Transcript of a Book Forum
"One Economics, Many Recipes: Globalization, Institutions, and Economy Growth"

Washington, D.C.
Thursday, November 29, 2007

Featured Speaker:
Dani Rodrik, Harvard University

Discussants:
Robert Davis, The Wall Street Journal
Arvind Subramanian, Center for Global Development
Roberto Zagha, The World Bank
View a Webcast of the Book Forum

MR. LOUNGANI: Good afternoon, everyone. Trevor Manuel, the South African Finance Minister, once said of the antiglobalizers, "I know what they are against, but what are they for?" And sometimes people have said the same thing about Dani Rodrik: we know what he is against, but what is he for. He is skeptical that one can show the effects of trade on growth. He is skeptical about globalization—has it gone too far?—and so on. But people have often said: if Dani were put in the shoes of advising policymakers, what would he do, what would he be for? I think that this book gives you the answer. It is a great read and I hope you will go out and order it. But while you are waiting for it to come in the mail, I also want to do a little promotion for "Finance and Development" which is available on the Web. The March 2006 issue has a summary article by Dani Rodrik and others on the growth diagnostics approach that he advocates in the book.

Let me just tell you about the other speakers. Bob Davis is a Pulitzer Prize-winning reporter at The Wall Street Journal. He is the author with David Wessel of a 1998 book called "Prosperity" and he has written very perceptively about the IMF and about globalization in many recent articles. Arvind Subramanian is a Senior Fellow at the Peterson Institute and also at the Center for Global Development. He is also Senior Research Professor at Johns Hopkins. He was at the IMF, most recently as Assistant Director in the Research Department, until mid-2007. Roberto Zagha is Senior Adviser in the World Bank Vice Presidency in charge of economic policies and poverty-reduction strategies. He is the Secretary to the Commission on Growth and Development, a commission chaired by Nobel Laureate Michael Spence that is doing a 2-year study which Roberto will tell you about.

The size of the audience turnout tells me that you do not need an introduction to who Dani is, so I will just turn the floor over to him. Dani, please.

TALK BY DANI RODRIK

MR. RODRIK: Thank you very much, Prakash, for the introduction, and in particular for organizing this event. Over the years I have been quite surprised by the assistance I have received from the IMF both in generating and also disseminating my ideas. My first book, "Has Globalization Gone too Far?", which was published in 1997 and which Prakash referred to, was actually written when I was a Visiting Scholar in the IMF's Research Department, not exactly the place that you would think that the ideas in that book would have necessarily originated. So there are interesting ways in which the IMF has been subversive and it gets too much criticism for not being open to different ideas--or maybe for surreptitiously promoting them in my case.

(Laughter)

I want to thank all of you for coming to listen to this. I also want to thank my discussants or my commentators, three of my favorite people in Washington, so it is good to see them and I am very happy that they are doing this.

I guess my job is to say enough about the book to convince you that it is an interesting read and that you should actually read the thing in its entirety, but not say so much that you feel that you've now got it and therefore you do not need to read the whole thing.

(Laughter)

So I am just going to leave many of the arguments incomplete. What a nice way to begin a talk, right? Don't you wish you could do that all the time?

(Laughter)

I will not of course tell you which arguments are incomplete because I have not figured them out, and which are incomplete because I think at least I know how to complete them.

But let me begin by giving you a sense of what is the context that motivates the book. I will draw a contrast between the fact that, in practice, economic development is really working, at least in the most populous parts of the world. We know this because the World Bank's studies on poverty around the world show us that there are 400 million fewer people who live in extreme poverty since the early-1980s. The amount of growth that has been achieved in East Asia, Southeast Asia, India, and many other parts, of course, to some extent in sub-Saharan Africa as well, is quite striking. The facts on the ground are that vast parts of the world economy have been doing quite well in the last quarter-century or so, with a few important exceptions of course.

But the contrast I think is with what I would say is not working, which is development policy. Because the successes have happened not necessarily in those parts of the world where there is a good fit with the ideas that have come out of this town or, for that matter, the town that I live in, Cambridge, Massachusetts. So our North American, Cambridge, Massachusetts, Washington, D.C., conception of what good policy is, what is a good reform sequence, is not exactly the one that has produced those development successes. In fact, one of the points that I make in the book is that if a Martian were to descend and sort of tried to line up actual experience on the ground with how closely different countries followed or tried to follow the type of policies that were in vogue or the types of reforms that have been in vogue and supported by North American policy advisers in economics, that the Martian would probably find a very big disjuncture, not to say even a negative correlation. So one big puzzle, and this is how I would put it, is development is working while development policy in a real sense is not. Of course, what I mean by development policy is our North American sense of external advisers, multilaterals kinds of economic policy. Obviously, Chinese development policy has worked very well but it bears a very awkward resemblance to the kinds of recommendations that have come out of here.

Similarly, I think it is possible to say that globalization has been working. I think it is very hard to imagine how countries like China and India would have grown so rapidly if we did not have the kinds of opportunities today that world markets present, but there is also again a sense that globalization policies are not working. On the financial front I think there is now increasing recognition that financial liberalization policies, financial globalization policies, have not produced the kinds of results that were expected. And on the trade front of course we have this increasing backlash and increasing legitimacy problem that the world trading organizations and international trading system is facing, with many people in both rich countries and poor countries alike essentially expressing displeasure with the types of policies that we associate with globalization. So this is the context, I think this is the paradox that we need to face up to, and I think it is the paradox that motivates my attempt at trying to both provide a perspective on the last quarter-century or so of economic development as well as provide an agenda for how we think about the future—both in the area of development policy proper as well as in the area of thinking about globalization and institutions that govern globalization.

