Video
A staff member leads a training session as part of the Targeting the Ultra Poor program. Photo: BRAC.

Can the Graduation Approach Help to End Extreme Poverty? 

An intensive development intervention focused on helping the destitute “graduate” into sustainable livelihoods has shown remarkable promise. Yale’s Tony Sheldon discusses the methodology’s development, extensive evaluation, and future potential.


For a time, microcredit was seen as a panacea to poverty. Billions of dollars in tiny loans poured into developing economies around the globe. Then critics tore into microcredit as usurious and too often burying the poor in debt. The debate is ongoing; a series of randomized control trials have shown that traditional microcredit doesn’t increase average incomes, but tweaks to the model may improve outcomes. Perhaps most importantly, the data showed that microcredit typically doesn’t reach the poorest of the poor. 

One effort to tackle the distinct and complex barriers to economic development faced by the “ultra-poor” comes from the Bangladesh-based NGO BRAC, which in 2002 launched an intensive program aimed at addressing extreme poverty. BRAC’s Challenging the Frontiers of Poverty Reduction—Targeting the Ultra Poor (CFPR-TUP) initiative used a carefully sequenced series of interventions to give destitute household the tools and opportunity to make sustainable change. After the program was carried out in hundreds of thousands of households within Bangladesh, rigorous evaluations showed clear and sustained positive impacts on participants’ ability to “graduate” out of extreme poverty into sustainable livelihoods. 

The core methodology, also referred to as the “graduation approach,” has since been adapted and implemented in more than 33 countries and evaluated through case studies and randomized control trials that have found it is cost-effective and produces long-lasting change.

Yale Insights talked with Tony Sheldon ’84, lecturer in the practice of management and executive director of the Program on Social Enterprise, about the initial development, documented impacts, and significance of the approach. Sheldon has been a consultant to the Ford Foundation on the development, evaluation, and dissemination of lessons learned on the graduation approach from the foundation’s initial explorations in 2006 through 2016. These efforts have been undertaken jointly with the World Bank, Innovations for Poverty Action, and several other organizations.

Q: The graduation approach has shown promise in helping people move from extreme poverty to a sustainable existence. Could you describe the methodology and where it came from?

BRAC is the largest NGO in the world; it’s based in Bangladesh. They’ve been working in microfinance and small business development for 35 years, long enough that some of their clients, today, are the children of people they first worked with. 

In returning to villages where it’s been working for over a generation, BRAC noticed that people excluded from self-selecting, peer-lending microfinance groups were from families that had been excluded a generation earlier. The excluded families typically live in a different part of the village; they are separated economically and socially. Basically, they are the extreme poor in these already poor villages. 

BRAC looked for appropriate development interventions in situations where microcredit—taking out a loan—would be inappropriate or insufficient. Through a number of fortuitous partnerships with the World Food Programme and other groups, BRAC developed a methodology, Targeting the Ultra Poor (TUP), which combines a series of sequenced interventions. 

The core is really that extremely poor households need resources and opportunity to be able to engage with the marketplace and create sustainable livelihoods that support them, their families, and their villages. The methodology involves consumption support, technical skills training, asset transfers, financial literacy training, and exposure to savings, in order to generate a trajectory to a sustainable livelihood. The program requires from 18 to 36 months. Everybody knows from the beginning that it is a limited-time intervention. 

Q: How do the elements of the program work together?

What’s unique about the graduation approach is that it merges aspects from three very distinct areas of development work: social protection, livelihood development, and financial services. Generally, development interventions have focused on one of those, and the experts in each area, and the organizations that focus on each of those areas, work fairly independently of one another. 

This methodology starts with consumption support, so that literal day-to-day survival is taken care of for a guaranteed period, ranging from 12 to 24 months. At the same time, the program staff work with the households to assess skills and viable livelihoods as well as providing any necessary training. Options might include livestock, agriculture, carpentry, or crafts, depending on the local context. 

Integrated into it all is weekly engagement with the program staff in what’s been termed “life skills coaching.” The intensive interaction helps to change aspirations, expectations, and healthcare practices, and to get kids into schools. 

Then there’s a linking with asset providers. In some cases, that might mean providing a sewing machine or carpentry tools. Frequently it’s livestock. Chickens provide a short-term source of both nutrition and income while an investment in goats or cows will take longer, but eventually offer greater profits. Again, there’s coaching on what’s needed in terms of livestock shelter and veterinary care. 

There’s an introduction to financial literacy: how to think about developing a business, how to set something aside every week to be able to draw on in the future, whether for emergencies, school fees, or health expenditures. The possibility of saving was really beyond the conception of most of these households before. 

Q: How should we understand what the program accomplishes?

It’s not “moving out of poverty.” It’s moving out of extreme, destitute poverty into sustainable livelihoods in a way that really alters the social and economic dynamics for the participants. 

After some trial and error and a number of iterations, the methodology has worked extremely well. The vast majority of households that go through BRAC’s TUP program ultimately join the microfinance groups that they had been excluded from for a generation because they now had viable livelihoods. 

For 10-plus years, formal randomized control trials of BRAC, mostly through the London School of Economics, have shown not only that the treatment group does better than the control group, which you’d expect after all the inputs, but the treatment groups continue to do significantly better over time. The improvement, in livelihoods, income, kids in school, and healthcare, actually gets better each year. 

Q: How replicable is it?

People looked at what BRAC was doing, and said, “BRAC can do that because they have a huge presence in Bangladesh, but none of this would work outside of Bangladesh.” So a small group of people from the World Bank and the Ford Foundation decided to adapt the methodology to different contexts and evaluate the outcomes.

