Monthly Archives: September 2015

Why is it so hard to make priorities in development?


Originally published on Humanosphere

British trains used to ‘slam doors’, metal slabs that swung outward, a latch on the outside. If the train was pulling into a station, passengers could reach out through the window, swing the door open and hop off without waiting the extra few seconds for the train to come to a full stop. During long delays, they could lean out, have a cigarette and shut it again when the train started moving.

The downside of the slam doors was the accidents. Every year, a few people fell off the back, pulled under the wheels. Passengers waiting at train platforms got bashed in the face by the doors as they swung open. The trains put up signs, of course, don’t open this, watch your step, but every year, the doors caused between 5 and 10 deaths, and dozens of injuries.

The need for replacing the doors seems obvious, but for decades, the UK stubbornly refused. Updating the doors would have required designing an automatic opening mechanism, then paying workers to replace each swinging door with a sliding one. With hundreds of trains, thousands of doors, the cost was in the billions. So Britain did nothing. It left the doors as they were, cleaned up the mess from the fatalities but did nothing to prevent them.

I spent last month in Seattle. The city has been in a decade-long debate about what to do about the Alaskan Way Viaduct, an elevated waterfront freeway, one of the city’s busiest north-south arterials. In 2001, an earthquake rattled the viaduct, weakened it. Even after the city added extra steel and sensors to all the weak points, everyone knows it’s not going to survive the next earthquake. As my friend, a Seattle city planner, puts it, ‘the next time someone sneezes on that thing, it’s coming down.’

And again, it seems obvious what the city should do. Close the viaduct, tear it down, build a safer one. But they haven’t. Fourteen years now, it simply remains, carrying just as many passengers as before. When the next earthquake happens—and in Seattle, it is indeed a when, not an if—a not-insignificant number of people will die in their cars, crushed by concrete and steel. There’s even a road underneath the viaduct, a popular tourist area, bike lanes, hot dog stands. Those people, if the earthquake is during the day, will probably die too.

Countries have a formula they apply to these sorts of problems, it’s called the value of preventing a fatality, of VPF. In Jonathan Wolff’s Ethics and Public Policy, where I read about the train doors, he notes that in the UK, the value of a human life is £1.4 million. In the United States, it’s apparently $6 million.

What that means is, since the train doors killed up to 10 people a year, Britain was willing to invest up to £14 million in retrofitting them. If the cost went over the VPF, it would leave them. It did, so it did. In Seattle, tearing down the viaduct, spending years rebuilding it, would interrupt the commute of millions of people, would cost billions in lost productivity. Whatever mayor or governor decided to do it would be voted out of office.

I’m not even sure I disagree with leaving the viaduct up. I biked beneath it almost every day in Seattle, I took that photo standing right under it. A small chance of, say, 60 people dying in exchange for keeping a major urban arterial might actually be a worthwhile trade-off.

What’s interesting to me isn’t that we make these choices, but that we are only allowed to make them invisibly. A politician who stood at a podium and said ‘saving 10 lives isn’t worth more than £14 million’ would be seen as a monster. Yet that is indeed the decision Britain’s politicians reached, and the one we live with intrinsically in things like our drinking age, our speed limits, our pharmaceutical regulations, our sentencing laws. At the population level, almost every decision means lives lost. Since 1979, 10 people have apparently been killed by Bic cigarette lighters. Is banning them worth the inconvenience of millions of people taking slightly longer to light their cigarettes? Meh, probably not.

This month, the international community will come together to sign the Sustainable Development Goals, an ambitious framework to end poverty, achieve gender equality and improve global health. As I’ve written before, it’s a mess, a soup of unmeasurable indicators and undefined targets, things like ‘halve per capita global food waste’ and ‘encourage companies … to adopt sustainable practices’

One of the reasons it’s so bad, I’m convinced, is that in development, we aren’t allowed to talk about these trade-offs, the kinds governments and citizens make every day. With the viaduct, with cigarette lighters, we traded a small risk of fatalities for the inconvenience of preventing them. With train doors, Britain decided there were more pressing risks to spend its resources on, more passengers it could save for its pounds elsewhere.

Yet in development, we never talk like this. One of the main criticisms of the Millennium Development Goals, the precursor to the Sustainable Development Goals, was that they neglected some development issues in favor of others. Domestic violence, human trafficking, corporate tax evasion, all of them got left behind.

I remember a meeting at my last human rights job. We were a department of four people, trying to plan our activities for the next year. We brought in a strategy consultant, he gave us each a matrix of organizational priorities, stuff like land resettlements in Africa, foreign direct investment in Myanmar, sexual harassment in the Middle East. He asked us to rank them in priority from high to low.

