World Development Report 1998--99

Knowledge for Development

by the World Bank

Geeks vs. Wonk on H Street

The World Bank, it should be said first of all, is not an ordinary commercial bank. (A joke circulates by e-mail, sometimes in the first person and sometimes in the third, about a man who, shortly after moving to Washington, chanced upon the Bank's headquarters, at 1818 H Street, stepped in to see about opening an account, and was hurt and perplexed when the guards laughed him out of the lobby.) Like its sister institution, the International Monetary Fund (with which it is often confused), the Bank was created by the Bretton Woods agreements concluded by the Allies in 1944, as part of a plan to restore international commerce and the world economy, and prevent a disastrous collapse of trade, as in the Great Depression, from recurring. The Bank is actually a cluster of organizations, of which the most salient is the International Bank for Reconstruction and Development, the IBRD. As its name indicates, its original purpose was the reconstruction of the European and Asian economies after the war, at least those which fell outside the Soviet orbit. (If memory serves, the Soviet satrapies were invited to participate in the IBRD, as in the Marshall Plan, but Moscow forbade both.) The Bank's capital is subscribed by donor governments, and lent, for long terms at low interest rates, to other governments, to be spent on specific development projects which are (ideally) carefully scrutinized by the Bank beforehand for their utility and feasability, and the borrower's ability to repay. A government which defaults on any loan from the international community is in trouble; one which defaults on a loan from the Bank is simply in a hopeless position. Accordingly, the Bank's bonds are very highly rated...

The Bank's original clientele --- the western European countries and Japan --- no longer need its services; indeed, they now lend it money, and so have a say in its policies roughly proportional to their contributions. Now it lends to what are optimistically known as ``developing'' countries, i.e. the Third World, and, in the last few years, the rubble of the Second. It is one of the tools by which rich and powerful countries manipulate poor ones, and an almost Byzantine bureaucracy. It is also genuinely concerned with helping poor countries become richer; at least as good at this as any other concern; and hands down the one with the most experience at it.

Since 1978, the Bank has published an annual set of ``World Development Indicators,'' which are one of the best sets of statistics on the economic and social conditions of the world's countries. These are poor gross crude things, of course, but they make do for lack of any better. The indicators are preceeded by a report on some important topic in development, and what can be and has been done about it: ``Investing in Health,'' for instance, or ``Workers in an Integrating World.'' This year the subject is ``Knowledge for Development.'' The Bank took firm stands in favor of Health, and of Workers, and now has officially come out in support of Knowledge and Information.

There are two intellectually respectable notions of information: the geeks' and the wonks'. Geek-information developed from control and communication engineering in the Second World War. It's a positive concept, a kind of quantification of variability. Just as distance is what you measure in meters (or picoparsecs, as you will), geek-information is what you measure in bits. Geeks are very interested in taking bits, storing them, shuffling them around, and moving them from place to place. Lately they've become extremely good at these things, and impressed with their own skill at them. More generally, when they think about knowledge, they think about what they themselves know --- technical knowledge, know-how, how to make things work.

Wonk-information, by contrast, is essentially a negative concept: lacking such information --- about future events, about other people, about opportunities, about goods and conditions --- people fail to act like the ideal Homo economicus. What such wonks study is what happens to people who rashly try to have economic lives with less than perfect information. A geek looks at a CD in a music store and sees a large chunk of stored information, which can be recalled, transmitted and manipulated in various ways. A wonk looks at the same CD and wonders how much less people are willing to pay for it without knowing whether or not they'll like it, or even whether or not it will play. [Ackerloff ref.?]

A number of attempts have been made to link geek-information to wonk-information theoretically --- usually trying to derive wonk-information from geek-information. None of them have come to anything. This doesn't mean that the two don't have anything to say to each other on a practical level. Informational problems, in the wonk's sense, can be mitigated by geek-information technologies, and there are wonk-information problems to the use of those technologies. (The CD-in-the-music-store problem is pretty easily solved by letting people listen to it in the store. The problem of buying large and customized pieces of software for, say, inventory control are more tricky.) The guiding idea behind Knowledge for Development is that one of the places the two notions meet is the problem of global poverty.

