## 1stpea2025003-source-pdf

## Source details

**Canonical URL:** [1stpea2025003-source-pdf](https://www.imf.org/-/media/files/publications/cr/2025/english/1stpea2025003-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2025/english/1stpea2025003-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2025/english/1stpea2025003-source-pdf.pdf.json)

---

### Inflation Developments — structure and recent trajectory
- CPI composition and weights:
  - CPI basket: 12 subindexes, 79 categories (updated in 2015).
  - Food: approximately 70 percent of the CPI basket.
  - Locally produced food: nearly half of food items in the CPI.
  - Fuel and electricity: around 6 percent of the CPI basket.
  - Core inflation (excludes food, fuel, electricity): about 24 percent of the CPI basket.
- Exchange rate regime and historical inflation:
  - Flexible regime (1999–2009): average inflation 16.4 percent; standard deviation 6.5 percent.
  - Peg regime (2010–2020): average inflation 8.4 percent; standard deviation 2.5 percent.
  - Under the peg, inflation fell from 16.1 percent in 2009 to approximately 3.9 percent in 2015.
- Recent inflation timeline:
  - Inflation averaged about 8 percent annually from 2017 to 2021, driven primarily by rising food prices.
  - By 2023, inflation surged to over 25 percent due to a global inflationary shock and imported food price increases, followed by substantial increases in locally produced food.
  - Core inflation rose from around 2 percent in 2021 to 15.9 percent by end-2023.
  - By December 2024:
    - Headline inflation: 11.6 percent year-over-year.
    - Core inflation: 3.0 percent.
  - Drivers of the December 2024 decline: unwinding of the 2023 food price spike, fiscal consolidation, and monetary tightening.
- VAT interaction:
  - VAT introduced in June 2023; estimated to have contributed about 5 percentage points to price increases in 2023 (see Box 1).

### Box 1 — Estimated VAT impact on inflation
- VAT design:
  - Main rate: 15 percent.
  - Reduced rate: 7.5 percent for certain products/services.
  - Some products/services fully exempt.
- Estimates:
  - Mechanical complete-coverage estimate: CPI inflation would have risen by 5 percentage points in June 2023 if coverage were complete and impact immediate.
  - Difference-in-differences empirical estimate:
    - June 2023 immediate association: inflation increase of 2.7 percentage points.
    - June–December 2023 (assuming broad VAT coverage by end-2023): estimated VAT impact on 2023 inflation = 4.9 percentage points.
  - Cross-country panel evidence (11 Sub-Saharan African countries that implemented VAT since 1990, excluding São Tomé and Príncipe):
    - VAT implementation dummy coefficient = 3.17 (P value 0.00).
    - Lagged inflation coefficient (Inflation (t-1)) = 0.52 (P value 0.00).
    - Constant c = 1.9 (P value 0.00).
    - R2 = 0.40; Adjusted R2 = 0.33; Observations = 362.

### External competitiveness and pass-through
- REER and NEER:
  - From 2010 to 2023, REER experienced substantial appreciation, reflecting persistent high inflation and the pegged exchange rate, signaling a decline in competitiveness.
  - NEER rose by 20 percent during the same period, driven largely by the Euro’s strengthening against the US dollar.
- NEER passthrough in comparable island microstates (Box 2):
  - Sample: seven micro-islands with fixed exchange rates, 2001–2023.
  - NEER (t) coefficient = -0.33 (P value 0.00): a 10 percent appreciation/depreciation of the NEER is associated with a decline/increase in annual inflation by 3.3 percentage points.
  - NEER (t-1) coefficient = -0.20 (P value 0.10).
  - Inflation (t-1) coefficient = 0.24 (P value 0.00).
  - Global crises coefficient = 2.10 (P value 0.02).
  - Model fit: R2 = 0.36; Adjusted R2 = 0.31; Observations = 140; Number of countries = 7; country fixed effects included.
  - Interpretation: passthrough occurs largely within the same year; inflation inertia and global crises are significant.

### Key findings on inflation determinants
- Drivers identified:
  - Supply-side: inflation inertia, import food prices, global/regional crises.
  - Demand-side: weakened macroeconomic discipline in lead-up to elections.
- Exchange rate regime effect:
  - Peg to the euro associated with lower average inflation (8.4 percent vs. 16.4 percent in flexible period) and lower volatility (standard deviation 2.5 percent vs. 6.5 percent).
- Recent shocks and tax policy:
  - Global food and energy shocks pushed inflation over 25 percent by 2023.
  - VAT introduction in June 2023 contributed an estimated 4.9 percentage points to 2023 inflation.
- External pass-through magnitude:
  - 10 percent NEER appreciation/depreciation → 3.3 percentage point decline/increase in annual inflation in comparable island microstates.

### Policy implications on inflation
- Comprehensive policy mix required to manage inflation:
  - Manage supply shocks: monitor and mitigate import food price volatility and strengthen food supply resilience given food’s ~70 percent CPI weight.
  - Maintain macroeconomic discipline, particularly around election cycles.
  - Monetary–fiscal coordination: use forecasting tools to time tightening or consolidation to counteract inflation inertia and respond to external shocks.
  - Monitor NEER and external currency dynamics given significant passthrough (-0.33 NEER coefficient).
  - Anticipate near-term inflationary effects when introducing broad-based indirect taxes (example: estimated 4.9 percentage point impact in 2023).

