An anti-poverty program tested extensively on three continents has produced sustained gains in individuals' income, wealth, and well-being, according to a study published today in the journal Science.
The program provides very poor people with productive assets, such as livestock, as well as job training, life-skills coaching, and health information. Known as the "Graduation" program, its intention was to examine whether helping the poor in multiple ways simultaneously could be especially effective in fighting poverty.
Overall, with more than 20,000 people enrolled across six countries over a three-year period, the experiment produced a 5 percent increase in per capita income, an 8 percent increase in food consumption, a 15 percent increase in assets, and a 96 percent increase in savings, compared with similar groups of people not enrolled in the program.
"The results show that three years after the intervention, hunger is down, consumption is up, and income is up," says Abhijit Banerjee, the Ford Professor of International Economics at MIT, and a co-author of the paper detailing the findings.
The "Graduation" program was targeted at substantial groups of very poor citizens in Ethiopia, Ghana, Honduras, India, Pakistan, and Peru; about 48 percent of households in the experiment had daily per capita consumption of less than $1.25. While the welfare of recipients was expected to increase in the short run, those gains proved durable.
"It seems to be an improvement that happens and stays intact," Banerjee says, noting that the self-reported mental health of participants improved as well: "They are happier, too."
Providing a "big push"
There are nine co-authors of the Science paper, including Banerjee and Esther Duflo, co-founders of MIT's Abdul Latif Jameel Poverty Action Lab (J-PAL). The paper's corresponding author is Dean Karlan, a development economist at Yale University and a member of J-PAL's executive committee.
J-PAL was founded in 2003 and named in 2005 after the father of MIT alumnus Mohammed Abdul Latif Jameel '78, whose substantial gift to the lab greatly expanded the scope of its activities.
The study examined results from 21,063 adults in 10,495 rural households. The "Graduation" program concept was first used deployed in Bangladesh by a large non-governmental organization known as BRAC. As the paper notes, the idea was to provide the poor with a "big push" intended to alleviate poverty.
"The intellectual impetus came from asking the question: Can we have durable consequences from a one-time intervention into the lives of the poor?" Banerjee says.
The "Graduation" program gave participants a one-time asset transfer, often providing people with animals, such as cows or chickens, from which they could earn income. It supplemented that asset donation with temporary spending support; training on running a business; frequent home visits from project staff; and information about health care. Participants were also encouraged to save money.
Some specifics were tailored to each country, but in each case, participants' results were compared to those of people with similar income levels who did not take part in the project. The wealth gains were observed in five of the six countries; only Honduras was an exception.
"The overall bottom line is that the program appears to be effective in most places," the authors write in the Science paper.
"A battle that could be won"
The "Graduation" project has also produced a number of questions the researchers hope to explore. Some pertain to costs; the program is intensive, with home visits and personalized coaching. Future variations, Banerjee notes, may modify some elements of the program, to see if the same outcomes can be achieved at lower cost.
For now, Banerjee thinks the experiment has demonstrated that the world's poorest people are not inherently incapable of improving their own lives.
"We wanted to show it's a battle that could be won," Banerjee says.
###
The other co-authors of the Science paper are Nathaniel Goldberg of Innovations for Poverty Action; Robert Osei of the University of Ghana-Legon; William Pariente of the Catholic University of Louvain in Belgium; Jeremy Shapiro of Princeton University; Bram Thuysbaert of the University of Ghent; and Christopher Udry of Yale University.
Funding for the project was provided by the Ford Foundation, 3ie, and the U.S. Agency for International Development.
Journal
Science