Transcript

International Monetary Fund
Washington, D.C.
Friday, November 10, 2006

Moderator:
Raghuram G. Rajan, Economic Counselor and Director of Research, IMF

Panel Presentation:
Dani Rodrik, Professor of International Political Economy, John F. Kennedy School of Government, Harvard University
Kristin J. Forbes, Professor of International Management, Sloan School of Management, Massachusetts Institute of Technology
Eswar Prasad, Division Chief, Financial Studies Division, Research Department, IMF
Joaquim Levy, Vice President, Finance and Administration, Inter-American Development Bank

MR. RAJAN: (In progress) -- one of the panelists is missing, but he will be coming shortly. Let me start without too much in the way of preliminaries. The order of speakers will be Dani first, then Kristin Forbes, after which we will have Joaquim if he is here by then, and then Eswar Prasad.

Just a short introduction. Dani who of course needs no introduction, but I will proceed to give one anyway, is a professor of international political economy at the John F. Kennedy School of Government and he has published widely in the areas of international economics, economic development, and political economy. There is really a long bio, too long to read at this point. But let me just say that Danny is one of the most innovative and challenging thinkers in this area. He has a habit of throwing grenades amongst academics and I thought it would be wonderful to start with him so that we can have the grenades thrown and see what happens. Dani? Shall we say about 15 minutes for the initial presentation?

MR. RODRIK: Thanks, I guess, for the introduction. I am not sure. I did not bring my full ammunition kit with me. I am not sure any of this has any explosives in it. What I thought I would do was present some thoughts and ideas on these issues as to sort of how I make sense of some of the puzzling consequences or puzzling features associated with financial globalization (View the presentation). I think regardless of which way you come out on the desirability of opening up the capital account, how soon and how quickly and things like that, I think pretty much everyone agrees that the process of financial globalization has been one that has been full of surprises, way too many crises compared to what we would have expected, way too little by way of improved economic growth and improved investment in the low-income or emerging market economies.

What I am going to do here is, and this is just a quick outline of what I want to talk about, to start by a perspective that says let's not start from the finance side, let's not start from the financial markets or capital flow side, but let's actually pose the question of what it is about growth, what are the growth fundamentals, what drives growth.

There what I would argue is that there is an important distinction between looking at developing countries as being fundamentally saving or finance constrained versus looking at them as being fundamentally investment constrained, and I will try to make that precise. And I am going to argue that I think the evidence we have seen in the last 10 to 15 years adds to a long list of other evidence from elsewhere in the growth literature, so that most developing countries are really investment constrained and that these economies respond quite differently from capital flows. In particular, there are very important second-best interactions through the real exchange rate consequences of capital inflows. I will talk a little bit about some of the evidence which I am sure that other panelists will talk about as well and conclude with some quick comments on policy.

Here is the standard story. This is what we thought we were doing with financial globalization and capital account liberalization. The basic story is that developing countries are poor in savings. By the way, I am going to do all of this in simple supply and demand, it is possible to do it also much more rigorously as well, but this will be enough to drive the main points that I want to get across home.

The main story is that in a saving constrained economy where the optimal level of investment is that I Star, given the true opportunity costs R Star of world interest rates, instead what you have is a much lower level of domestic investment, much higher domestic interest rates if you close the capital account. Of course, what happens if you open the capital account is the domestic interest rate is going to go down and you have capital flow in and then you can finance a much bigger capital account and much bigger levels of investment domestically, and that is where you are going to get good growth. That is the basic story and the main story that is repeated. When people talk about the benefits of capital account liberalization, this is number one on the list and there is going to be number two about portfolio diversification, there is going to be number three about consumption smoothing, there might be number four about the improved practices and benefits of having foreign financial firms enter, but this is still the number one explanation.

The problem is I think as we have learned that this picture does not seem to really represent what has happened. We have not had the big investment response to financial globalization and capital account liberalization. In other words, we have not had the big investment come in. We have not had, therefore, the large increase in growth, and we have had nothing like the large current account deficits that this model really predicts, that this is not reflected.

Let's look at the challenge of growth now from a different perspective, the perspective of what I said is an investment constrained economy rather than an economy that is constrained on the saving or the financing of the investment side. Think about the main problem in a developing country being that there is a big wedge between the social and private profitability of investment. So what we had before as the marginal product of capital schedule is now the social marginal product of capital, what private investors perceive is in fact a much lower return to private investment even though the social returns are quite high. What I have done also is draw the investment demand schedule for the private sector.

You might ask where is that coming from. There are at least two main stories as to why you might have this big wedge between the social and private return to investment in developing countries. One, of course, that everybody is familiar with and is probably the leading explanation in the literature on growth is poor institutions. Poor institutions means that there is a wedge between the social and private return to investment because property rights are not properly protected, contract enforcement is not adequate, and there all those sorts of institutional anomalies that developing countries face.

There is a second large group which I think is of a different nature which is that the kind of investment that we care about here is going to be mostly investment in tradables and that is because investment in tradables has all kinds of special things that go with it. Tradables are the source of learning, externalities of technological dynamism, and all these growth models are based on the notion of various externalities or learning effects associated with tradables. That will once again suggest that there is a difference between the private return to investment in tradables and the social return. There is actually a link between those two features as well because you might think that even though poor institutions is something that affects profitability across the board in a low economy even as long as tradables, and in particular manufacturers and various activities within manufacturers are much more intensive in contract enforcement and property rights protection, that in fact poor institutions is going to disproportionately act as a wedge on tradables compared to nontradables. So whatever reason, we have this big wedge.

What happens in the closed economy? Once you have the closed capital account, now you have once again a low level of investment and we are far away from what the socially optimal level of investment I Star would be. What happens when now you open this investment constrained economy to capital flow? What you are going to have is now domestic interest falls to R Star. Will domestic investment rise? Yes, domestic investment will rise on impact because we have capital coming in.

The first thing to notice though is even when the investment rises, it is going to rise only part way toward the social optimum because the capital account opening is only helping relieve part of the problem, it is not relieving the main distortion in this economy which is on the investment demand side, not only the investment supply side.

Furthermore, this is not the end of the story because as long as capital is coming in, we are experiencing a capital inflow. What we are also going to have is a real exchange rate appreciation, and that is really a key thing in my story. Once you have a capital inflow you have real exchange rate appreciation. What does that mean for the private investment for tradables? That is going to reduce the incentive to private return for investing in tradables. So in terms of what that does to the private investment demand schedule is that it actually shifts it further in. What that capital inflow induced real appreciation does effectively is to worsen a preexisting distortion. It does not matter here whether the preexisting distortion was due to poor institutions or whether it was due to externalities of some kind or another in tradables. It does not matter precisely why the wedge between the social and private return to investment in tradables is, but as long as there is such a wedge and as long as the capital inflow induced real exchange rate appreciation results in a further reduction in the location of the private demand for investment, what you are going to have is an exacerbation of the initial problem in this lower-income economy.

The way that I have drawn this is that the investment demand schedule shifts insufficiently, that actually in this new equilibrium you have lower investment with an open capital account than you had previously. You still have a current account deficit, not because investment has increased, but it is because saving has fallen. So this is the typical syndrome of the capital inflows leading to a consumption increase and a reduction in saving, but actually in equilibrium, the investment, particularly investment in tradables, going down.

