Thursday, June 29, 2017
Photo of Bjørn Lomborg

With answers that are bound to please and upset people on both sides of the political divide, paying more attention to unbiased analysis of costs and benefits could be a way of finding the most sensible policies on which all of us can agree.

Visit article 4

Using Data to Find the Middle Ground

PORT AU PRINCE – A sad reality of this hyper-partisan, politicized era is that many policy proposals are immediately identified as either “left-wing” or “right-wing” and lauded and derided by partisans as if by rote, with little room for discussion about soundness or impact. But there is an alternative that could help us get past this political divisiveness: using data to help us focus on the policies and investments that would have the biggest positive impact on society.

This may sound like an idealistic thought experiment dreamed up in an ivory tower, but data-driven policies are having a real-world impact in several countries.

In Bangladesh, the government is officially exploring measures that independent economists from there and around the world concluded would do the most to increase prosperity and wellbeing. And in Haiti, local and international experts are studying dozens of policy ideas to do the same.

In both countries, the well-established methodology of cost-benefit analysis is being applied to interventions as varied as anti-poverty measures, agricultural tariffs, child immunization, legal aid reform, e-government solutions, and rebuilding the armed forces.

Cost-benefit analysis sounds superficially cold or perhaps even cruel. It reduces a complex policy to a stark set of figures: if we invest one extra dollar (or euro, taka, gourde, or peso), we will get a return that is x times higher.

This simplicity belies some complex number crunching from experts. Consider one investment studied in the project Haïti Priorise: agro-silviculture, which assists farmers to grow trees among crops and pastureland.

There certainly are economic benefits: farmers receive extra income from selling leaves, seeds, and timber. There are also social returns, in the form of lower rates of malnutrition, because the additional crops generate more nutritional value than what is being grown currently. There are also strong environmental benefits, such as a reduction in CO₂, better water flow to reduce flooding, and a boost for natural pollination processes.

The cost-benefit analysis includes all of this: economic benefits, social benefits for the country, and environmental benefits (and perhaps costs) for the world. As a result, this approach cuts across the political spectrum.

A cost-benefit analysis of the United Nations’ development goals illustrates this. On the one hand, it found that freer trade would be one of the best development policies, lifting 160 million people out of poverty and making every person in the developing world $1,000 better off, on average. Although out of favor with the Trump presidency, free trade is a right-wing staple. But, by establishing the benefits to the world’s worst-off, the research also provides reasons for the left to reconsider their opposition to it.

Conversely, another highly recommended development priority, improved access to contraception and family planning, is strongly supported by the left and often opposed by the right. The research points not only to reduced maternal and child mortality, but also to economic benefits, perhaps giving the right more reasons to take another look.

Typically, governments do not conduct cost-benefit analysis except in limited ways, examining at one policy at a time. This scattered approach prevents policies from being compared. Instead, taxpayer money is often spent on whatever issues receive the most media attention or have the most engaged stakeholders.

Cost-benefit analysis should not be the only input to a government’s decision-making. But for countries like Bangladesh and Haiti, it provides a vital, independent injection of data on which sound decisions can be based.

In Bangladesh, three top priorities were identified by the research and by a panel of experts including a Nobel laureate and Bangladeshi development specialists: e-government solutions, improved TB response, and child nutrition. This work is also informing the National Nutrition Plan of Action, as well as “Digital Bangladesh” policies. The Bangladeshi government’s own think tank is helping to disseminate the information about top-ranked solutions to ministries and government entities.

In Haiti, researchers are still working – but early results show promise. A surprising study examines preventable deaths from trauma and illness. There are many health system challenges, which other researchers in the project explore. But one answer could be more ambulances. Only 2% of accidents and emergencies are reached by free ambulances, of which there are just 100 for a population of ten million.

R. Christina Daurisca, an economist at the Haitian Ministry of Public Health and Population, found that nearly 4,000 lives could be saved annually by creating a national ambulance network. But this comes with a high annual price tag of around $33 million – more than 1% of Haiti’s entire national budget. Each dollar spent will create nearly three dollars of social benefit – a solid, if not outstanding investment.

