Millions of poor people around the world are enrolled in safety net programs that hand out cash or food. What’s the best way to design these transfer programs?
A Bangladeshi woman receives emergency monetary assistance organized by the World Food Programme in the wake of flooding.
See sidebar: Spreading the Wealth
See sidebar: What Happens When You Give People a Windfall
The world’s largest food security assistance program—the Public Distribution System—provides subsidized rice and wheat to some 70 percent of India’s population. The country is in the midst of a vigorous debate over whether this system should be overhauled to provide poor people with cash instead of food.
Not only is food expensive to store and distribute, but, one argument goes, much of it is lost to corrupt officials and undeserving citizens. On the other hand, food transfers can encourage people to consume certain goods or to induce less needy people to self-select out of programs targeting the needy. Cash transfers are also susceptible to corruption. And poor families may spend the cash on things that don’t improve their well-being or that don’t help the most vulnerable.
The debate over how best to transfer resources to poor people has been simmering for decades. The question of food versus cash comes up in settings that need to respond to chronic food insecurity, such as Ethiopia and Niger, as well as in countries that often need to respond to sudden shocks, such as Bangladesh and the Philippines. Similar discussions are unfolding in Middle Eastern countries that are in the midst of reforming subsidies.
So what does the evidence say about which form of transfer—cash or food or something else—improves people’s well-being more and is more cost-effective? In fact, until recently, few studies sought to answer this question. Typically, says Paul Niehaus, assistant professor of economics at University of California, San Diego, researchers compare a “treatment” group of people who receive a transfer with a “control” group of people who receive nothing. According to Niehaus, “If we get some positive effect, we say, ‘Great, it works,’ and we call it a day. But we don’t often get as far as testing two different uses of funds head-to-head in the same setting and asking which does more per dollar.”
Recently, though, IFPRI researchers have conducted a number of studies not just comparing people who receive cash with people who don’t, but looking at what happens when people receive cash, food, or vouchers for food. Other research has examined the range of impacts of requiring recipients to meet certain conditions before receiving safety net transfers. The new findings provide many answers, but also raise further questions.
Weighing the Alternatives
The World Food Programme (WFP) is in the business of giving aid to people at risk of hunger, so the best way to design these transfers is a central concern. Starting in 2009, WFP and IFPRI developed a series of randomized control trials to find out whether transferring cash or food best improves households’ food security in Ecuador, Niger, Uganda, and Yemen. In Ecuador, researchers also compared cash and food with vouchers redeemable for certain foods. Communities in these pilot projects were randomly assigned to receive either cash, food, or vouchers. Within each country, researchers attempted to standardize the value, number, and frequency of transfers households received. A large team of IFPRI researchers—made up of Daniel Gilligan, Melissa Hidrobo, John Hoddinott, Amy Margolies, Vanessa Moreira, Amber Peterman, Shalini Roy, Susanna Sandström, and Joanna Upton—conducted the work and produced a report.
In three of the four countries, says Gilligan, deputy director of IFPRI’s Poverty, Health, and Nutrition Division, cash did a better job of improving household food security because it allowed people to diversify their diets. Food rations aren’t necessarily what people would buy if given the cash, he says, and many programs that provide in-kind rations hand out staples and salt or oil but rarely vegetables. “You can see that an equivalent cash transfer might allow people to diversify their diets much more,” he said.
But this result did not hold everywhere. In northern Ecuador, researchers found that people who received cash, food, and vouchers all consumed more and better-quality food. But people who received food transfers or vouchers had a bigger jump in calories than those in the control or cash groups, whereas people who got vouchers showed larger increases in the diversity of their diets, consuming more vegetables, eggs, and milk and dairy.
Researchers also compared the impact of food and cash transfers in Niger on people’s food consumption scores—a measure of food security that considers dietary diversity in terms of the frequency of consumption and nutrition quality of each food group. There, recipients of food transfers had larger increases in their food consumption scores than did cash recipients. People who received cash used much of the money to buy bulk grains, the cheapest calories available. This choice reflected not only the extreme poverty of the area, where markets for nongrains were limited or absent, but also uncertainty about future food prices.
