## 1.      Niger continues to face daunting development challenges and recent improvements

## Source details

**Canonical URL:** [1.      Niger continues to face daunting development challenges and recent improvements](https://www.imf.org/-/media/files/publications/selected-issues-papers/2023/english/sipea2023003.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/selected-issues-papers/2023/english/sipea2023003.pdf.md)
- [Structured JSON version](/-/media/files/publications/selected-issues-papers/2023/english/sipea2023003.pdf.json)

---

### Overview: development context and policy commitments
- GDP per capita remains below its 1980 level; Niger is one of the poorest countries globally.
- Human Development Index in 2022: ranked at the bottom 189th out of 191 countries.
- Under-five mortality rate: 78 per 1,000 live births.
- Life expectancy at birth: 63 years.
- Poverty incidence in 2021: 41.4 percent; about 10 million Nigeriens living below the national poverty line; two thirds of the poor live in Zinder, Maradi, and Dosso.
- Population growth: more than 3 percent per year, creating pressures on social spending and hindering women’s empowerment due to high fertility rates.
- Government commitments:
  - Scale up and preserve social spending to support poor and vulnerable populations.
  - Education: build durable classrooms and residential school facilities for girls; revise hiring process and career prospects of teachers to attract skilled teachers.
  - Social safety nets: increase coverage—especially for the most vulnerable—and reduce gender inequality; improve targeting by creating a unified social registry.
- IMF program emphasis:
  - Program objective: improve effectiveness of public spending, emphasizing social spending and poverty-reduction expenditure; program includes an on-budget floor on social spending.
  - Emphasis on revenue mobilization, quality of spending, and re-prioritization to create fiscal space for education, health, and social protection and to improve their efficiency.

### Education outcomes and efficiency
- Data and methodology:
  - Analysis based on FAD’s Expenditure Assessment Tool (EAT); dataset covers indicators from around 2000 until 2018-19.
- Key indicators:
  - Adult literacy rate: 35 percent (implying 6.5 adults over 10 are illiterate).
  - WAEMU adult literacy: 45 percent.
  - SSA and LIDCs adult literacy: 66.4 percent.
  - Net enrollment: primary 65.1 percent; secondary 20.1 percent.
- Spending and per-student metrics:
  - Education spending declined to 3.8 percent of GDP in 2020 after increases over the 2000s.
  - Niger’s education spending percent of GDP: above SSA average but slightly below WAEMU average of 4 percent of GDP.
  - Spending per student in primary and secondary education: less than half that of SSA and LIC countries.
  - Teacher-student ratio: lower than peers; secondary teacher-student ratio noted as 3.4 per 100 students.
  - Spending per student at tertiary level is much higher than for other levels.
- Efficiency findings:
  - Niger is well below the efficiency frontier for various measures of education spending and outcomes (DEA used).
  - Examples: Ethiopia primary net enrollment 84.6 percent vs Niger 65.1 percent, despite equal or lower spending per student in those countries.
  - Comparable teacher-student ratios in LIDC and SSA countries correlate with better outcomes (around 80 percent net enrollment in primary and close to 30 percent in secondary, versus Niger’s 65 percent and 20 percent).
  - The lower secondary teacher-student ratio (3.4 per 100 students) may explain part of the net enrollment gap.

### Health outcomes and efficiency
- Health spending and outcomes:
  - Health spending averaged 7.4 percent of GDP over the last two decades for Niger.
  - WAEMU average: 5.5 percent of GDP; LIDCs average: 5.9 percent of GDP.
  - Expenditure per capita: still below peers.
  - Niger’s ratio of physicians: 0.019 per 1000 people (one of the lowest in the world).
  - LIDCs and SSA countries exhibit fewer infant deaths, higher life expectancy, more hospital beds, nurses, and physicians.
- Efficiency findings:
  - Spending per capita and Health Adjusted Life Expectancy (HALE) are lower than LIDC and SSA countries.
  - Some countries (e.g., Madagascar, Ethiopia, Malawi) spend less per capita but have slightly better outcomes.
  - Inefficiencies in health exist but are less pronounced than in education (lower distance to the efficiency frontier using DEA).

