## 1. Socioeconomic Progress, 1996–2022

## Source details

**Canonical URL:** [1. Socioeconomic Progress, 1996–2022](https://www.imf.org/-/media/files/publications/cr/2023/english/1rwaea2023004.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2023/english/1rwaea2023004.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2023/english/1rwaea2023004.pdf.json)

---

### Summary and key context
- Rwanda’s economy averaged 8.5 percent a year over the 1995–2022 period.
- Social indicators:
  - Life expectancy: 40.6 years in 1996 → 66.1 years in 2021.
  - Under-five mortality: 131.1 deaths per 1,000 live births in 1996 → 29.7 deaths per 1,000 live births in 2021.
  - Adult literacy: 57.8 percent in 1991 → 75.9 percent in 2021.
  - Primary school enrollment: 31.9 percent in 1996 → 91 percent in 2023.
  - Gross National Income per capita: US$ 833.3 in 1996 → US$ 2427 in 2022.
- Financing needs and gaps:
  - SDG financing gap by 2030: between 18.7 and 21.3 percent of GDP (past IMF studies, 2019 and 2021).
  - Additional spending to meet Climate goals by 2050: 8.8 percent of GDP per year.
  - Main long-term financing sources: concessional financing, external borrowing, domestic revenue mobilization. External financing options are recently limited or constrained.

### Progress in domestic revenue mobilization
- Tax revenue ratio rose from just under 8 percent of GDP to over 16 percent of GDP between 1992 and 2020.
- Composition of revenue gains:
  - Direct taxation contributed around 5 percent of GDP of this increase (driven by personal income tax).
  - Goods and services taxation added over 4.5 percent of GDP (excise increases in late 1990s; VAT introduced in 2001).
  - Trade taxes fell as a share of GDP.
- 2010–15: revenues increased by 3¾ percent of GDP due to base-widening, incentive rationalization, PIT progressivity improvements, and enforcement.
- Empirical panel: original panel of 1,845 reform measures for EAC countries (1988–2022) distinguishes announced package and non-package measures.
  - Finding: isolated (selective) tax measures tend to yield protracted revenue loss; package reforms are revenue neutral near term and lead to significant revenue increases beyond 3 years when fully implemented.
  - Revenue loss arises when parts of a reform package are implemented without accounting for revenue-losing aspects.

### Institutional and administrative reforms implemented
- RRA formed in 1997; MINECOFIN responsible for tax policy.
- Direct taxation reforms:
  - Simplified PIT rates in 2005; turnover tax and presumptive tax for small taxpayers; PIT base expansion and reduced CIT rates in 2000s.
  - Electronic tax administration platform in 2003; electronic filing in 2010; new electronic tax administration system in 2016.
- Indirect taxes:
  - VAT introduced in 2001 at 15 percent then increased to 18 percent in 2003; VAT exemptions removed in 2010/11.
  - Quarterly filing and electronic payments; value‑based mining royalties; strengthened transfer pricing attention.
- Trade taxes declined after customs duties reductions and EAC customs union accession in 2009.
- Administration measures:
  - Electronic filing/payment systems (2010–11), electronic tax registration, risk management, direct bank payment, integration of tax and business registration, customs Single Window, ETD for VAT compliance, withholding VAT at source by government departments, increased arrears collection.
- Exemptions reform: investment tax code revised; VAT exemptions on imports for investment certificate holders removed.

### Recent developments and emerging challenges
- Tax-to-GDP ratio: 15.8 percent of GDP in FY20/21 → 15 percent of GDP by end-FY22/23; peak 16.1 percent of GDP in FY2018/19.
- Indirect taxes: 7.8 percent of GDP in FY18/19 → 6.7 percent of GDP in FY22/23.
- Policy drivers of revenue weakness:
  - MBRP (2021) exempted approved projects from VAT and import duties on construction materials; extended until end-2024.
  - Reductions in fuel levy and excises between 2021 and 2023.
  - VAT exemptions on electric vehicles.
  - PAYE thresholds adjusted downward.
- Tax expenditures:
  - 3.7 percent of GDP in FY21/22 (up from around 3 percent of GDP in 2018-19).
  - VAT represents almost half of all tax expenditures; largest VAT expenditures arise from exemptions on financial services and agriculture.
- Sectoral composition shift:
  - Agriculture and food manufacturing nominal share: 26 percent → 30 percent of total GDP in 2022–23; both sectors largely untaxed.
  - Excluding these sectors suggests the fall in tax-to-GDP in FY22/23 can be attributed to composition of nominal GDP relative to the tax base.
- Medium-Term Revenue Strategy (MTRS):
  - Mid-2022 Cabinet approved MTRS to raise tax-to-GDP by 1 percent of GDP over MTRS period; package included some tax‑losing reforms (PIT threshold adjustment and CIT reductions) offset by excise increases and a minimum alternative tax.
  - To date, only loss-making tax measures have been adopted; core revenue-enhancing measures remain planned but not adopted.

### Policy implications emphasized
- DRM is critical to close development and climate financing gaps; tax potential remains sizeable.
- Priorities:
  - Comprehensive strategy: rate adjustments, base broadening, compliance improvements, curb evasion.
  - Implement reforms as coordinated packages rather than selective measures to avoid protracted revenue losses.
  - Shift tax burdens from low-income households to higher income/wealth cohorts; use targeted expenditure (direct transfers) to compensate regressive effects.
  - Focus on tax policy changes as compliance alone has diminishing returns.

---

### Revenue administration, tax effort, and the tax gap
- RRA measures: taxing the shadow economy; improving voluntary compliance through taxpayer services; promoting compliance improvement plans (CIPs) for manufacturing, large businesses, customs; combating individual tax evasion.
- Tax gap and tax effort:
  - Rwanda’s tax gap in 2019: [9.7] percent of GDP (DART database).
  - Tax revenues have fallen around a percentage point of GDP since 2019.
  - Tax potential has increased implying a larger tax gap by end-FY22/23.
- Composition:
  - Tax gap = policy gap + compliance gap.
  - Substantial improvements in compliance gap since 2010; policy gap remains significant due to multiple rate structures and numerous exemptions.
  - Administration reforms show diminishing returns over time.

### Empirical dataset and stylized facts
- Novel compilation: EAC economies 1988–2022 with 1,845 tax measures; documents direction (INCREASE/DECREASE) across policy (RATE, BASE) and administrative (ADMIN) practices for PIT, CIT, VAT, SSC, EXE, TRADE, PRO using IBFD news clips.
- EAC and Rwanda patterns:
  - EAC: 24.9 percent of measures narrow the tax base; 28.5 percent strengthen administrative practices.
  - Rwanda: 21.7 percent reduce the base; 28.7 percent strengthen administrative conditions.
  - 94 percent of EAC tax changes announced as part of a broad package; Rwanda: 82.9 percent comprehensive measures; 17.1 percent selective.
- Tax revenue trajectory:
  - 9.1 percent of GDP in 2001 → 16.1 percent of GDP in 2022.
  - Strong implementation post-2016; several measures after 2019 were revenue losing.

### Methodology and principal empirical results
- Local projections (LP) framework with 7-year window post-announcement to estimate impulse responses.
- Key regression findings:
  - Equation (1): one or more comprehensive fiscal measures → 2 percent average increase in tax-to-GDP by year 6; one selective reform → around 1 percent loss in tax-to-GDP.
  - Equation (2) disaggregating intent:
    - Comprehensive strengthening → about 2.9 percent of GDP increase in tax revenue on average.
    - Comprehensive revenue-losing → almost neutral (decline contained).
    - Selective strengthening → losses of about 1.4 percent of tax revenue to GDP.
    - Selective revenue-losing → losses of about 1.9 percent of GDP.
- Interpretation: selective measures often reduce revenue via market distortions or avoidance; administration gains take years and face diminishing returns.