In Niger, people who received cash transfers tended to buy bulk grains, the cheapest source of calories available.
Accounting for Costs
When the WFP commissioned IFPRI researchers to evaluate the pilot projects in four countries, it gave them an unusual degree of access to detailed financial records so they could compare the relative costs of different programs. The researchers were thus able to study not only which kinds of transfers made the most difference to recipients, but also how cost-effective each type was.
When stacked up against other kinds of transfers, food does not always perform well. The additional cost of transferring food instead of cash is, for example, US$8.47 in Ecuador and US$7.38 in Niger. The lower cost of providing a cash transfer means more households can receive the benefit (see the sidebar “Spreading the Wealth”). In IFPRI’s four-country study, if the six-month pilot programs had transferred only cash and no food, another 44,769 people could have benefited at no additional cost.
Chris Blattman, associate professor of political science and international and public affairs at Columbia University, has underlined the need to compare the costs of delivering programs, whether they give training, capital goods, supervision, cash, or something else. “If you halved the cost of program delivery, you could help twice as many people,” he says. “Moving to cash is one way to do this.” But simply transferring cash is no guarantee of cost-effectiveness. Even when programs do give cash, he says, “they do it in a very costly way, because they are not accountable for efficiency.”
After evaluating the WFP pilot projects, IFPRI researchers concluded that there is no one “right” kind of transfer. The relative effectiveness of different types of transfer depends heavily on context. How severe is food insecurity? How well do markets for grains and other foods function in that location? What are the ultimate objectives of the transfer? These are the kinds of questions that policymakers need to answer before launching a transfer program.
In Ecuador, for example, the researchers who drilled down into the impact estimates and cost data found that in an urban setting with well-functioning food markets, vouchers are usually the most cost-effective transfer for improving any outcome measure. Least cost-effective? Food.
Sidebar
Spreading the Wealth
Which social transfer policymakers choose can affect the costs of a program and by extension, the number of people who benefit. In a study of pilot programs in four countries commissioned by the WFP, IFPRI researchers compared the costs of delivering food and cash transfers (and, in Ecuador, vouchers). A monthly transfer—whether food or cash—in Niger was worth US$55.00. The cost of delivering that transfer ranged from $2.89 for cash to $10.27 for food. For cash, costs relate to contract preparation, debit cards, and bank fees. For food, costs are associated with staff, in-country transport, ration preparation, and distribution. Because the costs associated with a food transfer are three times that of cash, a $2 million budget stretches further with cash transfers. You could serve another 651 households, or the equivalent of 4,427 more people.
IFPRI’s studies for WFP provided evidence that helped the organization be more confident in proposing cash and vouchers to its donors, according to
Annalisa Conte, deputy director of the Policy, Programme, and Innovation Division at WFP. Between 2008 when WFP introduced cash and voucher transfers and 2014, its spending on these types of transfer rose from about $10 million to $1.5 billion.Implications of cash versus food transfers in Niger
Cash transfer
Food transfer
Value of each monthly transfer
US$55.00
US$55.00
Cost of delivering the transfer
$2.89
$10.27
Total cost of one transfer
$57.89
$65.27
Number of households served by a US$2 million budget
5,758
5,107
Source: A. Margolies and J. Hoddinott, “Costing Alternative Transfer Modalities,” Journal of Development Effectiveness 7, no. 1 (2015): 1–16.
In Kenya, large unconditional cash transfers led some recipients to purchase assets like sewing machines.
With Strings Attached?
Although some governments and nongovernmental organizations dispense cash or food to poor people with no strings attached (see the sidebar "What Happens When You Give People a Windfall"), many governments require recipients to fulfill a condition, usually related to their children’s health, nutrition, or school attendance. Mexico was a pioneer of this approach, and its conditional cash transfer program—now called Prospera—has reached more than 24 million people. More than 50 countries have replicated the program in Africa, Asia, and Latin America.
