IMF Conference

Panel Discussion on

Friday, April 13, 2001, 3:30 p.m.
IMF Auditorium

Opening Remarks - Stanley Fischer, IMF

Participants:

Rob Townsend, University of Chicago

Nick Stern, World Bank

T.N. Srinivasan, Yale University

Angus Deaton, Princeton University

Open Discussion

Closing Remarks - Stanley Fischer

MR. FISCHER: This workshop is what is called in academic life a pre-conference, that is, it features papers that are in the process of being written. The intention was to have experts in the subject review the papers and give the authors suggestions for taking the research ahead. So that's what this day-and-a-half workshop has been about. But it would be a sad misuse of resources to have the experts here at the Fund—the people on the panel and many of you in audience—and not make use of the opportunity to have them speak to a broader audience. That's the reason that we decided to end with this panel.

I think a question which many people have asked out loud, or thought quietly to themselves, is: Why is the IMF doing anything about poverty? Some here have stronger views on that than others, and I'm sure that as this discussion gets going, you will figure out which is which.

We have four terrific panel members. Rob Townsend of the University of Chicago, has done extraordinary work with panel data on financial intermediation in Thailand in particular, and also work on India. He moves easily between very abstract theory and very detailed empirical work and was here as a discussant on other papers. Rob?

MR. TOWNSEND: Thank you very much. It is a pleasure to be here. The conference was about poverty reduction and macroeconomics, and the juxtaposition of those two concepts is either productive or argumentative. I want to think about my remarks as arguing that they shouldn't be juxtaposed with one another as if they were in opposition to one another; that is, one shouldn't think that one has standard IMF operating procedures, and then one has appended to that poverty reduction as an additional goal. I think that that would be a potentially unproductive way for the Fund to proceed, and I'll try to argue that case with some examples.

One way to think about this is something that was a little unclear during the workshop, although it did surface from time to time, which is the difference between moving along the Pareto frontier of the set of all possible Pareto optimal allocations and hence a question of redistribution, presumably from poor to rich, versus the idea of moving from the interior of a utility possibilities frontier onto the frontier, in which case one is using policy and doing an evaluation to try to remedy imperfect institutions or missing markets. And it may be a little bit of a stretch, but one could certainly conceive of the IMF being created in order to make up for some missing capital flows in the international markets. But the same principle applies for the analysis of domestic markets.

So let me be very specific and tell you about the Bank for Agriculture and Agricultural Cooperatives (BAAC). This is a bank in Thailand that is chartered to lend to agriculture and hence to the poor, but lest you think that it is tiny and insignificant, it turns out that it is actually the dominant formal rural lender in Thailand, with millions of customers as its clients.

During the crisis, this government bank was evaluated, and it looked as if it had, like the private commercial banks, a fairly unhealthy percentage of nonperforming loans. Then, there was a division of what to do about it. One group at the time was arguing that the policy was too stringent, and although the Bank for Agriculture was apparently showing losses, that it should be bailed out, that there should be government support and recapitalization of the institution.

Another group might have gone so far as to say that the bank should be privatized if not dispensed with but certainly would have argued at the very least for a reevaluation of the nonperforming loans.

As it turns out, the truth of the matter lies between those two extremes. The Bank for Agriculture and Agriculture Cooperatives in Thailand is running a risk contingency system which is not terribly clear in its books. When farmers suffer from local problems, like illness or having their house burn down, or regional problems like droughts and floods and so on, they will petition to the bank officers who in turn, actually at some cost, visit the creditor, make an assessment. If it is judged that the ability to repay the loan is indeed something beyond the farmer's control, then, there is a renegotiation of the loan, so to speak, the loan is rolled over, sometimes interest is not paid or is paid later, sometimes the principal is forgiven.

In effect, as I said, the bank is acting something like an insurance company in addition to its provision of loans. Is this performing a useful service? We know enough from economic theory to evaluate this bank on the basis of whether it is reallocating risk for the farmers, and one can look at consumption fluctuations relative to income fluctuations and make an assessment of the risk-smoothing, consumption-smoothing role of the institution. And indeed, with hard numbers and some guesses about utility functions and so on, and conceivably dynamic, stochastic optimization models, one can actually put some numbers on the willingness of Thai households to pay for this kind of service.

Then, we turn to the accounts that the Bank for Agriculture is keeping, and we notice, with some difficulty I must say, that there is a transfer under the other income category from the Government of Thailand. Then the issue becomes: Is this transfer really meant to cover the risk contingency system? If Thai farmers are not paying enough in premia and interest to cover the insurance system, can we view the transfer that is coming from the Bank of Thailand and Government of Thailand as basically an income transfer, paying the premia for the farmers.

So one ends up comparing the welfare gain based on theory and data with this government transfer. My point is that this is a way to evaluate a financial institution. In other words, we have gone from an institution designed to help the poor, thinking about the standards for evaluation and the way that institutions like the BAAC and commercial banks were treated during the crisis under IMF loan conditions and the way they might be evaluated somewhat differently based on this alternative standard.

So again, rather than having poverty tacked on, if somewhat artificially, to a standard IMF operating procedure, the hope is that the IMF does things that are helpful to the poor through a hardnosed assessment of the efficiency of these financial institutions. This does not rule out subsidies and transfers, but one has to do the arithmetic to make sure that the subsidy is warranted relative to the value of the transfers.

Thailand off and on continues to be on the watch list. I'm not sure if it is today. I know that it was several months ago, at least under the U.S. Treasury Watch List. Why is that? Because it continues to have rather anemic performance. It is not still in crisis, but it has not recovered very strongly, and the issue is is there a way to deal with it, and how should the IMF and the World Bank and the Asian Development Bank be thinking about Thailand. Is there something left to do, or are we stuck with the standard macroeconomic tools?

The answer is, again, there is much left to be done, or at least, so it seems. The same Bank for Agriculture still has a somewhat limited charter. If we use standard, general equilibrium models of occupation choice and estimate welfare gains from somewhat modified credit contracts, it is apparent that much growth remains in Thailand which could be drawn on, and presumably, if the models are correct, would happen to a certain degree.

There are social safety net programs put in place to help the poor in the crisis. That is still being implemented, even though I guess Thailand is really no longer under the IMF program. Village funds and village savings banks are being promoted throughout the countryside and funded by government expenditure without hardnosed evaluation of the role of these institutions that are supposed to help the poor. Again, it is a very shallow version of what one would really want

The point is that we know, collectively and professionally, how to do these evaluations, and even a small additional budget allotment, a fraction of a percent of the government revenues, would complement the Fund's programs and the World Bank's programs for these kinds of institutions.

Meanwhile, the government sets up a small business credit guarantee fund, which I fear—I struggle to get the details—basically has close to 100 percent guarantee of small business loans. Again, this is not something that economists would be drawn to because it does not deal with the moral hazard problems. And commercial banks themselves remain relatively undiversified and have historically not been servicing the rural or semi-urban areas and areas other than industry, if not services and real estate.

One can think about an evaluation of commercial banks, and this is being done with some data from Argentina and done by the Bank of Italy with credit registries in both countries. Again, the point is that the metrics that are used under Basel Standard Risk Credit Metrics need to be modified and placed in the context of these evaluation techniques, the use of micro data. It is not true that one wants to insist on strict repayment when we know that there are aggregate shocks which would not be subject to the moral hazard problem.

