## 2. Domestic Revenue Mobilization in Sub-Saharan Africa: What Are the Possibilities?

## Source details

**Canonical URL:** [2. Domestic Revenue Mobilization in Sub-Saharan Africa: What Are the Possibilities?](https://www.imf.org/-/media/files/publications/reo/afr/2018/may/pdf/chap2.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/reo/afr/2018/may/pdf/chap2.pdf.md)
- [Structured JSON version](/-/media/files/publications/reo/afr/2018/may/pdf/chap2.pdf.json)

---

### Chapter team and scope
- Prepared by a team led by Alex Segura-Ubiergo and composed of Chuling Chen, John Hooley, Gabriel Leost, Toomas Orav, Miguel Pereira Mendes, Ashan Rodriguez, and Manuel Rosales.
- Focus: developments in revenue-to-GDP and tax-to-GDP ratios in sub-Saharan Africa (SSA); structural determinants of lower tax-to-GDP ratios; lessons from revenue mobilization case studies; role of digitalization and distributional impacts (including focus on CEMAC countries).

### Regional trends and headline statistics
- Median SSA economy:
  - Total revenue excluding grants increased from around 14 percent of GDP in the mid-1990s to more than 18 percent in 2016.
  - Tax revenue increased from 11 to 15 percent of GDP (1990s to 2016).
- Two-thirds of SSA countries now have revenue ratios above 15 percent, compared with fewer than half in 1995.
- Since the mid-1990s, 15 SSA countries have transitioned to tax-to-GDP ratios of about 13 percent and above (minimum ratio associated with significant acceleration in growth and development).
- Resource vs nonresource:
  - Oil exporters (average 2000–16): average revenue-to-GDP ratio 27 percent (compared with 18 percent for non-oil economies).
  - Nontax revenue accounts for almost half of oil exporters’ revenue, compared with less than 20 percent for non-oil exporters.
  - Standard deviation of total revenue for oil exporters was seven times that of non-oil exporters during 2000–16.
  - World oil price decline since 2014 reduced oil exporters’ overall revenue-to-GDP from 31 percent in 2012 to 18 percent in 2016.
- Fragile states: median non-resource-revenue-to-GDP was less than 14 percent in 2015, compared with 18 percent for nonfragile states.

### Global and regional comparisons
- SSA has, on average, the lowest revenue-to-GDP ratio compared with other regions.
- Over the past three decades, the increase in SSA’s revenue ratio has been double that for all emerging market and developing economies.
- Median revenue-to-GDP ratio among all emerging market and developing economies is 23 percent, 5 percentage points higher than for SSA.
- Median tax-to-GDP ratio for SSA is only 2 percentage points lower than that of all emerging market and developing economies.

### Tax system characteristics and efficiency
- Top rates and revenue:
  - Average top PIT rate in SSA reduced from about 44 percent to 32 percent since 2000.
  - Average top CIT rates reduced by more than 5 percentage points over the same period.
  - Despite lower rates, total direct taxes (PIT and CIT) as a percentage of GDP have trended upward.
- CIT productivity: defined as (CIT Revenue as a share of GDP)/(CIT rate). On average SSA lags advanced and emerging markets, with substantial cross-country variation (examples: Senegal, South Africa; SEZ rates as low as 15 percent; Côte d’Ivoire, Rwanda, Tanzania offer zero CIT in SEZs).
- VAT adoption and constraints:
  - Most SSA countries have introduced a VAT.
  - VAT efficiency (VAT C-efficiency) is relatively low and varies widely.
  - Constraining factors: narrow bases (exemptions and zero-rating), registration thresholds set relative to per capita GDP, weaknesses in VAT refund systems and administrative delays.

### Quantitative thresholds and guidance
- Fixed sample: 40 SSA economies with data 1995–2016 used for median-based analysis.
- Tipping point (Gaspar, Jamarillo, and Wingender 2016): minimum tax-to-GDP ratio of 12.88 percent estimated to enable the state to perform key functions.
  - With nontax revenues typically averaging 2 percent of GDP, a tax-to-GDP revenue of 13 percent and overall revenue ratio of 15 percent of GDP should be viewed as a minimum threshold.
- VAT rate guidance:
  - VAT rate below 13 percent: a 2 percent rate increase would have virtually no negative impact on growth.
  - VAT rates between 13 and 18 percent: a 1 percent increase would not have much effect on economic activity.
  - VAT rates above 18 percent: even small increases can have a substantial negative impact on growth.

### Underexploited revenue sources
- Excise taxes:
  - In 2015, SSA collected on average 1.4 percent of GDP from all excise taxes, less than half the level in emerging Europe.
  - Several countries (Benin, Côte d’Ivoire, Madagascar, Mozambique, Nigeria, Sierra Leone) collected excise revenues of less than 1 percent of GDP.
  - Implementation choices: specific taxes vs. ad valorem; specific taxes often better for externalities and predictable revenue.
  - Typical excisable products: petroleum, cigarettes, alcohol, motor vehicles, sometimes telecommunications.
- Property taxation:
  - Underused; previous studies suggest SSA countries can raise 0.5 to 1 percent of GDP via property taxation.
  - Provides stable, hard-to-evade revenue and improves local service delivery and accountability.

### Customs, cross-border rules, and administration
- Customs role:
  - In 2015, on average, SSA countries collected a third of their nonresource revenue through customs at their border.
  - Customs collect VAT on imports, trade taxes, and excise on imported goods.
- Reform priorities: digitalization of transactions and payments; anti-smuggling units; channeling goods through major ports with adequate controls.
- Cross-border tax rules:
  - Thin capitalization rules by end-2016 set debt-to-equity ratios up to 4:1 in SSA; international trends suggest countries with ratios above 2 could further limit interest deductions.
  - Transfer pricing regulations should embed the “arm’s length” principle where absent.

### Structural determinants, tax frontier and tax gap
- Tax frontier concept: highest level of tax revenue (percent of GDP) achievable given macroeconomic and institutional conditions; estimated using a stochastic panel model covering 121 countries during 2002–16.
- Drivers associated with higher tax-to-GDP: higher income per capita, more trade openness, higher public spending on education, better government effectiveness; lower income inequality and lower corruption also associated with higher tax ratios.
- Comparative level: the average tax frontier for Sub-Saharan African countries is around 7½ percentage points of GDP lower than the average tax frontier for the rest of the world.
- Average tax gap/tax potential for SSA: ranges between 3 and 5 percent of GDP.
  - Oil producers: tax potential at 5 percent of GDP or more (lowest tax effort).
  - Other resource and nonresource countries: tax potential about 3 percent of GDP.
- Policy implications by tax-collection level:
  - Below tipping point (~12½–13 percent of GDP): need efficiency reforms and structural improvements (reduce corruption, improve governance, increase education spending) to raise the frontier.
  - Medium (13–18 percent): could mobilize on average about 3½ percent of GDP via efficiency reforms; some countries near frontier require structural reforms to raise it.
  - High (over 18 percent): still average distance to the frontier of about 4 percent of GDP.

