Transcript

International Monetary Fund
Multipurpose Room, B-702
Thursday, March 15, 2002
Washington, D.C.

View or listen to this press conference using Media Player

Panelists:

Anne Krueger, Chair, (First Deputy Managing Director, IMF) 3

Nicholas Stern, (Chief Economist, World Bank)

Santiago Levy (Director General, Instituto Mexicano de Seguro Social)

Nancy Birdsall (Center for Global Development)

Montek Ahluwalia (Director, Independent Evaluation Office, IMF)

Proceedings:

MS. KRUEGER: Thank you very much. This is the closing panel of the seminar conference on poverty and macroeconomic policies. This is, of course, a subject of great interest to the Fund because our primary concern—or our primary competence, I suppose it would be better to say, has always been macroeconomic policy, and good macroeconomic policy, as has been well documented, is essential for economic growth. It also is essential if one is going to avoid very high rates of inflation, and there is plenty of evidence, of course, that both of those things are directly and importantly linked to poverty alleviation. To a first approximation, real incomes go up approximately at the same rate across the spectrum. So to that extent, the Fund's business has been the business of poverty alleviation from the start.

On the other hand, there are things in addition to simply raising real incomes that can make a difference. And, of course, there's lots we need to learn both in terms of what has been effective and what has not been effective.

Given the Fund's past competence, I think we tend to stay very much in our sphere, hoping that we can make sure that the programs that are put in place are very effective in raising growth rates, and in targeting the poor, we rely, of course, on partnership with the World Bank for some of the more micro issues that are important in that regard. But there's plenty yet to be learned both by the Bank and by us as to what could be done, and that, of course, is the function of this conference.

So, without further ado, what I'm going to do is ask each of our four panelists to make an introductory statement of 10 to 12 minutes, and then after that open up the floor for questions, and then we'll have a final round at the end with the panel. After some brief haggling, we've finally got an order for the four speakers, and the first one is Nicholas Stern from the World Bank.

MR. STERN: Thanks very much, Anne.

As you emphasized, the Bank's objective is indeed directly the reduction of poverty, and it's been very clearly and explicitly that, at least since 1974 and Robert McNamara's famous and fundamentally important speech in Nairobi.

I understand that Mr. McNamara is here today, and I'd like to take the opportunity to pay tribute to him for that fundamental move in changing the way the Bank works and what its objectives are.

In my remarks, I want to really do two things. I want to talk about the design of policy—and I'll specifically talk about pro-poor growth—and the assessment of policy, which is particularly poverty impact analysis. So design and assessment, and they're obviously complementary, and they should obviously be based on evidence. So let me take policy design and particularly look at that in terms of pro-poor growth.

Now, at the Bank for some time, and certainly in our strategy that was endorsed by the Board in the early part of last year, we have emphasized two basic pillars: the investment climate for investment, productivity, entrepreneurship, opportunities, jobs. That's the key first pillar. The second is investing in and empowering poor people so that they can participate in the process of growth.

Both of those pillars are clearly dynamic. They're clearly interrelated and mutually supportive. They're both very heavily involved in processes—processes where learning and history and adaptation are all important. And they're both focused on, as it were, the environment for decisionmaking and not simply the allocation of resources to, say, physical capital or human capital. They're both about the environment in which decisions are taken.

The first of them I think quite clearly embodies a kind of shumpeterian approach to the whole investment process, placing entrepreneurship and the environment for entrepreneurship at the center. So that basic two-pillar approach is what we're following at the Bank, and I think everything we do should be and increasingly is subject to the test of whether it's making real advances on one or the other or both of those two pillars. And if you look at the language of President Bush's speech at the IADB yesterday, I think you'll find that kind of approach quite strongly reflected there also, and that is, of course, extremely welcome.

Now, I haven't obviously got time to develop the analysis of economic policy based on that approach in any detail, but it is possible to drill down from that basic two-pillar approach into the hard empirical and theoretical analysis that should support policymaking.

One of the things that you clearly realize when you do that is the interrelationship between those two pillars. There's strong evidence, for example, that educational achievement, including enrollment—on the way to that, of course, including enrollment in schools—depends strongly on key aspects of the investment climate.

In looking at the investment climate, what we do essentially when we're drilling down is to divide into three: macro and trade policy, which, of course, are of fundamental importance, as Anne emphasized; governance and institutions, the second part of the investment climate; and, thirdly, a functioning of the infrastructure. All three of those, of course—macro, governance and institutions, and infrastructure—place strong emphasis on the role of the government, even though we are focusing in looking at the investment climate on private sector—mainly on private sector entrepreneurship.

But if you look at education, you find from evidence in Eastern and Southern Africa that electrification plays a powerful role. Evidence in North Africa, you find that transport plays a powerful role in getting children, particularly girls, to school. Lots of evidence from Central America on the role of nutrition in education and so on. So looking at all these things, we have to be very careful to look at the interactions.

Before I go on to poverty assessment, let me just emphasize the kind of data that you need to look at the questions I've described.

First, you need decent household surveys, and the Bank has for long been working on the promotion of good household surveys, with the Living Standards Measurement Study being a shining example.

More recently, we've intensified the work on firm level surveys, not just collecting data on inputs and outputs and prices, but much more—we do do that, of course, but much more fundamentally looking at the problems in the investment environment facing the firm. That's something we've been doing within the Bank, in collaboration with partners, particularly at the EBRD and so on. So those firm level surveys now are becoming pretty standardized across countries in terms of some of the principles and questions that are asked, and we're getting very quite strong evidence now on the role of the investment climate.

The third set of data, also micro, is that the service—is on service delivery and looking at the functions of different kinds of agents and units, the way schools function, the way hospitals, clinics function, the way water delivery services function and so on. So what we're doing here, both with the firm surveys and with the service delivery surveys, is looking at the detail of how decisions are made and try to understand why. So if firms complain that bureaucratic harassment is a big problem, we're able to identify that and see what sort of bureaucratic harassment and then relate that to the behavior of different aspects of government, and similarly with weak infrastructure and so on, to try to understand how those difficulties come about and, therefore, to formulate policy to change them.

So we're getting really quite powerful results there, and that is the way in which, through these kinds of data sources, we're trying to deepen an understanding of pro-poor growth from the perspective of the two pillars I described.

Now, let me just say something fairly briefly about policy assessment and poverty impact assessment in particular. This is something which NGOs and our shareholders have, quite rightly, asked us about. If we declare for reduction in poverty, which we do, of course, we ought to be able to assess ex ante and ex post the impact of our policies on poverty.

We must be very careful, of course, to do that in a long-term, dynamic way as well as short-term. You cannot, for example, look at the impact of education on poverty and income distribution by just looking at what happens over the next couple of years. The impact of education on poverty and income distribution obviously has a very long gestation period. So that's just one of the problems.

Similarly, if you're looking at policies to improve the investment climate, then some of those, because they affect the kinds of decisions that firms make and the environment for entrepreneurship, they're going to take some time to come there as well. And they're going to have manifestations which are actually quite subtle relative to the kinds of models we're currently able to use.

