| Macroeconomics Term Paper | | |

[Social Safety Nets in Bangladesh] | This report looks at the overall structure of Social Security Net Plans in Bangladesh and their poverty impact at a macroeconomic level. The main objective of the study is to look into the overall impact of the existing safety net programs on poverty reduction in Bangladesh. The paper also looks into the impact of the existing safety net programs on reducing inequality. |

Social Safety Nets In Bangladesh
This report looks at the overall structure of Social Security Net Plans in Bangladesh and their poverty impact at a macroeconomic level. The main objective of the study is to look into the overall impact of the existing safety net programs on poverty reduction in Bangladesh. The paper also looks into the impact of the existing safety net programs on reducing inequality.

Institute of Business Administration, University of Dhaka

Table of Contents 1. Executive Summary i 2. Introduction 1 2.1. The Poverty Situation of Bangladesh 1 2.2. Social Safety Nets 1 3. Social Safety Nets of Bangladesh 2 3.1. Structure of Social Safety Nets 4 3.2. Trends in Social Safety Nets 7 4. Social Safety Nets and Poverty 9 4.1. Impact Assessment of Social Safety Nets on the Poverty of Bangladesh 9 4.2. Inequality Effect of Social Safety Nets 9 5. Conclusion 10 6. Recommendations 10 7. Bibliography 12

1. Executive Summary
Social safety net is a measure taken by the government in order to prevent the vulnerable section of its population to fall beyond a certain level of poverty. Social safety net programs (SSNPs) are designed with the aim to provide support for the vulnerable section of the society. With a vision to prevent transmission of poverty from generation to generation, the safety net programs opt for a more efficient society in terms of the choices made by individuals. The Social safety nets play both a redistributive and a productive role supporting moral philosophy as well as managing risks. These two are the major pillars that justify the existence of safety net programs. It should be mentioned at the outset that the safety net programs create a path towards poverty reduction in the long run. They do not reduce poverty directly; rather, these programs tend to reduce transitional poverty through ensuring proper nutritional intake, education, health care etc.
In other words, the safety net programs are methods through which poverty is expected to fall through investment in human capital.
With about 40 per cent of its population living below the poverty line and an increasing number of population being added below the lower poverty line, safety net programs in Bangladesh are more than a necessary element in fighting poverty. The government has been following a combination of direct and indirect safety net programs for poverty eradication addressing both human and income poverty. However, despite having a large number of programs under the safety net, the rate of poverty reduction has not been satisfactory. It is thus crucial to scrutinize the impact of the existing safety net programs on poverty reduction in Bangladesh and identify the kind of programs that would be more suitable to the socio-economic condition of the country.
This report looks at the overall structure of Social Security Net Plans in Bangladesh and their poverty impact. It does not go into detailed evaluation on which of the various types of safety net programs have been successful in Bangladesh. The paper discusses a number of SSNPs that have been considered to be successful in other countries of the world and highlights the current scenario of Bangladesh.
The main objective of the study is to look into the overall impact of the existing safety net programs on poverty reduction in Bangladesh. The paper also looks into the impact of the existing safety net programs on reducing inequality. In the final section, the study provides several policy suggestions on design, target and coverage of the existing safety net programs.

2. Introduction
Bangladesh had made some remarkable achievements in reducing poverty and other improving social and economic outcomes between 1991 and 2000 (The World Bank Group). During this period, poverty rates fell by 9%. Also, the percentage of population that completed grade five educations increased from 44% in 1991-92 to 56% in 2000 (The World Bank Group).
In spite of this, some 63 million people, which is roughly half of the country’s population, still live in severe deprivation, according to 2000 data (The World Bank Group). 55% of people over the age of seven couldn't read nor write.

