Search GDAE

Subscribe for
E-mail Updates

Recent Publications

GDAE in the News

Upcoming Events

Publicaciones en Espanol

Jobs and Resources

Leontief Prize

"Ruling Out National Development? States, Markets and Globalization"

Remarks by Dani Rodrik
on occasion of his receipt of the Leontief Prize
at Tufts University, on November 21, 2002

Thank you very much, Dr. Goodwin and Professor Moomaw. It’s really a very special honor to be given this prize, especially because Professor Wassily Leontief was such a special scholar. I wish I had had the opportunity to know him. I think I saw him in person only once; he was on his way out of Harvard when I had just come to Harvard as a freshman. And I’m intellectually somewhat related to him because one of my advisors at Princeton, Peter Kenen, was actually one of his advisees, so in some sense I am intellectually a grandson of Leontief.

Professor Leontief was a rarity in many ways in the economics profession because even though economists will always remember him for his technological contributions, in particular input-output analysis and something that’s very close and dear to the heart of every trade economist, the Leontief Paradox. What really sets him apart as a very special man and special scholar is that he always thought that economic analysis has to be embedded in the broader social context, that if one forgets that broader social context economic analysis is likely in fact to go very wrong. The way that he tried to combine sensible theorizing with hard nosed empirical work I think is to this day really quite exemplary in the economics profession.

One of the nice things that this award made me do is actually go back and either read or re-read some of Wassily Leontief’s work, and what a joy it was In fact as I was re-reading this work, or reading a little bit for the first time, I was struck by how we would have avoided many of the wrong turns that we took in the last couple of decades in issues of international economic if we had actually paid closer attention to some of the methodological precepts that he articulated in his work. Indeed I want to organize my own remarks here under two headings that are taken directly from Leontief ‘s own writings and I’ll call the first the problem of non-observed facts. That comes from his American Economics Association Presidential address in 1970, which was called ‘Theoretical Assumptions and Non-observed Facts’. And the second heading is the problem of implicit theorizing, and here again implicit theorizing is taken from an article he wrote called ‘Implicit Theorizing, the Methodological Criticism of the Neo-Cambridge School’, which was published in the Quarterly Journal of Economics in 1937. And these two ideas, one that relates to the realm of empirical analysis and the other to conceptual and theoretical analysis, I think nicely tie together some of the themes that I want to talk about today.

So let me talk about non-observed facts first. And I think a good place to begin when one is talking about non-observed facts is with the Washington Consensus. Just to make it concrete, the set of principles that John Williamson first articulated in 1989 are the original items in the Washington Consensus. But of course much has happened since then, and in fact one thing that has happened is that for a number of reasons, not the least of which is that things have not turned out as predicted, this original consensus has in some sense been augmented by the items on the second column, the right column. And although one can list these new requirements of the Washington Consensus in a number of different ways, to preserve symmetry I have classified them under ten headings. And what you see there is essentially that we now have a fairly long list of demanding institutional requirements that span all the way from labor markets to financial governance.

In many ways, the last two items here, the social safety nets and targeted poverty reduction, which are part of today’s common conventional wisdom of things that have got to be done, have a relationship to the secure property rights in John Williamson’s original Washington Consensus, in the sense that when John first thought about listing the kinds of things that considered as this consensus, what he came up with was only nine items, and he said a list of nine items doesn’t seem very good, so he had to find one more, and the last thing was secure property rights, and that sort of nicely rounded off the list to ten. But of course everything in the second column now is in some sense trying to make that happen, because we’ve understood that in fact, the institutional underpinnings of property rights, contract enforcement, and desirable modes of behavior between the public and private sector, are in fact at the heart of economic development.

Now where do the non-observed facts come into this? The reality is that actually no country in the world has actually developed following these principles. In fact it’s more accurate to say that to the extent that they capture any reality, they capture the reality of countries that are already developed, rather then being a set of useful recipes or useful guides for countries that want to embark on a period of high growth. You don’t have to go very far to see this point; simply ask yourself what are the countries that have done reasonably well in the past two decades. Two very large countries stand out, China and India. The fact is that India’s growth rate has doubled since the early 1980s, and China has been growing at rates that, even if one corrects for some exaggeration, are really quite remarkable by any standard. Those facts are the primary reason why global poverty has been declining, because these two very large countries have been doing very well; but they’ve been doing very well in the sense of experiencing very high overall economic growth rates by following policies that really sit very awkwardly with the ideas of the Washington Consensus, whether it’s the original one or the augmented one. And the question is what does one make of that?

