Macroeconomic Developments and Prospects for Low-Income Countries—2024—Online Annexes
Author:
International Monetary Fund
Search for other papers by International Monetary Fund in
Current site
Google Scholar
Close
and
World Bank
Search for other papers by World Bank in
Current site
Google Scholar
Close

MACROECONOMIC DEVELOPMENTS AND PROSPECTS FOR LOW-INCOME COUNTRIES—2024—ONLINE ANNEXES

Abstract

MACROECONOMIC DEVELOPMENTS AND PROSPECTS FOR LOW-INCOME COUNTRIES—2024—ONLINE ANNEXES

Title page

MACROECONOMIC DEVELOPMENTS AND PROSPECTS FOR LOW-INCOME COUNTRIES—2024—ONLINE ANNEXES

February 29, 2024

Contents

  • NEW METHODOLOGY FOR ESTIMATING FINANCING NEEDS

  • BOXES

  • 1. FAD’s SDG Costing Exercise and Financing Tool

  • 2. Defining Expenditure Ceilings for the Annual SDG-Related Public Expenditure

  • TABLES

  • 1. Low-Income Countries: Estimation of Additional Financing Needs to Meet the SDGs, Alternative Specifications of the Model

  • 2. Low-Income Countries: Estimation of Additional Financing Needs to Meet the SDGs6

  • 3. Possible Sources of Financing

  • CASE STUDIES: STRENGTHENING SOCIAL SAFETY NETS IN LOW INCOME COUNTRIES

  • A. Ghana

  • B. Mozambique

  • C. Tanzania

  • D. Uganda

  • E. Zambia

  • FIGURES

  • 1. Coverage and Targeting of Main SSN Programs in Ghana

  • 2. Adequacy of Benefits: Main SSN Programs in Ghana

  • 3. Coverage and Targeting of the BSSP Program in Mozambique

  • 4. Adequacy of Benefits: The BSSP Program in Mozambique

  • 5. Coverage and Targeting of the PSSN Program in Tanzania

  • 6. Adequacy of Benefits: The PSSN Program in Tanzania

  • 7. Coverage and Targeting of the SCG Program in Uganda

  • 8. Adequacy of Benefits: The SCG Program in Uganda

  • 9. Coverage and Targeting of Main SSN Programs in Zambia

  • 10. Adequacy of Main SSN Programs in Zambia

  • TABLES

  • 1. Poverty Reduction of Main SSN Programs

  • 2. Characteristics and Poverty Reduction Reform Scenarios

  • 3. Poverty Reduction of the BSSP Program

  • 4. Characteristics and Poverty Reduction Reform Scenarios

  • 5. Poverty Reduction of the PSSP Program

  • 6. Characteristics and Poverty Reduction Reform Scenarios

  • 7. Poverty Reduction of the SCG Program

  • 8. Characteristics and Poverty Reduction Reform Scenarios

  • 9. Poverty Reduction of Main SSN Programs

  • 10. Characteristics and Poverty Impact Reform Scenarios

New Methodology for Estimating Financing Needs

This annex explains a new methodology to estimate low-income countries’ (LICs) additional financing needs, taking into account the development agenda as informed by the United Nation’s Sustainable Development Goals (SDGs) beyond needs linked to foster macroeconomic stability.

1. Estimating the financing needs of low-income countries over the medium term is difficult, as there is no generally accepted methodology. At the Fund, consistent with its mandate, the emphasis in such analysis has been traditionally placed on the needs emanating from restoring/maintaining macroeconomic stability. This has typically been approximated by the financing required to cover external Gross Financing Needs, which could vary by country, under current or program policies.1 At the current juncture, these needs reside in the US$817 billion range for the period 2024–28 (Table 2). In addition to these baseline needs, and taking a broader perspective, recent analysis has included needs to (i) further increase international reserve buffers to a common level judged as “adequate” (ii) cope with the COVID-19 pandemic; and (iii) regain a path of growth convergence with more advanced economies (2022 LIC report).2 These additional financing needs were estimated at US$437 billion in the 2022 LIC report for the period 2022–26. Given financing needs specifically related to COVID-19 have effectively disappeared for most countries, updated estimates using the same methodology would yield US$439 billion (Table 2 Total additional Needs Approach A). Even this broader approach, however, may significantly underestimate the actual needs in LICs. For example, in Gaspar, et al., 2019, and Carapella, et al., 2023, the Fund’s Fiscal Affairs Department (FAD) derived an annual need of about US$500 billion in 2030 to meet the SDGs across 49 low-income countries by 2030.

2. This annex presents a new approach that incorporates elements of SDG costing to place a larger focus on development needs to estimate additional financing needs for LICs. Specifically, the proposed approach draws on the recent work in FAD on SDG costing (see Gaspar et al. 2019; Benedek et al. 2021) focusing on needs in education, health, roads, electricity, and water and sanitation, using nominal costs as exogenous inputs in the context of a dynamic macroeconomic framework (Box 1).

3. To increase the realism of the estimates, our approach includes a fiscal multiplier effect on growth and thus allows growth to be endogenously determined.3 This is a critical innovation as the higher growth generated by increased spending lifts fiscal revenues over time and thus reduces external financing needs compared to a static exercise.4 To preserve realism, we also cap the increase in the annual growth rate to 5.1 percent (one standard deviation above the [average] real growth rate for LICs over the period 1999–2019).

4. To avoid unrealistic estimates both from an absorptive capacity and a financing perspective, the approach imposes feasibility constraints on the scaling-up of spending. This is necessary as the unconstrained needs of reaching the SDGs over a relatively short time span are staggering (see Text Figure). For example, reaching the SDGs fully by 2030 would entail scaling up investment, education and health spending by some 19 percent of GDP for the median LIC. We thus moved the target date for achieving the SDGs further into the future (to 2040 for illustrative purpose) to gain more time for achieving those. However, even assuming a linear path of spending increases to achieve the levels consistent with that target would entail continuous spending growth at a pace never seen in LIC history (Table 1, rows 1 and 2). To derive a more realistic spending path, we imposed absorptive capacity constraints on the maximum possible annual and 5-year growth of spending, with the ceilings depending on the initial level of total public expenditure across countries (Box 2; Table 1, row3).5,6

uA002fig01

Estimates of Additional Spending for SDGs Under Alternative Approaches

(In millions of U.S. dollars)

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates
Table 1.

