## wpiea2019270-print-pdf

## Source details

**Canonical URL:** [wpiea2019270-print-pdf](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019270-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019270-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019270-print-pdf.pdf.json)

---

### Introduction and context
- Sub-Saharan Africa (SSA) faces large development challenges driven by rapid population growth, high persistence of inequality, and prevalence of fragility.
- Extreme poverty trends (1990–2015):
  - SSA: from 54 percent in 1990 to 41 percent in 2015.
  - World total: from 36 to 10 percent.
  - Asia and Pacific: from 62 to 2 percent.
  - Latin America and Caribbean: from 14 to 4 percent.
- Life expectancy at birth (2015):
  - SSA average: 60 years.
  - Rest of the world average: 74 years.
- SSA countries have endorsed the Addis Ababa Action Agenda and are aligning national development plans to the SDGs and Africa Union’s Agenda 2063.

### Key findings on development performance
- SDG and human development gaps:
  - Median SDG index score across SSA countries is on average 25 percent lower than the median in other regions.
  - Income per capita (2018): US$1,574 (US$3,906 in PPP terms) for SSA.
  - Human Development Index: SSA average 0.52 versus 0.76 in the rest of the world.
- Sectoral performance (median outcomes across education, health, roads, electricity, water and sanitation):
  - SSA median outcomes are behind medians across EMEs.
  - Water and sanitation concentrate around an average SDG score of 49 out of 100.
- Historical improvements (1990–2017):
  - Net primary enrollment rate increased by almost 30 percentage points.
  - Infant mortality has been halved.
  - Life expectancy at birth increased by 22 percent.
  - Access to electricity improved but half of SSA population still lacks reliable electricity.

### Public spending and infrastructure trends
- 2015 public expenditure averages across SSA:
  - Education: 4.6 percent of GDP.
  - Health: 2.8 percent of GDP.
  - Comparison: health spending in other EMEs averages 4.2 percent of GDP; public education across SSA reaches 4.1 in 2015 in comparison.
- Public real capital stock per capita:
  - Almost flat between 1990 and 2015 in SSA.
  - Other LIDCs and emerging markets increased public real capital stock per capita on average by 5.4 percent and 3.4 percent per year, respectively, over the same period.

### Costing methodology and additional spending estimates (Box 1)
- Methodology overview (three-step approach from Gaspar and others (2019)):
  - (i) identifying main cost parameters (inputs and unit costs);
  - (ii) benchmarking cost parameters to levels in countries with comparable GDP per capita and high development outcomes today;
  - (iii) estimating spending levels associated with these benchmarks given country GDP per capita and population growth projections until 2030.
- Reporting convention:
  - Estimates correspond to additional total expenditure required in the year 2030 compared to 2016 spending.
  - Reported as percentage points of 2030 GDP and refer to total expenditure (public and private).
  - Benchmarks use high-performing countries with comparable GDP per capita; estimates assume high spending efficiency and are interpreted as a lower spending floor.
  - After 2030, education and health spending would be mostly recurrent, while infrastructure spending would decrease to cover depreciation of the capital stock built through 2030.
- Aggregate regional findings:
  - Median SSA country must spend an extra 18.8 percent in 2030 in education, health, water and sanitation, roads and electricity to achieve high development outcomes.
  - These additional needs are shared equally between human and social investment and physical capital investment.
  - Median SSA additional spending: about 19 percent of GDP.
  - Comparisons: median additional spending of 12 percent of GDP for non-SSA LIDCs and 4 percent of GDP for EMEs.
- Sector distribution in SSA (five selected sectors):
  - Human capital sectors (health and education) represent about 9 percent of GDP.
  - Physical capital sectors (water, electricity, and roads) sum to 10 percent of GDP.
  - Additional physical infrastructure investment in SSA is 40 percent higher than average infrastructure needs across LIDCs and four times more than the EMEs’ average.
- Heterogeneity:
  - Fragile countries: additional spending of 24 percent of GDP in 2030 (6 percentage points higher than non-fragile SSA countries).
  - No major difference between resource rich and non-resource rich countries in aggregate additional spending.