### Lessons from successful mobilization episodes
- Definition: total increase of 2 percentage points of nonresource GDP over a three-year period, with no substantial declines during or immediately after.
- Episodes identified: six sustained episodes across 44 countries (2000–16).
- Magnitude and persistence:
  - Nonresource revenue gain during three-year episodes: 2.2 to 8 percent of nonresource GDP.
  - Average annual increase during episodes: 1.2 percentage points.
  - Average total revenue gain during episodes: 3.5 percentage points.
  - Subsequent years: gains averaged 1 percentage point a year over the next three years; 2016 data show current revenue at least at episode end level and on average 3.4 percent of GDP higher than episode end.
- Common features:
  - Diverse contexts (low to medium tax effort; varied geography, income, fragility, resource intensity).
  - Most episodes overlapped with intensified IMF engagement (lending, nonlending, technical assistance).
  - Reforms were comprehensive, multiyear, and focused on foundational institutions and base broadening.
  - Robust growth often present during episodes (tax buoyancy), though acceleration in growth was not required.

### Institutional development, administration, and digitalization
- Foundational investments common to successful cases:
  - Taxpayer identification numbers, semiautonomous revenue authorities, VAT introduction, taxpayer segmentation.
- Administration modernization:
  - Establishment/strengthening of revenue authorities, large-taxpayer units, medium-term strategic plans, taxpayer segmentation, risk-based compliance.
  - ICT measures: e-filing, e-payment, integration of social contributions, electronic billing machines, ASYCUDA, electronic single window systems.
  - Many countries rolled out first e-tax platforms during 2011–13.
- Digital revenue mobilization (Box 2.1 highlights):
  - Benefits: better information, lower costs, deeper tax base via reduced cash use.
  - Hurdles: low internet penetration, data quality issues, financial and reputational risks, weak enforcement, low trust.
  - Peer-to-peer learning examples: 2016 Hackathon in Senegal; 2017 Ideas Workshop in Uganda.
  - Recommendations: consider inclusive seminars when preparing medium-term revenue mobilization plans.

### Policy and administrative recommendations (priority actions)
- Improve efficiency of current tax systems:
  - Reduce tax exemptions and zero-rating that narrow tax bases.
  - Lower registration thresholds where appropriate as administrative capacity matures.
  - Strengthen VAT refund systems (for example, consider settling refunds out of gross VAT receipts via escrow accounts; use risk-based audit verification to expedite refunds).
- Tap underexploited taxes with administrable measures:
  - Expand well-designed excise taxes on appropriate goods; choose specific vs. ad valorem instruments based on objectives.
  - Develop property taxation to mobilize 0.5 to 1 percent of GDP where feasible.
- Focus on VAT efficiency rather than exclusively on rate increases; pair VAT-based measures with pro-poor spending and social protection to offset distributional impacts.
- Review international corporate taxation and accelerate customs administration reforms.
- Medium-term strategies and political economy:
  - Build broad-based support via outreach to public and private sectors.
  - Emphasize governance improvements and anti-corruption to unlock an estimated average potential of about 3 to 5 percent of GDP in additional tax collection through efficiency and institutional reforms.
  - Define medium-term revenue strategies with clear objectives, capacity-building plans, and credible public communication explaining why taxes are being increased.

### Distributional simulations for resource-rich countries (CEMAC — Box 2.2 summary)
- Model: IMF DIGNAR model (Melina, Yang, and Zanna 2016). Two household types: NFC (non–financially constrained) and FC (financially constrained).
- Simulated channels:
  1. Increase in VAT rates.
  2. Improvement in collection efficiency (expand base).
- Key results:
  - Both measures increase non-oil revenues and reduce public debt.
  - Both initially reduce non-oil GDP; long-run recovery and possible higher levels when efficiency improves.
  - Collection efficiency improvements have more favorable distributional outcomes (negative impacts concentrated on NFC consumption).
  - VAT rate increases disproportionately reduce FC household consumption (higher marginal propensity to consume).
  - Mitigation strategies: channeling a fraction (for example, half) of additional non-oil revenue to targeted transfers to FC households or to public investment reduces adverse distributional effects; combining VAT increases with public investment mitigates non-oil GDP effects; combining VAT increases with targeted cash transfers mitigates adverse effects on FC households.
  - Caveat: DIGNAR omits a channel where public investment reduces unemployment in poor households.

- DIGNAR calibration targets for CEMAC (Percent of GDP):
  - Exports: 40.1
  - Imports: 38.7
  - Government consumption: 14.6
  - Government investment: 11.9
  - Private investment: 16.2
  - Resource sector: 24.5
  - Government domestic debt: 12
  - Government external concessional debt: 13.2
  - Government external commercial debt: 10.4
  - Grants: 0.7

### Lessons on sequencing, political economy, and institutional prerequisites
- Sequencing and fundamentals:
  - Foundational institutions are prerequisites for complex administrative/technological reforms.
  - Revenue institution building, credible medium-term strategies, and sustained political commitment are key.
- Political and governance prerequisites:
  - Peace and stability are preconditions for success; fragile countries often have tax-to-GDP ratios below 10 percent.
  - Consistent leadership (long-tenured ministers) coincided with mobilization episodes in examples (Mozambique, Senegal).
  - Publication of exemption beneficiaries and tax-expenditure budgets can improve legitimacy and compliance.
- External engagement:
  - Prolonged IMF technical assistance and IMF-supported programs were present in all studied successful cases, but IMF engagement cannot substitute for political will.

### Outcomes, timing, and realistic expectations
- Potential additional revenue: SSA countries could mobilize on average up to 5 percent of GDP in additional tax revenues in the next few years.
- Pace of gains:
  - Countries rebuilding after conflict can expand revenue quickly (example: Liberia nonresource revenue rose by 2.6 percentage points each year over three years).
  - Strong performers: average annual increases in nonresource revenue about 0.9 percentage point of GDP during episodes; post-episode increases slow to about 0.7 percentage point.
- Gains are typically incremental and accrue over prolonged periods; sustained capacity and perseverance are essential.