Low-Income Countries: Estimation of Additional Financing Needs to Meet the SDGs, Alternative Specifications of the Model

(In billions of U.S. dollars, based on Oct 2023 WEO)

article image
Sources: Fund staff estimates.

5. Our best estimates suggest US$527 billion in additional financing needs for the public sector over the period 2024–28 to progress towards the SDGs.7 Table 1 (row 4) reports the annual breakdown, while annexes III and IV show country-specific results. The required median annual additional expenditure for all LICs is 4.3 percent of GDP. In terms of shares of GDP, the level of variation across the sample of LICs is relatively low. However, as one would expect, there is a high degree of concentration in terms of the share of total needs in nominal US dollars when the LIC sample of countries is broken down by institutional category (i.e., Frontier markets, FCS, Small-States, and others not included in the previous categories).8 With the total needs for frontier markets adding to US$167 billion for the period 2024–28 (or 32 percent of the total LIC additional needs), with a median of $9.5 billion. Needs for FCS total US$187 billion, for the same period, but with a much lower median of $3.3 billion with Ethiopia accounting for about a third of the amount (US$58 billion), whereas the median needs of small states is around US$300 million for the same period.9 Needs for the group of other LICs add to $169 billion, with a median need of US$3.8 billion, with Bangladesh accounting for around 60 percent of the total,

uA002fig02

Additional Spending Towards SDG by Country’s Institutional Category

(In billions of U.S. dollars)

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates.

6. Total additional financing needs are higher. Total financing needs for LICs resulting from the sum of SDG-related spending and the need to rebuild external reserve buffers would be US$557 billion over the period 2024–28 (Table 2, Total Additional Needs Approach B).10 This is a 120 billion increase compared to the US$437 billion estimated in 2022. The increase in financing needs would still be sizeable even if the same methodology developed in the 2022 report is applied to today’s data.11,12

Table 2.

Low-Income Countries: Estimation of Additional Financing Needs to Meet the SDGs

(In billions of U.S. dollars, based on Oct 2023 WEO)

article image
Sources: Fund staff estimates.

7. A mix of domestic revenue mobilization (DRM) and debt financing will be needed to cover the projected needs. DRM measures could help cover as much as US$292 billion of the projected financing needs, but significant domestic and external public financing would need to complement those efforts.13 As Table 3 shows, revenue efforts should start immediately. Other financing could come from reprioritizing spending towards SDGs. The Fund’s concessional lending under the PRGT would only cover a small part of these needs, consistent with its mandate as a BoP lender.

Table 3.

Possible Sources of Financing

(In billions of U.S. dollars)

article image
Sources: Fund staff estimates.

8. Given the limited capacity of the public sector in LICs, a major role in the financing of progress towards the SDGs would also fall on the private sector. This result is consistent with the efforts of LIC governments, supported by their external partners, to improve the environment for doing business and strengthening the role of the private sector as a key engine of growth in LICs. Specifically, the private sector would need to close a residual gap of some US$258 billion over 2024–28 (Annex II Table 1, row 4).14

FAD’s SDG Costing Exercise and Financing Tool

The goal of the SDG costing (Gaspar et al. 2019) exercise is to estimate the additional annual spending required for meaningful progress on the SDGs in four key areas. The resulting estimates denote additional annual spending in 2030, relative to a baseline of current spending to GDP in these sectors. The estimates are available for a sample of 155 countries: 49 low-income developing countries, 72 emerging market economies, and 34 advanced economies.

This study finds that delivering the SDG agenda for Low Income and Developing countries will require an annual spending flow for education and health expenditure in year 2030 of US$0.2 trillion; and an additional “annualized stock” of infrastructure in roads, electricity and water and sanitation estimated at US$0.3 trillion by year 2030. The total amount for both education and health and infrastructure in year 2030 is US$ 0.5 trillion.

For physical capital, additional spending is expressed in percentage points of 2030 GDP, and corresponds to the annualized spending (assumed to be a fixed level in percent of GDP) required to close infrastructure gaps between 2019 and 2030. Note that 2030 GDP, is expressed in constant 2020 dollars. After 2030, education and health spending would recur, whereas infrastructure spending would be expected to decline to cover depreciation of the capital stock built through 2030. To translate this exercise into required financing needs over any given time period, properly discounted values of infrastructure spending should be added to annual health and education spending flows.

The FAD tool framework and calculation macro routines were structured to calculate a path of expenditure consistent with the results of the FAD costing exercise. 1/ The FAD financing tool assumes that both the required recurrent spending level for education and health, and an appropriately discounted annualized fixed level of infrastructure spending (both in percent of GDP), are already reached in the first year of projections, before continuing to gradually increase in nominal terms in line with nominal GDP growth. While the tool offers a variety of customizations options, the application of this approach on its default set up resulted in an unrealistic increase in nominal and real spending in the initial year of projections.

1/ The FAD costing exercise expresses SDG financing needs as a recurrent education and health spending, and an annualized amount (fixed in percent of GDP) for infrastructure spending, both in percent of 2030 GDP in 2020 constant dollars (see Annex I).

Defining Expenditure Ceilings for the Annual SDG-Related Public Expenditure

The estimates derived from the SDG costing tool do not consider any limits in terms of absorptive or institutional capacity that would limit the execution of public sector in the sample countries. Imposing country-by-country constraints is beyond the scope of this exercise; we propose instead two common, alternative annual public expenditure limits based on: (i) the maximum annual increase that a public sector in a LIC country could execute; and (ii) the maximum cumulative increase that could be sustained over a five-year period.