### Reform revenue potential estimates and IMF TA
- IMF TA (2021 mission) integrated package potential: 7.2 percent of GDP total
  - 4.9 percent of GDP in the short to medium-term;
  - 2.3 percent of GDP over the longer-term.
- VAT reforms potential: 2.9 percent of GDP.
  - Rwanda’s VAT capacity: 10 percent of consumption expenditures.
  - Rwanda’s average VAT effort: 50.7.
  - Rwanda’s VAT effort ranks 55 out of 70 countries.
- CIT reforms potential: around 0.8 percent of GDP via:
  - reduction in corporate tax rate to 20 percent;
  - elimination/grandfathering of tax incentives;
  - full expensing for capital expenditure;
  - indefinite carry forward of losses;
  - introduction of a gross asset-based minimum alternative tax (MAT).
- Redundancy ratio for incentives: 0.98 (98 percent of the investment would have taken place had the incentive not been in place).
- Note: effectiveness is diminished if individual measures are selectively accepted or rejected.

### Key numeric facts and statistics (selected)
- Rwanda’s tax gap in 2019: [9.7] percent of GDP.
- Tax revenue: 9.1 percent of GDP in 2001; 16.1 percent in 2022.
- 1,845 tax measures in EAC during 1988–2022.
- EAC: 24.9 percent narrow tax base; 28.5 percent strengthen administration.
- Rwanda: 21.7 percent reduce base; 28.7 percent strengthen administration.
- EAC: 94 percent of changes as part of broad package; Rwanda: 82.9 percent comprehensive; 17.1 percent selective.
- Empirical effects:
  - Comprehensive measures → 2 percent increase in tax-to-GDP by year 6.
  - One selective reform → ~1 percent loss in tax-to-GDP.
  - Comprehensive strengthening → ~2.9 percent of GDP increase.
  - Selective strengthening → ~1.4 percent of GDP loss.
  - Selective revenue-losing → ~1.9 percent of GDP loss.
- Reform revenue potentials:
  - Integrated package: 7.2 percent of GDP (4.9 percent short/medium-term; 2.3 percent longer-term).
  - VAT potential: 2.9 percent of GDP.
  - VAT capacity: 10 percent of consumption expenditures.
  - VAT effort: 50.7; rank 55/70.
  - CIT potential: 0.8 percent of GDP.
  - Redundancy ratio: 0.98.
  - Proposed corporate tax rate: 20 percent.

---

### Personal Income Taxation and excises
- PIT structural weaknesses:
  - PAYE registered taxpayers ≈ 24 percent of estimated labor force, despite ≈ 70 percent of adult labor force employed outside subsistence agriculture.
  - High labor tax wedge for average formal salaried employee (PIT + social security contributions).
  - Recommendations: improve equity; transition to a global income tax system long run; harmonize withholding rates across capital taxation in short run; introduce a wealth tax.
- Excise design and recommendations:
  - Current: all excises ad valorem except fuel and partial specific excise on cigarettes.
  - Issues: excise revenues decreased; lack of indexation of specific excises; ad valorem use may reduce revenue stability.
  - Recommendations: change to only specific excises on alcohol, tobacco, and sugar beverages combined with indexation; introduce higher excise on old vehicles, reduce motor vehicle fee, and introduce higher taxes on fuels to incentivize fuel-efficient vehicles.

---

### Climate financing, NDCs, and DIGNAD modelling

#### NDC targets, pipeline, and costs
- Rwanda’s updated NDC: reduce GHG emissions by 38 percent during 2020–30.
- Overall NDC implementation cost: US$11,040 million (US$11 billion), implying investments amounting to 7 percent of GDP each year during the ten-year period.
- NDC Implementation Framework:
  - 15 outcomes; 56 outputs; 91 KPIs (41 adaptation, 45 mitigation, 5 cross-cutting).
  - Identifies 566 projects.
- Project status (Table 2):
  - Ongoing: 190 projects; Project value USD million 3,662.5.
  - Planned: 88 projects; Project value USD million 2,333.9.
  - Without support: 21 projects; Project value USD million 60.0.
  - Indicative: 48 projects; Project value USD million 1,192.2.
  - To be determined: 2 projects; Project value USD million 40.0.
  - Total: 566 projects; Project value USD million 7,188.7.
  - Ongoing + planned: 278 projects; Project value USD million 5,996.4.
- Sectoral composition by project value:
  - Energy 47 percent.
  - Transport 25 percent.
  - Water and sanitation 15 percent.
  - Agriculture and water and sanitation together account for almost 3/4 of total emissions but represent 5 percent and 15 percent of total project value, respectively.
- Implementing agencies: Central government 83 percent; Local government 15 percent.
- Funding composition: Grants and budget resources 89 percent.

#### Financing secured, gaps, and realization
- Total 566 projects value: US$7,188.7 million.
- Ongoing and planned projects gross: US$5,996 million.
- Projects started before 2020: US$1,481 million.
- Ongoing and planned projects net (excluding pre-2020): US$4,515 million.
- NDC financing need (USD million):
  - 2020-2025: 5,101
  - 2025-2030: 5,939
  - Total: 11,040
- NDC financing gap:
  - 2020-2025: -586 (USD million)
  - 2025-2030: -5,939 (USD million)
  - Total: -6,525 (USD million)
- IMF staff adaptation investment estimate up to 2030: about US$10 million per year (contrasts with authorities’ estimate of US$500 million per year).
- Realization shortfalls:
  - Analysis of 13 largest ongoing/planned projects: four major projects total value US$1,900 million have not been appropriated in FY20/21–FY22/23 budgets.
  - Only 12 percent of the total spending expected for 2020–25 has been realized in the first three fiscal years.
- Selected large-project planned vs realized spending:
  - Nyabarongo II Hydro Power Plant (2.43.5MW): Planned 2020-25 USD 157.7; Spending till end-FY22/23 USD 20.0; Actual vs planned 13%.
  - Rwanda Universal Energy Access Program (one entry): Planned USD 395.4; Spending till end-FY22/23 USD 197.7; Actual vs planned 50%.
  - Aggregate for large projects: Planned USD 3,135; Spending till end-FY22/23 USD 387.2; Actual vs planned 12%.

#### Ireme Invest (Green Private Investment Facility) and private climate financing
- Pledges/commitments:
  - Pledges from bilateral partners and GCPF: US$104 million.
  - Further commitments at June 2023 Paris Climate Summit: EUR 150 million.
- Expected disbursements:
  - Till end of this year: expected US$20 million by AFD; another US$20 million by EIB available next year.
  - Considering RDB resources and assumed 30 percent own private resources, facility expected to mobilize US$21 million this year and about US$245 million in 2024–30 period.
- Ireme project pipeline total project value: US$145 million (hydro power project 2023; green city Kigali and e-buses 2024).
- Outlook: lending operations likely to scale up gradually; Ireme to mobilize private financing via SLBs and green bonds blended with concessional financing; private green financing unlikely to significantly reduce Rwanda’s climate financing gap in next couple of years.

#### Fiscal space, crowding out, and DRM implications
- NDC cost is close to 70 percent of public investment for most of 2020–30 period.
- SDGs estimated cost: 18.7 percent of GDP.
- If fully implemented, climate strategy would crowd out other development investment absent additional fiscal space.
- External financing: externally financed (conditional) NDC component would amount to half of total loans and grants projected by end of 2020–30 period.
- DRM projection example: under ambitious MRTS-2 strategy tax-to-GDP expected to increase from current 15 percent to close to 20 percent by 2030.

