99250 From Gas to Cash: Policy Options for Transferring Resource Revenues to Citizens in Mozambique World Bank Policy Note June 2015 This policy note is a product of joint collaboration between the Macroeconomics and Fiscal Management, Poverty, and Social Protection & Labor Global Practices. The work was led by Natasha Sharma and the team included Pedro Olinto, Eric Zapatero, Vanessa Moreira da Silva, Carlos da Maia, Ruben Villanueva, Nada Choueiri, and Poorva Karkare. Feedback and support was provided by Enrique Blanco Armas and Julio Revilla, and the work was carried out under the overall guidance of Mark Roland Thomas. The team would like to acknowledge the financing provided by DFID for this work under the programmatic AAA 'Enhancing macroeconomic and fiscal policy making for inclusive growth in a resource-rich setting.' 1 2 Contents Acronyms ...................................................................................................................................................... 5 Executive Summary....................................................................................................................................... 6 Introduction .................................................................................................................................................. 9 Section 1: Growth and Poverty ................................................................................................................... 10 Section 2: The Social Protection System ..................................................................................................... 13 Section 3: Policy Options for Scaling-Up Cash Transfers in Mozambique’s Future Resource-Rich Environment ............................................................................................................................................... 16 (i) Rationale for a Scaled-Up Cash Transfer Program .......................................................................... 16 (ii) Design Options for a Scaled-Up Cash Transfer Program............................................................. 17  Option 1: Universal Cash Transfer Program................................................................................ 19  Option 2: A Cash Transfer Program Targeted to the Poor .......................................................... 21  Option 3: A Cash Transfer Program Targeted to the Poor with a Conditional Component ....... 24 (iii) Addressing Potential Risks of a Scaled-up Cash Transfer Program............................................. 27 (iv) Preparing for a Scaled-up Cash Transfer Program ...................................................................... 28 Conclusion ................................................................................................................................................... 30 References .................................................................................................................................................. 34 3 Figure 1: There is wide variation in how poverty responds to growth in resource-rich countries… .......... 12 Figure 2: ….While in non-resource-rich countries the dispersion is not as great ....................................... 12 Figure 3: Over time domestic financing for social protection is steadily increasing .................................. 14 Figure 4: The amount spent on progressive social action programs is increasing as subsidies decline ..... 14 Figure 5: With a budget constraint between 0 and 1.5% of GDP the poverty headcount does not vary much with different transfer levels* .......................................................................................................... 22 Figure 6: With a budget of 1% of GDP, the poverty gap index improves with a higher transfer value and lower coverage levels**.............................................................................................................................. 22 Figure 7: The poverty severity index also improves with a higher transfer value and lower coverage levels*** ..................................................................................................................................................... 23 Box 1: Characteristics of Poverty and Vulnerability in Mozambique ......................................................... 11 Box 2: Unconditional cash transfers improving poverty and welfare in selected peer countries .............. 16 Box 3: Methodology and Assumptions used in the Simulation Exercise .................................................... 18 Box 4: Determining eligibility for a targeted cash transfer program .......................................................... 19 Box 5: Natural Resource Wealth Redistribution through Cash Transfers in Alaska ................................... 20 Box 6: Cash Transfer Programs in Latin America Reducing Poverty and Inequality, and improving the use of Social Services ......................................................................................................................................... 25 Box 7: Transferring gas to cash in Bolivia to reduce poverty and promote development outcomes ........ 26 Box 8: Implementing an integrated cash transfer program in Mozambique ............................................. 26 Box 9: Universal distribution of cash transfers in Mongolia ....................................................................... 28 Table 1: Coverage of basic social protection programs implemented by INAS (number of Households).. 14 Table 2: Changes to monthly subsidies under the Basic Social Subsidy Program (PSSB) ........................... 15 4 Acronyms CCT : Conditional cash transfers CT: Cash transfer GEPR: Growth Elasticity of Poverty Reduction HDF: Human Development Fund INAS: Instituto Nacional de Acção Social (National Institute of Social Action) IOF: Inquérito aos Orçamentos Familiares (Household Budget and Expenditure survey) LNG : Liquefied Natural Gas MIS: Management Information System OVC: Orphan and Vulnerable Children PASP: Programa de Acção Social Produtiva (Productive Social Action) PASD: Programa de Acção Social Directa (Direct Social Assistance) PSSB: Programa de Subsídio Social Básico (Basic Social Subsidy) 5 Executive Summary While Mozambique’s growth rate is impressive, the extent to which growth impacts poverty reduction is declining, and this trend is likely to continue unless there are major policy changes. In the period after the end of the civil war (1996 to 2002) strong economic growth largely led by agriculture was associated with an impressive decline in poverty rates from 69 percent to 56 percent. In later years (2002 to 2009) economic growth continued to be strong, but as capital intensive investments increasingly contributed to GDP, including megaprojects with limited linkages to the rest of the economy, poverty fell slightly from 56 percent in 2003 to 52 percent in 2009. Therefore, the sources of growth matter, where labor-intensive sectors such as agriculture and manufacturing are typically more poverty reducing compared with resource extractive sectors. As Mozambique’s economy evolves to one where extractive industries will play an increasingly important role, this tenuous relationship between growth and poverty is likely to continue in the absence of major policy changes. In Mozambique’s future resource-rich environment, implementing a scaled-up cash transfer program could be a relatively cost effective policy option for using resource revenues to reduce poverty over a short period of time. The government is currently in a transition period, with an opportunity to plan for the use of resource revenues expected to come on stream from 2020 onward. Among the policy options under consideration, investing in human and physical capital, and strengthening institutions will undoubtedly continue to be government priorities, which should lead to improvements over the medium to long term. However, given absorptive capacity constraints it may not be possible for the pace of spending to increase at the same pace as government revenues. Implementing a scaled-up cash transfer program in Mozambique’s future resource-rich environment could complement these efforts through reducing poverty over a relatively short period of time. Evidence from both within the region and internationally has shown that the distribution of cash transfers can reduce poverty, improve welfare and promote development outcomes. In Africa, impact evaluations in Zambia, Kenya, Malawi and South Africa have all shown that the distribution of cash transfers can reduce poverty and have other positive impacts such as better agricultural productivity, increase in self-employment, enhanced health status and improved school attendance. The implementation of cash transfer programs in these countries have in part been motivated by the success of cash transfer programs worldwide, particularly in Latin America. Impact evaluations undertaken in Mexico, Colombia, Brazil and Nicaragua among others have demonstrated that the implementation of conditional cash transfer programs can be directly attributed to a reduction in poverty and improved welfare. Implementing a scaled-up cash transfer program could build on the achievements Mozambique has made in recent years in improving the social protection system. Over the last few years Mozambique has been engaging in efforts to modernize the social protection system, including the introduction of the Social Protection Law and the National Social Protection Strategy, which is underpinned by an increase in domestic financing. Furthermore, spending on social action programs which directly benefit the poor have been increasing. However, the coverage of existing social protection programs is low, where 6 approximately only 14 percent of the poor benefit. Implementing a scaled-up cash transfer program could build on existing achievements, although efforts would be needed to strengthen the overall social protection system, including: (i) strengthening capacity to administer methods for targeting the poor; (ii) establishing a management information system, which would include a single registry of beneficiaries; (iii) outsourcing the payment system; (iv) strengthening the capacity of the national Institute of Social Action (INAS) at the local level; and (iv) developing monitoring and evaluation mechanisms to include regular reports from the management information system, strengthened internal control measures and period impact assessments. Implementing a scaled-up cash transfer program targeted to a broad section of the poor, could rapidly reduce poverty and encourage wider support for this policy initiative from the population. Different options for implementing a scaled-up cash transfer program in Mozambique’s future resource-rich environment have been considered in this report. The optimal scenario for poverty reduction would be to distribute a relatively generous transfer (compared to the average consumption of the poor) to the entire population, which is estimated to reduce poverty from 52 percent to 36 percent. Assuming Mozambique was already in receipt of expected resource revenues this policy option would cost around 5 percent of GDP, which is unlikely to be affordable given other spending priorities. Instead, with a more limited fiscal envelope, a scaled-up cash transfer program could target the poor, and decisions would need to be made on whether to have a large target population or provide a more generous transfer to a smaller target population. With a budget of 1 percent of GDP, cash transfers could rapidly reduce poverty by 4 percentage points by covering the majority of the population under the food poverty line. Increasing the budget to 2 percent of GDP could further reduce poverty by 7 percentage points. One component of the scaled-up targeted cash transfer program could include a conditional cash transfer to encourage school attendance for children. While distributing cash without any conditions can reduce poverty and promote positive development outcomes, the evidence suggests that introducing conditions or co-responsibilities in order to receive cash can help to promote specific behaviors, such as increased school attendance. In Mozambique, school enrollment levels are high, expected to reach 100 percent, but pupil absenteeism is estimated to be 56 percent. Also, the majority of poor households have children and the current social protection programs do not specifically address their needs. Therefore, it is suggested that a conditional cash transfer could be introduced to encourage school attendance, but this would need to be complemented with broader efforts to improve the availability