This equilibrium has a number of features, one is a fall in investment in response to the capital account liberalization, which gives us another result that we actually wanted which is that we do not get these huge current account deficits that we normally would have gotten in the case of a saving constrained economy. Another interesting thing here is that what it gives us is that at the margin it looks like capital controls would actually hurt. I am sure Kristin is going to talk about some of the micro level studies which she has contributed to. What the micro level studies seem to point to is that when you look at a cross-section of firms, you do find that firms that require finance seem to be hurt by capital controls. This picture is saying, yes, of course you are going to see this because there is a difference between the interest rate R Star that you face in an open capital account economy and the interest rate pointed to with the arrow that you would have faced otherwise, conditional on the overvalued real exchange rate. But the real comparison you want to make is not conditional on that overvalued real exchange rate, but actually compared to equilibria, one where the exchange rate is significantly more depreciated and one that it is not. So this is entirely consistent with the finding that across firms at the micro level, capital controls appear like they have adverse effects on investment, whereas once you factor in the general equilibrium repercussions of capital inflow to the real exchange rate and overvaluation you might get a picture that is very different. The key point is that in saving constrained economies, large current account deficits and capital inflows would be good, in investment constrained economies these are going to be bad to the extent that they exacerbate the investment side distortion.

Let me just go very quickly through some of the pieces and bits of evidence. This actually I am pleased to say comes through the work of our Chair and one of the panelists, and this is from a recent paper of theirs along with Arvind Subramanian which has the remarkable finding that countries that have done the best are not the ones who have run large current account deficits but in fact are the ones who have run current account surpluses. So if anything, there is a negative relationship, or I should say it is a positive relationship between your current account surplus and the growth rate that you experience. Once again, it is very hard to square an idea with a model of saving constrained economies, but not that hard to understand in the context of investment constrained economies.

Is there evidence that overvaluation of the type that I focused on is a key transmission mechanism, that it matters for growth? Here once again I am very pleased to cite some IMF work by Johnson, Ostry and Subramanian which looks at the relationship between overvaluation and economic growth, no surprise really here, or there should not be any surprise whenever there are countries that have had sustained overvaluation, this is a cross-sectional relationship, and have had lower growth rates than others.

What about the time series evidence? Here are just four countries that looks at how often growth spurts or growth accelerations take place in the context of a significant and sustained real exchange rate depreciation. In fact, I would say that a significant and sustained real exchange rate depreciation could be as close to any single thing that we can identify as a necessary condition for engineering growth acceleration. Here I have the case of China. Of course, there is growth acceleration coming in 1978. I am not sure you can see the labels there. The pink line is the real exchange rate. It is the measure of overvaluation of the real exchange rate and when it goes down it is becoming undervalued, and the blue or black line is the log of the per capita GDP, so when the slope picks up it is an increase in growth.

The first picture is China with a big acceleration in 1978 in growth in the context of a huge and very sustained real exchange rate depreciation, and Chile in 1984-1985 with a big acceleration in growth preceded by a very big exchange rate depreciation that was maintained into the 1990s. The Malaysian growth acceleration in 1986, the Turkish growth acceleration in 1980, they are all telling the same story that these things took place in the context of big real exchange rate depreciations. Do capital inflows appreciate due to the real exchange rate? Once again, I rely on work by Prasad, Rajan and Subramanian which shows that that is the case.

There is also the issue of real exchange volatility which I have not said something about, but, again, it is key that to the extent that openness in capital flows and financial globalization have created some volatility in actual and anticipated capital flows, and those are reflected in real exchange volatility, that seem to be costly to growth as well.

Finally, when we look at the emerging market economies with a few exceptions like China which has maintained closed capital accounts or at least controlled capital accounts, we do see that the overall investment rate, the best that we can say is that it has not increased, if anything, this is a group of emerging market economies that actually the investment effort has fallen, again, something that is very hard to understand in the context of saving constrained economies.

I am sure that I am running out of time. The policy take that I take from all of this is that in economies where growth is constrained by distortions on the investment demand side, that there are important complications that arise from the presence of an open capital account and the simplest thing to say is that if you have not already liberalized your capital account, there is no particular reason to rush for it if you are in this category of an investment demand constrained economy.

The other thing I would say is that the models still consider it a mark of progress. I read too many Article IV Reports from the IMF which treat movement toward capital account openness as something that countries ought to be praised for no matter how the official line might be, that this is not something that we are pushing, it is clear that the subliminal message is something that you ought to do.

It is much harder in the case of countries that have already liberalized their capital accounts, and this I think really fundamentally changes the nature of the policy gain because it is really very, very difficult to put back the genie in the bottle. I think here we need to do a variety of things. There are not really easy answers. We need to invest in what I would say is a portfolio of capital account management policies, to use a euphemistic term for capital controls, and there is a whole range of possibilities there. One of the things that I reproached the IMF for is that actually the IMF is not working on this, that the IMF staff is actually not working on how you can do intelligent capital account management.

I think the other issue of course is that you need to work on policies in order to keep the equilibrium level of the real exchange rate undervalued. Let me be clear here unless anybody has doubts about it. The relevant real exchange rate here is the price of tradables to nontradables, so this is not a beggar-thy- neighbor. So keeping your exchange rate competitive in this sense is not a beggar-thy- neighbor policy. There is confusion about that. I think there of course the way to keep your real exchange rate undervalued is to have high domestic savings, so there you obviously roles both for fiscal policy and in an increasing number of emerging markets where the issue is no longer fiscal imbalances, actually working on increasing private savings, and I think that is another thing.

Finally, the final message, the key as I look around at the growth experience of countries is often the real exchange rate. If you have the right real exchange rate for growth, a lot of things go much easier. If you do not have the right real exchange rate, the right exchange rate for growth, then it becomes really a much messier and much more difficult process to engineer growth and I think we need to look at the issue of capital account management and the openness of the capital account from this particular perspective, and I think that actually changes the way that you approach it. Thank you. I will stop here.

(Applause)

MR. RAJAN: Thank you, Dani. We should record this lest anyone say we do not give our critics full and free forum. Let us now move on to Kristin Forbes. Kristin is from the MIT Sloan School of Management, and she has served as a member of the White House Council of Economic Advisers from 2003 to 2005, in which she was the youngest person to ever hold such a position. I can testify to the many breakfasts I had with her at the White House mess. I had to invite her here to the canteen upstairs which was not as good, but we are trying to compensate a little bit by having you here and giving us part of your wisdom on capital accounts. Thank you.

MS. FORBES: Thank you very much. I know a number of you in this room have been here for 2 days in this dark, warm auditorium and have had a big lunch, so I thought I would start on a slightly lighter note.

Capital account liberalization, is it a vehicle for growth? (View the presentation). I call it a car for growth. When you think about cars in developing countries, when people get cars, there are a lot of advantages initially. It means you can get from point A to point B a lot faster. It opens up a whole new set of flexibility and opportunities for consumers and for individuals. It also opens up a whole new set of opportunities and flexibility for businesses, lots of new ways you can run a business and be more efficient. All of this will make individuals and companies more productive, it can raise growth in the economy. The bottom line is for all the pros and cons, it is hard to imagine a developed country without cars and without trucks.

Similarly, capital account liberalization. When countries open up their capital accounts, it will allow them to grow or can allow them to grow much faster, to get to a developed state much faster. It does open up more opportunities for consumers, allowing them to invest abroad and diversify their portfolios. It does open up tremendous new opportunities for companies, allowing them more access to financing, more access to knowledge and expertise from other countries. And again, it is hard to imagine a developed country without capital account liberalization.

But the one catch with both cars and with capital account liberalization is you worry about the crashes. Just as with cars, if you did not have cars and if you just walked everywhere, you would not have as severe crashes, and one of the big concerns with capital account liberalization is when you liberalize your capital account, are you opening up your economy to more severe crashes. I actually thought that that would be something Dani would focus on more, the crashes, the crises that often are associated with capital account liberalization.