### Overview of social protection programs
- Main SPL program areas: social safety nets, social insurance, labor market measures, general subsidies.
- Trends (2008–2017):
  - Social insurance expenses more than doubled.
  - Safety nets tripled.
  - General subsidies (particularly petrol) declined by 78 percent.
- Social safety net components:
  - In-kind distributions: targeted food distribution and subsidized sales of cereals using community-based Household Economy Approach (HEA); geographic targeting for transiently food insecure households; OPVN handles sales of cereals at subsidized prices.
  - Nutrition programs: focus on chronic malnutrition solutions.
  - Unconditional cash transfers:
    - National Safety Net Project (NSNP) and WFP Resilience approach.
    - ASNP 2 (launched 2019): 120,000 households receive monthly transfer of 15,000 CFA for 24 months.
    - NSNP uses geographic targeting followed by proxy means test to identify poor households.
    - WFP targets poorest crisis-affected households over three years using HEA.
  - Conditional cash transfers for education: cover all education levels; largest share to university scholarships and allowances; programs to keep children in secondary and primary school; specific programs to support girls’ continuation and reduce early marriage via enrollment/attendance and family participation conditions.
  - Cash for work: temporary public works to provide income and skills, protect assets and consumption.
  - School feeding programs: run in over 2,000 primary schools; WFP supports school canteens in 7 out of 8 regions.
  - Targeted agricultural subsidies: free distribution of seeds and animal restocking; sales of livestock feed and fertilizers at subsidized prices; targeting issues noted; FAO is a leading partner.
  - Free health care for pregnant women and children under 5.
- COVID-19 scaling and evidence:
  - Adaptive social protection scaled up during COVID-19; emergency cash transfers covered about 65 percent of individuals impacted by the crisis.
  - World Bank impact evaluation: cash transfers induced lasting investments in productive assets and activities among the very poor and built resilience to shocks.
  - IMF staff analysis: cash transfers are more effective than fertilizer subsidies at safeguarding household welfare.

### Profile, impact, and efficiency of SPL (SPL-AT analysis)
- Methodology: Social Protection and Labor Assessment Tool (SPL-AT) using PovcalNet, ASPIRE, ILO, IMF WEO, etc., to analyze coverage, adequacy, targeting, and efficiency.
- Spending levels and composition (2017):
  - Social assistance spending: 0.42 percent of GDP in Niger.
  - Comparators/medians: SSA median 0.89 percent of GDP; LIDC median 1.15 percent of GDP; WAEMU median 0.69 percent of GDP; Senegal 1.85 percent of GDP; Burkina Faso (median) 2.46 percent of GDP.
  - Niger’s social assistance spending concentrated on food and in-kind assistance: 0.36 percent of GDP (84.8 percent of all social assistance expenditures).
  - Other components in 2017: conditional cash transfers 0.04 percent of GDP; school feeding 0.02 percent of GDP; public works 0.01 percent of GDP.
- Coverage and targeting:
  - SPL coverage in 2017: 21.3 percent of the population.
  - Coverage lower relative to WAEMU median but similar to SSA and LIDC medians.
  - Approximately 80 percent of Nigeriens are not covered by any SPL program.
  - Very low coverage for labor market and social insurance programs: less than 1.5 percent.
  - Private transfers reach 46.1 percent of the population—higher than comparators.
  - Coverage by quintile: bottom quintile coverage 17.2 percent; other quintiles between 19.5 and 27.3 percent (top quintile 19.5 percent; third quintile 27.3 percent).
  - The richest quintile is more covered by SPL benefits than the poorest quintile.
  - The poorest are less covered relative to SSA and LIDC medians.
  - Difficulties in reaching populations in conflict-affected areas are noted as an impediment.
- Adequacy and poverty impact:
  - Total adequacy of benefits: 3.4 percent of beneficiaries’ income.
  - Comparators/medians: WAEMU 21.7 percent; SSA 39.1 percent; LIDC 33.4 percent.
  - Poverty incidence context: 40.8 percent.
  - Adequacy by program type: social assistance 1.6 percent; in-kind programs 1.1 percent; social insurance 13.1 percent; private transfers 8.6 percent.
  - Conclusion: despite nominal increases in social protection expenditures (more than doubled over 2008-2017), funding remains insufficient to significantly reduce poverty or cover the population in need; scope exists to increase efficiency through better targeting.

### Adequacy and coverage of SPL benefits (detailed)
- Adequacy for bottom quintile: 3.8 percent.
- Adequacy for top quintile: 4.2 percent.
- Adequacy for overall SPL programs for the bottom quintile in Niger is 7 times lower than SSA median.
- Adequacy for overall SPL programs for the bottom quintile in Niger is 19 times lower than LIDC median.
- Private transfers adequacy for bottom quintile: 13.3 percent (almost like SSA median but more than 3 times lower than LIDC median).
- Conclusion: SPL programs provide relatively weak support for poor and vulnerable populations in Niger.

### Beneficiary incidence and benefit incidence (targeting outcomes)
- Share of beneficiaries in bottom (poorest) quintile: 16.1 percent (lowest share across quintiles).
- Share of beneficiaries by quintile:
  - Q3 (third quintile): 25.4 percent.
  - Q4 (fourth quintile): 20.9 percent.
  - Q5 (top/richest quintile): 18.2 percent.
- Share of total SPL benefits received by bottom quintile: 7 percent.
  - This is 7 times lower than the share received by the top quintile.
- Finding: Richer segments are more likely to participate in SPL programs than poorer ones; SPL programs are regressive and half of SPL benefits go to the richer and wealthier.