The poor are, of course, different from you and me: they have less money. They are also less informed, in both senses. Obviously, if the poor got money they would no longer be poor. That they would not be poor if they were well-informed is not obvious, but worth looking into. For one thing, informing the poor would much less costly than directly giving them money, even if there were enough to go around. Moreover, the recent successes of the geeks have made information, and the gadgets that handle it, the notions of the hour, so the combination of low cost and conceptual sexiness means there's a chance of actually making viable propositions of information-based poverty reduction schemes.

It is here that the difference between wonks and geeks rears its head. Geeks see information as a positive good thing, and see a technical problem before them: how to make the Net catch, not just Singapore and San Jose, but every favela and rice-paddy in the world. The benefit from this will be spreading around know-how, and so giving poor countries and poor people the capacity to do more and become better off. The wonks see information as something conspicuous by its absence, something the lack of which must be remedied, lest markets fail. Their problem is one of social engineering (though they're unlikely to use the phrase), of designing institutions and social mechanisms to either supply information, or evade the need for it.

But which sort of information? This actually makes a big difference when it gets down to the brass tacks of policy decisions --- do we worry about telecoms and putting computers in villages, or about the way lenders decide who's credit-worthy? The Report is, pretty obviously, the work of two different groups who did not see eye to eye on this, and is split into two parts: the first is by the geeks, and the second by the wonks. (There is also a third, on what international institutions ``can'' and developing countries ``should'' do. We'll get back to the third part.)

So: the geeks, and, as they put it, ``acquiring, absborbing and communicating knowledge.'' Knowledge, here, is used in the sense in which one knows calculus, or how to unblock a drain, or apply fertilizer properly, not in the sense in which one might know the works of Shakespeare or know Islamic jurisprudence. (The report goes so far --- p. 43 --- as to note a negative correlation between growth rates and the proportion of university students who study pre-law. This is a joke in America, but potentially a serious problem in countries where law is the dominant field of study.) Probably no one, in these last days of the twentieth century, would deny that a country which wants to be wealthy and powerful, or just not helpless and destitute, needs such knowledge; it must, as the report's authors put it, close the ``knowledge gap'' between itself and the rich, strong and learned countries. Nonetheless, ch. 1, ``The Power and Reach of Knowledge,'' sets out to convince of this; also that knowledgable societies are better ones to live in general. Again, this hardly needs saying; but still, some of the evidence --- like exponential declines in infant mortality rates, after controlling for real income, over the course of the century --- is very compelling. Probably the item which will do the most to catch eyes is the graph of real per capita income in Ghana and South Korea, from the mid-1950s to the present. Starting at very nearly the same level, Ghana's income has gone nowhere, whereas South Korea's has increased by a factor of seven, less than half of which can be explained by greater investment. [Presumably ``explained'' means ``explained by linear regression analysis''?] Some of the examples chosen are curious; one (box 1.3, p. 23) seems to show that whatever effects education may have on GNP growth are swamped by those of openness to trade and a large number of telephone lines per person. Still, the heart, or the homologous organ in bankers, is in the right place: abundant technical knowledge is a Good Thing, necessary to building a semi-tolerable society. (That it is not sufficient is shown by Russia, and the Soviet Disunion generally.) How then to get it?

Basically there are two ways of learning something: you can figure it out for yourself, or you can copy someone who already knows. Discovery does wonders for self-esteem, and may even have real benefits besides, but it is, in essence, a process of making and eliminating enough mistakes to hack out something which works: hence chancy, hence slow, hence expensive. Copying, on the other hand, is quick, reliable, and offers one of the few areas in which late-comers to the industrialization game are not decidedly handicapped, since they can have others make their mistakes for them. Accordingly, the report stresses local R&D much less than it does copying. (We shall return to local R&D in due course.)