#### DIGNAD model description, scenarios, and key simulation results
- DIGNAD extends two-sector DIG by introducing natural disasters and allowing government investment in standard infrastructure, adaptation capital, and financial buffers; calibrated to Rwanda.
- Natural disaster shock: modeled as once-in-100-years flooding hitting in 2027.
- Scenario setups and financing assumptions (exact values where reported):
  - Baseline: no climate adaptation investment.
  - Scenario 1 (NDC full implementation): adaptation conditional overall 2.1 percent of GDP per year (cumulative USD 5.3 billion); financing mix external concessional 1.4% and grants 0.7% (historical ratio 2:1).
  - Scenario 2 (funding squeeze without policy adjustment): concessional financing and grants limited to 1.1 percent of GDP.
  - Scenario 3 (funding squeeze with policy adjustment): offsets include DRM 0.5% of GDP (VAT) and spending savings 0.5% of GDP; table summary lists adaptation conditional overall 3.9%, external concessional 0.8%, grants 2.1%, DRM 0.5%, spending rationalization 0.5%.
  - Scenario 4 (policy adjustment + full RSF implementation): adaptation conditional overall 4.3%, external concessional 0.8%, grants/private 2.5%, DRM 0.5%, spending rationalization 0.5%, public investment efficiency +20 percentage points.
- Key model results (exact reported values):
  - Baseline no adaptation with disaster in 2027:
    - real GDP decline: 4 percent.
    - reconstruction implies 4 percentage points increase in public investment to GDP ratio.
    - debt-to-GDP ratio increases by 2 percentage points.
  - Scenario 1 (full NDC):
    - GDP drop after shock: 1 percentage points.
    - reconstruction investment needs: 1.5 percent of GDP.
    - debt-to-GDP ratio increases by 6 percentage points (due to concessional borrowing).
    - fiscal deficit increases by 1.5 percentage points during additional climate spending; current account widens.
  - Scenario 2: partial NDC with limited concessional/grants (1.1% of GDP) → larger adverse GDP and sustainability impacts than Scenario 1 (specific magnitudes not restated).
  - Scenario 3 (0.5% VAT + 0.5% spending cuts): GDP decline only slightly more pronounced than Scenario 1 and much smaller than Scenario 2; increases in debt, fiscal deficit, and current account are more muted than in Scenario 1.
  - Scenario 4 (full RSF on top of Scenario 3): higher public spending efficiency and private mobilization yield most positive macro outcomes; more medium/long-run GDP, unchanged debt levels lead to lower debt-to-GDP ratio; fiscal and current account deficits similar to Scenario 3.
- Policy conclusions:
  - Adaptation investment mitigates GDP losses from disasters but if financed mainly by concessional debt increases debt, deficits, and current account deficits.
  - Financing mix matters: DRM and spending rationalization limit output loss while safeguarding debt sustainability.
  - Improving public investment efficiency and mobilizing private financing (RSF) yield best macro outcomes.
  - Recommended package: increase fiscal space via DRM; spending rationalization; enhance public investment efficiency; mobilize private climate investment.

---

### Monetary policy, NBR strategy, and FX market issues

#### NBR inflation objective and toolkit
- Inflation benchmark: range of 5±3 percent.
- MPS explicit objective: keep headline CPI within band of 2 to 8 percent, focus on being close to 5 percent in medium term.
- Policy instrument: set Central Bank Rate (CBR) to stabilize inflation in the medium term under a forward-looking framework.
- MPS published February 2023.
- Instruments: OMO, repo (1 to 7 days), reverse repo, Central Bank Bills (28, 91, 182, 364 days), overnight deposits facility, standing lending facility (overnight), refinancing facility up to 30 days.
- Reserve requirements: base = all bank liabilities up to one year excluding interbank; reserve averaging allowed since March 2005; current maintenance period two months; excess reserves not remunerated.

#### FX intervention practice and classification
- NBR sells USD to domestic banks in regular bilateral FXI under a de facto crawl-like exchange rate arrangement.
- Interventions almost weekly at preannounced intervention rates; USD sale amount derived from NBR’s available USD supply and banks’ self-reported USD needs.
- AREAER classification: de facto crawl-like arrangement; de jure arrangement floating.

#### Frictions to transmission and recent experience
- LLMIC frictions: weak asset/financial markets, ineffective communication, high financial illiteracy; result: transmission mainly via interest rate, credit, and exchange rate channels.
- Rwanda evidence:
  - Transmission from policy rate to money market rates improved; transmission to lending rates not yet statistically evident per NBR 2017–18 report.
  - Kwizeraa and Ndarihoranye (2022): evidence of an operational bank lending channel.
  - Forthcoming NBR research (quarterly 2006–2022): exchange rate channel exists.
- Inflation dynamics and food role:
  - Inflation volatile and more volatile than SSA median.
  - 300-bps cumulative hike between February 2022 and August 2023; policy tightening since December 2021 small compared to other SSA countries.
  - High food weight in CPI; two consecutive low domestic agricultural production years contributed to high food inflation since 2022 alongside higher international food, fuel, and fertilizer prices.
- Bank concentration and pass-through:
  - Average lending rate flat Jan 2016–May 2023.
  - High bank concentration and recent mergers reduce pass-through from policy rate to lending rates.
  - Market dominated by few large, well-capitalized borrowers with significant collateral.
- FX shortages and interventions:
  - Structural FX shortages due to persistent current account deficit; intensified as donor financing declined.
  - Interbank FX activity muted; NBR regularly sells FX to banks (one-sided interventions) indicating disequilibrium.
  - IMF TA (September 2023) recommends addressing structural USD backlog before reducing interventions.

#### Box 1: FX market structural issues and transmission mechanism findings
- Structural issues:
  - Rwanda runs trade deficits; capital inflows mainly insensitive to interest rates.
  - High remittances and foreign aid share of GDP; FDI and portfolio inflows marginal.
  - NBR intermediary role converting external concessional USD to RWF contributes to FX market structure.
  - Dollarization of deposits reduces monetary policy transmission to deposit rates; self-sustaining depreciation expectations and USD hoarding.
- VAR and Granger causality results (sample/periods reported):
  - VAR sample: Q4 2010–Q1 2023; ordering: nominal ER depreciation rate, interbank rate, private credit growth, inflation, GDP growth.
  - VAR: credit growth and interbank rate increase with M2 growth; exchange rate response not statistically different from zero.
  - Granger causality (Jan 2009–Aug 2023): policy rate Granger-causes interbank rate (7-day) and lending rate (average) but not vice versa; weak pass-through to T-bill and deposit rates.
  - After Jan 2019 transition to interest-rate framework: pass-through from central bank rate to interbank and T-bill rates seems strengthened; central bank rate now affects lending rate though magnitude remains limited.
- Exchange rate pass-through (ERPT):
  - At 12-month horizon, cumulative response of inflation to 1 percent depreciation in bilateral ER vs USD ≈ 0.4 percent for Rwanda (Kemoe et al., Forthcoming).
  - Rwanda’s ERPT almost double SSA average; ERPT declined since 2017.
  - Managed floats (including Rwanda) have higher ERPT on average than pegs or floats in SSA.
- Policy implications:
  - First-stage transmission (policy → ER) weak due to de facto crawl-like regime; second-stage (ER → inflation) strong, amplifying depreciation effects.
  - Strengthening monetary policy effectiveness limits second-round effects.
- NBR efforts to strengthen transmission:
  - Communication, financial inclusion, financial education initiatives: guideline for women’s financial inclusion; mobile app "GERERANYA"; NBR quiz challenge; university monetary policy challenges; savings week; insurance week; global money week; Financial Consumer Protection Law awareness.
  - Quarterly meetings of Chief Economist with bank treasurers; bond market development in H1 2023: three new bonds and four reopenings; average subscription rate 117.1 percent (up from 109.4 percent year earlier).
- Concluding policy recommendations:
  - Use policy rate more proactively; interest-rate based frameworks yield more effective transmission.
  - Ensure monetary policy decisions are data driven and forward-looking; improve forecasting and policy analysis integration into policymaking.
  - Allow greater exchange rate flexibility while addressing structural FX market issues.

*Source: 1rwaea2023004 - IMF staff report (Selected Issue Paper) as provided in the content unit.*

### 1. Socioeconomic Progress, 1996–2022 _________________________________________________ 4

### 1. Socioeconomic Progress, 1996–2022

### Summary and Key Context
- Rwanda’s economy grew rapidly over three decades, averaging 8.5 percent a year over the 1995–2022 period.
- Significant social improvements accompanied growth: life expectancy rose from 40.6 years in 1996 to 66.1 years in 2021; under-five mortality fell from 131.1 deaths per 1,000 live births in 1996 to 29.7 deaths per 1,000 live births in 2021; adult literacy increased from 57.8 percent in 1991 to 75.9 percent in 2021; primary school enrollment increased from 31.9 percent in 1996 to 91 percent in 2023; Gross National Income per capita rose from US$ 833.3 in 1996 to US$ 2427 in 2022.
- Financing gaps remain large: past IMF studies (2019 and 2021) estimate the financing gap to achieve the SDGs by 2030 ranges between 18.7 and 21.3 percent of GDP. Estimates also suggest Rwanda will need an additional 8.8 percent of GDP in spending each year to meet Climate goals by 2050.
- Rwanda’s main long-term financing sources are concessional financing, external borrowing, and domestic revenue mobilization. External financing options are recently limited or constrained.