### Simulated impacts and efficiency of SPL programs
- Overall SPL simulated reductions (Niger):
  - Poverty headcount reduction: about 2 percent.
  - Poverty gap reduction: close to 1.8 percent.
  - Gini (inequality) change: increase by 0.07 percent.
- Comparators:
  - SSA median: poverty headcount reduction 2.7 percent; poverty gap reduction 5.07 percent; Gini reduction 0.41 percent.
  - LIDC median: poverty headcount reduction 7.3 percent; poverty gap reduction 14.06 percent; Gini reduction 2.46 percent.
  - WAEMU median: poverty headcount reduction 0.6 percent; poverty gap reduction 0.82 percent; Gini increase 0.79 percent.
- Interpretation: SPL programs in Niger are less efficient at reducing poverty and inequality than SSA and LIDC medians; they perform better than WAEMU median on some metrics but are somewhat regressive.

### Performance by program type
- Social assistance (in-kind and public works) simulated impacts (Niger):
  - Poverty headcount reduction: 1.71 percent.
  - Poverty gap reduction: 1.56 percent.
  - Gini reduction: 0.28 percent.
  - These impacts are lower than SSA and LIDC medians, but higher than WAEMU medians.
- Contributory pensions simulated impacts (Niger):
  - Poverty headcount reduction: 0.34 percent.
  - Poverty gap reduction: 0.21 percent.
  - Gini change: increase by 0.35 percent.
- Comparator medians for contributory pensions:
  - WAEMU: poverty headcount reduction 0.46 percent; poverty gap reduction 0.82 percent; Gini increase 0.72 percent.
  - SSA: poverty headcount reduction 1.36 percent; poverty gap reduction 1.92 percent; Gini change 0.01 percent.
  - LIDC: poverty headcount reduction 5.17 percent; poverty gap reduction 9.24 percent; Gini reduction 1.76 percent.
- Conclusion: Contributory pensions in Niger have low impact on poverty and may increase inequality.

### Determinants of access and efficiency of targeting (EHCVM 2018–19 analysis)
- Data: Harmonized Survey on Households Living Standards 2018–19 (EHCVM) covering more than 6,000 households.
- Reported access rates:
  - Food programs (any household, last 12 months): 12 percent.
  - Food programs (households that experienced any shocks): 14.9 percent.
  - Health programs (any household): 38 percent.
  - Government cash transfers (any household): 0.71 percent.
  - Government cash transfers (households that suffered any shocks): 0.83 percent.
- Regression findings (linear probability/OLS):
  - Evidence of geographical targeting for food and health programs (regional dummy significance), but not for government cash transfers.
  - Poverty, hunger/food insecurity, and household characteristics generally do not explain access to social programs.
  - Households affected by shocks have higher likelihood to participate in programs.
  - Indications of leakage: non-poor households and households in regions with fewer poor have higher likelihood to benefit.
- Specific program findings:
  - Food programs:
    - Poverty and food insecurity indicators tend not to be significant predictors of access.
    - HH head tertiary education reduces likelihood of inclusion: -0.079*** (0.019) and -0.060* (0.024) in reported specifications.
    - Family shocks, natural disaster shocks, and conflict shocks increase likelihood of participation.
    - Compared to Niamey, households in Diffa and Dosso have higher likelihood of selection; those in Zinder and Maradi are less likely.
  - Health programs:
    - Households in Zinder and Maradi (higher poverty) are more likely to receive health programs.
    - Inverted U relationship between log consumption per adult equivalent and selection probability: coefficient 0.872* (0.385) and squared term -0.035* (0.014) in one specification.
    - Food insecurity does not determine health program targeting.
    - Agriculture and non-agriculture employment shocks increase likelihood; natural disaster shocks decrease likelihood.
    - Other notable coefficients: HH size 0.024*** (0.003); HH gender (1 = Female) -0.059** (0.022); HH head age -0.015*** (0.004).
  - Government cash transfers:
    - Targeting issues pronounced; poverty, food insecurity, and most household characteristics do not predict access (except age of household head).
    - Conflict shocks increase likelihood: 0.032** (0.011) and 0.034** (0.011).
    - Regional examples: Diffa 0.040*** (0.010) / 0.043*** (0.010); Dosso -0.008* (0.004); Tillaberi -0.007* (0.003).
- Notes on regression tables:
  - Model: Linear probability models for determinants of access to food programs, health programs, and government transfers.
  - Columns: (1) poverty criteria only; (2) hunger and food insecurity criteria only; (3) shock criteria only; (4) other household characteristics criteria only; (5) all criteria included.
  - Significance notation: * 10% level, ** 5% level, *** 1% level. Robust standard errors in parentheses.
  - Number of households entries appear as repeated sequences: 59985998599859985998 and 5998599859985998 as shown in the source.

