IN RECENT weeks The Economist has been following and supporting the Copenhagen Consensus project—an unusual, ambitious and, some have argued, misguided attempt to set priorities among a range of ideas for improving the lives of people living in developing countries. Starting on April 17th, we began publishing, both in print and on our website*, reviews of essays commissioned by the organisers from leading economic researchers. Each of the papers addressed one of ten global challenges, and proposed possible responses. During May 24th-28th, a panel of distinguished economists assembled in Copenhagen. Their task was to review these papers alongside critical commentaries commissioned from other researchers, to question the various authors, and to decide what to make of it all.

The organising idea was that resources are scarce and difficult choices among good ideas therefore have to be made. How should a limited amount of new money for development initiatives, say an extra $50 billion, be spent? Would it be possible to reach agreement on what should be done first?

The drive behind this venture was supplied by Bjorn Lomborg, author of that modern classic of green demythology, “The Skeptical Environmentalist”. Mr Lomborg is a figure of controversy around the world and especially in his native Denmark, where he is currently head of the Environmental Assessment Institute. The meeting was hosted by the institute, with the support of the Danish government.

Mr Lomborg's role, and the somewhat hubristic character of the undertaking (who do these economists think they are, and when have economists ever agreed about anything?), helped to draw attention to the event. So did the calibre of the various challenge-paper authors and discussants, and of the “expert panel” that met to judge their submissions. This panel of eight included three Nobel prize-winners—Robert Fogel of the University of Chicago, Douglass North of Washington University in St Louis, and Vernon Smith of George Mason University. And the other five, who may collect a few more Nobels in due course, are also eminent in their respective fields—Jagdish Bhagwati of Columbia University, Bruno Frey of the University of Zurich, Justin Yifu Lin of Beijing University, Thomas Schelling of the University of Maryland, and Nancy Stokey of the University of Chicago.

At an earlier stage, the panel had narrowed a much larger number of development challenges, drawn from assessments of the United Nations and its agencies, down to just ten:

•Civil conflicts

•Climate change

•Communicable diseases

•Education

•Financial stability

•Governance

•Hunger and malnutrition

•Migration

•Trade reform

•Water and sanitation

That list itself was somewhat controversial, mainly because of what it left out. (“Why nothing on the role of women?” the panel was repeatedly asked by the press during the course of the week.) But the list of ten challenges will strike many as less surprising than the prioritised list of policies that the panel then drew up.

The top of this list (see table) was not the problem. Ranked first was a package of measures aimed at controlling HIV/AIDS. Next came a set of interventions aimed at fighting malnutrition. The third-ranked policy did raise a few eyebrows, among economists and NGO sceptics alike: “multilateral and unilateral action to reduce trade barriers and eliminate agricultural subsidies.” (“Why so low?” ask economists. “How come so high?” reply the NGOS.) In fourth place, also unlikely to arouse much protest, were new measures to control malaria.

The panel rated all four of those proposals “very good”, as measured by the ratio of social benefit to cost. In fact, by the ordinary standards of project appraisal, they are not just very good but extraordinarily good, with benefits exceeding costs by a factor of ten or more, and sometimes much more. That proposals this good should fail to be adopted for lack of finance is a scandal, especially when you reflect on some of the projects that governments are currently financing.

The bottom of the list, however, aroused more in the way of hostile comment. Rated “bad”, meaning that costs were thought to exceed benefits, were all three of the schemes put before the panel for mitigating climate change, including the Kyoto protocol on greenhouse-gas emissions. (The panel rated only one other policy bad: guest-worker programmes to promote immigration, which were frowned upon because they make it harder for migrants to assimilate.) This gave rise to suspicion in some quarters that the whole exercise had been rigged. Mr Lomborg is well-known, and widely reviled, for his opposition to Kyoto.

These suspicions are in fact unfounded, as your correspondent (who sat in on the otherwise private discussions) can confirm. A less biddable group would be difficult to imagine. The challenge-paper on climate change was written by William Cline of the Centre for Global Development; Mr Cline is pro-Kyoto, and in fact favours even stronger measures to abate carbon emissions than the protocol requires. But the panel insisted on making their own minds up on the issue. Right or wrong, there was no dissent among any of the eight.

Interestingly, an invited gathering of young people from around the world attending a “youth forum” run in parallel with the main event, and hearing the same submissions from challenge-paper authors and their discussants, ranked climate change only ninth out of the ten global challenges. So much for the view that age blinded the expert panel to the long-term dangers of global warming (“they won't be around to suffer the consequences”). Perhaps this should give pause to governments dedicated, or claiming to be dedicated, to Kyoto's implementation.