But what we thought was the right thing to do was to start by assessing what kinds of models there are already that we can use, so one doesn't want to bemoan the lack of models when, in fact, there are quite a lot of models out there already, and what we should be doing is make the best use, with proper judgment, as to the quality and effectiveness of those models and what they can offer.

We're in the process of developing and we hope to provide, as it were, a little toolkit for the economists in the region within the next two or three months, a little toolkit for analyzing poverty impact of different kinds of interventions, both macro and more structural and micro.

This grew out of the gathering which I'm sure a number of you attended last year, which was organized by John Page at the Bank and Masood Ahmed of the Fund on poverty impact assessment, which Stan Fischer and I chaired. So this is a program which has been going for a while of collaboration between Fund and Bank on this crucial issue, which we welcome very much.

Now, there's a paper on this which I think has been made available to you, written by largely Francois Bourguignon and Luis Perreira Da Silva (ph), with some cooperation from me. And that sets out in a systematic way the kind of techniques that are now available. It looks at that from the point of view of micro incidence, from the point of view of macro impact, and the third layer, if you like—in some way it's the most difficult one—the intersection between the two. It leans heavily on household surveys as a technique.

As I say, I haven't got the time to go into the techniques which were reviewed there, but I think it is a strong basis for the kind of toolkit that we'll be making available to our economists in the field over the next two or three months.

Fundamentally what we emphasize there, after reviewing—and there are lots of different useful ways of doing these things. Fundamentally what we emphasize is the importance of combining the macro impact analysis with the micro simulation, the micro simulations through household data. Any of you have worked seriously on poverty know that when you get into the micro surveys, you find enormous variation within them. To look just at blocks here and blocks there does not really reveal to you what is going on. And one of the challenges I pose—and obviously we travel a lot in these jobs. One of the challenges I pose to our people in the field, whether they be the IMF resident representative, who I usually meet, of course, or the World Bank country director, is: Who are the poor people in this country? What are their characteristics? Where are they? The good ones can give an answer, but there are others, and I think that the use of the micro data here are absolutely fundamental in simulating the poverty impact effects of what we do.

As I've mentioned, there are lots more dynamic problems down the track. Our modeling capabilities are limited. But they're not zero, and I think we can produce for the people in the field a reasonable toolkit, at least in the short run. But it's more than that. It's a full program of research because the effect of our policies on poverty reduction should be what we're here doing all day every day from the analytical side. There's no sense in which one review paper or one toolkit can close it. What they can do, I think, by trying to do things in a systematic way, is help chart a research agenda for poverty impact assessment.

Thank you very much, Anne.

[Blank spot on tape.]

MR. LEVY: Thank you. I'd like to begin by thanking the IMF for the opportunity of being here in this conference, and probably center my remarks mostly on some of the topics that were talked about in the day and a half on safety nets and the relationship between safety nets and macroeconomic crisis.

I just want to make a few caveats before I begin, which is when I'm thinking about safety nets, I'm thinking about safety nets mostly for people in extreme poverty, and I'm thinking mostly about the problems of middle-income countries. I don't know sufficiently about very, very low-income countries to make these remarks generally, general for them.

The premise here is that macroeconomic shocks are unpredictable and up to now unavoidable, and that maybe in the future hopefully policy will improve substantially, but that we should count on the fact that maybe in countries ahead they will be facing very different types of macroeconomic shocks.

These shocks affect the poor through very many channels. We saw some papers in the last two days here about some of the channels through which they affect them. But it's not easy to make generalization because, first, the shocks vary from country to country. Sometimes it might be a fiscal crisis, sometimes an exchange rate crisis, sometimes it's contagion effects from other countries. Sometimes it's an oil shock, sometimes it's terms of trade. The crises vary in many ways, and the transmission mechanisms of the crisis to the poor vary, again, from country to country and from experience to experience. That's why I think safety nets play an important role on trying to protect poor people from these macroeconomic shocks that, again, we would not like them to but most likely they will be part of the future agenda for the next ten years or so.

Most of the papers in this conference centers on the reaction of these macroeconomic crises on consumption and various indices of inequality Ginis, on poverty indices, on some Foster(?) indexes, and there were some examples for Indonesia, Ethiopia, and Mexico. I wish there had been a little bit more work on the reaction to public consumption and public expenditure.

Clearly, one of the important components of most macroeconomic adjustment programs is an unavoidable increase—an unavoidable reduction of the budget deficit and the search for fiscal equilibrium. And in most cases, it's a combination of either tax increases and budget reductions, and the budget reductions that take place are very important for the long-term effects on the poor.

We have seen quite a bit on private consumption, but the emphasis also needs to be put on what are the adjustment implications on the provision of public goods, particularly public health, nutrition, education, because these not only have immediate effects on consumption, but they have long-term effects on the poor.

And the reaction of the public sector to the crisis can make the crisis worse from the point of view of the poor, or the reaction of the public sector to the crisis can actually make the situation a little bit more manageable.

From this perspective, I think that safety nets play a double role, and I want to discuss a little bit this double role. They serve as a mechanism to generate income transfers to the poor and protect their income and allow them to do a little bit of consumption smoothing. But they can also serve a double role as a mechanism for investment in human capital and, therefore, avoid the long-run effects that these shocks have on the poor.

Now, if you look at some developing countries—and, again, I don't want to generalize, but at least the case of Mexico and very many other countries in Latin America—what you see is that there are multiple mechanisms for income transfers. There are tax exemptions for some basic commodities. There are sometimes price controls on telephones. There are sometimes price controls on transportation. There are food subsidies. There is sometimes free delivery of goods. There's a multiple set of mechanisms through which these mechanisms work.

But they all at the margin are mostly (?) marginal transfers, and they basically try to transfer income to poor people through all these mechanisms.

Now, if the countries that you're looking at have a very unequal income distribution—and these are the characteristics of many countries in Latin America, that income distribution is very unequal—what you're going to find is that most of these mechanisms for income transfers are extremely ineffective and inefficient. They're inefficient because if you subsidize broad categories of goods—tortillas, bread—broad categories of consumption, then the Lorenz curve for most of these broad categories of consumption is not going to be very different from a 45-degree line. And a lot of the subsidy is going to be captured by non-poor groups, which you don't want to be captured.

If you instead go into subsidizing very narrow categories of goods—coarse tortillas or just this type of bread—then they're going to be extremely ineffective as opposed to inefficient because the poor only devote a small share of their budget to that, and the income transfer that can be associated through the subsidy of the particular individual good is not going to be very large.

In addition, these entail very high administrative costs. They sometimes lead to a lot of arbitrage, not to say corruption. And of the share of the budget devoted to most of these food subsidies or, you know, subsidies to transportation or whatnot, a very large percentage is taken up by administrative cost, and the actual transfer, the actual income transfer received by poor people is very, very low.