### Education outcomes and conditional cash transfers (CCTs)
- Recommendation: Step up conditional cash transfers for primary and secondary education to reduce elevated dropout rates.
- Recommendation: Extend existing programs supporting girls’ education to include more beneficiaries.
- Recommendation: Continue supplementary workshops to educate parents and reduce incentives for early marriage.

### Fiscal space, governance, and program targeting recommendations
- Fiscal space and social spending:
  - Finding: Niger has lower expenditure per capita in education and health compared to peers, and very low social assistance spending.
  - Recommendation: Create fiscal space to scale-up social spending through reprioritization within total spending, given strong population growth.
  - Recommendation: Complement increased funding by improving the quality of education and healthcare.
  - Recommendation: Implement policies aimed at reducing informality to improve low coverage in social insurance.
- Governance, public financial management, and procurement:
  - Recommendation: Emphasize transparency and competitive bidding in public procurement.
  - Recommendation: Reinforce the capacity of the Cour des Comptes to audit a very large share of government’s expenditure through public procurement and to fight corruption.
- Targeting and program design:
  - Move away from untargeted subsidies (e.g., distribution of food and fertilizers at moderate prices) toward targeted social safety nets.
  - Aim for full coverage of the poorest and most vulnerable (starting with those at the bottom quartile) using means testing or proxy-means-testing, combined with geographic methods as needed.
  - Improve individual and geographic targeting through establishment of a unified social registry (USR) with World Bank assistance; progress ongoing on institutional, regulatory, and operational frameworks for the USR.
  - Improve geographic targeting in food and government cash transfer programs toward regions where poverty is higher.
  - Reallocate benefits away from richer segments of the population to address regressivity and improve adequacy and coverage for the most vulnerable.
  - Better balance shock-response and long-term social assistance:
    - Increase funding for long-term assistance programs to reduce poverty.
    - Orient social assistance toward enhancing productivity and resilience via cash transfers used for lasting investments and income-generating activities rather than only post-crisis support.
- Evidence on targeting effectiveness:
  - Schnitzer (2019): Proxy-means testing (PMT) performs more effectively in identifying persistently poor households; Household Economy Analysis (HEA) performs better for transiently food insecure households.
  - Premand and Schnitzer (2021): A large-scale randomized experiment confirms PMT is more efficient in identifying households with lower consumption per capita (poor households).
  - Finding: PMT and a formula to identify the food-insecure (FCS) can improve targeting and efficiency relative to community-based targeting (CBT), which is more affected by manipulation and information imperfections.

*Prepared by Jean-Marc Atsebi (FAD) and Paola Ganum (AFR); IMF staff analysis and tools (FAD EAT, SPL-AT) cited in source.*

### 1.      Niger continues to face daunting development challenges and recent improvements

### 1.      Niger continues to face daunting development challenges and recent improvements

### Overview: development context and policy commitments
- GDP per capita remains below its 1980 level; Niger is one of the poorest countries globally.
- Human Development Index in 2022: ranked at the bottom 189th out of 191 countries.
- Under-five mortality rate: 78 per 1,000 live births.
- Life expectancy at birth: 63 years.
- Poverty incidence in 2021: 41.4 percent, with about 10 million Nigeriens living below the national poverty line; two thirds of the poor live in Zinder, Maradi, and Dosso.
- Population growth: more than 3 percent per year, generating pressures on social spending and hindering women’s empowerment due to high fertility rates.
- Government commitments:
  - Scale up and preserve social spending to support poor and vulnerable populations.
  - Education: increase quantity and quality of education by building durable classrooms and residential school facilities for girls; revise hiring process and career prospects of teachers to attract skilled teachers.
  - Social safety nets: increase coverage—especially for the most vulnerable—and reduce gender inequality; improve targeting by creating a unified social registry.
- IMF support and program emphasis:
  - Program objective: improve effectiveness of public spending, emphasizing social spending and poverty-reduction expenditure; program includes an on-budget floor on social spending.
  - Emphasis on revenue mobilization, quality of spending, and re-prioritization to create fiscal space for education, health, and social protection and to improve their efficiency.