I will just say a few words about monetary policy, because much is made of the role of the Fund, rightfully so, in maintaining macro stability and price stability. It is certainly true that the poor would suffer from inflation, and we presume—and some data measurement that I have done with Indian villages confirms—that indeed the poor are the heavy users of currency so that indeed they would suffer more in inflation.

However, the issue is why are they using currency and not other means of smoothing, and do they have capabilities to smooth through these other means. That is very important when one actually tries to put some hard numbers to the cost of inflation. Or indeed, people may be using currency to smooth idiosyncratic and aggregate shocks. This is something that Angus Deaton has worked a great deal on.

So then, imagine that there is a regional shock of some magnitude, and these markets are incomplete—the whole point of the buffer stock models is that there aren't complete markets for credit and insurance. If one believes that money is being used by households, as it has been estimated to be used, as a buffer stock, then there would be pressure on the monetary authority in any country, particularly in developing countries, to engineer payments of some kind or other transfers—they would take the form of loans, maybe, to agricultural development banks—but the pressure would be there.

Of course, in these days, we know that central banks need to resist pressure to just start inflating, because otherwise, expectations could take over, but the pressure will make its way to pressure for loosening monetary policy. Then we have the question of should the central bank engage in open market operations or some other means to try to accommodate these local shocks which are causing a problem. Or indeed, the local shocks could be aggregate macro shocks, as they were in the Asia crisis, or shocks that repeatedly hit the Latin American countries.

So the debate rages at those times about whether monetary policy should be loose or tight—tight to keep the interest rates high and attract foreign investment, or loose, presumably because people are genuinely worried that credit markets are limited and that monetary policy is needed to stimulate the economy.

My point is that we need much better measurement of the extent of the credit and insurance markets in these countries—that is an integral and an intimate part of what the Fund needs to be doing. It needs the data, it needs the measurement, it needs to couple these with theory to gain a better understanding.

Another related example would be, say, the work of Ricardo Caballero, or Bengt Holmstrom, or several of my temporary colleagues as I visit MIT. Notions of liquidity constraints are very much on the minds of these people in understanding how countries respond to crises, and presumably, that should make its way into IMF policy.

So in terms of where does that leaves us, it seems to me that poverty is not something that should be appended on and added to the standard agenda. We should emphasizing the movement of the Fund in its standard operations toward things that have inevitable consequences on the distribution of income—as I said, inflation for the poor and financing small business enterprise are very related to growth and welfare, anyway, but we need to make sure that the metrics that are being used to guide policy are consistent with that.

Now, can this be done? How would it be done in practice? It seems to me that we need much more consistent use of data archives. In Thailand, for example, in addition to the data that I have been collecting, one could—and I am doing this in Thailand—assemble and put in one spot data from the Bank of Thailand, data from the national income accounts, which is not just national, but also provincial. There is data from a village census, from the community development department. Thailand has labor force data, it has population census data, and indeed, many of these data can be vectorized and put into a geographic information system.

I think that professionally, we tend to look at one question with one database. These questions are complex, and there is great gain to be made from having many of the datasets in front of us. In particular, as you can see, and from my comments, not just having standard macro data that one would get from a central bank or from the national income accounts but having as well the micro data, because the micro underpinnings are essential, as I hope my earlier comments indicated.

Now, I know that the Fund has really been moving very rigorously and quickly to improve the data archives in these countries, so hopefully, this is not inconsistent with that, but rather is perhaps a little more ambitious vision, and I think something that the Fund could continue to do, that it could leverage in the sense of the much used word "conditionality."

The point is that this is not just for crisis countries. This understanding is necessary for countries before they get or if they—hopefully never would—get into crisis. Nick Stern made a comment last night that it would be useful to be ready with a short, two-page document and an ordered set of priorities, and it seems to me that the ongoing analysis that would come from having these data would be useful.

The point is not just to have the data but actually to start posting on web sites the research papers and conversations among researchers about the underlying situation in these countries, especially analysis that integrates the macro and micro elements. Collecting and organizing data would not be the only role; presumably, the Fund may wish to commission special studies which highlight areas of relative ignorance where we need more work.

And finally, I would say that it is clear—at least, it seems clear—that this is very ambitious. Let's just say that I'm very optimistic. I know that it is hard work, but I know that it can be done. I think perhaps it would be a mistake for the Fund to think that it either shouldn't be done because it is too much work or that the Fund should do it all on its own. I think there are still great gains to trade of having the Fund integrated in perhaps a more formal way with networks of researchers that could help bring this vision on line.

Thank you.

MR. FISCHER: Thanks very much, Rob.

Nick Stern, Chief Economist in the World Bank for nine months now, was previously at the London School of Economics; before that, he was Chief Economist at the EBRD; before that, a professor at Oxford and a visitor also at many universities in the States, including MIT; and also has spent considerable time doing research on a village in India whose greatest claim to fame, I think, are the books that Nick has written about it.

Nick is actually reflecting that every, single one of the panelists here combines theoretical virtuosity, typically earlier in their careers, with detailed empirical work. The spirit of cooperation that Nick has brought to the World Bank's Chief Economist job is really very highly appreciated by all of us—those who knew him before, as I did, and those who have just met him now.

We are delighted that you can take part in this, Nick.

MR. STERN: Thank you very much, Stan. At our age, we have to believe that the theoretical pyrotechnics of our youth have been replaced by the ripe wisdom of experience.

This was billed as a panel discussion, but because I've got six points and only ten minutes, it is actually going to be a set of assertions. So you can think of these as panel assertions which we may have time to discuss. These assertions, I should say, are based largely on the experience of studying particular countries and not particularly from cross-country regressions, which I am happy to read, and I believe in freedom of pleasure, but it's not my favorite way of learning.

The first assertion—and I think it is something that we would share—is that inflation hits the poor, for the reasons that Rob described, and that it is best stopped reasonably quickly. But I don't think we should be neurotic about this. I would take some convincing that cutting India's inflation rate from the 5 percent or so that it is at the moment down to 2 percent is going to be of really strong benefit to the poor, either in the short term or the medium term. So one must not go over the top on these things, but I think the basic statement that inflation hits the poor is clear and strong, and anybody who watched what happened to poverty in Russia during the rapid inflation of the mid and early 1990s would, I think, see that evidence very strongly.

The second point that I wanted to make was about the two-page document that Rob alluded to, and for consistency, I will repeat a bit of what I said last night, only faster. Most crises are not total surprises, and when the crisis comes, we ought to be ready with two things. We ought to be ready with key priorities for structural and social changes at the time of a crisis—crises are times when you can make changes. And the second part of the story would be which areas or which programs do we think are particularly important to protect at a time of crisis. I think it is a fair question to expect the Bank to be able to answer those two questions at short notice of any country where it is working. That was the two-pager that Rob referred to.

Kemal [Dervis] and I were really going to push this idea strongly in the next months. Kemal went off to his own crisis, but I will have to push it anyway. It's an idea which is not yet Bank policy; I am sharing it with you in the interest of transparency and openness.

