72455 from EVIDENCE to POLICY Learning what works, from the Human Development Network September, 2012 BEFORE CRISIS HITS: Can Public Works Programs Increase Food Security? Fighting famine is basic to ending poverty and saving lives. At the World Bank, we’re committed to creating new op- Emergency aid, which arrives after the food has run out, isn’t portunities for people to end poverty and improve lives, helping enough. Households most in need of emergency aid often don’t countries meet the United Nations Millennium Development have enough food during other times of the year, posing a broad- Goals. Social protection is a crucial part of success and we are ac- er challenge for devising tively engaged in working with countries to develop, implement programs that can cut and analyze programs that work. To understand the role these hunger and build food programs can play in helping people feed themselves and their security. Social protec- families, the World Bank supported an evaluation of an ongoing tion programs, including program in Ethiopia. This public works employment program, grants, social assistance designed to ensure the poorest households have enough to eat and public works pro- year-round, and to avert the conditions leading to famine in grams, are one way to times of drought, is supplemented by cash grants for those who social protection transform people’s lives can’t work. A complementary initiative works to build household and protect them both agricultural assets so families can better provide for themselves. before and when disas- The evaluation found that these measures boost food security, ter strikes. What works and under what circumstances is what helping households better manage risk year-round. Now, Ethio- policymakers and development experts want to know, especially pia, with the assistance of donors, is expanding the program those focused on famine breakouts in Africa and Asia. to reach even more people. Context For many, the word famine was defined by Ethiopia in 1984-1985, Ethiopian government partnered with donors and NGOs to create when hundreds of thousands of people died from lack of food a working coalition to improve food security for the poor. The result brought on by a severe drought. Emergency aid saved countless lives, was the Productive Safety Net Programme (PSNP), which went into but it didn’t end the cycle. Over the next two decades, extreme pov- effect in 2005. This program, the largest of its kind in Africa, initially erty and harsh weather conditions, both floods and droughts, forced targeted 7.6 million people (8 percent of Ethiopia’s population) who Ethiopian leaders to make regular appeals for emergency aid to feed suffered chronic food shortages and lived in areas prone to drought. people and in extreme cases, stop them from starving. Through a public works component and direct grants for those who Donor fatigue, coupled with the Ethiopian government’s desire can’t work, the program aims to help households meet their food to better assist its people, pushed both to consider a new approach needs, keeping people fed and reducing the need to sell off produc- to stopping hunger and relieving extreme poverty. In 2003, the tive assets. People in the program are employed in public works projects 6 What does a food gap mean? When there isn’t enough food, house- months out of the year, corresponding to the so-called “lean� period, holds cut down on the number of meals—on average, adults will get when households tend to run out of food because it’s a few months two meals and children three—and serve less desirable food. When after harvest season and planting is just beginning. The total pay- food supplies are sufficient, households average three meals a day for ments vary depending on household size, but the median payment adults and three to four for children. is about 186 birr (about $10) for the first year for a household, to a total of 3,370 birr (about $200) for the whole five years. Payments limited resources and lack of clarity over who was targeted, it was are made in cash, food, or a mix of the two. Households without extensively revised in 2008. The renamed Household Assets Build- workers (such as those headed by the disabled or elderly) qualify for ing Program, put into effect in 2009, includes more demand-led ac- direct grants. tivities in agriculture and a focus on beneficiaries of the public works A complementary government program, initially called the program, among other changes. Ultimately, through the public Other Food Security Program, provided access to agricultural credit works program, the assets building program and other government and related services to help people build up assets and improve agri- efforts, the goal is to build up people’s resources so that they can culture work. Because of implementation difficulties, partly due to manage without aid. Evaluation Three surveys, in 2006, 2008 and 2010, were carried out propensity score matching methods to evaluate the pro- covering beneficiary and non-beneficiary households in the gram’s impact, taking into account duration of participa- four major regions of Tigray, Amhara, Oromiya, and the tion. Researchers thus were able to account for selection social protection Southern Nations, Nationalities and People’s Region. The biases in levels of program participation and measure the same 3,700 households were surveyed each time, allowing impact of the program by comparing people who had re- researchers to see how their circumstances changed during ceived benefits for five years versus those who had received the period. The program was not randomly allocated and the benefits for only one year (benefits were minimal in the recipients had different characteristics when compared to first year). Researchers evaluated the impact of the public non-recipients (who didn’t qualify for the program because works program and the assets-building program on food they were not as food insecure), making it difficult to com- security. pare the two groups. Instead, researchers used generalized Findings People enrolled in the program are better able to reported living through two or more droughts during the meet their food needs. survey period. In these areas, households that had received the public works benefit for five years saw a 25 percent At the start, households reported a food “gap� of 3.6 decline in the periods in which they went without enough months, during which they couldn’t meet the food needs food, compared with households that received the program of their families. After five years in the program, the food for only one year. Households living in areas less affected “gap� dropped by just over one month, reflecting a 29 by drought reported a 42 percent decline in the time in percent decline in the time that households couldn’t meet which they couldn’t meet their food needs. their food needs. This translated into an increase in food And during “lean� periods, households in the program security of 32 days a year. for five years were better able to feed their children. The The gains were even more apparent when researchers number of meals children ate during the lean season in- focused on households in more drought-prone regions, creased by 5 percent, compared with households in the defined as areas where at least 50 percent of households program for only one year. This policy note summarizes the report, “The Impact of Ethiopia’s Productive Safety Nets and Household Asset Building Programme: 2006-2010,� by Guush Berhane, John Hoddinott, Neha Kumar, Alemayahu Seyoum Taffesse (International Food Policy Research Institute, Dec. 30, 2011) Households in the program also experienced terraced fields, laying down roads and setting up small-scale a slow but steady rise in livestock and in the irrigation systems.