of schools and the quality of teaching. Over time, capacity would also need to be developed to monitor compliance with the conditions or co-responsibilities. There may be risks to implementing a scaled-up cash transfer program which should be addressed through sound fiscal management and program design. Concerns are sometimes put forward that cash transfers can have unintended consequences on the labor market and the economy. On the contrary, evidence from other countries suggests that recipients of cash transfers are more likely to participate in the labor market and engage in productive agricultural activities. Other concerns could be related to a scaled-up cash transfer program being financed through resource revenues because of the potential risk of volatility. These risks can be avoided through sound macroeconomic and fiscal management, such as 7 avoiding earmarking resource revenues for cash transfers. Instead, the increased fiscal space should be used as an opportunity to increase financing through the regular budget process for cash transfers. There may also be concerns that a cash transfer program could stoke inflationary pressures, which could be avoided through ensuring the transfer is capped at a modest level. In sum, once resource revenues are on stream, a scaled-up cash transfer program could be an affordable and effective way of significantly reducing poverty over a short period of time. A scaled-up cash transfer program could be part of the government’s broader strategy toward ensuring impressive economic growth is having an impact on poverty reduction, which is likely to be even more challenging in a resource-rich environment. With revenues from natural resources on stream, a scaled-up cash transfer program would be affordable and a relatively cost-effective policy option. Therefore, implementing a scaled-up cash transfer program provides an opportunity to significantly reduce the number of people living below the poverty line, and improve other welfare measures, while sharing the benefits of resource wealth more widely amongst the population. 8 Introduction If developments in the extractive industries are implemented as planned, Mozambique could have unprecedented opportunities to reap the benefits of a natural resource boom. Based on proposed developments in the liquefied natural gas (LNG) sector, government revenues are expected to increase significantly by 2020 onward. By 2032 resource revenues could be as high as US$9 billion, representing 7 percent of GDP and 21 percent of total government revenues. If channeled correctly, resource revenues could significantly reduce poverty and provide a basis for inclusive growth and shared prosperity. Mozambique has enjoyed strong economic growth but poverty levels are still unacceptably high. Economic growth has been strong over the last two decades, with an average annual growth rate of 7.4 percent in real terms. Despite the sustained growth levels an estimated 52 percent of the population lives below the national poverty line (Alfani et al, 2012). Over time, the extent to which poverty has reduced in response to economic growth has declined, coinciding with the extractive and capital- intensive sectors increasingly contributing to growth. Looking ahead, it is likely that this pattern of growth will continue, suggesting that in the absence of major policy changes strong economic growth may continue to have disappointing results on poverty reduction. Mozambique is now in a transition period with an opportunity to plan for how resource revenues can contribute to poverty reduction and inclusive growth. A clear priority will be to invest in human and physical capital and to strengthen the institutional environment. Through investing resource revenues effectively, benefits could be accrued to current and future generations by promoting higher and more inclusive growth. However, due to absorptive capacity constraints the pace at which public spending can be increased will be limited, suggesting that a strategy to alleviate poverty more quickly could be called for. The distribution of cash transfers could form a key part of such a strategy, enabling the government to significantly reduce poverty over a relatively short period of time. Indeed, the experience in other countries suggests that cash transfer programs can directly reduce poverty, improve welfare and promote development outcomes. Impact evaluations undertaken in the region, and further afield in a number of Latin American countries, provide wide-ranging evidence on the positive effects of cash transfer programs in reducing poverty and promoting welfare. In Mozambique, while there have been some improvements to the social protection system, coverage levels are low reaching only an estimated 14 percent of the poor. The onset of resource revenues provides an opportunity to significantly scale-up the distribution of cash transfers, which could be a comparatively cost-effective way of reducing poverty in a relatively short period of time. A scaled-up cash transfer program could also be a way of distributing the benefits from natural resources amongst a wide segment of Mozambicans, thereby promoting broader popular support for this policy initiative. Any policy to scale-up a cash transfer program will operate with a limited budget, meaning that decisions will need to be made on the optimal design choice in Mozambique. In a resource-rich environment, the possibility of implementing a scaled-up cash transfer program will be much more 9 affordable. Nevertheless, due to other spending priorities, there is likely to be a limited fiscal envelope, implying the need to make decisions on the optimal design of a cash transfer program in order maximize poverty reduction and improve welfare. Therefore, there are likely to be trade-offs. For example would it be better to have a more generous transfer and cover a smaller segment of the population, or would wider coverage and a less generous transfer promote greater gains? Similarly, there may be questions on how the target population should be defined, for example with limited resources should a scaled-up cash transfer program attempt to cover all of the poor, or to target the very poorest? Furthermore, from a political economy perspective, ensuring support for a cash transfer program may involve increasing coverage of the population so that the benefits are felt more broadly. Finally, it is important to consider the potential risks of a scaled-up cash transfer program, and the steps that could to be taken to prepare for implementation. The objective of this policy note is to generate debate on implementing a scaled-up cash transfer in Mozambique’s future resource-rich environment, as part of a broader strategy to reduce poverty. The scope of this note is focused on distributing resource revenues through a scaled-up cash transfer program, and not the broader management of resource revenues.1 Section one discusses growth and poverty dynamics. Section two presents the existing social protection system. Section three discusses policy options for implementing a scaled-up cash transfer program using a simulation exercise to estimate poverty and welfare effects for a given fiscal envelope. Section four discusses how to address the risks of financing a scaled-up cash transfer program from resource revenues. Section five focuses on the practicalities of how the social protection system should be strengthened to implement a scaled-up cash transfer program. And the final section concludes. Section 1: Growth and Poverty Mozambique has enjoyed high levels of economic growth, where extractive industries are increasingly playing an important role. Over the last two decades an average real growth rate of 7.4 percent has been sustained. From 1992 to 2000, following the end of the civil conflict, strong growth was supported by agricultural activity. While agriculture continues to play an important role, over the last decade capital intensive investments including megaprojects have become increasingly important, and are expected to continue growing as plans to develop coal and LNG sectors come underway. However, capital and import-intensive megaprojects have developed with limited linkages to the local economy. While economic growth has been impressive, job creation has been limited, leading to high levels of underemployment. Mozambique has an estimated population of 25 million in 2014, with an economically active population of around 11 million (Cunha et. al, 2013). There are high rates of labor force participation where virtually everyone of working age is economically active, albeit not to their full potential. Unemployment largely affects the urban youth with an estimated 39 percent being unemployed (INCAF, 2012). Overall, there are limited formal job opportunities, where approximately 10 percent of the population work in the formal sector and are salaried employees (Jones and Tarp, 2012). 1 Separate policy notes on natural wealth accounting and resource revenue management have been developed which complement the findings in this policy note 10 Informal sector employment is dominated by the agriculture sector, with an estimated 70 percent of the population involved in subsistence farming. The extent to which poverty reduction is responsive to economic growth is declining over time. The rapid economic growth at the end of the civil war, which was supported by agricultural activities helped raise living standards and per capita incomes from a very low base. Poverty rates fell from 69 percent of the population in 1997 to 56 percent in 2003. During this period, a 1 percent increase in GDP was associated with a reduction in poverty by 1.44 percent (a relationship described as the growth elasticity of poverty – GEPR). In comparison, from 2003-2009 poverty only fell slightly from 56 percent in 2003 to 52 percent in 2009, where a 1 percent increase in GDP was associated with a reduction in poverty by 0.17 percent. During this latter period capital intensive megaprojects increasingly contributed to GDP, and it is likely this trend will continue, implying that in the absence of major policy changes growth may continue to have a disappointing impact on poverty reduction (see box 1 for a description of poverty and vulnerability in Mozambique). Box 1: Characteristics of Poverty and Vulnerability in Mozambique Poverty is associated with a number of factors in Mozambique including demographics, education and employment. The poorest households are typically larger, have a higher proportion of children, people with disabilities, greater dependency rates, are female headed with lower education levels and are predominantly rural. Given the high prevalence of poverty, consumption levels are comparatively flat for quintiles 1 to 4, making it relatively challenging to identify the absolute poorest. The age-dependency ratio in poor households is above 60 percent, which is considerably higher than non-poor households at 51 percent. Of the poorest 10 percent of households, 91 percent have children. Persons over 60 years comprise 10 percent of the poorest quintile. About 25 percent of children (5 to 14 years) and more than 80 percent of the senior population (60 years and above) are economically active. The poor are largely engaged in the agriculture sector, without access to any kind of insurance or social benefit. Levels of access to financial and credit markets are also low. In rural areas only an estimated 19 percent of people report using a financial product from a bank, compared to 43 percent for the urban population. Households with higher education levels have greater consumption levels. For example, female literacy rates in the poorest two quintiles are 12.7 percent and 17.0 percent respectively, compared with 83.9 percent in the least poor quintile. However, the quality of education is a concern where the youth literacy rate is 67.2 percent, and is one of the lowest compared with peer countries. Rural households are vulnerable to seasonal and frequent climate-related shocks as they are heavily dependent on agriculture-related income. During the “lean season” affected households are estimated to consume 40 percent fewer calories. Poor households in urban areas are particularly vulnerable to protracted underemployment, low income, and food price increases. With shifts in labor movements from low productivity rural agriculture to even lower productivity urban activities, the vulnerabilities of urban households are of increasing concern. Sources: IOF 2008-09; Fox et. al 2012; Demographic and Health Survey 2011; and World Bank, 2012 and 2013a, WB FinCap Survey 2013 11 Evidence from the region also suggests that growth led by extractive industries will not