3.1. The Poverty Situation of Bangladesh
While today, fewer people lived in acute poverty, inequality within and across region have progressively increased. The country's major challenge is to keep the momentum in some successful areas and expand these successes into other more challenging areas of social and human development.
It has also been noticed that the inequality is not only persistent within the urban and rural areas, but regional disparity also prevails strongly. Statistics clearly show a division between the eastern and western parts of the country. Rajshahi, Barisal and Khulna, some of the countries eastern divisions, appear to be high poverty-prone zones, while western divisions like Dhaka, Sylhet and Chittagong record fast poverty reduction. Hence, it can be concluded that poverty reduction accompanies increase in inequality.
The World Bank has been providing technical assistance to the Government for improving its poverty measurement and deepening understanding of bottlenecks for pro-poor growth and poverty alleviation. The Bank's work program entails: * Poverty assessment based on Household Income and Expenditure Surveys * Improving the data processing and analyzing facility in the Bureau of Statistics * Providing technical assistance for designing a safety net survey geared toward improving the efficacy of safety-net programs

3.2. Social Safety Nets
Social Safety Nets are non-contributory reassignment programs that seek to prevent the poor or those vulnerable to shocks and poverty from falling below a certain poverty level. Safety net programs can be provided by the public sector, that is the State and aid donors, or by the private sector, for example the NGOs, private firms, charities, and informal household transfers. Safety net transfers include: * Cash transfers * Public works * Conditional cash transfers * In-kind transfers such as school supplies and uniforms * Price subsidies for food, electricity, or public transport * Fee waivers and exemptions for health care, schooling and utilities * Food-based programs such as supplementary feeding programs and food stamps, vouchers, and coupons
On average, the spending on safety nets accounts for about 1 to 2 percent of GDP across developing and transition countries [ (Grosh M, 2008) ]. This proportion may be much less or much more at times and at different contexts. Over the last decade, a visible growing expertise in various areas of safety nets has been seen to take place. On the other hand, though an escalating number of safety net programs are extremely well thought out, correctly implemented, and demonstrably effective, numerous other safety net programs face, and often create, serious challenges. Often, government assistance has been criticized as it is said that the government ultimately deprives a person of an incentive to work.
Social Safety Net Programs should vary from one social context to another. That is to say, each area or location of use should adapt their programs to suit the social situation that exists in that area. Both the program mix and shape of individual programs should vary from place to place.
Safety nets in low-income countries, like Bangladesh, are increasingly being recognized as effective tools to reach out to the most vulnerable. At their worst, they protect households facing hard times from falling into deeper poverty and help them manage risk by allowing them to maintain assets on which their livelihoods are based. At their best, they can provide households with a cushion to invest resources more efficiently and effectively in human capital. Common interventions vary from public works and food-based interventions to more recently cash and conditional cash transfer programs. Low-income states may face institutional capacity and financial constraints. 3. Social Safety Nets of Bangladesh
The government of Bangladesh views poverty from two broad perspectives – income poverty and human poverty. While income and employment generating programs as well as direct transfers towards the poor are taken to deal with income poverty, the other type of programmers to help human development are there to mitigate human-poverty like education, health, nutrition and water and sanitation programmers.
The government identifies direct and indirect measures to address these two types of poverty. The direct measures are considered those that are targeted towards the poor while indirect measures are growth oriented and hence expected to leave indirect effects on poverty reduction. Safety nets are considered as direct measures and different safety net programmers are taken to address both income poverty and human poverty.
Examples of indirect or growth oriented measures cover mostly infrastructure development and rehabilitation programmers. However there are safety net programmers that merge the two concepts of direct and indirect measures. For an example, direct measures like Food for Work programmers that are targeted towards the poor are used to construct infrastructure services, falling in the category of indirect measures.
The safety net programmers of the country, if analyzed from the structure of the programmers for FY2008, addresses poverty mainly from the broad perspectives of education, health, vulnerability of some special groups, employment, special risk reduction, subsidy and micro-credit among the poor. Total safety net programmed grossly amounts to 12.08 per cent of the total public expenditure of the year, which is 1.8 per cent of the national GDP (Government of Bangladesh, 2008). However, the programmers that are listed as safety nets by the government might justify further scrutiny to be categorized either as safety nets or general public expenditure. Perceptibly, all the programmers are there with expectations to leave positive impacts on poverty reduction and might even assist the poor directly or indirectly to deal with the desolation of poverty. But as the government denotes itself, social safety nets programmers are generally devoted to the hardcore poor, the programmers listed as safety nets are sometimes subject to further scrutiny as some of them are generally poverty reduction oriented and cover both hardcore poor and general poor, even sometimes simply growth oriented. For example, donor assisted “Agriculture research fund” or “Training fund for RMG workers” might be subject to debate whether they are targeted towards the hardcore poor or not. The government, however, provides a different explanation of social protection in Bangladesh Economic Review with wider scopes for safety nets as it states “Social Safety Nets (SSN) are based on the government's policy to (a) reduce income uncertainty and variability; (b) maintain a minimum standard of living; and (c) redistribute income from the rich to the poor”
Nevertheless, there is a common understanding in the study of safety nets, particularly for a low income country like Bangladesh, that the social security programmers should emphasize the “very poor”, rather “poor” in general.
Three factors that generally constrain the feasibility of safety net programs are; (i) the availability of information for identifying potential recipients (ii) administrative capacity to deliver the services, and (iii) Fiscal affordability of the programmers needed.
Unfortunately for Bangladesh, all three are obligatory.