I am just going to list a few of the points that I tried to make in the book, and I will just go through this rather quickly. The first point is that the Washington Consensus was really not sound economics, that the Washington Consensus was a set of rules of thumb, whereas what the circumstance of economic policy reform in developing countries requires is really thinking of solid second-best reasoning. Instead the Washington Consensus is basically a set of rules of thumb that if you do 10 things that are on the agenda, then that is what your reform is, without really a sense of to what extent these reforms interact with either second-best problems arising from market imperfections or market distortions or market incompleteness, or other second-best interactions that result from administrative or institutional shortcomings. But I think when the history of the Washington Consensus is definitively written we will recognize that the Washington Consensus was actually not grounded in economics at all, but it was actually grounded on some loose political model of what it was possible to do in developing countries and why you should do those things. And this is an experience I have had repeatedly, that when I discuss policy reforms with people and ask why do you recommend this, the first line of the answer is always economic. And I say but economics does not actually say that, and then the second line may still be an economic answer. By the time I have questioned the reasoning or the rationale sufficiently, then eventually it goes to some political argument as to why this is the only thing that can be done or this is the thing that ought to be done because of some loose argument about how the politics of a country work.

I would like to underscore that the problem with the Washington Consensus was not that it just overlooked the institutional underpinnings of sound policy. So it is not that the Washington Consensus missed that you really need to get the institutionals right and not just the policies right, it was a much more fundamental one, that it overlooked the impossibility of removing all distorted margins rather institutional or otherwise simultaneously and that is the key thing, that is, that it was based on a first-best model of the economy, that if you could identify the eight, nine, or ten distortions and those were the main distortions, then you could attack them all simultaneously and that is what would get you the impact that you have. But of course, in reality we have many, many more distortions that are context-specific and you never have the capacity to do it all at once. Therefore you need a strategic approach rather than a laundry-list approach.

The second point that follows on the first is of course a realization that the sound practice of economics is one that is always context-specific. So when you are talking about or when you say the answer depends on the context you are not being heterodox, you are being orthodox. You are being orthodox because policy prescriptions that arise from the proper application of economic principles are always contingent. They are contingent because opportunities and constraints that policymakers and economies face depend on initial conditions that have been shaped by history, that are shaped by endowments, that are shaped by administrative capabilities and everything else. So in that context, one of the big parts of the book's important argument is that you need to think of the practice of development policy, if you want to call it the art of development policy, as that of trying to undertake an exercise of diagnostics where you are trying to identify areas of reform where you get the biggest bang for the buck. Secondly, once the diagnostic step is accomplished, then to undertake a process of policy design that economizes on administrative resources that are required to tackle the identified constraint while taking into account the second-best complications that might arise with the exercise of what is effectively a partial or sequential approach. That is why when we go back and look at the history of successful development, successful economic growth, what we find is that the reality corresponds not to a case where governments all of a sudden go with a big bang and do all kinds of things together and then the economy grows happily thereafter. Quite to the contrary, what we observe in practice is that successful reform efforts tend to be selective, they tend to be sequential, and often they take unorthodox forms precisely because of the need to design policy approaches that economize on scarce resources at home while taking advantage of, again, context-specific capabilities.

The third point that I underline in the book is to go back essentially to the very origins of thinking about development and to underscore a very old point about development, that before there was neoclassical economics there was the study of economic development. And the early thinkers of development understood that what development is about and is different from other areas of economics is that economic development is about structural change, it is about structural transformation and that markets generically underprovide the incentives for structural transformation. One of the important hypotheses of the book is that the market failures associated with the creation and expansion of nontraditional economic activities are often as important and sometimes perhaps even more important in the short-run as binding constraints as government failures on which the original Washington Consensus and now increasingly the governance agenda focuses. So it may well be that in some setting, government-imposed entry regulations may lead to binding constraints, but in many other settings it could be that the constraints are not that entry regulations create rents and too little competition, it may well be in fact that there is too much competition and not enough rents for entrepreneurs and investors to be willing to invest in new nontraditional areas. So you need a way of keeping your mind open to the possibility that the binding constraint might be different, and also ways of developing diagnostic tools where you can try to figure out which is which and in what kind of setting, without taking a prior stand on saying that entry regulations or bureaucratic regulations or corruption are always the binding constraint.

I think that explains why in all successful countries that I know of, with possibly one unimportant exception, industrial policy has been the norm rather than the exception. It is just a fact that you cannot avoid facing when you are analyzing countries that have been successful.

If industrial policy is the norm rather than the exception in successful countries, it raises the question of how it should be done and part of the book deals with this issue. Rather than take this stand, which is the common one among economists, that industrial policy is too difficult and poor countries do not have the capacity to do it, I instead suggest ways in which institutions or policy arrangements can be designed to undertake industrial policies in ways that might prove beneficial even in environments with low institutional or administrative capacity. I stress a number of design principles. One of them is the need to combine sticks with carrots, the need to combine rents with discipline. We think that markets are important because they provide competition—that is important, that is the stick, that is the discipline. But structural change requires rents, that is the underlying Schumpetarian idea that you are not going to get innovation unless there is the prospect of rents. Too many sticks, too much discipline, too much competition undermines the rents which are required to induce investors and entrepreneurs to undertake structural transformation generating activities. So optimal industrial policy arrangements will neither maximize on eliminating the rents nor obviously try to maximize the rent at the risk of market discipline and disregard the need for risk.

The second fundamental idea with respect to the design of industrial policies is the need to combine autonomy with embeddedness. And there I am getting to the idea that industrial policies always are always undertaken in a context where the government knows too little and that is of course a valid criticism of industrial policy, but you design the institutions in the understanding that the government knows too little. Therefore you design it in a way that is going to enable the government to elicit information from the private sector about where the opportunities are, where the constraints are, and what are the best mechanisms for removing them. That is why successful industrial policy has always been a combination of some bureaucratic autonomous capacity to decide on what is good for the economy as a whole so that bureaucrats are not in the pockets of industrialists. On the other hand, they are not totally autonomous and do not totally hold industrialists and businessmen in the private sector either because that pure principal agent model of economic regulation minimizes the amount of information that can be elicited from the private sector. So you need to create institutions, deliberation councils, using industrial associations and other institutional mechanisms where there is a process of eliciting information from the private sector. That way of thinking about industrial policy suggests that the way to think about it is as a process and not as a set of priority sectors A, B, and C, and a set of policy instruments X, Y, and Z. So the question about industrial policy is not which sectors have you prioritized and which incentives are you using, it is have you created an institutional setting whereby the government is in touch with the private sector about opportunities and constraints. Secondly, that it has instruments at its disposal to respond quickly to identify the opportunities in ways that would remove those constraints. Thinking about industrial policy in that way makes clear that what is required for success is not picking winners, so that that bugaboo of industrial policy is really totally irrelevant. The real question is whether you can design a set of incentives and institutional arrangements that enable governments to let losers go because it is a given that successful industrial policy will end up making mistakes, that an industrial policy that only picked winners would be an unsuccessful one because it would be one that is not sufficiently adventurous. So the issue is whether you actually are able to turn your back on the losers enough of the time.