### Education outcomes and efficiency
- Data and methodology:
  - Analysis based on FAD’s Expenditure Assessment Tool (EAT); dataset covers indicators from around 2000 until 2018-19 (latest available year for many indicators).
- Key education indicators:
  - Adult literacy rate: 35 percent (implying 6.5 adults over 10 are illiterate).
  - WAEMU adult literacy comparison: 45 percent.
  - SSA and LIDCs adult literacy comparison: 66.4 percent.
  - Net enrollment: primary 65.1 percent; secondary 20.1 percent.
- Spending and per-student metrics:
  - Education spending declined to 3.8 percent of GDP in 2020 after increases over the 2000s.
  - Niger’s education spending percent of GDP: above SSA average but slightly below WAEMU average of 4 percent of GDP.
  - Spending per student in primary and secondary education: less than half that of SSA and LIC countries.
  - Teacher-student ratio: lower than peers; specific secondary teacher-student ratio noted as 3.4 per 100 students (see efficiency discussion).
  - Spending per student at tertiary level is much higher than for other levels.
- Efficiency findings:
  - Niger is situated well below the efficiency frontier for various measures of education spending and outcomes (Data Envelopment Analysis used to construct frontier).
  - Several countries (e.g., Ethiopia, Rwanda, Madagascar, Malawi, Burundi) spend equal or less per student but achieve better net enrollment rates (example: Ethiopia primary net enrollment 84.6 percent vs Niger 65.1 percent).
  - Comparable teacher-student ratios in LIDC and SSA countries correlate with better outcomes (around 80 percent net enrollment in primary and close to 30 percent in secondary, compared to Niger’s 65 percent and 20 percent).
  - The lower teacher-student ratio in secondary education (3.4 per 100 students) may explain part of the net enrollment gap.

### Health outcomes and efficiency
- Health spending and outcomes:
  - Health spending averaged 7.4 percent of GDP over the last two decades for Niger.
  - WAEMU average: 5.5 percent of GDP; LIDCs average: 5.9 percent of GDP.
  - Expenditure per capita: still below peers.
  - Niger’s ratio of physicians: 0.019 per 1000 people (one of the lowest in the world).
  - LIDCs and SSA countries exhibit fewer infant deaths, higher life expectancy, more hospital beds, nurses, and physicians.
- Efficiency findings:
  - Spending per capita and Health Adjusted Life Expectancy (HALE) are lower than LIDC and SSA countries.
  - Some countries (e.g., Madagascar, Ethiopia, Malawi) spend less per capita but have slightly better outcomes.
  - Inefficiencies in health exist but are less pronounced than in education, indicated by a lower distance to the efficiency frontier (DEA methodology applied).

### Overview of social protection programs
- Main SPL program areas: social safety nets, social insurance, labor market measures, general subsidies.
- Trends (2008–2017): social insurance expenses more than doubled; safety nets tripled; general subsidies (particularly petrol) declined by 78 percent.
- Social safety net components and characteristics:
  - In-kind distributions: targeted food distribution and subsidized sales of cereals using community-based Household Economy Approach (HEA) for vulnerable areas; geographic targeting for transiently food insecure households; OPVN handles sales of cereals at subsidized prices.
  - Nutrition programs: focus on chronic malnutrition solutions.
  - Unconditional cash transfers: National Safety Net Project (NSNP) and WFP Resilience approach.
    - ASNP 2 (launched 2019) increased cash transfers: 120,000 households receive monthly transfer of 15,000 CFA for 24 months.
    - NSNP uses geographic targeting followed by proxy means test to identify poor households.
    - WFP targets poorest crisis-affected households over three years using HEA.
  - Conditional cash transfers for education: cover all education levels; largest share to university scholarships and allowances; programs to keep children in secondary and primary school; specific programs to support girls’ continuation and reduce early marriage via enrollment/attendance conditions and family participation conditions.
  - Cash for work: temporary public works to provide income and skills, protect assets and consumption.
  - School feeding programs: run in over 2,000 primary schools; WFP supports school canteens in 7 out of 8 regions.
  - Targeted agricultural subsidies: free distribution of seeds and animal restocking; sales of livestock feed and fertilizers at subsidized prices; targeting issues noted; FAO is a leading partner.
  - Free health care for pregnant women and children under 5.
- COVID-19 scaling and evidence:
  - Adaptive social protection scaled up during COVID-19; emergency cash transfers covered about 65 percent of individuals impacted by the crisis.
  - World Bank impact evaluation: cash transfers induced lasting investments in productive assets and activities among the very poor and built resilience to shocks.
  - IMF staff analysis: cash transfers are more effective than fertilizer subsidies at safeguarding household welfare.