One of the key points I would like to make is that I think focusing too much on some of the specific potential disadvantages of capital account liberalization such as the crashes or some of the arguments that Dani raised is not the right approach. We all agree that there are some potential risks from capital account liberalization, just as there are some potential risks from driving, the crashes. But we really need to look at capital account liberalization in a cost-benefit framework. We cannot just take one disadvantage or one benefit of capital account liberalization and draw conclusions based on that. Just like when we talk about buying cars, we are not just going to focus on the crashes or the pollution or just the benefits, you need to weigh the costs and benefits, and I think that is the framework we really need to bring to our analysis of capital account liberalization. So I am going to quickly talk about some of the costs, especially the relationship between capital controls and crashes, talk quickly about the benefits, and some of the less-than-compelling macroeconomic evidence on the benefits of capital account liberalization. I want to focus more on I think a new and exciting literature on the micro evidence on capital controls and capital account liberalization. I think that is where we are really starting to learn more about liberalization.

But then I also want to go to the next step and say how do we ensure the benefits outweigh the costs. There are certainly some examples in some countries where the benefits of capital account liberalization may be smaller, so what can we do to encourage countries to take the steps so that the benefits of globalization will outweigh the costs, and I will talk about a few thoughts that I think that that is actually an area where much more work needs to be done.

And one just quick caveat, the last part of the car joke, one of the problems with talking about capital account liberalization is that it incorporates so many different aspects of capital controls and liberalizing capital accounts, just like when you talk about cars you are talking about all sorts of different vehicles. And I am going to because of time constraints unfortunately gloss over a lot of important distinctions between different types of capital account liberalizations. For example, liberalizing inflows on capital versus outflows, liberalizing inflows of FDI versus inflows of portfolio investment, et cetera. So there really are lots of distinctions when we talk about capital account liberalization and those distinctions do matter, but unfortunately I am going to brush over a lot of those just in the interest of time.

First, what are the costs of capital account liberalization? There are four major costs that people have developed very nice models of and that people worry about. One is increasing vulnerability to crises. Another is that capital inflows can lead to bubbles and misallocation of resources. Liberalization gives you less freedom to use monetary policies, the standard Mandel -- framework. And capital account liberalization can cause pressure for currencies to appreciate or depreciate, so it really does matter if you liberalize restrictions on capital inflows or outflows, and this can have implications for exchange rate competitiveness as Dani was talking about.

But the one potential cost of capital account liberalization that seems to have gotten the most attention and at least in the policy world has made more governments back away from liberalization is this first the potential cost, that liberalizing can increase your vulnerability to crises. Here I think the evidence is actually much less compelling than one might think from just the anecdotal evidence that say during the Asian crisis, countries that liberalized such as Malaysia, the Philippines, Thailand and Korea, were more vulnerable to the Asian crisis, and countries that kept under capital controls such as India and China were less affected by the Asian crisis. This has led to a general sense that capital controls might provide a security against crises, but when you look at the data, look at the evidence, look at a broader subset of crises in countries, I think you become more convinced that capital controls cannot protect a country against a crisis if it does not have sound policies in fundamentals. There are numerous examples of countries that have had capital controls in place and had very severe crises.

India is again a great example. The capital controls people claim might have prevented it from being as affected by the Asian crisis, but India had even stricter capital controls in place in 1991 and had a severe crisis then. So capital controls cannot provide protection against crises. So that potential benefit of keeping capital controls on I think is overstated.

There has been some empirical work trying to find a relationship between capital account liberalization and crises, and this empirical work could actually find no evidence that capital controls prevent a crisis. In some cases, they have actually found the opposite, the countries with capital controls are more likely, not likely, to have crises. I think you cannot read too much into that because of a causality problem, that countries with capital controls are probably more vulnerable anyway, and that is why they have the controls in place. But the bottom line is the empirical evidence at least is not supporting that capital controls will prevent crises. Maybe it will make you a little less vulnerable on the margin, but the empirical evidence really is not there.

Again, even if capital controls can prevent crises, you need to get back to a cost-benefit framework. Even if capital account liberalization does have a cost of increased vulnerability to crises, what are the potential benefits you need to weigh that against.

So then that leads to what are some of the benefits of capital controls. As for potential benefits, I could do a long list. I think there are five that have gotten the most attention. Capital account liberalization can lead to increased capital inflows financing investment. Liberalization can provide improved technology in management techniques, especially capital inflows in the form of FDI, can facilitate the diversification of risk, reducing volatility in consumption and income. I believe the next speaker will talk about that. Capital account liberalization can increase market discipline leading to a more efficient allocation of capital. And all of the above can lead to higher productivity and growth. So potentially there are very substantial benefits of capital account liberalization. If you believe this, if you believe the models that lay all this out in detail, you should see higher growth, higher productivity with capital account liberalization.

But then the catch is, the macroeconomic evidence is not nearly as convincing as you would think. If you look at extensive series of papers looking at how capital account liberalization affects growth, some papers find a positive effect, some find no effect, some find mixed effects. So if you are an optimist, you might read all this literature and say there is no evidence capital account liberalization is bad for growth. But if you are a government talking about liberalizing your capital account, you probably want something stronger than that, and there is some weak evidence but it is not terribly convincing at the macro level. Recently there has been somewhat more success in a few studies that use finer measures of capital account liberalization with better measures of the types of controls in place. But the bottom line is, the macro evidence on the benefits of capital account liberalization is not as convincing as you would expect. There have been two excellent surveys of all this literature, one written by the next speaker or co-authored by the next speaker, so I will not get into the details of this macro empirical evidence and leave that to him, but there are imposing challenges to this cross-country work and I think we should not write off capital account liberalization just because some of the macro evidence is not as convincing as you would like.

Instead, I think this literature stands right now somewhere similar to where the literature on trade liberalization and growth was about a decade ago. Many economists believed trade liberalization is good for growth so people ran a lot of cross-country growth progressions, put in trade liberalization on the right-hand side and again found mixed evidence that trade liberalization will affect growth. And it really took a lot more work to convince people that there really was empirical evidence that trade liberalization was good for growth, and I think that is sort of where we are right now on this literature on capital account liberalization and growth.

But I think there are some more promising approaches that people are starting to work on now. One is you cannot just lump capital account liberalization together. You really need to focus on what is being liberalized, the sequencing, what types of controls were in place, and there are a lot of other factors that interact with capital account liberalization that will lead to greater benefits and greater costs and you need to take the whole package into consideration. So just a rough numerical measure of capital account liberalization is very difficult to get sound evidence and solid results on. And I think even more promising is there has now been an emerging microeconomic literature that does not just look at the aggregate impact of liberalization on growth, but looks at specific country examples, specific company examples, looks at the microeconomic evidence as exactly how capital account liberalization will have spillover effects throughout the economic. So that is what I am going to talk about next, microeconomic evidence on capital account liberalization.

I recently did a survey which pulled together about 40 different papers which have tackled in some way or other how capital account liberalization has microeconomic effects throughout countries and different experiences, and I just want to now summarize. All this microeconomic evidence can be broken into three main categories to pull it all together, and I will do this very quickly.

First, microeconomic evidence shows that capital account liberalization tends to reduce the cost of capital and ameliorate financial constraints so it is more likely to benefit smaller companies, companies without access to capital markets, and firms in less distorted markets. This has very implications for some countries where small firms are very important engines of growth. Capital controls can limit the access of smaller companies to financing and, therefore, limit the ability to effect growth.

Another set of microeconomic evidence shows that capital controls can reduce market discipline and insulate an economy from competitive forces, and this evidence shows that these effects on market discipline can occur through a whole variety of different channels.