The third point is looking back on what I learned from part of my youth spent on fiscal issues, and that is that the benefits to the poor are usually on the expenditure side of the budget. It is certainly true of the UK, where we have looked at this issue in some sort of technical detail, but I think it is true of most countries that if you want to see what really benefits the poor from fiscal activities, it is the expenditure side of the budget. The scope for redistribution on the tax side is limited. That doesn't mean that you are cavalier or flippant about the distributional aspects of taxation—it is serious—but the big hitter is on the expenditure side of the budget.

So I think that strong taxation is a crucial part of the story of fighting poverty. You have to have some resources. The question of how you do it, at what levels, and so on, again is something for close analysis. But if you want to have resources to spend on programs that are going to benefit the poor, obviously, a measure of serious taxation is crucial.

And again, if you look at the area I worked on while at the EBRD, I think you could argue that taxation in the Western part of that region is too high, and taxation now in the Eastern part of that region is too low to sustain serious expenditures on social fabric and the structural side. So that's the third point.

The fourth point is that I think we now see macroeconomics—or, at least we probably have for 15 years or so—as very much about the growth on the supply side in the medium term, as well as obviously shorter-term demand issues. And from that perspective, I think we have learned clearly that growth is bad for poverty, and growth is good for poverty reduction.

It is a simple lesson. Again, just looking at countries rather than cross-country regressions, the biggest poverty reduction event in the last 15 years or so has been the 150 million, or 250 million, depending on how you count it, people who have come out of poverty in China, and that has been a growth story.

So I think it is very important to remind ourselves—it seems astonishing to have to say it, but sometimes you do—that growth is a key element in poverty reduction. So those of us who are out there promoting growth are also promoting poverty reduction. As I said, it is almost embarrassing to say it because it is so obvious, but it is occasionally questioned, so no harm in underlining that.

But of course—and this is the fifth point—you have to go on from there and say, well, we can actually think of different types of growth, and a crucial element in poverty reduction is the empowerment of poor people to participate in the growth process. And there is lots analytical that you can say about that.

So again, I won't go over it, because I stressed it last night but I think the twin tracks of a pro-poor growth policy are getting the investment climate right to boost productivity, investment, jobs and growth—and there are lots of specific details that one can bring to bear to that analysis—and the empowerment of poor people to participate in the growth process.

Education and health, of course, are key to that, but it is much more than that. It is about the whole way in which projects and programs and economic life function. And again, there is lots that we can say both theoretically and empirically about that. And I do think that a lot of the charting of the agenda of what pro-poor growth means is developing these twin ideas of what is good for the investment climate and productivity and jobs on the one hand, and on the other, what is good for participation of poor people in that story.

It is tempting to think of the former being structural and the latter being social, but that is actually to miss the richness of the interrelationships. A strong investment climate is particularly good for agriculture, for rural off-farm employment, the small firms generally. Bad investment climate hits them particularly hard. So, as it were, there is a lot of the empowerment story in a good investment climate. And if you look inside the empowerment and investing in people story, you can see that those have very powerful influences on growth itself. So while it is tempting to think of investment as structural and empowerment issues as social, they are both structural and social, and the interrelationship is very strong.

Finally, what can IFIs do about all this. Well, I would like to underline a point which Robert alluded to or made right at the beginning. I think the name of our game is largely expanding frontiers. The IFIs are agents of change, and if you follow that line of argument, that we are there to support and promote change, you see that it is absolutely crucial that in our actions we meet the costs or share the costs of reforming; we don't meet the costs or share the costs of not reforming. And again, if you look back on our record in the area where I used to work, in Eastern Europe and the former Soviet Union, I think one has to recognize that some of what we were doing was supporting the costs of not reforming. It's not what we meant to do. We certainly always meant to support the costs of reforming, but sometimes, it turned out the other way. I think it is very important that we keep that very strongly in mind, that we are agents of change.

What follows from that? Well, there are some very specific things that follow from that and are embodied in the strategy of the World Bank which we have been sharing with the Board over the last two or three months, in particular in our Strategic Framework Paper.

It means, for example, that we support strongly countries with track records of change, and I think that is a fair description of HIPC, PRSP, PRSC, and so on. If you believe in having agents of change being at the center of the story, one line is to support countries with these strong track record, and that's what we learned from the "Aid and Effectiveness" work which Paul Collier and David Dollar at DEC [World Bank's Development Economics Group] and others at the World Bank have done.

The second is the importance of the demonstration effect of projects. We had a slightly irritating description which those of you who were at the EBRD will remember (I see at least one of them back there): the plain vanilla project doesn't really contribute much. It has to have a very powerful demonstration effect, and we have to ask that of our projects, because it is only then that they will be agents of change rather than simply a resource transfer. It's a strong challenge, but it is a very important one. And again, I think that one can get serious and analytical about what constitutes demonstration effects of projects.

A third thing that follows from the agent of change is the serious analysis, where we bring our experience from outside. You have heard a lot about the "knowledge bank" and the "ideas bank". I believe in the power of the idea and that we can really bring serious analysis to the table. That's what our partners really look for. Often, they can't admit that they are reading our documents or that they have been influenced by them, but that's what they look for, and that's what we have an obligation to offer—and again, as an agent of change, that's what we have a duty to supply.

Lastly, I have to say that if you want to be an agent of change, you also have to take conditionality seriously, and again, it's surprising that sometimes you have to emphasize that, but it does follow from being an agent of change. It has to be simple, of course, and I must say that many of us from the other side of the street welcome strongly the progress that the Fund has made on simplifying conditionality. It has to be simple, but it also has to be serious, and I think that that is something which shouldn't get lost in the discussion of country ownership and aid and effectiveness. Strong though those ideas are, it's important that we don't throw out all conditionality, because I think we would be throwing away an important leverage of change and thus the crucial aspect of expanding the frontier, and thus we would not be doing our job.

So those are my six assertions. In questioning, I might be forced to co me up with a few more arguments, but let's see if we have time.

MR. FISCHER: Thanks very much, Nick, and let me just say that if there is one area where joint research between the Fund and the Bank is obviously desirable, it is this one. One of the papers in this workshop, at least, is joint Bank-Fund research, and I expect that we'll cooperate in that regard as we move forward with the research agenda.

T.N. Srinivasan of Yale needs no introduction, but I'll give you a short one, anyway. T.N. is—I'm not sure if it's a compliment, but it's how we think of it—one of the "grand old men" of development and of the policy world. He knows more about more of what goes on in developing countries and in the field of development economics than almost anyone else. He has stood for certain principles of the use of markets and of sensible macro policies throughout his career. He never hesitates to say it if asked, even sometimes if not asked, and he has been a prominent intellectual force in the change in the development paradigm that has taken place over the past 40 years, over the past 20 years, over the past 10 years. T.N.?

MR. SRINIVASAN: Thank you very much, Stan, for your kind words.

In inaugurating this workshop yesterday, Stan Fischer gave an argument for why the IMF should be involved in poverty alleviation. That reminded me of the answer that the bank robber, Willy Sutton, gave to the question, Why do you rob banks? He said: "That's where the money is." Just as Stan, in his introductory remarks, said that most borrowing members of the IMF are poor, so the IMF should be concerned with poverty alleviation.

But if you take that argument one more step, many of the poor or a large proportion of the poor in the world, however measured, live in two countries, India and China, and even within those two countries, they live in rural areas of those two countries. Now, if poverty alleviation would be the only driving force of IMF activity, and given that no institution has unlimited resources, then it should shut down everything else and concentrate on India, and that probably will have, hopefully, a large impact on the reduction in poverty in the world.