4.3. Structure of Social Safety Nets
Programs under the safety net of Bangladesh can be classified into different categories considering different criteria. Keeping in mind the nature and intensions of different programs, the study in this section categorizes the safety net programs of Bangladesh into some broad dimensions like education, health, vulnerability reduction, employment creation, risk reduction etc. However, a different categorization is also possible and programs may overlap between them. The overall structure of safety net evolves from the past and usually do not radically change in subsequent years. Each year safety net allocations are made in the national budget, mostly indicating carryover of programs from the previous year with few adjustments or inclusions. The following structure is based on the safety net programs as have been targeted in the national budget of FY2008.
A major portion of the support through safety nets is provided as food assistance, covered through different categories of programs like direct feeding programs, employment creation programs and others. Transfers in the form of food support constitute almost per cent of the total safety net spending. However, apart from this food support, there is also food subsidy included in the safety net programs of Bangladesh which is more or less evenly targeted among different income deciles. What is acting as a barrier for the non-poor to participate is the social status issue, as the non-poor groups are usually reluctant to stand in line to get the limited offer per person a day. While the fact is that free or subsidized food distribution has been a popular choice politically in the low income countries over cash, what is needed to be kept in mind is that it is recognized in the literature that untargeted food transfers/subsidies usually tend to prove fiscally unsustainable with gradual increase in cost. Free or subsidized food distribution also tends to distort markets, create dependency, and involve large inclusion errors and leakage to the non-poor. However, the justification for food assistance is that society as a whole considers it unacceptable for people to be living below the food poverty line owing to the threat of starvation [ (Subbarao, 2003) ] 4.4.1. Health
Unexpectedly, safety nets on health in Bangladesh constitutes only a little more than 2 per cent of the total protection. The health programs particularly focus on nutritional protection of women and children.
Based on World Bank study of 2005, the report said that the expansion of public and NGO health services together appears more generally to have been critical factors driving these gains.
3 programs are there in the safety nets to provide health support to the poor. Major program in health includes Health, Nutrition and Population Sector Program (2003 – 2010). The program is the modified version of Health and Population Sector Program (1998-2003). 2 out of the three 3 health programs under the safety nets are being implemented under this HNPSP program. 4.4.2. Education
As the GoB documents states, safety net programs relating education of the poor targets increasing the number of primary school enrolment, reducing drop outs as well as increasing female student enrolment at secondary schools. There are long term social development targets as well like reducing incidence of under age marriage by educating the poor female children. More than 11 per cent of the total safety net is dedicated to education programs.
The choice varies between countries regarding the mode of transfers (cash or food) for bringing the poor children to the school. In most of the low income countries, food is provided to the students either through school feeding or by providing packs of food grain to the families. Bangladesh have discontinued food assistance in education programs and gone for cash support as feeding programs usually suffer from large inclusion errors, because of the difficulties in feeding only the poor in a given class, as well as leakages. The safety net coverage includes special education programs for special vulnerable children like orphans or disabled. However, the share of this special support is only about 4 per cent of the total education program. Apart from the Mosque Based Child and Mass Literacy Program covering 30,000 centers from where the service is provided, the whole education program under safety net covers about 9 million children per annum. 4.4.3. Vulnerability Reduction
There is well justification in support of programs targeting the vulnerable groups like edged, widowed and others and there is also scope for targeted cash transfers within the specific groups. These groups are generally accepted, even demanded, by the society as deserving candidates of support with recognition of correlation between poverty and their vulnerability.
Safety net programs of Bangladesh give special consideration for the vulnerable groups within the poor. These groups include edged poor people, widowed or deserted women, disabled persons and others. Bangladesh safety net programs also provide support to the insolvent or wounded freedom fighters.
Without the Food Aid for Chittagong Hill Tracts program and the honorarium program for wounded freedom fighters, the total beneficiaries per annum exceeds 3.5 million through these programs, covering about 16.5 per cent of the total safety net in monetary terms. Other than the program for Chittagong hill tracts and the VGD program providing support in kind, all programs distribute cash among the beneficiaries.