Finally with respect to globalization, I think about the implications of these kinds of ideas, and one-third of the book deals with thinking about global economic arrangements, where I make the argument that the best kind of globalization is not necessarily the one with the fewest economic controls at the border, but the one that most enables countries to pursue their developmental, and I think for the rich countries of the North, their social objectives. And an important part of the book of course is aimed at convincing you, the potential reader, that those two objectives do not always coincide. They do not coincide because transaction costs associated with national borders and sovereignty cannot really be eliminated even if tariffs and capital controls are. This is one of the big, important findings of the literature in international trading and international finance over the last decade, that all these capital controls and all these import tariffs have gone down, yet we find that there are huge transaction costs still associated with borders. I think the right way to think about it is that, even in the absence of former border controls, the jurisdictional and other discontinuities associated with national borders continue to impose important transaction costs even in the absence of explicit import controls or capital controls—which means that you are inherently in a second-best setting with respect to the practice of your international economic policies. That means that you still are in a position to generate the kinds of high-productivity employment at home for your own labor force. In a truly transaction-free world where your labor could go to the rich countries and where foreign capital could come in without any transaction costs, you would not have to worry about economic policy or industrial policy or development policy because the forces of convergence would be operating at full force. Because we live in a world that is divided up into different nation-states and that is going to be the reality for the vast majority of developing countries going forward with just a few exceptions—potentially of those in the immediate periphery of the European Union—the reality is that you need to have a domestic strategy, and that domestic strategy is going to require developmental policies and having the appropriate maneuvering space, the policy space, for their deployment. That is why I think a thin model of globalization, a sort of shallow type of economic integration to use my colleague Robert Lawrence's term, will actually work better for most countries than a thick model. One way to restate what this says in the context of current experience is that China has done much better under a GATT model of global economic arrangements than it is likely to do under a WTO model. And that is because a set of international disciplines did not constrain China from following the kind of heterodox and sequential reforms that it followed ever since the late-1970s and proved not only very useful for growth and poverty reduction alone but actually proved very good for generating trade and investment. This is one of the paradoxes of the postwar order that in fact by having less-restrictive international regimes you can actually end up having more trade and more foreign investment if in fact that less-restrictive regime is one that is more consistent with the pursuing of appropriate growth and developmental strategies, which of course will also generate increased trade and investment opportunities.

Let me go back to where I started. I think it is commonplace to discuss the backlash against both mainstream economics and against economic globalization. There is a never-ending stream of books being written on how mainstream economics has gotten it wrong, with Naomi Klein's "The Shock Doctrine" being one of the more better-known recent examples. The argument that I am really making is that the problem is not with mainstream economics per se, I think the problem has been with the inappropriate application of mainstream economics and with the specific globalization agenda we have chosen to pursue rather. So ultimately what I think the book is about is, and here I am putting myself in the role of the good guy, to save mainstream economics and globalization from their cheerleaders ...

(Laughter)

... in that it is not clear to me anymore that those cheerleaders are actually doing the best things to save mainstream economics and globalization from their supporters. Of course then more seriously and going forward, what I hope the book will contribute to is a new way of thinking about development and globalization policies which in many respects may seem heterodox but, once I again I would want to emphasize, in fact is better grounded in mainstream economics, than what you think is conventional wisdom and conventional approaches of the day - so, I am the true mainstream economist.

(Applause)

REMARKS BY BOB DAVIS

MR. DAVIS: Good afternoon. My name is Bob Davis and I am a reporter with The Wall Street Journal. I just wanted to say about Dani that I have admired his work for many years. He writes with incredible clarity. He writes in a way that is amazingly useful for journalists in the sense that you can follow what he says, and his work is not models that God knows journalists can never follow, and I thought that the book in particular is just a really good read. It distills a lot of what Dani has been thinking about and writing about for many years and puts it into a cohesive, logical whole so that it does not read like a set of individual essays but reads like a book. So there is my plug of Dani's book: It is really a fine read.

I want to take my journalist's prerogative and try to poke some holes in what he was saying and giving some food for thought and for further questions later on. One of the things Dani was saying was that he found that industrial policy was the norm in successfully developing countries, but I think the reality, which he also acknowledges, is it is the norm in all countries. Some use industrial policy more, some use industrial policy less, but they all use it in one way or another even if they say they do not. It is like the U.S. saying it is for a strong dollar when it is not for a strong dollar. So the question really is: how do you apply it well and where does it really work.

Although Dani does not like to talk about one-size-fits-all kinds of approaches, I think what he is looking to do is to try to figure out a way to spur entrepreneurship. That seems to come through a lot of the writing in the book. He is unconvinced that in many countries there are sufficient market incentives to spur entrepreneurship and so he sees a bigger role for government in doing that. The book also talks about the many times and many places that there are spurts in growth, that growth goes on, I forget what the exact number is, but something like 7 to 8 percent for 5 or 6 years, it is substantial, and then it peters out for one reason or another, and that there are many, many, many fewer examples of sustained growth. I would argue to that, although he does not make this argument quite explicitly, that if the creation of this entrepreneurial class, if one could do it, would be a big boost to that second goal which is incredibly important of sustaining these growth bursts, because presumably you would have a lobby that was arguing for the kinds of policies that would lead to the kinds of outcomes that you want. That is the case I would say for industrial policy.