Boosting expenditures in countries with a smaller public sector (as measured by the total amount of public expenditures to GDP) poses greater implementation challenges. First because the proportional increase would be greater relative to its current size (likely generating undesired macroeconomic consequences), and second, because smaller-sized public sectors would likely have weaker capacity to manage and execute larger amounts of public expenditure. Because of this, we first compute quartiles based on the level of total expenditure to GDP on LICs in year 2019 (pre-pandemic). Then we compute the 80th percentiles of the annual increase (change) in public expenditure within each of the previously defined quartiles over the period 1999–2019 using WEO data. An examination of the results shows that the 80th percentile of the one-year change in public expenditure (only counting years in which change is positive) in countries in the first quartile is 3.0 percent of GDP, while the analogous statistic for countries in the fourth quartile is 8.2 percent of GDP (the presence of SIDS makes spending growth more volatile in the fourth quarter; Annex II). Likewise, the maximum cumulative increase of total public expenditure over a 5-year period is significantly larger for countries in the fourth quartile.

Based on this, we use the one-year cap as the binding constraint for the first year, while subsequently, for years 2–4, spending grows to the five-year maximum. Because the caps result on a lower level of public expenditure than required to reach the target stock infrastructure SDGs, the resulting gap would need to be added to the gap of the subsequent five-year period. At year 6, we allow public spending to grow again by the annual cap, and increase to a new 5-year cap (incremental to the previous cap) from year 7. The process is repeated every 5 years, until public spending reaches a level required in the unconstrained estimations (with no expenditure caps) and all the additional gaps have been closed.

Change in Total Public Expenditure, 80th percentile (within Quartile)

(Percent Change)

article image

Case Studies: Strengthening Social Safety Nets in Low Income Countries1

This Annex includes four case studies analyzing the functioning of social safety nets in low-income countries using household survey data. The analysis focuses on Ghana, Mozambique, Tanzania, Uganda, and Zambia. Each case study describes selected social safety net programs, including an assessment of coverage, adequacy, and incidence, and examines possible reform scenarios for the main cash transfer program, by increasing coverage and adequacy, aiming at strengthening the impact of social safety nets on poverty alleviation. For some countries, some social assistance programs are not included in the analysis (i.e., Ghana), reflecting limitations from the household survey. The analysis employs the tax-benefit microsimulation model SOUTHMOD, developed by UNU-WIDER in collaboration with partners, considering existing policy rules and using representative survey microdata. The analysis is complemented by data from the World Bank ASPIRE database and reports from the Committed to Equity (CEQ) project.

A. Ghana

Description of Main Social Safety Net Benefits

1. Ghana spending in SSN programs covered in this Annex amount to 0.3 percent as of 2022.2 The livelihood empowerment against poverty (LEAP) program, the school feeding program, and the free senior high school (SHS) initiative are the primary social safety programs in Ghana. The largest allocation was for the school feeding program at 0.18 percent of GDP, followed by the SHS at 0.06 percent of GDP, and the LEAP, despite being the government’s flagship cash transfer program, received a mere 0.05 percent of GDP. For Ghana, these estimates do not include the allocation for the National Health Insurance Authority.3

2. Ghana’s primary unconditional cash transfer program is the Livelihood Empowerment Against Poverty (LEAP) program. Transfers are targeted to the poorest households, and eligibility is assessed through proxy means tests (PMT) aimed at identifying the bottom 20 percent of the poor. Alongside the PMT criteria, beneficiaries must fall into one of the following categories: individuals aged 65 and above, individuals with disabilities, expectant mothers, or individuals responsible for orphaned and vulnerable children. The benefit amount varies based on the number of eligible individuals within the recipient household. As of 2022, the monthly transfers ranged from GH¢ 64 (US$ 5.73) to GH¢ 106 (US$ 9.49). The average annual benefit amounted to GH¢ 845 (4.6 percent of GDP per-capita). In total, the program benefitted 2.06 million individuals in 344 thousand households.

3. The Ghana School Feeding Program aims to increase school enrolment, attendance, and retention by providing meals that can account up to 30 percent of the daily calories’ intake. The program has showcased its effectiveness in addressing hunger and malnutrition while also encouraging local food production. The program targets pupils in public schools at the pre-secondary school level (i.e., students aged between 2 and 12 years old). Since its inception in 2005, the school feeding program has seen a substantial increase in its reach. The number of children receiving school meals has expanded significantly from 1,900 to 3.8 million in 2023. The program extends its benefits to 43 percent of children within the targeted age group. In line, with this expansion, there has been a notable increase in the program’s expenditure, mirroring the growth in coverage and the annual cost per pupil. As of 2022, the estimated annual cost per pupil stood at GH¢ 313 (1.7 percent of GDP per-capita) and the benefits reaches 3.6 million pupils.

4. The Free Senior High School program was instituted in the 2017–18 academic year to support students attending public senior high schools. The benefit is not means tested and its amount differs between resident students (GH¢ 1,002.47) and non-resident students (GH¢ 648.47). In the 2022–23 academic year, all students enrolled in Public Senior High Schools received the benefit. This encompassed approximately 400 thousand students, reaching one-third of the 1.1 current students. The average annual benefit per student amounted to GH¢ 985 (5.4 percent of GDP per-capita).

Assessment of Social Safety Nets

5. LEAP targets the poor more than the other SSN programs. About 58 percent of total LEAP expenditures reach household in the poorest quintile. This is significantly higher than the corresponding figure for the School Feeding program (31.3 percent) and SHS (10.4 percent). A decreasing share of LEAP reaches households in richer quintile groups, until reaching 3.5 percent for households in Q5. A similar trend although less progressive, is observed for the School Feeding program. On the contrary, only 10.4 percent of the total spending in SHS reaches households in the lowest quintile group, against 24.6 percent reaching household in Q5 (Figure 1).