---

### Model specification, estimation, and diagnostics
- Framework:
  - Phillips Curve within an Autoregressive Distributed Lag Model.
  - Inflation equation (as specified):
    - 휋 = c + α 휋_{t−1} + β1 Portugal_food_inflation_t + β2 Portugal_food_inflation_{t−1} + δ GDP_growth_{t−1} + θ fuel_change_t + φ D1 + γ D2 + ω D3 + μ
  - Definitions:
    - 휋: end-of-period inflation rate in São Tomé and Príncipe.
    - Portugal_food_inflation: end-of-period food inflation rate in Portugal.
    - GDP_growth: real GDP growth in São Tomé and Príncipe.
    - fuel: percentage change in domestic fuel prices.
    - D1: election dummy.
    - D2: global crises dummy (2007–2008, 2020, 2022).
    - D3: pegged exchange rate regime dummy.
- Data and sample:
  - Annual data from 1998 to 2023.
  - Trade context: imports from Portugal average 49 percent of total imports over the past 10 years; European countries accounted for 55 percent of total imports.
  - Stationarity tested with Augmented Dickey-Fuller and Phillips-Perron tests; lags by AIC.

### Estimation results (core coefficients and fit)
- Core coefficients and p-values:
  - Lagged end-of-period inflation (휋_{t−1}): Coeff. 0.485; P value 0.0001
  - End-of-period Portugal food inflation: Coeff. 0.761; P value 0.0005
  - Lagged end-of-period Portugal food inflation: Coeff. 0.167; P value 0.2433
  - Lagged Real GDP growth: Coeff. 0.454; P value 0.0490
  - STP fuel price change: Coeff. 0.063; P value 1.1766
  - Dummy – elections (D1): Coeff. 2.365; P value 0.0480
  - Dummy – global crises (D2): Coeff. 3.172; P value 0.0317
  - Dummy – pegged regime (D3): Coeff. -4.014; P value 0.0176
  - Constant (c): 3.766; P value 0.1125
- Model fit:
  - R^2: 0.964835
  - Adjusted R^2: 0.941392
  - Observations: 21

### Diagnostics and robustness
- Residual tests:
  - Breusch–Godfrey Serial Correlation LM Test (up to 2 lags):
    - F-statistic Prob. F(2,10): 0.8262
    - Obs*R-squared Prob. Chi-Square(2): 0.6747
  - Heteroskedasticity Test: Breusch-Pagan-Godfrey:
    - F-statistic Prob. F(8,12): 0.4022
    - Obs*R-squared Prob. Chi-Square(8): 0.3351
    - Scaled explained SS Prob. Chi-Square(8): 0.9699
  - Normality Test: Jarque-Bera: 0.331581; Probability: 0.847224
  - Conclusion: no serial correlation; homoskedastic errors; residuals normally distributed.
- Multicollinearity (VIFs and variances):
  - Lagged inflation: Uncentered VIF 11.88; Centered VIF 2.68
  - Portugal food inflation (t): Uncentered VIF 3.58; Centered VIF 2.58
  - Portugal food inflation (t−1): Uncentered VIF 2.59; Centered VIF 1.87
  - Lagged GDP growth: Uncentered VIF 7.08; Centered VIF 1.50
  - ICRISIS: Uncentered VIF 2.29; Centered VIF 1.85
  - ELECTION: Uncentered VIF 1.94; Centered VIF 1.47
  - PEG: Uncentered VIF 10.00; Centered VIF 3.33
  - FUEL: Uncentered VIF 3.60; Centered VIF 2.20
  - C: Uncentered VIF 34.06; Centered VIF NA
- Variables excluded due to poor performance: government expenditure, nominal effective exchange rate, international reserves, BCSTP credit to government (all statistically insignificant and violated residual diagnostics if included).

### Forecasting performance and projected path
- Forecast evaluation metrics (Forecast sample: 1998-2029; Adjusted sample: 2003-2029; Included observations: 27):
  - Root Mean Squared Error: 1.538
  - Mean Absolute Error: 1.258
  - Mean Absolute Percentage Error: 14.810
  - Theil Inequality Coef.: 0.052
  - Bias Proportion: 0.000
  - Variance Proportion: 0.094
  - Covariance Proportion: 0.905
  - Theil U2 Coefficient: 0.405
  - Symmetric MAPE: 13.631
- Pseudo out-of-sample forecasts: reported as reasonably good with low bias and variance.
- Model projections (assuming all other factors remain constant):
  - End-of-period inflation expected to decline to 7.7 percent in 2025.
  - Anticipated to rise to 9.1 percent in 2026 due to national elections.
  - Gradually decline to 6.5 percent by 2029.