I understand this is arguing economist to economist, and for an economist the idea of industrial policy as a positive thing as a first best policy is very controversial, so he makes the case where the examples where industrial policy worked really well in some of the Asian countries in particular, and Chile although it is not seen usually as a model for industrial policy is another one, and I think he may underplay the many examples of it failing essentially. I think one way to look at the wave of privatization in Latin America and in Central and Eastern Europe is that these are essentially the hollowed wrecks of industrial policy that went awry and that they were efforts to build a domestic economic by government-run electricity, power, and so on and so forth, and when it did not really work very well, what was left was selling it off.

What I would argue is something a little different than what Dani does. I have looked in my reporting career a lot at U.S. industrial policy. I realize that the U.S. is very different than developing countries, but I think there are lessons in how the U.S. industrial policy and when it worked well that would be useful in thinking about how to design an industrial policy scheme anywhere. I think the Japanese did some variation on the U.S. model or maybe it was vice versa, but essentially the most successful agency in the U.S. government that has pushed industrial policies is the part of the Pentagon called the Defense Advanced Research Projects Agency. They are the ones who spur the research that led to the Internet, that led to jet engines, and that led to a lot of the advances in rocketry and so on. The question is how did they do it. What they did not do is go out and say `we want to invent the Internet' because that sort of thing never works. Or they did not say that `we want to start a commercial airplane industry' because again that never works. When you look at the failures of American industrial policy it is when the government sets out to do that sort of thing, when it sets out to create a commercial nuclear breeder reactor, when it sets out to create a commercial supersonic transport. Under the Clinton administration, Al Gore led this big effort to create a superefficient automobile. Detroit took the money and it was kind of the payment to keep the administration off their backs when it came to CAFE standards.

I think industrial policy works best when the government is trying to buy something for itself. In the case of DARPA and in the case of the Internet for instance, what they were trying to do is they had a lot of big fast computers, they were trying to maximize the use of these computers, and so they were trying to figure out a way to link them together. With the jet engine, it was obvious, we were at war and wanted faster airplanes and that eventually led to a commercial industry, but it was not meant that way. It was a sort of serendipitous kind of event.

The reason that it is important for the government to act this way is because then the incentives naturally align themselves in the proper way. When the government is trying to buy something for itself it can better tell the difference between what is a useful kind of project to pursue than what is not. When Dani was talking about the sticks and carrots, what are the proper sticks to design into the system, how to cut off company A and company B when you know they are going to lobbying like hell to keep the kinds of projects that they have.

The other thing that I found that was interesting there too is essentially what they do is they will create a community of companies who are the project winners and that they get together and they are required as part of the contracts to share information. It creates on the one hand a fair amount of interplay between the companies which are also competing very fiercely to go on and to go past round one, to round two or round three and actually be the ones to develop whatever it is the government is looking for.

Also I think that Dani had some very useful requirements that companies involved in industrial policy regarding the sticks which were: invest only in new projects, sunset the projects, demonstrate the possibility of spillovers, but I think maybe he underplayed the basic idea which is to create competition. The U.S. does it as I am saying in this kind of way where there is DARPA giving out contracts for a dozen different companies and setting up this competition, the Japanese do it differently where they will have enlightened bureaucrats force competition on companies, and often times they make mistakes. There is the famous example of the Japanese trying to convince Honda not to get into the automobile industry, or get out of the automobile industry, because there was too much competition, but again it is I think the kind of defining characteristic.

One could say: it is the U.S. and Japan, they are rich countries, how much does that apply? Poor countries are not going to be looking for the newest and latest technology and that is also certainly true. But there are many, many things that poorer countries need. They need ports, they need to buy uniforms for their soldiers, they need to fix up their roads, they need to figure out how to adapt certain technologies. So instead of being a system where they are trying to create technologies, it might be a system where they are trying to adapt technologies, but nevertheless, I think the idea would be the same. The principle would be, first of all, what does the government need for itself and then the spillovers and serendipity will lead from that. Thanks very much.

(Applause)

REMARKS BY ARVIND SUBRAMANIAN

MR. SUBRAMANIAN: Let me say what a great pleasure it is to be here to speak on Dani's book. I am going to do three things. Praise Dani of course. And then interpret Dani for those who think they have not completely understood Dani, which I think probably is a null set, but I will try and do that just a same. Then I want to criticize Dani as well, so I hope to do all three things and do it in the quickest possible way so that I do not come between you and your conversation with Dani.

I am a big, big fan of Dani's. To be a co-author I think at least you have to be. I can say that one of the finest minds in the profession is Dani's. It is really a very subtle mind, and it is unconventional but not the kind of unconventional mind about which you say "interesting points" and then you move on. You have to stop and address what Dani's saying and you cannot move on. Many of you may not know this, or probably you do, but he is the first recipient of the Albert Hirschman Award, and one cannot think of a better recipient for that than Dani. So that is the praise, Dani.

But let me then go and interpret Dani for you. The heart of it all, which is the key problem for the Washington Consensus—let's call it WC for the moment—is the following. There is some new work being done at the Research Department by Alessandro Prati, Antonio Spilimbergo, Jonathan Ostry and other people. [It shows that if]you look at any kind of normal measure of reform as it were, either in terms of levels or change, Asia does not grab you as having "reformed" the most. If you take banking, if you take openness to capital, if you take trade, the picture is the same. What in some ways sets the stage for Dani to say he does, and he has said that is his motivation, is that [the notion that if] you reform more and you get more growth sits somewhat awkwardly. The way I like to put it is that the irresistible force of ideology meets the immovable force of inconvenient fact, reality does not play ball, and so you need someone like Dani to point that out.

Let me interpret Dani. Dani's claim I think is that the WC is a template, it is a kind of mindless template, and it is a biased template. At an analytical level the template that if you do more reforms and you get more growth is an intellectual straightjacket because it forces you into a somewhat incorrect reading of the growth history. Dani's contribution of course here is that successful growth experiences have been extremely varied and rich, and he may want to call it unorthodox, but let's just call it heterodox for the moment, if you look at what China did or what South Korea did, and the book has it also, so you should read it. A different way of reading the history has been I think extremely influential and certainly has made people sit up and ask new questions and fresh questions.