6. Coverage rates vary substantially across programs. LEAP and SHS has the same coverage among the program analyzed. In total, about 4.7 percent of households receive LEAP in 2022, similarly 4.8 percent of the households benefit from SHS, and 27.9 percent benefit from the school feeding program. Both LEAP and the School feeding program show a progressive coverage pattern, with higher coverage at the bottom of the welfare distribution than at its top. About 21.7 percent of households in the poorest welfare quintile receive LEAP, against 0.5 percent of households in the richest quintile. Figures for the School Feeding program amount to 60.6 percent and 7.1 percent respectively. The coverage of the SHS shows a flat distribution, with 4 percent of household in Q1 receiving the support, similarly 4 percent of the households at the top of the distribution benefit from the program.

Figure 1.
Figure 1.

Coverage and Targeting of Main SSN Programs in Ghana

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

7. All SSN benefits show highest adequacy at the bottom of the income distribution (Figure 2). LEAP represents about 7.7 percent of the consumption of households receiving it. The same figure amounts to 3.1 percent and 3.5 percent for the School Feeding program and the SHS respectively. For all benefits the adequacy declines with household resources. Specifically, LEAP’s adequacy declines from 12.7 percent in Q1 to 4.1 percent in Q5, the School Feeding program share of equivalized income declines from 7.9 to 1.1 percent; the adequacy of the SHS declines from 9 to 2.2 percent.

Figure 2.
Figure 2.

Adequacy of Benefits: Main SSN Programs in Ghana

(In percent of household consumptions)

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates using SOUTHMOD

8. The school feeding program has the largest poverty reduction properties, followed by SHS and LEAP. Out of a baseline poverty rate of 24.6 percent, the school feeding shows poverty reduction properties of 1 percentage point, i.e., poverty would have been one percentage points higher in the absence of the program. The SHS reduces poverty by 0.4 percentage points, while LEAP by 0.2 percentage points. The school feeding program reduces the poverty gap by 0.81, followed by LEAP (0.31) and SHS (0.18).

Table 1.

Poverty Reduction of Main SSN Programs

(In percent)

article image
Sources: Fund staff estimates using SOUTHMOD

Reform Scenarios for LEAP

9. We evaluated two scenarios to enhance the SSN impact on poverty reduction: (i) increasing benefits while keeping the coverage and (ii) expanding coverage while keeping benefit amounts constant, aiming for a LEAP program budget of 0.5 percent of GDP. Table 2 shows the outcomes in poverty reduction, coverage, adequacy, and the percentage of the total benefits going to the households in Q1. In the benefit increase scenario, achieving 0.5 percent of GDP expenditure would mean decupling the current transfers, representing up to 60 percent of beneficiary households’ consumption in Q1, well above the 20 percent international benchmark. This is expected to reduce poverty by 2.9 percentage points. Conversely, relaxing eligibility criteria to increase coverage could lower poverty by 4.1 percentage points. However, it would also decrease the benefit share for the poorest from 57.5 percent to 35.6 percent, indicating higher leakage. Improvement in targeting will allow the government to achieve similar poverty reduction with less than half the proposed budget.

Table 2.

Characteristics and Poverty Reduction Reform Scenarios

article image
Sources: Fund staff estimates using SOUTHMOD

B. Mozambique

Description of the Main Social Safety Net Benefits

10. Mozambique’s primary social transfer is the Basic Social Subsidy Programme (BSSP). The BSSP is a cash transfer to low-income households belonging to one of the following categories: a) having a member who is permanently unable to work due to age, chronic illness, or disability; b) containing malnourished children or orphan children living in poverty; c) households headed by an orphan aged 14 to 18. The assessment of benefit eligibility includes means tests at both the individual and household levels. The BSSP is implemented in all districts, but geographical gaps in coverage exists. As of 2022, the monthly transfer depends on household composition, ranging from MZN 540 (USD 8.46, equivalent to 1.56 percent of per capita GDP) to MZN 1000 per household (USD 15.66, 2.89 percent of per capita GDP). The benefit reaches 492 thousand households, about 8.7 percent of the household population. Annual spending on the benefit amounts to 0.4 percent of GDP.4

Assessment of Social Safety Nets

11. The BSSP has limited efficiency in targeting and coverage. The targeting and coverage indicators of BSSP are relatively uniform across welfare quintiles, with the exception of the richest. Quintile groups from Q1 to Q4 receive between 20 and 24 percent of the total benefit each; coverage ranges from 12 percent of household in Q1 to 9 percent in Q4. The poor targeting observed could originate from the fact that both means tests are based on incomes from employment and pensions, while welfare quintiles are based on consumption.

Figure 3.
Figure 3.

Coverage and Targeting of the BSSP Program in Mozambique

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

12. The BSSP adequacy decreases steeply with household resources. Adequacy ranges from 34 percent of household consumption in Q1 to 3 percent in Q5. The BSSP account on average for 8.54 percent of household consumption at the country level.

Figure 4.
Figure 4.

Adequacy of Benefits: The BSSP Program in Mozambique

(In percent of household consumptions)

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates using SOUTHMOD

13. The poverty reduction properties of the BSSP are modest. With a baseline poverty rate of 47.49 percent, the BSSP contributes to a reduction of 0.5 percentage points (Table 3). The program reduces the poverty gap by 0.58. It should be noted that both poverty rate and poverty gap are well above the value observed in other countries analyzed in these case studies.

Table 3.