Tying a condition to a cash handout is one way governments get more for their money by inducing behavior change. Poor people are often so focused on meeting today’s needs that they tend not to save or invest in the future. Conditions can help overcome that tendency by helping to build up people’s capacities and health and alleviating poverty in the long run. Moreover, conditions can make transfer programs more politically palatable to taxpayers because it doesn’t look like the poor are getting something for nothing.
At one time, conditional cash transfers were called “as close as you can come to a magic bullet in development.” But reality is more nuanced. While conditional transfers have led to sometimes substantial jumps in the use of health and education services, they do not necessarily affect final outcomes. For example, conditional cash transfers raise the likelihood that households will take their children for preventive health checkups, but that does not always boost nutritional status.
What’s more, while conditions often achieve specific program goals, they can have unintended consequences. In a study of Brazil’s Bolsa Família, IFPRI researchers noted that a program that determined eligibility based on income earned in the formal sector created an unexpected effect. Some people switched hours of work from the formal to the informal sector, possibly because they feared losing eligibility for the program.
Another example of an unintended consequence comes from one of the first studies to randomize access to conditional and unconditional cash transfers. The study took place in Zomba, Malawi, where cash transfers were targeted to never-married adolescent girls. One group received cash transfers only if they attended school, another group got the cash unconditionally, and a control group received no cash. As expected, school-age adolescent girls who got the conditional transfer tended to stay in school longer than did the girls who got the unconditional transfer. But researchers were surprised to find that girls who received the unconditional transfer had significantly lower teenage pregnancy and marriage rates compared with girls in both the conditional group and the control group.
These reductions were entirely due to what happened when girls dropped out of school during the intervention. When girls whose households received the unconditional transfer left school, the transfer didn’t go away. “They still had these protective resources, which seem to have allowed them to make different decisions around marriage and pregnancy,” says Sarah Baird, coauthor of the study, assistant professor of global health and economics at The George Washington University, and an IFPRI consultant. In contrast, once the girls in the conditional group dropped out of school, the program no longer exerted any protective effect.
“One of the main takeaways was that conditional cash transfers are great, but one of the potential costs is that they may leave a very vulnerable group out of the program,” said Baird, “because if you don’t meet the condition, you don’t get the cash.” Put another way, conditions can exclude the very people the transfers are designed to help.
A field-worker in Uganda administers an early childhood development assessment. Transfers were distributed only to households in which children attended early childhood development programs.
The Next Wave of Safety Nets
No matter how effective proponents find cash transfers in reducing poverty, they can’t do the whole job on their own. “In order for people to escape poverty, lots of things need to happen in the right way,” says John Hoddinott, the H.E. Babcock Professor of Food and Nutrition Economics and Policy at Cornell University and a former senior research fellow at IFPRI.
Cash transfers in many countries allow people to send their kids to school. But if the schools are of poor quality, the children may not learn much and the long-term benefits of schooling may be minimal. Similarly, there is plenty of evidence that cash transfers allow people to buy more food. But for cash transfers to help people diversify their diets, Hoddinott says, the local markets must carry a wide variety of food.
In the next stage of her research, says Baird, to really compare conditional transfers, unconditional transfers, and noncash transfers, it will be important to evaluate interventions at a larger scale and to replicate interventions across countries to figure out which approach outperforms the other and why. “The answer is not going to be the same across countries,” she says. “We’re not going to say unconditional cash transfers or conditional cash transfers work best everywhere. We’ll say unconditional cash transfers work best in countries with these kinds of institutions that are trying to tackle this type of outcome.”
In the next round of program design, policymakers are already considering what types of services to add on top of cash transfers to help people not just get by, but transform their lives, says Berk Özler, a senior economist at the World Bank who cowrote the Malawi study and also studies cash transfer design. He envisions a world where the ultra poor will have a menu of programs and services that will help them climb their way out of poverty, graduating from one program to another, until they eventually reach a stage where they no longer need government assistance. “That,” he said, “is the next wave.”
For more information on this topic:
- John Hoddinott, Daniel Gilligan, et al., "Enhancing WFP’s Capacity and Experience to Design, Implement, Monitor, and Evaluate Vouchers and Cash Transfer Programmes: Study Summary," IFPRI report, June 2013.