### Progress in Domestic Revenue Mobilization — Lessons from Past Reforms
- Rwanda increased its tax revenue ratio substantially: from just under 8 percent of GDP to over 16 percent of GDP between 1992 and 2020.
- Composition of revenue gains:
  - Direct taxation contributed around 5 percent of GDP of this increase, driven by personal income tax.
  - Goods and services taxation added over 4.5 percent of GDP (through excise increases in the late 1990s and the introduction of a VAT in 2001).
  - Trade taxes fell as a share of GDP.
- The period 2010–15 saw revenues increase by 3¾ percent of GDP due to reforms widening the tax base (tax incentives rationalization), improving progressivity of personal income taxes, and enforcing tax compliance.
- Dataset and empirical finding:
  - An original panel of 1,845 reform measures for EAC countries over the 1988-2022 period distinguishes announced package and non-package tax measures.
  - Findings: isolated (selective) tax measures tend to yield protracted loss of revenue; tax measures implemented as a package are revenue neutral in the near term and, beyond 3 years, lead to significant increases in revenue. Comprehensive reforms, when fully implemented, are designed to yield net positive revenue gains.
  - Revenue loss arises when parts of a reform package are implemented without accounting for revenue-losing aspects.

### Institutional and Administrative Reforms Implemented
- Rwanda Revenue Authority (RRA) was formed in 1997; MINECOFIN responsible for tax policy.
- Major administrative and policy reforms:
  - Direct taxation: simplified PIT rates in 2005; turnover tax and presumptive tax introduced for small taxpayers; base expansion for PIT and reduced corporate tax rates in the 2000s; electronic tax administration platform in 2003; electronic filing in 2010; new electronic tax administration system in 2016.
  - Indirect taxes: VAT introduced in 2001 at rate of 15 percent then increased to 18 percent in 2003; VAT exemptions removed in 2010/11; quarterly filing and electronic payments introduced; value‑based scale for mining royalties; strengthened attention to transfer pricing.
  - Trade taxes: declined after customs duties reductions and EAC customs union accession in 2009.
  - Tax administration measures: electronic filing and payment systems (2010–11), electronic tax registration, risk management approaches, direct bank payment of tax, integration of tax and business registration, customs Single Window, ETD for VAT compliance, withholding VAT at source by government departments, increased collection of tax arrears.
- Reforms to exemptions: investment tax code revised to streamline exemptions; VAT exemptions on imports for investment certificate holders were removed to broaden the tax base.

### Recent Developments and Emerging Challenges
- Domestic resource mobilization has stalled: tax-to-GDP ratio fell from 15.8 percent of GDP in FY20/21 to 15 percent of GDP by end-FY22/23; ratio peaked at 16.1 percent of GDP in FY2018/19.
- Falls observed across tax types, more pronounced for indirect taxes: indirect taxes fell from 7.8 percent of GDP in FY18/19 to 6.7 percent of GDP in FY22/23.
- Sectoral and policy drivers of recent revenue weakness:
  - Policy measures during COVID-19 and thereafter reduced revenues: Manufacturing Build to Recovery Program (MBRP) introduced in 2021 exempted approved projects from VAT and import duties on construction materials and was extended until end-2024; reductions in fuel levy and excises between 2021 and 2023 in response to global oil prices; VAT exemptions on electric vehicles; PAYE thresholds adjusted downward.
  - Tax expenditures rose: last tax expenditure report shows tax expenditures were 3.7 percent of GDP in FY21/22, up from around 3 percent of GDP in 2018-19. VAT represents almost half of all tax expenditures; largest VAT tax expenditures arise from exemptions on financial services and agriculture sectors.
  - Sectoral composition shift: agriculture and food manufacturing nominal share increased from 26 to 30 percent of total GDP in 2022–23; both sectors are untaxed. Excluding these sectors from GDP suggests the fall in the tax-to-GDP ratio in FY22/23 can be attributed to changing composition of nominal GDP relative to the tax base.
- Medium-Term Revenue Strategy (MTRS) and implementation gap:
  - In mid-2022 Cabinet approved an MTRS designed to raise the tax-to-GDP ratio by 1 percent of GDP over the course of the MTRS; the MTRS was designed to be implemented as a package including some tax‑losing reforms (adjustment to PIT thresholds and reductions in CIT rates) offset by revenue from excise increases and introduction of a minimum alternative tax.
  - To date, only loss-making tax measures have been adopted by the authorities, while core revenue-enhancing measures have not yet been adopted but remain planned.

### Policy Implications and Recommendations Emphasized
- Domestic revenue mobilization is critical to close development and climate financing gaps; tax potential remains sizeable.
- Priority policy direction:
  - Pursue a comprehensive strategy to increase tax revenues combining rate adjustments, broadening the domestic tax base, improving tax compliance, and curbing tax evasion.
  - Implement reforms as coordinated packages rather than selective isolated measures to avoid protracted revenue losses and to capture medium-term revenue gains.
  - Shift higher tax burdens from low-income households to higher income/wealth cohorts to advance distributional fairness amid growing inequality.
  - Focus on tax policy changes as improved compliance alone has diminishing returns.

*Source: 1rwaea2023004 - 1. Socioeconomic Progress, 1996–2022*

### 13.      Despite the setbacks on the tax policy side, Rwanda has made improvements in

### 1rwaea2023004 - 13.      Despite the setbacks on the tax policy side, Rwanda has made improvements in

### Revenue administration, tax effort, and tax gap
- Rwanda Revenue Authority (RRA) has implemented measures aimed at:
  - taxing the shadow economy;
  - improving voluntary compliance through better taxpayer services;
  - promoting compliance improvement plans (CIPs) targeted at the manufacturing sector, large businesses, customs, and to combat tax evasion by individuals.
- Rwanda’s tax effort performs consistently below its tax potential, producing a significant tax gap:
  - Rwanda’s tax gap in 2019 was [9.7] percent of GDP (DART database).
  - Tax revenues have fallen around a percentage point of GDP since 2019, and Rwanda’s tax potential has increased, implying a larger tax gap by the end of FY22/23.
- Composition and dynamics of the tax gap:
  - The tax gap comprises a policy gap and a compliance gap.
  - Rwanda has seen substantial improvements in the tax compliance gap since 2010, while the policy gap remains significant due to multiple rate structures and numerous exemptions.
  - Administration reforms show diminishing returns over time and should not be relied upon exclusively for short-term revenue gains.

### Empirical analysis, dataset, and stylized facts
- Dataset and scope:
  - Novel compilation covering the EAC economies over 1988–2022 with 1,845 tax measures.
  - Documents direction of changes (INCREASE/DECREASE) across policy (RATE, BASE) and administrative (ADMIN) practices for PIT, CIT, VAT, SSC, EXE, TRADE, PRO.
  - Source of news clips: International Bureau of Fiscal Documentation (IBFD).
- EAC and Rwanda patterns:
  - For the EAC, 24.9 percent of measures aimed at narrowing the tax base; 28.5 percent aimed at strengthening tax administrative practices.
  - For Rwanda, 21.7 percent of reforms reduced the base; 28.7 percent strengthened administrative conditions.
  - During 1988–2022, 94 percent of tax changes in EAC were announced as part of a broad package; in Rwanda, 82.9 percent were comprehensive and 17.1 percent selective.
- Rwanda tax revenue trajectory:
  - Tax revenue moved from 9.1 percent of GDP in 2001 to 16.1 in 2022.
  - Strong implementation of tax measures after 2016; several measures after 2019 were revenue losing.

### Methodology and principal empirical results
- Framework and modelling:
  - Local projections (LP) framework applied with a 7-year window post-announcement to estimate impulse response functions.
  - Primary regression specification (Equation (1)) models change in tax revenue (% of GDP) as a function of comprehensive (P) and selective (D) measure dummies, with country and time fixed effects.
  - Extended specification (Equation (2)) disaggregates measures by comprehensive/selective and strengthening/loosening intent: CI, CD, SI, SD dummies.
- Main empirical findings:
  - Comprehensive measures are associated with medium-term revenue increases; selective measures tend to yield protracted revenue losses.
  - From Equation (1):
    - Having “one or more” comprehensive fiscal measures yields a 2 percent on average increase in tax-to-GDP percentage by year 6.
    - Having one selective reform tends to yield around 1 percent losses in tax revenue to GDP.
  - From Equation (2), controlling for intent:
    - Comprehensive measures aimed at strengthening taxes lead to about 2.9 percent of GDP increase in tax revenue on average.
    - Comprehensive measures judged revenue losing are almost neutral (decline contained).
    - Selective measures aimed at strengthening taxes generate losses of about 1.4 percent of tax revenue to GDP.
    - Selective measures assessed as “revenue losing” amplify losses to 1.9 percent of GDP.
  - Interpretation notes:
    - Selective tax measures often reduce revenue by distorting market dynamics or enabling avoidance behaviors.
    - Gains from tax administration may take several years to materialize and face diminishing returns; recent digitalization could have shifted the administration frontier.