Nobody would seriously advocate that sort of policy, but let me address this issue more from an institutional framework. Since Stan has described me as "a grand old man," I will play the role of a grand old man and hark back to the Articles of Agreement that established IMF and the World Bank. These are intergovernment institutions, and the separation which was traditionally advocated was that the World Bank and the IMF suggest what the policy choices might be for the country, and the country is the one which chooses what policy it might want to pursue, whether it might want to follow the advice or not. This has a very fundamental foundation, it seems to me. As we well know, economic outcomes, and particularly the extent of poverty and the fortunes of the poor, however you measure poverty, depends on the interrelated functioning of the social, economic, and political processes. Simply focusing on a subset of economic processes, those that might be amenable to the influence by the World Bank and the IMF, would be to abdicate one's responsibility, it seems to me, toward the poor.

So what does that mean? It seems to me that both Rob and Nick have mentioned a number of things, such as, for example, that having good, sustainable, rapid growth is good for the poor; having a sound macroeconomic framework, with the appropriate inflation, taking into account the circumstances of the country, is good; having a financial sector that is functioning well, that has appropriate prudential regulations is good for the country and for the poor.

So if you took that view, the traditional role of the IMF in advising governments under Article IV consultations on appropriate macroeconomic policies or financial sector policies or what-have-you, and at the World Bank, what are the micro policies with respect to trade, social sector or what-have-you, would affect the fortunes of the poor in the right direction. You don't have to add something onto this and say, "Look, I am going to do this for poverty alleviation." You are doing the best thing that you could possibly do for the fortunes of the poor anyway if you were to establish a framework where the growth of the good kind were to take place.

Now, if you accept this separation, you might ask what is the role of conditionality, and how does the Article IV advice process come in, there are again two ways of thinking about it. My, in a way extreme, position would be to take the conditionality very seriously in both ways—in the direction of "rewarding" good performance and "punishing" bad performance. What does that mean? It means that IMF and World Bank are willing to withdraw from areas or countries where, demonstrably, the advice is not being taken, the resources being transferred are not being spent effectively. It is not simply saying that we are going to be behind you if you do the right thing; you have to also give the signals of withdrawing where things are not functioning well. Unless you do that, I don't think it will have major impact on anything, let alone poverty alleviation.

Now, it is one thing to say, Look, in your economy, certain trade policies, macroeconomic policies, subsidy policies, are not functioning well, and given your own objectives, they are not reaching the people you want to reach. Some of the public distribution of food programs in India are intended to reach the poor, and all the empirical evidence that one has is that it does not reach the poor.

This knowledge can be brought to bear—you can explain that this is what is happening. Should you go beyond that? Should the World Bank or the IMF give a signal that by its action it is going to tilt the domestic political processes toward one group, the so-called reformers, against another group, the so-called resisters of reform. In the short run, if it is successful, this may have beneficial effects, but in the long run, it will undermine the whole system of relationship between the sovereign governments and the intergovernmental agency, whether it is the IMF or the World Bank.

Of course, it is true that any policy if implemented has losers and gainers, so that advocating one policy, implicitly, you are taking a political stance. That is accepted, but the choice has to remain with the domestic political processes, and it has to be transparently seen as the domestic political processes that are choosing one or the other policies with whatever distributional implications they might have.

Now, as Nick was suggesting, if we are behind reformers, if it is pushed in the direction of intervening more than otherwise appropriate in the domestic political processes, that would be a dangerous road to take.

So where do I come out on the IMF's role in poverty alleviation? It seems to me you don't need anything more than what you ought to be doing, namely, in your systemic responsibility, so that the international financial system doesn't melt down. If that involves giving advice about domestic financial sector, so be it. But that is a systemic responsibility that the IMF has which it has to discharge if the world is to be on an even keel and to provide an environment in which the developing countries can grow effectively and alleviate poverty. Now, if you have a world financial system which you see lurching from one crisis to another, that is not an environment where sustained growth can take place and poverty alleviation will come about. So there is a very significant responsibility on the IMF on the systemic role.

Giving sound advice to the member governments during Article IV consultations requires research, and the advice has to be based on knowledge of the social, economic, and political institutions of the country and the circumstances that the country faces. So Article IV consultation is a vehicle by which sound policies at the macro level can be promoted, and that, in my judgment, without doubt will have a far greater and sustained impact on the poor than anything you might try to do on an ad hoc basis targeted to the poor here or there or elsewhere.

The last point I want to make is that comparative advantage is a great concept in economics, and it applies not only for choice of activities, but it also applies to activities that international institutions engage in. IMF getting into an activity which properly is better discharged by WHO, or IMF getting into an activity which is better discussed at the WTO, or IMF getting into areas where the World Bank is better-placed, does good to neither institution and certainly does no good to the poor of the world.

Thank you.

MR. FISCHER: Thanks, T.N.

Angus Deaton of Princeton, formerly of Bristol and Cambridge, has worked for a long time on detailed micro datasets in developing countries with the World Bank. He is renowned for his work on consumption. He was the first winner of Econometrica's Frisch Medal for the best empirical article in any two-year period, and he is one of the great experts on development and on the behavior of consumers in particular in developing countries. Angus, please.

MR. DEATON: Rob also got the same award.

MR. FISCHER: Oh, sorry Rob.

MR. DEATON: Thank you very much, Stan. Those of you who read Econometrica on a regular basis will know that writing the best applied paper over two years is not a very difficult task. Of course, getting it in there is another matter altogether.

[Laughter.]

When Stan kicked us off yesterday morning on this conference, he made a compelling case, I think, that it is impossible for the Fund to go on as before simply doing what T.N. just told it to do. On the other hand, when I listen to T.N. for more than a couple minutes I get very persuaded by him too, and I guess that's why I am an empiricist. When I can't make up my own mind, I turn to the data.

I'm not sure, though, that you can neatly separate things as well as T.N. implied at the end of his comments—in fact, I really don't think you can. Let me start with a broad issue and then make it a little bit narrower. And I do come to this as someone who has been interested in poverty, particularly the measurement of poverty over a long period of time, and who most recently has been concerned a great deal with health in one respect or another.

The first thing that I think is really, really important to realize is that poverty cannot really be treated as low income, as though it is just sort of the tail end of the income distribution, and the middle of the income distribution is what macroeconomic policy is about.

Poverty is a great deal more than that. There are many other dimensions. Inequality is obviously one, given the mean, but also education, and I think the most important other dimension is almost certainly health.

Right now in Africa, of course, a lot of health means AIDS, and indeed, as many of us I think have discovered, once you start thinking about AIDS in Africa, it actually becomes quite hard to think about anything else for a period of time.

It is true that if you look at the correlations across countries or within countries or over time, historically, there is a very, very strong relationship between any reasonable measure of health and income. They are very closely tied together, historically, between poor and rich countries, and even within countries. Rich people in the U.S. or in Britain live a lot longer than poor people in the U.S. or in Britain.

Even so, that doesn't mean, as sometimes it is taken to mean, that income is sort of a proxy for those other dimensions of poverty or of human well-being, and that if you look after income, it sort of means that the other things will look after themselves. It is just not true. I think it simply clouds our thinking if we think of these two things as being the same thing or as being substitutable for one another, which I think means that if the Fund is going to get really seriously into the poverty business, it is going to be very hard for it not to get into the health business, if only because health and income are just so totally related.