After a few minutes of scribbling, one of my colleagues reported that she had marked every issue as ‘high priority.’ The consultant looked confused. ‘Those are all really critical issues,’ she said, ‘with profound impacts on peoples’ lives. We should be working on all of them.’ I looked around, everyone else in the department was nodding.

It’s understandable, this. No one wants to argue that one development issue is more pressing than another, to stand up and declare ‘state surveillance of political dissidents affects fewer people, and less severely, than human trafficking. Lets prioritize the latter.’ No one wants to admit that working on one problem leaves all the other ones in place.

When you work at Nike, when you have to decide on launching an ad campaign for sandals instead of sneakers, you’re allowed to make arguments why one should take precedence over the other. But in development, lives on the line, you can’t. So we say yes to everything, we plan our years without differentiating between priorities, we stretch ourselves thin. And we fail, again and again.

I don’t know what’s going to happen with the Sustainable Development Goals. Maybe governments will pick the ones they want to reach, will defend the choice to leave others behind. Maybe they will be honest about the choices they make, we make, all the time, the trade-offs that come with resource constraints and political realities.

But I think, I fear, that they won’t. That the international community will fail to make the decisions that governments do every day, that we will give developing countries rules and principles, but no tools for choosing between them. That we will, once again, tell poor countries to replace their train doors and rebuild their viaducts all at once.


Filed under Random

The Millennium Development Goals Were Bullshit. And That’s OK.

All year I’ve been trying to decide what I think about the Millennium Development Goals. You remember those, right? In 2000, 189 countries and 23 international organizations committed to eradicating poverty, promoting gender equality and improving global health by 2015.

As the deadline approaches, the internet has filled up with equally unconvincing arguments for and against the MDGs. Most of the ‘they’ve failed!’ condemnation is by people who think foreign aid shouldn’t exist at all, and most of the ‘they’ve succeeded!’ cheerleading is by people who were there for creating them.

So a few months ago, I started reading institutional and academic reports on the Goals. Their creation, their progress, their data, I wanted to know what the evidence, what the people gathering it, actually said.

I came away even more conflicted than when I started. Defenders of the Goals say they were great PR, an excuse for the global north to start sending money southward again. Critics of the Goals say they were unrealistic, a top-down tickbox exercise inflicted upon the developing world without their consent.

I think they’re both right! Here’s the arguments for and against the Millennium Development Goals, and why it’s so hard to pick a side.

1. The MDGs resurrected development aid

Let’s start with the non-arguable stuff. In the mid-1990s, development aid was in crisis. The Cold War had just ended, and without communism-prevention giving rich countries a reason to give money to poor ones, the air was slowly leaking out of the field.

International organizations needed a big idea to shake governments out of their ennui, to inject enthusiasm—and more importantly, money—back into poverty reduction. After years of deliberations, they come up with the MDGs, eight quantitative(-ish) targets for the world to rally around. By 2015, they pledged, they would halve extreme poverty, cut maternal mortality by three quarters and reverse the spread of HIV/AIDS. Oh, and reduce hunger and battle child mortality and improve sanitation and provide safe water and achieve universal education.

Almost immediately, donations started increasing. Between 2000 and 2005, aid flows went from $60 billion per year to $120 billion. Health spending doubled; primary education spending tripled. Donor countries started coordinating their projects, rallying around specific outcomes and quantitative monitoring rather than the ad hoc before-and-afterism they used to work under. As one evaluation puts it, ‘a cascade of statistical and analytical work got underway once the MDGs gained currency.’

The MDGs increased donor commitments and coordination; that part’s undisputed. But just as fast as the new money came in, though, so did the question of whether it was actually making a difference.

2. The MDGs aren’t going to be reached

Look, we’re not going to make the MDGs, not even close. I’m not going to go into a whole big thing where I talk about each Goal and how X number of countries are falling behind or whatever. Even a cursory glance at the Goals themselves shows that reaching them was never the point.

Take Goal 1, ‘eradicate extreme poverty and hunger.’ It’s split up into a few targets, components defining what reaching the Goal means in statistical terms. The first target for eradicating poverty and hunger is pretty reachable: Halve the proportion of people living on less than $1.25 a day. We did that years ago. Check.

But the next target under that Goal is ‘achieve full and productive employment and decent work for all.’ Oh is that it, MDGs? A job for every single person on the planet?

It’s like this going down the right down the list, reasonable targets alternating with utter fantasy. Goal 2 is ‘achieve universal primary education.’ Denmark doesn’t have universal primary education. The rest of the world was never going to get there with 15 years and a 60 extra billion dollars split 40 or so-odd ways. One analysis points out that 38 countries started the MDGs with enrolment rates below 80 percent. Achieving the goal by 2015 would have meant ‘improv[ing] enrolment at a rate that has not been achieved by a single country for which post-1960 data is available.’