### Annex highlights — Estimating tax effort and tax potential
- Definitions:
  - Tax frontier: maximum theoretical tax revenue (percent of GDP) given structural conditions.
  - Tax effort: ratio of actual tax revenue to the frontier.
  - Tax potential: difference between frontier and actual revenue.
- Estimation strategy:
  - Stochastic panel model of log tax/GDP as a function of lagged log real GDP per capita, its square, trade openness, agriculture share, Gini, public education spending, corruption, government effectiveness, and oil dummy.
  - Tax frontier computed from these regressions; tax effort and tax potential derived from frontier minus actual.
- Main regression coefficients (selected exact values reported):
  - Log of real GDP per capita: 2.939 ***, 2.866 ***, 2.885 ***, 2.781 ***, 2.691 ***, 2.716 ***
  - Log of real GDP per capita squared: –0.152 ***, –0.148 ***, –0.150 ***, –0.142 ***, –0.138 ***, –0.140 ***
  - Trade openness: 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***
  - Gini coefficient: –0.006 ***, –0.006 ***, –0.007 ***, –0.006 ***, –0.006 ***, -0.006 ***
  - Education: 0.015 ***, 0.016 ***, 0.016 ***, 0.016 ***, 0.018 ***, 0.017 ***
  - General government: 0.105 **, 0.109 ***, 0.110 ***, 0.091 **, 0.093 **, 0.098 **
  - Corruption: 0.117 ***, 0.083 *, 0.134 ***, 0.100 **
  - Government effectiveness: 0.091 *, 0.088 *
  - Oil dummy (selected values): 0.080 **, 0.035, 0.031, 0.043, 0.030, 0.026
- Model diagnostics (selected):
  - Sigma_u: 0.515 ***, 0.515 ***, 0.516 ***, 0.525 ***, 0.526 ***, 0.526 ***
  - Sigma_e: 0.099 ***, 0.098 ***, 0.098 ***, 0.106 ***, 0.106 ***, 0.105 ***
  - Number of observations: 1,366; 1,360; 1,360; 1,109; 1,031; 1,03
  - Number of countries: 122; 121; 121; 99; 98; 98
- Goods and Services tax regressions (selected coefficients):
  - Log of real GDP per capita: 2.379 ***, 2.332 ***, 2.353 ***, 2.228 ***, 2.173 ***, 2.207 ***
  - Log of real GDP per capita squared: –0.122 ***, –0.120 ***, –0.122 ***, –0.113 ***, –0.111 ***, –0.114 ***
  - Trade openness: 0.002 *** (all specifications)
  - Gini coefficient: –0.008 *** (all specifications)
  - Oil dummy: –0.642 ***, –0.626 ***, –0.634 ***, –0.646 ***, –0.581 ***, –0.587 ***
  - Sigma_u (Goods and Services): 0.660 ***, 0.665 ***, 0.665 ***, 0.667 ***, 0.672 ***, 0.672 ***
  - Sigma_e (Goods and Services): 0.162 ***, 0.159 ***, 0.159 ***, 0.177 ***, 0.174 ***, 0.174 ***
  - Number of observations (Goods and Services): 1,152; 1,146; 1,146; 930; 924; 924
  - Number of countries (Goods and Services): 105; 104; 104; 85; 84; 84
- Country-level tax-to-GDP examples (selected exact sequences as reported):
  - Nigeria: 5.9 11.1 11.1 2.0 10.7 10.4 12.0 8.1 8.3 8.5
  - Central African Rep.: 6.2 8.4 8.5 9.7 8.0 8.2 8.8 8.1 7.9 8.8
  - Madagascar: 9.9 16.7 17.3 19.5 16.6 16.7 19.4 14.8 15.8 18.4
  - Tanzania: 12.4 20.3 20.9 19.5 20.2 19.8 19.7 18.3 18.6 19.4
  - Ethiopia: 12.7 13.8 14.2 13.2 13.9 13.8 13.1 13.3 13.3 13.1
  - South Africa: 24.7 26.9 26.9 31.1 27.9 27.6 30.5 25.5 25.4 26.2
  - Zimbabwe: 26.9 27.7 27.6 27.5 27.7 27.6 27.5 27.8 27.7 27.6
  - Swaziland: 28.3 30.4 29.8 30.1 30.4 30.4 30.5 30.3 30.4 29.6
  - Seychelles: 29.2 36.2 34.8 49.4 39.4 39.4 39.0 48.3 34.5 34.2 37.1
  - Namibia: 32.1 33.5 33.4 33.9 34.2 33.9 35.4 33.7 33.8 33.2
- Average row values reported (exact sequence): 16.2 19.6 19.9 21.3 19.9 19.8 20.9 18.7 18.9 19.6

### Final actionable steps (concise)
- Prioritize VAT efficiency reforms and base broadening over blunt rate hikes; apply VAT-rate guidance with distributional safeguards.
- Strengthen tax administration foundations (taxpayer IDs, large-taxpayer units, semiautonomous revenue authorities, ICT platforms) and sequence reforms to country context.
- Expand administrable excises and develop property taxation to mobilize 0.5 to 1 percent of GDP where feasible.
- Address structural factors (governance, corruption, education spending, trade openness) to raise the tax frontier and close the 3 to 5 percent of GDP tax gap.
- Design medium-term revenue strategies, build a constituency through transparency and outreach, and pair revenue measures with targeted social protection to mitigate distributional impacts.

*Source: Regional Economic Outlook: Sub‑Saharan Africa — Chapter 2 (excerpt).*

### 2. Domestic Revenue Mobilization in Sub-Saharan Africa:

### 2. Domestic Revenue Mobilization in Sub-Saharan Africa: What Are the Possibilities?

### Chapter team and scope
- Prepared by a team led by Alex Segura-Ubiergo and composed of Chuling Chen, John Hooley, Gabriel Leost, Toomas Orav, Miguel Pereira Mendes, Ashan Rodriguez, and Manuel Rosales.
- Focus: developments in revenue-to-GDP and tax-to-GDP ratios in sub-Saharan Africa (SSA), structural determinants of lower tax-to-GDP ratios, lessons from revenue mobilization case studies, and the role of digitalization and distributional impacts (including focus on CEMAC countries).

### Regional trends in revenue mobilization
- Over the past three decades, substantial gains in revenue mobilization for many SSA countries.
- Median SSA economy:
  - Total revenue excluding grants increased from around 14 percent of GDP in the mid-1990s to more than 18 percent in 2016.
  - Tax revenue increased from 11 to 15 percent of GDP (1990s to 2016).
- Gains driven primarily by nonresource revenues, which increased sharply in the past 10 years.
- Resource revenues have not increased substantially and have been volatile (notably in late 2000s and since 2014).
- Since the mid-1990s, 15 SSA countries have transitioned to tax-to-GDP ratios of about 13 percent and above (a minimum ratio associated with significant acceleration in growth and development).
- Two-thirds of SSA countries now have revenue ratios above 15 percent, compared with fewer than half in 1995.
- Sources of gains: increases in direct and indirect taxes; indirect taxes boosted by VAT introduction in several countries; taxes on imports declined as a share of GDP reflecting trade liberalization.

