I finally visited a Millennium Village, the Koraro Cluster, in the northern Tigray region of Ethiopia (estimated regional population is 4.5 million people). The cluster is located in the Hawzien district (population 117,954) and is made up of 11 villages: Koraro plus 10 neighboring villages (68,000 people total). I’d been invited by John MacArthur who until recently managed the Millennium Promise project, an effort inspired and led by Jeff Sachs to create and support as many as 14 Millennium Villages (MVs) in poor, rural Africa. For the few in the development community who don’t already know: the idea of the Millennium Villages is to demonstrate Sachs’s longstanding contention that with enough money to fund adequately enough interventions (agriculture, health, schooling, water, protection of natural resources) in the poorest parts of the world, the poverty and other Millennium Development Goals can be achieved by 2015.
Koraro thrilled and discouraged me. My priors about the Millennium Village project were partly upended and partly confirmed.
I had two questions on my mind. The first one is the one my CGD colleague Michael Clemens has raised: Are there indications that the rate of progress in Koraro is greater than it would have been without the big push (money, expertise, scrutiny from above and afar and so on) from which Koraro has been benefiting since 2005? And if so, is the MV approach scalable and sustainable? Sachs and other proponents of the villages sometimes argue that scalability and sustainability are not the objective – though in the past he and colleagues have predicted that the short-burst village-level package of interventions is a way to end the poverty trap. The first outside assessment of the project by ODI does hone in on those two issues; with good reason those are the objectives against which most students of development want to assess the project. (A side comment: Clemens and his co-author Gabriel Demombynes asked a larger question about the project’s approach. Will or could there be good evidence attributing to the project greater progress on key indicators (infant mortality, farm productivity, etc.) than there would have been anyway without a more experimental (and ideally less insider) evaluation? A more independent evaluation appears to be on the books for a single new Millennium Village to be supported by DfID in Ghana. I hope the DfID-sponsored evaluation does address explicitly the fundamental questions about scale and sustainability – as well as other questions I refer to below.)
In any event this first question (Can the big push be scaled up, and at what cost; and when the Millennium Promise money is gone, will its benefits be sustained?) cannot be addressed on the basis of a single visit to a single Millennium Village.
My second question was less well-formulated: Even if the cost per person is high and the evidence that people are immensely better off than they would have been otherwise does not in itself demonstrate scalability and sustainability, are there other lessons (not only to the affected populations) but to the larger development “project” of reducing poverty worldwide? What are they, and how might they be extracted and shared among the wider interested development community of researchers, advocates and activists?
Before I answer that second question, let me thank the staff of the project in New York who made my visit possible, and most especially in Ethiopia. The word “visit” is inadequate to our Ethiopian hosts’ kindness and energy (answering an endless stream of questions in the field.) It was a whirlwind, intense, dawn to dusk two-day tour: micro-dam; village market; school (students there for tutoring help on a Saturday; a computer room with Internet access), health clinic; dug-wells, conservation ponds and terracing on spectacular Arizona-like mountains to minimize run-off and raise the water table, and other impressive waterworks for mini-irrigation schemes (2 farmers, 1 hectare); papaya/mango farm and seri-culture (silkworms) in Koraro where hadn’t been before; modern more efficient beehives replacing traditional ones; a 17 kilometer road (a rough road providing “African massage”); a house visit (one of the better-off houses in the area the local team explained) to see a smokeless stove, smart in-house latrine technology, and cows kept inside small compound walls to reduce land degradation. . . and finally from afar a spotting of the cave from which now Prime Minister Meles and guerrillas operated in the war against the terrorizing autocrat Mengistu in the 1980s.
Alem Abay (the MVP leader in Ethiopia based in Addis) and Aregawi Aklilu (the team leader in Koraro, an MD who works in the village leaving his family three hours away in the city of Mekele) are a smart, dedicated and energetic pair, and their water engineer, Gigar Kibede and other local team leaders are impressive. Our visit cost them not only their weekend. Under MV-wide smart guidelines to enable visits like ours it was their and not NY’s decision to help make our visit worthwhile in every respect; their own direct costs (vehicles and gas for example) come from the 10 percent of their budget dedicated overall to local management costs. They welcomed me and colleagues, answered questions eagerly and thoughtfully, and were generous in sharing their own insights.
A lot and yet very little can be learned from a two-day tour of a single MV. With apologies for so much throat-clearing, here are my reflections and questions.