immediately contribute to poverty reduction. Looking at a sample of 15 countries in Sub-Saharan Africa there is wide variation in the extent to which poverty reduction is responsive to changes in growth in resource-rich countries.2 Poverty reduction has been more responsive to growth in resource-rich countries such as Niger, Botswana and the Republic of Congo (see figure 1). However, in other resource-rich countries such as Zambia, Cameroon and Guinea increases in per capita growth rates have occurred in parallel with worsening levels of poverty. In comparison, for the majority of non-resource rich countries increased per capita growth has led to a reduction in poverty (see figure 2). These trends can partly be explained by the sources of growth being important for poverty reduction, where growth in labor intensive sectors such as agriculture and manufacturing is typically more poverty reducing compared with capital intensive sectors (World Bank, 2013b). Therefore, growth related to the extractives will not have an immediate impact on poverty reduction, suggesting the importance of adopting specific policy measures to reduce poverty in Mozambique. Figure 1: There is wide variation in how poverty Figure 2: ….While in non-resource-rich countries the responds to growth in resource-rich countries… dispersion is not as great3 5 5 3 3 1 GEPR 1 GEPR -1 -1 -3 -3 -5 -5 Mozambique Malawi Benin Ethiopia Burundi Ghana Rwanda Tanzania Uganda Swaziland Cabo Verde Togo Lesotho Chad Burkina Faso Zambia Guinea Niger Namibia Sierra Leone Mauritania Senegal Mali Botswana Nigeria Angola Congo, Rep. Cameroon Source: World Development Indicators Initial levels of inequality also affect the extent to which economic growth impacts poverty reduction. Studies using cross-country data have shown that where initial inequality levels are lower, growth has a larger impact on poverty reduction (Bourguignon, 2003 and Chhibber and Nayyar, 2007). Latest available estimates show that levels of inequality are actually higher in Mozambique compared with a number of resource-rich countries in SSA such as the Republic of Congo, Guinea, Mali and Sierra Leone (WDI). Therefore, for Mozambique to promote pro-poor growth, reducing inequality should be a priority, where redistributive income policies can play an important role. In sum, in order for strong economic growth increasingly related to the extractive industries to impact poverty reduction it is likely that specific policy measures will be required. While economic growth has 2 The analysis is based on differences in poverty headcount ratios using a sample of 28 spells (the period between two consecutive surveys) which vary by five to nine years, depending on data availability 3 For the purposes of this analysis Mozambique, Ghana, Uganda and Tanzania are categorized as not being resource-rich as the recent discoveries in natural resources were not significantly contributing to growth during the survey periods 12 been strong in Mozambique, there has been limited creation of formal job opportunities, resulting in high levels of underemployment. Following the end of the civil war, strong economic growth supported by agricultural activity was associated with a significant decline in poverty. However, more recent data shows that the pace of poverty reduction has slowed, coinciding with capital intensive megaprojects increasingly contributing to GDP. Evidence from the region shows that in resource-rich countries growth does not always result in poverty reduction, and that performance varies considerably. Mozambique has relatively higher inequality levels compared with other countries in the region, which is also likely to affect the extent to which economic growth reduces poverty. Therefore, as the extractive industries grow in importance, specific policy measures to directly reduce poverty are called for. Section 2: The Social Protection System The social protection system is in the process of being modernized as part of the government’s response to tackling poverty and vulnerability. The government has implemented new institutional arrangements for social protection, including a Social Protection Law (2007) and a National Social Protection Strategy 2010-2014. This strategy is being operationalized through three core social action programs: (i) The Basic Social Subsidy Program (PSSB), a social pension for poor households without adults able to work (elderly, disabled and chronically ill); (ii) the Productive Social Action Program (PASP), a labor intensive public works program for poor households with adults able to work, and (iii) the Direct Social Support Program (PASD), a temporary support program for households experiencing idiosyncratic shocks affecting their consumption and income. These three core cash transfer programs are complemented by the provision of social services for targeted vulnerable groups in institutions like orphanages, community centers for the elderly or daycare centers for vulnerable children. The progress in the social protection sector has been facilitated by an increase in domestic financing for social protection. The increase in the government’s contribution to social protection is a shift in trend from previous years when donors largely provided financing to the sector. Importantly, the increased spending is being channeled toward social protection programs, while the amount allocated to subsidies is reducing. This trend is important as subsidies were found to be regressive and not pro- poor, indicating that spending on social protection is increasingly benefitting the poor (World Bank, 2012). 13 Figure 3: Over time domestic financing for social Figure 4: The amount spent on progressive social protection is steadily increasing4 action programs is increasing as subsidies decline 1.5% 1.5% 1.0% % of GDP 1.0% % of GDP 0.5% 0.5% 0.0% 0.0% 2011 2012 2013 2014 2011 2012 2013 2014 Recurrent External investment Internal investment Social Action Programs Subsidies Sources: 2011 to 2013 CGE; 2014 REO Overall, the coverage of social protection programs in Mozambique is low. There has been some expansion of the major social assistance programs (see table 1),5 but coverage levels still remain low. An estimated two million people benefit from one of the major programs thereby covering approximately 14 percent of the poor, or eight percent of the entire population.6 Furthermore, important programmatic gaps still remain, particularly for poor households with children, which represent a large share of the poor. Table 1: Coverage of basic social protection programs implemented by INAS (number of Households) Programs 2012 2013 2014 (Planned) Direct Social Assistance (PASD) 33,988 51,414 43,698 Basic Social Subsidy (PSSB) 274,025 304,576 341,188 Productive Social Action (PASP) 9,592 11,345 33,680 7 Social Services for Social Action 4,470 8,971 8,818 Source: INAS 2012 and 2013 Annual Reports and Economic and Social Plan, 2014 Coverage levels are also low compared with peer countries. In SSA coverage levels vary significantly where middle income countries have a much larger coverage compared with low income countries. For example, South Africa’s grant system reaches approximately 27 percent of the entire population, where the Child Support Grant covers 8.8 million children, comprising 70 percent of the target population. Coverage of the target population is around 70 percent in Botswana and 20 percent in Malawi.8 4 This figure shows the level of resources channeled for social protection through the government’s treasury system and therefore does not include donor financed projects that do not use government systems 5 The Mozambique Social Protection Assessment, 2012 has a detailed assessment of all social protection programs and coverage levels including those implemented by the Ministry of Education, Ministry of Health and development partners 6 This calculation is based on a national average household size of 4.7 people 7 This is not a cash transfer program, but institutional support for targeted vulnerable groups 8 Aspire database 14 Furthermore, coverage levels are very low compared with countries where cash transfers have been used as an instrument to significantly reduce poverty. For example, Mexico’s Oportunidades program covers 72 percent of the poor population and Brazil’s Bolsa Familia covers 84 percent of the poor (Garcia and Moore, 2012). A study examining cash transfers in Eastern Europe and Central Asia found that the mean coverage rate of the poorest quintile is 42 percent (Fiszbein and Schady, 2009). Therefore, while coverage levels have been improving in Mozambique they are still far below countries in the region and countries that are using cash transfers as a key policy instrument to reduce poverty. There have been some improvements to the generosity of transfers. As seen in table 2 the transfer value of the PSSB subsidy is improving over time. However, not all social protection programs have increased in value. For example, the transfer value of the labor intensive public works program has remained unchanged since 2012, where households receive 650 MT per month for a maximum of four months. For a household solely relying on this transfer for its annual income, the value is equivalent to 217 MT per household per month or 1.6 MT per household member per day, which is only 17 percent of median per capita consumption of the poor.9 Therefore, while generosity of the PSSB has been improving this is not the case for other non-contributory social protection transfers. Table 2: Changes to monthly subsidies under the Basic Social Subsidy Program (PSSB) Size of the Household (number of Amount (MT) in 2012 Amount (MT) in Amount (MT) in 2014 household members) 2013 1 130 250 280 2 190 320 350 3 260 380 420 4 320 440 480 5 or above 380 500 550 Source: INAS Economic and Social Plan, 2012, 2013 and 2014 In sum, while improvements have been made to the social protection system, the scale of implementation is still relatively low. Important progress includes improving the institutional environment for social protection programs and increasing domestic financing for programs that benefit the poor. However, overall coverage levels are low, reaching only 14 percent of the poor. This coverage level is below other countries in the region, and countries that have used cash transfer programs to significantly reduce poverty levels. 9 This value is calculated as follows: 650 MT multiplied by the four month limit, and divided by twelve months. This amount is then divided by 4.7 as the average household size and then taken as a percentage of the median per capita consumption of the poor for each household member 15 Section 3: Policy Options for Scaling-Up Cash Transfers in Mozambique’s Future Resource-Rich Environment (i) Rationale for a Scaled-Up Cash Transfer Program A growing body of evidence from other countries demonstrates that cash transfer programs can reduce poverty and improve welfare. Cash transfer programs have successfully contributed to reducing the number of people living below the poverty line, and the depth and severity of poverty.10 Amongst the best known examples are in Latin America, where conditional cash transfer (CCT) programs have reduced poverty and inequality, and enabled households to smooth consumption to cope with shocks. For example, CCT programs in Colombia and Nicaragua significantly reduced poverty among program beneficiaries, and in Brazil and Mexico the wide-scale implementation of CCT programs has reduced poverty more broadly at the national level. Furthermore, by making the receipt of cash conditional on the utilization of social services, CCT programs have encouraged school enrollment and use of health facilities (see box 6 in the section on CCT programs for further details). Following these experiences, a number of other countries have implemented cash transfer programs both with and without conditions, also showing positive results. Amongst peer countries in Africa the emerging evidence shows that distributing cash reduces poverty and also leads to other benefits such as improved agricultural productivity, food security, and better education and health statuses (see box 2 for further details). Box 2: Unconditional cash transfers improving poverty and welfare in selected peer countries  In Zambia an evaluation of the Child Grant Social Cash Transfer program, which targets transfers to poor households with children without conditions was found to reduce extreme poverty by 5.4 percentage points. The depth and severity of poverty also reduced by 10 percentage points. Other productive impacts were observed where the use of agricultural inputs and cultivated land increased, contributing to a higher value of the harvest by 50 percent.  