### Policy recommendations and sector-specific measures
- Overarching strategy:
  - Domestic resource mobilization must be led by tax policy reforms; revenue administration gains are diminishing and cannot substitute for policy reform.
  - A comprehensive strategy should adjust rates, broaden the domestic tax base, improve compliance, and curb tax evasion.
  - Reforms should shift higher tax burden from low-income households to higher income/wealth cohorts to advance distributional fairness.
  - Targeted expenditure (direct transfers) should compensate low-income households for regressive effects.
  - Reforms are most effective if implemented as an integrated package rather than piecemeal.
- IMF technical assistance estimates (2021 mission):
  - An integrated package of reforms could raise 7.2 percent of GDP in revenues:
    - 4.9 percent of GDP in the short to medium-term;
    - 2.3 percent of GDP over the longer-term.
  - Effectiveness would be seriously diminished if individual measures are selectively accepted or rejected.
- VAT-specific recommendations:
  - VAT reforms have the greatest revenue potential: an estimated 2.9 percent of GDP from VAT reforms.
  - Rwanda’s VAT capacity: 10 percent of consumption expenditures.
  - Rwanda’s average VAT effort: 50.7 (collects 50.7 percent of feasible VAT revenues).
  - Rwanda’s VAT effort ranks 55 out of 70 countries.
  - Policy actions: standard-rate certain zero-rated or exempt domestic goods; remove exemptions to reduce input VAT allocations and refunds; base-broadening to reduce policy and compliance gaps.
- Corporate Income Tax (CIT) recommendations:
  - Address distortions: bias for debt financing, transfer pricing, gold plating of revenue expenses, and plethora of tax holidays/reduced sector rates.
  - IMF technical assistance medium-term levy estimate: around 0.8 percent of GDP could be raised by:
    - reduction in corporate tax rate to 20 percent;
    - elimination/grandfathering of all tax incentives;
    - full expensing for capital expenditure;
    - indefinite carry forward of losses;
    - introduction of gross asset-based minimum alternative tax (MAT).
  - Redundancy ratio estimated at 0.98 (98 percent of the investment would have taken place had the incentive not been in place).

### Key numeric facts and statistics (as reported)
- Tax gap and revenues:
  - Rwanda’s tax gap in 2019: [9.7] percent of GDP.
  - Tax revenues fallen around a percentage point of GDP since 2019.
  - Tax revenue: 9.1 percent of GDP in 2001; 16.1 in 2022.
  - Rwanda’s tax effort around 15 percent of GDP in FY22/23 (fallen by around one percentage point).
- Dataset and measures:
  - 1,845 tax measures in EAC during 1988–2022.
  - EAC: 24.9 percent of measures narrow the tax base; 28.5 percent strengthen administrative practices.
  - Rwanda: 21.7 percent reduce the base; 28.7 percent strengthen administrative conditions.
  - EAC: 94 percent of tax changes announced as part of a broad package.
  - Rwanda: 82.9 percent comprehensive measures; 17.1 percent selective measures.
- Empirical effects:
  - One or more comprehensive measures → 2 percent increase in tax-to-GDP by year 6.
  - One selective reform → around 1 percent loss in tax-to-GDP.
  - Comprehensive strengthening → ~2.9 percent of GDP increase.
  - Selective strengthening → ~1.4 percent of GDP loss.
  - Selective revenue-losing → ~1.9 percent of GDP loss.
- Reform revenue potentials:
  - Integrated package potential: 7.2 percent of GDP total (4.9 percent short/medium-term; 2.3 percent longer-term).
  - VAT potential: 2.9 percent of GDP.
  - VAT capacity: 10 percent of consumption expenditures.
  - VAT effort: 50.7; rank 55 out of 70.
  - CIT potential from reforms: around 0.8 percent of GDP.
  - Redundancy ratio for incentives: 0.98.
  - Proposed corporate tax rate: 20 percent.

*Source: IMF staff calculations.*

### 33.      Personal Income Taxation could be made more progressive. PIT suffers from important

### 33.      Personal Income Taxation could be made more progressive. PIT suffers from important

### Personal Income Taxation: structural weaknesses and recommendations
- Structural weaknesses hamper PIT progressivity and increase opportunities for tax avoidance and evasion.
- PAYE registered taxpayers account for only about 24 percent of the estimated labor force, even though about 70 percent of the adult labor force is estimated to be employed outside of subsistence agriculture.
- There is a high labor tax wedge for the average formal salaried employee, including both PIT and social security contributions.
- Recommended focus:
  - Improving equity.
  - Transitioning to a global income tax system in the long run.
  - In the short run, harmonize withholding rates across capital taxation.
  - Introducing a wealth tax.

### Excise tax reform recommendations
- Current design:
  - All of Rwanda’s excises are ad valorem except for fuel and a partial specific excise on cigarettes.
- Observed issue:
  - Excise revenues have decreased in recent years, potentially due to not indexing the specific excises.
  - Another reason may be the use of ad valorem rather than specific excises.
- IMF technical assistance recommendations:
  - Change to (only) specific excises on alcohol, tobacco and sugar beverages combined with indexation to adjust to inflation.
  - To incentivize the purchase of fuel-efficient vehicles:
    - Introduce a higher excise on old vehicles.
    - Reduce the motor vehicle fee.
    - Introduce higher taxes on fuels.

### Summary: CHALLENGES AND POLICY OPTIONS TO FINANCING RWANDA’S CLIMATE AGENDA
- Purpose: Illustrate macroeconomic impacts of planned climate-focused projects and reforms and discuss sustainable financing strategies.
- Model: Simulations using the IMF’s Debt-Investment-Growth-Natural-Disasters (DIGNAD) model.
- Main conclusions:
  - Full NDC implementation has sizeable fiscal costs but could improve Rwanda’s macroeconomic stability if accompanied by proper policy and financing mix.
  - Domestic Revenue Mobilization (DRM) and spending rationalization are key to increasing fiscal space and safeguarding debt sustainability.
  - Advancing reforms under the Resilience and Sustainability Facility (RSF) would increase public investment efficiency and help catalyze additional climate financing.

### Rwanda’s vulnerability and NDC targets
- Rwanda is vulnerable to climate shocks; according to Notre Dame Global Adaptation Initiative (ND-GAIN) index, Rwanda is more vulnerable to climate disasters than SSA and LIDC averages.
- Rwanda’s updated NDC aims to reduce Green-House Gas (GHG) emissions by 38 percent during the 2020–30 period.
- Overall NDC implementation cost estimated at US$11 billion, implying investments amounting to 7 percent of GDP each year during the ten-year period.

### NDC Implementation Framework: pipeline, sectors, and funding
- NDC Implementation Framework composition:
  - 15 outcomes; 56 outputs; 91 key performance indicators (41 KPIs for adaptation, 45 for mitigation, 5 for cross-cutting).
  - Identifies 566 projects.
- Project status (Table 2):
  - Ongoing: Number of projects 190; Project value USD million 3,662.5.
  - Planned: Number of projects 88; Project value USD million 2,333.9.
  - Without support: Number of projects 21; Project value USD million 60.0.
  - Indicative: Number of projects 48; Project value USD million 1,192.2.
  - To be determined: Number of projects 2; Project value USD million 40.0.
  - Total: Number of projects 566; Project value USD million 7,188.7.
  - Ongoing + planned: Number of projects 278; Project value USD million 5,996.4.
- Sectoral composition by project value:
  - Energy 47 percent.
  - Transport 25 percent.
  - Water and sanitation 15 percent.
  - Agriculture and water and sanitation (together) account for almost 3/4 of total emissions but represent 5 percent and 15 percent of total project value, respectively.
- Implementing agencies:
  - Central government 83 percent.
  - Local government 15 percent.
- Funding composition:
  - Grants and budget resources 89 percent.

