Lauren Whitehead

NexThought Monday – How We Can Improve Ultra-Poor ‘Graduation’: The next generation of innovations required to end the most extreme poverty

There is no great innovation without skepticism, much less in the tireless crusade against global poverty. Currently gaining ground as a viable path out of extreme poverty, the ultra-poor ‘graduation’ approach pioneered by Bangladesh-born global non-governmental organization BRAC is no exception.

In 2002, BRAC (where the author works as a Program Manager) began piloting the Targeting the Ultra-Poor program (TUP). The two-year intervention consists of consumption support, asset transfer, rigorous training, access to services, and regular supervision and monitoring for some of the country’s poorest individuals. Evaluations demonstrated that 95 percent of participants not only improved their economic welfare through increased income from livelihood-generating activities and access to savings and social services, among other factors, but also continued that upward trajectory after “graduating” from the program. It was clear that as a response to re-engaging the poorest of the global poor, the ultra-poor graduation approach was tremendously effective in Bangladesh.

By 2010, the Ford Foundation and the World Bank’s Consultive Group to Assist the Poor (CGAP) set out to determine whether the model could be translated and replicated across varied geographic, cultural and systemic contexts with governments, NGOs and financial institutions collaborating at various levels to implement ten pilots in eight countries. Recently, the results of six randomized control trials (RCTs) done one year after completion (three years after initiation) in Ethiopia, Ghana, Haiti, Honduras, India and Peru have underscored at least one very clear finding: the graduation approach works. And it works across contexts.



Testing the Model

In each of the pilot countries (with the exception of Honduras),[1] participants experienced notable improvements across all key indicators and sustained those improvements one year after the close of the program. Policymakers in particular have reason to be delighted as economic variables such as consumption, productive and household assets, household income and revenue, and financial inclusion not only persisted over time, but also retained their statistical significance at least one year later. This presumably reduces the costly burden to the state of fully supporting these households. Evidence from Bangladesh shows that these results could be sustained over time, which would justify the short-term cost of this time-bound intervention as it has prolonged impact for households. In addtion to these economic gains are notable social improvements in political involvement, women’s empowerment, and mental and physical health, each of which may be attributed to the sense of self-esteem, pride and social integration participants gradually build.

As is to be expected, these results have been met with a heavy yet healthy dose of pragmatic skepticism. The critiques have varied: sample sizes are too small; the evaluation timeframe is too short; and, in the budget-constrained environment implementing governments and organizations face, just which components matter most? Wherein lies the greatest bang for the ever-scarce buck?

As leading development and graduation approach practitioners agree, a new set of questions have emerged in what is now being referred to as “Graduation 2.0,” the next generation of graduation iterations around the world. Not too long ago, GiveDirectly presented the findings of an RCT examining the implications of their core business model: enabling donors to “give directly” by placing cash in the hands of the poor (in this instance, in Kenya). Overall, that money was wisely spent on precisely what it was intended for—consumption smoothing and investment in the household; as a result, a new “cash versus cows” debate ignited. Given the potentially cumbersome delivery mechanisms implied for a mass-scale asset transfer of livestock, would it not in fact be more cost-effective and less labor-intensive to simply substitute a cash transfer in lieu of providing a physical asset?

As Esther Duflo, director of the Abdul Latif Jameel Poverty Action Lab (J-PAL) and a key researcher on the CGAP/Ford Foundation pilots rightly asserts, such a comparison between graduation programs and the GiveDirectly model is spurious at best.

“These are not comparable programs. At all,” Duflo said speaking at an event on June 4 at CGAP. “They don’t have the same objectives. The evaluations don’t have the same timeframe. What is really powerful out of the GiveDirectly experiment is that we find if you give money to people they use it pretty responsibly, but the ambition of the graduation program is different. The ambition of the graduation program is a long-term change.”

And one might add, a change to produce a multiplier effect on the long-term welfare of the poor and their families.

By building a holistic model that addresses a multiplicity of vulnerabilities, the graduation approach is intended to supply an injection of capital alongside key elements to combat food insecurity, social isolation, poor access to health care and financial services, and a lack of productive skills for a marginalized subset of the extreme poor. The solution is neither cash nor cows, but a broader question of purpose and scope of the chosen intervention in elevating ultra-poor households. We have to first ask what we hope to gain from implementing an intervention in X region of Y country to reach Z population of the ultra-poor. If the goal is to provide income support and consumption smoothing in an environment with widely used digital financial services such as Kenya’s M-PESA (where appropriate delivery channels do not present a liability), then cash may be a more effective means of delivering assistance. If the aim is holistic improvement in welfare with prolonged impact over time, the scales might tip in favor of a graduation approach.

“If it was as simple as credit constraints,” Dean Karlan, president and founder of Innovations for Poverty Action (IPA), said at the June 4th CGAP event, “we would have solved this long ago by simply doing microcredit or cash transfers. The basic idea here is what we see as a proof of concept. Is it sufficient to do lots of things as a big push at a micro-level to lift (the ultra-poor) up and have it last?”

This might take the form of a full-scale and robust graduation program that addresses a spectrum of needs of ultra-poor families, or it may be a modified derivative. In Latin America, the Colombian social enterprise Fundación Capital is investigating whether computers can make in-person training more efficient by reducing staff costs, creating a more light-touch program. Other programs are investigating similar permutations. “The basic theory [of the graduation approach] is that there are complementarities,” said Karlan. “The training and assets are better when they are together than the sum of the treatment effect of each individually.” Savings behavior is likely bolstered by the introduction of an income-generating asset; healthy participants can better focus energies on their new economic pursuits; and so forth. Further examination will only deepen this knowledge.


Where do we go from here?

We know from BRAC’s experience that the program produces impact over time. We know from the RCTs that these findings hold external validity in a variety of country contexts with diverse implementing actors. What remains is closer examination of the nuances of the program and the scalability of the approach at the national level.

As BRAC founder and chairperson Sir Fazle Abed said, when we address extreme poverty, graduation is “one of the solutions, not the one solution.” It cannot be heralded as a panacea for addressing the needs of the extreme poor. Similarly, the ways in which we approach graduation programs must be conscientious and adaptive to enhance learning and improve outcomes in each iteration.

Can cash transfers improve service delivery and retain outcomes when still paired with home visits and training? Can reduced home visits instill sufficient confidence and guidance to help participants succeed? Testing such premises is a crucial next step for Graduation 2.0.

Looking forward to the next generation of graduation programs and the ambitious marching orders of the Sustainable Development Goals to eradicate extreme poverty by 2030, we can rest assured that healthy skepticism aside, even with continued learning to come, graduation is in fact one proven way forward for the world’s extreme poor.


Lauren Whitehead is Program Manager at the Ultra-Poor Graduation Initiative of BRAC USA


[1] In Honduras, participants received a newly-introduced breed of chickens as an asset transfer; unfortunately most succumbed to disease and died, leaving households bereft of an income-generating mechanism for the program.


Impact Assessment
graduation approach, impact measurement, poverty alleviation