Will a scientific trial show the value of cash donations?

A charity is awaiting a verdict on its practice of sending money directly to poor Kenyans, reports Kath Nightingale.
 
One way for aid agencies to help those in hardship is to see that they get livestock or textbooks; in effect, to decide what people need and then buy it for them. But now this traditional approach has a competitor: some charities are giving aid directly to the beneficiaries and letting them spend it however they want. Yet there are doubts about whether this newer approach is as effective as the old one. To help find out, one charity is putting the new ideology to the test, scientifically.

The charity in question, the US-based GiveDirectly, already gives unconditional cash payments — typically of US$1,000 in a few instalments — straight to poor people in Kenya using the M-Pesa system for transferring cash via mobile phone. In Kenya, where M-Pesa originated, that sum is roughly equivalent to a family’s yearly income.

The idea of providing aid directly is gaining traction. Last December, Google handed GiveDirectly a US$2.4 million impact award via its Google Giving arm. This grant, for “innovative”, “game changing” ideas according to Google’s website, is designed to help the charity scale up its work and expand it to a second country.

In addition, charity evaluator GiveWell ranks GiveDirectly second in its league table of international charities, noting its transparency and commitment to self-evaluation as particular strengths.

But the development community is unaccustomed to distributing cash in this way — and some see the charity’s approach as controversial. There has been a long-held belief in the development sector that cash is at best a temporary fix, and at worst people would simply spend it on weddings or luxuries such as cigarettes and alcohol.

A long history          
 

“People ask me if I’m a nihilist who wants to tear down the development sector and replace it with mobile cash transfers,” says Paul Niehaus, a director and co-founder of GiveDirectly and assistant professor of economics at the University of California, San Diego, United States.

Niehaus thinks such opinions are unjustified. “We’re generating all kinds of controversy and argument by doing something that governments have been doing on a fairly significant scale for a long time,” he says.

“If you’re providing in kind, you are imposing a particular type of consumption, whereas with cash you are letting households determine for themselves what is the best use for the money.”

Armando Barrientos, University of Manchester 

Others share his opinion. Armando Barrientos, professor in poverty and social justice at the University of Manchester in the United Kingdom, notes that social assistance from governments via small regular payments to families has a long history in developed countries.

A 2011 report by the UK Department for International Development describes cash transfers as “one of the more thoroughly researched forms of development intervention” and says there is “convincing evidence” that cash transfers can reduce inequality and the depth or severity of poverty. [1]

What is unusual, says Barrientos, is simply that the organisation doing the direct giving is a charity. Previously, it has only been practised by international charities during humanitarian emergencies.

There are also ethical reasons for giving people cash rather than goods, according to Barrientos. “If you’re providing in kind, you are imposing a particular type of consumption, whereas with cash you are letting households determine for themselves what is the best use for the money,” he says.

Direct giving might also be economically favourable. It reduces charities’ costs and so could enable them to give a higher proportion of donations to those in need.

Aid on trial

Yet Barrientos is unsure whether direct giving can work in the long term. “Most social assistance is guaranteed for a period of time,” he says. “If you give a one-off payment, what happens when the money runs out?”

In an effort to allay such fears, GiveDirectly has commissioned US not-for-profit organisation Innovations for Poverty Action to run a randomised controlled trial of its cash transfers in a district in western Kenya.

Such trials are the tried-and-tested ‘gold standard’ test of interventions in both social and medical research. In theory, they eliminate any bias on the part of researchers in selecting who receives the intervention, and provide direct comparisons between the ‘haves and have-nots’.

Niehaus says he initially doubted the need to run the trial because he felt direct transfers already had a strong track record. But in the end, he says, the charity decided to run it partly to add rigour to the development sector.

“Most people don’t do rigorous evaluation or bring an external person to look at what they’re doing,” he says. “We are very comfortable with doing that.”

To conduct the trial, the research team randomly assigned families in a village to either receive cash or not. Before the start of the trial, and then after six months and again after a year, the 1,000 households involved in the study were surveyed to assess various aspects of their lives: their levels of hunger and the type of food they ate; their physical and emotional health; whether their children were in school; their household spending on non-food items; and their incomes and investments. Their saliva was also tested for the presence of cortisol, a hormone that indicates stress levels.

Niehaus hopes the results will be published by early November.

Nutrition impacts

In the meantime, preliminary results published in the charity’s 2012 annual report suggest that the biggest impacts were on nutrition and investment in assets such as houses and livestock. [2] Crucially, spending on activities such as ceremonies, weddings and funerals, and recreational activities, including smoking and gambling, only increased in proportion with other spending. Niehaus says this shows that, given the freedom to choose how to spend money, people in Kenya make prudent choices.

Yet Barrientos questions whether a trial in which some people randomly receive money while others miss out is unethical.

“There are very strong ethical reasons for not randomising anti-poverty programmes. If you say: ‘For the purposes of attaining knowledge, we’re going to support only half the people’, that’s really problematic. That exercise would not get through my university’s ethics committee,” he says.

Niehaus says GiveDirectly is considering assisting those families that did not receive transfers the first time around, and that any future trials will randomise at the village level, rather than within villages.

So will the outcome of the trial show the intervention to be effective? Barrientos suspects that one isolated strategy can never be entirely effective on its own: a mix of interventions will always be needed.

But if the trial on GiveDirectly’s work demonstrates the effectiveness of directly providing cash, such views may no longer be justifiable.