### Financing secured, gaps, and realization
- Total value of 566 projects: US$7 billion.
- Ongoing and planned projects value: US$5.9 billion.
- When projects started before 2020 are deducted from ongoing and planned projects, net ongoing and planned projects value: US$4.5 billion.
- Financing gap for 2020–25 period: US$0.6 billion.
- Overall financing gap for 2020–30 period (authorities’ official estimate): US$ 6.5 billion.
- IMF staff estimates of adaptation investment needs up to 2030: about US$10 million per year (Aligishiev, Bello, and Massetti (2022)) compared to the authorities’ estimate of US$500 million per year.
- Realization shortfalls:
  - Analysis of 13 largest ongoing and planned projects: four major projects with total project value of US$1.9 billion have not been appropriated in the FY20/21–FY22/23 budgets.
  - Only 12 percent of the total spending expected for the 2020–25 period has been realized in the first three fiscal years.

### Fiscal space, crowding out, and DRM implications
- The overall cost of NDC implementation amounts to close to 70 percent of public investment for most of the 2020–30 period.
- Rwanda’s Sustainability and Development Goals (SDGs) estimated cost: 18.7 percent of GDP.
- If fully implemented, the climate strategy would crowd out other development investment absent additional fiscal space.
- External financing conditions:
  - The externally financed (conditional) NDC component would amount to half of the total loans and grants projected by the end of the 2020–30 period.
- DRM projection example:
  - Under assumption of an ambitious MRTS-2 strategy the tax to GDP ratio is expected to increase from the current 15 percent to close to 20 percent by 2030.

### Private climate financing and Ireme Invest
- Ireme Invest (Green Private Investment Facility) objectives: mobilize additional green financing, both public and private.
- Pledges and commitments:
  - Pledges from bilateral development partners and the Green Climate Partnership Fund (GCPF) of US$104 million.
  - Further commitments at 2023 June Paris Climate Summit: EUR 150 million.
- Disbursement status and expectations:
  - Till the end of this year, only the US$20 million by Agence Française de Développement (AFD) is expected to be disbursed, while another US$20 million by the European Investment Bank’s (EIB) will be available next year.
  - Considering Rwanda Development Board’s own resources, and assumed 30 percent of own private resources, the facility is expected to mobilize US$21 million this year and about US$245 million in the 2024–30 period.
- Ireme’s project pipeline total project value: US$145 million (hydro power project in 2023; green city Kigali and e-buses projects for 2024).
- Outlook:
  - Lending operations are likely to be scaled up only gradually.
  - Ireme Invest expected to mobilize private sector financing by issuing sustainability linked bonds (SLBs) and green bonds and blend these with concessional financing.
  - Even with sizeable SLB/green bond issuance, private sector green financing is unlikely to significantly reduce Rwanda’s climate financing gap in the next couple of years.

### Key numerical tables and figures (selected exact figures)
- NDC financing need (USD million):
  - 2020-2025: 5,101
  - 2025-2030: 5,939
  - Total: 11,040
- Ongoing and planned projects, gross: 5,996 (USD million)
- Projects started before 2020: 1,481 (USD million)
- Ongoing and planned projects, net: 4,515 (USD million)
- NDC financing gap:
  - 2020-2025: -586 (USD million)
  - 2025-2030: -5,939 (USD million)
  - Total: -6,525 (USD million)
- Table 2 project-value totals:
  - Ongoing project value: 3,662.5 (USD million)
  - Planned project value: 2,333.9 (USD million)
  - Total project value: 7,188.7 (USD million)
- Selected large-project planned vs realized spending (from Table 4 examples):
  - 2.43.5MW Nyabarongo II Hydro Power Plant: Planned spending in 2020-25 USD 157.7; Spending till end-FY22/23 USD 20.0; Share of actual vs planned spending 13%.
  - Rwanda Universal Energy Access Program (one entry): Planned spending USD 395.4; Spending till end-FY22/23 USD 197.7; Share of actual vs planned spending 50%.
  - Aggregate for large projects: Planned spending USD 3,135; Spending till end-FY22/23 USD 387.2; Share of actual vs planned spending 12%.

*Source: IMF staff report (Selected Issue Paper) as provided in the content unit.*

### 10.      This section uses the Debt-Investment-Growth-Natural-Disasters (DIGNAD) model to

### 10.      This section uses the Debt-Investment-Growth-Natural-Disasters (DIGNAD) model to 

### DIGNAD model description and calibration
- DIGNAD extends the two-sector DIG framework (traded and non-traded goods) by introducing natural disasters and allowing government investment in:
  - standard infrastructure (e.g., roads) and adaptation capital (e.g., seawalls or climate-proof roads)
  - financial buffers.
- Natural disasters in the model cause:
  - permanent damages to public and private capital,
  - temporary losses of productivity,
  - inefficiencies during reconstruction,
  - damages to the sovereign's creditworthiness.
- Features and mechanisms captured:
  - adaptation infrastructure depreciates at a lower rate than standard infrastructure,
  - adaptation raises the marginal product of other capital and crowds in private investment,
  - interactions among public investment, growth, and debt (investment-growth nexus, fiscal adjustment, private sector response).
- Calibration to Rwanda:
  - country-specific macroeconomic indicators and historical averages used where possible,
  - the average return on standard infrastructure is set at 5 percentage points lower than that of adaptation infrastructure,
  - adaptation capital assumed to depreciate at a lower rate than standard infrastructure,
  - model allows calibration of size and timing of a natural disaster and the recovery process.
- Natural disaster shock used:
  - calibrated to mimic a once-in-100-years magnitude of flooding (as in DSA tailored stress test in IMF Country Report No. 22/381),
  - modelled to hit the economy in 2027,
  - EM-DAT indicates Rwanda was affected by disasters almost every year, with floods most prevalent.

### Simulation setup and financing assumptions
- Baseline: “no climate adaptation investment.”
- Scenario 1: “NDC full implementation scenario.”
  - Adaptation conditional overall: 2.1 percent of GDP per year over the next 5-year horizon (corresponds to cumulative USD 5.3 billion).
  - Financing mix: external concessional financing and grants in historical ratio 2:1.
  - External concessional: 1.4%
  - Grants: 0.7%
- Scenario 2: “Funding squeeze without policy adjustment.”
  - Concessional financing and grants limited to 1.1 percent of GDP (half of what was assumed under full NDC).
- Scenario 3: “Funding squeeze with policy adjustment.”
  - Lower external financing offset by domestic measures:
    - 0.5 percent of GDP increase in tax revenues (VAT),
    - 0.5 percent of GDP savings from spending cuts.
  - Under scenario 3 in table summary:
    - Adaptation conditional overall: 3.9%
    - External concessional: 0.8%
    - Grants: 2.1%
    - DRM: 0.5%
    - Spending rationalization: 0.5%
- Scenario 4: “Funding squeeze with policy adjustment and full RSF implementation.”
  - Full RSF implementation assumptions:
    - improves public investment efficiency by 20 percentage points through PIMA and green PFM reforms,
    - catalytic role in mobilizing additional private financing,
    - allows higher climate adaptation investment.
  - Scenario 4 summarized:
    - Adaptation conditional overall: 4.3%
    - External concessional: 0.8%
    - Grants / private: 2.5%
    - DRM: 0.5%
    - Spending rationalization: 0.5%
    - Public investment efficiency: +20bps (presented as 20 percentage points improvement)