Like export processing zones (building on an observation of Chorching Goh, the delightful country economist from the World Bank Addis office who joined us), do the Millennium Villages produce islands of greater prosperity in a sea of poverty – with different rules and expectations - - that via their success for people produce larger and larger wakes and tides, changing the sea itself? That is similar to the idea of Charter Cities proposed by Paul Romer (though Romer is focused on the rules by which rules that matter for governing are made and on private investments to generate wealth rather than on publicly funded inputs.)
Maybe in short the Koraro MV will be most important for its demonstration effect – if not the big comprehensive Sachs vision then site-specific, locally appropriate innovations that take root because they work for people. There are several ways that might happen. One is that the Tigray regional government takes up and promulgates outside the Millennium Village some idea – the micro-dams or the better technology to kick-start sericulture or the particular approach to water conservation. Here’s a small example: The MV team persuaded the Tigray government to try a different approach to incentives to farmers for watering newly planted trees – a tough task where water is scarce and distant. Monthly payments under the safety net program (under which all beneficiaries provide labor to public works of one kind or another) were held back pending evidence over the four-month dry season that farmers had watered the trees adequately. That approach worked. But in the end all farmers got their payments one way or another so it is less likely to work next round. Still, something about rules and incentives might have been internalized at the regional level… or not…. We may never know if no one keeps track.
That something about pricing and incentives is demonstrated and applied elsewhere is one kind of positive spillover. Another type is a change in the demand for MV-type public services in neighboring villages. People living in the other half of the district but outside the 11 “villages” of the Koraro project are using the better health facility inside the project. Will that translate into new political demands for comparable services in their own villages? Is the political environment one where that is possible? In the MV clinic we saw, a well-trained and committed health community worker said one difference from pre-MV work is that now she can send the sputum of a person she fears has tuberculosis for testing at the Tigray regional level. Will the Tigray government make that possible in non-MV settings? An NGO partner is helping cover the initial costs of introducing a more productive breed of chicken for the poultry program the team has introduced – will neighbors in close-by villages want to try those better chickens and if so will the NGO or regional government provide the necessary credit and initial training?
Another kind of spillover might result if young people from Koraro move elsewhere with new ideas and a can-do spirit of how to get things done learned from observing how the MV team operates. Will more young MV beneficiaries move to Mekele, the city two hours’ drive away, than their young non-MV neighbors? Why not follow a sample of Koraro’s beneficiaries and their counterparts for the next 10 years?
What about sustainability inside the villages themselves when the MV money ends? Will differences in the packages of inputs (as well as initial conditions) across villages in different countries matter? What about those differences might be observed and shared by the NY MV team that has not been (I don’t think, at least not on their website or in their preliminary evaluation reports) so far? For example, in Koraro, the Ethiopia team decided to devote less of its budget than originally envisioned for all Millennium Villages to health interventions and more to livelihood interventions. The budget blueprint for all MVs is 40 percent of the total budget to health and 10 percent to livelihood. They reversed those percentages, and also have spent more on infrastructure – including a road, water works and environmental conservation, and they are eager going forward to put more effort into trying new agri-business ventures. All of these are geared to increasing villagers’ income in an extraordinarily difficult environment, where income of about $0.70 a day is heavily dependent on micro-irrigation for farming or on supplementing agricultural income with other income-earning activities. Is the Koraro MV team spending more on livelihood interventions than teams in other villages in other countries? Are Koraro residents enjoying greater gains in income than their counterparts in the non-MV villages? If so what are the implications for any hope of sustaining some of any greater progress on social indicators they are enjoying?
Even with greater income gains, it’s difficult to be optimistic. We talked briefly with four women spinning cotton thread outside the building where silkworms in various stages of development were busy doing their thing. Five hours of spinning would, roughly, add up to about $2 a day of additional income – without spinning wheel technology hardly a likely sustainable escape from poverty.
Still the emphasis on livelihoods struck me as entirely appropriate even absent a permanent escape from poverty -- if it helps ensure that at least some of the benefits paid from the Koraro project budget are sustained because people have more income: more income to replace their worn-out smokeless cook stoves, repair and replace the water pumps they buy on currently cheap and directed credit – and more income for sanitation and other health inputs too.
This question of sustainability would ideally be looked at across the 14 Millennium Villages in different settings; it would require that the project team in New York make available costs and allocations across all Millennium Villages in the last five years.