That's why I think there is some solid arguments to move strongly in the future towards monetary transfers as opposed to subsidizing individual goods or individual commodities of goods or individual public services, because I think what we could achieve is delink the relative price structure and consumption taxes from income transfers.

If you delink consumption taxes and if you delink relative price structure from subsidies, then these can be used strictly on the basis either for generating revenue for the government, which helps the adjustment program, or for generating efficiency and competition, avoiding cross-subsidies. And if you put income transfers in monetary terms, they can be much more efficient because the amount of transfer that you get per fiscal piece of revenue spent is substantially much higher.

Secondly, they have much lower administrative costs, and the chances of diversion to other users and arbitrage are substantially reduced. In some cases, when these income transfers are given in rural areas, they can also stimulate the local market and avoid the problem of dumping that sometimes occurs when goods are brought in at below-market prices into rural areas from urban areas. And they can actually in budgetary terms—and this is an important point from the macro point of view—imply less money, less revenues than the money that is actually being spent through a multiple set of mechanisms in about five or seven different categories of, you know, subsidies to food, subsidies to local telephone rates, subsidies to transportation, subsidies to whatnot. So that they can from a macro point of view be smaller, and, therefore, when the time for a budget cut comes—and the premise here is that inevitably it will come. When the time for a budget cut comes, they can be protected from the budget cuts.

Now, in arguing for transferring and changing all the structure of subsidies and consumption exemptions and consumption tax and all that, for a very clean relative price structure and putting all the transfers into—monetary income transfers to poor families, I think you have to associate and think a little bit also about the incentive effects that those income transfer might have.

In some cases, you do not necessarily want to associate income transfers with a particular incentive or a behavioral condition. Brazil had a very interesting experiment about seven years ago. They decided to legislate transfers for very poor people in the rural areas as part of social security for people who are 60 or older, which given life expectancy in the northeast of Brazil is fairly old people. And there was no behavioral contingency put into the transfers.

In most cases, however, I would argue that you want to make income transfers conditional on some kind of behavior by the poor people that are receiving these transfers, and, therefore, you can get a double role if you make the conditions associated with some sort of investments in human capital or some sort of work that generates future assets. You get a role in generating present consumption, because you get an income transfer giving to these people, but you also generate conditions for future consumption because you create assets, either individual human capital or public assets, public goods in the works that are being done through them.

In Mexico, we've been transferred—we've been transiting in this direction over the last five or six years. Over the last five or six years, we have eliminated all generalized food subsidies. We've eliminated quite a few cross-subsidies at a very different rate. And we've created a program named Progresa (ph), which basically tries to give an income transfer in monetary terms, in cash, to families that is indexed on the consumer price index, so, therefore, it's protected from all the shocks. It can be fine-tuned in terms of a crisis because you can actually increase the amount or lower the amount. And, amazingly, we are now reaching almost 20 million Mexicans through the program, which is about 20 percent of the population, and the cost of the program is one-third of 1 percent of GDP, which is something that you can then protect in the case of macroeconomic crisis and not go into the sort of trade-offs that were put into some of the papers before, which is you have to cut, and if you cut at the margin, are you going to let more people in or out.

The point here, we have to have these programs, that you actually do not cut, even in the case of the program, because these programs are taking one-third of 1 percent of GDP, is something that you can have.

Let me close my remarks by making three political economy considerations on this.

First, if safety nets are going to be safety nets, safety nets must be there in advance. So, therefore, what I think they should do and I think is an important point of the agenda for IFIs is to help countries set up safety nets before the crisis. If you're doing it through the crisis, it's going to be too late. It's going to take you three years to get it running, and by then it is not going to work very well and people are going to suffer a lot. But if in times of peace, so to speak, in times of tranquility in a macroeconomic scenario, one can put into the work agenda of these countries setting up well-defined, well-structured—with the right incentives—safety nets, it can help a lot.

It could also create new tools for making transfers that can help adjusting not only to macroeconomic shocks, but also to trade shocks or to other kind of price reforms or even tax reforms and can enhance the toolkit that we currently have through finer mechanisms of reaching it. So the timing is a very important problem.

Secondly, maintenance of the programs is very important in times of macroeconomic crisis. I think here also IFIs can help a lot if we are much more careful in identifying exactly what are the programs that we want to protect in the budget. Social spending is not a very useful category. Even spending in health services is not a very useful category because in many cases what happens here is that wages take a big chunk of that, and sometimes the unions get a bigger chunk of that. And then if you cut the budget of health care and what you end up cutting is the provision of the inputs, even though the budget cut doesn't look very large, the actual pain that you inflict on people can be very large. So one has to be very, very careful in identifying these items within the budget and protecting them.

And a question for the audience is whether, in fact, conditionality in adjustment programs not only should be conditionality on the size of spending and the size of the budget, but maybe one can think of—and this is speculation—setting up conditionality on the protection of very carefully defined individual programs that don't represent a big chunk of the budget and that are going to set up a safety net and a medium-term protection for the poor.

And the last comment that I'll make—and I'll conclude with this—is that safety nets have to be permanent.

Now, in most OECD countries, this is done through social security legislation. The question is whether in some LDCs we are ready to start making a commitment, a more permanent commitment to poverty alleviation and poverty protection by legislating safety nets.

Many of these countries are going through democratic transitions, and now what you would like is that these transfers not be the unilateral act of the executive power, but maybe be the act of the legislative power in a societal consensus in favor of protecting the poor, which is what really implies putting it into a law and making it permanent, and this would avoid the very volatile nature that these transfers have had in the past and would allow extending the concept of social security to protecting the very people who need it the most and because of the nature of social security today are excluded from that protection.

Thank you.

MS. KRUEGER: Thank you, Santiago.

[Applause.]

MS. KRUEGER: Our third panelist is Nancy Birdsall, the founding director of the Center for Global Development. Nancy?

MS. BIRDSALL: Thank you very much, Anne.

I was thinking, as Santiago finished, that some of what I'll say will sound inconsistent with what he's saying, but he saved me by invoking democracy there at the end. And the other savior for what will seem some inconsistency between his optimism that relatively poor countries can succeed with safety nets and what might come across as a little bit of my pessimism, at least in the short run, the other reason is that he is not telling you that he is the architect of the program he's describing, that it is exceptional, and that I would still worry that if and when he leaves government, there won't be the protector in the executive branch for this program.

Now, I have three propositions that I'd like to lay out, and I'm talking about things I don't know enough about, mostly politics, but I think it's still—I still may have thought more about it than many of the people in this room—at least the IMF staff. So let me first—

[Laughter.]

MS. BIRDSALL: Let me start with the first one, which is that politics matters. And that's just another way of saying in the new world of the Poverty Reduction Growth Facility and the Poverty Reduction Strategy Paper that ownership matters. So it helps if the IMF and the multilateral banks and the economists understand the technicalities of how to reconcile necessary macroeconomic policy with socially responsible macroeconomic policy. That helps countries. But ultimately I think we've got to rely on ownership by the citizens of developing countries and the accountability of their political leaders if we're going to end up with what I think is good shorthand for what I'm talking about, namely, a socially responsible mix of macroeconomic policy.