4.4.4. Employment Creation
Measures that are seasonally targeted can potentially have a major welfare benefit for the poor, especially for agriculture dependant third world country like Bangladesh. Like most very low income countries, a majority of the poor population in Bangladesh are dependent on their own production of basic food crops. As a result, geographic and environmental factors cause loss of welfare for the people of some of the areas in the mean seasons. Therefore, even though employment programs are a relatively expensive way of making transfers and managerially complex compared to pure transfer programs (Smith and Subbarao 2003), they plays a vital role in the social protection of a country. 8.6 per cent of the total safety net of Bangladesh for FY2008 goes to creating employment opportunities for the vulnerable poor, apart from other micro credit programs that also aid employment creation. While there are programs to bring the extreme poor to the production process through employment, programs are there as well to create employment opportunities for those who suffer from seasonal unemployment due to geographic and environmental concerns. “Monga” affected areas in northern Bangladesh has been given special attendance in this regard. These areas yield only one or sometimes two annual harvests, in contrast with three crops per year in more fertile areas of the country. Employment opportunities, particularly for the landless and the poorest segment participating in agricultural wage labor, are limited from September through December in average years. Food for Work (sometimes cash for work) is a traditional safety net program in the country. It is arguably best to opt for self-targeted workfare because the intervention minimizes the tradeoff between investment in growth and safety nets through infrastructure building. There is also program to develop small entrepreneurs. Regarding employment programs, ensuring non-participation by the non-targeted population is of concern and calls for self-targeting mechanism with employments at a lower than market wage. Works program of Bangladesh provides 100 tk or 6 kgs of grain a day which fulfills this consideration [ (Government of Bangladesh, 2008) ]. 4.4.5. Special risk reduction
Risk reduction is increasingly seen as the primary function of public safety nets (Holzmann and Jorgenson 1999). The objective of such programs is to help the households withstand sudden shocks or disasters and avoid steep fall in consumption. This calls for action in different phases; direct support is provided after the shock or disaster; preventive support is provided before the occurrence of such shocks. Support may also be provided by allowing them to take on “riskier,” but higher return activities in the rehabilitation phase.
Bangladesh bears serious consequences from natural calamities, the poorest segment being the worst victim. In 2007 alone the country faced two floods and a devastating cyclone. During the disaster affected years, usually the expenditure on safety nets regarding such shocks exceeds the initial allocation. Safety net for FY2008 keeps Tk 1505 crore for disaster related activities which are about 18 per cent of the total. Major programs include VGF, Test Relief, immediate disaster fund and Gratuitous relief programs that are targeted to provide direct and immediate support to disaster victims. A fund for small farmers who are victims of disasters is kept to assist them in returning to normalcy after the disaster. 4.4.6. Subsidy
Largest share, around 34.5% within the total safety net goes to subsidies for the poor. Within the subsidy programs, 24.7 per cent is provided as food subsidy while the rest are provided as electricity and fuel subsidy. Food subsidy is considered as direct support to the vulnerable groups justified by the shortfall of food production of the country and associated price hike in food items. The energy and fuel subsidy on the other hand is growth oriented, providing inputs towards future production and long-term food price support with the expectations to ensure poor farmers with access to production inputs.
Regarding other types of subsidies, one of the intension of safety nets is considered to be assisting the vulnerable groups in adjusting with any possible adverse implications of macroeconomic policy changes. Social safety net programs designed for FY2008 includes Tk 750 crore subsidy for marginal farmers (within the total subsidy of Tk 2990 crore kept as safety nets) aiming at such intensions as electricity price adjustment was undertaken in FY2007. 4.4.7. Micro-credit
Reputation of the NGOs of Bangladesh in providing micro-credit to the poorest segment of the society played a role to ensure government-NGO collaboration in poverty reduction efforts. Safety net programs in micro-credit are one such example. While transfers in the form of micro-credit are carried out by the government body PKSF, the government runs micro-credit programs through the NGOs as well. Tk 238 crore micro credit programs (about 3 per cent of the total safety net) for FY2008 includes Tk 218 crore to be distributed by the NGOs and Tk 20 crore for the self employment of women through government ministries.
Overall, the structure of the SSNPs of Bangladesh indicates that poverty reduction and social protection of the poor have been well linked with growth facilitation through infrastructure development and employment creation.