On choice of interventions, the local and national teams decide – enjoying access to the best advice from experts sent by “the professor in New York”. That makes sense. Still it would be useful to have traditional project analysis done ex ante and ex post on a menu of interventions across villages, as rough guidance to teams working in different settings. That would require more information on costs and prevailing prices and inputs, and on expected and actual returns (in higher income or productivity) than is now available for analysis from the Millennium Promise. Should more go into the seri-culture and poultry projects in Koraro, or into improving the road so the farmer we visited could transport his mango and papaya out not by donkey but by bicycle?
The Ethiopia team views sustainability in a different way than Sachs and colleagues have emphasized in describing a permanent escape from poverty. They noted that the maintenance cost of the 17-kilometer road (an MV outlay of about $800,000) which connects the otherwise isolated Koraro village to a paved road, has been picked up by the regional government. The regional government also covers the salary costs of the community health workers and school staff. And it is the national and regional “safety net” food for work program that pays the villagers to terrace the steep mountainside that rise up as though exploded from below (not for farming them as in Asia or Rwanda but simply to slow the flow of water when precious rains come); and pays the young men who are digging the dug wells and conservation ponds that are helping maintain the water table. That these public works and their maintenance are funded by the region not the project makes eminent sense (the World Bank and major donors indirectly cover many of the salary costs through various programs amounting to as much as $1 billion a year). The regional government also helps enforce security for the new conservation zones (where grass grows at the foot of those mountains as long as water runoff is slowed and goats and cattle are kept out).
Government take-up is one form of sustainability that seems reasonable – assuming overall that this arid zone can continue to support a still-growing population. That, of course, is the larger question of sustainability. Among other changes will climate change bring even less rain in future years? Is the Koraro area, like the dry lands of Arizona, meant to support fewer not more people in the long run? Is it a fool’s game to offer MV incentives that effectively discourage people from moving sooner to small towns and cities, where they can participate in the benefits of specialization and economies of scale and scope in production and consumption of basic services?
A last remark, on implementation. The Ethiopia team has smart and committed staff who are getting things done. They said they are paid about halfway between what comparable staff in the government earn and what comparable staff working for international NGOs earn. In Africa smart and committed people with the necessary training are not the scarce resource. Putting them in place is thoroughly scalable all over rural Africa. The $2.5 million that the Ethiopia team has to deploy this year is not a lot (just over $35 per capita in the Koraro cluster –it excludes of course the costs of the New York office and any costs this year of soil, engineering and other experts from outside Ethiopia) – as Jeff Sachs has forcefully argued.* But even if that amount could be spent everywhere in rural Africa, could it be spent equally well everywhere? What is not scalable, even if there were more money on the table, is the flexibility that the MV project based in New York gives to the Ethiopia team – in how it spends that budget, how many and what staff to employ and what interventions to try.
The team is allowed to try and fail and adjust. I assume that is the case for all Millennium Village teams. For the seri-culture program mulberry trees were grown – their leaves ensure higher quality silk rather than the cotton thread from custer tree leaves. But the team is giving up on the mulberries – they require too much water. Can that approach be scaled up in the public sector, even in advanced economies? I wonder – and if so only with a lot of patience with a process of trial and error, and a lot of emphasis on citizen empowerment as opposed to simple solutions.
The ability to fail and adapt is a rare asset in aid-funded programs. It is why local private and charter schools provide some hope in inner U.S. cities. It is why the private sector has been the engine of growth in China and India. In this singular and unusual respect, the MV experiment seems like a rare exotic bird in a vast flock of highly planned, input-dominated aid programs. It is this aspect of the MV approach—that relies on implementation savvy and flexibility at the local level that cannot easily be replicated let alone scaled up -- even were more money and more fine staff across all of rural Africa fully available to manage MV-style interventions.
The best hope for learning about development from the MVs lies in its potential combination of demonstration effects and positive spillovers that don’t rely on MV flexibility but on local people’s initiatives and demands. The best hope is that the project highlights, through trial and error, a few ideas and interventions that inspire people to be agents of change themselves and thus might just work at scale and over time in a few specific settings.
I urge the project staff in New York to start thinking about that – recording and studying and reporting on the process of trial and error and failure and adaptation across all the villages, extracting for all the world to assess when and where there are those demonstration effects or those spillovers (those positive externalities that economists invoke), and when and why beneficiaries became agents of local change.
A special note to our hosts in Koraro: I apologize if I got the facts wrong or emphases misplaced. Please send corrections, complaints and comments. At the least I hope I have adequately conveyed the passion and commitment and savvy with which you have approached your work.
*This is a correction. My earlier language referred to $350 per capita and reflected my bad arithmetic; I missed a zero in the numerator. Thanks to keen observers and apologies.
Photos taken by Emily Putze