### Global and regional comparisons
- SSA still has, on average, the lowest revenue-to-GDP ratio compared with other regions.
- Over the past three decades, the increase in SSA’s revenue ratio has been double that for all emerging market and developing economies.
- Median revenue-to-GDP ratio among all emerging market and developing economies is 23 percent, 5 percentage points higher than for SSA.
- Median tax-to-GDP ratio for SSA is only 2 percentage points lower than that of all emerging market and developing economies, though SSA still has the second lowest ratio among regions.

### Heterogeneity within SSA: oil exporters and fragile states
- Oil exporters (average 2000–16):
  - Average revenue-to-GDP ratio: 27 percent (compared with 18 percent for non-oil economies).
  - Nontax revenue accounts for almost half of oil exporters’ revenue, compared with less than 20 percent for non-oil exporters.
  - Revenue volatility much higher: during 2000–16 the standard deviation of total revenue for oil exporters was seven times that of non-oil exporters.
  - Decline in world oil price since 2014 led overall revenue-to-GDP for oil exporters from 31 percent in 2012 to 18 percent in 2016.
- Fragile states:
  - Median non-resource-revenue-to-GDP ratio was less than 14 percent in 2015, compared with 18 percent for nonfragile states.
  - Weak institutions, security, and governance challenge non-resource revenue mobilization.

### Tax system characteristics and efficiency
- Trend of reduced top tax rates:
  - Average top PIT rate in SSA reduced from about 44 percent to 32 percent since 2000.
  - Average top CIT rates reduced by more than 5 percentage points over the same period.
- Despite lower rates, total direct taxes (PIT and CIT) as a percentage of GDP have trended upward.
- CIT productivity definition: (CIT Revenue as a share of GDP)/(CIT rate).
  - On average, SSA CIT productivity lags advanced and emerging market economies, with substantial cross-country variation.
  - Some countries (e.g., Senegal, South Africa) offer reduced CIT rates (15 percent) for companies in SEZs; other countries (Côte d’Ivoire, Rwanda, Tanzania) offer zero CIT rates in SEZs.
- VAT adoption and performance:
  - Most SSA countries have introduced a VAT, replacing general sales taxes.
  - VAT efficiency (VAT C-efficiency = actual VAT collections as a share of potential base) in SSA is relatively low compared with other regions and varies widely across countries.
  - Factors constraining VAT efficiency:
    - Narrow tax bases due to proliferation of exemptions and zero rating.
    - Different registration thresholds for taxpayers; thresholds often set relative to per capita GDP with substantial cross-country differences for PIT and VAT.
    - Weaknesses in VAT refund systems, including varied practices (VAT credits against future payments; quarterly refunds; refunds after audit verification) and administrative delays that can build up unpaid claims.

### Quantitative assessments and thresholds cited
- Fixed sample of 40 SSA economies with data 1995–2016 used for median-based analysis due to skewed distribution of revenue ratios.
- Tipping point from Gaspar, Jamarillo, and Wingender (2016):
  - Minimum tax-to-GDP ratio of 12.88 percent estimated to enable the state to perform key functions.
  - With nontax revenues typically averaging 2 percent of GDP, a tax-to-GDP revenue of 13 percent and overall revenue ratio of 15 percent of GDP should be viewed as a minimum threshold.
- VAT rate guidance (summary of findings):
  - In countries where VAT rate is below 13 percent, a 2 percent rate increase would have virtually no negative impact on growth.
  - In countries with VAT rates between 13 and 18 percent, a 1 percent increase would not have much effect on economic activity.
  - With rates above 18 percent, even small increases can have a substantial negative impact on growth.

### Underexploited revenue sources and administrative reforms
- Excise taxes:
  - In 2015, SSA countries collected on average 1.4 percent of GDP from all excise taxes, less than half the level in emerging Europe.
  - Wide cross-country differences: several countries (Benin, Côte d’Ivoire, Madagascar, Mozambique, Nigeria, Sierra Leone) collected excise revenues of less than 1 percent of GDP.
  - Implementation choices: specific tax (monetary amount per quantity) vs. ad valorem (based on value/price); specific taxes often better for addressing externalities and producing predictable revenue.
  - Typical exciseable products: petroleum, cigarettes, alcohol, motor vehicles, sometimes telecommunications.
- Property taxation:
  - Underused in SSA; provides stable and reliable revenue, hard to evade, and improves local service delivery/accountability.
  - Previous studies suggest SSA countries can raise 0.5 to 1 percent of GDP via property taxation.
  - Growing use across SSA, but many countries still rely on one-time payments for property revenue.

### Policy and administrative recommendations emphasized
- Improve efficiency of current tax systems, including:
  - Reduction of tax exemptions and zero-rating that narrow tax bases.
  - Lower registration thresholds where appropriate as tax administration capacity matures.
  - Strengthen VAT refund systems (e.g., consider settling refunds out of gross VAT receipts by establishing escrow accounts; use risk-based audit verification to expedite refunds).
- Tap underexploited taxes with targeted, administrable measures:
  - Expand well-designed excise taxes on appropriate goods and set taxes (specific vs. ad valorem) according to country objectives.
  - Develop property taxation as a stable local revenue source, aiming to mobilize 0.5 to 1 percent of GDP where feasible.
- Focus on VAT efficiency rather than exclusively on rate increases to be more growth friendly; pair VAT-based revenue measures with pro-poor spending and social protection to offset potential distributional impacts.
- Review international corporate taxation policies and accelerate customs administration reforms.
- Medium-term strategies should incorporate political economy considerations:
  - Build broad-based support through proactive outreach to public and private sectors.
  - Emphasize institutional changes (governance improvements, anti-corruption measures) to unlock an estimated average potential of about 3 to 5 percent of GDP in additional tax collection through a combination of efficiency reforms and institutional improvements.