You can take this from both ends of it. One is the argument which Jeff Sachs and his colleagues have been pushing very hard, that health is the primary determinant of income and that low incomes, particularly in Africa, have more to do with the disease environment than they have to do with the investment environment, or indeed, that the investment environment is basically conditioned by the disease environment. That is certainly an enormously powerful argument right now if you look at what is happening in Southern Africa and the mass cancellation of investment plans that is going on surrounding the AIDS epidemic. You can tell lots and lots of other stories.

I'm not sure I buy into that. I think this is one of these nexuses where there is a mutual determination of health and income going on simultaneously, with income determining health very powerfully and health being a powerful determinant of income. Sorting that out intellectually is one of the great unfinished businesses of the last 50 years, and at the policy level it is even less clear where we should be intervening.

It is not even clear to me that the appropriate response—even if you believe, as Jeff Sachs does, that health is the primary determinant of the macro economy in Africa—that you want to deal with it by health measures. It might be that you want to deal with it through macroeconomic or income measures. And it might be that you want to deal with it through transfers. I think a powerful case can be made that properly directed income transfers within the country are actually a more powerful way of affecting individual health, especially in circumstances where health delivery is weak—and that is not necessarily because people spend the money on health. It is just that there are direct and health expenditure-related correlates of income.

So I find it very difficult to think about the Fund getting into poverty without addressing these questions in a very serious way. And then, when I go down that path, I begin to think, well, there is a lot to be said for T.N.'s position. It has been very difficult for the World Bank, I think, to get into these issues in a serious way, and the World Bank has a much clearer responsibility than the Fund does for thinking about these. But ultimately, bankers are not doctors.

And then you say, okay, it should be the WHO that does this. But the WHO doesn't know any economics, and that is a catastrophic problem, too. Look, for instance, at the latest issue of the American Journal of Public Health—its editorial says "Economic Policy is Health Policy" and calls for massive redistribution as a public health measure. This is for the U.S., but I'm sure they would happily argue it for other countries as well.

So you can't leave these things to WHO, and you can't deal with health as a health problem. Health is an economic problem, and economics is a health problem. There is going to have to be some integration and some outreach across the institutions if we are going to make any progress with an issue that is this difficult and this interdisciplinary.

All right. Having made the broad spiel, let me get back down to something much narrower and I think probably less controversial. I was trying to think as I was listening over the last couple of days where I would consider the IMF has a clear advantage in helping in what seem to me really crucial issues in poverty right now.

Stan yesterday talked about the Fund's interest in discovering the links between sustained growth and sustained poverty reduction and how one can devise strategies which guarantee both of those things simultaneously.

Some of us have been much concerned over recent years about what has been going on in India where, since the reforms in the early 1990s, there has been very rapid growth, at least by Indian historical standards, and actually by pretty much anyone's standards, over the last decade or so. Yet none of it shows up in the poverty statistics in India.

So we have what looks like a case of sustained growth with very little poverty reduction. If you read the Indian newspapers or the commentary in India, this is used as a rod for beating reforms in general, the World Bank and the Fund, and as evidence that the reform process doesn't work, not that there is a very clear policy on the other side.

One of the main problems at the root of that is that the poverty data are calculated from a different database than the macro data are calculated from. The macro data are calculated from the national account statistics, which show all the growth. The poverty data come from the national sample survey—certainly the oldest sample survey organization in the world, which collects very high-quality data and has been doing so for close on 50 years now.

But the means of those two series—the mean consumption from the national accounts and the mean consumption from the survey data—are diverging over time. One is showing a great deal of growth, and the other is not showing any growth. So when you look at the one that is not showing any growth, which is the survey data, of course, there is no poverty reduction, either.

So the lack of sustained poverty reduction accompanying sustained economic growth is not a fact at all. It is a statistical discrepancy, or at least, that's its current status.

Which is right? Of course, the reformers argue that the survey data are wrong, and the anti-reformers argue that the national accounts data are wrong.

It is pretty clear when you look at these in any detail that both of them are in pretty bad shape. I was really quite surprised to come to this workshop for two days and not really during that time have any really serious discussion of data reform and data initiatives on this thing. An enormous amount of work needs to be done on reconciling national accounts and on reconciling them with survey data.

What is more, this is not just a problem with the Indian statistics. This is true in China, where the survey is coming apart from the national accounts, and it is even true in the U.S., where the consumer expenditure survey and the national accounts consumption are growing apart at about one percent a year. The U.S. doesn't use that survey for calculating poverty figures, so it is less of an immediate concern, but there seems to be something of a worldwide problem, and I think it is something that the Fund would be in a very good position, jointly with the Bank, to address.

If you look back 20 or 25 years, the way the Bank used to calculate its poverty figures was by taking the national accounts, taking some inequality data from somewhere else, usually, and calculating the tail area and saying that that was poverty. Well, of course, in that case, if you got growth, then you got poverty reduction. There's no way it could come out any other way.

Since then we have collected a lot of survey data. But now we've got a situation where we've got a lot of survey data, and it's sort of coming apart. Well, we paid a lot of money for those surveys—people paid a lot of money for those surveys. A lot of them are very good surveys, but we have got to make this next step if we are going to make them really, really useful.

There were a lot of papers at this workshop dealing with inequality of various sorts. There is a tremendous literature around the world using inequality measures to explain things and explaining inequality things. The dataset that is now enormously widely used is the Deininger-Squire dataset, which is indeed enormously better than anything that was available before, but the truth is it is not very good. If you look at it in relationship to the OECD data, it gives wildly incorrect numbers which are quite at odds with what we know about inequality in Europe. This is work that Tony Atkinson and his colleagues have done.

Let me make one final point, apart from the national accounts and the survey data, but it's another data point, and that has to do with price indices. Price indices are something that bankers are really interested in and know a lot about. Controlling inflation has been a mandate of the Fund and of bankers forever. So they know a lot about inflation; inflation is one of their central things. And yet the measurement of poverty around the world is enormously hampered by the low quality of price index data, especially the fact that for most countries, there is very little reliable data beyond the national level. We don't know, for instance, in most countries whether urban and rural prices are moving together, even though most of the poor people are located in the rural areas. The Bank in its poverty numbers has separate rural and urban price indices only for China and for India and for nowhere else. There is very little at the regional level outside of make-shift series, which are often cobbled together from a few observations of a few times where people happened to collect the data.

There is very little known on the crucial issue of whether price indices are the same for the poor as they are for the rich.

There is very infrequent updating of weights for index numbers. When the Indian numbers were recently revised, the weighting base was 35 years old. It was regarded as a major scandal in the U.S. at the time of the Boskin Commission that the CPI was using 12-year-old weights. And it was a scandal. It is not a good idea for any country to be using 35-year-old weights.

So in addition to thinking about the national accounts and the survey data, I would like to put a plea in for a serious research program to give ourselves a much better base of price index numbers around the world. One could extend this to thinking of PPP numbers and the calculation of international prices, but I would be pretty content if we could do a much better job at the country level than we are doing now.

Thank you very much.

MR. SRINIVASAN: We got nowhere making that plea with the Bank for the last 30 years.