### Sector costing approaches (summary)
- Education:
  - Total spending function: number of teachers, teacher salaries, share of non-compensatory current expenses, and capital spending.
  - Benchmarks: median values in countries with comparable per-capita income and high education outcomes.
  - Assumed enrollment targets: 50 percent for preprimary and tertiary, 100 percent for primary and secondary.
  - Benchmark GDP per capita buckets used: USD 0–3,000; USD 3,000–6,000; USD 6,000–15,000.
- Health:
  - Total spending function: doctor salaries, number of doctors and other medical personnel, ratio of non-doctor to doctor wages (assumed 0.5), share of non-compensatory current expenses, and capital spending.
  - High-performing low-income developing countries defined as SDG3 health index above 70.
- Roads:
  - Road density regressed on GDP per capita, population density, agriculture and manufacturing shares, urbanization rate, and the World Bank’s Rural Access Index (RAI).
  - Additional kilometers estimated by raising the RAI to at least 75 percent in LIDCs.
  - Unit cost per kilometer set at a minimum of USD 500,000 unless country-specific estimates suggest otherwise.
- Electricity:
  - Additional network corresponds to 100 percent access of projected population in 2030, accounting for per-capita consumption increases with GDP per capita.
  - Unit cost per kilowatt of generation capacity set at USD 2,250 (World Bank 2013).
- Water and sanitation (WASH):
  - Costs derived using the WASH World Bank methodology (Hutton and Varughese (2016)); unit costs calibrated at the country level, including capital investment, operations, and major capital maintenance.

### Drivers of SSA vs other LIDCs/EMEs differences (selected findings)
- Education:
  - SSA additional spending needs average 4.2 percent of 2030 GDP, vs 0.3 percent in other LIDCs and EMEs.
  - SSA demographics explain 91 percent of the difference; projected student-age population share in 2030 is 47 percent in SSA versus 30 percent in other LIDCs and EMEs.
- Health:
  - SSA additional spending needs average 4.7 percent of 2030 GDP, vs 1.5 percent in other LIDCs and EMEs.
  - Almost half the difference driven by lower GDP per capita requiring a costlier combination of health staff density and salaries.
- Roads:
  - SSA additional spending needs average 6.9 percent of 2030 GDP, vs 0.9 percent in other LIDCs and EMEs.
  - Lower 2030 GDP per capita and initial lower RAI explain 84 percent of the difference.

### Benin case study: additional spending and sectoral estimates (2030)
- Aggregate additional spending required in five sectors: about 21 percent of GDP in 2030.
- Education:
  - 2030 required spending: 8.7 percent of GDP (US $395 per student).
  - Current spending: 5.5 percent of GDP (US$135 per student).
  - Student-to-teacher ratio: decline to 15.1 from 22.4 today.
  - Teacher compensation share: decline from 60 percent to 45 percent of total spending.
- Health:
  - 2030 required spending: about 9.3 percent of GDP (US$119 per capita).
  - Current total spending: 4.2 percent of GDP (US$33 per capita).
  - Staffing: recruit 8 times more doctors and 4 times more support staff than today.
  - Policy: ARCH launched to provide universal health insurance (see source).
- Roads:
  - Additional kilometers to build: 12,276 km.
  - Additional spending required: 8.1 percent of GDP in 2030 (of which 1.9 percent of GDP for maintenance).
  - Road transport: 93 percent of people transport and 73 percent of goods transport; paved roads account for 45 percent of total roads.
  - Transport-related infrastructure: account for 25 percent of all investment under the PAG.
- Electricity:
  - Annual cost to reach universal access: estimated at 2.4 percent of GDP.
  - Domestic production meets only 12 percent of domestic consumption; reliance on imports from Côte d'Ivoire, Ghana, and Nigeria.
  - Constraints: lower cost of imported electricity relative to domestic generation; low tariffs that only partially cover production costs.
  - PAG: electricity-related infrastructure to account for 10 percent of all investment under the PAG.
- Water and sanitation:
  - Total cost to reach universal access: estimated at 2.5 percent of GDP per year—0.8 percent in water and 1.7 percent in sanitation—through 2030.
  - Sanitation accounts for 70 percent of the total estimate because only about 20 percent of the population has access to improved sanitation.
  - PAG target: reach universal access to water by 2021; 3 out of 45 flagship projects relate to this goal.

### Rwanda case study: sectoral additional spending estimates for 2030
- Aggregate additional spending required in five sectors: about 19 percent of GDP.
- Education:
  - Required spending to meet the SDGs: 7 percent of 2030 GDP.
  - Current public spending: 3.6 percent of GDP.
  - Student-to-teacher ratios: current 41 students per teacher vs. peers with better performance 15 students per teacher.
  - Government plan by 2030: nearly double current spending to reach 6.3 percent of GDP.
  - SDG index 3 score: 61.
- Health:
  - Estimated additional spending needed: 2 percent of GDP by 2030.
  - Current per capita public spending planned increase by 2024: from US$38 to US$52 per capita per year (a 36 percent increase).
  - Required change: increase in number of health workers, mostly doctors.
- Roads:
  - Additional required spending: about 4 percent of 2030 GDP.
  - RAI: 52 out of 100.
  - Current network: only 5 percent paved; only 10 percent reaches all-season riding quality.
  - Investment focus: upgrading and maintaining existing roads, especially in remote and high areas.
- Electricity:
  - Additional spending for universal access: 2 percent of GDP.
  - Current household electricity access: 43 percent.
  - Assumptions: 60 percent of new connections on-grid and 40 percent off-grid; electricity consumption growth annual rate of 9 percent; projected consumption 730 kWh in 2030 from an average of 238 kWh in 2016.
- Water and sanitation:
  - Required spending (WASH model): 4.5 percent of GDP per year to safely provide water to all households and provide fixed-point latrines to all.