*Source: Regional Economic Outlook: Sub-Saharan Africa, chapter “Domestic Revenue Mobilization in Sub-Saharan Africa: What Are the Possibilities?”*

### 2. DOMESTIC REVENUE MOBILIZATION IN SUB-SAHARAN AFRICA: WHAT ARE THE POSSIBILITIES?

### 2. DOMESTIC REVENUE MOBILIZATION IN SUB-SAHARAN AFRICA: WHAT ARE THE POSSIBILITIES?

### Customs, property taxation, and cross-border tax rules
- Customs administration is key:
  - In 2015, on average, sub-Saharan African countries collected a third of their nonresource revenue through customs at their border.
  - Customs administrations collect VAT on imports, trade taxes, and excise taxes on imported goods.
  - Reforms that can deliver relatively rapid results include modernization of customs processes (digitalization of transactions and payments), measures to combat corruption and fraud (strengthening clearance procedures and creating anti-smuggling units), and channeling goods through a few major ports with adequate custom controls.
- Recurrent property taxation:
  - Rollout requires significant capacity-building around property registries and annual appraisal systems and stronger coordination between central and subnational governments.
  - Relatively rapid progress is possible in urbanized areas where ownership information and reference valuations exist and can be supplemented by geo-spatial data from global positioning systems.
- Cross-border tax rules:
  - Thin capitalization rules have been adopted to limit tax deductions on interest; by the end of 2016 thin capitalization rules across sub-Saharan Africa had set debt-to-equity (or “gearing”) ratios of up to 4:1.
  - Recent international trends suggest countries with rules allowing for ratios above 2 could look to further limit interest deductions (Botswana, Equatorial Guinea, Namibia, Rwanda, Tanzania, Zambia, Zimbabwe).
  - Transfer pricing (intragroup transactions) can distort taxable income; new regulations typically embed the “arm’s length” principle. Tax rules and monitoring frameworks covering transactions between related parties need to be introduced where absent.

### Excise taxes and revenue composition (selected statistics)
- Excise taxes, 2015:
  - Comparative note: figures presented for Emerging Europe, Advanced economies, Asia, Latin America, and SSA (sub-Saharan Africa) in percent of GDP (see Figures 2.20 and 2.21 in source).
- Share of nonresource revenue collected at customs, 2015:
  - Sub-Saharan Africa average: one-third of nonresource revenue collected at customs (Figure 2.22).
  - Country-level variation shown in source charts (percent of GDP and percent collected at customs).

### Structural factors affecting tax effort and tax frontier methodology
- Tax frontier concept:
  - Defined as the highest level of tax revenue (percent of GDP) a country can be expected to achieve given macroeconomic and institutional conditions.
  - Computed using a stochastic panel data model covering 121 countries during 2002–16 (following Fenochietto and Pessino 2010, 2013).
  - Independent variables include income per capita, trade openness, share of agriculture in GDP, income inequality, public spending on education, measures of corruption, and government effectiveness.
  - Higher income, more trade openness, higher spending on education, and better government effectiveness are associated with higher tax-to-GDP ratios; lower income inequality and lower corruption also tend to be associated with higher tax ratios.
- Comparative level:
  - The average tax frontier for Sub-Saharan African countries is around 7½ percentage points of GDP lower than the average tax frontier for the rest of the world.

### Tax gaps, country groups, and implications
- Average tax gap and potential:
  - The average tax gap (tax potential) for sub-Saharan African countries ranges between 3 and 5 percent of GDP.
  - On average, sub-Saharan Africa’s tax gap is slightly lower than elsewhere, implying similar average inefficiency conditional on structural factors.
  - Because overall tax revenues are lower, the cost of inefficiency may be higher in sub-Saharan Africa.
- Variation by country group:
  - Oil producers: lowest tax effort and highest average tax potential, at 5 percent of GDP or more.
  - Other resource and nonresource countries: lower levels of tax potential of about 3 percent of GDP.
- Policy implications by tax collection level:
  - Low tax collection levels (below tipping point):
    - Countries that have not reached a minimum threshold of about 12½ to 13 percent of GDP will need reforms to increase efficiency and to push the tax frontier higher by improving structural factors (for example, reducing corruption, improving governance, increasing spending on education).
    - Example: Nigeria could double its tax-to-GDP ratio and exceed 10 percent of GDP with efficiency reforms, but surpassing the tipping point would likely require improvements in structural factors.
  - Medium tax collection levels (13–18 percent of GDP):
    - These countries tend to have larger tax gaps and could mobilize, on average, about 3½ percent of GDP in additional revenues through efficiency reforms (for example, review of existing taxes and exemptions).
    - Some countries (Côte d’Ivoire, Ethiopia, Mali) appear relatively close to the tax frontier; efficiency gains alone could be limited, and structural reforms to raise the frontier are also needed.
  - High tax collection levels (over 18 percent of GDP):
    - These countries have a relatively elevated tax frontier. Despite comparatively high tax-to-GDP ratios, there is still an average distance to the frontier of about 4 percent of GDP, indicating potential for additional revenue mobilization.
    - Some countries may opt for lower taxes as a public policy choice (for example, on the desired size of government).

### Lessons from successful revenue mobilization episodes
- Definition and identification:
  - A successful episode is defined as a total increase of 2 percentage points of nonresource GDP over a three-year period, with no substantial declines in the revenue ratio within or immediately following the period.
  - Using a data set covering 44 sub-Saharan African countries from 2000–16, the analysis finds only six episodes of sustained revenue mobilization.
- Magnitude and persistence of gains:
  - Nonresource revenue gain during the three-year episodes ranges from 2.2 to 8 percent of nonresource GDP.
  - Average annual increase during episodes: 1.2 percentage points.
  - Average total revenue gain during episodes: 3.5 percentage points.
  - In all cases, gains continued in subsequent years, with increases averaging 1 percentage point a year over the next three years.
  - Data for 2016 indicate that the current level of revenue is at least at the same level it was at the end of the episode, and on average 3.4 percent of GDP higher than the episode end point.
- Characteristics of successful episodes:
  - Success occurred in a variety of circumstances and initial conditions, spanning low to medium tax effort, diverse geography, income levels, fragility, and resource intensity.
  - Common factor: countries tended to experience robust growth during the revenue mobilization episode (tax buoyancy may have contributed), although acceleration in growth was not required.
    - Example: only Liberia saw a significant acceleration in growth; other countries’ growth decelerated modestly from an average growth rate of 6.7 percent prior to the episode to 5.7 percent during the episode.
  - Most episodes overlapped with intensified engagement with the IMF in the form of both lending and nonlending programs and substantial technical assistance efforts.
- Reform patterns:
  - No single template—reforms were tailored to country circumstances.
  - Common elements across cases:
    - Comprehensive and multiyear reform strategies.
    - Focus on basic institutions and building blocks of the tax system (foundational measures).
    - Measures to broaden the tax base.
    - Modernization of tax administration institutions.

### Pursuing a comprehensive reform strategy: foundational elements
- The tax system as a pyramid:
  - Foundational institutions provide the base for more complex administrative and technological transformations.
  - Countries invested significant effort in foundational elements such as:
    - Taxpayer identification number.
    - Semiautonomous revenue authority.
    - Value-added tax (VAT).
    - Taxpayer segmentation.
- Sequencing depends on country circumstances, but investments in basic building blocks were common across successful cases.