In Kenya, an evaluation of the GiveDirectly unconditional cash transfer program attributed a 58 percent increase in assets built by the poor to the program, an increase in food consumption by 20 percent and a 42 percent reduction in the number of days children go without food. Transfers also increased revenue from animal husbandry and self-employment.  In Malawi, an evaluation of the unconditional Mchinji Social Cash Transfer Scheme showed that recipients of the cash transfer registered an increase in health status, had improved hygiene practices, greater school enrollment rates with reduced absences, improved agricultural production and greater food security.  In South Africa, an impact assessment of the Child Support Grant suggests that children enrolled in the program reported improved height-for-age scores for children, improved school attendance and learning outcomes particularly for girls, and reduced likelihood of illness especially for boys. Sources: Seidenfeld et. al, 2013, Haushofer and Shapiro, 2013, Miller et. al, 2008, and UNICEF South Africa, 2012 10 The depth of poverty is also known as the poverty gap, and is defined as the mean shortfall from the poverty line expressed as a percentage of the poverty line. The severity of poverty is also known as the squared poverty gap and is calculated by averaging the squares of the poverty gap relative to the poverty line 16 Implementing a scaled-up cash transfer program in Mozambique would provide a means to reduce poverty in a relatively short space of time. Given low levels of human and physical capital, it is a priority to use the increased fiscal space from resource revenues to support development goals such as investing in public goods, infrastructure, agriculture and strengthening the institutional environment. These efforts are likely to produce results over the medium to long-term. Due to absorptive capacity constraints there will also be limitations to the pace at which public spending can be effectively increased to enhance public service provision. A scaled-up cash transfer program would provide a relatively rapid way to increase household consumption and reduce poverty, enabling the current generation of Mozambicans to increase their welfare, which in turn could generate further growth. The experience in other similar contexts (for example, in Tanzania), shows that scaling up a cash transfer program would take four to five years to achieve full national coverage. A scaled-up cash transfer program would also be a more effective intervention for reducing poverty compared with other policy measures. One potential policy option under consideration is to provide gas at subsidized rates to domestic industries to promote local employment opportunities. However, the number of potential jobs that can be created is limited and there is a high subsidy for job creation.11 For example, in the methanol industry potentially 420 jobs could be created, and the average annual cost of creating one job through providing subsidized gas is estimated at US$ 114,000. Similarly, a fertilizer plant megaproject is estimated to generate 300 jobs at an approximate annual cost of US$ 166,000 per job. If the subsidies that would be used to create these potential 720 jobs were instead distributed as cash transfers it could be possible to lift over 720,000 people out of poverty in one year, thereby scaling up the benefits a thousand fold. In sum, a scaled-up cash transfer program is a relatively quick and cost effective policy option for reducing poverty over the short term. The international evidence demonstrates that cash transfer programs have been directly attributed to reducing poverty and inequality, as well as promoting productivity and the use of social services. A scaled-up cash transfer program could reduce poverty in a relatively short space of time, which would complement continued efforts to improve service provision in priority areas such as health, education and infrastructure. Furthermore, in comparison with other policy interventions such as the creation of local jobs through providing subsidized gas, a scaled-up cash transfer program is a relatively cost effective policy measure that could potentially benefit a broad section of the population. The next section will discuss different design options for a scaled-up cash transfer program, and the likely costs in a resource-rich environment. (ii) Design Options for a Scaled-Up Cash Transfer Program It is assumed that the primary objective of a scaled-up cash transfer program should be poverty alleviation. Cash transfer programs can be designed in different ways to address specific issues. Given high levels of poverty, it is assumed that the main objective of a scaled-up cash transfer program in 11 These estimations are based on data from the Gas Master Plan (2014) which specifies the amount investors are willing to pay for gas, the quantity of gas they are requesting, the netback value of LNG to estimate opportunity costs of subsidizing gas, and the approximate number of jobs that could be created 17 Mozambique should be to reduce poverty and improve welfare. However, other secondary objectives may also be considered. For example, a cash transfer program may be a way of sharing resource wealth amongst the population or the program may be designed in a way that will promote the utilization of public services such as education and health facilities to address low levels of human capital. Experience also shows that cash transfer programs have resulted in enhanced productive activities by beneficiaries (see box 2). Therefore, in discussing different design options, poverty and extreme poverty alleviation will be the main objective, while other secondary objectives may also be considered. To inform discussions on how a scaled-up cash transfer program can maximize poverty reduction for a given fiscal envelope, a simulation exercise has been undertaken. The exercise estimates the likely impact that a scaled-up cash transfer program targeted to the poor could have on poverty reduction for a given budget. The exercise uses the latest available data on poverty which is from the IOF 2008/09. To assess the affordability of a cash transfer program in a resource-rich environment, a counterfactual GDP has been estimated which assumes the revenues expected from coal and gas in 2025 were available in 2009. The scenarios use different transfer values in relation to the median per capita consumption of the poor, to see how poverty levels vary as the generosity of the transfer increases. Details on the assumptions used for the simulation exercise are presented in box 3. Box 3: Methodology and Assumptions used in the Simulation Exercise  Due to data availability, the exercise estimates the likely impact on poverty if the cash transfer program had been implemented in 2008/09, using the IOF 2008/09.  A counterfactual GDP has been estimated that assumes the revenues from coal and gas expected in 2025 were available in 2009, based on all four LNG trains being built.  A proxy means test (which is a targeting method used to estimate household income by associating indicators or ‘proxies’ with household expenditure or consumption) is applied to estimate household consumption in urban and rural areas.  The predicted values of daily per capita household consumption are used to calculate the probability of a given household being poor.  These household consumption values are used to determine eligibility to receive a cash transfer.  The estimated median per capita consumption of the poor in the 2008/2009 IOF survey is MT 9.48 /day in 2009 prices. This value is equivalent to $10.47 /month in 2009, and $14.24 /month in 2015.  Four different transfer values have been considered which correspond to 50, 100, 150 and 200 percent of the median per capita consumption of the poor.  It is assumed that all cash transfers are consumed. Source: IOF 2008/09 and World Bank staff calculations The simulation exercise shows how the extent to which a scaled-up cash transfer program can reduce poverty and enhance welfare for a given budget will depend on three factors: coverage, generosity and targeting accuracy. For a given level of resources, decisions will need to be made on whether a scaled-up cash transfer program can maximize poverty reduction through having a wider coverage level so that more poor people could benefit, or by providing a more generous transfer, which may also affect the depth and severity of poverty. Therefore, for a given fiscal envelope there are trade-offs between 18 coverage levels and generosity. However, even once a target population has been defined using criteria for eligibility based on poverty, it may not be possible to reach all of the intended beneficiaries due to targeting errors. By targeting individuals with a higher probability of being poor, the number of non- poor individuals benefiting would reduce but some of the intended poor individuals would also be excluded by mistake (see box 4 for further details). Therefore, the design of a cash transfer program focused on poverty alleviation would need to consider all three factors in determining how welfare can be maximized. The results of the simulation exercise are discussed in relation to two options: (i) a universal cash transfer program; and (ii) a cash transfer program targeted to the poor. Box 4: Determining eligibility for a targeted cash transfer program Under the simulation model developed, the estimations of daily per capita consumption values are used to determine the probability of an individual being poor, which in turn is used to assess if an individual is eligible to receive a cash transfer. For example, if it is decided that a cash transfer program should cover all of the poor this would mean selecting a probability of being poor of greater than or equal to 0 percent, which essentially means universal coverage. Under such a scenario all 52.08 percent of the poor would benefit, but 47.92 percent of Mozambicans who are not poor would also receive a transfer, which would be the error of inclusion or leakage rate. Instead, to reduce the number of non-poor receiving transfers, it may be decided that the cash transfer program should target the poor. Knowing the probability of an individual being poor through estimating daily household consumption helps to define the target population. For example, there could be a scenario where beneficiaries are eligible if they have a probability of being poor of greater than or equal to 25 percent, which would mean 87.67 percent of the poor are covered (see annex 1 for details). However, because targeting is not perfect, this scenario could result in 35.84 percent of the non-poor being included, and for some of the poorest being excluded by mistake, also known as exclusion errors. Attempts to reduce errors of inclusion or leakage would also mean that the coverage of the poor reduces, indicating trade-offs between inclusion and exclusion errors. A full presentation of different eligibility points, coverage of the poor and leakage rates is provided in the annex. A discussion on how to reduce targeting errors is included in section iv. Source: World Bank staff projections  Option 1: Universal Cash Transfer Program A universal program could significantly reduce poverty, but it is unlikely to be affordable or the most cost effective option. The greatest impact on poverty is when the highest transfer value is distributed universally. Under such a scenario 200 percent of the median per capita consumption of the poor is distributed to the whole population, resulting in the poverty headcount ratio falling from 52.08 percent to 36.56 percent. However, the cost of this program is estimated to be around 5 percent of GDP, which is unlikely to be affordable given other spending priorities. If instead the second highest transfer value is distributed universally the poverty headcount ratio would reduce from 52.08 percent to 44.69 percent and cost an estimated 2.57 percent of GDP. However, for a similar cost it would be possible to reduce poverty even further under a cash transfer program targeted to the poor. For example, for a budget of 2.46 percent of GDP it could be possible to cover approximately 78 percent of the poor and the poverty headcount ratio would reduce to 42.98 percent. Therefore, a more affordable and cost-effective approach would be to target a cash transfer program to the poor. 19 A universal cash transfer program could also have the objective of sharing resource wealth among citizens, to help create a broad-based constituency for resource revenue management. It has been argued that transferring cash directly to citizens could be an effective way of ensuring the benefits of natural