---

### Banking sector: structure, liquidity, and monetary instruments
- Sector concentration and systemic risk:
  - Largest bank controls 65 percent of total assets.
  - High concentration elevates systemic risk and hinders competition.
- Private sector credit and lending constraints:
  - Private sector credit declined to less than 8 percent of GDP after nearly five years of contraction, down from 24 percent pre-pandemic.
  - High interest rates close to 20 percent; limited financial infrastructure, weak contract enforcement, absence of robust credit information system, and widespread informality deter lending.
- Non-performing loans and regulatory response:
  - NPLs reduced to 10 percent of total loans by Q1:2025 from a pandemic peak of approximately 34 percent in Q3:2020 following 2022 regulatory changes.
- BCSTP instruments and settings:
  - OMOs via BCSTP Certificates of Deposit (CDs); CDs maturities 1 to 360 days; since July 2022 issuance shifted to fixed-quantity, variable-rate tenders.
  - Required reserves (RR): unremunerated reserves set at 28 percent for local currency liabilities (increase from 18 percent prior to mid-2022); 21 percent for foreign currency liabilities (composed of 90 percent in local currency and 10 percent in foreign currency).
  - Standing facilities: policy reference rate 10 percent; lending facility 10.5 percent; deposit facility 0 percent; facilities limited to five days and recently unused.
  - Collateral: total collateral must be at least 125 percent of requested loan amount; eligible instruments accepted only if their value exceeds the bank's reserve requirement at the BCSTP.
- New monetary policy framework and MPC:
  - Core inflation target: six-month moving average at 3.1 percent.
  - Threshold for banks’ excess reserves: STN 275 million (previously STN 200 million).
  - Commitment to introduce 12-month CDs; retain 1-month, 3-month, and 6-month options.
  - MPC convenes quarterly; composition includes governor, deputy, administrators, directors, and three senior advisors.
- Interbank money market and liquidity forecasting:
  - IMM inactive due to insufficient trust and underdeveloped infrastructure; eligible collateral includes foreign currency deposits, T-bills, BCSTP CDs.
  - Liquidity forecasting framework initiated in 2011; constrained by data and resource limitations.
- Recent liquidity and interest rate developments:
  - Excess reserves peaked at STN 783 million by early 2017; fell to STN 494 million by end-2019; increased 2020–mid-2022 due to external support.
  - Decline in excess liquidity since mid-2022 driven by fiscal consolidation, PFM improvements, VAT introduction, BCSTP tighter liquidity management, increased reserve requirements, and cessation of BCSTP credit to government.
  - Weighted average CD interest rates: 1.3 percent in July 2022 → 4.6 percent by December 2024.
  - T-bill rates: approximately 4 percent in mid-2022 → 6.2 percent by April 2024.
  - MEF issued two-year T-bonds offering 4 to 6 percent.
  - Short-term deposit rates: 2.3 percent in mid-2022 → 2.6 percent by December 2024.
  - Six-month moving average of core inflation: approximately 12 percent mid-2022–end-2023 → 3.3 percent by December 2024.
- Early 2025:
  - Excess reserves increased in Q4 2024 and early 2025 despite sizable CD issuance; stock of CDs rose to STN 335 million by mid-February 2025—50 percent higher than excess reserves in January 2025.
  - Increase attributed to foreign exchange inflows from external disbursements and BCSTP regulation requiring banks to sell 25 percent of foreign exchange revenues to the BCSTP.
  - Excess liquidity unevenly distributed; one commercial bank holds around two-thirds of total liquidity surplus.

### Monetary policy recommendations for BCSTP and markets
- Open Market Operations:
  - Introduce 12-month CD options and use T-bills for monetary policy to absorb structural excess liquidity and reach STN 275 million excess reserves target.
  - Offer limited CDs once excess reserves dip below target or core inflation falls below 3.1 percent.
  - Continue one-month CD auctions or trial 7/14-day maturities to assess interest rate signals.
  - Promote transparency, diversify counterparties, leverage electronic trading, and issue longer-term instruments.
- Required Reserves:
  - Recalculate RR base using the average of the last month’s data.
  - Consider remunerating RR at or near the targeted policy rate to reduce spread between deposit and lending rates if correction of distortions is intended; otherwise maintain unremunerated RR to drain liquidity cheaply.
- Standing facilities and ELA:
  - Make standing lending facility permanently available; lift access restrictions; set lending rate above BCSTP CD rate to act as penalty and ceiling for interbank rates.
  - Operationalize standing deposit facility linked to BCSTP CD rate; ensure deposit rate remains below CD rate to create a corridor.
  - Clarify ELA framework categories and standardize rates for transparency.
- Interbank Money Market:
  - Strengthen legal/regulatory framework, invest in electronic trading and settlement, make CDs tradable, and establish robust collateral framework.
- Liquidity forecasting:
  - Initiate one-month liquidity forecasting, expand to one year; improve data collection and coordination between MEF and BCSTP; invest in IT and human resources.
- BCSTP recapitalization, communication, capacity:
  - Recapitalize BCSTP via government capital injection or government bonds and retain higher profits.
  - Enhance communication with financial institutions, government, and public.
  - Strengthen analytical tools and staff capacity; IMF available for technical assistance.