*Source: REGIONAL ECONOMIC OUTLOOK: SUB-SAHARAN AFRICA — Chapter 2 (excerpt).*

### introduction of such reforms has been associated

### 2. DOMESTIC REVENUE MOBILIZATION IN SUB-SAHARAN AFRICA: WHAT ARE THE POSSIBILITIES?

### Institutional development and administrative modernization
- Revenue institution building and modernization were central to successful mobilization episodes:
  - Establishment and strengthening of revenue authorities, large-taxpayer units, and medium-term strategic plans.
  - Reorganizations and medium-term strategies to strengthen capacity and coverage were common even where institutions were already established.
  - All countries in the study adopted some form of taxpayer segmentation; Rwanda, Tanzania, and Uganda dedicated resources to risk-based compliance strategies for different taxpayer segments.
- ICT and automation advances supported administration reforms:
  - Fast adoption of automating systems across domestic tax and customs administration; several countries rolled out first e-tax platforms during 2011–13.
  - Examples of ICT measures: e-filing and e-payment systems; integration of social contributions into e-tax; electronic billing machines; ASYCUDA customs automation; electronic single window systems.
- Focus areas in administration reforms:
  - Improve compliance and enforcement through taxpayer segmentation and risk management.
  - Simplify registration, filing, payment, audit, collection, enforcement, and appeals via ICT platforms.
  - Combine domestic tax and customs operations and simplify customs clearance.

### Tax policy, base broadening, and exemptions
- Emphasis on expanding and protecting the tax base rather than on frequent rate adjustments:
  - Measures to reduce base-narrowing exemptions (voiding or suspending exemptions in Liberia and Uganda; publishing beneficiaries in Tanzania and Uganda).
  - Revision of investment codes (Mozambique, Rwanda, Senegal, Tanzania) to limit tax incentives.
  - Elimination or revision of distortions in VAT (Rwanda, Senegal, Uganda).
- Targeting hard-to-tax sectors and small taxpayers:
  - Introduction of simplified tax regimes for small businesses (Mozambique, Rwanda, Senegal, Tanzania).
  - Changes to VAT thresholds to better target high-value businesses (Tanzania, Uganda).
  - Expansion of withholding agents (Uganda) and strengthening specialized taxes (property, investment income) in Rwanda and Senegal.
- Policy design principle:
  - Successful experiences relied on implementing broad-based VATs, gradually expanding the base for direct taxes (CIT and PIT), taxing small businesses, and levying excises on key items.

### Political commitment, stability, and governance prerequisites
- Political and security conditions:
  - Peace and stability are preconditions for success; fragile countries often have tax-to-GDP ratios below 10 percent of GDP.
  - Consistent commitment of political leadership matters; long-tenured ministers coincided with mobilization episodes in Mozambique and Senegal.
- Human resources and management:
  - Rapid turnover in revenue administration staff or inadequate HR practices can hinder progress.
- Governance and transparency:
  - Improving governance, controlling corruption, and enhancing transparency and efficiency of public spending support compliance.
  - Publication of beneficiaries of tax exemptions and tax expenditure budgets were used to increase legitimacy and accountability.
- Outreach and communication:
  - Transparency and taxpayer education helped build public support and improve compliance (examples: Liberia published financial accounts of revenue-generating agencies; Rwanda and Uganda launched taxpayer education programs).

### Role of external engagement and medium-term strategies
- Technical assistance and IMF engagement:
  - All countries received prolonged IMF technical assistance and maintained IMF-supported programs emphasizing revenue mobilization.
  - IMF engagement provided a sounding board but could not substitute for political will.
- Medium-term revenue strategies:
  - Adoption of multi-year revenue mobilization strategies enhanced commitment and focus (Senegal 2003, Tanzania 2003, Mozambique 2006, Rwanda 2013).
  - Successful strategies emphasized taxpayer-centric policies, private sector consultation, and accountability of tax authorities to taxpayers.

### Outcomes, magnitude, and timing of gains
- Potential additional revenue:
  - Sub-Saharan African countries could mobilize on average up to 5 percent of GDP in additional tax revenues in the next few years.
- Pace of revenue gains:
  - Countries rebuilding institutions after conflict can expand revenue relatively quickly: Liberia’s nonresource revenue ratio rose by 2.6 percentage points each year over three years.
  - Among strong performers with foundational reforms in place before the episode:
    - Average annual increases in nonresource revenue were about 0.9 percentage point of GDP a year during the episode.
    - After the episode, gains tended to slow to 0.7 percentage point.
- Gains are typically incremental and accrue over prolonged periods; perseverance and capacity to sustain reform momentum are essential.

### Key policy implications and actionable steps
- Get the basics of tax policy right:
  - Ensure a sound VAT, limit excessive tax incentives, and establish frameworks to enforce tax compliance.
- Prioritize institutional reforms and medium-term planning:
  - Develop capacity for risk-based allocation of enforcement resources and build reliable registries of large taxpayers.
  - Implement medium-term revenue mobilization plans that focus on taxpayer-centric policies.
- Strengthen governance, transparency, and public financial management:
  - Publish tax-exemption beneficiaries and pursue PFM reforms to increase legitimacy and willingness to pay.
- Maintain strong political commitment:
  - Secure sustained leadership support, manage turnover in key positions, and align incentives and sanctions for officials to preserve integrity.
- Tailor reforms to country-specific contexts:
  - Define reform sequencing and policies using local knowledge, led by country authorities.

*Source: IMF Regional Economic Outlook: Sub‑Saharan Africa — chapter content on domestic revenue mobilization.*