### Simulation results — macroeconomic impacts and tradeoffs
- No climate adaptation (baseline) with disaster shock in 2027:
  - real GDP decline: 4 percent,
  - reconstruction implies a 4 percentage points increase in the public investment to GDP ratio,
  - debt-to-GDP ratio increases slightly by 2 percentage points,
  - fiscal deficit widens then corrected via model fiscal rule (reductions in expenditures and increase in consumption tax),
  - current account shows only minor deviation from steady-state.
- Scenario 1 — Full NDC implementation (upfront adaptation 2.1 percent of GDP per annum, financed by concessional loans and grants):
  - GDP drop after shock: 1 percentage points,
  - reconstruction investment needs: 1.5 percent of GDP,
  - debt-to-GDP ratio increases by 6 percentage points due to concessional borrowing,
  - fiscal deficit increases by 1.5 percentage points during the period of additional climate spending,
  - simultaneous widening of the current account deficit.
- Scenario 2 — Funding squeeze without policy adjustment:
  - partial NDC implementation,
  - larger adverse GDP impacts compared to scenario 1 (specific magnitudes not restated in text),
  - financing shortfall worsens output and sustainability outcomes.
- Scenario 3 — Funding squeeze with policy adjustment (0.5% VAT increase and 0.5% spending cuts):
  - GDP decline is only slightly more pronounced than scenario 1 and much smaller than scenario 2,
  - increase in debt, fiscal deficit, and current account deficit is more muted than in scenario 1,
  - lower debt financing and smaller current account deterioration help offset some adverse GDP impacts of tighter fiscal stance,
  - conclusion: full NDC implementation when financed by DRM and spending rationalization yields more favorable macroeconomic imbalances than heavy debt financing.
- Scenario 4 — Full RSF implementation on top of scenario 3:
  - higher public spending efficiency and mobilization of more private resources produce most positive macroeconomic outcomes,
  - more climate resilient investment reduces the natural disaster impact on economic activity,
  - higher public investment efficiency reduces required public investment for same outcome, limiting crowding out of private investment,
  - higher GDP in medium to long-run and unchanged debt levels lead to a lower debt-to-GDP ratio,
  - additional financing in the form of private financing keeps fiscal deficit and current account deficit similar to scenario 3.

### Policy findings and recommendations
- Investing in adaptation infrastructure mitigates GDP losses from natural disaster shocks but:
  - if financed primarily by concessional debt, adaptation investment will increase debt levels, fiscal deficits, and current account deficits.
- Fiscal and financing policy mix matters:
  - supplementing financing gaps with domestic revenue mobilization (DRM) and spending rationalization can limit output loss while safeguarding debt sustainability,
  - raising taxes (e.g., VAT) and cutting expenditures can finance full NDC implementation with smaller adverse macro impacts than debt financing.
- Improving public investment efficiency and mobilizing private financing (RSF reforms) yield the best macro outcomes:
  - increased public investment efficiency leads to more efficient construction and post-disaster reconstruction,
  - higher productivity of private capital and crowding in of private investment,
  - greater confidence may attract additional private climate financing.
- Recommended policy package:
  - increase fiscal space through DRM,
  - spending rationalization,
  - enhancement of public investment efficiency,
  - utilize programmatic synergies across development objectives to enhance returns on climate investment and lower pressure on development spending,
  - continue efforts to mobilize private climate investment.

### Key exact numeric values reported
- Adaptation conditional overall under full NDC: 2.1 percent of GDP per year (cumulative USD 5.3 billion referenced).
- Historical financing ratio (concessional loans to grants): 2:1.
- External concessional under Scenario 1: 1.4%
- Grants under Scenario 1: 0.7%
- Natural disaster modeled to hit economy in: 2027.
- No adaptation scenario GDP decline after shock: 4 percent.
- No adaptation scenario reconstruction public investment increase: 4 percentage points (public investment to GDP ratio).
- No adaptation scenario debt-to-GDP increase: 2 percentage points.
- Scenario 1 GDP drop after shock: 1 percentage points.
- Scenario 1 reconstruction investment needs: 1.5 percent of GDP.
- Scenario 1 debt-to-GDP increase (from adaptation financing): 6 percentage points.
- Scenario 1 fiscal deficit increase during additional climate spending: 1.5 percentage points.
- Scenario 2 concessional financing and grants limited to: 1.1 percent of GDP.
- Scenario 3 domestic policy offsets: 0.5 percent of GDP increase in VAT and 0.5 percent of GDP savings from spending cuts.
- Scenario 3 adaptation conditional overall (table): 3.9%
- Scenario 4 adaptation conditional overall (table): 4.3%
- Public investment efficiency improvement under full RSF implementation: 20 percentage points (presented as +20bps in one table context but described in text as 20 percentage points improvement).

*Source: Staff’s calculations and DIGNAD model simulations as presented in the section.*

### 2. The NBR is committed to its price stability objective and developed strategy in line

### 2. The NBR is committed to its price stability objective and developed strategy in line with best practices for inflation targeting regimes

### NBR’s inflation objective and strategy
- Inflation benchmark: range of 5±3 percent.
- Explicit objective in the Monetary Policy Strategy (MPS): “maintain price stability by keeping headline consumer price inflation within the band of 2 to 8 percent, with a focus of having it close to 5 percent in the medium term”.
- Policy instrument: “the NBR will set the Central Bank Rate (CBR) to stabilize inflation in the medium term” under a “forward looking” monetary policy framework.
- MPS published on the NBR’s website in February 2023 to improve communication.

### Monetary policy implementation toolkit
- Instruments used:
  - Open market operations (OMO) to satisfy banking system demand for liquidity and keep the 7-day interbank rate within the money market rates corridor of ±1 percent around the central bank rate.
  - Repo operations: mop up liquidity for periods from one to 7 days; reverse repos inject liquidity for the same period.
  - Central Bank Bills: mop up liquidity for 28 days, 91 days, 182 days, and 364 days.
  - Overnight deposits facility: invest excess liquidity overnight for remuneration.
  - Standing lending facility: overnight borrowing from the NBR.
  - Refinancing facility: borrowing facility for periods up to 30 days.
- Reserve requirements (RR):
  - Depository institutions obliged to hold minimum reserves at the central bank.
  - Base for RR calculation: all bank liabilities up to one year excluding interbank transactions.
  - Reserve averaging over a maintenance period allowed since March 2005; current maintenance period is two months.
  - Excess reserves are not remunerated.

### Foreign exchange intervention (FXI) practice
- The NBR sells USD to domestic banks in regular bilateral FXI under a de facto crawl-like exchange rate arrangement.
- Intervention mechanics:
  - Direct interventions with commercial banks at a preannounced intervention rate based on historical official exchange rate data.
  - FXI occurs almost weekly.
  - USD sale amount derived from the NBR’s available USD supply based on a macro framework and banks’ perceived USD needs based on self-reporting.
- Classification: IMF’s upcoming 2023 Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER) classifies Rwanda as having a de facto crawl-like arrangement; the de jure arrangement is floating.

### Frictions and impediments to effective monetary transmission in LLMICs and Rwanda
- General LLMIC frictions:
  - Weak financial market development reduces effectiveness of the asset channel.
  - Ineffective central bank communication and high financial illiteracy impede the expectations channel.
  - Result: policy rate changes transmit mainly via interest rate, credit, and exchange rate channels.
- LLMIC-specific structural factors weakening transmission:
  - Underdeveloped securities and interbank markets weaken policy rate transmission to market rates.
  - Weak institutional environment (poorly defined property rights, inefficient legal systems, weak creditor protection, weak accounting/disclosure) and limited pool of viable borrowers weaken bank lending channel.
  - Extensive central bank intervention in FX markets often weakens the exchange rate channel.
- Evidence for Rwanda:
  - MTM operates through interest rate, credit, and exchange rate channels.
  - NBR in-house research (annual report 2017–18): transmission from policy rate to money market rates has improved, but transmission from money market rates to lending rates not yet statistically evident — interest rate channel likely weak.
  - Kwizeraa and Ndarihoranye (2022): evidence of an operational bank lending channel in Rwanda.
  - Forthcoming NBR research using quarterly data 2006–2022: monetary policy transmission via exchange rate channel exists.