Poverty Reduction of the BSSP Program

(In percent)

article image
Sources: Fund staff estimates using SOUTHMOD

Reform Scenarios for BSSP

14. Consistent with other case studies, we evaluate two reform scenarios. First, we increase the BSSP generosity by 25 percent, with unchanged coverage; second, we expand coverage to those fulfilling eligibility conditions but not receiving the benefit in 2022, maintaining benefit levels constant. In both scenarios, we target a BSSP budget of 0.5 percent of GDP (i.e., extra 0.1 percent compared to baseline). Table 8 shows that increasing benefit generosity, improves marginally the poverty reduction properties of the scheme. Under this scenario, the transfer would account for 40 percent of Q1 beneficiary households’ consumption, potentially reducing poverty by 0.58 percentage points. Expanding coverage would have similar poverty reduction properties. Different from other case studies, the poverty reduction properties remain limited even in the reform scenarios. A mix of factors contribute to this finding. First, considering the high level of poverty rate and poverty gap in the baseline, the BSSP is not generous enough to lift from poverty households from the bottom of the income distribution. Only households already close to the poverty threshold exit poverty. Second, the total benefit already amounts to 0.4 percent of GDP in the baseline, so there is limited fiscal space for benefit increase.

Table 4.

Characteristics and Poverty Reduction Reform Scenarios

article image
Sources: Fund staff estimates using SOUTHMOD

C. Tanzania

Description of the Main SSN Benefits

15. The Productive Social Safety Net (PSSN) program is the main social program in Tanzania. The PSSN consists of a basic cash transfer paid to low-income households and a top-up cash transfer related to household composition (number of children of various age groups, number of people with disability). Eligibility is based on PMT and community inputs. As of 2022, about 685 thousand households (6 percent of the household population) benefit from PSSN, with total spending amounting to 0.14 percent of GDP. The basic cash transfer amounts to 12 thousand TZS per month for households with at least one adult (0.4 percent of per-capita GDP) and an additional 5 thousand TZS per month for households with at least one child (0.18 percent of GDP). The top-up amounts depend on the household composition, with the average benefit amounting to 30.6 thousand TZS per month (1.1 percent of per capita GDP).

16. This case study focuses solely on the PSSN and does not analyze other SSN components. The SSN encompasses various other initiatives, such as the public work component within the PSSN, the National Agriculture Input Voucher Scheme (NAIVS), and the Bed Net Program, which distributes anti-mosquito nets. However, these components are not included in the scope of this case study analysis.

Assessment of Social Safety Nets

17. The PSSN demonstrates high efficiency in beneficiary selection, effectively directing 83 percent of its benefits to households within the lowest two income quintiles (Figure 5). This precision in targeting ensures that most of the program’s resources are allocated to its intended recipients, effectively minimizing leakage to non-poor households. However, coverage remains a critical area for enhancement. Currently, the PSSN extends to merely 25 percent of households in the lowest quintile and only 8 percent in the second quintile. As household consumption levels rise, the coverage of the PSSN diminishes, extending to just 1 percent of households in the fourth and fifth quintiles. This robust targeting mechanism is primarily attributed to the program’s dual strategy, which blends means-tested targeting with community-based selection, effectively identifying and enrolling predominantly impoverished families, aligning closely with the program’s core objectives of poverty alleviation and social protection.

Figure 5.
Figure 5.

Coverage and Targeting of the PSSN Program in Tanzania

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates using SOUTHMOD.

18. The PSSN exhibits its highest benefit adequacy at the bottom of the consumption distribution (Figure 6) yet remains modest compared to international best practices. For households receiving PSSN benefits, these account for approximately 11 percent of their consumption. The benefit adequacy declines with household consumption and represents a significant proportion of the consumption for households in Q1. Specifically, PSSN adequacy decreases from 14.8 percent in Q1 to a mere 3.1 percent in Q5, highlighting its targeted support to the most vulnerable households.

Figure 6.
Figure 6.

Adequacy of Benefits: The PSSN Program in Tanzania

(In percent of household consumptions)

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates using SOUTHMOD

19. The poverty-reducing properties of the PSSN are modest. Despite the good work of the program in selecting poor households, the program fails to lift a significant portion of households out of poverty. This limited impact is attributable to the relatively modest benefits, as well as low coverage. With a baseline poverty rate of 25.62 percent, the PSSN contributes to a marginal reduction of 0.62 percentage points (Table 5). Furthermore, the program reduces the poverty gap by 0.56. It should be noted that both the poverty rate and the poverty gap are in line with the values observed in other countries analyzed in these case studies.

Table 5.

Poverty Reduction of the PSSP Program

(In percent)

article image
Sources: Fund staff estimates using SOUTHMOD

20. When looking at the tax-benefit system, we found that, on average, all households are net contributors in Tanzania. The tax system in Tanzania imposes a progressively heavier burden with increasing household consumption. Even those in the first quintile, the lowest income group, are net contributors to the system, contributing 4.1 percent of their consumption. For the poorest 20 percent, benefits constitute 3.9 percent of household consumption and 0.03 percent for the top quintile. Direct taxes represent 12.2 percent of consumption for the highest quintile, compared to 3.2 percent for the lowest. The burden of indirect taxes also increases with consumption, rising from 4.7 percent in the lowest quintile to 7.7 percent in the highest.

Reform Scenarios for BSSP

21. In line with the other case studies, we evaluate two reform scenarios for the PSSN: enhancing benefit generosity and expanding coverage, both targeting a budget of 0.5 percent of GDP. The first scenario involves a 173 percent increase in PSSN generosity without altering coverage, while the second entails extending coverage to all eligible households not currently receiving benefits in 2022, keeping benefit levels constant. Enhancing benefit generosity significantly reduces the poverty rate (Table 6). In this scenario, the transfer would represent 37.5 percent of consumption for households in the first quintile (Q1), potentially reducing poverty by 3.16 percentage points – a fivefold compared to the current impact of the program. The poverty gap will be reduced by 1.37 percentage points.

22. Expanding PSSN coverage, while less impactful in poverty reduction, significantly increases coverage and reduces the poverty gap. Expanding coverage under the PSSN would lead to a slightly lesser effect on poverty reduction compared to increasing benefits, but it would significantly enhance the program’s coverage. This expansion would result in the program covering 84.3 percent of households in Q1, given the efficiency of the current targeting mechanism. Due to this extensive coverage among poor households, the poverty gap would be notably reduced by 1.92 from a baseline of 6.27. The reason for the lower poverty reduction effect, relative to the benefit increase scenario, is attributed to the fact that while increased coverage brings more households closer to the poverty line, the amount transferred is insufficient to fully lift them out of poverty.