### Tailoring costing estimates to country context
- Rationale:
  - Significant variation across SSA countries justifies tailoring cost parameters to country circumstances.
  - Methodology enables benchmarking against SDG-high performing peers for input mix and potential efficiency gains.
- Parameters that can be tailored:
  - Unit cost to build a km of road.
  - Unit cost to generate one kilowatt of electricity.
  - Input-mix variables (e.g., student-to-teacher ratio, number of doctors per 1,000 population).
- Process:
  - Tailoring requires discussion with authorities and stakeholders and good understanding of national development challenges.
  - Costing exercises can catalyze coordination among development partners and ground development agendas in shared knowledge.
  - Benin and Rwanda analyses were refined during costing missions in 2018 involving line ministries and development partners.

### Baselines, data gaps, and monitoring
- Key step: establish clear baselines to measure and monitor progress.
- Data actions/examples:
  - Rwanda: line ministries and development partners gathering data on SDG target baselines; UNICEF and UKDFID helped establish common template in education.
  - Benin: authorities prioritized SDGs fitting their context and estimated needs with technical and financial partners.
- Data limitations:
  - 60 percent of all 17 SDG indicators are currently not documented in SSA countries; existing data often outdated or not comparable across countries (SDG Center for Africa, 2019).
  - Lack of updated baselines complicates planning and transparency.

### Short-term implications: prioritize, raise resources, and spend efficiently
- Use of costing:
  - Rank and prioritize competing projects according to countries’ priorities and capacity to deliver.
  - Map spending gaps to existing sectoral projects and sort by implementation timeline and financing capacity.
- Financing strategy elements:
  - Ramping up domestic resources is core: average tax-to-GDP ratio about 14 percent in 2017; room to reach the average of 18 percent for emerging economies.
  - Additional tax revenues alone insufficient to cover additional investments required by 2030.
  - Importance of official development assistance (ODA) and creating an enabling environment to attract private investment.
- Efficiency imperative:
  - Costing assumes both more and more efficient spending.
  - If spending efficiency is not improved, additional spending required to reach the SDGs will be much larger.
  - Example capacity needs: increasing spending in the medium term by an additional 19 percent of GDP would require substantial improvement in capacity to deliver public services and execute budgets.
  - Projected public spending trajectories: assuming current public-private split continues, Benin and Rwanda expected to more than double public spending in education and health by 2030.

### Medium-to-long term implications: anchoring in budgets and financing strategies
- Budgetary integration:
  - 2030 spending estimates can anchor medium-term budget strategy and discussion of sustainable financing paths.
  - Rwanda: MTEF includes a subset of aggregated costs from sectoral NST chosen based on absorptive capacity and projected revenues.
- Functional classification and accountability:
  - Presenting budget according to functional classification could strengthen accountability in delivering SDGs.
- Integrated National Financing Strategy (INFF) building blocks:
  - Improve political stability, promote good governance, and improve business environment.
  - Map spending estimates to commensurate resources with a clear public/private investment split in each sector.
- Country-specific constraints and options:
  - Rwanda: limited potential for additional tax revenues in the medium term—about 2 to 3 percent of GDP—and a declining trend in ODA not compensated by increasing private investment; MTEF includes “contingent” spending items depending on additional resources.
  - Benin: room to increase tax revenue via consumption taxation (VAT and excises); reforms to boost private investment continue but private sector participation remains limited.
  - Recommendation: further country-level analysis to align national development plans with assessed spending needs and a comprehensive financing strategy detailing private and public sector roles.

### Appendix Table 3: Estimates of Roads Spending in 2030 in Benin and Rwanda (selected figures)
- Unit cost (USD per km) reported in table: 1,100,000 and 609,905.
- GDP per capita (2016 / 2030) reported for examples:
  - Rwanda: 1,210 (2016) and 2,172 (2030).
  - Benin: 734 (2016) and 1,363 (2030).
  - Additional listed GDP per capita numbers in table: 915, 791, 1,206.
- Road density (km per km2) examples shown: 164, 406, 191, 678, 142, 277.
- Population density (pop per km2) examples shown: 75, 89, 504, 640, 104, 138.
- Results (selected):
  - Additional annual spending (percent of 2030 GDP) values shown in table: ---, 3.9, -8.1.
  - Depreciation (percent of 2030 GDP) examples: 1.5, 3.0.
  - Rural Access Index (RAI) examples: 54, >75, 52, >75, 32, >62.5.
- Contextual entries:
  - Electricity: unit cost per kilowatt set at US$2,250 (World Bank 2013).
  - Water and sanitation: estimates derived using WASH World Bank methodology (Hutton and Varughese 2016); unit costs calibrated at the country level.

*Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019270-print-pdf.pdf*

### References _______________________________________________________________________________________ 25

### wpiea2019270-print-pdf - References _______________________________________________________________________________________ 25

### Introduction and context
- Sub-Saharan Africa (SSA) faces large development challenges driven by rapid population growth, high persistence of inequality, and prevalence of fragility.
- Extreme poverty:
  - SSA: from 54 percent in 1990 to 41 percent in 2015.
  - World total: from 36 to 10 percent (1990–2015).
  - Asia and Pacific: from 62 to 2 percent (1990–2015).
  - Latin America and Caribbean: from 14 to 4 percent (1990–2015).
- Life expectancy at birth:
  - SSA average: 60 years (2015).
  - Rest of the world average: 74 years (2015).
- SSA countries have endorsed the Addis Ababa Action Agenda and are aligning national development plans to the SDGs and Africa Union’s Agenda 2063.

### Key findings on development performance
- SDG and human development gaps:
  - Median SDG index score across SSA countries is on average 25 percent lower than the median in other regions.
  - Income per capita (2018): US$1,574 (US$3,906 in PPP terms) for SSA.
  - Human Development Index: SSA average 0.52 versus 0.76 in the rest of the world.
- Sectoral performance (median outcomes across education, health, roads, electricity, water and sanitation):
  - SSA median outcomes are behind medians across EMEs.
  - Water and sanitation concentrate around an average SDG score of 49 out of 100.
- Historical improvements:
  - Between 1990 and 2017, net primary enrollment rate increased by almost 30 percentage points.
  - Infant mortality has been halved over the same period.
  - Life expectancy at birth increased by 22 percent.
  - Access to electricity improved but half of SSA population still lacks reliable electricity.

### Public spending and infrastructure trends
- 2015 public expenditure averages across SSA:
  - Education: 4.6 percent of GDP (2015 average across SSA countries).
  - Health: 2.8 percent of GDP (2015 average across SSA countries).
  - Comparison: health spending in other EMEs averages 4.2 percent of GDP; public education across SSA reaches 4.1 in 2015 in comparison.
- Public real capital stock per capita:
  - Almost flat between 1990 and 2015 in SSA.
  - By contrast, other LIDCs and emerging markets increased public real capital stock per capita on average by 5.4 percent and 3.4 percent per year, respectively, over the same period.

### Costing methodology and additional spending estimates
- Methodology:
  - Builds on IMF methodology to quantify additional spending in five sectors: education, health, water and sanitation, roads and electricity.
  - Estimates correspond to the additional total expenditure required in the year 2030 compared to 2016 spending.
  - Reported as percentage points of 2030 GDP and refer to total expenditure (public and private).
  - Benchmarks use high-performing countries with comparable GDP per capita; estimates assume high spending efficiency and are interpreted as a lower spending floor.
  - After 2030, education and health spending would be mostly recurrent, while infrastructure spending would decrease to cover depreciation of the capital stock built through 2030.
- Regional and country-scale findings:
  - Median sub-Saharan African country must spend an extra 18.8 percent in 2030 in education, health, water and sanitation, roads and electricity to achieve high development outcomes.
  - These additional needs are shared equally between human and social investment and physical capital investment.
  - Estimates cover both public and private spending in each sector, implying the need to mobilize all sources of financing—public and private, internal and domestic.
- Country examples:
  - Benin: additional spending estimate of 21 percent of GDP in 2030.
  - Rwanda: additional spending estimate of 19 percent of GDP in 2030.
  - Historical growth comparison: Between 1996 and 2017, Benin recorded an average growth rate of 4.4 percent against 8.1 percent in Rwanda.
  - Historical GDP per capita gap: Benin’s GDP per capita was 1.7 times higher than Rwanda’s in early 1990s; by mid-1990s onward Rwanda reduced the gap to a factor of 1.1.
  - Benin is on par with SSA averages on key state capacity and development indicators; Rwanda overperforms relative to SSA peers on those indicators.

### Policy implications and operational lessons
- Financing:
  - Achieving the SDGs will require mobilizing all financing sources—public and private, internal and domestic—given that additional spending estimates cover both public and private expenditures.
- Governance and institutions:
  - Strengthening national ownership of development goals, improving governance and the business environment are critical complements to financing.
- Prioritization:
  - Country-specific investment priorities must reflect differences in economic endowment, policy choices, and resilience to shocks, as illustrated by Benin and Rwanda’s differing paths and priorities.
- Efficiency:
  - Costing assumes high spending efficiency; therefore, improving spending efficiency remains essential to minimize required additional resource mobilization.

*Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019270-print-pdf.pdf*

### Box 1. Costing Methodology in Five SDG Sectors

### Box 1. Costing Methodology in Five SDG Sectors

### Methodology overview
- Three-step approach (Gaspar and others (2019)):
  - (i) identifying the main cost parameters, including inputs and their associated unit costs;
  - (ii) benchmarking the cost parameters to their levels in countries with comparable GDP per capita and reaching high development outcomes today; and
  - (iii) estimating the spending levels associated with these benchmarks, given individual country’s GDP per capital and population growth projections until 2030.

### Sector costing approaches
- Education
  - Total spending expressed as a function of: number of teachers, teacher salaries, share of non-compensatory current expenses, and capital spending.
  - Main inputs benchmarked against median values observed today in countries with comparable per-capita income and high education outcomes.
  - Education spending in 2030 estimated using benchmarked inputs and unit costs and countries’ projections for economic growth and school-age demographics, assuming full enrollment for at least 2 years of preprimary and tertiary education, and 12 years of primary and secondary education.
  - Benchmarking mapping: three 2016 GDP per capita buckets: i) between USD 0 and 3,000, ii) between USD 3,000 and 6,000 and iii) between USD 6,0000 and 15,000. High-performing low-income developing countries are those with an SDG4 education index above 80.
  - Assumed enrollment rates: target rates of 50 percent for preprimary and tertiary education, and 100 percent for primary and secondary education.

- Health
  - Total spending calculated as a function of: doctor salaries, number of doctors and other medical personnel, ratio of non-doctor to doctor wages, share of non-compensatory current expenses, and capital spending.
  - Main inputs benchmarked against median values observed today in countries with comparable per-capita income and high healthcare outcomes.
  - Health spending in 2030 estimated using benchmarked parameters with countries’ projections for growth and demographics.
  - Assumptions and imputations: ratio of non-doctor to doctor wage assumed to be 0.5; shares of capital and other current spending to total spending imputed using the World Bank income group averages.
  - High-performing low-income developing countries are those with an SDG3 health index above 70.

- Roads
  - Road density regressed on GDP per capita, population density, agriculture and manufacturing sector shares, urbanization rate, and the World Bank’s Rural Access Index (RAI), for a cross section of low-income countries and emerging economies.
  - RAI used as proxy for adequate access to the transport system.
  - Additional kilometers of roads needed estimated by raising the RAI to at least 75 percent in LIDCs, accounting for projected changes in population and GDP per capita through 2030.
  - Total cost derived by multiplying additional kilometers by a unit cost per kilometer set at a minimum of USD 500,000 (following Imi and others (2016)) unless country specific estimates suggest otherwise, and accounting for depreciation.

- Electricity
  - Additional electricity network corresponds to 100 percent access of projected population in 2030, accounting for an increase in per-capita consumption in line with GDP per capita.
  - Total cost estimated using unit cost per kilowatt of generation capacity set by the World Bank (2013) at USD 2,250.

- Water (WASH)
  - Costs derived using the WASH World Bank methodology described in Hutton and Varughese (2016).
  - Model estimates the cost of meeting SDG 6 WASH targets using unit costs calibrated at the country level, including capital investment, operations, and major capital maintenance to sustain infrastructure life span.

### Aggregated findings and regional comparisons
- Reporting conventions:
  - For education and health, results reported as the difference between the share of 2030 GDP in spending consistent with high performance and the current level of spending as a share of 2030 GDP.
  - For physical capital, spending to close the infrastructure gap between 2019 and 2030 is annualized and expressed in percent of 2030 GDP.
- Sub-Saharan Africa (SSA) median additional spending:
  - The median SSA country faces additional spending of about 19 percent of GDP.
  - Comparisons: median additional spending estimated at 12 percent of GDP for non-SSA LIDCs and 4 percent of GDP for EMEs.
- Sector distribution in SSA (five selected sectors)
  - Human capital sectors (health and education) represent about 9 percent of GDP.
  - Physical capital sectors (water, electricity, and roads) sum to 10 percent of GDP.
  - Additional physical infrastructure investment in SSA is 40 percent higher than average infrastructure needs across LIDCs and four times more than the EMEs’ average.
- Heterogeneity within SSA
  - Fragile countries: additional spending of 24 percent of GDP in 2030, which is 6 percentage points higher than non-fragile SSA countries.
  - No major difference between resource rich and non-resource rich countries in aggregate additional spending.
- Drivers of SSA vs other LIDCs/EMEs differences:
  - Education: SSA additional spending needs average 4.2 percent of 2030 GDP, against 0.3 percent in other LIDCs and EMEs. SSA demographics explain 91 percent of the difference; SSA projected student-age population share in 2030 is 47 percent versus 30 percent in other LIDCs and EMEs.
  - Health: SSA additional spending needs average 4.7 percent of 2030 GDP, against 1.5 percent in other LIDCs and EMEs. Almost half the difference driven by lower GDP per capita requiring a costlier combination of health staff density and salaries; remainder driven by factors including differences in current share of non-compensation spending.
  - Roads: SSA additional spending needs average 6.9 percent of 2030 GDP, against 0.9 percent in other LIDCs and EMEs. SSA lower 2030 GDP per capita and initial lower RAI explain 84 percent of the difference.