Now, that's my first proposition. It's kind of mundane.

The second proposition is the following: In most developing countries, there's something that I would call a missing political alliance between the poor and the non-rich. Now, who are the non-rich? They are a lot of people. This is what Lant Pritchett was describing in his remarks. They are the people who were poor a year or two ago and are not poor now, the people who will be poor a year or two hence and are not poor now. They are the people who are at risk in the working class, especially. They're not exactly Marxist proletariat because most of them in middle-income countries and in poor countries are in the informal class. They're the rising, emerging entrepreneurial people who run small businesses and who rely on not too tight an economy in order to have jobs and to generate jobs. The point is that they're most people in most developing countries.

We know—and maybe it came up in the discussions, for example—a good paper on China showing that if you looked in the mid-'90s at the number of poor or the proportion of poor, it was 6 percent. But if you took into account how many households had been poor over three or four years, it was closer to 50 percent.

In Brazil, we know that people at the median income are a lot poorer than you would think looking at average income per capita. In Brazil, average household income per capita in surveys in the '90s was about $4,000 in PPP terms. But the median was closer to $1,700 in PPP terms. And the education level of the top—of the people, the heads of households in the 90th percentile of income in Brazil in the mid-'90s was only about four years. So we are talking about—you know, Brazil is on the extreme because income inequality is very high. But it's also one of the better-off middle-income countries.

So the point I want to make is that most people in most developing countries are at risk in one way or another, even if they're not poor today. And most of them would, therefore, like the kind of safety rope approach that Lant Pritchett referred to, to prevent them from falling into poverty, not just a safety net which would alleviate their misery after they have become poor because of some kind of shock or crisis or because of the way the economy is failing to work.

So safety rope programs, as he described them, make more sense economically in terms of protecting those who are vulnerable, and in democratic societies they ought to be more politically attractive since they would appeal to a much larger group. But we don't see any sign, oddly enough, of this missing political alliance in most developing countries between the poor and this huge vulnerable group of non-rich.

Why? One possibility is to try and understand where this alliance, this implicit alliance came from in Western economies that now have safety rope programs—unemployment insurance, all kinds of social insurance that affect a large swath of the population.

In the Western economies, it surprised me to learn that these safety rope programs really only got started in the 1950s. In 1940, Norway spent 4 percent of its GDP on social insurance type programs and the U.S. 2 percent of its GDP. Those numbers tripled in the 1950s, which is when average per capita income in PPP terms rose above $5,000 per capita. And those numbers doubled again as a proportion of GDP between 1960 and 1995. So the big surge came when average income was around $5,000 per capita.

Don't forget, I said the median income of—the income of the median household in Brazil in these PPP 1990 terms is about $1,500 per capita. And when in the 1950s the Western economies started to expand their safety rope programs, the median income was much, much higher.

My third proposition is that the standard recipe that we have been talking about for the last day or two abstracts too much from the missing politics of this missing alliance. The way we're setting it up conceptually and, to some extent, in the programs that we talk about is pitting the middle or this big nebulous near-poor class against the poor. That's why Mexico is an exception because, with Santiago's excellent architecture, they've found a way to spend only one-third of 1 percent on this really highly targeted program. But I would emphasize that it is the exception.

Now, let me close with a couple of comments on this third proposition, first, for normal times in countries, and, second, during crisis.

First, with respect to normal times, countries with low and medium income—with low incomes have a difficult time implementing pro-poor programs because pro-poor programs, with the exception of Mexico, are fundamentally a luxury good. They didn't really get rolling in the safety rope sense in the Western world until income was much higher than it is in most of the developing world. So a few countries in the developing world in normal times are getting there, but I think we have to be much more creative and much more sympathetic about the difficulty politically right now for most poor countries to manage safety rope programs.

What about in times of crisis? I think the problem is even greater. Many countries dealing with a crisis have to respond with contractionary monetary and fiscal policies. This exacerbates the political tension between the poor and the near-poor or the non-rich. For one example, there is a tension between monetary tightening which hurts the middle group, the rising entrepreneurs, those running small businesses in the informal sector, disproportionately; and fiscal tightening, which is the alternative—or we want some mix of monetary and fiscal tightening—fiscal tightening which hurts the poor disproportionately.

Martin Ravallion's paper made it clear that the poor are the last to gain from fiscal expansions and social programs and the first to lose from fiscal contractions.

Second, when we compensate fiscally with a targeted program, as in Mexico, even if it works, it probably at the margin, at least, implies somewhat higher interest rates. And those higher interest rates affect this at-risk middle group who are the most likely to have borrowed in times of pre-crisis growth and who are the most likely to need working capital to sustain and main—to use their assets and avoid stripping their assets during the crisis.

Furthermore, the adjustments that they are often faced to make, the countries in this combination of monetary and fiscal policy would never survive politically in rich countries because of their impact not on the poor but on this big middle group.

When Brazil risked financial crisis in 1997, real interest rates increased—real interest rates increased to 30 to 35 percent and remained there for about five or six months. Maybe that was in '98. I'm not sure.

During the Volcker years in the U.S., when real interest rates skyrocketed, for those of you who are old enough to remember, and Carter lost the presidency because people didn't like what was going on with their variable rate mortgages, real interest rates peaked, at the most, at 9 percent, briefly, for a month or two.

So let me conclude saying these papers, this conference, this attention of the IMF to poverty, it's an enormous step forward that I think is reflected in all of the work of the IMF and the World Bank and the regional development banks in the last decade. But I think we have new challenges. In terms of research, I think a lot more attention has to go to the middle quintiles in poor countries, how they're affected by different mixes of macroeconomic policies. That's kind of from a program point of view. I'd say that safety net programs can work in countries like Mexico, but to be sustainable politically, even a safety net program probably has to be complemented by safety rope programs so that the vulnerable middle gets on board politically, and yet those safety rope programs are a luxury good for most countries. And, second, that in the IMF, I suppose I would say that we've got to move on, or I would urge the IMF to move on from poverty and macroeconomic crisis to bigger thinking about politics, this middle group, and macroeconomic policy, particularly during crises.

Thank you very much.

[Applause.]

MS. KRUEGER: Thank you, Nancy.

Our final panelist is Montek Ahluwalia, who is Director of the Independent Evaluation Office here at the IMF, also a founding director. Montek?

MR. AHLUWALIA: Thank you, Anne for the introduction. I should actually clarify that nothing that I say is being informed by what I have learnt as Director of the Independent Evaluation Office, since the work has only just begun. I was asked to comment pretty much from the perspective of experience in India, and I must say that I agree with a great deal of what has been said, so my remarks will be organized in a slightly disconnected way to make four or five points, which touch on some of issues that have also been discussed by the others.