Table 6.

Characteristics and Poverty Reduction Reform Scenarios

article image
Sources: Fund staff estimates using SOUTHMOD

D. Uganda

Description of the Main Social Safety Net Benefits

23. Uganda’s flagship social protection program is the Senior Citizens Grant (SCG), which provides cash benefits to senior citizens, and the only eligibility criterion is age. Initially, the program targeted individuals over 65 in specific districts (lowered to 60 in the Karamoja region). In 2020, it expanded to encompass all seniors aged 80 and above nationwide. This increased age criterion is now uniformly applied, although beneficiaries registered prior to July 2020 maintain their eligibility. As of 2022, the program disburses UGX 25,000 monthly per beneficiary (USD 6.75), accounting for 8.7 percent of per capita GDP. Currently, the SCG reaches about 645,000 individuals, representing 42 percent of those over 65, with its total spending constituting 0.12 percent of GDP.

24. Other social programs exist, but their analysis falls beyond the scope of this study. The Extremely Vulnerable Households Program offers in-kind and cash transfers to food-deprived households in the Karamoja region. The program reached 215,218 households in 2018, less than 2.5 percent of national households. The Youth Livelihood Program and the Uganda Women’s Entrepreneurship Program focus on empowering youth and women through skill training and project financing. Meanwhile, the Operation Wealth Creation initiative aims to transition subsistence agriculture to commercial levels by providing agricultural inputs.

Assessment of Social Safety Nets

25. The SCG program shows limited progressivity and experiences significant leakages toward better off households by design. While the largest proportion of the transfers goes to households in the lowest consumption quintile, those going to households in Q2 to Q5 are relatively uniform, with each quintile receiving, on average, 19 percent of the benefits. The program’s sole eligibility criterion is age, with no linkage to household consumption levels. This absence of a Proxy Means Test (PMT) or other targeting mechanisms accounts for the SCG’s moderate progressivity and significant leakages; around 76 percent of the total benefits go to non-poor households. On average, 6.4 percent of households benefit from the SCG, with the highest coverage among the poorest 20 percent, where 8.8 percent of households have at least one beneficiary. Coverage decreases with rising consumption levels, dropping to 4 percent in the highest quintile.

Figure 7.
Figure 7.

Coverage and Targeting of the SCG Program in Uganda

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates using SOUTHMOD

26. The SCG program shows the highest adequacy at the bottom of the consumption distribution (Figure 8). The SCG accounts for approximately 7.6 percent of consumption for beneficiary households. This adequacy diminishes as household consumption increases, ranging from 26.3 percent in the lowest quintile to 2.5 percent in the highest quintile. Notably, among poor households benefiting from the program, the SCG comprises a significant proportion of their total household consumption.

27. The poverty alleviation impact of the SCG program, while modest, aligns with trends observed in other SSA countries. With a baseline poverty rate of 20.4 percent, the SCG contributes to a reduction of 0.6 percentage points (Table 7). In other words, in the absence of the SCG, the poverty rate would have been higher by this margin. This level of poverty mitigation is consistent with outcomes in similar countries, considering the program’s budget allocation.

Figure 8.
Figure 8.

Adequacy of Benefits: The SCG Program in Uganda

(In percent of household consumptions)

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates using SOUTHMOD
Table 7.

Poverty Reduction of the SCG Program

(In percent)

article image
Sources: Fund staff estimates using SOUTHMOD

Reform Scenarios for SCG

28. We evaluated two reform scenarios: (i) increasing benefits with unchanged coverage; and (ii) expanding coverage to all seniors and newborn, maintaining benefit levels, targeting a SCG budget of 0.5 percent of GDP. Table 6 presents a detailed analysis of the potential impact on poverty reduction, coverage, adequacy, and the allocation of total benefits to households in the lowest consumption quintile. In the scenario of increasing benefits to achieve 0.5 percent of GDP in expenditures, a more than fourfold increase in the current transfer amounts is required. Under this scenario, the transfer would account for 60 percent of Q1 beneficiary households’ consumption, potentially reducing poverty by 1.72 percentage points. A larger reduction in poverty, estimated at 2.1 percentage points, could be achieved by expanding the coverage to include all individuals over 65 and all newborn while maintaining the transfer at UGX 25,000. However, this coverage expansion, lacking specific targeting measures and including children, leads to a decrease in the proportion of benefits reaching the poorest 20 percent, thereby increasing leakages. Consequently, incorporating a targeting mechanism could yield comparable reductions in poverty with significantly lower resource requirements.

Table 8.

Characteristics and Poverty Reduction Reform Scenarios

article image
Sources: Fund staff estimates using SOUTHMOD

E. Zambia

Description of Main Social Safety Net Benefits

29. Budget allocations within SSN programs in Zambia vary significantly, with the SCT program budget allocation being the largest, amounting to approximately 0.44 percent of GDP. According to SOUTHMOD simulations, agricultural benefits not covered in this note have the second-highest allocation. Meanwhile, the school meal program, benefiting from its policy of using local products, gets 0.1 percent of GDP despite broader coverage. Smaller allocations are seen in the SWL and KGS programs, complementing the SCT, at 0.05 and 0.01 percent of GDP, respectively. It should be noted that these figures only cover transfer costs and exclude the administrative expenses of the programs.

30. Zambia’s flagship social assistance program SCT targets labor-constrained individuals and extremely poor households. SCTs employs proxy means tests, categorical targeting, and community input, to identify potential beneficiaries. Moreover, beneficiary households must fall into one of the following categories: Households with an elderly person; households with members with severe disability; households with members who are chronically il and on palliative care; child-headed households; female-headed households with at least three children. As of 2022, the SCT is operational in all 116 districts, providing bi-monthly cash transfers of 400 ZMW. Households with severely disabled members receive a higher benefit of 800 ZMW. In 2022, the program benefitted 680 thousand households, with an average annual transfer of ZMW 3283 (equivalent to 13 percent of GDP per capita).