There were 10 pilots done in 8 different countries: Haiti, Pakistan, Ethiopia, Yemen, Peru, Honduras, Ghana, and three in India. The sites were very intentionally chosen to be geographically, culturally, and politically distinct settings. They all used the same basic architecture but adapted it where needed.

Innovations for Poverty Action—founded by a Yale economist, Dean Karlan—and MIT’s Poverty Action Lab worked with Ford Foundation and the World Bank to incorporate randomized control trials into six of the pilots. The results of those studies found that the series of interventions was extremely successful. The findings were ultimately published in Science, which tends to focus on hard science, but the results were so positive in sustaining viable livelihoods for the people who’d taken part, not just immediately but one, two, and three years later, that the magazine was very interested. In a cost-benefit analysis, the significant up-front cost was more than made up for by the longer-term economic viability and reduction in subsidy for these households. The researchers said this was one of two interventions that they had seen in the last 15 years that had such a remarkably positive effect.

This created great interest. Originally, we’d thought that it would mostly be large international NGOs that would integrate this model into their practice, which many have done. But there was also strong interest from governments.

Many countries have cash transfer programs where governments provide a subsidy conditioned on certain behaviors being implemented—usually it has to do with enrolling kids in school or going for regular healthcare visits. Policymakers overseeing these programs saw the opportunity to integrate the graduation approach, in order to provide their citizens with a ladder out of extreme poverty. 

Q: How much changes when the approach is used in a new context?

The randomized control trials showed the program is not context-dependent, but we have been finding that adapting the specifics to the context is crucial. In one of the country programs, they chose a variety of chicken that had to be imported from way farther north. The chickens died, so the program was not successful. But what it illustrated was not the failure of the graduation approach, but the crucial importance of making sure you’re selecting the right asset, the right livelihood, and have the right infrastructure in place to secure the asset—that is, vets, food supplies, and so on. 

The specifics need to be adapted to the local circumstances. In the highlands of Peru, the main economic activity has been raising of guinea pigs, which has been going on for hundreds of years. In Ethiopia, the program focused on honeybees. In Ghana, access to savings accounts is done in a totally different way than in Yemen. Program staff have to have local expertise and knowledge. But the core principles of consumption support, life skills coaching, technical training, asset transfer, and introduction to financial literacy—those have been consistent throughout. 

Q: What are the next steps?

We’ve seen big investments in integrating graduation work all over the world. In an effort to capture the lessons learned from these scale-ups, the Ford Foundation commissioned a series of case studies to look at government implementations in Peru, Colombia and Ethiopia as well as an NGO scale-up in India. We wanted to understand what happens when the program was implemented not for hundreds or a few thousand households, but for tens of thousands and ultimately hundreds of thousands of households. These cases were recently published by the foundation.  

What continued to work in the original design? What needed to be adapted? What could we learn about the different elements? Which elements are crucial and which might be secondary? 

There’s now a whole slew of implementations going on. We’re starting to gather more and more information on what good practices are, and which variations are still effective. So far, the government scale-ups have been quite positive. In Peru and Colombia and Ethiopia, we’re seeing significant investment to expand further. We now need to learn the minimum critical specifications that allow for success. 

Q: What are the key challenges to scaling the approach?

I’d say the toughest element is the intensive life skills coaching, which can go on for two to three years. It’s a huge investment on the part of the implementing organization and requires a mix of social and technical skills. 

As governments have gotten more and more interested in this, and because government staff often don’t have the time or appropriate skills, we have started to see the introduction of technology in an effort to standardize the coaching and reduce the cost. 

This is probably most developed in Colombia, where a leading development organization called Fundación Capital has been the technical assistance provider. They have developed a tablet-based application for the technical skills modules and a lot of the life skills coaching. They leave the tablets in the villages to circulate among the households involved in the program. 

What was weekly coaching is now more likely monthly, which obviously reduces the cost. It’s also standardized the quality of information that participants get because it isn’t dependent on a particular staff person’s personality or knowledge or skills. 

We’re still trying to evaluate the results. As the community of practice for the graduation approach expands, there are many open questions. We’re trying to learn and facilitate cross-pollination rather than each implementation figuring things out from scratch.

Q: Are there other hurdles?

I’d say there are two foreseeable hurdles. One is sufficient funding and willingness on the part of large NGOs, and especially of governments, to take on the risk of this kind of innovation. The evidence that this works is quite strong, but it requires a shifting of priorities, budgets, and criteria for hiring personnel. That’s a big investment, especially on the financing side. 

Another major concern, as there’s more experimentation, is, if there are poor results, will it be a reflection of poor implementation by a particular program, or an indication that the graduation approach doesn’t work in certain contexts? For example, if the program generates less positive impacts when life skills coaching is not just replaced with technology but really scaled back, and the households don’t feel the ongoing sense of contact and engagement, is it the underlying approach or the specific implementation that has not lived up to expectations? That’s a very hard thing to tease out.

Q: Are there elements to this approach that lead you to think differently about development?

In some ways, it suggests that, ultimately, to move out of extreme poverty into sustainable livelihoods means developing an entrepreneurial venture that can support a household. So, while international aid resources and government social protection programs are definitely crucial, the solution is not simply a further flow of funds. 

Beyond that, the metrics are there for improvements in income, assets, child nutrition, children enrolled in schools, and quality of housing, all measured over time. But what’s been most striking, I think, is the participants saying engagement in the program, and particularly the weekly coaching sessions, changed their whole mindset and gave them a basis for hope. They believe their children’s lives can be significantly different than what their lives and their parents’ lives had been. That empowerment and sense of opportunity is probably the most astounding outcome for those of us who’ve been working on this approach.

Lecturer in the Practice of Management & Executive Director of the Program on Social Enterprise