Finally, a third major set of papers shows that capital controls can lead to widespread distortions in the behavior of firms and individuals. It can affect multinational investment patterns, repatriation patterns, it can affect how companies finance their debt, it can affect all sorts of individual actions. But the bottom line is in pulling together this whole series of microeconomic papers, it is clear capital controls will decrease the supply and increase the cost of capital, it will reduce market discipline causing a less-efficient allocation of resources, and substantially distort the behavior of firms and individuals. Therefore, capital controls do appear to have small benefits at least at the micro level, and have substantial and pervasive costs, but these effects do vary across firms and countries depending on preexisting economic distortions.

That leads to the big question then, what can countries do to ensure that the benefits of capital account liberalization outweigh the costs. I think there is no doubt that certain reforms will increase the benefits and reduce the cost of capital account liberalization. There has been a whole bunch of papers that have made arguments on which reforms matter, and you see there is a wide variety. I could put another 20 papers on this list. But there is a wide variety of what different authors suggest is important to make sure with capital control liberalization the benefits outweigh the costs. Some focus on legal and institutional developments, some papers focus on the role of transparency, others on overall developments, others on macro imbalances, others on the financial sector. So that is a pretty imposing list for a government that might want to liberalize their capital account to tackle all of these areas all at once before liberalizing your capital account and is a pretty onerous task.

What should countries do? I am going to lay out a broad strategy and then a more focused strategy. The broad strategy is I believe that countries should move forward in addressing all of these areas that could help increase the benefits and reduce the costs of liberalization, strengthening financial systems, sound regulations, et cetera. These are all policies that make sense anyway and will have the incremental benefit of making capital account liberalization more beneficial, but I do not think these should be viewed as prerequisites as could people do. I think if you wait to get everything perfect in terms of your financial system, regulations, rule of law, fighting corruption, then you will never get around to capital account liberalization and you will lose the very substantial benefits of capital account liberalization in the process.

Instead, what I think countries should do is do some precommitments to liberalization. I think China's example with bank reform is an excellent example. They precommitted to liberalizing the banking sector, opening it up to FDI, and to make that commitment credible they made it as part of their WTO agreement, so they could not back off. Then everyone knew China was liberalizing and would open up. That accelerated the pressure to reform the banking system and it gives the policy makers more backing to move forward on some of these reforms that need to be done in order to increase the benefits and reduce the costs of liberalization.

Sequencing is another important area. I think it makes sense to liberalize inflows before outflows. Micro evidence suggests strong benefits of liberalizing FDI more than some of the other types of liberalization, although there is mixed evidence on that. With the financial system, I think we need to realize that the financial system has a special role. Banking crises are correlated with broader financial crises, so we need to think very carefully about how you liberalize the financial system. I will not say more on that because I am running out of time.

The final point or final strategy is countries do need to take advantage of the good time to liberalize. There is a tendency when growth is good, the economy is stable, countries do not want to touch liberalization, but that is the time you need to do it. If you wait until things turn around, liberalization will be more risky.

To conclude, capital account liberalization needs to be viewed in a cost-benefit framework. The macro evidence is mixed, including the evidence on how liberalization affects crisis. The micro evidence though does show evidence that the costs of capital controls are more widespread and pervasive. Capital account liberalization is key to growth and development. I think the benefits are more likely to outweigh the costs if you also tackle other reforms. Just like when people start to drive, you also need to build the roads, you need to train people to be drivers, if countries take steps to build the institutional framework, capital account liberalization will be more beneficial, but do not wait for perfection and some ideas on how to liberalize more safely. The bottom line is I think there is an important role for more research in all of this. I think there really are still a lot of open questions on how best to liberalize, what is the proper sequencing, how do you prioritize different reforms that will go along with liberalization to increase the benefits and reduce the costs, and much more work especially I think at the country level is necessary before we can have any solid conclusions. Thank you.

(Applause)

MR. RAJAN: Thank you very much. We have had two really interesting presentations, one saying capital account liberalization is really, really bad, the other one saying it is really, really good. We will have an opportunity for the panelists to ask questions of each other and also, of course, for you to ask questions.

Now we have what I believe is going to be a presentation which will fall in between, maybe good, maybe bad.

MR. PRASAD: I am still a faceless bureaucrat. You will not get from me any jokes or pictures of cars. I will instead attempt to be the voice of reason.

(Laughter)

MR. PRASAD: There has been an explosion in terms of the literature on the macroeconomic benefits and costs of financial globalization (View the presentation). What I plan to do today is to give you a brief overview of this literature which has really exploded, but in very different directions. So what I will try to do is put together this thing in a synthetic perspective and then lay out a conceptual framework which I think is useful in terms of organizing this literature.

The remarks I will make today are drawn from joint work with Ayhan Kose, Ken Rogoff, and Shang-Jin Wei, although I should note that not necessarily everything I say aligns with their views or those of the august institution where we are gathered today.

The theory is fairly clear about what the benefits of financial integration should be. By allowing for more efficient international allocation of capital, it should increase welfare both in the countries that export the capital, I believe the rich countries, and the capital poorest class countries that get the capital. In addition, you should get international risk-sharing benefits by allowing financial flows to smooth consumption related to idiosyncratic or country-specific income shocks. And these welfare effects should of course be larger for developing countries because they are capital poor and they have more volatile growth.

If you look at the evidence, at one level it does accord with what one might believe. The emerging market economies which is the group of developing countries that have received the most capital inflows, they have grown faster than average, and even if you take China and India out of the picture, they have done better. Indeed, if you look at this from a cross-country perspective, there is a correlation between changes in financial openness measured as --

QUESTION: (Off mike.)

MR. PRASAD: All right, take my word for it.

(Laughter)

MR. PRASAD: There is an unconditional correlation which is positive, but when you look at the conditional correlation it disappears. That basically is the problem, that if you control for determinants of growth, macroeconomic policies, institutions which Dani emphasized and a few other standard determinants, capital account liberalization at the margin does not really seem to have much of a role. In fact, if you look at the broad swath of evidence as Kristin correctly pointed out, the evidence is really very mixed. You have about four studies showing no effect, most of them mixed, meaning that the results are not really robust to specifications, definitions of the capital account integration measure and so on.

What is going on here? These are all perfectly good studies done by good economists and many of these studies have been published in top journals, so of course they must be right, but something is not quite kosher here if all of these studies are not coming up with the same results or even finding conflicting results.

There are good reasons why the results differ. It turns out as Kristin mentioned that if you use more disaggregated measures of capital account restrictiveness or actual measures of financial integration measured by flows rather than capital account restrictiveness, you start seeing more positive effects in the data. And if you include much poorer developing countries, you do not see good effects perhaps for the reasons that Dani mentioned. And the longer the time periods over which you study this phenomenon the more positive the effects are. But by and large, the bottom line is that you cannot really find robust macroeconomic evidence in terms of the growth benefits.

But what about the volatility side? Of course, in terms of volatility, crises have received the most attention, and there Kristin has already pointed out that the literature does not speak as loudly as one might expect based on anecdotal evidence about the effects of financial integration on crises. And Kristin pretty much laid out this evidence which I think basically suggests that there is no clear evidence that by itself financial integration can be identified as the proximate determinant of financial crises.

But this is not to say that all is well on the volatility front. In fact, two of the key predictions of theory, or one of the key predictions is not really borne out in the data. Some of my co-authors and I have detected that in fact the relative volatility of consumption growth tends to increase at low and intermediate levels of integration which is exactly counter to the predictions of theory. So the basic finding is that developing economies, including emerging market economies, have not really attained the true benefits of risk sharing.

The macroeconomic evidence does not seem very positive at one level because there are not robust growth benefits, and in terms of the risk-sharing benefits, again, it is far from obvious. But there are other ways of looking at the data. One is to think about different types of flows and to ask whether the fact that you do not see these positive effects in macroeconomic data is simply because you are lumping all kinds of flows together. Or because you are looking at this from the wrong perspective in terms of levels of desegregation.