4.4. Trends in Social Safety Nets
Obviously there is no answer to the question of what would be an optimum level of spending on safety nets. While this will be largely dependant on the poverty dynamics of a country and availability of resources, other concerns would be the efficiency of government in spending on growth oriented fields as well as the possibility of “leakages” or cost of delivering direct transfers. Smith and Subbarao [ (Subbarao, 2003) ] argued that if the efficacy of other public spending on health, education or infrastructure is low, direct transfer becomes an important candidate.
In case of Bangladesh, in monetary terms, extent of safety net programs is gradually increasing over the years. However, in terms of share of public expenditure, investment on safety nets has been falling since 1998-99, against the increasing number of people below the poverty line. On an average, during 1996-97 to 2004-05 period, Bangladesh has been spending on SSNPs to the tune of 0.8 per cent of the GDP and 5.7 per cent of the total public expenditure (based on world bank assessment of Bangladesh safety nets).
It is pertinent to mention here that actual expenditure data on SSNPs is not provided by the government.
Policy makers usually prefer transfers in kind over cash in most of the low income countries. It is a general perception that the hardcore poor in need of social protection suffers from starvation and need food support more than anything else. Even though cash can be more efficient for some programs, it is generally hard to sell politically. Bangladesh, however, seemed to have evolved through a more positive way. About 75 per cent of the transfer is made in cash. Programs that have shown weaknesses in kind support, for example Food for Education, has been converted into cash. The major social safety net programs used in Bangladesh are shown in Table 1.

Table [ 1 ]: Main Safety Net Programs of Bangladesh [ (F., 2007) ] Type Of Safety Net Program | Example of Safety Net Program | Cash Transfers | 1. Old Age Allowance 2. Widowed and Distressed Women Allowance 3. Disabled Allowance | Conditional Cash Transfers | 1. Primary Education Stipend Program/ Food for Education 2. Stipend for Female Secondary Students | Public Works/ Training Based Cash or Kind Transfer | 1. Rural Maintenance Program 2. Food for Work 3. Vulnerable Group Development | Emergency/ Seasonal Relief | 1. Vulnerable Group Feeding 2. Gratuitous Relief Test Relief |

4. Social Safety Nets and Poverty
Since the 1990s the growth of the economy started to find some momentum and that has also reflected in the poverty profile of the country. However during this period escalating income inequality has also turned out a foremost policy obligation. Various surveys conducted over time, have proved that social safety nets can be successfully used to restore income distribution among the poor and even prevent further deterioration.

5.5. Impact Assessment of Social Safety Nets on the Poverty of Bangladesh
Social safety net is especially designed and implemented for the poor and the vulnerable groups of the society. However the implication of these programs on poverty is often remained untouched. In recent times a number of countries and international agencies initiated researches to find out how safety net programs can effectively work for poor. However in Bangladesh such initiatives are somewhat limited. One of the major constraints is placed from the inadequacy of necessary information and data. At national level a survey regarding social safety net has yet to be commenced.

5.6. Inequality Effect of Social Safety Nets
The rising trend in inequality has perturbed the recent “fairly” improving poverty scenario. Over the years, academics around the world have been arguing on the relationship between poverty and inequality. However a number of studies concluded that inequality can influence the poverty structure in two ways – the initial low level of poverty can initiate the responsiveness of poverty to a given rate of income growth and at the same time changes in inequality during a growth spell affect the poor benefiting from a given level of growth. It has been demonstrated that in the short run, income growth accounted for just over two-thirds of the changes in relative incomes, and inequality or distributional change for the rest. Thus in any poverty discourse, inequality issue is considered as an essential agenda.