This is why I’ve spent the first six months of this year rolling my eyes at op-eds gloating about how the aid community has ‘failed’ to reach most of the Goals. Of course we did! Half them are ridiculous!

I should also mention here, speaking of ridiculous, that many of the targets don’t have particularly trustworthy data behind them. Lots of the statistics are based on household surveys, dudes with clipboards wandering through villages, asking people about their kids’ birth weights and whether they use mosquito bednets. Only one African country, Mauritius, even registers births and deaths according to UN standards. Maternal mortality rates for the year 2000, the year the MDGs were signed, were estimated to be between 210 and 620 per 100,000 births. Reducing something by 66 percent gets a lot harder when the baseline has a margin of error of 300 percent.

3. The MDGs might not have made a difference

But the real debate isn’t over whether the Goals, measured by their own science-fiction targets and fingers-crossed data, fail or succeed. It’s about whether they had a galvanizing effect, whether all those extra donations resulted in leaps forward for the indicators the international community decided to work on. It’s incontrovertible that nearly every indicator of human well-being—life expectancy, literacy, income, mortality—has improved in the years since the MDGs were adopted. The question is whether that would have happened without them.

By now, there’ve been a few studies on this, and it doesn’t look great for the MDGs. In 2010, an analysis found that only five indicators (out of 24) accelerated after the MDGs were adopted, and that was only in half to two-thirds of the countries where they were being applied. China, the greatest poverty-alleviation success story of the last generation—28 million Chinese people were lifted out of poverty every year between 1990 and 2008—barely participated in the MDGs. The latest MDGs Progress Report notes that when in 2000, only 6 percent of the world’s population had access to the internet. Now, it’s 43 percent. Considering all the technical and economic changes that have happened during that time, is anyone really going to argue that that it was a set of donor targets that were the critical factor in that rise?

That critique, though, only works when you look at the global picture. Zooming in, you find specific places, specific ways, where it seems like the MDGs have worked. The Center for Global Development’s Charles Kenny, for example, has shown that according to trends from before 2000, primary education rates in developing countries should have reached 76 percent by 2010. They actually reached 81 percent. Maternal mortality should have been 221 per 100,000 births; it ended up 203. That same analysis that found only 5 indicators improving globally post-2000 also found that, in Africa, 16 of them did.

But you can julienne the statistics however you want. The challenge of the MDGs, and why it’s so hard to make up my mind about them, is because the ways in which they’ve failed are so easy to measure, while the ways in which they’ve succeeded are so not. As Kenny and others have pointed out, coordinating donors around measurable goals, renewing the reasons for rich countries to invest in poverty reduction, these things matter. They’re just not as quantifiable as literacy rates or HIV prevalence. In the least developed countries, where aid makes up a significant percentage of the national budget, they may even have been decisive. The shittiness of the data, and the non-existence of the counterfactual, means we’ll never know for sure. 

4. The MDGs don’t measure what matters

Another, slightly more convincing, criticism of the MDGs isn’t about whether we reached them, but whether they were worth reaching at all.

Remember Goal 2, ‘Achieve universal primary education’? The way the MDGs chose to measure this was through enrollment rates, how many kids attend school every day. By that measure, poor countries have made significant headway toward the Goal. By the measure of whether they actually learned anything, however, the evidence is less inspiring:

In many cases the rapid expansion of schools aimed to grant an increasing number of students access to primary schools had in many cases a deteriorating effect on the learning quality, first and foremost due to teacher shortages, resulting in single teacher schools with one teacher responsible for one multigrade classroom, or the hiring of so called para-teachers with considerable less educational qualification as regular teachers. … 130 million children completed primary education but without being able to read or write.

This, according to MDG skeptics, is their real weakness: They focus on inputs, the ability of a country to provide a service, rather than outcomes, whether those services are actually improving people’s lives. In doing so, they’ve encouraged governments to work on means and ignore ends. It’s like pledging to lose weight but never actually weighing yourself, just counting how many Cinnabons you eat.

I’m tempted to accept this critique—I’ve been bitching about measuring ‘gender equality’ by the percentage of women in national parliaments for years—but it’s worth pointing out some caveats in it too. Not all the indicators measure inputs. Some of them, like the target on providing access to HIV treatment, really do measure the outcome the MDGs are trying to reach.