Of course, over time the patterns of flows have changed. These are FDI flows, equity, and debt inflows, and one pattern that you can see is that the composition of inflows to emerging market economies has been changing quite significantly over time, away from debt towards FDI with equity playing an important role but still not huge relative to debt and FDI, and the pattern is quite similar even for industrial economies and other developing economies.

Equity market liberalizations -- do seem to have very positive effects on growth, and equity market liberalizations here are defined as events where the country opens up its equity markets to foreign capital, and Becker (?) and his co-authors have found very large effects, in fact, some might argue implausibly large effects of about 1 percentage point of growth coming just from this one reform. These other authors have looked at the micro data and also found some evidence that there are micro reasons why at the firm or industrial level you might expect to see these positive effects of equity market liberalization.

But at some level these effects may be implausibly large, after all, as I showed you, portfolio equity flows at some level are still relatively small. And in addition, these studies, like I said, use -- liberalization events rather than actual flows.

The most important point in fact is that equity markets may be liberalized only when the countries feel that all their ducks are lined up, so to speak, that all the other macroeconomic elements are in order, so equity market liberalizations may be proxying for other good things happening on the ground. But still the evidence is there that equity market liberalization seems to have a positive effect.

What about FDI? Theory tells us that FDI is in principle the best kind of flow because it tends to be more stable, you get transfers of technology and expertise, you have the right sort of alignment of incentives between the foreign investors and the domestic recipients of the capital, but it has proven very difficult to detect the growth benefits of FDI using macroeconomic data. The microeconomic evidence is still weak, but it is turning out to be a little more positive recently. And there is evidence in fact that we may not have been looking for the benefits in the right places. If you think about spillovers of technology, you would not want an FDI investor to think about him having an incentive to spread out the technology in the same sector to his competitors, but backward linkages where there are spillovers of technological and managerial expertise to suppliers or downstream firms, that is where you might actually see these spillovers, and there is some evidence of accumulating of that, but still, rather weak. So by and large, one thing coming out of this work is that if you look in the micro data and look at the right kinds of benefits, you do start seeing them, but it is still something that requires a fair amount of detective work.

Attacking this problem from the other side is to think about the distortionary effects of capital controls, and there again Kristin's work has been very important in terms of highlighting these distortionary effects of capital controls using firm-level data. There are good reasons, as she mentioned, why you would expect to see capital controls having effects. These studies again use -- measures, but it is less of a problem because these studies typically are focused on one country or a very small group of countries and they can actually tell how much the capital controls actually bite and how effective they are.

Our reading of the evidence is that things are beginning to look a little more positive. Equity markets seem to help growth, FDI benefits are becoming more apparent, and the benefits by and large are becoming more evident in the micro data.

So let's go back to the theory and try to think about how to put all of this together. The traditional approach, again, is to think about how financial globalization through capital deepening is going to improve growth and volatility outcomes and generate better sharing. Clearly in the macro data this does not quite seem to work, so slight variation of this, not denying the importance of the traditional channels, but thinking about these intermediate, indirect channels, what we call the potential collateral benefits, may be a better approach. Essentially, the approach here is to think about financial globalization having these benefits, and these collateral benefits in turn affecting growth. This is just a slight reformulation of the standard theory, but it has very powerful empirical implications because if you think about regressing GDP growth and financial globalization, you get a positive coefficient, you have controlled for the stuff and it goes away. But the fact that the regression coefficient is no longer significant does not mean that financial globalization does not work, it simply means that it may be working in fact exactly the way it should be working.

This link is fairly well established in the literature, the fact that financial markets and institutions, and Dani's work has been very important here, macro discipline, the IMF mantra, all of these of course are good for growth and reduce volatility.

What about this link? This is where a fair amount of evidence has started building up. The problem in terms of establishing this link is that the links are really bidirectional, so disentangling the causality which we economists of course care a great deal about turns out to be very difficult. What we have here is a certain amount of evidence that is moving in the right direction, so let me just lay out these collateral benefits.

If you look at the relationship between financial development and financial integration, the correlation between measures such as private credit to GDP or stock market capitalization is pretty strong. Of course, the correlation is not a causation, but there are good reasons and channels that have been identified in the literature why foreign ownership of banks or equity inflows should in fact boost financial market development and, in addition, financial sector FDI has been shown to have a lot of benefits. So on that front there is a fairly large body of evidence suggesting that financial integration does help domestic financial development.

Likewise, on the matter of institutions and governance, there is a fair amount of work suggesting that openness to capital generates incentives for improving corporate governance and, in fact, making it easier to improve corporate governance by importing it from abroad. And in addition, there is some evidence that corruption and institutional quality can be improved, or at least the control of corruption can be improved. So, again, suggestive evidence of this collateral benefit.

Macroeconomic policy discipline is one other potential collateral benefit, and the logic here is that capital account liberalization acts as a commitment device. So if you open up your capital account and run bad policies, you can get hurt very badly by international foreign investors, so it gives you an incentive to run better policies. In fact, even the IMF's favorite macroeconomist Joe Stiglitz is on board. Sadly, the one thing Joe agrees with us on, there is very little evidence for. If turns out that if you look carefully in the data, there is some evidence that financial openness is associated with better monetary policy outcomes in the form of lower inflation, but the benefit that we all thought would be there, which is better fiscal policy outcomes, there is very little evidence at this stage. So the evidence on this is mixed.

Collateral benefits are all fine, but there is a complication. There is always a fly in the ointment, and in fact, it is probably bigger than a fly, it's a bug of unspeakable name, and this is in the form of threshold effects. There is a large literature suggesting that the benefits of financial integration really can be felt by a country only when it has certain levels of certain good things, and these good things include financial market development, institutional quality, macro policies and trade integration. If you are below the thresholds, bad things happen. GDP growth may or may not rise, but volatility certainly does. If you are above the thresholds, you get the sort of outcomes.

What is the problem here? Virtually everything in this list of threshold conditions is exactly what was on the list of potential collateral benefits, and therein lies the tension. There is some evidence that financial integration can catalyze a lot of these good things, the potential collateral benefits. That evidence, again, is strongly suggestive, but not quite conclusive yet. The problem, however, is that unless you have some of these good things to begin with, you run into trouble.

So how does one resolve this tension? First of all, the threshold literature I think is where a fair amount of work is needed. Kristin already made a call for more research, and I think this is where a fair amount of work needs to be done both in identifying the thresholds, thinking about the interactions among the different thresholds, and finally, be sure that the level of financial integration itself functions as a threshold. But I think the collateral benefits perspective may actually be a way of moving forward, and let me conclude with these thoughts. It has given us at some level I think a unified conceptual framework for thinking about how it is that financial globalization works through both direct and indirect channels. But the good thing is that within the context of this unified framework you can actually take the count of country-specific circumstances or initial conditions, and the logic there is that if you think about a country that is going to undertake financial integration, you could think about what collateral benefit might be most important for a given country given its initial conditions and specific circumstances. So you could potentially design a selective approach to capital account liberalization that makes it work, and this is not quite pie in the sky thinking, it is not easy, but there are ways that one can think about approaching capital account liberalization to get these benefits in a selective way.

But I think the bottom line is that during the transition to these thresholds there really is no way to eliminate the risks completely. Ultimately, as Kristin said, there is going to be a cost-benefit tradeoff and what one should be thinking about is how to manage the risks rather than eliminating them, but I think it can be done and ultimately may be worth doing. Thanks.

(Applause)

MR. RAJAN: Thank you. I did forget to introduce Eswar, but Eswar needs no introduction. Let me now call on Joaquim Levy. Joaquim comes to us from the IADB, and somewhere in here I had his bio which is missing. So I will leave Joaquim also without an introduction since he has been at the Fund and served in the Brazilian government before taking his current post. But Joaquim, you have the floor and for 15 minutes.