The first copying mechanism the Report mentions, and, according to it, one of the most important, is openness to global trade. At this point (if not before) skeptics and cynics will suspect that the talk of knowledge and information is so much patter, and wait for all the other Usual Recommendations --- cater to global markets, let in multinationals and other investors, privatize, shrink government, and let the poor shift for themselves --- to be rounded up. They are, of course (except for the last, which isn't a Bankism at all, being left to the IMF and the faculty of the University of Chicago Department of Economics), but, very remarkably, they are in fact justified in the context of improving the knowledge-base of poor and backward countries. Take foreign trade and investment, for instance. Assuming that one is trading in services or manufactured good more sophisticated than carving knives (and an increasingly large chunk of trade is in those things, even between the First World and the Third), then there are certain minimums of quality, reliability and efficency which must be met to be competitive, and firms which do not meet them will fail. (This does not guarantee that a country will have any exporters which can meet those standards, of course; it just gives them an added incentive to do so.) Again, much of the world's technical knowledge is the property of large companies from the developed world. (This is even more true if one consider tacit, institutional, and other unformalized sorts of know-how.) Leaving aside spying (which has high fixed costs and very uncertain returns), one can get at this knowledge either by having such companies set up for business in your country, or by getting them to sell you the knowledge. The first is foreign investment; the second is licensing, under some regime of intellectual property rights or other. The standard objection to foreign investment, as a way of gaining technical knowledge, is that it doesn't work: the multi-national sets up a factory for cheap labor, imports all the necessary machinery, imports the engineers and managers (and, sometimes, the all the comforts of home for the last-named --- spouses and spawn, refrigerators, Real Food, etc.), and leaves the host country no better off --- except for the drone-work --- than it was before. But this can be avoided, and has been in some countries, e.g. Malaysia.

Intellectual property rights are strange beasts. In an ideal market economy, they would not exist, because information has two properties which make it especially ill-suited to conventional trading. First, there is a fixed cost to producing it, which can be quite high, but once it's there, the cost of producing an extra copy --- the marginal cost --- is very low (though not zero), and in an ideal market marginal price equals marginal cost. Second, someone who sells a piece of information still (usually) has it; it doesn't exclusively belong to any one person; in fact, everyone who owns the information can sell it themselves. The net effect of this is that markets, left to themselves, will not let those who come up with information recoup their costs by providing it to others; consequently information will be under-supplied, and everyone will be worse off. There are two ways of dealing with such public-good type problems. One is to have the good supplied by a non-market agency --- charities (usually conscience money) or governments (taxes). The other is to legally change the properties of information. The owner of copyright in a given work is (as the name implies) the only one with the legal right to make copies of it, a right which can be, in the usual way, leased, sold or given away. Other forms of intellectual property are so many other state-imposed monopolies. The problem with such monopolies is that the monopolists want to exploit them. Generally they are limited by requiring the monopolist to disclose the information (as with patents, or filing copies of works with the national library), and --- this is the clever touch --- making them temporary. Naturally, since these monopolies are legal creations, the monopolists will attempt to manipulate the laws in their favor: lengthen the term of protection (now well over a century for US copyrights held by corporations), shift from patents (which require disclosure, and proof of advance over prior art) to copyrights (which do not), make reverse engineering illegal, and so on. There is, as the Report notes, no empirical evidence that such manuevers actually increase the pace of technical innovation or aid the adoption of innovations; also no evidence that they do not. (Actually, the only evidence that the companies make any money of this is what they're willing to spend on lobbyists, bribes, etc., to get the law changed...)