---

### Climate resilience: DIGNAD model, scenarios, and outcomes
- NDC and climate context:
  - 2021 NDC target: reduce GHG emissions by 27 percent by 2030 vs business-as-usual.
  - Estimated total cost of 2021 NDC: around US$150 million.
  - Energy sector largest emitter through 2018; “Forest and other land uses” sector acts as a CO2 sequester; on balance a net carbon sink through 2018.
- DIGNAD model purpose:
  - Debt, Investment, Growth, and Natural Disasters dynamic general equilibrium model to analyze impacts of investing in climate-resilient capital on growth and debt sustainability.
  - Five disaster channels: damages to public capital; damages to private capital; temporary productivity loss; reduced public investment efficiency; loss in creditworthiness.
- Calibration and initial values (selected):
  - Grants to GDP ratio: 7.8
  - Domestic public debt to GDP ratio: 42.8
  - Concessional public debt to GDP ratio: 59.2
  - External commercial public debt to GDP ratio: 12.3
  - Return to standard infrastructure: 30.0
  - Return to climate-resilient infrastructure: 40.0
  - Public standard infrastructure investment to GDP ratio: 6.0
  - Depreciation rate of standard public infrastructure: 7.5
  - Depreciation rate of resilient public infrastructure: 3.0
  - Public investment efficiency: 50.0
  - Initial calibration aligns steady state with projected values for 2025-28 (about 6 percent of GDP public capital infrastructure investment).
- Simulated disaster and scenario sets:
  - Baseline starting point: state in 2024.
  - Simulated disaster: 2029 one-time flooding with a one-off 6 percent point GDP loss.
  - Simulation set 1 (role and scale of adaptation investment):
    - 1.1 Baseline: no resilient infrastructure pre-disaster; recovery financed equally by domestic debt and external commercial borrowing; rebuild to pre-disaster level in ~5 years.
    - 1.2 Resilient infrastructure: invest 2 percent of GDP per year in 2025-28 in climate-resilient public infrastructure; financing equally split between grants and concessional loans.
    - 1.3 Resilience scaling-up: invest 4 percent of GDP per year in 2025-28 in resilient infrastructure; financed with ¾ grants and ¼ concessional loans.
  - Simulation set 2 (resilient vs standard infrastructure and PIM):
    - 2.1 Standard infrastructure: 2 percent of GDP per year in 2025-28, financed equally by grants and concessional loans.
    - 2.2 Resilient infrastructure: same as 1.2.
    - 2.3 Resilient + PIM reform: 2.2 plus increase in public investment efficiency by 10 percentage points.
  - Simulation set 3 (ex-post vs ex-ante financial preparedness):
    - 3.1 Ex-post disaster management: disaster in 2029; recovery spending 2 percent of GDP per year in 2029-2032, financed equally by grants and concessional loans.
    - 3.2 Ex-ante contingency fund: invest 2 percent of GDP per year in 2025-28 into contingency fund, funding split equally between grants and concessional loans; drawdown 2 percent of GDP per year in 2029-32.
    - 3.3 Resilient infrastructure: same resources invested ex-ante in climate-resilient infrastructure (same as 1.2).

### Simulation outcomes and policy implications
- Mechanisms and quantitative outcomes:
  - Pre-disaster resilient investment crowds in private investment, raises output, and has larger effects than standard infrastructure due to higher returns and slower depreciation.
  - Natural disaster reduces output via permanent public/private capital damage and temporary productivity loss; increases fiscal deficits from reconstruction costs.
  - Resilient infrastructure reduces damages and repair costs, containing debt increases and enabling quicker debt stabilization.
  - Higher public investment efficiency further improves recovery needs and debt dynamics.
- Quantitative highlights:
  - GDP decline when disaster hits:
    - Baseline scenario: 6.3 percentage points relative to the steady state.
    - Standard infrastructure scenario: about 5 percentage points.
    - Resilient infrastructure scenario: 2.3 percentage points.
  - Private investment: more resilient under resilient infrastructure scenarios.
  - Public debt: lower in long term under resilient infrastructure compared to baseline and standard scenarios; further improved with higher public investment efficiency.
  - Private consumption: decline post-shock is less pronounced under resilient infrastructure; resilient capital supports swifter recovery and rebound in private consumption that can exceed pre-disaster levels.
- Comparative assessment (Simulation 3 / Figure 5 conclusions):
  - Timing of financial inflows (ex-ante vs ex-post) does not change macro outcomes unless used to augment real capital/infrastructure.
  - Ex-ante contingency funds and ex-post grants/concessional loans yield identical macro outcomes if used only for financing.
  - Investing same resources in resilient infrastructure produces superior GDP, private investment, and private consumption outcomes short- and long-run.
  - Resilient infrastructure lowers disaster losses, repair costs, and long-term public debt.

### Policy priorities from DIGNAD simulations
- Secure grants and concessional financing for resilient infrastructure and adaptation plans.
- Strengthen public investment management to raise public investment efficiency.
- Accelerate preparation of a National Adaptation Plan.
- Invest in resilient infrastructure, early warning systems, and climate/environmental protection.
- Favor ex-ante investment in resilient physical capital over equal-sized contingency funds when aiming to maximize macroeconomic resilience.

---

*Source: IMF staff analysis from “1. Inflation Developments” (July 16, 2025).*

### 1. Inflation Developments  ___________________________________________________________ 6

### 1. Inflation Developments

### Introduction
- Paper objective: examine drivers of inflation in São Tomé and Príncipe and develop a forecasting model using the Phillips Curve approach within the Autoregressive Distributed Lag Model, integrating recent global market developments.
- Prepared by: Alsis Cruz and Dilson Tiny, led by Bahrom Shukurov.
- Key contextual points:
  - Inflation substantially declined under the euro peg established in 2010.
  - Recent global food and energy price surges pushed inflation into double digits and created a notable inflation differential with the euro area, straining the peg.