MR. LEVY: Thank you very much. It is always an enormous pleasure to be back at the Fund. I thank you for the invitation of my name of the IADB. It is also very good to have the opportunity of exchanging views of those who are at the inside and those are at the west side of the White House. And I understand that Professor -- was also here at some point. I found the discussion in this very stimulating here as usual at the Fund in particular because of the somewhat nuanced approach that we saw here. It was quite interesting especially for someone who at some point had to deal with these questions in a very concrete way while we are formulating or participating in the formulation of economic policy in a given country, to see this realization, this perception of the tradeoffs, of the considerations that this subject raises.

I think that the example of the car is certainly very inspiring as well as my immediate reaction was exactly what Kristin said that at the end, that if you buy a car, sometimes it is good that you also have roads and it comes with a lot of things. And I think as far as to confirm this, it has to be complementary, and I think what I understood from the discussion of Dani on the investment constraints, I think that is exactly the reality we face in some countries.

I think that as in any case where you have a liberalization or you really open the doors and the windows, what happens is they exacerbate the problems if you have distortions, if you have problems when you open the windows you will see more and then sometimes the question is should you keep everything closed just because you have these problems. And I think that in terms of policy making, the question is exactly in the old times we have called sequencing or in real life it is how we balance the different steps and know that it needs to be effective is not just one measure that you sold everything. It is a combination of a lot of things of which might be here called as opening capital accounts is part of it and you have to see how you move all the pieces.

I think in particular that financial integration is the real issue more than some constraints and controls. If you do have a big financial sector and it is an integrated financial sector, I venture to say that most of the literature says that controls are ineffective. I remember back in the 1900s and you look at Latin America you had a big financial sector with big foreign banks, specifically in Chile there are a lot of studies showing there is so much you can attain, and Brazil was the same case. So I think in some ways the real issue is you have or you do not have integrated financial markets.

If you do have an integrated financial market, I think, and the case of Brazil is obvious, it is very difficult to come back and you do not put, like someone said, the genie back in the bottle. So you really have to think hard what are the other actions that you do take to ensure that you are not driving on an unsafe road. There I think that a number on this list of reforms are very important. Of course, this is part of a general move of more transparency, more governance, and I think the thing at least in Brazil, and I might be here modest here in my conclusions, is that you have to have the two things or all the things together to do it.

I will take one just instance, but I think this is understood. I will take just one example of what could be called capital market liberalization. Again, first, it is an example that this is more complex than a yes/no situation. The example is the following. At the beginning of this year, the Brazil government decided to change the taxation for foreign investments in public sector securities, basically government debt. To some this was seen because you had the taxation were seen as a restriction in capital mobility. It was not really a restriction, but reduced the movement of capital because it was an inconvenience. Locals do pay, but for foreigners to pay it means additional costs and it was seen by many as a restriction.

We found that as you moved the inflation was being overcome and it was good to have more participation of foreign investors, although we knew that it of course made the country more in some sense vulnerable, but it was useful. Then the discussion, and I remember some public discussion with the president, the point was, yes, we know we are taking more risk, but it is worth doing it. And I think that many of you realize that in life you have to take some risk and the art in some ways is to weigh how much risk you take and how this fits in a more general policy.

So I think that the way, and again this is one experience, it is not a broad scientific evaluation of pros and cons, really the decision at the bottom was can we afford this risk, do we have enough other elements that would justify it. Also as I mentioned that time to the president is this will work as long as we continue with the other policies. I am fine even if the world is not as benign as today that investors are not going to disappear from Brazil or there will not be a big outflow as long as you have the same policies. Is this a precommitment? Maybe, maybe not. I think it is not enough to be called a precommitment, but certainly this is a kind of decision that has to be taken with a broader framework and view. And I think that probably with someone will look at all these events, the ones that went fine, there are a lot of other things going on and it was part of a larger package.

Talking about packages and roads and cars and all of that, I can't help mentioning another thing that usually, and I recently had to do it actually at a very high cost, but the one thing that we have to do sometimes is to buy insurance. I am with the IMF so I do not have to say much more what could be thought as insurance, and I think that now maybe the issue of international architecture is not in the forefront and this is more of an academic say meeting, but as someone who of course was in government and now I am on the other side of the street, certainly this is important. I think it goes with buying a car and not having the things that will minimize or mitigate the problems if you do have a crash or anything like this may not be such a good policy.

We can discuss if there are insurances that really work or do not work. Leslie I think is not here, but I think that it is certainly also a commitment if we are convinced that capital account liberalization is a threat, we should also be ready to consider some things that will help this process to move forward. Guillermo Cavo (?) from the IADB has some proposals of a mechanism how this works. I know here a lot of people have thoughts. And I will just say that Brazil certainly has a view about that, if you are taking the bat, if you are taking the risks because you believe that it is good, I think it is really worth it thinking of ways to deal effectively with the risks that do exist.

So I think that there is not much I can add to the very comprehensive discussion of my three predecessors, but, yes, I would agree that you can call it different ways, investment constraints or whatever, if you do not have certain things in place, it is probably more difficult to think that you are going to have risk benefits just by liberalizing. Actually, sometimes just creates a lot of incentives. I was just reading the paper yesterday of Shaw (?) and at some point he mentions as well if you open a capital account and you know that you can have sudden stops, rational agents, that would be fine, except that we know that in real life we do not have one agent in a country, and you have a number of agents that can have very different incentives. Again, one of the things that the IADB is doing is the fact that if you just look at the current account deficit, you may have a different picture than if you look at the outflows and inflows because it could be different people and they could be protected in different ways when the hard times come. So I think that also when we think in liberalization this issue of looking at the micro and disaggregate a little bit the effect in different sectors or different types of agents can be extremely helpful in guiding the discussion for the policy makers.

Thank you again. I think it is a discussion that is certainly very helpful for policy makers and I think that it also says that the world of sequencing still has some value. We now have more to guide this process of sequencing. And yes, probably it is worth doing and I think in some cases and I think in terms of Latin America given that we already have so much financial integration, do the other reforms there is not much more choice if you want to reduce the macro and the general economic risks and also accelerate growth in the region. Thank you very much.

(Applause)

MR. RAJAN: We have had varied set of perspective. Dani started with looking at the macro effects which might be detrimental, and he suggested that policy measures should be explored to have selective capital account controls. Kristin focused more on the good micro effects and the bad distortions on the micro side created by capital accounts, and she suggested a precommitment to liberalization to get the benefits without incurring the costs of volatility and so on.

Eswar looked at the collateral benefits from capital account liberalization, the effects of capital account liberalization on financial development, and suggested an intermediate position in some sense between the two. And finally, Joaquim talked about those who had already liberalized and how it was very important that once you open the door, find ways to minimize the costs including through novel forms of insurance. I think we have sort of covered the ground.

Let me ask the panelists first if they have questions for each other just given that you have hopefully some contradictory opinions that you might have direct questions? Any of the panelists? Dani?

MR. RODRIK: I would like to hear what the Chair has to say on these issues.

(Laughter)

MR. RAJAN: I am a faceless bureaucrat. I have no opinions. I think this is your day and not mine, so I will refrain from my offering my opinions on the subject. But let me throw it open to the floor. I am sure that people have a lot of questions of the panelists and then we can come back to the panelists themselves if they have some questions in the middle. Jeremy?