resources are distributed amongst the population (Sandbu, 2005). Such an approach could also create support for the management of resource revenues, particularly in Mozambique where a large majority of the population may be vulnerable to falling into poverty at some point in their lifetimes. In practice, not many countries have prioritized this wealth sharing objective. The state of Alaska in the US is a rare example12 of where resource revenues are distributed to citizens, with everyone receiving a share of the income accumulated from earnings from the Alaska Permanent Fund. Part of the motivation for this policy measure results from apprehensions from the electorate that politicians would not use the revenues effectively based on past experience (see Box 5 for further details). Box 5: Natural Resource Wealth Redistribution through Cash Transfers in Alaska Alaska has institutionalized universal unconditional cash transfers linked to its resource revenues. One of the main motivations for the universal cash transfer was the fear that politicians would squander the resource revenues. At least 25 percent of resource revenues accruing to the state should be deposited with the Alaska Permanent Fund. The Fund can only invest in income producing assets, and only earnings on these investments can be spent. Since 1982, a share of these earnings is distributed to residents of Alaska (who need to apply and give proof of residency) following a clearly defined formula that ensures some stability in the annual payout as well as independence from political pressures to maximize the dividend in the short term for electoral purposes. Approximately 50 percent of annual returns from the Alaska Permanent Fund are distributed in equal amounts to all citizens, as an annual payment, regardless of their income level. Meanwhile, sales and income tax have been abolished. There is wide-spread support for the cash transfer program, which is largely viewed as an entitlement. Sources: Anderson, 2002; and Goldsmith, 2001 Another potential objective of a universal cash transfer program could be to promote governance outcomes through taxing a portion of the transfer, but it is unlikely this would be feasible in the Mozambican context. For countries rich in natural resources it may be less important to rely on taxation to finance public service provision, which is argued to weaken accountability between the state and its citizens, encouraging authoritarianism, corruption and poor service delivery (Moore, 2004). To address some of these concerns, a portion of the cash transfer could be taxed, which could encourage accountability since citizens are arguably more aware of their rights as a taxpayer, and an increase in direct taxation is associated with more democratic institutions (Devarajan et al., 2010; Sala-i-Martin and Subramanian, 2003; Sandbu, 2005; Gupta et al., 2014; and Moss and Young, 2009). However, it is unlikely that taxing cash transfers could work in Mozambique. Firstly, while the principle of using taxation to improve accountability is theoretically appealing there are no real examples of direct taxation from a universal distribution of cash from resource revenues. In the case of Alaska for example, sales and income taxes have been abolished, which is leading to growing apathy from the population on 12 The state of Alberta in Canada is also referred to in the literature as an example of where universal benefits have been implemented. However, Alberta does not have an established program of universal benefits. Rather, there have been two one-off payments to residents from an oil dividend (Gupta et. al, 2014). As part of efforts to restructure subsidies, Iran implemented a universal cash transfer program, where implementation is in the early stages. Finally, Mongolia has implemented universal cash transfers on an iterative basis which is discussed in Box 9. 20 public spending scrutiny and a decline in public goods (Devarajan et. al, 2010). Secondly, in Mozambique the amount that could be transferred would not be sufficient to allow for taxation, which would compromise the intended governance objectives.13 In sum, while a universal cash transfer program does have certain advantages; implementing such a program is unlikely to be feasible in Mozambique. Since a universal cash transfer program would cover the entire population, all the poor would benefit, and it would provide a way of sharing resource wealth more broadly. However, it is unlikely that a universal cash transfer program will be affordable if the highest transfer value is distributed; or the most cost effective option to alleviate poverty if a lower transfer value is distributed. Secondly, while a universal cash transfer program could theoretically provide a means to realize potential governance objectives through taxation, this is unlikely given the low transfer values. Therefore, a more feasible and cost effective option would be to implement a cash transfer program targeted to the poor, which will be discussed in detail below.  Option 2: A Cash Transfer Program Targeted to the Poor A second option is to implement a cash transfer program targeted to the poor, which could reduce poverty at a lower cost. Incorporating the likelihood of targeting errors, the simulation exercise shows that for a budget of 1 percent of GDP implementing a program with either (i) a generous transfer value and lower coverage level; or (ii) a less generous transfer value and higher level of coverage would yield the same result, with poverty reducing from 52 percent to 48 percent (see figure 5). With the first option 41 percent of the poor would be covered with an estimated 25 percent leaking to the non-poor and with the second option 72 percent of the poor would be covered but 30 percent would leak to the non-poor. It is important to note that with the second option, almost the totality of those under the food poverty line, comprising 40 percent of the population, would be included in the program. Going further, if the budget was increased to 2 percent of GDP, poverty could be reduced from 52 percent to 45 percent, again with either a lower transfer value and high levels of coverage or vice versa. Therefore, a scaled-up cash transfer program would be both a rapid and affordable way of reducing poverty in Mozambique’s future resource-rich environment. Overall, welfare can be maximized in a scenario with the most generous transfer value targeted to a more narrowly defined coverage of the poor. With a limited fiscal envelope of 1 percent of GDP, a scenario with a generous transfer value (for example one that is double the median per capita consumption) could be distributed to the very poorest. Since the consumption of the very poorest is even further from the poverty line compared with the median consumption of the poor, increasing the generosity of the transfer reduces both the depth and severity of poverty (see figures 6 and 7). This result indicates that the greatest impact on overall poverty alleviation and welfare is when a more generous transfer is distributed among a smaller share of the population. 13 The annual personal income tax threshold in Mozambique is MT 42,000 (equivalent to approximately US$ 1,360) and so the transfer would in most cases not be taxed. 21 Figure 5: The poverty headcount does not vary much with different transfer levels* 53 Transfer amount per household per year % of median consumption of the poor (in 2015 US dolars) 51 50% 100% 150% 200% 49 Poverty Headcount 47 45 43 41 39 37 35 0 1 2 3 4 5 Cost (as % of GDP) * The red/horizontal line indicates the baseline poverty headcount. Figure 6: The poverty gap index improves with a higher transfer value and lower coverage levels** 22 Transfer amount per household per year % of median consumption of the poor (in 2015 US dolars) 50% 100% 20 150% 200% 18 Poverty Gap Index 16 14 12 10 0 1 2 3 4 5 Cost (as % of GDP) ** The red/horizontal line indicates the baseline poverty gap index. 22 Figure 7: The poverty severity index also improves with a higher transfer value and lower coverage levels*** 11 Transfer amount per household per year % of median consumption of the poor (in 2015 US dolars) 50% 100% 10 150% 200% 9 Poverty Severity Index 8 7 6 5 4 0 1 2 3 4 5 Cost (as % of GDP) *** The red/horizontal line indicates the baseline poverty severity index. Source: World Bank staff calculations based on IOF 2008/09 However, to ensure broad support for a cash transfer program targeted to the poor, an optimal strategy would be to increase coverage and distribute a less generous transfer. With a budget of 1 percent of GDP, welfare gains would be maximized through having a higher transfer value at the expense of lower coverage, but this scenario may be less desirable for three main reasons. Firstly, less than a third of the total population would benefit, meaning that the majority of the population eligible to vote would be excluded, which could compromise support for a scaled-up cash transfer program. Secondly, given that consumption levels are relatively flat for quintiles 1 to 4 a significant share of the population may be vulnerable to falling into poverty, and would be excluded from a narrowly defined target population. By distributing a lower transfer value, which would cover the median consumption of the poor but not go beyond this, the welfare impacts on the depth and severity of poverty would be similar but 53 percent of the population would be included. Going further, distributing a transfer equivalent to half of the median consumption of the poor could mean that for a budget of 1 percent of GDP 98 percent of people below the food poverty line and 87 percent of the population are covered, with the poverty headcount ratio still dropping by four percentage points. Third, providing a generous transfer could potentially introduce distortions in the labor market if transfers are close to labor market wages. Therefore, distributing a less generous transfer which would reach a greater proportion of the poor is more likely to encourage support for a scaled-up cash transfer program. In sum, with a relatively modest budget, a scaled-up cash transfer program could be designed with a lower transfer value and still cover a significant proportion of the population to maximize welfare and ensure broad-ranging support. The simulation exercise shows that for a limited budget of around 1 23 percent of GDP, the poverty headcount ratio could potentially reduce by 4 percentage points. Increasing the budget to 2 percent of GDP could reduce poverty from 52 percent to 45 percent. These results prevail through having either: (i) a lower coverage level and more generous transfer, or (ii) wider coverage and a less generous transfer. Overall, to ensure a scaled-up cash transfer program has broad support from the population, a higher coverage level may be more appropriate as it would achieve very similar results in reducing the poverty headcount ratio, while including a large share of beneficiaries below the food poverty line.  Option 3: A Cash Transfer Program Targeted to the Poor with a Conditional Component The evidence shows that imposing conditions or co-responsibilities to receive cash transfers could help to promote specific behaviors. While distributing cash with no conditions can reduce poverty and improve broader development outcomes, in cases where conditions or co-responsibilities are imposed, it can be possible to promote specific behaviors such as attending school or completing health check- up’s (Adato and Hoddinott, 2010). For example, an impact evaluation undertaken in Malawi showed that imposing conditions contributed to better school enrollment, attendance, and test scores. However, the group that received a cash benefit without conditions also experienced improved school attendance (although not to the same extent) and also importantly, noted other positive effects such as delayed marriage and child-bearing (Baird et. al, 2010). The idea of a conditional cash transfer (CCT) is that while cash alleviates poverty today and compensates for the costs of children’s education, the conditions help guarantee that the children in the household will have better human capital when they become earners in their own right, and thus be more likely to escape the intergenerational transmission of poverty (Grosh et. al, 2008). Therefore, cash transfers with and without conditions have positive impacts, although conditions can improve specific and intended behaviors. The positive impacts of CCT programs have been well documented. A number of impact evaluations undertaken in Latin America provide evidence of the positive impacts of CCT programs in reducing poverty, inequality and improving the utilization of health and education services. Gains have been particularly impressive in countries where the school enrollment rates were relatively low before the CCT program was implemented (see box 6 for further information). Emerging evidence in Africa also highlights the broader positive outcomes of CCT programs. An evaluation of Kenya’s Cash Transfer Program for Orphans and Vulnerable Children demonstrated a number of positive impacts, where participating households have higher expenditures on food, health, and clothing, with significantly less spending on alcohol and tobacco. And in Tanzania an impact evaluation of the Community-Based Conditional Cash Transfer program showed that grant recipients were healthier and more likely to complete primary school. The cash grant was also used to invest in livestock, children’s shoes, insurance, and for the very poorest it was also possible to realize some modest savings (Evans et. al, 2014). 