A more modest option could, in the short term, do much more for every dollar spent. Training teachers and community leaders across the country to act as paramedics and first responders would have a one-time cost of just $1.1 million but would save around 700 lives every year. Each dollar would produce $16 of social good, helping Haiti much more.

The research for Haiti, Bangladesh, and, soon, two Indian states helps provide inputs that allow any of us to answer the question, “Where will extra resources achieve the most good first?” It is an approach that I believe should inform every country’s policy discourse.

With answers that are bound to please and upset people on both sides of the political divide, paying more attention to unbiased analysis of costs and benefits could be a way of finding the most sensible policies on which all of us can agree.


Next opinion

Photo of Steven J. Klees

The costs and benefits being analyzed, it too often seems, are in the eyes of those beholding the data.

Visit article 6

The False Promise of Cost-Benefit Analysis

COLLEGE PARK, MARYLAND – Can we really use data to overcome the left/right ideological divide and identify policies that, as Bjørn Lomborg puts it, “would have the biggest impact on society?” Lomborg is a leading advocate of cost-benefit analysis (CBA) as the best way to choose between policy options. He is not alone in this belief, though he is perhaps extreme in his faith in CBA.

Lomborg is the Director of the Copenhagen Consensus Center (CCC), and both there and in his regular commentaries, including for Project Syndicate, he reports on the results of CBA to recommend development priorities. Over the years, Lomborg has hired more than 300 economists, including seven Nobel laureates, to carry out these economic evaluations. Unfortunately, these – indeed, all – CBA results are infused with the ideology of those funding and conducting them – and thus offer very limited information for public policy choice.

CBA, and its common use in studies of the rate of return on investment, originated in the United States in the 1930s to assess water resource projects. But, while the general methodology has become part of orthodox economics, it is mostly confined to academic research. Institutions like the World Bank use it to make decisions, but governments rarely do (and most often to offer a post hoc rationale for some policy choice).

The fundamental problem is that these types of economic analyses are so loosely specified in practice that the result is almost always full of analyst bias. Indeed, CBA provides a misleading basis even for prioritizing policies within a sector. One well-known example is that, starting in the 1980s, CBA was used to argue that the returns on primary education were much higher than the returns on higher education. The World Bank went around the world encouraging or requiring governments to cut subsidies for universities and expand those for primary schooling.

But this CBA-derived conclusion was based on a very narrow, mechanistic, and simplistic calculation. In 2000, a joint World Bank/UNESCO report argued that many of the economic benefits higher education generates were not counted in the CBA. These included: technology development, discoveries and invention, private-sector innovation, a better investment climate, and democratic functioning, among many others.

While difficult to measure, especially in monetary terms, such factors are among the most critical benefits that higher education brings to societies. Analysts have concluded that basing the choice to favor primary education on a narrow CBA has disadvantaged many developing countries, by preventing them from competing internationally on the basis of a more educated labor force, rather than on low-wage labor. Yet, in 2014, Lomborg commissioned a study by an economist that, with another narrow CBA, came to the same – if not discredited, at least debated – conclusion favoring investment in primary education over higher education.

The costs and benefits being analyzed, it too often seems, are in the eyes of those beholding the data. A 2004 conference that Lomborg initiated, for example, focused on global warming; for the three widely touted climate-change policies examined, Lomborg’s panel concluded that “costs were likely to exceed benefits.”

Columbia University’s Jeffrey Sachs, criticized these findings, arguing that Lomborg “failed to mobilize an expert group that could credibly identify and communicate a true consensus.” And John Quiggin of the University of Queensland was even more blunt: the 2004 conference amounted to “exercises in political propaganda,” and the “panel members were slanted toward conclusions previously supported by Lomborg.”

The CCC continues to assess climate-change policies. On its website is a testimonial by the American Council on Science and Health, a pro-petrochemical industry group, praising Lomborg’s advocacy for “cheap fossil fuels for those living in impoverished countries.”