### Profile, impact, and efficiency of Social Protection and Labor (SPL)
- Methodology:
  - Analysis uses Social Protection and Labor Assessment Tool (SPL-AT), drawing on World Bank PovcalNet and ASPIRE, ILO, IMF WEO, etc., to analyze coverage, adequacy, targeting, and efficiency.
- Spending levels and composition:
  - 2017 social assistance spending: 0.42 percent of GDP in Niger.
  - Comparators/medians: SSA median 0.89 percent of GDP; LIDC median 1.15 percent of GDP; WAEMU median 0.69 percent of GDP; Senegal 1.85 percent of GDP; Burkina Faso (median) 2.46 percent of GDP (chart labels reflected).
  - Niger’s social assistance spending is concentrated on food and in-kind assistance: 0.36 percent of GDP (84.8 percent of all social assistance expenditures).
  - Other components in 2017: conditional cash transfers 0.04 percent of GDP; school feeding 0.02 percent of GDP; public works 0.01 percent of GDP.
- Coverage and targeting:
  - SPL coverage in 2017: 21.3 percent of the population.
  - Coverage is lower relative to WAEMU median but similar to SSA and LIDC medians.
  - Approximately 80 percent of Nigeriens are not covered by any SPL program.
  - Very low coverage for labor market and social insurance programs: less than 1.5 percent.
  - Private transfers reach 46.1 percent of the population—higher than comparators.
  - Coverage by quintile: bottom (poorest) quintile coverage 17.2 percent; other quintiles between 19.5 and 27.3 percent (top quintile 19.5 percent; third quintile 27.3 percent).
  - The richest quintile is more covered by SPL benefits than the poorest quintile.
  - The poorest are less covered relative to SSA and LIDC medians.
  - Difficulties in reaching populations in conflict-affected areas are noted as an impediment (not fully captured in analysis).
- Adequacy and poverty impact:
  - Total adequacy of benefits: 3.4 percent of beneficiaries’ income (i.e., total transfer amount received by beneficiaries as a share of their total income).
  - Comparators/medians: WAEMU 21.7 percent; SSA 39.1 percent; LIDC 33.4 percent.
  - Poverty incidence context: 40.8 percent (noted in adequacy discussion).
  - Adequacy by program type: social assistance 1.6 percent; in-kind programs 1.1 percent; social insurance 13.1 percent; private transfers 8.6 percent.
  - Conclusion: despite nominal increases in social protection expenditures (more than doubled over 2008-2017), funding remains insufficient to significantly reduce poverty or cover the population in need; scope exists to increase efficiency through better targeting.

*Prepared by Jean-Marc Atsebi (FAD) and Paola Ganum (AFR); IMF staff analysis and tools (FAD EAT, SPL-AT) cited in source.*

### 19.      In addition, the adequacy of SPL benefits is relatively lower for the poorest (Figure 7).

### sipea2023003 - 19.      In addition, the adequacy of SPL benefits is relatively lower for the poorest (Figure 7)

### Adequacy and coverage of SPL benefits
- Adequacy for bottom quintile: 3.8 percent.
- Adequacy for top quintile: 4.2 percent.
- Adequacy for overall SPL programs for the bottom quintile in Niger is 7 times lower than SSA median.
- Adequacy for overall SPL programs for the bottom quintile in Niger is 19 times lower than LIDC median.
- For private transfers programs:
  - Adequacy for bottom quintile: 13.3 percent.
  - This is almost like SSA median but more than 3 times lower than LIDC median.
- Conclusion: SPL programs provide relatively weak support for poor and vulnerable populations in Niger.

### Beneficiary incidence and benefit incidence (targeting outcomes)
- Share of beneficiaries in bottom (poorest) quintile: 16.1 percent (lowest share across quintiles).
- Share of beneficiaries by quintile:
  - Q3 (third quintile): 25.4 percent.
  - Q4 (fourth quintile): 20.9 percent.
  - Q5 (top/richest quintile): 18.2 percent.
- Share of total SPL benefits received by bottom quintile: 7 percent.
  - This is 7 times lower than the share received by the top quintile.
- Finding: Richer segments are more likely to participate in SPL programs than poorer ones; SPL programs are regressive and half of SPL benefits go to the richer and wealthier.

### Simulated impacts and efficiency of SPL programs
- Overall SPL simulated reductions (Niger):
  - Poverty headcount reduction: about 2 percent.
  - Poverty gap reduction: close to 1.8 percent.
  - Gini (inequality) change: increase by 0.07 percent.
- Comparators:
  - SSA median: poverty headcount reduction 2.7 percent; poverty gap reduction 5.07 percent; Gini reduction 0.41 percent.
  - LIDC median: poverty headcount reduction 7.3 percent; poverty gap reduction 14.06 percent; Gini reduction 2.46 percent.
  - WAEMU median: poverty headcount reduction 0.6 percent; poverty gap reduction 0.82 percent; Gini increase 0.79 percent.
- Interpretation: SPL programs in Niger are less efficient at reducing poverty and inequality than SSA and LIDC medians; they perform better than WAEMU median on some metrics but are somewhat regressive.

