 18,000 rupees per year, job guarantee for life, pension assured, you’ve got health benefits assured, in a few years, you also get housing. In the private sector, a lecturer will probably get just 18,000 rupees with no job security, no health benefits covered, no housing, nothing, and really have to work very, very hard. Whereas as a Class IV employee in the government of India, the so-called multitasking staff, half the time, they are totally free. Not very much to do.

RAJAGOPALAN: Here, I understand the argument in the aggregate. We understand that free trade will basically increase the size of the pie. As the economy grows, there will be more jobs and so on. Now, I want to dig in a little bit into the sectoral question. This has proven true more for developed countries than developing countries like India.

You have developed countries where you had large manufacturing. As free trade opened up, places like China and Taiwan, Bangladesh, they started developing a comparative advantage in those sectors. Those industries, factories went to those countries, and therefore the jobs went to those countries. You had some sectoral churn.

If 50 years ago, you were working in a shoe factory in the United States, for instance, today that job simply doesn’t exist. Now, the question in developed countries is not so much protectionism, but how do we make sure that people who are employed in one sector where they are no longer as productive, can be reintegrated into the rest of the economy in some other sector?

This is also related to arguments about automation, all the technology, artificial intelligence, all these new sorts of things which are coming in and taking away traditional jobs. Trade is a very large part of the argument against this kind of sectoral mass unemployment in the short term, which really devastates families, cities, regions and so on. What is your view on this kind of very specific sectoral unemployment and free trade?

PANAGARIYA: Here, in the developed country context actually, even at the aggregate level, what you do observe, where this pressure, at least in the United States, seems to manifest more is in terms of the wages. We have this wage inequality debate. The skilled wages to unskilled wage ratio has been rising continuously. Some exceptional periods have been there, but broad trend, particularly if you really take from late 1970s to 2000 or something, you see that the so-called skilled premium has been on the rise. It’s at the bottom, which is where you really care for people. That’s where the poor people are. There, the wages have not risen. In real terms, perhaps they’ve risen marginally. There is a difference oftentimes that people conflate the relative wages to real wages. That distinction should be made, but at least the evidence I saw from some work by Rob Feenstra and all, the real wages were not negatively impacted, but the wage gap is very clear.

The question simply is that, can protection solve that? There’s no evidence really that protection will successfully solve that problem because a lot of this, at least the view of trade economists, has been that the very significant part of this shift in favor of the skilled labor in the wages has been driven by the shift in technology.

It is the production processes that have become progressively more and more skill intensive. They have progressively increased the demand for skilled labor relative to the unskilled labor. That is the main driver of the wage premium or skilled premium in the wages. Obviously, you don’t want to stop the technological change. Ultimately, it’s for the good of everybody, but it does have that effect of wages at the bottom.

There is a lot of a pain, there is no doubt, at the bottom in the United States, Europe and so forth. Economists have traditionally always argued for good social safety nets. For political reasons, maybe they have to be trade related to some degree, but I think that should be more generalized because trade is not the only policy that is at play.

There are other policies at play as well, and technology is at play as well. So generalized social safety nets, now, that’s easier said than done. Obviously, there’s still a lot of things, a lot of issues that remain, but one can blame trade. The simple example I give is, if trade had not played its big role, look what would have happened to the iPhone.

At the end of the day, the very fact that everybody can afford—almost in the developed countries and the developed world—at least some version of iPhone, is because the Chinese manufacturing has brought prices so far down, and the massive scale. If production of iPhone had been actually confined to the U.S., there’s no chance.

RAJAGOPALAN: I couldn’t afford an iPhone then.

PANAGARIYA: There will not be that many iPhones. But for Korea and China and Taiwan—it’s spreading a bit to Vietnam and maybe India also and all, but principally those three countries—but for them, mobile revolution across the world couldn’t have happened. It’s those countries that massively, at a very low cost, manufactured these mobile phones.

Even when we replace these feature phones by the smartphones, smartphones also very quickly fell in cost. That’s because of the manufacturing prowess of these countries. And just imagine how much good it has done to almost everybody. That gets lost in this.

RAJAGOPALAN: I think also more fundamentally, free trade in developing countries—to switch gears back—brings structural transformation. That’s fundamentally what we need. Structural transformation is not painless. It does require people to stop doing their traditional trade or agriculture or something they’ve been engaged in for a long period of time, and start doing something else, which is a higher productivity job.

On the one hand, we want the structural transformation because that’s the path to higher GDP, higher GDP per capita, better wages and a growth trajectory. On the other hand, it’s not painless. Now, since that’s the prescription for developing countries, it would be weird if that’s not also the prescription of growth for developed countries.

The margins on which they can afford to ease their distress by providing unemployment, welfare schemes, or universal basic income or something else—of course, that’s a different magnitude. The prescription for growth and productivity can’t be that different across both. Right?

PANAGARIYA: That is true. As you can see, that’s a choice most developed countries have made.

RAJAGOPALAN: They have made it, yes.

PANAGARIYA: They’re not generally impeded. There are some protectionist noises here and there, but largely the markets have remained open. If you look at the flow of trade, that flow has not been impeded. If you look at the global exports of either goods or services, boy, that has continued to expand dramatically.

Today, merchandise exports are in the range of $18, $19 trillion, and services is another $6 trillion. Together, they are more than a quarter of the global GDP. Then they have chosen their paths. Look, advanced economies certainly are in a much better position to build social safety nets. Tax revenues are quite substantial.

RAJAGOPALAN: Even to build the social security net, you need free trade and economic growth because that’s what gets you the higher revenue. At some point, there are very few options if the goal is prosperity for large numbers of people.

PANAGARIYA: All you can say is that if the developed countries resort to protectionism, the cost of it to them might be a little smaller than what it is to developing countries.

RAJAGOPALAN: Absolutely.

PANAGARIYA: I think developing countries pay a huge cost. We know the Indian experience, particularly of how much we paid for almost 50 years.

RAJAGOPALAN: It’s almost a luxury that only a few countries can afford, and even they can’t afford it for too long, right?

PANAGARIYA: Yes.

RAJAGOPALAN: Protectionism as a luxury good is an interesting way to think about it. This is a good point to set up our next conversation. I think you’ve been very patient with us in explaining and laying the ground for arguments in favor of free trade, and walking us through the merits, mostly demerits, of arguments in favor of protectionism. Hopefully, we can continue this conversation in that direction.

PANAGARIYA: We shall do that.

RAJAGOPALAN: Thank you so much as always, Arvind.

PANAGARIYA: Okay, thank you.