### Inflation Structure and Recent Developments
- CPI composition:
  - CPI basket: 12 subindexes, 79 categories (updated in 2015).
  - Food: approximately 70 percent of the CPI basket.
  - Locally produced food: nearly half of food items in the CPI.
  - Fuel and electricity: around 6 percent of the CPI basket.
  - Core inflation (excludes food, fuel, electricity): about 24 percent of the CPI basket.
- Exchange rate regime effects:
  - Flexible regime (1999–2009): average inflation 16.4 percent; standard deviation 6.5 percent.
  - Peg regime (2010–2020): average inflation 8.4 percent; standard deviation 2.5 percent.
  - Under the peg, inflation fell from 16.1 percent in 2009 to approximately 3.9 percent in 2015.
- Recent inflation trajectory:
  - Inflation averaged about 8 percent annually from 2017 to 2021, driven primarily by rising food prices.
  - By 2023, inflation surged to over 25 percent, driven by global inflationary shock and imported food price increases, followed by substantial increases in locally produced food.
  - Core inflation rose from around 2 percent in 2021 to 15.9 percent by end-2023.
  - By December 2024:
    - Headline inflation fell to 11.6 percent year-over-year.
    - Core inflation decreased to 3.0 percent.
  - Drivers of the December 2024 decline: unwinding of the 2023 food price spike, fiscal consolidation, and monetary tightening.
- VAT introduction interaction:
  - VAT introduced in June 2023; estimated to have contributed about 5 percentage points to price increases in 2023 (see Box 1).

### Box 1 — Estimating the VAT Impact on Inflation
- VAT design:
  - Main rate: 15 percent.
  - Reduced rate: 7.5 percent for certain products/services.
  - Some products/services fully exempt.
- Mechanical complete-coverage estimate:
  - If coverage were complete and impact immediate: CPI inflation would have risen by 5 percentage points in June 2023.
- Difference-in-differences empirical estimate:
  - Method: compare São Tomé and Príncipe 2023 monthly variations to 2015–2022 average and use Cabo Verde as counterfactual.
  - June 2023 immediate association: inflation increase of 2.7 percentage points.
  - June–December 2023 (assuming broad VAT coverage by end-2023): estimated VAT impact on 2023 inflation = 4.9 percentage points.
- Cross-country panel evidence (11 Sub-Saharan African countries that implemented VAT since 1990, excluding São Tomé and Príncipe):
  - Panel regression result: VAT implementation dummy coefficient = 3.17 (P value 0.00).
  - Other panel model estimates:
    - Lagged inflation coefficient (Inflation (t-1)) = 0.52 (P value 0.00).
    - Constant c = 1.9 (P value 0.00).
    - R2 = 0.40; Adjusted R2 = 0.33; Observations = 362.

### REER and NEER Developments
- Trends and drivers:
  - From 2010 to 2023, REER experienced substantial appreciation, reflecting persistent high inflation and the pegged exchange rate, signaling a decline in competitiveness.
  - NEER rose by 20 percent during the same period, driven largely by the Euro’s strengthening against the US dollar.
- Figure reference: REER and NEER series show appreciation from 2000–2023 (visuals in source).

### Box 2 — NEER Passthrough on Inflation in Island Microstates
- Sample and period:
  - Seven micro-islands with fixed exchange rate regimes, 2001–2023 (Antigua and Barbuda; Barbados; Dominica; Grenada; St. Kitts and Nevis; Saint Lucia; Saint Vincent and the Grenadines).
- Model specification:
  - OLS panel regression of annual average inflation on lagged inflation, annual NEER change (current and lagged), and global crises dummy; k=1 based on Akaike and Schwarz criteria.
- Key empirical findings:
  - NEER (t) coefficient = -0.33 (P value 0.00).
    - Interpretation: a 10 percent appreciation/depreciation of the NEER is associated with a decline/increase in annual inflation by 3.3 percentage points.
  - NEER (t-1) coefficient = -0.20 (P value 0.10) — negative but not statistically significant at conventional levels.
  - Inflation (t-1) coefficient = 0.24 (P value 0.00) — evidence of inflation inertia.
  - Global crises coefficient = 2.10 (P value 0.02) — global crises significantly raise inflation.
  - Constant = 1.63 (P value 0.00).
  - Model fit: R2 = 0.36; Adjusted R2 = 0.31; Observations = 140; Number of countries = 7; country fixed effects included.
- Interpretation:
  - Passthrough occurs largely within the same year.
  - Inflation inertia and global crises are significant contributors to inflation dynamics in micro-islands.

### Methodology and Data (summary of approach)
- Core modeling approach: Phillips Curve within Autoregressive Distributed Lag framework for forecasting end-of-period inflation (2003–2029 covered in broader analysis).
- Controls and external factors: import food prices, global/regional crises, VAT introduction, NEER changes, election-related demand pressures.
- Model evaluation and robustness:
  - The paper includes model results, residual evaluation, multicollinearity evaluation, and pseudo out-of-sample forecasts (tables and diagnostics referenced in the source).