I'd like to make a distinction between longer-term issues and shorter-term crisis management issues which have come up in the discussion.

On the longer-term issue, commenting from the perspective of a borrowing country, I must point out that there is a bit of a disconnect between the way objectives are presented and formulated by international financial institutions and the way they are formulated domestically.

Internationally, the focus is exclusively on the objective of poverty alleviation, and I think the reason for that is the perception amongst the shareholders of the institutions that the principal objective of these institutions is to serve a public good, and the public goods usually identified for the Bretton Woods institutions are handling market failures at times of crisis management, which is clearly one class of public good function, and the other is poverty alleviation, which is clearly seen to be an international public good. The more general issues of development increasingly tend to be seen as capable of being handled through market mechanisms. It is, therefore, logical that the World Bank focuses on poverty alleviation as the principal international public good which it is promoting.

From a developing country perspective, however, the development objective includes poverty alleviation as a very, very important objective, the more so in countries where there are very large numbers of poor people, but the development objective is formulated not only as reducing poverty—not that that's an easy thing to do by any means-but in terms of a much broader definition of development, and growth is central to the formulation of development objectives, from this perspective.

If you read the documentation that comes out of the international institutions carefully, the central role of growth is recognized, but it is recognized in a manner which I do not think reflects adequately the way developing countries view it. It's recognized as kind of an instrumental variable, that growth will deliver poverty reduction and, therefore, we must be for growth.But countries are interested in a much broader sense of social and economic transformation, going beyond poverty alleviation and driven by a very firm conviction that this cannot be achieved unless there is a rapid growth.

It is entirely appropriate for international institutions, whose existence is predicated on the belief that the principal public good they are serving is poverty reduction, to focus on poverty reduction but from an ownership point of view they have to recognize that developing countries will have a much broader objective. They can then target their own activity to assist in this process, but focusing very largely on what to do for poverty alleviation.

Now, looking at it from this perspective, I am concerned that the total picture that comes out of some of the literature often appears to understate the role of growth. At times it appears to warn countries: don't get carried away too much by growth because if you are not very careful, you actually won't reduce poverty.

If we go back historically, I think—well, Nick mentioned that [former World Bank President] Bob McNamara is here, and he will certainly recall that when he delivered his famous 1973 speech in Nairobi, the intellectual world at that time was very much dominated by an article by Al Fishlow in the American Economic Review commenting on Brazil, in which Al had said that although Brazil grew at almost 9 percent in the '60s-and that sounds kind of incredible if we look at more recent growth rates, but that is exactly right, just under 9 percent in the '60s—there was a massive deterioration in the income distribution of Brazil. Al never actually said that it led to worsening poverty, but there was an implication questioning the use of such rapid growth if the income shares of the poor are really going to be compressed so much.

In retrospect, the fear that rapid growth may not yield benefits to the poor has proved to be misplaced. All the evidence we have for the next 20, 30 years suggests that there's no known case where the achievement of rapid growth has not led to a massive reduction in poverty. Unfortunately, I don't think this is adequately being projected, and I think it's important that that projection comes across.

Let me make a few comments on India's experience in this specific context. India entered into a process of economic reforms in '91. It was obviously controversial. The reforms were projected as designed to accelerate growth, but there was a lot of criticism that perhaps they would not do a good job of poverty reduction, and interestingly enough, through the 1990s, there was a great deal of uncertainty as to what was actually happening on poverty.

India is unique in having virtually annual surveys conducted by the National Sample Survey organization, but the quality of these surveys varies. There are annual surveys,which are called "thin sample" surveys, and every five years there are larger "full sample" surveys.

It turned out that a lot of these thin surveys were generating information which appeared to suggest that, although there was as lot of growth, there wasn't much poverty reduction. Many of us had argued that we should wait for the full survey of '99-2000, which has now been released before coming to a conclusion. There has been a huge debate in India on the comparability of these surveys,but the latest one was examined in great depth, by Angus Deaton, among others, who is present here. And his conclusion at the end of it all was that, although the official figures might be overstating the extent of poverty reduction, he said, and I quote,"They are not seriously misleading."

In the early stages of this debate, the Bank itself seemed to be misreading the data in the '90s and was worried that rapid growth had not led to an adequate reduction in poverty. I should acknowledge that when I brought this to the attention of Nick Stern, this led to a more nuanced assessment subsequently, which turns out to have been validated

I think this also illustrates another important point, and that is, one tends to think that if one had a survey, one would get all the answers. But, actually, the quality of the information that one gets from surveys can be open to a great deal of questioning both ways. So I think the issue of data and how one concludes what's happening is a very difficult one. But I do think that with what we know about China, what we know about East Asia, and now if you add to that the latest conclusions on India, one can confidently say that there is virtually no serious danger that if countries could actually get into rapid growth, they can be reasonably certain that there'd be a huge amount of poverty reduction.

Now, that's obviously not to say that the growth process couldn't be made more poverty-reducing, and that's a different class of issues. We know the kinds of policies that lead to rapid growth, and to these we should add those that would make the growth more poverty-reducing. Clearly, what happens in agriculture is very important. Clearly, what happens in education, in the social sectors, and these need to be added on. But it's quite important that we reiterate not just the importance of growth, but the importance of the policies needed for growth.

And before I conclude, I just want to make two other very brief points.

There is a lot of concern about poverty,

but what policymakers frequently have to worry about is impoverishment. All the evidence suggests that growth leads to a reduction in poverty and an increase in the per capita income of the poor, which may be the bottom 10 percent,or 20 percent or 30 percent, depending on the country. That is not the same thing as saying that growth does not lead to an impoverishment of certain socioeconomic groups. We have known about this for some time. General equilibrium models of Korea, done by Irma Adelman and Sherman Robinson way back in the '70s, showed that, as you shock the model with all kinds of policy changes, the end result in terms of income distribution didn't change very much, but the relative position of socioeconomic groups could change quite a bit. So you can have people slipping into the bottom 20 percent because the particular economic activity they are involvede in is not doing well, but this movement is offset by other people going into higher deciles. In other words, you can have a situation where the per capita income of the bottom 20 percent is rising, but many of the people in the bottom 20 percent have actually been impoverished in absolute terms because they were earlier in a higher quintile or decile.

If one is really concerned not just about making sure that the incomes of the poor rise, but also making sure that there is no impoverishment of any group, that's a hugely more difficult objective. This raises the issue that any successful development strategy is going to involve structural change.

Certain occupations, certain types of activities are simply going to get phased out. Now, it is very difficult for politicians to actually acknowledge that. It's very difficult to go before a group of people and say that if development succeeds, your association of people doing X, Y, or Z would actually need to be disbanded in 20 years' time because this activity won't exist, and yet that is actually what is happening very often.

I think a lot of the unhappiness that one sees—and it's legitimate, and this is where the politics and political economy comes into it—a lot of the unhappiness with rapid growth is not based on any careful calculation of what has happened to income levels per decile. It's really based on the perception that different groups are benefiting to different degrees. And if some groups are going to find that the income level of that particular socioeconomic group is going down, those unable to shift out of that socioeconomic group to other groups are going to find it very difficult to approve of this process.