With China and India, the globalizers appropriated these two countries to their camp: China globalized more and therefore it grew more, India globalized more and therefore it grew more. I think Dani makes two very, very interesting remarks which again force you to stop and think. One is: how have they done it. I think one of the really nice things that Dani says in his book is that if China had been underperforming, believe me, all of us would have had a hundred reasons as to why China is underperforming: we would say they do not have a proper property rights system, they are not open to capital inflows, they allow too much industrial policy, et cetera, and similarly with India.

The other aspects about China and India which I think this reinterpretation is very important is because all these reform metrics that we can eventually use, under no circumstances can you say that India and China are more open or more globalized in some policy sense as any of these other countries, so in that sense, even though the globalizers appropriate these country experiences to their own, I think there is a sense in which their experience does sit awkwardly with the Washington Consensus.

The critique of the mindless template is that when you say do everything, [here are] rules of thumb, it leads to a kind of intellectual laziness and sloppiness and points to just do what the fad says, and that is why we had these fads in development policy prescribed from the outside which in the 1950s and 1960s praised capital fundamentalism, moved on to structural reform, moved on to human capital, moved on to institutions.

And Dani comes along with his positive contribution here to say: let's try and identify the binding constraints, do them sequentially, and so on. So it does force you to be much more sharp and hard in your thinking about what the constraints are. Of course, as an aside I cannot help but say that when you look at these other fads that the World Bank followed, growth diagnostics could well be the latest cottage industry and the fad that the World Bank is going off with, but that is something that I will leave to Roberto to talk about—whether or not it has become a cottage industry.

On the biased template I think again Dani's critique is that it is founded on this first best world with an inherent bias toward markets, and I think Dani's real contribution is to say that development is about doing different things and to stop doing old things, what you do is stop doing the status quo, the market generically or intrinsically militates against the pursuit of new things and that is why you need industrial policy, and of course Dani has some interesting variants on this new industrial policy.

This is where I would like to start my critique of Dani. I think Dani is utterly conventional in two respects. What is the first respect in which he is utterly conventional? Take this biased template, for example. Dani is an unconventional thinker, but in talking about this biased template of the Washington Consensus there is no room for conspiracy theories or multinationals and rich country governments pursuing their agendas using the Bretton Woods Institutions or the WTO as a front. So there is a real contrast between him and Paul Krugman, for example, who sees Cheney and dark forces behind every corner. Dani I think has too sunny and good-natured a disposition to go after conspiracy theories. One thing is that I think there are good plausible conspiracy theories one can say about the Washington Consensus. Jagdish Bhagwati, the world's biggest globalizer, thought that the IMF push for capital account convertibility was a kind of Wall Street/Treasury complex conspiracy. Many people thought that the WTO agenda on intellectual property was driven by rich country governments and strong corporate interests. Many countries also noticed why the World Bank through the AIDS crisis said nothing about intellectual property and the harm it could do to developing countries. So there is sense in which I think Dani has disappointed many of those to the left of him for not pointing out some of the interest-driven agenda of the Washington Consensus. So that the sense in which he kind of utterly conventional.

There is a second sense in which I think there is more in common with him and the Washington Consensus than he lets on, and the way I like to phrase it is the following. So where Dani differs from Jeff Sachs or the Washington Consensus is in terms of what needs to be done. Jeff will say throw money at the problem, the Washington Consensus says reforms, but Dani says, no, you have to go after different things, industrial policy and so on. But what is common to all of them is what I call the Nike premise that underlies Dani Rodrik's Mark II and the Washington Consensus which is "just do it," we know what it is, but government should go ahead and implement this. So a sense in which there are policy levers out there that can be pulled, with a fairly top-down view of the whole process.

I call this Mark II because I could bemoan with Simon and Garfunkel, "Where have you gone, Dani Rodrik Mark I?"—and let me explain what I mean by that. Dani Rodrik Mark I was someone who was very deeply interested in understanding why policies were the way they were and understanding the deeper determinants of outcomes. In my view, outcomes are surely a combination of long-term, slow-moving processes, the policy reforms and what I call the triggers. They are a combination of history and can-do hubris, a combination of persistence and possibility. We do not really know how many degrees of freedom we have and how much is actually just predetermined and I think that is the fascinating area of economics that is open. I think Dani falls too much for me on the side of possibility, hubris, and top-down can-do.

Let me give you an example. Can we really understand Latin America today without addressing inequality? Dani has a great paper with Alberto Alesina on the long-term consequences of inequality. Can we really understand Africa today without the conflict and the ethnic fragmentation determined by history and geography? Can we understand India's growth without investments made many, many years ago? South Korea and Japan, can we understand their growth without early land reforms about which Dani wrote a beautiful piece in The Journal of Economic Literature called "Understanding Policy Reforms"? That is at the analytical level.

Even at the prescriptive level, Dani has all these things about industrial policy, but there too I really would like to push Dani by saying: industrial policy, can we really implement it without knowing where it will succeed, how it will succeed, the circumstances which will allow it to succeed in some cases and not others? Dani, for example, has some very persuasive arguments about the importance of undervalued exchange rates, but can we really control undervalued exchange rates? Are they really a policy lever? Or are they in turn determined by some deeper political economy factors. For example, lots of Africa had overvalued exchange rates which many argue were driven by ethnic conflict and so on.

So Dani Rodrik's Mark II was a little bit what Prakash said: the urge to understand and critique gave way to this kind of itch to prescribe and put forward a positive agenda and I think a lot of insightful analysis and unconventional prescriptions have come up. But I would like him to marry them with Dani Rodrik's Mark I and then we will get Dani Rodrik Mark III which has all the unconventional aspects but with an effort to go behind and understand what are the circumstances under which some of the prescriptions can work, when they cannot work, and what other preconditions need to exist. So I cannot wait for Dani's next book.

(Applause)

REMARKS BY ROBERTO ZAGHA

MR. ZAGHA: Like my two predecessors, I am also a great admirer of Dani and an avid reader of his works and books and papers which I have been reading for the past 15 years and always with the same sense of discovery and admiration for his complexity and the depth of his thinking. So I was very pleased when Prakash asked me to come and participate in this session.