31. Other social assistance programs work alongside the SCT in Zambia’s effort to combat poverty. Key among these programs are:

  • Supporting Women’s Livelihoods (SWL): The SWL is one component of the Girls Education and Women Empowerment and Livelihoods (GEWEL) Project. SWL provides training, grants, and support, primarily to women in SCT beneficiary households in targeted districts. In 2022, a total of 50.2 thousand households received a one-time transfer of ZMW 4442 (USD 225, equivalent to 15 percent of GDP per capita).

  • Keeping Girls in School (KGS): Operating under the GEWEL initiative, the KGS program aim to enhance secondary education access for girls aged 14 to 18 from extremely poor households in 39 districts. Targeting girls in SCT households, KGS provides school fee vouchers and, since 2021, an additional lump-sum equivalent of 15 per cent of the annual SCT amount for school-related incidentals like uniforms and shoes. In 2022, the program reached almost 56 thousand girls or 2.7 percent of all girls aged 14 to 18 in the country. Beneficiary girls received an average transfer of ZMW 1466, amounting to 5.81 percent of GDP per capita,

  • Home-Grown School Meal Program (HGSM): Managed by the Ministry of General Education, the HGSM initiative operates in 70 districts, providing daily free meals to public school students using locally sourced products. The program, linked to school attendance. In 2022, 1.3 million students received benefits from the program, comprising 54 percent of the school-aged population (5 to 14). The value of the provided meals amounted to ZMW 397 per child annually, equivalent to 1.57 percent of GDP per capita.

Assessment of Social Safety Nets

32. In Zambia, the reach and accuracy of different SSN programs vary substantially. On average, about 60 percent of these programs’ benefits go to the poor, those in the bottom 40 percent of the consumption distribution (Figure 9). The school feeding program and the SCT are the biggest and most widespread, reaching 41 and 38 percent of the poorest families, respectively. Smaller programs like the SWL and the KGS don’t reach as many households. Importantly, SWL does a great job of targeting the right people, giving 94 percent of its benefits to the bottom 40 percent of the consumption distribution. However, there is room for improvement on the targeting instrument used by the other programs, as around 20 percent of their benefits go to the households on the top 40 of the consumption distribution.

Figure 9.
Figure 9.

Coverage and Targeting of Main SSN Programs in Zambia

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates using SOUTHMOD

33. SSN program transfers form a notable portion of beneficiary household’s consumption, especially for the households in the bottom 20 percent of the consumption distribution. SCT program transfers account for 28 percent of the household consumption for households in the lowest quintile (Figure 10). This importance decreases as household consumption increases. The school feeding program, with the broadest coverage, provides smaller benefits, averaging 7 percent of the household consumption of households in the bottom 20 percent, and 1 percent for those in the highest quintile. SWL program benefits are comparable to the SCT, and households benefiting from both programs can see transfers amounting to as much as 57 percent of their consumption. Benefits from the KGS initiative are similar in scale to the school meal program.

Figure 10.
Figure 10.

Adequacy of Main SSN Programs in Zambia

(In percent of household consumptions)

Citation: Policy Papers 2024, 011; 10.5089/9798400272400.007.A002

Sources: Fund staff estimates using SOUTHMOD

34. The effectiveness of SSN programs in poverty alleviation in Zambia is closely linked to their respective budget allocations. This correlation is evident in the substantial contribution of the SCT to poverty reduction, where it effectively lowers the poverty rate by two percentage points. Without the program, the poverty rate would escalate to 40.8 percent (Table 9). The KGS program, being the smallest in terms of budget, has a correspondingly lesser impact on poverty reduction, reducing it by only 0.08 percentage points. The targeted nature of these programs not only reduces the poverty headcount but also significantly narrows the poverty gap. The magnitude of the poverty gap reduction mirrors that of the reduction in the poverty rate.

Table 9.

Poverty Reduction of Main SSN Programs

(In percent)

article image
Sources: Fund staff estimates using SOUTHMOD

Reform Scenarios for SCT

35. We evaluated two SSN reform scenarios: (i) increasing the SCT benefit while keeping the coverage, and (ii) expanding coverage while keeping the benefit constant, aiming for SCT program expenditure at 0.5 percent of GDP. Table 4 details the effects on poverty reduction, coverage, and benefit distribution, particularly for households in the first quintile. Boosting spending from 0.44 to 0.5 percent of GDP translates into a 9.14 percent rise in household transfers in the first quintile, resulting in an additional 0.39 percentage point drop in the poverty rate. Alternatively, broadening coverage by easing eligibility nearly doubles first quintile beneficiaries to 65.5 percent of households. This strategy reduces leakages to non-poor households, with 66 percent of benefits reaching the poorest 40 percent, up from the current 56 percent. Such a reform could significantly reduce poverty, potentially decreasing the poverty headcount rate up to 3.52 percentage points.5

Table 10.

Characteristics and Poverty Impact Reform Scenarios

article image
Sources: Fund staff estimates using SOUTHMOD
1

Operationally, a rough estimate has been provided by various Fund publication as the sum of current account deficits and external amortization from the WEO medium-term projections. Importantly, the financing needs so defined (“Gross Financing Needs”, GFN) are by definition “already met” by a combination of capital and financial account transactions, official support, and active IMF programs.

2

See also 2021 LIC Report.

3

The FAD model presents a Cobb-Douglas production function where TFP is set as a residual so that growth matches pre-determined growth assumptions by country teams. We considered the initial expenditure multiplier and persistence approach suggested in FAD’s fiscal multiplier template. Given that the multipliers in the template were not specific to LICs, which would likely have lower multipliers due to efficiency considerations and the size of the informal sector (Colombo, et al, 2022 Fiscal Multipliers and Informality), we assumed 50 percent of those values (resulting in: 0.2, 0.25, 0.2, 0.15, 0.1, and 0.05, for years 1 to 6, respectively). An alternative scenario assuming 75 percent of the whole-sample values would also be offered. The scaling up of spending, associated with the multiplier effects above, would endogenously increase growth beyond the starting WEO projections for each country.