Politicians who are trying to formulate policy cannot ignore this by simply saying, well, but the percentage of people below the poverty line has gone down. So it really indicates that although one often talks only about poverty reduction, in practice politicians are concerned with a much wider set of objectives.

Finally, let me make just one point on poverty impact considerations that are relevant for short-term policy. I agree entirely with what Santiago has said on the need for social safety nets. One can imagine that by its very nature an adjustment program is going to have an element of contraction somewhere, whether fiscal or monetary or something. If the objective is to reduce absorption, somebody's absorption is going to be reduced, and, regrettably, it is almost impossible to reduce the absorption of the rich because they tend to have marginal saving rates of unity. So whatever you do to their income, their consumption level remains more or less the same, and what is worse, it remains visibly more or less the same, while other sectors suffer a contraction.

Understndably there is a concern to protect income levels of particular groups. But one of the issues that we have to recognize is that vulnerability may not be totally correlated with income poverty. It may well be the case in many developing countries that the groups that are poor, and to whom the longer-term poverty alleviation programs are directed, may be the groups that are not effectively linked with the market economy and the growth process. But vulnerability in the event of a short-term shock will probably be greatest for those that are actually linked to the growth process because they're the ones who lose jobs, they're the ones who suffer from a contraction of credit, a slowdown in expansion, et cetera.

It's not easy to react to this problem by saying that the short-term solution lies in intensifying the poverty alleviation programs that have been built for the long-term reduction of poverty. Those programs often focus on groups that are not linked effectively to the market economy; hence, the great importance of a social safety net, which is what Santiago said. This is not something that can be constructed overnight for programs whose time horizon would be 12 to 18 months.

It seems to me, therefore, that if we want to be able to protect the vulnerable groups during short term crises, which would include not just the poor but many of what Nancy described as the non-rich, non-poor category, then probably the only way it can be done is to identify certain programs which can be operated countercyclically. A social safety net program would be ideal for that purpose, but many countries don't actually have it, and putting one in place is a matter which will probably take a decade.

We also need to recognize our limitations in handling such problems. That doesn't mean that we shouldn't be doing something about it, but it is often too easily assumed that the solution lies in simply strengthening the existing anti-poverty programs but may not get to the groups that are most vulnerable in the short term.

Thank you, Anne.

[Applause.]

MS. KRUEGER: Thank you, Montek.

We now have about 20 minutes for questions from the floor, after which I'll allow each of the panelists about four minutes to—or two minutes each, about 10 minutes at the end, for wrap-up. T.N.? Identify yourself. Everybody knows who you are, but if you don't set a good example, we don't know what will happen next.

MR. SRINIVASAN: [Inaudible - off microphone] address what the international financial institutions [inaudible] IMF, World Bank, were to do in the context of poverty [inaudible] poverty alleviation, things that they are not doing or they have not done before. Instead, most of the panelists, until Montek Ahluwalia [inaudible], focused on essentially domestic policy changes, and it is unclear to me whether aggregate resources from the [inaudible] and multilateral lending institutions, the volume of transfer from those institutions is a constraint on bringing about the policy changes that need to be brought about. And on this question none of the panelists had anything to say [inaudible] hear more.

And, in particular, on Nancy Birdsall's point [inaudible] given whatever evidence there is of [inaudible] economic advice provided by the IFIs and their adoption and their results [inaudible] one doesn't have much confidence that [inaudible] mandate of the IFIs related to politics—

[Laughter.]

MR. SRINIVASAN: I don't think the IFIs, the World Bank [inaudible] dispense political advice, let alone [inaudible] advice. I don't see this as a way [inaudible] poverty alleviation.

The last point is what Montek said. Right from the word go, many of the developing countries, particularly in India, had a clear notion of what the role of growth was and what development is all about. And if you bring the bulk of the population into the income generation process and enable them to do what is best for themselves, remove impediments across the board [inaudible] able to achieve what they want, this includes trade policy, macro policy, and whatever. And there is a small proportion of the population which are not well connected with the income generation process. For them you [inaudible]. To think in terms of a permanent, a forever transfer program as a poverty alleviation strategy doesn't make any sense to me. And I wish that you would comment on this PRSP process [inaudible].

Now, if you put all the eggs in one basket, call it a PRSP paper, and when the objectives of developing countries are not only poverty alleviation but a much broader development, I don't see what this is going to achieve. And why the IFIs are moving in this direction rather than thinking about the broader issues of development is beyond me.

MS. KRUEGER: Thank you.

[Applause.]

MS. KRUEGER: Are there other questions, comments from the audience? Okay. Back here.

MR. : Mohammed (?) , Fiscal Affairs Department, IMF. If you look at the World Bank indicators like either the last four or five years or the latest one, you are struck by two things. One is the paucity of poverty data, and second is the timeliness. And my question is: Why do you think so much resources are allocated to studying poverty and so little to actually measuring it?

MS. KRUEGER: Okay. Others—yes? Please identify yourself.

MR. PRITCHETT: I'm Lant Pritchett from Harvard University and Nancy's center.

I've never been very clear why, when the IMF says it wants to get involved in poverty reduction, the IMF doesn't take the point of view of the IMF shareholders and define the poverty line relevant for IMF success in poverty reduction as being the poverty line used by its shareholders in setting their own poverty reduction initiatives. In other words, we're interested in reducing poverty. Nearly everyone in India is desperately poor relative to the standard at which we invoke poverty programs in our own country. Therefore, what IMF means by poverty reduction is poverty reduction in India relative to OECD standards.

What this would do is essentially, instead of getting the IMF worried about whether the IMF program benefits the 37th percentile versus the 42nd percentile of people in India, it would get IMF back on a lot, I think, what Montek and T.N. are talking about, a lot more deeply involved in the broad aggregate growth strategy. So the poverty reduction strategy would be a strategy to reduce poverty, broadly speaking, which is, yes, we can cut off the concern about the upper tail, but to get the IMF too deeply involved in setting an entire national strategy based on a definition of poverty that's narrow relative to the country at hand and gets them into distributional issues and away from growth issues, it's not clear to me that you wouldn't have a winning strategy of saying the IMF is a thousand percent concerned about poverty reduction, exactly to the same degree our shareholders are. So our poverty line is 5,000 per capita, and we're concerned about poverty reduction below that strategy, whether it occurs in the U.S., Europe, or India.

MS. KRUEGER: Over here.

MR. PERNIA: Ernesto Pernia (ph) from Asian Development Bank. Just a little comment on Nancy's point about the missing political alliance between the poor and the non-rich. I thought that is not the critical missing link because, in fact, there's already some alliance going on between the poor and the middle class or the poor and the not-so-poor in terms of the middle class buying, for example, from the services of the poor, and also in terms of the middle class being active in NGOs and civil society groups, you know, having alliances with the poor. I thought the more critical missing link was between the poor and the rich or the upper class, because I think this group still don't understand exactly why poverty reduction is a public good and it benefits everybody.