I thought I was supposed to discuss his book, but I was told that rather I should speak about the project I am engaged in, which is the Commission on Growth and Development. That was a surprise at the beginning, but then thinking further I thought there is a logic to this request, which is that in many respects Dani prepared the work of this commission, in some respects the commission itself is a consequence of the critiques and work that Dani has been doing with his colleagues over the years, and it is also a measure of his influence.

I am going to cover a few aspects of the commission. It is a 2-year exercise, very complex, involving a large number of people, a large number of academics, a large number of practitioners, and in the 7 to 10 minutes I have I am going to be kind of flying.

On its origin, I think it would be accurate to say that the origin started when Dani came to the spring conference at the World Bank in April 2004 and gave the keynote lecture at spring conference. He brought a range of novel ideas many of which you have heard today about how to interpret the growth experience of developing countries, novel ideas on growth strategies. The concept of binding constraints to growth, and several other novel and challenging ideas which had a lot of influence on the report that we prepared at the World Bank on what we learned from the reforms in the 1990s and the growth consequences of these reforms.

Then we had invited Professor Spence to give a lecture in 2005 at again the Spring Conference which is usually an annual event, and there were lots of similar echoes. We had asked Mike Spence to give this lecture first because he is a theoretician, second because he is a theoretical macroeconomist, he had never been involved in development, but had been interested in development and had been working in China and some developing countries quite extensively as he was very fascinated by the practice of policies and development policies. So again he brought a series of interesting concepts to this discussion and from there the idea of the commission was born.

Essentially the goals of the commission as it defines it are to address directly the top leadership in development countries, prime ministers, presidents, on what are the conditions, policies, strategies, most likely to lead to rapid and sustained economic growth with the ultimate goal of influencing economic growth rates, if ideas matter for what we do and the ultimate consequences of what policymakers do, that is certainly an important exercise. As part of the objective, the commission undertook to consult and assemble academic assessments of the existing sets of knowledge in a variety of areas relevant for economic growth. The intent is not to do new economic research, but to assimilate and understand the state of the art in different parts of the discipline.

In terms of the planned output of the commission, we expect to have two or three reports of different lengths addressed to different levels of leadership. We also expect to have a large number of published papers written by academics, some in collaboration with researchers at the World Bank, and we hope the commission sees itself finishing its activities in mid-2008.

We started a year-and-a-half ago brainstorming with economists and in March 2006 the commission prepared a statement of purpose on what it intended to do and why and motivations. We had several meetings in different places of the world. There was the process of commissioning, getting written, reviewing, and assimilating a large number of papers in different areas.

There are only two academics on the commission, Professor Solow and the Chair, Mike Spence, and the commission itself is populated with people who have experience in policies, people who have gone through the blood, sweat, and tears of implementing policies, reform, change, and reaping the rewards or the lack of rewards. There are some business leaders. The work of this commission is supported by a working group. These are the commission members. It is a long list of 21 persons. All of them except for the two academics have some form of policy experience coming from developing countries, Boediono and Kemal Dervis, Foxley who was the first Finance Minister of Chile after the return to democracy, and Kuczynski, the former Prime Minister of Peru. Interestingly enough, there are many people who have had a history at the World Bank or the IMF, giving a sense that those institutions have been a nursery for economic leaders. The funding comes from several sources.

Now we are one year and a few months into the process, we are reviewing the first draft of the report, and we have completed 11 of the 12 workshops that were planned and if you are interested, there is a website which details all of this. We have about 50 to 75 papers in all. And we have had several meetings, small meetings, large meetings, with the chairman [involved], in different parts of the world.

It is too early to tell you what the commission's final report will say or not say, but what is quite interesting is that in these workshops which bring together academics and some members of the commission, and sometimes large a number of the members of the commission, there are a number of issues that have emerged and a number of concepts. I am not going to go through all this list. I think that we will make it available on the website. Just for the benefit of time, I will just focus on a few.

The commission defined itself that it was interested in long, rapid, and sustained economic growth, defined arbitrarily as 7 percent over a period of 25 years or more. There are two interesting things about this concept. One, it is possible. Some countries have achieved such as China, Korea before China, and Japan before Korea. And second, it is possible, but at the same time it is very difficult as very few countries have achieved it. So there were a number of discussions whether or not this was the right benchmark or what is wrong with 5 percent, why not 6 percent, but it is an organizing device for our thinking. I think some of the discussions that were interesting here, which echo some of the points that Dani made, is that there is not a uniform model which can capture the evolution of an economy throughout its process of development, and at different phases of development we have to think in different models of reality.

I think contrary to what sometimes the Bank and also the IMF thinks, there is a great uncertainty about our understanding of economic growth, and this is in sharp contrast with the confidence with which both the Bank and the IMF give advice to developing countries. It was interesting to see that all the leaders of developing countries easily agree on this idea, at least most of them.

Other things that were discussed include: stabilize, privatize, liberalize. The Washington Consensus is not a growth strategy, at best it is a stabilization strategy. Professor Solow put it in words that were similar to Dani's, perhaps he had seen an earlier draft of your book, that while we are clear that the ingredients are common across different growth experiences, the recipes are very country specific.

I do not have enough time, and I think the main purpose of this meeting is to talk about Dani's book, so I will leave the rest aside, with just a last word that comes from a recent paper by Professor Solow which echoes I thought quite remarkably some of the issues raised:

"Some of the literature gives the impression that it is after all pretty easy to increase the long-run growth rate. Just reduce a tax on capital here or eliminate an inefficient regulation there, and the reward is fabulous, a higher growth rate forever, which is surely more valuable than any lingering bleeding-heart reservations about the policy itself. But in real life it is very hard to move the permanent growth rate; and when it happens, as perhaps in the USA in the later 1990s, the source can be a bit mysterious even after the fact. (Solow, 2007)

So that is some of it. There is much more to say, but there is not enough time today. I enjoyed Dani's book enormously, and like the others I recommend it. It is a very important read. Thank you.

(Applause)

QUESTION AND ANSWER SESSION

Mr. SUBRAMANIAN: Prakash has asked me to moderate, which was meant to be my original role, so without further ado, please, the floor is open now for questions. Raise your hand and identify yourself, and please ask brief questions so that as many people can get a chance to interact with Dani and the others if you want as possible.