I think its clear that measured by the Washington Consensus criteria, neither country really fits very well. For example, just to take two areas of reform, trade reforms and privatization. Clearly countries like China and India have not undertaken nearly as much trade liberalization as, for example, a country like Haiti has done, which in a very short period of time actually has eliminated virtually all quantitative restrictions and now has very low tariff rates. Neither have they done as much privatization as, for example, a country like Argentina or many others in Latin America have done.

So what to you make of that? There are entirely reasonable positions one can take on the experience of this high performance by these countries in the last couple of decades. One is essentially to try to eliminate several contentions between this recipe and the experience of China and India by saying: well at the very least we can say that China and India have done well by moving somewhat closer to this recipe than they were at the outset, and therefore the liberalization and opening at the margin, while falling very far short of what much less successful countries have done, at least because they’ve moved in that direction, that’s what explains really their success. That is more or less the prevailing view in some sense in organizations like the World Bank and the IMF. The alternative view, and the one I feel most empathetic to, is that essentially this is not a good way of looking or understanding their behavior. That view basically is that all developing strategies are based on a combination of orthodoxy and heterodoxy. And these countries show that, provided you pay importance to private incentives and that you do not turn your back on markets, a certain amount of heterodoxy is often required to get you into high growth rates.

Whichever position you take, whether it’s the fact they are moving in the Washington Consensus direction view or the alternative view I just articulated, I think there are some defensible arguments you can make. But what is I think clearly indefensible is to observe China and India’s success, and on that basis recommend to other countries the kind of trade policies that Haiti has had, and the kind of privatization policies that Argentina has had, but that’s essentially what is happening today.

So there is a very big difference between saying that, even if you take the first position that these countries are moving in the Washington Consensus direction, that other countries ought to be doing it the same way that Haiti and Argentina did. Maybe what follows is that countries ought to be doing this thing in the same partial two-track gradualist way [that China and India did]. But those policies tend to be the so-called best practices. So what I would say here is that policy with regard to development has been driven by what Leontief would call un-observed facts, the un-observed facts being that countries with policies such as Haiti’s or Argentina’s ought to have done very well, but in fact they have not.

Now what is the actual reality of the development experience? I think it would be useful to explore, although I’m not going to obviously have time to go into this in any depth. I think it’s useful to have a clear idea of what actually the empirical record is. And I’m just going to very quickly mention the main headings here without going into detail, but I want to emphasize four key things here. Number one is something that I think is very important in the context of these Washington Consensus discussions, that when you look at transitions to high growth, what you actually find is nothing like the extensive set of institutional reforms that are called for under either the Washington Consensus view or the augmented Washington Consensus view. Whatever you look at, whether it’s the East Asian tigers of the early 60’s, or China and India in late 70’s and early 1980’s, or Mauritius in the early 1970’s, what you find actually is that it’s a relatively narrow range of reforms that kick start the process of growth. This is very, very good news. Its very good news because it says that actually you don’t have to do all those twenty extremely demanding set of institutional reforms to get some significant amount of economic growth going, starting from a relatively low level. That I think is very, very good news.

The second empirical fact that I will mention, and I said this in passing a while ago, is that the policy changes that initiate these growth transitions typically combine elements of orthodoxy with unconventional institutional innovations or, if you will, local heresies, local departures from what the conventional wisdom is. Of course Alice Amsden has done very good work in the area of documenting and analyzing this in the case of South Korea’s transformation in the 60’s and 70s: how they combined a certain amount of liberalization and output orientation with an extensive set of industrial policy conventions. The same is true of a country like Mauritius, which combined an export-processing zone with a very highly protective rate of the domestic sector. It’s certainly true of China, which turned to the market in some ways, but fell far short [of complete marketization] and came up with its own institutional innovations in the form of township and village enterprises, household responsibility system, two-track pricing system, and so forth. And even Chile, which is sort of everybody’s favorite example of the most Washington Consensus type country, has departed from the consensus in a number important ways. The best known are Chile’s capital controls, but in areas other than that such as public ownership and also a certain amount of industrial policies in export sectors, I think there have been important departures.

Now the bad news, and the third point, is that actually these institutional innovations do not travel well. Many of the kinds of innovations and local heresies that were tried in these countries often did not work very well at all when they were emulated in other settings. So the household responsibility system worked very well in China. But when Gorbachev tried to implement something similar to that in the mid 1980s in the former Soviet Union, it did not work at all. Import substitution worked in Brazil but not in Argentina, and so forth.