### Benin case study: additional spending and sectoral estimates
- Aggregate additional spending required to achieve SDGs in the five selected sectors: about 21 percent of GDP in 2030.
- Education
  - 2030 required spending: 8.7 percent of GDP (US $395 per student).
  - Current spending: 5.5 percent of GDP (US$135 per student).
  - Implications: decline in student-to-teacher ratio to 15.1 from 22.4 today; decline in share of teacher compensation in total spending from 60 percent to 45 percent to free resources for other current and capital expenditure.
  - Enrollment targets: improve secondary and tertiary enrollment and quality; primary already at high enrollment.
- Health
  - 2030 required spending: about 9.3 percent of GDP (US$119 per capita).
  - Current total spending: 4.2 percent of GDP (US$33 per capita).
  - Staffing implications: recruiting 8 times more doctors and 4 times more support staff than today.
  - Policy: Assurance pour le renforcement du capital humain (ARCH) launched to provide universal health insurance (see Box 2 summary in source).
- Roads
  - Estimated additional roads to build: 12,276 km.
  - Additional spending required: 8.1 percent of GDP in 2030 (of which 1.9 percent of GDP for maintenance cost).
  - Context: road transport accounts for 93 percent of people transport and 73 percent of goods transport; paved roads account for 45 percent of total roads.
  - Transport-related infrastructure to account for 25 percent of all investment under the PAG.
- Electricity
  - Annual cost to reach universal access: estimated at 2.4 percent of GDP.
  - Context: domestic production meets only 12 percent of domestic consumption; reliance on imports from Côte d'Ivoire, Ghana, and Nigeria.
  - Constraints: lower cost of imported electricity relative to domestic generation; low tariffs that only partially cover production costs.
  - PAG: electricity-related infrastructure to account for 10 percent of all investment under the PAG.
- Water and sanitation
  - Total cost of reaching universal access to safe and affordable drinking water and adequate sanitation: estimated at 2.5 percent of GDP per year—0.8 percent in water and 1.7 percent in sanitation—through 2030.
  - Sanitation accounts for 70 percent of the total estimate because only about 20 percent of the population has access to improved sanitation.
  - PAG target: reach universal access to water by 2021; 3 out of 45 flagship projects relate to this goal.

### Rwanda case study: context and preliminary assessment
- Policy alignment
  - National Strategy for Transformation (NST, 2017–2024) structured to align with SDGs and part of Vision 2050.
  - Line ministries have produced sectoral strategies and preliminary sectoral spending estimates; sector-specific costing estimates not reconciled with top-down multiyear NST budget estimates.
- Development outcomes and gaps
  - Life expectancy at birth reached 65 years in 2015.
  - Primary gross enrollment rate: 100 percent.
  - Remaining challenges (as of 2016/2018): 38 percent of children suffered from stunting; secondary gross enrollment rate at 40 percent; only 13 percent of the population has access to safely managed sanitation; 43 percent has access to electricity.
- Aggregate additional spending
  - The sizeable additional spending required to meet the SDGs in the five selected sectors amounts to about 19 percent of GDP.

*Source: Box 1. Costing Methodology in Five SDG Sectors (authors’ calculations based on Gaspar and others (2019)).*

### 2030. Appendix 1 provides detailed sectoral estimates and results are summarized below (Figure

### wpiea2019270-print-pdf - 2030. Appendix 1 provides detailed sectoral estimates and results are summarized below (Figure

### Sectoral additional spending estimates for Rwanda (2030)
- Education
  - Required spending to meet the SDGs: 7 percent of 2030 GDP.
  - Current public spending: 3.6 percent of GDP.
  - Student-to-teacher ratios: current 41 students per teacher vs. peers with better performance 15 students per teacher.
  - Government plan by 2030: nearly double current spending to reach 6.3 percent of GDP.
  - Assessment of education performance: SDG index 3 score 61 (just above the SSA median); lack of test-score data means this mainly reflects access rates.
  - Conclusion: Additional spending beyond planned levels would be needed to reach high performance in education by 2030.