MR. DAVOODI: Hamid Davoodi of the IMF. You mentioned that industrial policy is not about picking winners, but letting losers go. But this sort of misses the dynamic sense, because today's current losers and winners were picked a while ago, so it is hard to understand what that means. To give one example, do development banks work? Industrial policies often exist through development banks which are primarily funded from the budget.

MR. RODRIK: The distinction between the notion that industrial policy has to be able to pick winners in order to be a success is basically evaluated by asking whether every area in which there was either a particular investment made or a particular incentive provided whether those areas generated productive enterprises and firms. My point is that that is not the right benchmark for judging the success of industrial policy because you are necessarily operating in an imperfect information environment and it is a little bit like the way that pharmaceutical firms will invest in a whole number of different compounds and some of them will end up making very big money, the vast majority will end up being losers and at some point you keep investing in the winners and then you winnow out the losers. It is informationally obviously much less demanding to say that the benchmark for success is that you winnow out the losers because all that you then have is that can evaluate which are the winners and which are the losers. Some of that information will be financial but others might be through benchmarking or through the kind of competitive pressure that Bob Davis was talking about—which is the way that the Koreans and the Taiwanese figured it out is because the price of getting the incentive was to compete in world markets and so that sets the standard that only the successful eventually end up getting the resources.

In terms of development banks, what it means is that you can imagine many of them increasing the kinds of public venture fund roles that they play, which is to invest in a wider portfolio of nontraditional activities and to act as a lab that spurs new activities rather than supporting a small number of very large economic activities or assisting privatization, many of the new roles that they have taken on today. The criterion there is: are they going to have monitoring and information systems in place and the political capability to at some point cut their losses short in those areas which do not prove to be successful. So it was a very different way of thinking about it, and I would argue it is much less demanding and it does not rule out—if you say you cannot do industrial policy until you can convince me that you only pick winners, then I would say no country should do industrial policy. If instead you say you should not do industrial policy unless you have the system in place that will be able to let losers eventually go, a lot of countries can do that or can actually put systems in place that have a greater chance of doing that.

QUESTIONER: Just a couple of brief questions. Have you thought about what to do when some of the constraints have been engraved in bilateral trade agreements or etched into different types of regulations that have already been officialized and we cannot really move out of them? For instance, I will refer to the bank regulations which by bringing in minimum capital requirements that are strictly based on risk averseness has converted commercial banks from being financing entrepreneurs, the risk taking you want, into only financing the public sector and consumers. How do you move out of these types of constraints, prohibition on capital controls, intellectual property, how do you move out when you have built up these big walls?

MR. RODRIK: That is a great question. The short answer is that if the cost of getting out of these commitments is high which typically it is, then you need to take that as your new set of initial conditions and then you need to work around those things. I think one of the positive messages from the book and my work is that you can achieve a given end in a number of different ways. If you are interested in risk finance, you can do it in a number of different ways. For example, in Chile a lot of this has been done by Fundacion Chile which operates as a sort of public venture fund which is investing in a lot of areas, and some of them, including copper [?] of course has been extremely successful, effectively paying for all the other mistakes and all the other bad investments that they have made. Or you can use your tax systems. So the issue is that if you cannot get out of those commitments: obviously it is going to be harder because your freedom of movement is limited, but if you cannot do it through your commercial banking system, you might do it through other mechanisms and how will depend on what other assets you have in your portfolio in terms of capabilities.

But that general question I think is interesting. I think it highlights an important dilemma for a lot of Latin American countries in particular where implicitly I think that with a lot of the reforms you can read into them the idea that the binding constraint on growth was effectively achieving stability and conveying credibility and for some of them it was indeed the binding constraint so you can argue that in Argentina in 1991, the binding constraint was macroeconomic credibility. How do you do that? Here is a wonderful, totally heterodox policy, the convertibility law. Nobody in this town liked convertibility at the time. Only after Argentina started to grow, it came too late of course [because] by then the binding constraint had moved from credibility to competitiveness, and then you needed to do something else and unfortunately the convertibility law tied your hands. So I think that suggests also in looking forward you need to find ways of flexibility, and with a little bit of self-criticism here, precisely because the nature of the binding constraint will change over time if you are successful, if the constraint is binding and you remove it, something else will be the constraint as the economy grows. The question is whether what you have done in stage one may make it harder for you to remove the next constraint at stage two. But I am an optimist. I am like Arvind. I think that there is in fact room and there are always $100 bills lying around that governments can pick up.

QUESTIONER: Thank you. I am a regular reader of the Dani Rodrik blog, and in recent days you have been making a rather sharp distinction between social policy and growth policy which has surprised me somewhat. I was particularly surprised by a comment you made yesterday in which you argued that an increase in social expenditures will lead to Dutch disease effects which will slow export growth. It seemed to me that one would need to qualify that somewhat for a variety of reasons. In the work of Peter Lindert and others, one does not see any relationship between social spending and growth in OECD countries, presumably it operates through wages and there might be a labor supply response, and of course there is a human capital effect which seems to have been important in the East Asian story. So I was very surprised by the lack of qualification.

MR. RODRIK: One main thing is that with blogs one can say outrageous things without having to put in all the footnotes and the caveats. I have argued myself as well that in the long-run there is no determinate relationship between the size of the public sector or social spending and economic performance, but I think the Lindert type evidence you are talking about is really a very long-term relationship and I think the issue that you face when you are thinking about your growth strategy in the short-run is what is going to get you the growth. For a number of countries, increasing the profitability of tradables is a very important way in which you are going to get a growth boost in the short- to medium-term. That has been the experience of most if not all of the successful high-growth countries over the last three or four decades. Let's take that as a starting point for a second. You need to increase the profitability of tradables. How are you going to do that? The most direct way of doing it is by changing the relative prices in favor of tradables. That is what of course a real exchange rate depreciation means. But again as Arvind reminds us, a real exchange rate is not some policy lever. Even the central bank governor cannot just pull a policy lever called the real exchange rate and move it one way or the other. On the short-run, I think the signals that a central bank sends has a lot to do with the level of the real exchange rate, but in the medium- to long-term the only way you can get it to change in a more depreciated or undervalued direction is by having an increased level of savings relative to investment. That is at the economy level. How are you going to achieve that? It is going to be hard to do that without running a smaller deficit or a bigger surplus than you were planning to otherwise before this consideration came into your mind. So you cannot avoid that, and the question is do not spend on social services but realize that if you are going to spend more on social sectors without cutting back somewhere else and then your growth strategy on the other hand is saying that we need a more competitive real exchange rate, then you have just set up yourself with incompatible objectives.