How valid was this exercise even in economic terms—to say nothing of whether economists should give advice on choices in which ethics and politics play a great part? So far as the economics goes, the main issue is whether the proposals can properly be compared with one another.

A partial answer is that they will be so compared—if not explicitly, then implicitly—whether economists like it or not. Governments do in fact set priorities among the well-intentioned projects they can choose to finance, merely by deciding to do some things and not others. A ranking could be inferred from the choices they actually make. So why not face the question of priorities head-on? If it has to be done, it is best to do it well—and that, despite the undeniable difficulties, means giving the issue of priorities some thought.

Cost-benefit analysis must be the organising method for any such analysis—even if one thinks of this approach as inconclusive on its own, or as little more than a way of ordering one's thoughts more logically. Trying to calculate costs and benefits has the virtue of forcing assumptions into the open. If it turns out that different projects are being compared on the basis of different assumptions about, for instance, the value of a life saved, or the rate at which future costs and benefits should be discounted, it will become obvious, and allowance can be made.

And this, in fact, is exactly what the expert panel found. The challenge-paper on climate change, for instance, proposed the use of a far lower discount rate than the other studies (though it offered reasons for doing so). And estimates of the value of life (or of a “disability-adjusted life-year”, the standard yardstick) likewise tend to vary from study to study. The panel allowed as best it could for differences such as these. It also had to take two other broad points into consideration.

One is that there is a great deal of interdependence among the proposals. For instance, one of the challenge areas was civil conflicts. Several proposals for reducing their frequency, duration and severity were discussed. If such measures were to succeed, however, the benefits would extend far beyond reducing direct losses in lives and livelihoods. The effectiveness of many other forms of aid depends on civil order. In countries riven by conflict, little can be done to speed economic progress. Given peace and security, well-directed aid has a chance of working. In principle, the calculation of costs and benefits should take such linkages into account. In the case of civil conflicts, the implication is that the benefits that would flow from reducing their incidence are much larger than they might at first appear.

A second difficulty, frequently emphasised by members of the panel during their discussions, is that the likely effect of most if not all of the interventions under review depends on the local institutional context. For instance, spending more money on education may yield good results in countries where schools are already held accountable, in one way or another, for the quality of the service they provide. Where they are not held accountable, which is the position in most of the least developed economies, additional spending may simply be wasted.

Since the institutional setting varies greatly from country to country, generalising about the prospects for any particular policy is dangerous. Throughout, the panel leaned towards recommendations which made the smallest demands on government agencies and other institutions—that is, towards proposals where there seemed less to go wrong. If one could assume “good government”, many of today's intractable development problems would be soluble. Needless to say, one can make no such assumption.

Even if these questions of interconnectedness and institutions could be waved away, and basic parameters such as the value of life and the discount rate were clear-cut, the cost-benefit approach would still be difficult to implement. After all, it requires forecasts of costs and benefits, often extending far into the future—in the case of global warming, literally centuries into the future. Such forecasts will always be prone to error. And none of the panel was willing to let a mechanical assignment of costs and benefits drive the process to the exclusion of everything else. Their concluding statement emphasised that weight had been given as well to “the demands of ethical or humanitarian urgency”.

You might think this would have condemned the Copenhagen Consensus project to irrelevance—or at any rate to inconclusive squabbling among members of the panel. Surprisingly, that did not happen. Admittedly, sceptics could argue that the meeting had been designed to yield an apparent consensus even where there was none: the group's ranking of proposals was to be derived by taking the median of all the members' individual rankings, assuring a result, at least in an arithmetical sense. Yet each member was still going to have to endorse the collective ordering—which they did, and which was not to be taken for granted.

Also, since the individual rankings are to be published in due course, along with members' commentaries explaining their choices, it was desirable that the different orderings should look quite similar. Otherwise, the result would command little respect. In fact, the individual orderings were pretty much alike. Despite clear differences of intellectual approach and of ideological tendency, the group did indeed arrive at a broad consensus—both in setting priorities and in agreeing that some projects could not be ranked.

Altogether the challenge-paper authors offered 38 proposals for action. The panel chose to rank only 17 of these, deeming that for the other 21 there was too little information to make a clear judgment about the relative merits. (A proposal was included in the group ranking only if five of the members had included it in their individual ranking. Again, however, there was surprisingly broad agreement about which proposals to rank and which not. With only one or two exceptions, policies tended to be ranked either by all of the members or by none.)