- Health
  - Estimated additional spending needed to reach top-level performance: 2 percent of GDP by 2030.
  - Current per capita public spending increase planned by 2024: from US$38 to US$52 per capita per year (a 36 percent increase).
  - Current relative performance: better outcomes than the median peer in comparable income group.
  - Required change in input mix: overall increase in number of health workers, mostly doctors.
  - Data limitations: need more granular data on spending composition (compensation vs. other spending).

- Roads (Infrastructure)
  - Additional required spending: about 4 percent of 2030 GDP.
  - Rural Access Index (RAI): Rwanda’s RAI is 52 out of 100.
  - Current road network characteristics: only 5 percent paved; only 10 percent reaches all-season riding quality.
  - Investment focus: upgrading and maintaining existing roads, especially in remote and high areas.

- Electricity
  - Additional spending for universal access: 2 percent of GDP.
  - Current household electricity access: 43 percent.
  - Current situation: excess capacity and low access.
  - Assumptions for projection:
    - 60 percent of all new connections will be on-grid; 40 percent off-grid.
    - Electricity consumption growth: annual rate of 9 percent.
    - Projected consumption: 730 kWh in 2030 from an average of 238 kWh in 2016.

- Water and Sanitation
  - Required spending (World Bank WASH costing model): 4.5 percent of GDP per year to safely provide water to all households and provide fixed-point latrines to all.
  - Alignment: broadly in line with authorities’ own costing exercise, driven by current service coverage gaps and projections for a growing urban population with increased water consumption.

- Aggregate visualization (Figure 11 in source)
  - Sectors and percent of 2030 GDP listed: Health, Education, Electricity, Water, Roads (with numerical sector estimates detailed above).

### Tailoring costing estimates to country context
- Rationale
  - Variation across SSA countries justifies tailoring cost parameters to country circumstances.
  - Methodology enables benchmarking against SDG-high performing peers for input mix and potential spending efficiency gains.
- Parameters countries may tailor
  - Unit cost to build a km of road.
  - Unit cost to generate one kilowatt of electricity.
  - Input-mix variables (e.g., student-to-teacher ratio, number of doctors per 1,000 population).
- Process
  - Tailoring requires discussion with authorities and stakeholders and good understanding of national development challenges.
  - Costing exercises can catalyze coordination among development partners and ground development agendas in shared knowledge.
  - Example: analyses for Benin and Rwanda were refined during costing missions in 2018 involving line ministries and development partners.

### Baselines, data gaps, and monitoring
- Key step: establish clear baselines to measure and monitor progress.
- Example actions
  - Rwanda: line ministries and development partners gathering data on SDG target baselines; UNICEF and UKDFID helped establish common template in education.
  - Benin: authorities prioritized SDGs fitting their context and estimated needs with technical and financial partners.
- Data limitations
  - 60 percent of all 17 SDG indicators are currently not documented in SSA countries; existing data often outdated or not comparable across countries (SDG Center for Africa, 2019).
  - Lack of updated baselines complicates planning and transparency.

### Short-term implications: Prioritize, raise resources, and spend efficiently
- Use of costing
  - Rank and prioritize competing projects according to countries’ priorities and capacity to deliver.
  - Map spending gaps to existing sectoral projects and sort by implementation timeline and financing capacity.
- Financing strategy elements
  - Ramping up domestic resources is core: average tax-to-GDP ratio about 14 percent in 2017; room to reach the average of 18 percent for emerging economies.
  - Additional tax revenues alone insufficient to cover additional investments required by 2030.
  - Importance of official development assistance (ODA) and creating an enabling environment to attract private investment.
- Efficiency imperative
  - Costing assumes both more and more efficient spending.
  - If spending efficiency is not improved, additional spending required to reach the SDGs will be much larger.
  - Example capacity needs: Increasing spending in the medium term by an additional 19 percent of GDP would require substantial improvement in capacity to deliver public services and execute budgets.
  - Projected public spending trajectories: assuming current public-private split continues, Benin and Rwanda expected to more than double public spending in education and health by 2030.

### Medium-to-long term implications: anchoring in budgets and financing strategies
- Budgetary integration
  - 2030 spending estimates can anchor medium-term budget strategy and discussion of sustainable financing paths.
  - Rwanda: MTEF includes a subset of aggregated costs from sectoral National Strategy for Transformation (NST) chosen based on absorptive capacity and projected revenues.
- Functional classification and accountability
  - Presenting budget according to functional classification could strengthen accountability in delivering SDGs.
- Integrated National Financing Strategy (INFF) building blocks (Inter-Agency Task Force on Financing for Development, 2019)
  - Improve political stability, promote good governance, and improve business environment.
  - Map spending estimates to commensurate resources with a clear public/private investment split in each sector.
- Country examples and financing constraints
  - Rwanda: limited potential for additional tax revenues in the medium term—about 2 to 3 percent of GDP—and a declining trend in ODA not compensated by increasing private investment; MTEF includes “contingent” spending items that depend on additional resources.
  - Benin: room to increase tax revenue via consumption taxation (VAT and excises); reforms to boost private investment continue but private sector participation remains limited.
  - Recommendation: further country-level analysis to align national development plans with assessed spending needs and a comprehensive financing strategy detailing private and public sector roles.