24 Box 6: Cash Transfer Programs in Latin America Reducing Poverty and Inequality, and improving the use of Social Services Between 1995 and 2010 poverty and inequality declined in Latin America, where redistributive policies including cash transfers are estimated to contribute to 34 percent of the decrease in poverty. Progress was particularly impressive in countries where the cash transfer value is relatively generous compared with consumption of poor households. For example, in Colombia the Familias en Acción program reduced the depth of poverty by almost 7 percentage points. In Nicaragua the Red de Protección Social program (RPS) reduced the poverty headcount index by 5-7 percentage points. CCT programs have contributed to a reduction in poverty more broadly at the national level. For example, the severity of poverty reduced in Mexico by 29 percent and in Brazil by 15 percent respectively. Cash transfers have contributed to a reduction in inequality. In Brazil, since 2001 the government increased the average transfer amount to the elderly and disabled and broadened the coverage of the Bolsa Familia CCT program. Even though these programs only represent 0.5 percent of household income, they are estimated to have contributed to a 10 percent decline in household income inequality. And in Mexico inequality declined by 4.2 percentage points from 0.5472 to 0.5052 between 1996 and 2008, where the Oportunidades cash transfer program is estimated to account for 18 percent of this change. CCT programs have led to significant and substantial increases in the use of services, particularly for school enrollment rates, especially in countries where enrollment rates were low before the program intervention. For example, school enrollment increased in Ecuador by 10.3 percentage points. In Nicaragua school enrollment increased by 12.8 percentage points, where the impact was greatest amongst the extreme poor. CCT programs have also had a positive impact on health center visits by children, where for example in Colombia there was an increase in children taken to growth and development monitoring by 22.8 percent, and in Honduras a child taken to a health center at least once a month increased by 20.2 percent. Furthermore, over the long-term, in some countries a portion of the transfers have been saved and invested in productive assets, enabling households to smooth consumption in the face of negative shocks. For example, in Mexico agricultural income increased by almost 10 percent after 18 months of receiving the cash transfers and in Nicaragua the RPS acted as a social safety net during the coffee crisis, by protecting income. Sources: Ferreira et. al, 2013, Fiszbein and Schady, 2009, Gertler et. al, 2012and Lustig et. al, 2011 The number of countries implementing CCT programs in SSA is increasing. Examples include Mali’s Bourse Maman and Nigeria’s Kano CCT for Girls’ Education which focus on improving educational attainment of vulnerable groups, Eritrea’s Results-Based Financing program which aims to improve health outcomes for mothers and children, and Nigeria’s In Care of the Poor CCT which aims to improve both education and health care. As many of these CCT programs are in a pilot phase the evidence base on program impact is gradually improving. In practice, countries can implement both conditional and unconditional cash transfer programs targeted to the poor. For example, in Kenya both conditional and unconditional cash transfer programs have been implemented. Bolivia is an example of a resource-rich country which uses a portion of gas 25 revenues to implement a combination of conditional and unconditional cash transfer programs, with positive effects on poverty reduction and development outcomes (see box 7). Box 7: Transferring gas to cash in Bolivia to reduce poverty and promote development outcomes In Bolivia, the success of Evo Morales’ re-election to office in October 2014 was partly seen to be due to the effectiveness of cash transfer programs. Both conditional transfers (to encourage school attendance and regular antenatal check-ups for pregnant women) and unconditional cash transfers for the elderly, as well as expansion of the pension system have contributed to a reduction in extreme poverty, from 38.2 percent in 2005 to 21.6 percent in 2012. A recent impact evaluation of the transfer to the elderly (Renta Dignidad or Old Age Bonus) undertaken by the government’s think tank showed that poverty among the elderly reduced from 53.6 percent to 40.1 percent. Source: IMF, 2013 and Unidad de Analisis de Politicas Sociales y Economicas, 2014 In the Mozambican context a CCT program focusing on children could be introduced as one component of the overall social protection system. Given that the majority of poor households have children, a CCT program would be well placed to focus on children and improve school attendance rates. While primary school enrollment has been increasing over time, children from poor households have a higher dropout rate meaning they are less likely to progress to secondary and tertiary education (World Bank, 2014a). A CCT could also help to improve the generosity of transfers received by the poorest households which are not currently covered by the PSSB. How a CCT program could be integrated with the existing cash transfer programs is discussed in further detail in Box 8. Box 8: Implementing an integrated cash transfer program in Mozambique It is suggested that one integrated scaled-up cash transfer program could be implemented, combining the basic transfer (the current PSSB) and the transfer for the public works program with a transfer targeted to households with children. The transfer for children could have one unconditional household child benefit and the other conditional on school attendance. The advantage of having an integrated program is that a poor household could be entitled to all the cash transfer programs if they meet the eligibility criteria, therefore building on the existing programs in place as well as addressing gaps in the current system. Even if this integrated scaled-up cash transfer program was to be implemented gradually it could be possible over time to reduce poverty by 11 percentage points. With the CCT component, it may take time to build sufficient capacity to enforce program conditions. In other countries, in the early stages of implementing a CCT the conditions are presented as highly encouraged behaviors, where penalties for non-compliance are gradually introduced. Source: World Bank staff projections A CCT program could help to address demand side constraints to accessing education but efforts are still needed to improve the quality of teaching. There are a number of factors that affect demand for education including direct costs such as fees and scholastic materials, having to work (estimated to contribute to half of children dropping out of secondary school), pregnancy and early marriage for girls, and migration to South Africa for boys (Fox et al, 2012). Therefore, while school enrollment is estimated to be close to 100 percent, around 56 percent of pupils are absent (World Bank, 2014b). Implementing a CCT could potentially help to improve school attendance, but is only likely to be 26 effective if combined with efforts to improve and increase the supply of education services. For example, teacher absenteeism is estimated to be around 45 percent and there are serious gaps in teachers’ knowledge and teaching ability (World Bank, 2014b). If the quality of education services improves, households could expect the returns to education to be higher which may increase responsiveness to a cash transfer conditional on attending school (Fiszbein and Schady, 2009). In sum, a conditional cash transfer component could be introduced as part of a scaled-up cash transfer program. By focusing on children it could be possible to address existing programmatic gaps, given that the majority of poor households have children. Even though Mozambique has made impressive progress in improving school enrollment, school attendance rates are low. Therefore, a CCT could help to address demand side constraints, but this should be combined with efforts to improve the quality of teaching. A CCT component could be introduced as part of an overall scaled-up cash transfer program, integrated with other unconditional transfers, which would mean that a household could benefit from different transfers if it was eligible. Such an approach would address existing program gaps and build on the current system. (iii) Addressing Potential Risks of a Scaled-up Cash Transfer Program There may be concerns that a scaled-up cash transfer program could have unintended impacts, but this is not supported by the evidence. While there may be concerns that cash transfer programs can create disincentives to work and have adverse effects on the economy, such claims are not supported by the evidence. Firstly, if a relatively modest transfer value is distributed to ensure a broader coverage of the population (as suggested earlier in section ii) potential risks of distortions to the labor market would be minimized.14 Secondly, international evidence on cash transfer programs suggests that labor market prospects and productivity can actually improve. For example, recipients of cash transfers in South Africa increased their employment rates faster, and had more rapid wage increases compared with households that did not receive cash grants. Furthermore, recipients of cash grants spent more on basic items such as food, fuel and housing, which in turn is associated with a reduction in hunger and better nutritional outcomes (Samson et al, 2004). In Mexico, beneficiaries of cash transfer programs invested more than 25 percent of the cash in productive activities, leading to positive effects on long-term consumption (Gertler et. al, 2012). In countries such as Zambia, Kenya and Malawi cash transfers also had positive effects on agricultural productivity (see box 2 for further details). Implementing and financing a wide-scale cash transfer program from resource revenues has a number of potential risks. Firstly, if substantial levels of resource revenues are distributed through cash transfers macroeconomic distortions may be created such as inflation and consumption booms that risk turning into busts. Secondly, if cash transfers are financed from earmarked resource revenues there is a risk that volatility of commodity prices will be reflected in unpredictability of cash transfer payments, which could lead to greater household uncertainty and vulnerability. Early experiments with cash 14 For example, if cash is distributed at a value which is half the consumption of the poor this would be only 8 percent of the agriculture minimum wage. Similarly a transfer value covering the median consumption of the poor would be 16 percent of the monthly agriculture minimum wage. These modest transfer values would be very unlikely to cause distortions to the labor market. 