### Key Findings (enumerated)
- Both supply and demand factors drive inflation in São Tomé and Príncipe:
  - Supply-side: inflation inertia, import food prices, global/regional crises.
  - Demand-side: weakened macroeconomic discipline in lead-up to elections.
- Exchange rate regime matters:
  - Peg to the euro associated with lower average inflation (8.4 percent vs. 16.4 percent in flexible period) and lower volatility (standard deviation 2.5 percent vs. 6.5 percent).
- Recent shocks:
  - Global food and energy price shocks contributed to inflation surging to over 25 percent by 2023.
  - VAT introduction in June 2023 contributed an estimated 4.9 percentage points to 2023 inflation.
- External pass-through:
  - A 10 percent NEER appreciation/depreciation is associated with a 3.3 percentage point decline/increase in annual inflation in comparable island microstates.

### Policy Implications and Recommendations (derived from findings)
- Comprehensive policy mix required to reduce and manage inflation, addressing both supply and demand:
  - Manage supply shocks: monitor and mitigate import food price volatility and strengthen food supply resilience given food’s ~70 percent weight in the CPI.
  - Maintain macroeconomic discipline, particularly around election cycles, to limit demand-driven inflation pressures.
  - Monetary and fiscal policy coordination: use forecasting tools to time tightening or consolidation to counteract inflation inertia and respond to external shocks.
  - Consider exchange-rate-related externalities: given significant NEER passthrough (-0.33 NEER coefficient), monitor NEER movements and external currency dynamics (Euro–USD moves) for inflation implications.
  - VAT introduction impacts: when implementing broad-based indirect taxes, anticipate near-term inflationary effects (example: estimated 4.9 percentage point impact in 2023) and plan mitigating measures.

*Source: IMF staff analysis from “1. Inflation Developments” (July 16, 2025).*

### 11.      The model captures inflation by accounting for inflation expectations, as well as

### 1stpea2025003-source-pdf - 11.      The model captures inflation by accounting for inflation expectations, as well as

### Model specification and key drivers
- Model framework:
  - Phillips Curve approach based on the Autoregressive Distributed Lag Model (methodology similar to Lanau, Robles, and Toscani (2018)) with enhancements.
  - Inflation equation:
    - 휋 = c + α 휋_{t−1} + β1 Portugal_food_inflation_t + β2 Portugal_food_inflation_{t−1} + δ GDP_growth_{t−1} + θ fuel_change_t + φ D1 + γ D2 + ω D3 + μ
  - Definitions:
    - 휋: end-of-period inflation rate in São Tomé and Príncipe.
    - Portugal_food_inflation: end-of-period food inflation rate in Portugal.
    - GDP_growth: real GDP growth in São Tomé and Príncipe.
    - fuel: percentage change in domestic fuel prices.
    - D1: dummy variable for elections in São Tomé and Príncipe.
    - D2: dummy variable for structural breaks associated with global crises (2007–2008, 2020, 2022).
    - D3: dummy variable for the introduction of the pegged exchange rate regime.

- Key modeled drivers of inflation:
  - Inflation inertia/expectations (adaptive, backward-looking: past inflation used as proxy).
  - Demand-side shocks (lagged real GDP growth; election dummy capturing looser macroeconomic discipline).
  - Supply-side shocks (Portugal food inflation as proxy for imported inflation given heavy reliance on imported food and peg to the Euro).
  - Domestic fuel price adjustments.
  - Crisis dummy for 2007–2008, COVID-19 (2020), and Russia–Ukraine invasion (2022).
  - Dummy for pegged exchange rate regime.

### Data, tests, and sample
- Sample and variables:
  - Annual data from 1998 to 2023.
  - Variables: end-of-period inflation and real GDP growth in São Tomé and Príncipe; end-of-period food inflation in Portugal; changes in fuel prices in São Tomé and Príncipe; election and crisis/pegged regime dummies.
- Trade context:
  - Over the past 10 years, imports from Portugal have represented an average of 49 percent of total imports to São Tomé and Príncipe; European countries collectively accounted for 55 percent of total imports.
- Pre-estimation checks:
  - Stationarity tested with Augmented Dickey-Fuller and Phillips-Perron tests.
  - Number of lags chosen using Akaike Information Criterion (AIC).

### Estimation results (model coefficients and significance)
- Core coefficients and p-values (Table 1):
  - Lagged end-of-period inflation (휋_{t−1}): Coeff. 0.485; P value 0.0001
  - End-of-period Portugal food inflation: Coeff. 0.761; P value 0.0005
  - Lagged end-of-period Portugal food inflation: Coeff. 0.167; P value 0.2433
  - Lagged Real GDP growth: Coeff. 0.454; P value 0.0490
  - STP fuel price change: Coeff. 0.063; P value 1.1766
  - Dummy – elections (D1): Coeff. 2.365; P value 0.0480
  - Dummy – global crises (D2): Coeff. 3.172; P value 0.0317
  - Dummy – pegged regime (D3): Coeff. -4.014; P value 0.0176
  - Constant (c): 3.766; P value 0.1125
- Model fit and observations:
  - R^2: 0.964835
  - Adjusted R^2: 0.941392
  - Observations: 21