- Sarah Baird, Craig McIntosh, and Berk Özler, “Cash or Condition? Evidence from a Cash Transfer Experiment,” Quarterly Journal of Economics 126, no. 4 (November 2011): 1-44.
- Johannes Haushofer and Jeremy Shapiro, “Household Response to Income Changes: Evidence from an Unconditional Cash Transfer Program in Kenya,” November 2013.
Sidebar
What Happens When You Give People a Windfall
Should poor people have to meet any conditions to get a cash or food transfer? A simple economic model would say that just giving people cash is the best way to improve their welfare. Let them have cash, and they’ll do what they think is in their best interest. “From a purist perspective regarding welfare, it’s hard to improve on that,” says Daniel Gilligan, deputy director of IFPRI’s Poverty, Health, and Nutrition Division.
But policymakers still harbor reservations about giving people handouts. Some worry that poor people will squander the money on vices like alcohol and cigarettes or that they will sit around doing nothing.
So what happens when you give the equivalent of one year’s household budget—no strings attached—to extremely poor people in rural Kenya? Since 2011 the nonprofit GiveDirectly has provided about US$1,000, unconditionally, to more than 12,000 households. The money was disbursed in multiple installments from a mobile money agent.
After buying a blanket, one person was finally able to sleep well. Another bought a secondhand motorbike to use for business. A grandmother paid the school fees for her orphaned grandchildren. “My life has generally changed as I now eat well also,” another recipient told a fieldworker. Some of the recipients invested in livestock and small businesses. Revenue from animal husbandry rose by 48 percent, and total revenue from self-employment climbed by 38 percent a month.
Many GiveDirectly recipients spent their large unconditional transfers on new tin roofs.
“This runs counter to the idea that cash transfers get wasted,” said Johannes Haushofer, assistant professor of psychology and public affairs at Princeton University, who co-wrote the study on GiveDirectly’s program.
Cash grants generally were not frittered away on a descent into dissipated living, according to the research. The GiveDirectly recipients spent their money on basic items like food, better shelter, healthcare, and education. With food consumption up by 20 percent, the likelihood a respondent went to bed hungry in the preceding week dropped by 30 percent, and the number of days children went without food fell by 42 percent.
“To the casual donor in the Western world, giving money to the poor and letting them do what they want with it is really challenging,” says Paul Niehaus, cofounder and president of GiveDirectly, as well as assistant professor of economics at the University of California, San Diego. “It flies in the face of received wisdom.”
Niehaus wasn’t surprised that people were able to take the money, invest it, and see their incomes rise. What surprised him was that the number-one investment people made was a roof upgrade.
Many Kenyans in the program replaced thatched roofs with a corrugated tin roof. The upgraded roofs were a wise investment because they were cheaper than buying new thatch for the roof every year, he says. It helps that the metal roofs last longer, do not require yearly upkeep, and do not leak. What’s more, rainwater from the roof can serve as a water source.
Among other benefits, the cash infusion led to documented improvements in psychological well-being. The researchers saw evidence of increases in happiness and decreases in depression. In some of the treatment groups, the researchers also saw decreases in cortisol, a major stress hormone.
“We should think about the impact of programs not just in terms of economic outcomes, but also in terms of welfare on the psychological level,” says Haushofer. Such indicators, he says, shed light on facets of well-being beyond commonly measured economic outcomes such as consumption, assets, and labor market outcomes.
One thing that makes GiveDirectly’s program stand out is the size of the transfer, says Gilligan: “They’re massive.” People tend to use one-time, large person-to-person transfers differently than they use periodic cash grants, he notes. “People’s behavior around those two modalities could be really different,” he says.
Findings on GiveDirectly reveal as much. Monthly transfers were more likely to increase food security, while lump-sum transfers led to larger purchases, such as metal roofs.
“Part of what I like about GiveDirectly is they’re playing with the formula,” says Gilligan. “And I think a lot of people should be doing that.”
More experimentation, he says, would help us figure out what works best, given what we’re trying to achieve.