27 transfers in Mongolia showed this was the case where cash transfer payments to households were affected by unpredictability and uncertainty due to volatility (see box 9). It can be possible to mitigate the risks of financing from resource revenues through sound macroeconomic and fiscal management. The financing arrangements for a scaled-up cash transfer program should not be tied to earmarked resource revenues, as this inevitably implies the risk of volatility. Rather, the government should adopt an appropriate fiscal policy framework for managing resource revenues, including measures to address sustainability and volatility such as the introduction of a resource fund or fiscal rules.15 The cash transfer program should be allocated resources through the standard budget preparation process. If such measures are not undertaken, there is a risk that a separate financial management process could be created for cash transfers, which is not subject to transparency, accountability or controls mechanisms that apply to the regular budget process. Box 9: Universal distribution of cash transfers in Mongolia Mongolia established the Human Development Fund (HDF) in 2009 as part of efforts to manage the resource boom. The HDF mission is to channel part of resource revenues to the population through universal cash transfers and social benefits (tuition fees, pensions, healthcare, education and housing). The law however does not specify the size of revenues to channel to the HDF nor the size of payments it can make. This lack of predictability and transparency defeats the purpose of using a cash transfer system to avoid political capture. Indeed, competition amongst politicians led to wide variations in the type and distribution of benefits, which were subject to volatility of copper prices and contributed to inflation. The cash benefits were not fiscally sustainable and contributed to uncertainty over household income levels. To improve overall fiscal discipline and stability, in 2010 Mongolia approved a fiscal stability law (FSL) and three complementary fiscal rules, which should help contain potential macroeconomic risks associated with disbursements from the HDF. Sources: Isakova, A et. al., 2012 and Fritz, 2013 Other concerns over a scaled-up cash transfer program can be addressed through program design and sound economic policy. One potential risk is that injecting cash into the economy through transfer payments could stoke inflationary pressures. To address this risk it will be important for the design of the cash transfer program to appropriately cap the transfer value. Given volatility in food prices (which can particularly affect the poor) inflation will have to be managed as part of the government’s overall monetary policy. (iv) Preparing for a Scaled-up Cash Transfer Program To ensure the poor benefit from a scaled-up cash transfer program it will be important to strengthen capacity to implement the targeting system. Targeting errors can arise because it is relatively challenging to identify the very poorest as the distribution of poverty across the first four quintiles is relatively flat. Also, as a large share of the population is involved in the informal sector and subsistence farming it is challenging to identify beneficiaries on the basis of consumption (World Bank, 2012). To 15 See a separate policy note ‘Generating Sustainable Wealth from Mozambique’s Natural Resource Boom’ for more details, World Bank, 2014c 28 help address these constraints a common targeting system has been developed which includes: (i) geographical targeting through the use of highly disaggregated poverty maps; (ii) community based targeting using local councils; and (iii) a proxy means test to minimize inclusion errors and ensure that only the poor are enrolled in the program. However, INAS needs to invest in its operational capacity to correctly implement these targeting methods. In particular, it is important for INAS to develop a grievance and redress mechanism that will allow for feedback from communities on beneficiary selection. Establishing an effective Management Information System (MIS) is critical to the program’s success and its potential to scale up. INAS is currently developing a comprehensive MIS, where all databases generated by the different program units will produce reports on resources used and outputs achieved. All beneficiaries of the scaled-up program would be recorded in a single registry as a module of the MIS. The single registry of beneficiaries will make it possible to monitor the progress of households that receive support from the social protection system. The single registry will also be linked with other databases from social sectors enabling the referral of beneficiaries to other programs and improving coordination. The registry and the MIS will be used for the enrollment of beneficiaries (beneficiaries will be given ID cards and a tax number linked to the civil registry) to produce payroll lists and to generate progress reports. Another area for improvement is the payment system. At present, cash transfer payments are done manually with no reconciliation, leading to high fiduciary risks. To address this, INAS is currently in the process of outsourcing the payment system for all cash transfer programs. In the initial stage it is expected that the payment system will not be able to reach all areas in Mozambique due to information technology challenges, such as lack of mobile and internet network, and unavailability of energy. However, INAS should ensure that the combination of outsourced and manual payments will allow for timely and predictable payments to beneficiaries. Moving forward, it will also be important to develop adequate monitoring and evaluation mechanisms. Such measures should address progress with the scaled-up program, including whether the intended beneficiaries are receiving payments. A proper M&E system should combine a number of features including: (i) reports on outputs and resources used from a fully operational MIS; (ii) process and impact evaluations; and (iii) periodical control systems including internal audits, quality control mechanisms and spot checks. An effective monitoring system might require costly investments in information technology and can take time to develop and mature. Evaluations will complement the monitoring system and are essential to be able to improve and adapt the program, inform stakeholders on the program’s performance and draw lessons for other programs. Furthermore, if a CCT component is implemented, it will also be important to monitor compliance with conditions and ensure the information is fully integrated in the MIS (Fiszbein and Schady, 2009). A scaled-up cash transfer program will require additional operational capacity. During the last years, INAS has invested in the development of operational tools (such as the targeting system, MIS and payment system) that will facilitate the scale-up of existing programs. However increasing coverage in 29 an efficient and effective manner will only be possible through an increase in human capacity, especially at the province and district levels. It is expected that the introduction of information technology and outsourcing of key activities such as beneficiary registration and payments will facilitate the implementation of social protection programs. However, the allocation of additional human resources at district and delegation level will be essential for ensuring that cash transfers are implemented effectively and that proper monitoring and evaluation of the programs is undertaken. In the initial phase of the scale up, operational and administrative costs are expected to be higher, which needs to be considered in the government’s planning and budgeting processes. For the specific case of the CCT program, clear institutional linkages need to be established between the social protection, health and education sectors. Sustained impacts of the scaled-up program on poverty and human capital will require long term interventions. Cash transfer programs are expected to mitigate the impact of short term poverty, shocks and reduce extreme poverty. However, sustained reduction of extreme poverty and enhanced capacity of poor households to better manage risks require that those programs operate for substantially longer periods of time. Beyond the short term cash benefit impact, the scaled-up program should also involve medium to long term processes to enhance beneficiaries’ productivity and management of future risks and shocks. The scaled-up program should be part of a broader strategy for poverty reduction and as such needs to be complemented by other government measures to ensure that households leaving the program will not fall again into extreme poverty16. Conclusion Mozambique’s impressive economic growth has not translated into significant poverty reduction in recent years, a trend that is likely to continue as extractive industries increasingly contribute to growth. In the period following the end of the civil war economic growth, supported by agricultural activity, was associated with a significant decline in poverty. However, while economic growth continues to be impressive the pace of poverty reduction has slowed, coinciding with capital intensive megaprojects increasingly contributing to growth. Evidence from resource-rich countries in the region also confirms that economic growth does not always result in poverty reduction. Furthermore, Mozambique has relatively higher inequality levels compared with other countries in the region, which is also likely to affect the extent to which economic growth reduces poverty. Therefore, as the extractive industries grow in importance, specific policy measures to directly reduce poverty are called for. Implementing a scaled-up cash transfer program could be a relatively quick and cost effective way of reducing poverty in Mozambique’s future resource-rich environment. In recent years, improvements have been made to Mozambique’s social protection system, where the institutional environment has been strengthened, and increased domestic financing signals the government’s commitment to the 16 International experience shows that successful programs such as Brazil’s Bolsa Familia and Mexico’s Oportunidades not only have steadily increased program beneficiaries over the last 10-15 years, but also that a high proportion of beneficiary households have been in the program for long periods of time. 30 sector. However, coverage levels are low, benefiting only 14 percent of the poor. Implementing a scaled-up cash transfer program could build on these achievements and enable Mozambique to significantly reduce poverty over a relatively short period of time. International experience shows that other countries have used cash transfer programs to reduce poverty, improve welfare and promote development outcomes. Compared to spending on priority areas such as health, education and infrastructure which will yield results over the medium to long term, a scaled-up cash transfer program could reduce poverty in a relatively short space of time. Furthermore, in comparison with other policy interventions, a scaled-up cash transfer program is a relatively more cost effective policy measure that could potentially benefit a broad section of the population. One option could be to implement a cash transfer program universally, but such an approach is unlikely to be feasible in Mozambique. A universal cash transfer would mean that the entire population benefits, thereby covering all of the poor, and would also provide a way of sharing resource wealth more broadly. With an estimated cost of 5 percent of GDP, a universal cash transfer distributed at the most generous transfer value could reduce poverty by 16 percentage points. However, given other spending priorities this option is unlikely to be affordable or the most cost-effective way of reducing poverty. Therefore, while a universal cash program has certain advantages; it is unlikely to be the best choice for Mozambique when resource revenues start to flow. A second option which would be more feasible in the Mozambican context would be to target the distribution of a scaled-up cash transfer program to the poor. For a given fiscal envelope, decisions would need to be made on whether poverty reduction and welfare can be maximized through either having a relatively generous transfer value, or by covering a larger share of the poor. Results from a simulation exercise show that with a budget of around 1 percent of GDP (assuming Mozambique was already in receipt of resource revenues) it could be possible to reduce poverty by 4 percentage points through either having (i) a low transfer value and high levels of coverage, or (ii) a high transfer value and low levels of coverage, where both options would yield the same result. Increasing the budget to 2 percent of GDP would reduce poverty by 7 percentage points. It is also important to factor in political economy considerations, where distributing a slightly less generous transfer would enable a greater share of the population to be covered, which would have similar effects in reducing the poverty headcount ratio while ensuring broader support from the population. A further consideration could be given to introducing a conditional component to the overall cash transfer program, which could help to address program gaps and important human capital concerns. Through introducing a conditional cash transfer component as part of a scaled-up cash transfer program it could be possible to improve school