MR. TRUMAN: Ted Truman from the Peterson Institute. You have half-answered my question but I have to ask it anyhow. I think anybody who has been close to policymaking would agree fundamentally with you that there is no one formula and that you always have to take context into account, and anyone who does not I think is wrong. Where I wonder whether you go overboard is the question of—if you think about the Washington Consensus or any other set of lists, if you want to put it that way, including the fact that the list is now probably twice as long as it was—whether these are reasonable things to think about when you are looking for your bottlenecks. The way you have partly answered the question is: is it reasonable to think about macroeconomic stabilization and you answered that, yes, it is reasonable to think about macroeconomic stabilization. It is reasonable to think about having a competitive exchange rate—and leaving aside for the moment exactly what that means since we economists do not agree on that, rather than obviously an overvalued exchange rate which disadvantages trade—and you have answered that question. If you think about it as a menu for where you look at bottlenecks, are you sure you are actually that much different than the people who have been trying to do this in a serious way around here or in Cambridge or elsewhere around the world?

MR. RODRIK: I think, Ted, that probably the smartest and most-effective employers of that framework effectively were doing something along [the lines of] what I am saying ought to be done. One interpretation is saying that what I am trying to do is generalize from the actual practice of what successful practitioners of that doctrine did, but I want to emphasize exactly so we understand what they did, which might differ from most people's reading and rendition of what that doctrine stood for—in two directions. One is that you actually need to have a way of picking what the priorities are. So it is not that all 10 things or all 30 things all have to be done at the same time. Macroeconomic stability may be the binding constraint at some point, but it might be the competitiveness of your tradables at another point, it might be lack of human capital at another, and it might be corruption at another point in time. The issue is recognizing that you will be much more impact by focusing on where the biggest impacts are going to be and with that you need to have a methodology for how you are going to figure out where it is. That is important because with the Washington Consensus and it's sort of present day variant. Nothing is built in that that suggests that you need to pick and if you force people they say: `of course in practice we are not doing everything.' And then you ask how are you picking, it ends up being: `we do whatever we can or we do whatever politics allows or we try to focus on where the biggest distortions might be or we benchmark countries and look where they are far away from cross-country benchmarks', all of which are economically unsound ways of doing this.

One is read the book. One is to get at the point that when you start thinking seriously that you will need to choose among those things, there is a whole new machinery that needs to come into play which was absent in the Washington Consensus and in the second-generation reforms and all of that. The second aspect of it is to realize that even when you are doing something on the Washington Consensus agenda, that the way that you are going to be doing it may be very different from how it was presented in the Washington Consensus. Let me give you the most blatant example which is trade liberalization. You read the Washington Consensus and you take the trade liberalization to heart. What does that mean to your advice to the country? Do you go and tell them to liberalize, you remove or lower import barriers, import tariff restrictions. Look at the most successful liberalizers, the most successful outward countries that have turned outward. None of them did that at the outset. That was not their strategy. They did it through special economic zones, they did it through export subsidies, or they did it through export processing. So there is a whole variety of mechanisms which eventually achieve the same end. You look at trade to GDP ratios that have increased. You look at outward orientation and it increased. You look at import discipline and it increased. But the particular form that the reform took even when it was pursuing a very Washington Consensus type of objective differed across countries and there I think is a very important lesson which the Washington Consensus and today's focus on best practices, institutional reform, doing business agenda and all that totally misses, which is that in fact the way that you will actually implement even an orthodox objective has to take into account highly the second-best context you are finding yourself in and that you need to open up your mind to the differences, that a certain institutional objective maps into an almost infinite variety of institutional designs and if you do not see how in fact you can achieve those kinds of things in very unusual ways and often will be much more effective. Play in your mind the history which has China in 1978 liberalizing its imports, privatizing its public enterprises. Imagine what kind of hyperinflation China was going to go into if it did that at the time, but that would have been the advice and we know because that is the advice Russia got.

MODERATOR: Time for one last question.

MR. SANDERS: Greg Sanders with CSIS's Defense Industrial Initiative Group. I was interested in Bob Davis's DARPA example. I was wondering if you have any thoughts on the government acquisition or DARPA model as a way of doing things.

MR. RODRIK: It is interesting, Bob Davis I think was the most heterodox amongst us all because he suggested a public procurement model for industrial policy for developing countries. I must say that even I had not gone that far given how highly held public procurement practices are in developing countries. I think developing countries are trying to get away from that model entirely and now to say that we are going to add an additional layer of issues having to do with industrial policy into your public procurement scheme, I think is an interesting idea. But in fact, this is the issue that South Africa is confronting right now in the sense that there is a big public investment drive that is going on and the question is can you use your publicly held railroads and transport companies and your ports, and your airline is going to be buying, can you use that opportunity to transform your local economy and increase domestic technological capabilities. It is very hard to do when we do not even know how to do the first job properly which is to source their inputs competitively and run that process relatively smoothly. So the answer is it depends. If you can do it and you build the safeguards in it, it is worth trying.

Mr. SUBRAMANIAN: I think we have to stop because there is a function upstairs which the Managing Director wants Fund staff to attend. It is pity we have to close, but on behalf of Prakash and the IMF and everyone, it is important to give Dani a hand, but it is more important to go out and buy his book.

(Applause)

* * * * *



IMF EXTERNAL RELATIONS DEPARTMENT

Public Affairs    Media Relations
E-mail: publicaffairs@imf.org E-mail: media@imf.org
Fax: 202-623-6220 Phone: 202-623-7100