- Interpretations of key coefficients (as reported):
  - A one-percentage point increase in lagged end-of-period inflation raises end-of-period inflation by 0.49-percentage point.
  - A one-percentage point increase in Portugal's food inflation raises end-of-period inflation in São Tomé and Príncipe by 0.76-percentage point; the one-lagged Portugal food inflation coefficient (0.17) is not statistically significant.
  - National elections increase end-of-period inflation by 2.37-percentage points in the election year.
  - A global crisis increases end-of-period inflation by 3.17 percentage points.
  - A one-percentage point increase in lagged GDP growth raises end-of-period inflation by 0.45-percentage point.
  - Implementation of the pegged regime contributed to a -4.01-percentage point slowdown in end-of-period inflation.
  - A one-percentage point increase in domestic fuel price raises end-of-period inflation by 0.06-percentage point (coefficient not statistically significant but retained to capture supply-side shocks).

### Model diagnostics, multicollinearity, and robustness
- Residual evaluation (Table 2):
  - Breusch–Godfrey Serial Correlation LM Test (up to 2 lags):
    - F-statistic Prob. F(2,10): 0.8262
    - Obs*R-squared Prob. Chi-Square(2): 0.6747
  - Heteroskedasticity Test: Breusch-Pagan-Godfrey:
    - F-statistic Prob. F(8,12): 0.4022
    - Obs*R-squared Prob. Chi-Square(8): 0.3351
    - Scaled explained SS Prob. Chi-Square(8): 0.9699
  - Normality Test: Jarque-Bera: 0.331581; Probability: 0.847224
  - Conclusion: No serial correlation; homoskedastic errors; residuals normally distributed.

- Multicollinearity (Table 3, VIFs and variances):
  - Lagged inflation: Coefficient Variance 0.01; Uncentered VIF 11.88; Centered VIF 2.68
  - Portugal food inflation (t): Coefficient Variance 0.03; Uncentered VIF 3.58; Centered VIF 2.58
  - Portugal food inflation (t−1): Coefficient Variance 0.02; Uncentered VIF 2.59; Centered VIF 1.87
  - Lagged GDP growth: Coefficient Variance 0.04; Uncentered VIF 7.08; Centered VIF 1.50
  - ICRISIS: Coefficient Variance 1.70; Uncentered VIF 2.29; Centered VIF 1.85
  - ELECTION: Coefficient Variance 1.15; Uncentered VIF 1.94; Centered VIF 1.47
  - PEG: Coefficient Variance 2.13; Uncentered VIF 10.00; Centered VIF 3.33
  - FUEL: Coefficient Variance 0.00; Uncentered VIF 3.60; Centered VIF 2.20
  - C: Coefficient Variance 4.84; Uncentered VIF 34.06; Centered VIF NA

- Variables tested but excluded due to poor performance or diagnostics:
  - Government expenditure (negative coefficient, statistically insignificant).
  - Nominal effective exchange rate (positive coefficients contradictory to theory, statistically insignificant).
  - International reserves and BCSTP credit to the government (positive but statistically insignificant).
  - Inclusion of these variables would violate residual diagnostic tests for serial correlation, heteroskedasticity, and normality.

### Forecasting performance and projected inflation path
- Forecast evaluation metrics (Forecast sample: 1998-2029; Adjusted sample: 2003-2029; Included observations: 27):
  - Root Mean Squared Error: 1.538
  - Mean Absolute Error: 1.258
  - Mean Absolute Percentage Error: 14.810
  - Theil Inequality Coef.: 0.052
  - Bias Proportion: 0.000
  - Variance Proportion: 0.094
  - Covariance Proportion: 0.905
  - Theil U2 Coefficient: 0.405
  - Symmetric MAPE: 13.631
- Pseudo out-of-sample forecast performance: reported as reasonably good with low bias and variance.
- Model projections (assuming all other factors remain constant):
  - End-of-period inflation expected to decline to 7.7 percent in 2025.
  - Anticipated to rise to 9.1 percent in 2026 due to national elections.
  - Gradually decline to 6.5 percent by 2029.

### Conclusions on inflation drivers
- Empirical findings:
  - Both supply and demand factors significantly drive inflation in São Tomé and Príncipe.
  - Identified key influencers:
    - Inflation inertia (past inflation as proxy for expectations).
    - Import food prices proxied by Portugal's food inflation.
    - Global and regional crises causing supply chain disruptions and commodity price increases.
    - Weakened macroeconomic discipline during domestic elections.
  - Despite stability provided by the pegged exchange rate regime, inflation has risen substantially in recent years.

### Policy recommendations
- Overall approach:
  - Implement a comprehensive policy mix addressing both demand and supply factors to stabilize prices and reduce inflationary pressures.
- Demand-side recommendations:
  - Maintain coherent macroeconomic policies, including ongoing fiscal consolidation and managing public debt levels.
  - Pursue prudent monetary policy and liquidity management: manage liquidity carefully and raise interest rates where appropriate to alleviate local demand pressures.
- Supply-side recommendations:
  - Increase local food production to reduce reliance on imported food.
  - Improve processing of primary agricultural products to add value and reduce waste, making local food more competitive against imports.
  - Develop infrastructure such as greenhouses and other agricultural facilities to enable year-round production and improve yields.
  - Promote sustainable agricultural practices to boost production and resilience against climate change.