### 2. DOMESTIC REVENUE MOBILIZATION IN SUB-SAHARAN AFRICA: WHAT ARE THE POSSIBILITIES?

### 2. DOMESTIC REVENUE MOBILIZATION IN SUB-SAHARAN AFRICA: WHAT ARE THE POSSIBILITIES?

### Priority policy actions and institutional measures
- Identify the taxes that offer the greatest potential:
  - Improving the VAT offers substantial potential in most sub-Saharan African countries given its current low efficiency in most cases.
  - Systematically assess potential associated with other taxes, including the CIT (where excessive tax exemptions/incentives have been eroding the base), the PIT (where there should be an effort to gradually expand coverage), and excise taxes.
  - Stricter enforcement of customs rules and procedures could help mobilize additional revenues despite the general decline in customs duties.
  - Potential also exists in other areas, such as real estate taxes, though many countries have so far achieved limited progress.
- Review the legal framework and tax policy design:
  - Align tax policies with established revenue objectives.
  - Possible measures include the introduction of a VAT, the reduction of exemptions, and the introduction of sanctions for noncompliance.
- Assess the institutional framework at two levels:
  - Underlying supporting framework covering governance aspects: countries with weak governance are less likely to be effective in revenue mobilization; emphasis on improving governance and controlling corruption is crucial.
  - Operational framework covering institutional arrangements that have proven effective, such as the establishment of a revenue authority that follows specific principles.
  - Empirical note: seven of the 10 countries ranked highest in the control of corruption dimension of the World Bank Worldwide Governance Indicators have a relatively high tax-to-GDP ratio (above 18 percent of GDP). These include Botswana, Cabo Verde, Mauritius, Namibia, Senegal, Seychelles, and South Africa. Rwanda also scores high in control of corruption and has made great progress in revenue mobilization.
- Define a medium-term revenue strategy:
  - Strategy should provide medium-term objectives and short-term goals, and define capacity-building needs.
  - A convincing strategy needs to explain why the state is seeking to collect additional taxes.
- Build a constituency for reform:
  - Success depends on horizontal and vertical accountability structures.
  - Horizontal accountability: capacity to convince other political parties that revenue mobilization is in the broader interest to avoid reversals after elections.
  - Vertical accountability: the social contract between the state and citizens—state collects taxes in exchange for effective and transparent government spending.
  - Public outreach should be based on credible commitment to better governance and transparency.

### Box 2.1 — Digital Revenue Mobilization (summary)
- Digitalization impacts tax policy and administration by:
  - Empowering policymakers with quick access to more reliable information.
  - Reducing costs for administrators and taxpayers via digital infrastructure that eliminates manual processes.
  - Deepening the tax base by reducing cash use and facilitating analysis of transaction chains.
  - Benefiting the business climate by clarifying tax rules and speeding processes.
- Examples of digital measures adopted by some sub-Saharan African tax authorities:
  - Online e-tax portals.
  - Mobile tax payments.
  - Online reimbursement of VAT credits.
- Implementation hurdles in the region:
  - Low levels of internet penetration limit reach of some platforms.
  - Inherent complexity: platforms require extensive development and adaptation amid incomplete or low-quality data, with significant financial and reputational risks.
  - Sociopolitical challenges: weak enforcement and little trust in government.
- Peer-to-peer learning and homegrown solutions:
  - Examples: 2016 Hackathon in Senegal; 2017 Ideas Workshop in Uganda.
  - Senegal: expanding the menu of mobile options to improve e-tax accessibility.
  - Uganda: encouraging deployment of electronic fiscal devices to improve compliance with sales taxes and the VAT.
  - Suggestion: in preparing medium-term revenue mobilization plans, country authorities should consider organizing similar seminars to draw inputs from a broad range of stakeholders.

### Box 2.2 — Modeling the economic impacts of revenue mobilization in resource-rich countries (CEMAC application)
- Purpose and scope:
  - Analyze macroeconomic and distributional impacts of improved non-oil revenue mobilization in the Central African Economic and Monetary Community (CEMAC).
  - Investigate how undesirable distributional effects can be addressed using newly created fiscal space.
- Two revenue-raising channels simulated:
  1. Increase in value-added tax (VAT) rates.
  2. Improvement in the efficiency of collection of existing taxes (expansion of the tax base through greater efficiency).
- Model used:
  - IMF Debt, Investment, Growth and Natural Resources (DIGNAR) model (Melina, Yang, and Zanna 2016).
  - Key features: real small open economy, three production sectors, productive public capital, three types of debt (commercial, external, concessional), two types of households:
    - NFC (non–financially constrained) households with access to capital and financial markets.
    - FC (financially constrained) households that are poor and consume all their disposable income each period.
- Key simulation results:
  - Non-oil revenue mobilization helps reduce government debt and can increase long-term growth, but can have potentially undesirable distributional effects.
  - Both an increase in VAT rates and an improvement in collection efficiency:
    - Increase non-oil revenues.
    - Reduce public debt.
    - Reduce private consumption for NFC households (initially).
    - Initially reduce non-oil GDP, consistent with short-term fiscal multipliers; non-oil GDP recovers in the medium term driven by increased private investment and net exports and reaches a higher-than-initial level in the long run when revenue gains are realized via improved tax collection efficiency.
  - Differences between the two measures:
    - Improvement in revenue collection efficiency allows for lower tax rates for a given level of debt and has more desirable distributional properties: negative impacts fall largely on NFC consumption and not on FC consumers.
    - Increase in VAT rate negatively affects FC consumer consumption particularly, because FC households have a larger marginal propensity to consume than NFC households.
  - Mitigating distributional effects:
    - Channeling a fraction (for example, half) of additional non-oil revenue from higher VAT rates to targeted transfers toward FC households or to public investment can mitigate adverse effects.
    - Combination of increased VAT rates with additional public investment is especially effective at mitigating negative effects on non-oil GDP.
    - Combination of increased VAT rates with targeted cash transfers is powerful at mitigating adverse effects on FC households.
    - Caveat: the DIGNAR model does not include a channel whereby public investment reduces unemployment in poor households; thus potential mitigating mechanisms that reduce inequality via employment effects are absent in the model.
- Preparation note:
  - Box prepared by Giovanni Melina and Marcos Poplawski-Ribeiro with support from Mathilde Perinet.

### DIGNAR calibration targets for CEMAC (Table 2.1.1)
- Targets (Percent of GDP) — Value
  - Exports: 40.1
  - Imports: 38.7
  - Government consumption: 14.6
  - Government investment: 11.9
  - Private investment: 16.2
  - Resource sector: 24.5
  - Government domestic debt: 12
  - Government external concessional debt: 13.2
  - Government external commercial debt: 10.4
  - Grants: 0.7

### Annex 2.1 — Estimating tax effort and tax potential (definitions and estimation strategy)
- Definitions:
  - Tax frontier: the maximum theoretical level of tax revenues (measured in percent of GDP) that a country can achieve given underlying structural conditions (level of development, trade openness, sectoral structure, income distribution, institutions, etc.).
  - Tax effort: the ratio of actual tax revenue to corresponding frontier tax revenue.
  - Tax potential: the distance between the tax frontier and the actual tax revenue level; can be achieved through higher taxation or better collection efficiency.
- Estimation strategy (steps summarized):
  - Step 1: Estimate the tax frontier from a cross-country panel data set where the log of the tax revenue-to-GDP ratio for country i at period t is modeled as a function of a vector of independent variables that affect taxes, an inefficiency term correlated with the tax frontier but independent from the regressors, and a residual error term.
  - Step 2: Determine the tax effort as the ratio of actual tax revenue to the estimated frontier tax revenue.
  - Step 3: Determine the tax frontier and tax potential as the difference between the frontier and actual tax revenue.
- Data and variables used:
  - Log of tax to GDP: World Economic Outlook (WEO).
  - Log of tax on goods and services to GDP: WEO.
  - Lag of log of real GDP per capita: WEO.
  - Lag of log of real GDP per capita squared: WEO.
  - Trade openness—sum of imports and exports in percent of GDP: WEO.
  - Agriculture: Value added of agriculture in percent of GDP: World Bank, World Development Indicators (WDI).
  - Gini coefficient: WDI.
  - Oil: dummy for oil exporters.
  - General Government: dummy for General Government tax revenues.
  - Corruption and Government Effectiveness: Worldwide Governance Indicators (WGI).