QUESTION: I had a question for Dani Rodrik. You mentioned there is some evidence that undervaluation episodes get growth going, and there is also some evidence including from people at the Fund that overvaluation is bad for growth. What do you of the scope for monetary policy in getting these undervaluation episodes going? On your slide you had some language about private and public savings which is kind of the apple pie aspect of it. Would you encourage countries to use monetary policy to achieve undervaluation, and if they succeed, do you think these episodes can generate some kind -- and be sustained, because many times they are trying to do that right now?

MR. RODRIK: Yes, I think these are exactly the right questions to be asking and I would include in this arsenal of policies things like sterilized intervention as a short-term measure to lean against the wind when the exchange rate is appreciating. I have written about the high cost of reserve accumulation, but since I do think that the exchange rate has implications for growth, I think many countries do have room due probably in the end to accumulate reserves if that means preventing the exchange rate.

I think there was one whole quite unexplored frontier in monetary police which is how to make inflation targeting in the emerging market economies sound like they could not care less about the real exchange rate. I think that gives the wrong message to the real economy and we need to find the right language and the right operational way of where you are not throwing all the good things that inflation targeting and central bank independence -- to the monetary policy framework in various economies that have adopted that regime. But we find ways of thinking in incorporating the employment and tradables, the health of the export sector, the current account, and therefore the real exchange rate as one of the indicators that we are thinking about systematically, and, further, communicating that when the central bank talks about monetary policy because I think the signaling is very important because of these -- effects that you mentioned.

Then for countries that have not gone already that way there are I think broader issues. We have countries like Tunisia, for example, with closed capital accounts, and despite IMF skepticism has had a real exchange rate targeting policy for a very long time and has done it very well without apparently losing control over the price level. But we do not exactly understand all of these things. I wish there was more research being done on this including in the Fund where there is a lot of talent.

I would also add that there is some new evidence from my colleague Federico Sturzenegger that actually shows that sterilized intervention can have an effect on the real exchange rate in the short- to medium-term and then have the growth effects are positive. But I think if you start thinking about those in this direction, you also understand how many other things that we do not really quite understand very well that we need to do research on.

QUESTION: How do you take this up to the macro level -- we should have the right conditions, but we should do financial intervention to get these conditions going. It is not obvious to me why policy makers with limited political capital should focus on finances related to capital account liberalization as the most -- as opposed to say they are getting growth going and then getting -- as a consequence to growth --

SPEAKER: What about people who work on growth?

SPEAKER: They should not be allowed to say growth is good for the poor.

(Laughter)

MS. FORBES: You raise a very good point on all of these micro studies, especially ones that focus on one country, what else is going on, is this any experience that can be generalized to other countries, and that is the bit shortcoming of a micro approach.

QUESTION: (Off mike.)

MS. FORBES: Right, but I think how it be useful to group things together and get the aggregate effects is I think that micro studies are helping us to understand better exactly how liberalization affects growth, what channels they work through. Eswar showed the collateral benefits, and the micro studies can help you sort that out, and then hopefully when we do have a better understanding of exactly how they work, then we will be able to get some better macro evidence. That is what I am hoping.

But also I think there it is useful to look back at the evidence on trade liberalization and growth. Initially the macro evidence was more skeptical, and then it took some of these micro studies to really better understand exactly how trade liberalization will affect growth and work through other variables, and I am hoping the micro evidence will have that sort of effect and then you can control for the spillovers.

MR. ESWAR: May I comment on this as well?

MR. RAJAN: Sure.

MR. ESWAR: Let me take Arvind's second question. I think you are hitting upon an issue that thinking about financial integration as the way to achieve growth is something that is not supported by the data, but catalytic benefits, too, cannot be overstated. It is my view at least that some of these benefits like financial market development over time. The question is whether something like financial integration helps the process along, and I think the political economy may be an important angle here.

The eminent faceless bureaucrat on this panel has in fact done I think very interesting work suggesting, I mean Ragu of course in case it was not obvious, that in fact things like financial integration can break political deadlocks eventually by giving interest groups that may be blocking reforms a kick in the pants so to speak by making some of the benefits clearer to the broader segments of the population, thereby dislodging some of these effects. So, again, I would not want to overstate the benefits of financial integration.

I think what is clear in the data is that the direct benefits, the financing benefits, are not obvious if you think about just the raw financial capital helping the growth. But if there is anything in the data, it is that collateral benefits may be the way that growth -- but even there I think one can really overstate the case and the evidence, as I said, is strong suggestive but certainly not conclusive.

MR. RAJAN: Dani?

MR. RODRIK: I think the question that Arvind Subramanian asked is really right on point, and I wanted to make a conceptual point related to that. The problem with the micro studies from our perspective is or the perspective from the question about what does this do to growth is the following. Suppose that you observe a country like Chile during the period of the taxes on capital inflows and you look across firms and what you see is that firms that are let's say more dependent or are smaller face the higher cost of capital and, therefore, they were disadvantaged. This is what theory would lead you to leave and that is perfectly consistent with that.

But the sense in which that does not tell us anything about what we actually would want to know is the following, which is that if you actually had not had those capital controls, what would the environment actually have been like? And if it is the case where in fact the real exchange rate would have been significantly more appreciated, then that is an equilibrium where in fact all the tradable sector gets a hit and therefore you have the investment and employment in all of those sectors essentially being less than what they would have been.

So what the micro cross-section study is not taking into account is not comparing the right counterfactual. It is taking as a given the existing level of the real exchange rate and doing a comparative static around that, whereas the growth story is really about what would have happened to the real exchange rate and what consequences overall would that have been.

So the micro studies that find some distortions or misallocation or high cost to capital is perfectly consistent with the story that I gave which is that without those capital controls you would have been far worse in all those things that you care about.

If I can also add the point of look at trade liberalization. For so many years we did not have any evidence that it helps, and now we do, and it has got to be the same about this other area of policy where we do not have the evidence, all we have to do is wait and we will find the benefit. I just do not understand that argument. If we believe that something works and it is just that we have gotten the right evidence, then we just believe it works, but we believe it on faith, not because of evidence. And certainly if the argument is going to be that we have to wait long enough for us to get the right kind of evidence but that it is surely around the corner, what kind of science-based or evidence-based policy are we talking about?

MR. RAJAN: I was going to give you a few seconds to think about a response. If you want a few seconds, I can give you a few seconds.

MS. FORBES: That's okay.

MR. RAJAN: Go ahead.

MS. FORBES: I will take the last one first. I am not saying that just because of trade globalization -- the evidence, we will, but my main point is do not give up hope. Just like we did not write off trade liberalization as bad because it took us a while to get the evidence, I do not think we should write off capital account liberalization as bad just because we are not as convinced by the evidence yet. I think you created that -- than I meant it at least.

The other issue on the micro evidence versus the exchange rate effects, there I think then this is where you get back to my first point which is we need to think about this is a cost-benefit framework. The micro evidence really highlights some of the costs of capital controls. You are pointing out some of the potential costs of liberalization, but I think again we need to weigh them both in your what is the relative weight of the different costs and the benefits. For example, your specific example with Chile and the exchange rate, maybe there were these microeconomic costs, but you need to then know what was the aggregate effect on the exchange rate of the controls. And there then we need to look at other studies, and there has been actually a series of papers trying to look at the effects of the Chilean encage, the capital controls, on the exchange rate. There the evidence had found no significant effect of the controls. The majority of the papers on this found no significant effect of the Chilean capital controls on the exchange rate. So now we have concrete micro evidence of the cost of the capital controls, but there is no compelling macro evidence on the benefits in terms of reducing exchange rate appreciation. So there I agree we need to look at both, but that is a clear example where doing the cost-benefit analysis makes me think the micro evidence at least gives us something concrete and the exchange rate effect is not too concrete to weigh against that.

MR. RODIK: May I ask a question?

MR. RAJAN: Sure.