4

If we would only made this revision to the tool, and would not consider any absorptive capacity constraints, the estimated financing needs are US$2.2 trillion for 2024–28 (Table 2 row 1).

5

We have assumed a cap equal to the 80th percentile change in total expenditure for the relevant sample of LICs from the period 1999–2019. This timeframe set to avoid distortions that could be introduced by including atypical patterns resulting from the COVID pandemic. The threshold is computed for each of the quartiles of the distribution of LIC, based on the initial level of total public expenditure over 1999–2019 (see Annex II, Box 2).

6

This, however, would imply that even by 2040, there would be a gap between what the public sector can finance and the actual spending that should be executed to achieve the SDGs. This gap could be filled by the private sector (see paragraph 8).

7

Given that the FAD database covers only 47 countries out of a total of 69 LICs, our estimate relies on an extrapolation by applying the median estimate of additional SDG needs in percent of GDP for the sample of LIC countries covered, to the missing countries’ nominal GDP levels.

8

For this exercise some of the institutional categorizations reported in Annex I, were modified to define non-overlapping groups of countries.

9

It should be noted that small states have particularly large needs for climate change adaptation, which are not captured yet as part of the SDG costing exercise, hence their needs are likely significantly underestimated.

10

The reference to total financing needs is useful to compare results with previous exercises, but it relies on the strong underlying assumption that all public spending increases external financing needs 1:1. However, the import propensity of SDG-related spending may vary, but it is likely below 1, especially for health and education.

11

It is important to note that there are arguably significant overlaps between investments needed for convergence and SDG-related spending.

12

Materialization of macroeconomic shocks would exacerbate financing requirements for LICs, beyond the estimates described in the previously described baseline scenarios. LICs are particularly susceptible to external shocks, owing among others. to their typically less diversified economic structures and weaker macroeconomic buffers. Annex II presents an assessment of scenarios involving negative terms-of trade shocks such as a one-standard deviation increase in global oil prices and food prices. Such shocks could potentially anywhere from US$ 5 to 20 billion. Assessing the potential impact of more extreme shocks, akin to the once-in-a-decade pandemic, is methodologically challenging but it would increase the financing needs from standard shocks by several folds. As reported in Table 2, the previous LIC Report provided an estimate of the financing toll of the COVID-19 pandemic at US$ 150 billion.

13

In our exercise, we assume revenue efforts of about 5.0 percentage point increase in the Tax to GDP ratio for the median LIC over 5 years, with the country-specific increase being calibrated depending on the gap between current tax-revenue ratio projected in the medium term and the tax potential estimated by a Benitez et al. 2023. This implies a size and pace of adjustment higher than observed in the historical period examined (1999–2019, at around 3.1 percentage points of GDP), but substantially lower than the potential revenue increase from reforms that the FAD study finds to be feasible through tax system and capacity building reforms (9.0 percentage points of GDP).

14

An alternative way to think of this is that the US$274 residual gap can be “caught up with” by the public sector over a longer timeframe. Our model allows for this, as caps in spending increases would become less binding beyond 2028.

1

Prepared by Julieth Pico and Alberto Tumino.

2

For Ghana, SOUTHMOD employs survey data from the Ghana Living Standard Survey 2017 (GLSS-7). Furthermore, the results presented in the note are based on the 2022 tax-benefit, for that the monetary variables reported in the survey, e.g., earnings and expenditure items, are updated to 2022 values using CPI index. Socio-demographic variables such as labor market status or family composition are not updated, i.e., the values reported in the survey are kept. LEAP eligibility in SOUTHMOD is scaled up to reach 344 thousand households, in line with official statistics.

3

For 2024, it is projected that Ghana would spend about 0.5 percent of GDP in highly targeted social spending, including (Ghana School Feeding Program, NHIS transfers, LEAP, and Capitation Grant). (IMF 2024).

4

Other social programs exist beyond the scope of this study, including the Programa do Apoio Social Directo (time limited in kind support to vulnerable households), the Programa dos Serviços Sociais de Acção Social (social welfare services activities to communities), the Acção Social Escolar and the Acção Social da Saúde (access to education and health care to the vulnerable population), and the Programa Acção Social Produtiva (public works).

5

The table below incorporate variations in other benefit linked to SCT eligibility. For example, SWL spending increases from 0.05 percent of GDP to 0.16 percent because of the increase in the coverage of the SCT.

  • Collapse
  • Expand
Macroeconomic Developments and Prospects For Low-Income Countries—2024
Author:
International Monetary Fund
and
World Bank
  • Estimates of Additional Spending for SDGs Under Alternative Approaches

    (In millions of U.S. dollars)

  • Additional Spending Towards SDG by Country’s Institutional Category

    (In billions of U.S. dollars)

  • Figure 1.

    Coverage and Targeting of Main SSN Programs in Ghana

  • Figure 2.

    Adequacy of Benefits: Main SSN Programs in Ghana

    (In percent of household consumptions)

  • Figure 3.

    Coverage and Targeting of the BSSP Program in Mozambique

  • Figure 4.

    Adequacy of Benefits: The BSSP Program in Mozambique

    (In percent of household consumptions)

  • Figure 5.

    Coverage and Targeting of the PSSN Program in Tanzania

  • Figure 6.

    Adequacy of Benefits: The PSSN Program in Tanzania

    (In percent of household consumptions)

  • Figure 7.

    Coverage and Targeting of the SCG Program in Uganda

  • Figure 8.

    Adequacy of Benefits: The SCG Program in Uganda

    (In percent of household consumptions)

  • Figure 9.

    Coverage and Targeting of Main SSN Programs in Zambia

  • Figure 10.

    Adequacy of Main SSN Programs in Zambia

    (In percent of household consumptions)