MR. DEATON: Well, let me try.

[Laughter.]

MR. FISCHER: Thanks very much, Angus.

What we'll do now is take a few questions from the floor and then ask each panelist to make a concluding one- or maximum two-minute summing up, and then I'll sum up.

Does anybody want to start from the floor? Please.

MR. [Unidentified]: A popular topic at the moment seems to be the role of microfinance in poverty alleviation. I'd be interested in the views of particularly Mr. Stern on the role of the financial institutions in this.

MR. FISCHER: Nick, do you want to fire away?

MR. STERN: I think that there is great potential in the role of microfinance, and if you look at the track records of various microfinance banks, on the whole, there have been a lot of successes. There have been failures as well, but there have been a lot of successes.

I would be very keen to splice it with the work on governance and investment climate issues, because if you look at the problems facing small firms and ask them what their problems are, you often get that it is the various forms of harassment and problems with the infrastructure that are just as important, if not more important, than the finance side. So I think it's important to marry it with an analysis of what are the things stifling the growth of small firms or farms—and I include small agricultural activities in that.

So I think there are two sides of the story. One is the detail of how you organize the microfinance side, which is extremely important, and in many ways, Rob was referring to some of those issues. But I think we have learned a lot about how to organize that.

The other side is, as I emphasized, the whole understanding of what it is that is stopping the growth of small firms, and there, I think the whole governance and investment climate issues and rural infrastructure are also very important.

MR. FISCHER: Thanks, Nick. Rob, do you want to add something?

MR. TOWNSEND: I would just say that actually microfinance is a good way to think about how things get politicized quickly and how that can become unproductive.

You have the advocates of micro enterprise finance, and group lending in particular as the miracle that will lift so many people out of poverty, and you have the skeptics on the other hand who, looking at particular problems, only seem to see subsidies.

So the point is that there are various mechanisms under which these groups may or may not be succeeding in making an improvement over what was there before.

There is actually a fair amount of theoretical modeling now that distinguishes various motives for forming these groups—risk-sharing within groups, for example, or internal monitoring. On the other hand, other people have argued that it is just selection of good clients because of the joint liability aspect.

It is possible to test these various theories with some of the data we now have and try to identify how these groups are working. I don't think there is a presumption that they work the same way in all places, even among the successful ones—and it must be said that some of them are performing rather dismally.
MR. FISCHER: Thanks. Yes?

MR. [Unidentified]: It seems to me that a lot of what you guys are discussing is the relationship between the international financial institutions and national sovereignty. And it seems to me there are a couple of issues in both Mr. Srinivasan's presentation and in Mr. Stern's presentation that are at odds, or that seem to be a bit at odds with each other.

Mr. Srinivasan, you said that one thing we should be considering is punitive conditionality. I'd like to point out that if the IMF pulls out of a given country, chances are that that country's aid package is likely to go down to zero, and chances are that that country would bend over backward to have the IMF not do that.

And then, in Mr. Stern's presentation, on your points 5 and 6—5 being empowerment and 6 being the IFIs as agents of change—it seems to me that if we are serious about empowerment, if we are serious about, say, the PRSP initiative, and we are serious that the people involved in these should really be making the decisions at a certain level, doesn't that seem to say that they should be dictating the role that the IFIs would play? And that question really links back to what is the history. That is another question that I think we are skirting around here—what has been the history of these institutions with regard to poverty, and why is it that growth rates have gone down. If we look at growth rates between 1960 and 1980 and 1980 and now, they have gone down substantially almost everywhere except China and India.

So if I could get both of your comments on that.

MR. FISCHER: Thanks very much. Are there any other questions? Yes?

MR. [Unidentified]: I would follow up on the issue that Mr. Srinivasan talked about—overstepping the boundaries, more or less. I would ask you, do you really see that the IMF is now trying to become the doctors and the World Health Organization economists? I have been appointed recently as a representative of the World Food Program to the Bretton Woods institutions, and my intention is not to exchange our professions, but to come closer together. Is it your impression that we are trying to step into each other's professions, or just to work more closely together in order to learn each other's language and paradigms?

MR. FISCHER: Thanks very much. We have time for one more question. Are there any others? Okay.

MR. [Unidentified]: A question to the panel. I think both Nick and T.N. talked about the quality of growth aspects. One quality issue which didn't come up was the quality of fiscal adjustment, very often in the context of macro programs advocated by the Fund and the need for short-term fiscal adjustment and the fiscal retrenchment accompanying it. There are direct implications for the expenditure programs, the sorts of programs that Nick would like to preserve, actually, in times of crisis. So what are the views of the panel in terms of how we could trade off the shorter- and the long-term and medium-term issues?

Mr. FISCHER: Thanks very much. I'll now ask the panelists, those who got direct questions, to respond to them and add anything else you want to. I guess we'll go in the same order. We'll go with Rob first. These are your general concluding comments as well, if you have any—it's not compulsory.

MR. TOWNSEND: Yes, I understand.

There is a lot of pressure, I think, on the Bank and now on the Fund to be more visible in terms of poverty and poverty reduction. I fear there is a great risk now that a partial and rather shallow movement will be made toward appending onto otherwise standard operations some goal for poverty reduction in these countries.

I have not heard any of the panel members saying that they would endorse that. We have all in our own way had complementary suggestions for a much more thorough review and implementation of the very productive things that the Fund could be doing that would have a direct impact on poverty.

For my part, I have been emphasizing this rather rigorous analysis of financial markets and financial institutions, something which presumably was meant to be part of the Fund's mandate anyway. But it is clear that it is not an easy thing to do and will require not only the models but the data to do it. In that respect, my comments at the end about data archives and having micro and macro data on line were certainly clarified and added to substantially by Angus. He was saying that you not only need the micro and the macro data there, but that we need to be thinking more about why they are inconsistent and try to remedy those inconsistencies.

I also think there is a danger in what I said about price stability being misunderstood as an endorsement of standard operating procedures. It was not a standard endorsement. It is no doubt true that the poor would suffer a great deal from inflation, but the point is we need a metric to judge the welfare costs of inflation and a metric for standards for monetary policy that come from the presumably incomplete markets for credit insurance that are very much in place in these countries.

So I am really calling for a rather substantial investment in research, not for an immediate change, because I fear that if something is done immediately, it will just be cosmetic and shallow. I think we are all trying to lobby for something more substantive.

MR. FISCHER: Thanks very much, Rob. Nick?

MR. STERN: First, the question on empowerment and agents of change. I think partnership is a crucial part of the story, and we have learned much more about how important it is that good policies—by that, I mean policies which benefit the poor over the medium term—be shaped and be committed to by countries if they are going to be stuck with them and if they are going to be effective. So part of the story of partnership is supporting countries following those kinds of policies, and I think that that is what the international institutions have been trying to do with HIPC. And that is why supporting countries with good track records on the kinds of policies that can lead to poverty reduction was the first on my list of the actions for IFIs being agents of change.

So too is help with analysis and sharing ideas and innovative projects. That is all completely consistent with empowerment and partnership.

The last condition that I referred to is conditionality, and there, there is a tension. You have to ask yourself the question: Is the country following policies which, taking account of all the importance of ownership and so on, are going to help the poor?. And if you think they are not, regardless of the ownership and empowerment ideas, I think it may be right to stand back. How you do that, of course, is a challenge to professionalism and our involvement, and we have to learn and try to do it better.