And yes, enrollment rates are not a perfect measure of learning and women in government is not a perfect measure of gender equality. But what is? ‘Education’ and ‘equality’ are inherently qualitative concepts—so is ‘development’, while we’re at it. Maybe the Goals should have used test scores rather than enrollment rates to measure education, or the gender pay gap to measure equality, but those are just as jukable, just as subject to over-emphasis by logframes and donor tickboxes, as any other proxy. These are problems with quantification itself, the map versus the territory, not the MDGs in particular.

5. The MDGs were for donors, not governments

The MDGs might have been signed by a huge number of developing countries, but they were written almost entirely without them. The original idea, the Millennium Declaration, was developed by a country-club of rich development agencies in hotel conference centers throughout the 1990s. By the time the rest of the world was presented with the Goals, donors had already identified the problems they wanted to solve and the indicators they would use to measure them.

The result, condensing all the world’s development challenges into fewer than 10 goals, has encouraged countries to zero in on donor-approved problems, rather than solving the ones they actually have. Rwanda, according to one analysis, devoted 24 percent of its health spending to HIV/AIDS, even though only 1.6 percent of its population has it. Malaria might be a huge cause of death globally, but in Mongolia, one of the poorest countries in Asia, it doesn’t even register.

Again, it’s easy to say that donors picked too few goals, conducted too little consultation. But consider the opposite scenario, a set of Goals that included every development problem, that were perfectly applicable to every country in the world.

Actually, don’t. Just look at the sequel to the MDGs, the Sustainable Development Goals. Where the MDGs were primarily a tool for donors, the SDGs (stick with me on the acronyms here) have been the most inclusive, taking shape over a five-year, international consultation process that deliberately sought feedback from every institution with an incentive to push their pet issue onto the list.

The result is a jambalaya of impossible ambitions, utopian targets and unmeasurable indicators. Where the MDGs sharpened their attention down to 8 goals and 24 indicators, the SDGs leave no societal challenge behind, comprising 17 goals and 169 targets. Check out everything we, the world, will achieve before they’re finished:

In just sixteen years’ time we will have been able to end poverty in all its forms everywhere; achieve full and productive employment and decent work for all; end hunger and malnutrition; attained universal health coverage; wipe out AIDS, tuberculosis, malaria, and neglected tropical diseases; provide universal secondary education and universal access to tertiary education; end gender discrimination and eliminate all forms of violence against all women and girls; ensure adequate and affordable housing, water, sanitation, reliable modern energy, and communications technology access for all; and (strangely) both prevent and significantly reduce marine pollution of all kinds alongside preventing species loss. If that’s not enough, we will have also eliminated all discriminatory laws, policies, and practices.

This is why I have trouble dragging the MDGs for condensing development challenges down to just a few issues. The MDGs worked, to the extent they did, by coordinating donors around a discrete set of objectives, a consensus on what the world needed to fix and how. That necessarily meant leaving some development problems un-addressed, prioritizing some issues over others. It may sound crass in development, when you’re talking about letting people live with one disease while you work on curing another, but in every other area of human endeavor this is called having a strategy.

6. The time for development goals has passed

The closest thing I come to having a conclusion about the MDGs is that yes, they were bullshit. And yes, they were probably worthwhile.

But I’m not sure the next round of bullshit is going to be. During the 15 years we’ve spent debating the MDGs, the nature of the problem they set out to solve has changed. In 1990, the ostensible start date for the MDGs, 79 percent of the world’s poor lived in stable low-income countries. By 2010, only 13 percent of them did. These days, the vast majority of the world’s poor remain that way either because their countries are riven with conflict (Yemen, Syria, Somalia) or because they have political systems too captured or too gridlocked to be worthy of the term (Zimbabwe, Bangladesh).

In other words, the MDGs may—may—have been the right development initiative for the world of the late 1990s, but they are increasingly irrelevant to the one we have now. Only 1 in 10 poor countries get more than 20 percent of their budget from aid. Even in the poorest countries, domestic health and education are orders of magnitude greater than aid flows. Poor people in China are not poor because their country lacks to resources to make them not be. They are poor because their government would rather spend those resources on high-speed trains.

Maybe that’s a defensible decision for the long-term and maybe it’s not; we shall see. But what the MDGs never did, never could, was pressure governments to develop their own systems to solve their own problems. In 2030, only 8 percent of the world’s population will live in countries classified as “low-income.” Most of the world’s poor will live in cities; many of them will be employed. Informal employment, exploitative working conditions, dysfunctional education and healthcare, they will persist in other countries for same reasons they do in our own.

So did the Millennium Development Goals fail or succeed? I still don’t know. What I do know is that rallying around a set of utopian, un-enforced, top-down targets seems to have worked in the places where development agencies, where we, mattered. If we want to solve the next generation of global poverty, we should ask ourselves where we still do.


Filed under Development, Serious, Work