And finally the fourth point, I think the critical point, is that my first three points really go to the issue of how do you kick-start growth, but it’s important also there to bear in mind that that’s not the end of the road. I think we have many, many examples of how periods of high growth are eventually stifled by the absence of high quality institutions, so that these economies can suddenly find themselves in crises and stagnation as a result of their inability to handle adverse shock. So a strategy of institution building I think is extremely important over the medium to longer term and that’s where to some extent I would agree with the augmented Washington Consensus, with some important qualitative differences. It’s important to see this institution building as a different strategy, as something that’s a medium to long term strategy, and not one that is designed really to get the economy to kick-start. So I think there are important differences between a short-term strategy and a medium to long term strategy.

Now, let me briefly say a few more words about the second heading, which is the heading of implicit theorizing. And the question that arises, given this kind of empirical evidence, this empirical record which is laid out for everybody to see, is why has policy advice not been more imaginative, why do economists so often retreat to simplistic text book prescriptions, rather than apply the real lessons of cases such as China, India or before them South Korea, Taiwan, Singapore, Mauritius, and so on. The reason I think seems to be that our simplistic view of the evidence is matched by a simplistic theory of the state and of state capacity. I like to say that scratch any policy economist and you will find underneath an amateur political scientist, and a very bad one at that. And what happens here is that economists often make prescriptions that do not survive even the first level of economic analysis; eventually if you ask them why are they recommending such things it turns out there is a very crude political theory about what the state can do and cannot do, that underlies [their prescriptions] but has never really been fully articulated, let alone actually tested. So to give you an example, why is it that policy economists always recommend uniform incentives, uniform taxes, uniform tax incentives, uniform subsidies, when all the economic theory we have calls for differentiation and selectivity. The answer in practice, I think, is that policy economists do not trust governments to use selectivity and discretion to good purpose. So the implicit theory of the state is that the most that the state is capable of doing is to restrain from doing anything bad. So forget about playing a proactive role -- protect property rights, enforce contracts and that’s it.

Now conceptually what’s wrong with this view is that it’s a very incomplete theory, and when you try to make sense of it in a more fully articulated way, it falls far short. So consider this, if on the one hand the state is in fact beholden, as many economists fear, to special interests and therefore needs to be restrained, then there is really no reason to presume that the limited intervention recommendation can ever be made to stick, and if it appears as if it is sticking it will often be the case that huge distortions and rent transfers are hidden behind a façade of laissez faire. Probably the best example in recent times of this is the experience with privatization in Russia in the 1990s. So if your view is that the state has this structural problem, then simply saying “don’t do anything” has no relevance whatsoever, because the state will do what it wants to do, which is steal for the benefit of individual privileged groups. On the other hand, if the state is in fact capable of acting autonomously and in pursuit of social welfare, we can then certainly think of ways that this autonomy can be used creatively and to good ends. And I can go on talk about the empirical evidence on to what extent discretion and selectivity has been associated with bad and good outcomes. But clearly the empirical evidence does not suggest that there is this very clear alignment of the effectiveness of policies on the basis simply of whether in fact they allow discretion or not, whether they are uniform or not. In fact the evidence, to the extent that we have it, sometimes goes very much in the opposite direction.

So the lesson here really is to think in terms of enhancing the institutional capacity of the government, not to presume that the absence of institutional capacity has determinative implications for policy. And I think that again gives us a different way of thinking about policy. In particular, it has two sets of implications for areas where policy autonomy is very valuable and is steadily being eroded. One is the area of the World Trade Organization, where WTO agreements are increasingly constraining the ability of states to undertake policy and institutional innovations, which I think are actually quite important. It’s not a coincidence that the only countries that have done actually well under the WTO are countries that either were not members of WTO such as China, or countries that have effectively played by the GATT rules rather than the WTO rules, such as India.

The second area where I think we have to reevaluate the importance of policy autonomy is with regard to policy conditionality as a whole. I’d be much more in favor of an ex-ante type of conditionality, or a meta-conditionality where you make some broad distinctions concerning countries that are politically democratic, that have certain basic institutions, and then make the distinction on the basis of that, and not get into the very nitty-gritty details of actual policy.

So let me simply end by saying that I think that had we listened to Wassily Leontief’s advice on avoiding the problems of implicit theorizing, and not assuming that non-observed facts are actually observed, by now the quality of our thinking on development policy and international economic policy would have been much better than it is currently.

So once again thank you very much, and thanks very much for this honor, I will cherish it.

Top of Page