Biases are often not explicit or visible. For example, in a recent commentary, Lomborg argued that his CBA “found that freer trade would be one of the best development policies, lifting 160 million people out of poverty and making every person in the developing world $1,000 better off, on average.” In fact, the costs and benefits of freer trade are highly contested by economists, and these numbers have to be taken with a grain of salt, not as reliable inputs for policymakers.

I have taught courses on CBA for many years. My examples above are not exceptions. All applications of CBA are controversial when economists with different explicit or implicit biases study the same topic. Dozens, often hundreds, of assumptions must be made to estimate just the costs of almost any policy. Well-intentioned analysts can and do differ on what assumptions to make and often arrive at very different estimates.

This is even truer on the benefits side of the equation, where the assumptions concern not only how to measure returns, but what counts as a return in the first place. The bottom line for me is that economics, like all fields of study, is irreducibly ideological, and the tools that it has developed are far from value-free.

Lomborg has been very critical of the 2030 Agenda for Sustainable Development and the broad democratic processes that led to its 17 goals and 169 targets. He greatly preferred the Sustainable Development Goals’ predecessor, the Millennium Development Goals, which were fewer in number and reflected technocratic decision-making.

I am sympathetic to Lomborg’s view that the SDGs’ sheer number of targets makes prioritization extremely difficult. But a return to technocratic decision-making is not a valid way to reduce 169 targets to the 19 he thinks will have the most impact.

Yes, governments and policymakers need evidence before they act. But evidence generated by CBAs should always be contested. In choosing our priorities, we need to honor and reconcile different views about what counts and how to count it, in what will be – and should be – messy, participatory, democratic processes.


The Costs of Cost-Benefit Analysis

Read Comments (4)
  1. Comment Commented

    I believe there is merit in what both experts says. That said, there is a lot of potential in using data. Perhaps, it is better when applied to a narrower scope such as in Lomborg's examples of Haiti and Bangladesh. On the other hand, using CBA to arrive at a general conclusion regarding education misses out the cultural and political differences in each country.
    Also, I feel it is hard to identify costs and benefits that easily. I believe in Hayek's theory of asymmetric information. In such an environment it may not be possible to come at a dollar value for cost and benefit.
    At present, it maybe better to be cautious about CBA's role in large scale projects. It would be nice if future research focuses on the effectiveness of those policies guided by CBA. Read more

  2. Comment Commented

    The irony of benefit cost analysis is that it is often embraced by non-economists more than by economists, many of whom understand fully its shortcomings. For a review of the literature on its inappropriate use in environmental areas, see my article:
    The Incompatibility of benefit-cost analysis with sustainability science, in the journal Sustainability Science, January 2015 Read more

  3. Comment Commented

    All CBA does is reduce a future stream of costs and benefits into a dollar number you can compare against money in the hand earning interest. And the assumptions about what the costs and benefits are? And future interest rates? They are a particular view of the future, the view of the author of the study.

    In relatively shorter time frames for well defined and costed projects, fine. For anything else a CBA is just a way of putting a number of best guesses about the future in today's dollar terms.

    CBA also fails pretty dramatically when you look at the effect of rather long term outcomes. A climate change catastrophe, for example, affecting much of life on the planet is discounted by CBA, by both the time it is in the future and the discount rate used to devalue future events. Plus whatever estimates of future damage are used. So long as you can make a buck in the near term, the long term doesn't matter so much with CBA.

    USeful in it's proper place. A nice simplification to compare different timeline streams of future costs and benefits. But otherwise no more valuable than any other view of the future. Read more

  4. Comment Commented

    Cost/benefit analysis is utilitarianism in an ineffective disguise, and the limits of the former are well known. That's not to say that such analysis should be banished from policy conversations -- but to insist that its many problems and limitations be full taken into account.

    An elaborate cost/benifits argument can be made in favor of eugenics, for example, but I don't think many of us would like to live in a society where that argument had been won. And in any event, the selection of factors that will be weighted in a CBA are inevitably made subjectively by whoever is doing the analysis.

    Leaping the normative divide requires more than an optimization algorithm, and it is in my view dangerous to pretend otherwise. Like it or not, tough decisions taken in an ocean of hard data are still made on the basis of who's ox is being gored more often than we would like to admit.
    Read more