The relevance of all this is to subject at hand is that intellectual property rights make technical knowledge expensive, except for basic research (which is generally and properly conducted by public bodies and in the public domain), and such pieces of knowledge as are thoroughly obsolete. Countries could try to get around this by not having IPRs, or enforcing them laxly, or not recognizing those filed in other countries and making the re-filing process long, expensive and difficult for foreigners, but all this has the disadvantage of making holders of IPRs less willing to do business in your country, and reviving the public-good problem, which the governments of poor countries are not in a position to address. But, once again, holders of IPRs are monopolists, and will price accordingly: as much as the traffic will bear. Since the traffic comes mostly from First World companies, which can pay First World prices, they have little incentive to do business with potential Third World users. The Report does not, in the end, claim that current IPRs, and the new and more stronger rights to be imposed by international treaties, work against development, but is reduced to suggesting that those countries which adopt such laws will show a willingness to do business on the terms of the first world, thereby attracting trade and foreign investment... Since, as I said, there's no evidence that strengthing IPRs (beyond levels reached in the First World decades ago) has any effect on technical innovation or economic development, we are, in short, wandering in the dark, making decisions on the basis of short-sighted prejudices, but what else is new?

It may be possible to keep IPRs and still encourage the dissemination of knowledge to the Third World. In India, for example, many western publishers contract with local firms to produce cheap editions of their books; everyone benefits, except for publishers of pirated editions. (Shipping and transaction costs are high enough that it's not worthwhile to mail-order Indian books from the West, though --- traveller's tip --- it makes eminent sense to stock up if you're there in the flesh.) Extending such arrangements to other countries and other sorts of intellectual property, and institutionalizing them, would be Good Things. (Notice, however, that it only exists because of the possibility of piracy.) Technical developments may also make formal IPRs (as opposed to secrets) impossible to enforce. What this would do to intellectual and cultural life in the First World, let alone to development efforts, nobody knows.

Instead of copying from somebody who already knows, you can learn something by figuring it out for yourself. Formalized and made effective, this is called research and development, or R&D. Basic research --- asking, in essence, what makes things tick the way they do --- is a very absorbing pursuit, and one to which researchers naturally gravitate. It's even somewhat prestigious. Ultimately, it's extremely important to technological and economic progress, since some parts of today's basic research are certain to be technical applications tomorrow. The problem is that nobody has any idea in advance what those parts are, and that sometimes ``tomorrow'' is decades away. (Quantum mechanics made solid-state electronics possible, but was devised for entirely different purposes, and the gap between theory and commercial application was around thirty years.) This makes basic research a very bad poverty-reduction strategy, and one would advise poor countries to abandon it entirely to those with money to burn, were it not that there is no known substitute for it as a way of nurturing the skills needed to employ the next cutting edge, whatever it proves to be. There's lots of more applied R&D, on the other hand, which probably must be done by Third World countries (or international organizations), since the rich ones have little use for knowledge about, e.g., tropical agriculture. The Report stresses the importance of reforming the funding systems for research in such countries, with the twin aims of making them more responsive to the needs of their countries, and improving its intellectual quality (e.g. strengthening, or in really bad cases simply introducing, peer review). In keeping with the sensitive-New-Age-technocrat tone of the Report, the section on R&D includes some pieties about ``local knowledge,'' meaning, roughly, what people without credentials know. The examples given, however, look like they either don't involve knowledge, as opposed to sweat, at all (the Rwadan beans, for instance), or look like incompetence on the part of credentialed outsiders (like the foreign advisers in Nepal who, designing an irrigation system, neglected to notice that there already was one: p. 121). In any case, the problem with this sort of local knowledge is that, to be blunt, it doesn't really work, doesn't really tell people how to make things tick: if it did, the societies possessing it would already be rich and powerful, and not in need of the Bank's ministrations. (This is not to say that people without credentials are stupid or even ignorant.)

We thus come naturally to an important point, that raw access to knowledge is insufficient; there must be people in the developing country who can understand it. The report considers this under the heading of ``Absorbing Knowledge,'' meaning education, meaning acquiring technical skills, especially as measured by test-scores. The only other benefit of education given attention is the fact that it improves the lot of women who receive it. (The claim that it makes them better mothers, i.e. that educated mothers have lower rates of infant mortality, would be more persuasive if the data controlled for the correlation of wealth with education.) The Report has the grace to note that there are other ideas as to the proper end of education, but doesn't even bother to justify its neglect of them. (Here is an easy argument for doing so: the report is about reducing poverty, and for that all the humanistic understanding in the world isn't worth a working knowledge of basic mechanics.) Education has been a staple of development efforts since they began, and the Report has little that is innovative, or even surprising, to say about it. (Wonk-information breaks in, to point out that uneducated parents are in a poor position to decide how much education their children ought to receive.)