### Performance by program type
- All social assistance (in-kind and public works) simulated impacts (Niger):
  - Poverty headcount reduction: 1.71 percent.
  - Poverty gap reduction: 1.56 percent.
  - Gini reduction: 0.28 percent.
  - These impacts are lower than SSA and LIDC medians, but higher than WAEMU medians.
- Contributory pensions simulated impacts (Niger):
  - Poverty headcount reduction: 0.34 percent.
  - Poverty gap reduction: 0.21 percent.
  - Gini change: increase by 0.35 percent.
- Comparator medians for contributory pensions:
  - WAEMU: poverty headcount reduction 0.46 percent; poverty gap reduction 0.82 percent; Gini increase 0.72 percent.
  - SSA: poverty headcount reduction 1.36 percent; poverty gap reduction 1.92 percent; Gini change 0.01 percent (neutral).
  - LIDC: poverty headcount reduction 5.17 percent; poverty gap reduction 9.24 percent; Gini reduction 1.76 percent.
- Conclusion: Contributory pensions in Niger have low impact on poverty and may increase inequality.

### Determinants of access and efficiency of targeting (EHCVM 2018–19 analysis)
- Data: Harmonized Survey on Households Living Standards 2018–19 (EHCVM) covering more than 6,000 households.
- Reported access rates:
  - Food programs (any household, last 12 months): 12 percent.
  - Food programs (households that experienced any shocks): 14.9 percent.
  - Health programs (any household): 38 percent.
  - Government cash transfers (conditional and unconditional; any household): 0.71 percent.
  - Government cash transfers (households that suffered any shocks): 0.83 percent.
- Regression (linear probability/OLS) findings on targeting determinants:
  - Some evidence of geographical targeting for food and health programs (regional dummy significance), but not for government cash transfers.
  - Poverty, hunger/food insecurity, and household characteristics generally do not explain access to social programs.
  - Households affected by shocks have higher likelihood to participate in programs.
  - Indications of leakage: non-poor households and households in regions with fewer poor have higher likelihood to benefit.
- Specific program findings:
  - Food programs:
    - Poverty and food insecurity indicators tend not to be significant predictors of access.
    - Households with household head having tertiary education are less likely to be included than those with no education.
    - Family shocks, natural disaster shocks, and conflict shocks increase likelihood of participation.
    - Agricultural shocks do not determine targeting of food programs.
    - Geographical note: compared to Niamey, households in Diffa and Dosso (fewer poor) have higher likelihood of selection; those in Zinder and Maradi (more poor) are less likely to receive food programs.
  - Health programs:
    - Households in Zinder and Maradi (higher poverty) are more likely to receive health programs.
    - Inverted U relationship between household consumption per adult equivalent and selection probability for health programs (both poorest and richest less likely; middle more likely).
    - Food insecurity does not determine health program targeting.
    - Agricultural and non-agricultural employment shocks increase likelihood; natural disaster shocks decrease likelihood.
  - Government cash transfers:
    - Targeting issues pronounced; neither poverty, food insecurity, nor most household characteristics predict access (except age of household head).
    - Only households experiencing conflict shocks have higher likelihood of participation.
    - Caveat: analysis cannot distinguish conditional vs. unconditional cash transfers; targeting issues are especially relevant for unconditional transfers.

### Policy recommendations and concluding remarks
- Move away from untargeted subsidies (e.g., distribution of food and fertilizers at moderate prices) toward targeted social safety nets.
- Aim for full coverage of the poorest and most vulnerable (starting with those at the bottom quartile) using means testing or proxy-means-testing of social assistance transfers, combined with geographic methods as needed.
- Improve individual and geographic targeting through establishment of a unified social registry (USR) with World Bank assistance to consolidate beneficiary information and improve efficiency and targeting; progress is ongoing on institutional, regulatory, and operational frameworks for the USR.
- Improve geographic targeting in food and government cash transfer programs toward regions where poverty is higher.
- Reallocate benefits away from richer segments of the population to address regressivity and improve adequacy and coverage for the most vulnerable.
- Better balance shock-response and long-term social assistance:
  - Increase funding for long-term assistance programs to reduce poverty.
  - Orient social assistance toward enhancing productivity and resilience via cash transfers used for lasting investments and income-generating activities rather than only post-crisis support.

*Sources: FAD EP SPL-AT tool, and IMF staff calculations*

### 32.      Supporting better education outcomes, particularly for girls, would require an increase

### 32.      Supporting better education outcomes, particularly for girls, would require an increase

### Education outcomes and conditional cash transfers (CCTs)
- Recommendation: Step up conditional cash transfers for primary and secondary education to reduce elevated dropout rates.
- Recommendation: Extend existing programs supporting girls’ education to include more beneficiaries.
- Recommendation: Continue supplementary workshops to educate parents and reduce incentives for early marriage.

### Fiscal space and social spending
- Finding: Niger has lower expenditure per capita in education and health compared to peers, and very low social assistance spending.
- Recommendation: Create fiscal space to scale-up social spending through reprioritization within total spending, given strong population growth.
- Recommendation: Complement increased funding by improving the quality of education and healthcare.
- Recommendation: Implement policies aimed at reducing informality to improve low coverage in social insurance.