But for much of what I said, there isn't a tension, and where there is a tension, I think it is right, although it is difficult to handle.

You also asked about economic history over the last 50 years and trying to understand that properly in terms of the role of different kinds of policies. I think that if we do that and look at the first 15 or 20 years after the Second World War, and we take an economic history perspective, we'll see that as essentially a period of rapid recovery. And a lot of the rapid periods of growth that you see in the world are periods of rapid recovery. That started to unravel in the seventies, and the seventies and the eighties were obviously very difficult periods for much of the developing world—although, as you said, India and China both started growing in the eighties. They started growing as a result of greater openness, greater market-oriented policies. If you study India and China closely, as I'm sure you have done, and I have, I think that that is a conclusion that you would come to. We often actually date India's liberalization to the early nineties, but it has a longer history than that; it goes back to the eighties.

It's not right to somehow see China and India as exceptions, putting them to one side—there are a billion people in each, they are key examples, and they are both examples that have started to grow more rapidly as a result of creating more open and market-oriented policies, and with benefits to the poor which are clear in China and I think actually will become clear in India as the data become more readily available.

I meant to talk to both Angus and T.N. about some of those data aspects before today. I did get back from India just a week ago, where we were discussing some of those data issues directly. The 55th round data at the household level will be released, we are told, in September. We are gathering a group together at the end of this year precisely to look at how we should go forward—by "we," I mean this intellectual community; it's India's choice—on the collection of further data to try to get somewhere in resolving some of the really very serious problems which Angus described so clearly.

So we are certainly working on that one, and I suspect you'll both be involved in that story. My apologies that I forgot to tell you before this panel, and maybe we can carry on this discussion afterward. One of the advantages of having a panel discussion with your intellectual colleagues is that you get a chance to take some of these discussions forward. But it is remiss of me not to have mentioned it before.

Also, on the points that Angus raised on price indices, I think it is extremely important to push ahead. We are actually starting to pass the hat around now for a major set of activities on updating the purchasing power parity numbers; they are very old now, the weights that we are using, and we take seriously the kinds of problems that Angus was drawing attention to. So we hope that we'll have a strong move forward on that front.

But that's very broad brush, and all that PPP stuff is very broad brush, and I certainly take the point that you need to look much more closely inside countries and look at the price indices for different groups. India for long has had an agricultural workers' price index, probably not revised as often as it should be, but it is the kind of detail that you have to go down to if you want to understand what's happening to income poverty.

MR. FISCHER: Thanks very much, Nick. T.N.?

MR. SRINIVASAN: Let me start with your question about whether the IMF is trying to do the work of WHO or whether it is going in that direction. My intention was that there is expertise about specific areas of specific institutions, but as Angus rightly pointed out, WHO doesn't have economic expertise. One way to do this would be to bring in economic expertise in the WHO and have coordination between IMF and the WHO on matters in which there is an overlap, rather than IMF trying to do health policy. And the same thing applies to labor. There is an International Labor Organization. Trying to push labor in the WTO is an extremely bad idea. But that doesn't mean that one should not discuss the labor standards issue in the ILO, and there could be coordination between the institutions, each with its own specific responsibility. One institution trying to do things in which it has no competence or mandate is a recipe for bad performance. That is what I meant.

Now, on your question on empowerment, I have to differ from my distinguished colleague here. In my review of the World Bank's latest World Development Report called "Attacking Poverty," I have argued that it is no business of the World Bank talking about empowerment. This is a quintessential domestic political process issue, and there is no way that any sustainable change can be brought about by external agencies, however well-intentioned these agencies might be. So on empowerment, I differ from Nick.

On your question on when IMF requires that expenditures be cut, in the old-fashioned way—in my way—of thinking, when IMF says that so much needs to be cut in the overall budget, it is you, the government of the country, who decides whether it should be cut from the capital expenditure or whether it should be cut from the current expenditure, and within the current expenditure whether it should be cut from the health expenditure or something else, or from employing more bureaucrats. This is the individual government's decision.

If you go the other way, and give the IMF the line-item veto on the budget of each country, you undermine the whole capacity in the country for thinking through and arriving at its own priorities about expenditure. It might appear that IMF might have a better allocation of expenditure, but institutionally that's a very bad way to go.

MR. FISCHER: Thanks, T.N.. I should also tell you that T.N. has a plane to catch, and if he gets up at some point when somebody is criticizing him, it has nothing to do with that.

[Laughter.]

Angus?

MR. DEATON: I have very little to say, but let me just respond a little bit to T.N. I think one thing that we have learned in academia is that as new problems come along, and as we think of new ways of thinking about them, the institutional structures within which we operate have to change. How economic departments run their course and sometimes whole disciplines have to be reshaped. We are learning in health that doctors have to become economists, and economists have to become doctors and that it is hopeless to try to deal with these problems without that sort of deep interchange.

So I am not suggesting that the IMF and the WHO change roles, but I think that to address some of the real problems that are facing the world right now in a real way, those bridges are going to have to be crossed, and that having hostile camps, one called the WHO and one called the Bank/Fund, is not a great idea. We are really going to have to get much more deeply into that and fuzz the borders more than we have tried to do and not retreat back to what we know how to do best.

Thanks.

MR. FISCHER: Thanks very much, Angus. There is a tradition in the Fund Board, and I'm sure in the Bank Board, of a two-handed intervention, which means 30 seconds. Nick wants a two-handed intervention.

MR. STERN: T.N. has a tradition of saying outrageous things and then running for airplanes.

[Laughter.]

We have known each other for long enough that it doesn't matter. There were two aspects on empowerment. One is that any funding institution has choices over programs and projects which it supports, and it has to exercise those from the perspective of what it thinks, given its analysis and understanding, is likely to be helpful to the poor. And from that perspective, we know that if parents are involved in the running of their schools, if farmers are involved in the water users' associations, if patients and staff are involved more in the way hospitals are run, that they work better. So this is in large measure about involving poor people in the story but also about development effectiveness. It is our business to support those projects which we think are going to help the poor the most. The District Primary Education Program in India, which now covers around 55 million children, is an example of where there has been a major change in the education of girls and their enrollment in schools as a result of support for a program which was created in India, but strongly supported by the World Bank, in terms of changing the way in which education services were delivered and the involvement of poor people in the story, and that's what I think we mean by empowerment.

At the macro level, you can mean by empowerment people embracing and sticking with policies. And if you look at macroeconomic policies in the developing world over the last 15 years, they have changed remarkably for the better, and it is no coincidence that it is now the case that developing countries as a whole are growing faster than developed countries. If you look at per capita income in developing countries, it is growing about one percent or so faster than in developed, and I think that that is as a result of policy changes in developing countries which they themselves have embraced. And I think that that differential in the rate of growth in favor of the developing countries and per capita income will continue, and it is our job to encourage it to do so.

MR. FISCHER: Thanks. I'm not going to give T.N. the right of reply.

[Laughter.]

MR. FISCHER: Let me sum up. I think the questions that worry a lot of people are: "Why the Fund should take any interest in poverty, isn't that the Bank's job?"—that is T.N.'s question; and, secondly, "Why should the Fund do research on the relationships between macroeconomics and poverty?"