The geeks are, of course, very taken with their new communications technologies, which, inevitably, they compare in importance to ``Movable type --- the Gutenberg Bible --- [which] is widely considered to have ushered in the Renaissance.'' (The comparison would be more persuasive were it not that the Renaissance was already about a hundred years old when Gutenberg printed the Bible c. 1454.) Now, there's no doubt that things like satellite communications and the Internet are neat, useful and important technologies, but they're not in the same league as the printing press or the telegraph, even in highly wired countries like the United States, Japan and Holland. (At least, if they are in that league, they've not shown it yet.) They are also integrally tied to large, complicated pieces of infrastructure, like power grids and phone networks and semiconductor fabrication plants.

Uses of Internet & c. by the poor: these are entirely ancedotal, with essentially no consideration of whether they can be scaled up, or are cost-effective, or even whether they work at all. This is not up to the Bank's usual standards!

Fully digitized phone networks outside developed world: on its own showing, either fairly small islands, or oil satrapies. ``Leapfrogging'' is not in the cards for Indonesia or Nigeria...

One occasionally has to wonder what planet the authors of the Report are describing.

The picture of a society committed to lifelong learning, then, presents more than the familiar scene of 8-year-olds engaged full-time in learning the basics of reading, writing and math. It also includes grandparents passing on their language and value systems to their grandchildren, while they in turn introduce their elders to the intricacies of the Internet, helping them gain access to information that will enlighten and give sustenance to their later years. [p. 54]

In fact, censorship is conspicous by its absence from the report; it's not mentioned even once. It has been recognized for centuries that free debate and the free exchange of ideas are vital, not just to making governments accountable to their citizens, but to the advance of knowledge itself. Since many --- maybe even most --- of the countries the Bank has to deal with have truly disgraceful records of trying to block the free flow of information and opinion, we may charitably attribute this silence to mere politics, rather than to obtuseness.

One of the noticable flaws of Knowledge for Development is that it is blinded by electronics. Older media have many advantages for developing countries. Take books, for instance. (The Report mentions them once, pointing out that the poor spend less on them than the rich do.) They don't need electricity, or temperature-controlled or dust-free rooms. They don't need hardware-upgrades every few years and they don't have compatibility problems. They don't need operating manuals, and the special skills needed to use them are required for more recent technologies as well. The largest and most bulky among them aren't much less portable than a laptop, let alone a TV, and most of them are much more compact. They do not crash. They can last a very long time. We have centuries of experience of making them useful information sources. They are, in bulk, much more accurate than TV or the Net. Most of all, they are cheap, and they do not go away when funding runs out. Putting a copy of Encyclopedia Britannica in every village school in (say) Nigeria is much less sexy than bringing in the Internet, but might well do more good. (You can buy access to Britannica on-line, but this is the kind of follow-up cost which is particularly likely to be neglected or cut.) The last time anyone looked into the matter carefully --- which, admittedly, was in the mid-1970s, and they confined their attention to the US --- they found that, the more people relied on electronic media, they less information they had, and the more likely that information was to be wrong. The Net probably hasn't changed this very much. What you can get on it, for free, are technical reports and working papers (nice if you're doing basic research), gossip, propaganda and a little old art. (The economic reasons why such things are readily distributed are left as an exercise for the reader's ingenuity.) Everything else, even (or rather especially) smut, costs money.

Wonk-information: Credit and finance, mostly. Also: environmental indicators. Institutional conditions needed to shame polluters into cleaning up are neglected. (Namely, it must be less costly for polluters to reduce their emissions than for them to silence those who complain.)

Economics / the Information Society

Currently in print as both a hardback, US$49.95, ISBN 0-19-52119-7, and as a paperback, US$25.95, ISBN 0-19-521118-9, LoC HC59.7 W659 1998