5. Conclusion
There is nothing like an ideal or universal safety net program design that can be guaranteed for successful poverty reduction. While most countries use indirect transfers in the form of subsidies, a large number of safety net programs are conducted through direct transfer programs. Among these, the conditional cash transfer programs have been successful and replicated in a number of countries while following the same route the unconditional cash transfer programs in some other countries have also yielded positive results. On the contrary, despite having any particular highly successful safety net program, Vietnam’s poverty reduction strategy through various policy variables and more particularly through agricultural diversification has managed to set a benchmark in this area. This actually points out the need for a coordinated policy framework for poverty reduction along with effective implementation of the safety net programs. 6. Recommendations
A well judged choice on modality is crucial for providing social protection in a country like Bangladesh. With the absence of an optimum mix, the government should opt for more to cash as the mode of transfer in view of the efficiency as well as the inbuilt flexibility that cash contain for the poor in its utilization. However, in case of risk management, distribution in kind would be the preferable choice since the nature of disasters in Bangladesh causes starvation more than anything else and mere survival becomes the prime concern. Regarding conditional Vs non conditional transfers, even though the performance of CCTs are well depicted in other countries, resorting only to conditional transfers will be no option for Bangladesh. While unconditional transfers will have to be there for the old and the disabled (the vulnerable groups) and disaster victims, conditional transfers can prove to be a useful tool in relating growth (through infrastructural and human resource development) and poverty reduction. Taking the lessons from other countries, conditional transfers might also help in child nutrition along with the ongoing education programs. Keeping in mind the declining enthusiasm in population programs, an effort to device a method linking it with conditional transfers might prove to be praiseworthy.
One important aspect of safety net programs demanding improvement in Bangladesh is the “regional disparity” issue. Uneven distribution of wealth and variation in poverty incidence between regions has to be incorporated in targeting the poor. Employment creation in Monga areas in a more permanent basis should be emphasized. At the same time, the growing urban poverty needs to be taken into account and social protection for the urban poor needs to be improved. One concern would be the food subsidy. In view of the steep rise in food price, the subsidy program should continue with due address to the growing number of lower-middle income group (who does not really belong below the lower poverty line) resorting to subsidized distribution of food grain. Financing post disaster security programs among the victims is always a big issue for a government of a third world country like Bangladesh, calling for donor assistance. However, for other types of safety nets, financing should depend on domestic resources as much as possible since donor contribution in safety nets tend to be discontinuous or not lasting for long, being replaced by other initiatives. This may result in loss of efficiency, a scarce resource for any developing country, gained by the implementing authorities from the previous program. Moreover, often the donor driven programs are not based upon requirement, rather on what is available from them. Lack of coordination between the funding partners also hold the chance of overlapping target groups, while some of the potential beneficiaries may remain outside the net.
Most important issue would be to manage leakages in transferring resource to the poor. Along with administrative reforms, this will call for better targeting. Avoiding participation of non-poor in the transfers require self targeting mechanism along with community targeting to be in place as much as possible. The local government bodies will need to be strengthened and work closely with potential beneficiaries to further enhance the targeting efficiency.
All together, a national policy on Social Safety Net will have to be developed with particular focus on strategic options for channeling aid to the safety nets and modality as well as targeting options. The national policy will need to be characterized by its promotional role in public-private collaboration in providing social security. At the same time it will facilitate a common aim and understanding between political regimes to avoid change of nature and means of safety nets with a change of government. Obviously, poverty reduction is not confined to, rather supplemented by, safety nets. However, safety nets undoubtedly are vital options to fight damages from natural disasters and for preventing transmission of poverty from generation to generation in the long run. While on one hand the safety nets provide immediate relief during emergencies, by ensuring employment, education and nutritional intake of the poor, these programs can often play leading role in addressing poverty and inequality in the long run. Success of any such attempt will depend on how, where and for whom the government decides to design the safety net programs. There are a lot of lessons to be learnt from the various successful conditional and unconditional transfer programs taking place in South America, Africa and India. Successful replication of such programs in the Bangladesh context may help to identify safety net programs that can leave radical impact on poverty reduction in Bangladesh.

7. Bibliography
F., H. N. (2007). Politics and Governance in the Social Sectors in Bangladesh, 1991 - 2006. In BRAC, Research Monograph Series No. 34. Dhaka.
Government of Bangladesh. (2008). Unlocking the Potentials: National Strategy for Accelerated Poverty Reduction.
Grosh M, d. N. (2008). For Protection and Promotion: The Design and Implementation of Effective Safety Nets. Washington D.C.: The World Bank.
Subbarao, S. W. (2003). What Role for Safety Net Transfers in Very Low Income Countries? Washington: The World Bank.
The World Bank Group. (n.d.). Retrieved from http://go.worldbank.org/JDLCWC6GR0