### Recent experience and challenges for Rwanda’s monetary policy
- Inflation dynamics:
  - Inflation has been quite volatile and more volatile than the SSA median.
  - Inflationary pressures persisted despite a 300-bps cumulative hike between February 2022 and August 2023.
  - Rwanda’s policy tightening since December 2021 is still small compared to other SSA countries given its large increase in inflation.
- Role of food in CPI:
  - High weight of food in the CPI basket makes inflation highly sensitive to weather-related supply shocks.
  - Rwandan agriculture mainly rain-fed; two consecutive years of low domestic agricultural production contributed to high food inflation since 2022, alongside higher international food, fuel, and fertilizer prices.
  - Monetary policy can play a role in containing second-round effects (e.g., non-food price increases via wage demands in construction).
- Bank concentration and lending-rate pass-through:
  - Average lending rate appears flat over January 2016–May 2023.
  - High bank concentration and recent bank mergers reduce pass-through from the policy rate to lending rates and weaken the credit channel.
  - Market dominated by a small number of large, well-capitalized borrowers with significant collateral; extending lending to marginal borrowers is costly and reduces interest rate pass-through.
  - More detailed analysis (e.g., using credit registry data) is warranted.
- Improvements and implementation challenges:
  - Transition to interest-rate based monetary policy in January 2019 improved framework: narrowed interest rate corridors, repo rate converged to the policy rate, and interbank rate steered within the corridor.
  - Operational QPM (quarterly projection macroeconomic model) forms the core of the central bank’s Forecasting and Policy Analysis Systems (FPAS).
  - Actual policy decisions have often deviated from core model recommendations, especially in the tightening cycle since February 2022.
  - Need to strengthen internal organization to better integrate model analysis in policymaking and MPC process.
- Exchange rate adjustment and FX market functioning:
  - Exchange rate adjustment increased somewhat beginning of 2023 but remains relatively small compared with other SSA countries.
  - Structural FX shortages due to persistent current account deficit; shortages intensified as donor financing declined.
  - Interbank FX trading activity has become more muted; NBR regularly intervenes to supply FX to banks.
  - FX interventions are one-sided (central bank selling FX to banks), indicating disequilibrium: FX demand by bank clients higher than supply.
  - Overly active central bank participation can create moral hazard, inhibit market risk awareness and risk management, and confuse participants about nominal anchors.
  - IMF TA on FX market diagnosis provided recommendations to allow more FX flexibility while addressing structural issues to improve FX market functioning.

### Policy implications and options highlighted
- Enhance a forward-looking monetary policy framework.
- Provide clear communication of data-driven monetary policy decisions.
- Allow greater exchange rate flexibility while addressing structural FX market issues.
- Strengthen internal organization of the NBR to better integrate model-based analysis into policy decisions and the MPC process.
- Conduct more detailed analysis of bank lending dynamics (for example, using credit registry data) to diagnose and address pass-through weaknesses.

*Source: IMF staff report (RWANDA).*

### Box 1. Structural Issues on the FX Market and Policies to Strengthen the Monetary Policy

### Box 1. Structural Issues on the FX Market and Policies to Strengthen the Monetary Policy Transmission Mechanism

### Structural issues in the FX market and implications
- Rwanda typically runs trade deficits while capital inflows are mainly not sensitive to interest rates.
- The share of remittances and foreign aid to GDP are high in Rwanda while foreign direct investment and portfolio inflows remain marginal.
- As the government converts its external concessional funding from USD to RWF at the central bank, the NBR has become an intermediator in the FX market.
- Dollarization of deposits reduces the impact of monetary policy transmission to deposit rates; a self-sustaining cycle of gradual depreciation expectations and USD hoarding resulted in relatively high deposit dollarization.
- An IMF TA team conducted an FX market diagnostic in September 2023 and suggested FX market-related reform options to help the exchange rate channel of monetary policy.
- The mission noted that a structural USD backlog is behind the demand-supply mismatch and USD shortage causing the steady Rwandan franc depreciation; policies should address those structural issues before the NBR could start reducing interventions.
- Monetary policy effectiveness is linked to FX market development: a stronger interest rate transmission channel may reduce the exchange rate’s importance as a tacit nominal anchor, and exchange rate flexibility is important to allow monetary policy to play an independent role.

### Empirical analysis of transmission channels (overall)
- Sample period used for VAR analysis: Q4 2010–Q1 2023.
- Assumed ordering of responses to a shock on the monetary aggregate (growth of M2): nominal ER depreciation rate, interbank rate, private credit growth, inflation, and GDP growth.
- VAR results:
  - Evidence of operational credit and interest rate channels: credit growth and the interbank rate increase in response to an increase in M2 growth.
  - Exchange rate response to the M2 shock is not statistically different from zero, indicating the exchange rate channel is weak.
- Caveats:
  - Structural breaks in policy frameworks and limited sample size (including the transition to an interest-rate based framework) potentially undermine precision of impulse responses.
  - Alternative specification replacing M2 with the central bank rate showed similar direction of responses but statistically insignificant results.
  - Selection between M3 and M2 had little impact on results.

### Interest rate / credit channel findings
- Granger causality analysis period: January 2009–August 2023 (monthly).
- Rates used:
  - Interbank rate: average 7-day interbank rates (operational target).
  - T-bill rate: 91-day T-bill rate selected for longest time series.
  - Deposit and lending rates: simple averages across maturities.
- Granger causality results:
  - Policy rate Granger-causes interbank rate (7-day) and lending rate (average) but not vice versa, suggesting the central bank rate successfully drives those market rates.
  - Weak policy rate pass-through to the T-bill rate and the deposit rate.
- Change after monetary policy framework transition (transition in January 2019):
  - Before transition: interest channel weaker; fail to reject hypothesis that central bank rate does not Granger-cause market rates for all rates except lending rate.
  - After transition: interest rate pass-through from central bank rate to interbank rate and T-bill rate seems to have strengthened despite shorter sample; central bank rate now affects lending rate though magnitude of lending rate adjustment remains limited.

### Exchange rate channel findings
- Exchange rate pass-through (ERPT) to inflation is strong in Rwanda.
- Evidence:
  - Kemoe et al. (Forthcoming) found that at a 12-month horizon, the cumulative response of inflation to a 1 percent depreciation in the bilateral exchange rate against the US dollar is around 0.4 percent for Rwanda.
  - Rwanda’s ERPT is almost double the Sub-Saharan Africa (SSA) average, despite a notable decline since 2017.
  - ERPT varies by exchange rate regime; countries with managed floats (including Rwanda) have higher ERPT on average than conventional pegs or fully floating regimes in SSA.
- Policy implication:
  - First-stage transmission (monetary policy shock → exchange rate) is weak due to a de facto crawl-like exchange rate regime, but second-stage transmission (exchange rate → inflation) is high, which amplifies inflationary effects of depreciation.
  - Strengthening monetary policy effectiveness limits second-round effects of exchange rate depreciation on domestic inflation.

### Recent NBR policy efforts to strengthen transmission and expectations
- Measures to improve communication, financial inclusion, and financial education to enhance the expectations channel and market functioning:
  - Development of a guideline for women’s financial inclusion for the financial sector.
  - Introduction of a mobile app "GERERANYA" for consumers to compare costs of financial services.
  - Implementation of initiatives and awareness programs: NBR quiz challenge, university monetary policy challenges, savings week, insurance week, global money week, and Financial Consumer Protection Law awareness campaign.
  - Quarterly meetings of the Chief Economist with bank treasurers, economic cluster and taskforce market analysis to discuss Monetary Policy outcomes.
  - Bond market development: during the first half of 2023, NBR issued three new bonds and reopened four existing ones; average subscription rate was 117.1 percent, up from 109.4 percent in the corresponding period year.

### Concluding remarks and policy recommendations
- Key frictions impairing monetary policy transmission in Rwanda:
  - Vulnerability to climate shocks.
  - Less developed financial markets.
  - Incomplete integration of forecasting and policy analysis into policymaking.
  - Structural shortage of foreign exchange.
  - continued need for exchange rate flexibility.
- Positive developments:
  - Monetary policy transmission mechanism has improved since the transition to an interest-rate based framework in 2019, with short-term rates broadly tracking the policy rate and the interest rate channel generally operational.
- Policy recommendations:
  - Use the policy rate more proactively; literature suggests interest-rate based frameworks yield more effective transmission than monetary aggregates.
  - Ensure monetary policy decisions are data driven and forward-looking; improve NBR forecasting and policy analysis capacities and fully incorporate the system into policymaking.
  - Allow more exchange rate flexibility; a more flexible exchange rate, coupled with more effective monetary policy, would shield domestic prices from exchange rate depreciation and help maintain price stability.

*Source: Box 1. Structural Issues on the FX Market and Policies to Strengthen the Monetary Policy Transmission Mechanism, IMF staff.*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2023/english/1rwaea2023004.pdf_