In Thailand, I understand there's already some effort in bringing about this alliance between the business sector, the businessmen, and the poor in rural areas in terms of bringing advice and assistance to rural poverty alleviation in Thailand.

I would like you to comment on that. Thank you.

MS. KRUEGER: Thank you.

Other comments, questions? Yes?

MR. MORRISSEY: I'm David Morrissey with the National-Louis University. I wonder if it's time to shift our thinking or to add to our thinking about income statements as the way we look at poverty, personal consumption expenditures or GDP per capita, to think about the balance sheet. We learned today from some of the studies that gold plays an important role in the individuals to help them and a safety net.

I wonder if there's a market failure in land ownership. I've mentioned it to a few people. Hernando de Soto has mentioned that people are living on land that's worth $10 trillion but they don't have title to it. And it's not because they're poor. It's because the process of gaining title to land ownership involves a number of years and going through 150 different steps which a poor working person can't do.

Now, that's really micro for IMF concern, but maybe there's a role for IMF in facilitating the market for land and helping people own land and grow.

MS. KRUEGER: Thank you.

In the back?

MS. : Maya (?) from Kenya, from a group called the Mazin-(?) Institute. I was wondering if Santiago would explain more about how the costs of corruption and administrative costs of your social safety net are low, because being from Kenya it's very—corruption is a major issue there.

And also talking about how do you manage to distribute the money to people in informal areas, you know, because that's the majority of the people who we work with.

Thank you.

MS. KRUEGER: Thank you.

Other—yes, over here?

MR. : Sheikh Kar-(?) , Africa Department. I had a question for Mr. Santiago Levy. I certainly agreed with his analysis that monetary transfers are probably more useful than our system of taxes and subsidies for transparency and efficiency. But you made a point about tying the transfers in some circumstances to various types of behavior such as accumulation of human capital.

I was wondering why that's necessary because if you have funds, then you will presumably maximize your welfare given your funds, and if education is a better use of your funds than other things, you will automatically spend on education. If the rate of return to education is higher than your discount rate, presumably you will spend the funds on that. And if it isn't, then it's not clear to me that you aren't reducing welfare by making that conditionality.

MS. KRUEGER: Thank you.

Other comments, questions?

[No response.]

MS. KRUEGER: Okay. At this point let's let each of the panelists respond in the same order in which they spoke initially, starting with Nick Stern.

MR. STERN: Thank you very much, and thank you for all the questions and statements.

First, let me say that the emphasis on the medium term and growth is absolutely fundamental. That was the way I started. That was the way I meant to start. But also as important is involvement of poor people in that process, and the challenge is to proceed on both at the same time and not to see one as a conspiracy against the other. Actually, I think that empowering and investing in poor people is a key part of the growth process, so we have to keep those two things going constantly.

A number of the panelists used the words "poverty alleviation." I've banned that from my staff. We're in the business of poverty reduction and growth, and poverty alleviation for me smacks too much as sort of a handout approach to development, when what we're trying to do is to drive forward and lift people out of poverty in a strong and sustained way.

Now, a lot of the discussions has been very macro and aggregate, and you're forced into that in this kind of surrounding. But I just want to close with what I have to say by talking for one minute about one village in UP in India where I've lived and worked on and off since 1974, and the most recent time I went back was with my wife one year ago. And we have data, 100 percent sample data on that village for every decade since Indian independence, the last three of those collected by ourselves.

If you look at what has driven the movement out of poverty of some people in that village over these last nearly 30 years since I've been going there and working with my colleagues, it is exactly the kind of thing I pointed to. It's the outside jobs that have appeared because the investment climate has improved in terms of less—there's still a great deal of bureaucratic harassment, but a bit less bureaucratic harassment. There's still rotten infrastructure, but it's been a little bit less rotten. The jobs are there in the sugar mills, in factories, in construction, and the domestic service. That school had a teacher that turned out 50 percent of the time. It turned out to be about the average for UP.

When parents and citizens get involved in the management of the schools, then teachers turn out more often, and that starts to appear 10, 15 years down the line in terms of poverty reduction.

That's the reality on the ground of the investment climate for growth and empowering poor people to participate in that process.

The classification into social classes and all this often leaves me cold. What we're interested in is, of course, poor people, but we're interested in the whole population across the spectrum. And it's by focusing on living standards, measurement studies, individual household surveys, that you get the real distributional impact of all this.

Thanks very much.

MS. KRUEGER: Thank you, Nick.

Next, Santiago Levy.

MR. LEVY: Thank you. I'll make three comments.

First, on the question from Kenya, I believe, on administrative costs and corruption, corruption is relative. It is going to be very difficult in any program to have zero corruption. It's just a question whether the incentives for more corruption and the price differences or the opportunities for arbitrage are bigger in one program vis-a-vis another one.

If you have the same commodity being sold at two different prices, then the incentives are much higher for that to take place, particularly if it can be stored or things like that.

So it's not that you transit from one in which there's a lot of corruption to one where there's zero, but you transit from one where there's more to one in which there's less.

You also try to use a lot of information for households and make them participatory. That comes to the second question up here. So that they realize that they have a right to these transfers in return for some behavior, and through community participation, corruption so far—the problem has been known for five years now, corruption so far has not been a big problem.

Administrative costs are a lot smaller because you don't set up any public enterprise. You don't produce anything; you don't distribute anything. There's no bureaucracy that does this.

We did a program on top of the health services that were already there and on top of the educational services that were already there, so there is no additional cost because the services were already being provided.

The efficiency we have is for every peso spent, about 93 cents are transfers directly going to the pocket of the persons that we want to help, and seven cents are more or less the administrative costs.

The second question over here is a very interesting question: Why is it that you want to tie transfers to some kind of behavior? And it also has to do with Professor Srinivasan's question, and I'll end with that.

I think you want to tie them for a whole bunch of reasons. First, there is some sense of co-responsibility. This is not—you want to get away from the sense that it's a pure handout, but this is some payment in return for something that you're doing constructive for your family and for yourself. And the people who participate, particularly the women, the mothers who participate in the program, feel that they are now responsible for bringing their kids to a health clinic every two months or every three months. They are responsible for making sure that the kids go to school. And it gives them a sense of empowerment and it gives them a sense of participating. They're doing something in return for the help that they're getting. It's not simply a free handout.

You can go deeper into sort of informational questions and inter-household inequality issues and inter-temporal issues. Parents may not be willing to invest fully in kids' education if the rate of return is too high—if the discount rate is too high, so you want to ensure, in fact, that parents do put kids into school, and particularly—and this was in some of the papers in here—they don't pull them out of school in the time of a crisis. So you might want to change the size of the transfers to make sure that they're tracking the opportunity costs of the time of children to make sure they stay in school, if what you want is an investment in the future. Also, because of inter-household inequality, you want to make sure that some health care is delivered particularly to pregnant mothers and to lactating mothers and to children between zero and four years of age that are particularly vulnerable to undernutrition risk and things like that. So there's a whole bunch of reasons. People feel empowered about this.