attendance for children. While school enrollment rates are high, pupils are absent for around 56 percent of the time. By targeting children it could be possible to address programmatic gaps in the current social protection system since the majority of poor households have children. While distributing cash addresses poverty in the short-term, a CCT component aimed to improve school attendance could help to improve the longer-term prospects of children. It is important to mitigate potential risks from implementing a scaled-up cash transfer program through program design, and sound macroeconomic and fiscal management. Concerns are sometimes 31 put forward that cash transfers can have unintended consequences on the labor market and the economy. On the contrary, evidence from other countries suggests that recipients of cash transfers are more likely to participate in the labor market and engage in productive agricultural activities. There may be potential risks to financing a scaled-up cash transfer program from earmarked resource revenues, as this could risk volatility. Rather, the cash transfer program should be allocated resources through the standard budget preparation process. It will also be important to ensure the transfer value is capped at a modest level to avoid inflationary pressures. Alongside this, efforts to prepare for a scaled-up cash transfer program should include continued strengthening of the overall social protection system through improving mechanisms for targeting the poor, payments, and developing a management information system. In sum, implementing a scaled-up cash transfer program in Mozambique’s future resource-rich environment is a relatively cost-effective policy option for reducing poverty and improving welfare over a relatively short period of time. In a resource-rich environment implementing a scaled-up cash transfer program would be an affordable and rapid way of reducing poverty, where a program costing 1 percent of GDP could reduce poverty by four percentage points in one year. Compared to other policy measures, a scaled-up cash transfer program would be a more cost-effective way of reducing poverty. This policy initiative could be part of the government’s broader strategy toward ensuring impressive economic growth is having an impact on poverty reduction, while sharing the benefits of resource wealth more broadly among the population. 32 Annex 1: Coverage and Leakage Rates Total Cost as a % of GDP Eligibility Transfer amount per person per year: % Impact on Poverty % of Pop. cut-off % of the Population targeted of median consumption of the poor (in (2009 National Poverty Rate: 52.08%) Targeted point 2015 US dollars) Probability of being Below Coverage Coverage Coverage US$ US$ US$ US$ US$ US$ US$ US$ poor Coverage food of of of 85.50 171.00 256.00 342.00 85.50 171.00 256.00 342.00 Leakage (greater of poor poverty bottom bottom bottom (50% of (100% (150% (200% (50% of (100% (150% (200% than or line 20% 40% 60% p50) of p50) of p50) of p50) p50) of p50) of p50) of p50) equal to) 0 100.00 47.92 100.00 100.00 100.00 100.00 1.28 2.57 3.85 5.13 48.36 44.69 40.81 36.56 100.00 5 98.86 43.80 99.03 99.27 98.95 98.54 1.12 2.25 3.37 4.49 48.48 44.95 41.18 36.98 91.61 10 96.89 41.87 97.76 98.73 97.55 95.91 1.04 2.07 3.11 4.14 48.91 45.52 41.82 37.85 86.80 15 95.07 39.83 96.05 97.60 95.70 93.42 0.96 1.91 2.87 3.83 49.02 45.84 42.34 38.48 82.29 20 92.09 37.77 93.34 95.69 92.57 89.63 0.87 1.74 2.62 3.49 49.23 46.20 42.99 39.35 77.07 25 87.67 35.84 89.11 91.93 87.97 84.75 0.78 1.56 2.34 3.12 49.53 46.87 43.87 40.50 71.15 30 82.87 33.99 84.95 88.21 83.35 79.52 0.70 1.39 2.09 2.78 49.77 47.62 44.92 42.03 65.38 35 77.90 31.98 80.04 84.58 78.52 74.11 0.61 1.23 1.84 2.46 50.04 48.04 45.66 42.98 59.64 40 71.60 30.30 73.87 78.72 72.52 67.85 0.54 1.08 1.62 2.15 50.36 48.50 46.53 44.14 53.50 45 63.93 29.27 66.71 72.13 65.18 60.16 0.46 0.92 1.39 1.85 50.75 49.16 47.54 45.52 47.07 50 57.55 27.71 60.36 65.78 58.72 53.72 0.39 0.79 1.18 1.58 50.91 49.57 48.21 46.56 41.46 55 49.45 26.21 52.33 57.33 50.29 45.65 0.32 0.64 0.96 1.28 51.14 50.05 49.09 47.83 34.90 60 41.32 24.80 44.31 47.54 41.85 38.02 0.26 0.51 0.77 1.02 51.35 50.61 49.82 48.79 28.61 65 33.28 23.13 35.59 38.99 33.24 30.30 0.19 0.39 0.58 0.77 51.63 51.00 50.37 49.60 22.55 70 25.77 22.77 27.38 30.06 25.48 23.37 0.15 0.29 0.44 0.58 51.74 51.25 50.84 50.23 17.38 75 18.15 21.17 19.34 22.25 17.99 16.25 0.10 0.19 0.29 0.38 51.89 51.61 51.33 50.93 11.99 80 12.04 19.38 13.28 15.41 12.26 10.54 0.06 0.12 0.18 0.23 51.98 51.85 51.74 51.55 7.78 85 7.04 16.52 8.03 9.99 7.47 6.15 0.03 0.06 0.09 0.13 52.07 52.01 51.98 51.86 4.39 90 2.93 …* 3.61 4.90 3.29 2.55 0.01 0.03 0.04 0.06 52.08 52.06 52.06 52.05 2.04 95 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 52.08 52.08 52.08 52.08 0.00 100 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 52.08 52.08 52.08 52.08 0.00  Source: WB staff calculation based on IOF 2008/09  *Note: Data was omitted due to the low number of observations (81 cases) affecting the precision of the estimate 33 References Adato, M. and Hoddinott, J. (2010) “Conditional Cash Transfers in Latin America,” International Food Policy Research Institute Alfani, F., Azzarri, C., d’Errico M., and Molini, V. (2012) “Poverty in Mozambique: New Evidence from Recent Household Surveys,” Policy Research Working Paper 6217, Washington, DC: World Bank Anderson, J. (2002) “The Alaska Permanent Fund: Politics and Trust,” Public Budgeting and Finance, Vol. 22 No.2 pp. 57-68, Blackwell Publishers Inc. Baird, S., Ferreira, F. H. G., Özler, B., Woolcock, M. (2013) “Relative Effectiveness of Conditional and Unconditional Cash Transfers for Schooling Outcomes in Developing Countries: A Systematic Review,” Campbell Systematic Reviews Birdsall N. and A. Subramanian (2004) “Saving Iraq from Its Oil”, Foreign Affairs, V83 No 4 pp 77-89 Bourguignon, F. 2003 “The Growth Elasticity of Poverty Reduction: Explaining Heterogeneity across Countries and Time Periods,” in: T. Eicher and S. Turnovsky, eds. Inequality and growth. Theory and Policy Implications. Cambridge: The MIT Press Chhibber, A. and G. Nayyar, “Pro-Poor Growth: Explaining the Cross-country Variation in the Growth Elasticity of Poverty,” Brooks World Poverty Institute Working Paper 14, 2007 Cunha, N., Pellerano, L., Mueller, J., Lledo, V., Xiao, Y., Gitton, P. (2013) “Towards a Mozambican social protection floor: consolidating a comprehensive social protection system in Mozambique: analysis of policy alternatives and costs,” International Labour Office, Geneva Department of Social Development, South African Social Security Agency and UNICEF (2012) “The South African Child Support Grant Impact Assessment: Evidence from a survey of children, adolescents and their households,” Pretoria: UNICEF South Africa Devarajan, S., Tuan Minh, L., and Raballand, G. (2010) “Increasing Public Expenditure Efficiency in Oil- rich Economies: A proposal,” Policy Research Working Paper, World Bank: Washington DC. Devarajan, S., and Giugale, M. (2013) “The Case for Direct Transfer of Resource Revenues in Africa,” Center for Global Development Working Paper 333, Washington, DC: World Bank Evans, David K., Hausladen, S,. Kosec, K., and Reese, N., (2014) “Community-Based Conditional Cash Transfers in Tanzania: Results from a Randomized Trial. World Bank Study,” Washington, DC: World Bank Ferreira, F., H., G., Messina, J., Rigolini, J., López-Calva, L., Lugo, M., A., and Vakis, R., (2013) “Economic Mobility and the Rise of the Latin American Middle Class,” Washington, DC: World Bank 34 Fiszbein, A. and Schady, N., (2009) “Conditional Cash Transfers Reducing Present and Future Poverty,” World Bank, Washington DC. Fox, L., Santibanez, L., Nguyen, V., and Andre, P. (2012) “Education Reform in Mozambique: Lessons and Challenges,” World Bank, Washington DC. Fritz, V. (2014) “Dealing with a Resource Economy Shock: Political Economy Analysis and Its Impacts in Mongolia,” in Problem Driven Political Economy Analysis: The World Bank’s Experience eds. Fritz, V., Levy, B., and Ort, R. World Bank: Washington DC. Garcia, M., and Moore, C. (2012) “The Cash Dividend: The Rise of Cash Transfer Programs in Sub- Saharan Africa,” World Bank: Washington DC. Gertler, P., Martinez, S. and Rubio-Codina, M., 2006. Investing cash transfers to raise long-term living standards, Report WPS 3994-IE, Washington DC: World Bank Gillies, Alexandra (2010) “Giving Money Away? The Politics of Direct Distribution in Resource-Rich States,” Center for Global Development Working Paper 231, Washington DC. Goldsmith, S. (2002) “The Alaska Permanent Fund Dividend: An Experiment in Wealth Distribution,” The Basic Income European Network, Geneva Gupta, S., Segura-Ubiergo, A., and Flores E., (2014) “Direct Distribution of Revenues: Worth Considering?” IMF Staff Discussion Note: Washington DC. Haushofer, J. and Shapiro, J. (2013) “Policy Brief: Impacts of Unconditional Cash Transfers,” Abdul Latif Jameel Poverty Action Lab (J-PAL), Massachusetts Institute of Technology IMF (2014) “Bolivia: Staff Report for the 2013 Article IV Consultation,” IMF Country Report No. 14/36: Washington DC. Isakova, A., A. Plekhanov, and J. Zettelmeyer (2012) “Managing Mongolia’s Resource Boom,” Working Paper No. 138 (London: European Bank for Reconstruction and Development) Jones, S., Tarp, F. 2012. “Jobs and Welfare in Mozambique. Country case study for the 2013 World Development Report,” Copenhagen: UNU-WIDER Lustig, Nora, López-Calva, L., and Ortiz-Juarez, E., (2011) “The Decline in Inequality in Latin America: How Much, Since When and Why?” Economics Working Paper 1118, Tulane University, New Orleans McGuirk, E., Rajaram, A. and Giugale, M. (2014) “On the Political Economy of Direct Dividend Transfers in Resource-Rich Countries: A Theoretical Consideration,” unpublished manuscript Miller, C., Tsoka, M. and Reichert, K., (2008) “Impact Evaluation Report: External Evaluation of the Mchinji Social Cash Transfer Pilot,” Boston University and Zomba: Centre for Social Research, University of Malawi 35 Moore, M. (2004) “Revenues, State Formation and the Quality of Governance in Developing Countries,” International Political Science Review 25 (3): 297-319 Moss, T. and Majerowicz, S. (2013) “Oil-to-Cash Won’t Work Here: Ten Common Objections,” Center for Global Development Policy Paper 024, Washington DC. Moss, T. and Young, L. (2009) “Saving Ghana from Its Oil: The Case for Direct Cash Distribution,” Center for Global Development Working Paper 186, Washington DC. Repulic of Mozambique: Instituto Nacional de Acção Social (2014) “Economic and Social Plan 2014,” Maputo: INAS Sala-i-Martin, Xavier and A. Subramanian, 2003, “Addressing the Natural Resource Curse: An Illustration from Nigeria” IMF Working Paper No 03/139, International Monetary Fund: Washington DC Samson, M., Lee, U., Ndlebe, A., Mac Quene, K., van Niekerk, I., Gandhi, V., Harigaya, T. and Abrahams, C., (2004) “The social and economic impact of South Africa’s Social Security System,” Commissioned by the Department of Social Development, Cape Town: EPRI Sandbu, M. (2005) “Natural Wealth Accounts: A Proposal for Alleviating the Natural Resource Curse,” Philadelphia, Wharton School of Business: University of Pennsylvania Seidenfeld, D., Handa, S. and Tembo, G. (2013) “Zambia Social Cash Transfer Scheme: 24-Month Impact Report for the Child Grant Programme,” American Institute for Research: Washington DC. Selvester, K., Fidalgo, L., Ismael, C. 2014. “Nutrição, Saúde Materna e Protecção em programas de Protecção Social em Moçambique,” Maputo: OIT The Kenya CT-OVC Evaluation Team (2012) The impact of the Kenya Cash Transfer Program for Orphans and Vulnerable Children on household spending, Journal of Development Effectiveness, 4:1, 9-37 UK Department for International Development (2011) “Cash Transfers Literature Review,” DFID: UK Unidad de Analisis de Politicas Sociales y Economicas (2014) “Impact Evaluation of Renta Dignidad,” La Paz: Bolivia World Bank. (2012) “Mozambique Social Protection Assessment: Review of Social Assistance Programs and Social Protection Expenditures,” Washington DC: World Bank World Bank (2013a) “Project appraisal document on a proposed credit to the Republic of Mozambique for a Social Protection Project,” Washington DC: World Bank World Bank (2013b) “Africa’s Pulse,” Volume 7, April 2013, Washington DC: World Bank World Bank (2014a) “Mozambique Public Expenditure Review. Addressing the Challenges of Today, Seizing the Opportunities of Tomorrow,” Washington DC: World Bank 36 World Bank. (2014b). “Mozambique Service Delivery Indicators,” Washington DC: World Bank World Bank (2014c) “Generating Sustainable Wealth from Mozambique’s Natural Resource Boom,” World Bank Policy Note 37