### Appendix: costing methodology overview (sectoral input–outcome approach)
- General approach
  - Based on input-outcome approach developed by Gaspar and others (2019): development outcomes are a function of a mix of main cost factors.
  - For Benin and Rwanda, key cost factors and unit costs set at values observed in 2016 in countries with GDP per capita below US$3,000 that reach high development outcomes.
- Education costing components
  - Main factors: number of teachers, their wages, share of non-compensatory current expenses, share of capital expenses.
  - Benchmarked values set at medians observed in 2016 for countries with GDP per capita below US$3,000 and SDG index score above 80 in education.
  - 2030 assumptions: full enrollment for at least 2 years of preprimary and tertiary education and 12 years of primary and secondary education.
  - Appendix Table 1 (selected figures preserved from source):
    - GDP per capita examples: 2016 1,210; 2030 projections and groupings shown in table.
    - Students per teacher ratio examples: 30, 32, 15, 41, 15, 22, 15.
    - Teacher wages (ratio to GDP per capita) examples: 2.8, 3.7, 2.1, 1.5, 2.1, 2.7, 2.1.
    - Other current and capital spending (% total spending) examples: 44, 44, 55, 71, 55, 40, 55.
    - Student age population (% total population) examples: 56.9, 57.7, 39.2, 57, 47, 46, 39.
    - Enrollment rate (preprimary to tertiary) examples: 51, 73, 70, 73.
    - Results in table:
      - Total education spending (percent of GDP) examples: 7.6, 9.7, 9.8, 3.6, 10.7, 5.5, 8.8.
      - Spending per student (USD 2018 or latest) examples: 203, 219, 674, 89, 441, 135, 395.
      - SDG4 index examples: 54, 47, 86, 61, >80, 47, >80.
- Health costing components
  - Main factors: number of doctors and other medical personnel, salaries for doctors and other health staff, share of non-compensatory current expenses, share of capital expenses.
  - Benchmarked values set at medians observed in 2016 for countries with GDP per capita below US$3,000 and SDG index score above 70 in health.
  - 2030 estimates use benchmarked values plus country-specific projections for economic growth and demographics (Appendix Table 2 referenced).
- Roads costing components
  - Method: regression to derive determinants of network needs (road density regressed on GDP per capita, population density, agriculture and manufacturing shares, urbanization rate, and RAI) for a cross-section of low-income developing countries and emerging markets.
  - Use regression results to estimate additional kilometers needed to raise RAI to at least 75 percent, accounting for projected population and GDP per capita changes through 2030.
  - Compute total cost of additional road network and depreciation using country idiosyncratic inputs on unit costs and current baselines (Appendix Table 3 referenced).

*Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019270-print-pdf.pdf*

### Appendix Table 3. Estimates of Roads Spending in 2030 in Benin and Rwanda

### Appendix Table 3. Estimates of Roads Spending in 2030 in Benin and Rwanda

### Methodology
- Source: Authors’ calculations tailoring the costing methodology of Gaspar and others (2019) to Benin and Rwanda.
- Road spending estimates use country-specific main factors (road density, population density) and unit costs (USD per km).
- Unit cost (USD per km) reported in table: 1,100,000 and 609,905.

### Main factors (reported by country / scenario)
- GDP per capita (2016 / 2030):
  - Rwanda: 1,210 (2016) and 2,172 (2030)
  - Benin: 734 (2016) and 1,363 (2030)
  - Additional listed GDP per capita numbers: 915, 791, 1,206 (as presented in table)
- Road density (km per km2):
  - Values shown: 164, 406, 191, 678, 142, 277
- Population density (pop per km2):
  - Values shown: 75, 89, 504, 640, 104, 138
- Unit cost (USD per km):
  - --, 1,100,000, 609,905

### Results (reported figures)
- Additional annual spending (percent of 2030 GDP):
  - Values shown in table: --- , 3.9, -8.1
- of which depreciation (percent of 2030):
  - 1.5, 3.0
- Rural Access Index (RAI):
  - Values shown: 54, >75, 52, >75, 32, >62.5

### Contextual table entries (adjacent data appearing in same appendix)
- Electricity: unit cost per kilowatt set at US$2,250, following World Bank (2013).
- Water and sanitation: estimates derived using the WASH World Bank methodology (Hutton and Varughese 2016); unit costs calibrated at the country level (Appendix Table 5).

*Source: Authors’ calculations tailoring the costing methodology of Gaspar and others (2019) to Benin and Rwanda.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019270-print-pdf.pdf_