Now, the tying also has to do with—a very good question asked by Professor Srinivasan, and I'll end here. What you would like is for these transfers to be transitory and for these transfers to open the way for people to increase their human capital and eventually not to need them. And that's why it's important to keep them low so that when people have a better set of opportunities, then they can actually not require them and open themselves into getting something else.

We would do a lot of damage if you set up programs with high transfers that generate a poverty trap, some kind of welfare trap, and keep people in here. So the incentives have to be looked at not only within the program, but the incentives have to be looked at across the whole distribution. What do I mean by this? That the government has many other programs that transfer income through social security, through taxation, and you want to make sure that the whole distribution has the right incentives so people eventually want to move from this one to that one. And in principle, you would like the program to be without targeting.

It's very difficult to run it without targeting initially because you don't cover everybody and you don't have enough administrative capacity. We're now reaching the point where we're probably going to be covering in a year from now the whole poor population, and we might even want to change to a self-selection so that when your own characteristics allow you to participate in something else, you don't need it and you have a way out.

And the last point is you certainly don't want this to be your poverty fighting strategy. You want this to be a component of a much broader strategy, of course, but you need it—and I brought it into this discussion because in macroeconomic crisis, which the premise here is that they're there and they're unavoidable, so to speak, they're very useful and they really help to protect.

MS. KRUEGER: Thank you.

Nancy?

MS. BIRDSALL: Well, T.N. is worried that I want to get the IMF and the World Bank into politics, and I guess my point is that the World Bank and the IMF are already in politics, and I want to get them out of politics. And let me explain what I mean.

I think if you rename the ESAF as the Poverty Reduction and Growth Facility, you're already conveying what Montek referred to. It's perfectly legitimate, but that the public good at the international level from the point of view of the IMF is poverty reduction, at least equally with or before the growth and transformation that Montek was talking about. So there is a kind of quiet politics that is going on there that clearly is appealing to the shareholders of the richer countries.

You know, I think it's hard to contest it because no one—and certainly I don't want to be against poverty reduction, but I think we ought to recognize that's going on.

The same is true with the Poverty Reduction Strategy Papers that the World Bank and the IMF are working with countries to develop. The fact is that we have World Bank and IMF staff reporting to the Boards on whether those PRSPs were developed in a participatory way in which civil society was engaged. And, you know, I think it might be a step in the right direction, and there's a lot of good things that seem to be coming out of it.

At the same time, many of my colleagues in Latin America are complaining that they're concerned that this could undermine local government efforts, that it is too much of a substitute for what is already supposed to be a representative political or democratic system.

So I think we have to be alert to the problem of intrusiveness, and I think the issue that I wanted to try and bring out is about that, in that, you know, in India the politicians are worried about impoverishment—I think it was an excellent use of the word by Montek—as much as they're worried about alleviating poverty, not to accuse anybody of—you know, Nick is talking about reducing poverty, you've got alleviating, reducing. But the politicians are worrying about impoverishment, and that's exactly why it reflects the fact that India is a democratic society in which many people are in the market economy, and if there is a shock, they face risks.

And I suppose there's a sense in which if the relevant advice of the World Bank and the IMF is focused heavily on a static definition of the poor, it doesn't sing politically. And it's not quite in the direction of forging this implicit alliance that I referred to, an implicit political alliance—and here I want to answer Ernesto's point—between the poor and this vast middle. And I say it's an implicit alliance in the following sense:

In Western democracies, the poor, luckily, are able to free ride on the political insistence of the near-poor, the might-be-poor, the vulnerable, for social insurance and for social—these safety rope programs. I like that expression a lot. And in the international community, we should not be interfering in that natural political outcome in democratic societies. We should be cautious about arriving and saying target, target, target. And I saw this at the time of the Asian financial crisis, when the first reaction—well meaning and enlightened—from the international community was to go to Indonesia and Malaysia and Thailand and say you must target the poor. And within a year, the information came up that those who were worst hit by the crisis in Asia were in the second and third quintiles of the income distribution.

This is not to say the poor didn't suffer and there shouldn't be a way to reach them. It's to say that in the long run, unless politically programs are designed that appeal to the second and third quintiles, who are also going to be poor or were poor, I think it's very hard to sustain the momentum to keep the programs. And it's certainly very hard to reduce the risk that the programs will be corrupted when people like Santiago Levy are no longer sitting on top of them.

So I hope I didn't create misunderstanding. I do think it's important for the IFIs not to enter into politics, but to avoid creating political constraints and to avoid being politically naive, particularly given the evidence that we need to explore more, that in the end it's in democracies that we're going to be able to see the poor protected and the near-poor protected. And it's because in those democracies the vast majority of people who are in the middle want those programs. We will never succeed in making them work from outside. They have to be politically motivated inside.

MS. KRUEGER: Thank you, Nancy.

Montek?

MR. AHLUWALIA: Thanks, Anne. Three very quick responses.

One to T.N. on the issue of what use is the additional transfer that IFIs could do. Clearly, it's useful only if the policy environment is right, which will achieve the broader objectives of a broad-based development-growth process. If those policy objectives are right, then presumably the resources they provide add just that much additional input into a, on the whole, good process, which should lead to better results.

Two other points. Someone made the point that many of the lower-income groups suffer from distorted markets, land markets, I think. This is absolutely a very important point. My own perception is that we have tended to underestimate the extent to which markets relevant for the poor are hopelessly distorted, very often because of lack of thinking, and quite often because of vested interests of a bureaucracy that is simply not very sensitive to what is needed there. And I have no doubt whatsoever—in fact, I think this is one of the areas where civil society organizations have made a major contribution in India. It varies—the nature of the problem varies from place to place, but this zeroing in on the burdens, the regulatory burdens, very often unthinking, which are placed on the poor, and the lack of participation in ground-level activity, including schools, which Nick mentioned, to my mind—both of which the CSOs have been leaders, really. And I think one of the good things in the Indian experience, anyway, is that their special role in this is now recognized and they're actively seeking—government actively seeks participation and inputs from them. I think they have a major—that is something we need to think about and do a lot more on.

Of course, it all depends on who and where and it's very micro, but it's very important.

Thank you.

MS. KRUEGER: Thank you, Montek.

At this stage we must reluctantly conclude this conference, which has, I think, brought forth a lot of interesting ideas. As usual in these things, there are more questions than there are answers, and so there's plenty of grist for both researchers and policymakers to move forward. We look forward to hearing more about the Mexican program, other efforts in the same direction, experiments will give us much more, and on the whole we look forward to doing even better than we have done already in terms of our growth and poverty alleviation objectives.

Thank you very much, panel, and thank you very much for participating.

[Applause.]

[Whereupon, the proceedings were concluded.]