### Governance, public financial management, and procurement
- Recommendation: Emphasize transparency and competitive bidding in public procurement.
- Recommendation: Reinforce the capacity of the Cour des Comptes to audit a very large share of government’s expenditure through public procurement and to fight corruption.

### Evidence on targeting effectiveness (qualitative citations)
- Schnitzer (2019): Proxy-means testing (PMT) performs more effectively in identifying persistently poor households; Household Economy Analysis (HEA) performs better for transiently food insecure households.
- Premand and Schnitzer (2021): A large-scale randomized experiment confirms PMT is more efficient in identifying households with lower consumption per capita (poor households).
- Finding: PMT and a formula to identify the food-insecure (FCS) can improve targeting and efficiency relative to community-based targeting (CBT), which is more affected by manipulation and information imperfections.

### Regression evidence — determinants of targeting (key statistically significant estimates preserved exactly)
- Table 1 (Access to food programs; linear probability model; robust standard errors in parentheses):
  - Family shocks (1 = Yes): 0.031* (0.013) and 0.029* (0.013) in specifications reported.
  - Natural shocks (1 = Yes): 0.037** (0.012) and 0.028* (0.012).
  - Conflict shocks (1 = Yes): 0.089* (0.035) and 0.088* (0.036).
  - Other shocks (1 = Yes): 0.044** (0.017) and 0.048** (0.016).
  - HH size: 0.004* (0.002) and 0.003 (0.002).
  - HH head level of education (3 = Tertiary): -0.079*** (0.019) and -0.060* (0.024).
  - Regional coefficients (examples): Agadez 0.099*** (0.014) / 0.108*** (0.013); Diffa 0.232*** (0.021) / 0.249*** (0.020); Dosso 0.214*** (0.019) / 0.235*** (0.018).

- Table 2 (Access to health programs; linear probability model):
  - Log of consumption per adult eq.: 0.872* (0.385) in one specification; squared term -0.035* (0.014).
  - Do not eat healthy and nutritive food (1 = Yes): -0.036* (0.018) in one specification.
  - Family shocks (1 = Yes): 0.050** (0.018) in a specification.
  - Natural shocks (1 = Yes): -0.034* (0.017) and -0.041* (0.017).
  - Agriculture shocks (1 = Yes): 0.057** (0.018) and 0.044* (0.018).
  - Non-Agri employment shocks (1 = Yes): 0.149*** (0.036) and 0.124*** (0.035).
  - Other shocks (1 = Yes): 0.057* (0.024) and 0.056* (0.024).
  - HH size: 0.024*** (0.003) and 0.022*** (0.003).
  - HH gender (1 = Female): -0.059** (0.022) and -0.066** (0.022).
  - HH head age: -0.015*** (0.004) and -0.014*** (0.004).
  - HH head age squared: 0.000** (0.000) in reported specifications.
  - HH head level of education (3 = Tertiary): -0.165*** (0.038) in one specification and -0.081 (0.044) in another.
  - Regional coefficients (examples): Niamey -0.120*** (0.029) / -0.072** (0.027); Zinder 0.071* (0.030) / 0.153*** (0.024).

- Table 3 (Access to government transfers; linear probability model):
  - Relative poverty: self-assessment (3 = Poor): 0.006* (0.003) and 0.006* (0.003).
  - Entire day without food (1 = Yes): -0.007* (0.003) in one specification.
  - Conflict shocks (1 = Yes): 0.032** (0.011) and 0.034** (0.011).
  - HH head age: 0.001** (0.000) and 0.001* (0.000).
  - HH head age squared: -0.000** (0.000) and -0.000** (0.000).
  - Regional coefficients (examples): Diffa 0.040*** (0.010) / 0.043*** (0.010); Dosso -0.008* (0.004); Tillaberi -0.007* (0.003).

### Notes on regression tables
- Model: Linear probability models for determinants of access to food programs, health programs, and government transfers.
- Column focus:
  - Column (1): poverty criteria only.
  - Column (2): hunger and food insecurity criteria only.
  - Column (3): shock criteria only.
  - Column (4): other household characteristics criteria only.
  - Column (5): all criteria included.
- Significance notation preserved: *Indicates significance at 10% level, **significance at 5% level, and ***significance at 1% level.
- Robust standard errors reported in parentheses.
- Number of households entries appear as repeated sequences: 59985998599859985998 and 5998599859985998 as shown in the source.

*Source: sipea2023003 - 32.      Supporting better education outcomes, particularly for girls, would require an increase*

---


_Source: https://www.imf.org/-/media/files/publications/selected-issues-papers/2023/english/sipea2023003.pdf_