*Source: chap2 - 2. DOMESTIC REVENUE MOBILIZATION IN SUB-SAHARAN AFRICA: WHAT ARE THE POSSIBILITIES? (PDF).*

### Annex 2.1. Estimating Tax Effort and Tax Potential

### Annex 2.1. Estimating Tax Effort and Tax Potential

### Main regression findings (Annex Table 2.1.1)
- Dependent variable: Log of tax/GDP (models A, B, C and subsamples reported).
- Key coefficients (selected, exact values as reported):
  - Log of real GDP per capita: 2.939 ***, 2.866 ***, 2.885 ***, 2.781 ***, 2.691 ***, 2.716 ***
  - Log of real GDP per capita squared: –0.152 ***, –0.148 ***, –0.150 ***, –0.142 ***, –0.138 ***, –0.140 ***
  - Trade openness: 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***
  - Gini coefficient: –0.006 ***, –0.006 ***, –0.007 ***, –0.006 ***, –0.006 ***, -0.006 ***
  - Education: 0.015 ***, 0.016 ***, 0.016 ***, 0.016 ***, 0.018 ***, 0.017 ***
  - General government: 0.105 **, 0.109 ***, 0.110 ***, 0.091 **, 0.093 **, 0.098 **
  - Corruption: 0.117 ***, 0.083 *, 0.134 ***, 0.100 **
  - Government effectiveness: 0.091 *, 0.088 *
  - Oil dummy (selected values): 0.080 **, 0.035, 0.031, 0.043, 0.030, 0.026
- Model diagnostics and samples:
  - Sigma_u: 0.515 ***, 0.515 ***, 0.516 ***, 0.525 ***, 0.526 ***, 0.526 ***
  - Sigma_e: 0.099 ***, 0.098 ***, 0.098 ***, 0.106 ***, 0.106 ***, 0.105 ***
  - Number of observations: 1,366; 1,360; 1,360; 1,109; 1,031; 1,03
  - Number of countries: 122; 121; 121; 99; 98; 98
- Dependent variable: Log of Goods and Services Tax/GDP (selected coefficients):
  - Log of real GDP per capita: 2.379 ***, 2.332 ***, 2.353 ***, 2.228 ***, 2.173 ***, 2.207 ***
  - Log of real GDP per capita squared: –0.122 ***, –0.120 ***, –0.122 ***, –0.113 ***, –0.111 ***, –0.114 ***
  - Trade openness: 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***, 0.002 ***
  - Gini coefficient: –0.008 ***, –0.008 ***, –0.008 ***, –0.009 ***, –0.009 ***, –0.009 ***
  - Oil dummy: –0.642 ***, –0.626 ***, –0.634 ***, –0.646 ***, –0.581 ***, –0.587 ***
  - Government effectiveness (selected): 0.170 **, 0.172 *
  - Sigma_u: 0.660 ***, 0.665 ***, 0.665 ***, 0.667 ***, 0.672 ***, 0.672 ***
  - Sigma_e: 0.162 ***, 0.159 ***, 0.159 ***, 0.177 ***, 0.174 ***, 0.174 ***
  - Number of observations: 1,152; 1,146; 1,146; 930; 924; 924
  - Number of countries: 105; 104; 104; 85; 84; 84
- Significance notation used in the table:
  - * p < .10; ** p < .05; *** p < .01

### Interpretation of model specifications and variables
- Models A, B and C correspond to the specifications listed in Annex Table 2.1.1:
  - Model A includes institutional factors and public spending on education.
  - Model B includes public spending on education but not corruption or government effectiveness.
  - Model C does not include corruption, government effectiveness or public spending on education.
- Notes on data timing for tax frontier estimates:
  - Data correspond to 2015 in most cases, with exceptions: Comoros, Seychelles, and Swaziland (all 2014), and Cabo Verde, Democratic Republic of the Congo, and Guinea-Bissau (all 2013).
  - Year selection requires data availability for the set of independent variables in the model.

### Estimates of countries’ tax frontier and country tax-to-GDP observations (Annex Table 2.1.2)
- Country-level reported Tax to GDP entries (selected reporting from table, exact values preserved):
  - Nigeria: 5.9 11.1 11.1 2.0 10.7 10.4 12.0 8.1 8.3 8.5
  - Central African Rep.: 6.2 8.4 8.5 9.7 8.0 8.2 8.8 8.1 7.9 8.8
  - Madagascar: 9.9 16.7 17.3 19.5 16.6 16.7 19.4 14.8 15.8 18.4
  - Tanzania: 12.4 20.3 20.9 19.5 20.2 19.8 19.7 18.3 18.6 19.4
  - Ethiopia: 12.7 13.8 14.2 13.2 13.9 13.8 13.1 13.3 13.3 13.1
  - South Africa: 24.7 26.9 26.9 31.1 27.9 27.6 30.5 25.5 25.4 26.2
  - Zimbabwe: 26.9 27.7 27.6 27.5 27.7 27.6 27.5 27.8 27.7 27.6
  - Swaziland: 28.3 30.4 29.8 30.1 30.4 30.4 30.5 30.3 30.4 29.6
  - Seychelles: 29.2 36.2 34.8 49.4 39.4 39.4 39.0 48.3 34.5 34.2 37.1
  - Namibia: 32.1 33.5 33.4 33.9 34.2 33.9 35.4 33.7 33.8 33.2
- Reported country-level range and variation are presented in the table for many Sub-Saharan African countries across the three model specifications and subsamples.
- Average row values as reported (exact sequence): 16.2 19.6 19.9 21.3 19.9 19.8 20.9 18.7 18.9 19.6

### Key methodological notes
- Source: IMF staff calculations.
- The tax frontier estimates use the models summarized in Annex Table 2.1.1 to compute potential tax-to-GDP (tax frontier) across Sub-Saharan African countries.
- Models incorporate macroeconomic variables (log of real GDP per capita and its square), openness, sectoral composition (agriculture), inequality (Gini), public spending on education, institutional variables (corruption, government effectiveness), and an oil dummy in various specifications.

*Source: IMF staff calculations (Annex 2.1, Regional Economic Outlook: Sub-Saharan Africa).*

---


_Source: https://www.imf.org/-/media/files/publications/reo/afr/2018/may/pdf/chap2.pdf_