Graduation: What’s next for ultra-poor programmes?

Blog Firms and Entrepreneurship

Results are in: Research shows that BRAC’s one-shot, ‘big-push’ intervention targeting the ultra-poor with assets and skills can lift extreme households onto a more sustainable path out of poverty. In this blog we presented updated findings on the long-term impact and sustainability of the approach from studies showcased at the recent IGC-BRAC conference: Tackling extreme poverty: Evidence from the field.

Strikingly positive results from the evaluation of BRAC’s ‘Targeting the ultra-poor’ programme have been echoed by subsequent adaptions of the original BRAC model in other low-income countries.

Professor Robin Burgess, Director of the IGC, described the programme as

“the one thing […] in the world which is capable of lifting people out of poverty”.

What remains to be understood is not if the approach works, but who it benefits, how much it benefits them, and fundamentally, why this approach has succeeded where others have failed.

Further research has been conducted in order to understand which combination of components matter most for eradicating extreme poverty? And how can the programme be scaled up and adapted to other settings to reach even more people?

Researchers at the recent IGC-BRAC conference presented new results from an extensive seven-year follow-up of the original BRAC programme (93% of households from the original study were tracked). Other studies evidenced the feasibility of translating this model to different contexts and shed new light on the complementarities of programme components.

Over the course of the conference, four key facts emerged.

#1 In-kind asset transfers are better at increasing labour market attachment than pure cash transfers

The provision of productive or in-kind asset transfers appears to be more effective at increasing the participation of ultra-poor households in local village labour markets. This supports conclusions from other studies that suggest that the extreme poor face a mired of barriers, beyond just credit constraints, that inhibit their access to productive employment opportunities. Finding effective ways to mitigate and remove such barriers is essential in combating extreme poverty and helping households work themselves up the economic ladder.

In an effort to better understand how different types of transfers can drive poverty reduction, researchers Imran Rasul (UCL) (with co-authors Attanasio, Bandiera, Burgess, Khan, and Khalid), conducted an experiment in Punjab, Pakistan where randomly selected households were offered either an unconditional cash transfer or an equivalently valued in-kind transfer (typically livestock).  Household outcomes were measured to assess the differential impacts of the two different types of transfers.

Initial results from mid-line surveys indicate that while both programmes increase household economic activity, asset transfers did so more than cash transfers. Figure 1 below compares labour market productivity levels (hours spent working) of households that received assets transfers (in-kind) to households that received unconditional cash transfers (UCTs).


The results suggest that labour market attachment is significantly higher for those households that receive in-kind transfers relative to those that received UCTs. However, as might be expected, total expenditures, investments, and ability to save increased slightly more for households receiving cash transfers. Results from Pakistan suggest that longer term tracking of households is needed to understand how combining cash and assets transfers can be used to reduce extreme poverty.

#2 Asset provision is not enough – complementary training and savings components may be crucial for the success of graduation approaches

Simply “dropping a goat” on someone does not work. For graduation approaches to be effective, complementary investments in human capital and efforts to improve capital accumulation by the extreme poor may be crucial to programme success. Training can teach households the skills needed to profitably manage and grow the provided assets; similarly, savings can be used to diversify and expand small businesses. Researchers suspected that these complementarities might drive the magnitude of impacts produced by the graduation approach.

Untangling the effects of complementary investments, Abhijit Banerjee (MIT) presented results from a three-pronged Ghanaian study (with co-authors, Karlan, Osei, Thuysbaert, and Udry). Households in Northern Ghana were randomly selected to receive one of three intervention models: the full model (assets and skills), a savings-only model, and an asset-only model. Half the households receiving the full model were also randomly selected to participate in the savings programme.

The follow-up of the study in Ghana found largely similar and positive effects from the full model. The savings-only model produced positive impacts on households’ food security, consumption, financial inclusion and improvements in their ability to weather shocks. The asset-only model (unsurprisingly) only had positive impacts on asset and livestock ownership; however, impacts on assets disappeared after the programme ended.

For graduation approaches to be effective, complementary investments in human capital and efforts to improve capital accumulation by the extreme poor may be crucial to programme success.

Results from these three interventions suggest that the effectiveness of the programme stems, at least in part, from the combination of transfers. Furthermore, incorporating components that can encourage households to save may produce even greater returns.

#3 Programme benefits can outweigh costs, even despite the high price tag

Providing ultra-poor households with large-scale assets, support and regular training for two years, does not come cheap. Implementation and staff costs are, as expected, higher than those of traditional cash transfer programmes. But do the benefits justify the costs?

The accuracy of any cost-benefit analysis (CBA) depends on the quality of assumptions driving it. We first need to estimate how long the benefits will persist. Evaluating returns to the Bangladesh programme, Oriana Bandiera (LSE) presented two CBA scenarios. The first scenario assumes programme benefits last forever. In this scenario, each dollar spent generates 5.7 dollars of benefits. The second scenario assumes programme benefits evaporate after four years. Under this scenario, each dollar spent generates only 80 cents worth of benefits.

Varying CBA assumptions and calculating returns from other studies produces similarly large benefit-to-cost ratios, reinforcing the Bangladesh results.

#4 Long-term results are both positive and sustainable

The graduation approach and others like it are attempting to unlock the extreme poverty trap through a cost-effective intervention. Although the model is not a magic bullet, results are highly promising. A follow-up to the Bangladesh programme, conducted seven years after implementation, suggests that gains and benefits to households are sustained, if not greater after seven years, than after four. Similarly positive and sustained outcomes from West Bengal and other contexts suggest that, at its core, the graduation approach has identified universal barriers (access to capital and skills) that prevent households from escaping extreme poverty.

The figures below based on the 7-year follow-up in Bangladesh, illustrate that while programme beneficiaries do not rapidly over-take their middle and upper class neighbours, they do begin to close the gap significantly over time. Ultra-poor households begin saving at higher rates, and increased capital accumulation can finance the purchase of additional productive assets


A follow-up to the Bangladesh programme, conducted seven years after implementation, suggests that gains and benefits to households are sustained, if not greater after seven years

In many of the discussions that followed the presentations of results, the importance of evidence in shaping policy design was increasingly apparent. Given the costs of implementing large-scale social protection programmes, particularly for countries with large populations of ultra-poor households, careful consideration must be given to both feasibility and costs of implementing programmes. Though costs remain a concern, results from South Asia and Africa serve as significant proof of concept that the graduation approach can significantly reduce extreme poverty.

Conference takeaways

Dr. Muhammad Musa, executive director of BRAC, closed the conference with a summary of his own takeaways from the findings and best practices presented by researchers and practitioners.

  • The graduation approach cannot work everywhere in the same way, but it can be effective if adapted and integrated to local contexts. Mobilising local resources and support is important for programme success.
  • The search for an intervention with lasting impact on extreme poverty is an on-going process of learning. Rigorous impact evaluations have allowed graduation programmes to evolve, scale up, and expand to other countries.
  • Ensuring cost effectiveness will continue to be a challenge in the scaling up of graduation programmes, but further research is helping BRAC and other organisations adapt their models and streamline costs.
  • ‘Graduation’ is not just a transition from extreme poverty to a higher level of income, it is a transformational process. Giving ultra-poor women the ability to change occupations – from working as agricultural labourers and maids to rearing their own livestock – is an often life-changing shift in identity.

Finally, on the conference theme of ‘tackling extreme poverty’, Dr. Musa was optimistic in his outlook that ending extreme poverty is possible in our lifetimes.