With something close to unanimity, the panel put measures to restrict the spread of HIV/AIDS at the top of the ranking. The challenge paper on communicable diseases, by Anne Mills and Sam Shillcutt of the London School of Hygiene and Tropical Medicine, having reviewed the literature, reckoned that a package of preventive measures costing some $27 billion (in purchasing-power-adjusted dollars) over eight years would prevent nearly 30m new infections (reducing expected infections from 45m over the period to 17m). One study has calculated that part of this package, condom distribution combined with treatment for sex workers who are suffering from sexually-transmitted diseases, would entail a cost of just $4 for each disability-adjusted life-year saved. The implied ratio of benefits to costs is nearly 500—and this is assuming a value of life, based on GDP per head, that is significantly lower than the figure of $100,000 which the panel said it preferred to apply.

Not only are millions of lives directly at stake. In sub-Saharan Africa, the toll of AIDS is so terrible that whole societies are in danger of breaking down. Despite the fact that the issue has received enormous attention of late, efforts to remedy the problem are still curtailed by lack of funds. The daunting scale and urgency of the issue, no less than the estimated costs and benefits of prompt action, persuaded the panel to make this its highest priority.

The same paper made a similarly compelling case for more to be spent on the control and treatment of malaria. Several different interventions were recommended, notably the wider provision of bednets treated with insecticide. Distribution of nets does require a basic health-service network, but experience in Tanzania, for instance, suggests that coverage of the population can be substantially increased given extra resources. Benefit-cost ratios are high. The panel earmarked more than $10 billion of its hypothetical $50 billion for this and a range of other anti-malaria initiatives, putting the package as a whole at number four in the ranking.

In second place, just behind control of HIV/AIDS, came a proposal to attack malnutrition—iron-deficiency anaemia, in particular—through a targeted programme of food supplements. Again the evidence suggests that the idea is feasible, and that it offers exceptionally high ratios of benefits to costs. This would account for the remainder of the putative $50 billion.

What about trade reform? The panel was keen on it, as you might expect, but the issue caused more disagreement than any of the other top four recommendations. The net global benefits of free trade, including the elimination of agricultural subsidies, would be enormous, according to the challenge-paper by Kym Anderson of the World Bank, maybe running into the trillions of dollars. According to one view, the budgetary outlays needed to secure these benefits are actually zero, implying not just a huge flow of net benefits but also a benefit-cost ratio of infinity. Beat that.

But two cautionary notes were entered. First, eliminating farm subsidies, often demanded of rich-country governments as a pro-poor policy, would, at least in the first instance, hurt some developing countries: the ones, often the very poorest, that are and expect to remain net importers of food. A way must be found to help them. Second, trade liberalisation hurts some workers even in rich countries. More generous trade-adjustment assistance may therefore make sense, especially if it is aimed at workers displaced from industries struggling to survive in any case.

Proposals for spending more on water and sanitation were approved, and ranked high, in places six to eight inclusive, with little to choose among them. No education projects were ranked. Nor was Barry Eichengreen's intriguing proposal (see our Economics focus of April 17th) for the fostering of new bond markets. Nor were any proposals for better governance, except for the proposal to lower state-imposed costs on new businesses, which got the nod because the costs are low, the institutional requirements modest, and the possible benefits very great. In all these cases, the panel reckoned there was too little research to go on.

At the foot of the list stand the three proposals on global warming. All require sharp reductions in carbon emissions starting soon, reflecting the view of the challenge-paper author, William Cline, that bold action on the problem is warranted, and quickly. The panel, all in agreement, simply refused to buy it. The issue is real, they said, but not so urgent that such massive abatement costs need to be incurred right now. One of the commentaries on Mr Cline's paper, by Robert Mendelsohn of Yale University, proposed starting with a much lower carbon tax than implied by Mr Cline's three variants—at say $2 a tonne (compared with $150 in Mr Cline's “optimal” carbon-tax plan), rising in later years as more information on both the hazards and the technological opportunities became available. The panel thought that was more like it.

So the Copenhagen Consensus ended, surprisingly enough, in consensus. Mr Lomborg is again to be congratulated for his intellectual entrepreneurship. If rich-country governments want excellent value in return for an increase in their taxpayers' dollars spent confronting global challenges, they could do a lot worse than look closely at the highest-ranked ideas from this exercise.

* See www.economist.com/copenhagenconsensus. See also www.copenhagenconsensus.com.

Copyright © 2004 The Economist Newspaper and The Economist Group. All rights reserved.