Let me start with three propositions which may reassure T.N. and the rest of you. First, the great bulk of research on poverty is going to be done somewhere other than the Fund; and as between the Bank and the Fund, the great bulk of research on poverty will be done in the Bank. We will never get a research effort on poverty that will be anywhere near the Bank's level.

Second, the great bulk of research in the Fund will be on topics other than poverty. We will continue to concentrate our research on good, old-fashioned macroeconomic issues of exchange rate regimes and stabilization policies and things like that.

And third, the Fund is in the process of trying to focus its conditionality, which is to say focus its areas of policy advice as well, to areas of its main expertise—macro, exchange rate systems, fiscal policy, monetary policy, and the financial sector, and in the financial sector we work closely together with the World Bank.

So those propositions are clear and will continue to be clear as we move ahead. Then the question is: what is all this about poverty? Well, first, we are active in all of our member countries, all 183. To belong to the World Bank, you have to belong to the IMF in the first instance. And most of the countries we lend to are poor countries, and problems of poverty alleviation and poverty reduction bulk large in those countries, and they bulk increasingly large in public perceptions in those countries and in the industrialized world, in the world of our major shareholders as well.

Well, fine, T.N. would say, and we used to say—we do macro stabilization, that's good for everyone. It didn't used to be thought to be the case, but it is now thought to be the case. It is conducive to growth, which we are all promoting, and therefore, leave all this stuff about poverty to the Bank—and that's essentially the T.N. argument.

But—and this is why we have to take an interest in poverty—first, aspects of things which are in the Fund's mandate affect poverty differentially. Fiscal spending, the incidence of expenditure, has different effects on different groups in the population. T.N. says leave it to the governments. This is fine where there are people who can figure it out and who do it well. But we also provide huge amounts of technical assistance to countries which don't have that capacity, and we need to have the analytical basis and the empirical basis on which to provide that advice. Of course, the Bank also does some of that.

On the tax side, we go out and we give advice on value-added taxes, for example. There are strong arguments about the distributional impacts of that. We need to have answers to those. Those are poverty-related, distribution-related questions.

Similarly, take inflation, on which there was a very long argument for many years about its impacts on the distribution of income and on poverty. We are now at a point where, as Nick said and as Rob said, it is generally accepted that high inflation is bad for the poor. And I think one of the great surprises of the last decade is that low inflation policies are politically winning policies. That never used to be thought to be the case. But we still need to figure out why. We need to figure out, as Rob said, what the relationship is between the impacts of poverty and the institutional structure, particularly the financial structure, of the country.

So for both monetary and fiscal policy, we need to know as we advise on those policies what the impacts are.

The second point—we discovered in crisis stabilization in Asia that what we do has huge impacts on poverty that are dependent on the institutional structure then in place—for instance, in Korea, whether there is or is not an unemployment insurance system—and we need, as we go in in emergency mode to give advice, to know what to do about that. If it were the case that a stabilization program in a particular environment was going to do a lot of bad things to health or to unemployment, we would need to know how to advise the country what to do about it when the stabilization program goes into effect.

You can say "Call in the World Bank", but calling in the World Bank or calling in the WHO is part of the things we need to know, and we need to get that collaboration in place in those areas in advance. So I don't think we can leave those things aside and say it is not a concern of ours, and somebody else will look after it, because if we take that view, there will be political difficulties which will affect the political sustainability of the programs that we are trying to support, and that has become clearer and clearer.

Similarly, we know that the links between growth and poverty are very complicated and very policy-dependent. We know that pro-market, pro-growth policies are more sustainable if they reduce poverty. So we need to know what pro-growth policies—and some of them relate to macroeconomic policies—are most effective in reducing poverty.

That is the intellectual basis. Now let me give you a political argument for why we should care. Policies will not be sustainable in poor countries if they are not equitable. Policies will not be sustainable among our financiers, among the countries that provide most of the financing for the IMF, if they are viewed as anti-poor in the countries where those policies work.

We took tremendous heat—unfairly, because I think it was not consistent with the facts—over the impact of the Asian crisis on poverty. That was a tremendous factor in the debate over whether the Fund should get more financing-the perception that we had supported policies that hurt the poor. I think it was vastly exaggerated, because the facts were not as they were portrayed at the time—the impacts were less.

It is not a defense to say, "Sorry, the World Bank takes care of that." It just doesn't work.
T.N. says, well, Stan made the argument that that's where the money is. Well, it is where the money is. And this institution needs money to survive, and that's part of the story. So on a political basis, we cannot avoid worrying about that.

There is another argument, which is morality. But everybody finds it more convenient—more comfortable—arguing realpolitik rather than morality. The fact is both go in exactly the same direction for what they say about the Fund and the fact that it needs to take an interest in the impact of what it recommends on poverty, without necessarily being the institution that undertakes the policies that will deal with that. The Bank frequently is the responsible institution.

The second question: Why should we be interested in research on poverty? Well, I've got a lot here, but the case was eloquently made by Rob Townsend a few minutes ago. If we care about poverty—and we do and we should—then we need to do research on the links between our policy environment, our policies, and poverty. Why? Because whether we know it or not, every institution, and this one not least, operates on the basis of an intellectual framework. Sometimes you have done it for so long you don't even know it, but you're doing it. And no such framework is ever satisfactory; no such framework is ever complete. As Angus said, they become less and less relevant. They need to be updated all the time. There are always new questions coming up. And if we as individuals and we as an institution don't have the capacity to rethink answers to questions of the moment, then we are bound to lose our effectiveness.

Well, why not leave it to outsiders? You should leave most of this work to outsiders. The ability of outsiders to see through issues independent of the institutional blinkers that we have is extremely important. We have to rely on what outsiders say, what other institutions do, but we also need the capacity to absorb the research that comes from outside, to reshape the results, to apply them to what we do. And for that purpose, we need a research capacity in this institution, and we need people in this institution to be working on these issues so that they can interact with people of the caliber of those who have been on this panel.

Let me conclude by thanking everybody who has helped to make this pre-conference the success that I believe it is from reports I have heard about it.

First, there are a lot of academics who have come from outside, either to coauthor papers or to comment on papers. I think, actually, that I would like to thank them by name—Joshua Comenetz, John Cuddington, Oded Galor, Toshihiro Ichida, Michael Keane, Nora Lustig, David Sarn, Xavier Salai-Martin, Eric Thorbecke—and if I left you out, I'm sorry, but thank you all very much for coming.

I thank our World Bank colleagues, coauthors of papers, discussants of papers, who have worked together with us.

I thank the panelists, who have given us such an interesting afternoon.

I thank the IMF staff who are participating in this research effort—it is not a research project—on macroeconomics and poverty and who have written papers. It took something to get this going. I think the institution should be very grateful to you, and the Management certainly is.

And I thank the organizers, Mike Mussa in his overall role as Director of the Research Department, and the committee for this conference, Paul Masson, Ratna Sahay, who really got it moving actively at the early stages, and Paul Cashin, Paolo Mauro, and Cathy Patillo; and Rosalind Oliver and Sylvia Brescia, who made it all happen.

So thank you to those of you who came to the workshop, thank you to those of you who have come to this panel and to all who have made it a success.

Have a good weekend.

[Whereupon, the proceedings were concluded.]