* value of their productive assets—in other words, they accumulated more (or better) tools, such as The biggest gain in food security was among hoes, sickles and ploughs, the very items that households that were in the public works help families work the land. And there was a program and had access to the program decline in the percentage of households forced to designed to increase household assets and im- sell assets to maintain their families. prove agricultural output. The program led to a boost in livestock holdings, such as Households that were in the public works program for sheep, goats and cattle. In areas where droughts were wide- five years and also benefited from the program to increase spread, households who participated for the whole five-year agricultural assets and agricultural output improved their period had an 11 percent increase in livestock holdings, com- food security by 1.53 months during a 12-month period, pared with those in the program for just one year. compared with households that received one year of pub- Beneficiaries also saw a slow and steady increase in the lic works employment alone. When compared with house- value of their productive assets over the period, compared with those outside the program, who saw the value of their assets fluctuate markedly during the period. And in a sign of increased stability, distress sales dropped. Shortly after the program began, 51 percent of beneficiary households re- ported distress sales of assets; by 2010, this had dropped to 34 percent. Among households not in the program (and can thus assume to have been better off to start with), the rate of distress sales dropped from 44 percent in 2006 to 28 percent in 2010. Fears that the transfers through the public works program would “crowd out� informal transfers among households, or discourage people from new income-generating activities, proved un- founded. In other words, there was no sign that the program created dependency. There was no reduction in informal transfers among house- holds, which usually refers to gifts and loans. In fact, although not statistically significant, there was a small gain in cash and in-kind transfers among households. Similarly, there was no sign that the public works programs reduced the likelihood of people starting up non-farm businesses. holds that were in the public works program for five years, The public works program doesn’t just employ those who had access to the public works and also the asset people, it has improved the country’s building program for five years raised their food security infrastructure. by an additional 17 days. Households that received both programs had higher Participants get involved in a variety of projects designed to output of grain—147 kilos more a year than households help improve agricultural output and arable land. More than that only qualified for the public works program. Likewise, 40,000 projects are undertaken each year, such as creating they were 21 percentage points more likely to use fertilizer *For further information on the public works program, see M.A Consulting Group. Impact Assessment Study Report for PW Component of the PSNP in Ethiopia. Addis Ababa: Government of Ethiopia, 2009 than households without access to either program; and house- The evaluation shows that a broad social safety holds in both programs were 19 percentage points more likely net can help people increase their food security to use fertilizer than those who only were in the public works and build their assets. But it takes time program. Households in both were also 13 percentage points more likely to invest in stone terracing, which improves pro- Many programs to improve people’s livelihoods and build ductivity by conserving topsoil. And these households were food security last only a year or two. Ethiopia’s program, 16.4 percentage points more likely to invest in fencing as well. by contrast, is now in its fifth year. It’s also been tweaked social protection The results underscore the complementary aspects of the and rejigged as needed, and attention has been paid to en- two programs: the asset building program offers the techni- suring people receive regular, predictable transfers. The at- cal assistance to support better agricultural methods, among tention to detail, and the commitment to continuing the other things, while the public works program helps house- program, are critical aspects of this successful social protec- holds maintain food consumption, freeing up other resources tion program. to improve both agricultural output and financial stability. Conclusion Ethiopian policymakers and international donors have long- ter prepare for and survive difficult weather conditions such as struggled with the challenge of reducing poverty amid weather droughts and floods. The program to build agricultural assets shocks that disrupt harvests and threaten households with helps households gain economic independence and stability. starvation. After years of emergency aid programs designed to Together, the two make up an innovative social protection pro- provide short-term relief, both Ethiopia and donors wanted to gram that has already shown strong success. create a program that could help people secure and build their Ethiopia, with the assistance of donors, is now expanding lives, rather than just react to disaster. the program to reach an additional 600,000 people by 2015. The result is Ethiopia’s Productive Safety Net Programme, With few large-scale social protection programs in place in Af- which uses public works employment, social transfers and an rica, Ethiopia’s experiences is a useful example for other coun- agricultural asset-building program, to stabilize and strengthen tries dealing with extreme poverty, with or without the even poor households. Giving households a more steady income, more destabilizing weather shocks. Social protection is only through a public works program for those who can work one element in the fight against poverty, but it’s a critical one. and cash transfers for those who can’t work, helps them bet- The Human Development Network, part of the World Bank Group, supports and disseminates research evaluating the impact of development projects to help alleviate poverty. The goal is to collect and build empirical evidence that can help governments and development organizations design and implement the most appropriate and effective policies for better educational, health and job opportunities for people in developing countries. For more information about who we are and what we do, go to: http://www.worldbank.org/sief. This Evidence to Policy note series is produced with the generous support of SIEF. THE WORLD BANK, HUMAN DEVELOPMENT NETWORK 1818 H STREET, NW WASHINGTON, DC 20433 Produced by Office of the Chief Economist, Human Development Network, Communications/Aliza Marcus amarcus@worldbank.org