MR. RODIK: If the encaje has not had an effect on the real exchange rate, it could have been because the encaje did not have an effect on the amount of capital inflows which is presumably what you have in mind. Is that what you were thinking? Let me ask the other question which I was asking. Do you believe that capital inflows have any effect on the real exchange rate?

MS. FORBES: My reading of the Chilean events is that the encaje did not have a significant effect on the volume of capital inflows, so it was not surprising there was not an effect on the exchange rate.

MR. RAJAN: Questions from the floor? Mike.

QUESTION: Yes, I must say I am a little disturbed by this conversation. Let me go back to 1985 when I wrote a chapter for the Council of Economic Advisers Report on policies to promote economic development. I think there we surveyed the evidence and we did find there was a consensus at that time that relatively open policies toward international trade did tend to promote growth. Relatively open policies were not characterized as free trade, but South Korea had a relatively open policy, and North Korea did not, to take an extreme case.

However, I declined to bow to the pressure from within the Reagan Administration to say that capital account liberalization benefited economic growth because there was no evidence to support that proposition. Indeed, when one looked at the record, the Japanese economy had far and away the strongest economic performance of industrial countries from the end of the Second World War through the mid-1980s and it had a quite closed capital account and a highly repressed financial system. And the openness to international capital mobility in Europe came gradually over the course of time and generally somewhat later than their periods of maximum economic growth. So it seemed to be very difficult to reach a very powerful conclusion that somehow the be and end all of promoting rapid economic growth was to have a highly liberal financial system and a liberal regime of international capital movement.

I think it is true that as economies advance and as the financial system has become more sophisticated and advancing technology in computation and communication and so forth has contributed a lot over the past couple of decades to the growth and sophistication of the financial system, that a much more liberal financial system and a much more open financial system for the most advanced economies does make sense. But still, when I was at the Fund and I would still say now, one wants to be cautious about prescribing a general prescription for countries around the world without examining their individual circumstances.

I agree very much with Eswar there probably are some general rules that are useful in giving such advice, and with this observation I will conclude. For example, we know that Kristin has a beautiful young baby boy. I suggest that she not begin to teach him to drive for at least another decade.

(Laughter)

QUESTION: Joaquim Levy made an interesting about poor countries that had liberalized and needed some insurance mechanisms. So I would like to ask him two questions. One is kind of insurance mechanisms does he think in terms of national or international? And secondly, if these insurance mechanisms are not developed or are waiting for a contingency credit line that works or for insurance mechanisms through the market, if these insurance mechanisms do not happen, would his position or that of the other panelists change in relation to the balance of benefits and costs of capital account liberalization?

MR. RAJAN: Let's collect two more questions and then we will end with those.

QUESTION: My question to the panelists is to get their reaction on these benefits, somewhat short-term or if they are spread over a longer-term period? The reason why I ask this question is because there is potential perhaps of reconciling maybe the macro and micro evidence in that most of the micro evidence within a short-time window in reference Henry and work that Sheri Nusha (?) has done, and work done by Dani himself was a much longer time period. And if one was to look at just say a stripped down, real simple neoclassical model and one scenario of two otherwise identical countries and one that is converging toward a steady state and all of a sudden you open up capital accounts, what happens? You are likely to see an upward sloping accumulation of capital. So if you were to compute the growth rate in that window, you would see a much higher growth. But as they converge to the steady state and you look at a longer time horizon, you would see the growth differential was sort of narrow. And I would like to get the panel's reaction on what they think the importance of the horizon over which you are measuring these benefits is.

MR. RAJAN: Does anybody want to take the last question?

QUESTION: I was intrigued by Dani Rodrik's argument that because of the distortions in governance and externalities in the tradable sector that you have benefits to having an undervalued exchange rate, and the evidence you referred to was the correlation between real exchange rates and growth rates, but obviously that correlation could have a number of other possible interpretations other than the distortion sort of argument that you gave.

What I was wondering is, first of all, whether there is another body of evidence that you could refer to that would convince us that the distortion story is the correct one in interpreting that correlation or whether there is other empirical work that you could suggest that would help untangle that question.

MR. RAJAN: Let's go through the panel backwards this time, giving each one a minute to respond or maybe make some last-minute comments. Joaquim?

MR. LEVY: I think that in answering the question about the insurance, I think probably countries will need to be more careful in perhaps the speed of liberalization, so that is certainly a perception in many countries.

MR. RAJAN: Eswar?

MR. PRASAD: I have just one observation. In terms of thinking about capital account liberalization whether it should be pursued or not, I think one reality that a lot of countries, especially middle-income countries, is that whether they like it nor, the capital accounts are becoming more open over time. You get capital flowing in through the trade channel through under- and overinvoicing of trade transactions. And if capital has an incentive to find its way in or out, it is going to find its way in or out because there are increasingly more channels to do so. So I think it is not so much a choice about whether or not to open the capital account now, but how to manage the process and how to deal with these difficult issues later to transition. So in that sense I do not think the passive approach is really on the table anymore. Ultimately, it is more a matter of some of these countries controlling their own destinies.

MR. RAJAN: Kristin?

MS. FORBES: Also coming to a question related to insurance, which I think is a very important point of liberalization is how you can better insure against the risks. Here I think the type of liberalization you do can be very important in terms of buying a little bit as insurance, not as much as you would ideally like, but that is where if you were to open first to FDI and equity investment, then if the growth slows, profits slow, the exchange rate weakens, then some of the risks of that down side are shared with the foreign investors. So that in some sense is a partial form of insurance that you would not get if you say liberalize first to debt inflows. Then if you liberalize debt inflows, you still have to pay the full amount back immediately.

Then, of course, I am sure Stephanie was hinting at this, something like growth index bonds is another way to issue debt and share insurance better against down sides as well as up sides. So I think thinking carefully about how you can better insure against the risks should be an important part of capital account liberalization.

In terms of the benefits long-term or short-term, that is I think a very good question, and I am not sure we have great evidence on that. My guy reaction without any empirical evidence is that some of the benefits of capital account liberalization do occur through raising productivity growth, through better market discipline, through better pricing say in equity markets and financial markets, and you would think those would raise productivity growth which would be a long-term effect, but that is in turn going to be a lot harder to measure. Theoretically that it where I would expect to see it but, again, I do not think we have the evidence yet.

MR. RAJAN: Dani?

MR. RODRIK: Very quickly on the question about overvaluation or real exchange rates and growth, I think the answer is that, first of all, it is worth repeating, and I may be putting even more strongly, that after so much time, so much work on cross-country growth empirics, I think the only robust finding in the huge growth empiric literature to my mind is the negative relationship between overvaluation and economic growth and that is sort of the strongest and most robust relationship in the growth relationship.

I think the question you are raising rightly is do we actually understand why, and I gave a couple of reasons having to do with, first, overvaluation is the relative price of tradables to nontradables so it has got to be something that promoting tradables over nontradables has got to be good for growth, and I think there are two categories, as I said, of why that might be so, but I think we are probing deeper and saying do we actually have real evidence on each one of those. We have bits and pieces, but not overwhelming convincing evidence of that, but I think the cross-national regularity and correlation is extremely strong.

MR. RAJAN: Thank you very much. I think we should thank the panelists for a really extraordinary session. It was great.

(Applause)

MR. RAJAN: It was very insightful. I think if you ask what can we agree on at the end, we can certainly agree on the fact that industrial countries seem to benefit from foreign capital, and poor countries do not benefit as much as they could. So how do we get poor countries to benefit from foreign capital, that is the research agenda in a sense, and I think the panelists would all agree that if there is a way, we should think about how to achieve that. Maybe there is no way, in which case we need to think about a different approach.

Thank you very much, and thank you all for coming to the Annual Research Conference.

* * * * *