### Banking sector context (overview)
- Sector characteristics and challenges:
  - Banking sector is small, with high operating costs, low asset quality, and low profitability.
  - Sector history:
    - Liberalization in early 2000s; BISTP joined by three new banks in 2004; total reached eight by 2008.
    - Uncertain commercial oil prospects led BCSTP to place three banks under administration within a decade, indicating regulatory weaknesses.
  - Structural constraints:
    - Limited investment opportunities in the small domestic market hinder banks' financing capabilities.
    - A significant portion of deposits is held in foreign currencies, reflecting lack of confidence in the local currency and economy.

*Source: 1stpea2025003-source-pdf - 11.      The model captures inflation by accounting for inflation expectations, as well as*

### 6.      The banking sector is marked by significant concentration, with the largest bank

### 6.      The banking sector is marked by significant concentration, with the largest bank controlling 65 percent of total assets

### Banking sector concentration and systemic risk
- Largest bank controls 65 percent of total assets.
- High concentration:
  - Hinders competition and limits choices for consumers and businesses.
  - Elevates systemic risk—failure or distress of the dominant bank could have cascading effects throughout the economy.
  - Distorts assessment of system liquidity because excess reserves may be disproportionately concentrated in one institution.
  - Undermines the BCSTP's ability to effectively manage liquidity and implement monetary policy.

### Private sector credit performance and lending constraints
- Private sector credit:
  - Has consistently lagged regional peers over the past decade and has exhibited a marked decline since 2019.
  - After nearly five years of contraction, private sector credit has dropped to less than 8 percent of GDP, down from 24 percent pre-pandemic.
- Lending environment factors deterring bank lending:
  - Economic instability characterized by stagnation and uncertain macroeconomic conditions.
  - Limited financial infrastructure.
  - Structural deficiencies including weak contract enforcement and absence of a robust credit information system.
  - Widespread informal sector—many businesses and individuals lack formal registration and bank accounts.
  - High interest rates that remain close to 20 percent, reflecting macroeconomic imbalances and high operating costs, effectively excluding many economic agents from formal financing.

### Non-performing loans (NPLs), judicial uncertainty, and regulatory response
- High NPLs have made banks reluctant to extend new credit, prompting banks to hold more liquid assets.
- Judicial uncertainties, particularly enforcement of collateral, discourage lending.
- Regulatory changes introduced in 2022 on asset classifications, provisioning, and capital adequacy:
  - Contributed to a notable reduction in systemwide NPLs to 10 percent of total loans by Q1:2025, down from a pandemic peak of approximately 34 percent in Q3:2020.

### BCSTP monetary policy and liquidity management framework
- Instruments and features:
  - Open market operations (OMO) via BCSTP Certificates of Deposit (CDs).
  - Required reserves (RR).
  - Standing facilities for temporary liquidity absorption or lending.
  - Use of Treasury Bills (T-bills) permissible but primarily focused on financing budgetary requirements.
  - Refinancing operations in exceptional circumstances using the policy reference interest rate (not used as benchmark for market deposit or lending rates).
- Open Market Operations:
  - OMO regulation established in 2014; actively used since 2020 to absorb excess liquidity.
  - CDs maturities range from 1 to 360 days; since July 2022 issuance shifted to fixed-quantity, variable-rate tenders.
- Required Reserves:
  - Cornerstone tool since 1992.
  - Unremunerated reserves set at 28 percent for local currency liabilities (increase from 18 percent prior to mid-2022).
  - 21 percent for foreign currency liabilities (composed of 90 percent in local currency and 10 percent in foreign currency).
  - Calculation based on the monthly average of deposits recorded in the second month prior to the month of compliance; maintenance period of one month.
- Standing Facilities:
  - Deposit and lending facilities limited to a maturity of five days.
  - Policy reference rate set at 10 percent.
  - Lending facility interest rate is 10.5 percent; deposit facility offers 0 percent.
  - Recently neither facility has been used in policy implementation; due to unremunerated deposit facility and large excess liquidity, banks typically do not utilize them.
- Collateral requirements:
  - Acceptable collateral includes BCSTP’s securities, CDs, T-bills, and deposits in foreign currencies.
  - CDs and securities valued at face value and must be free from encumbrances.
  - Foreign exchange assets valued based on BCSTP exchange rate; securities valued at market value and free from charges.
  - Total collateral must be at least 125 percent of the requested loan amount.
  - Eligible instruments accepted only if their value exceeds the bank's reserve requirement at the BCSTP.

### New monetary policy framework and MPC
- New framework:
  - Targets core inflation calculated as the six-month moving average at 3.1 percent (the average during the period before the pandemic).
  - Seeks to keep banks’ excess reserves in local currency below a threshold by issuing CDs at variable rates until the six-month moving average of core inflation falls below the target.
  - Current threshold set at STN 275 million (previously STN 200 million).
  - Commitment to introduce 12-month CDs to complement 1-month, 3-month, and 6-month options.
  - If excess reserves fall below STN 275 million, the MPC may offer limited amounts of CDs monthly to extract interest rate signals.
- Monetary Policy Committee (MPC):
  - Established in 2013; convenes quarterly and issues press releases on decisions.
  - Composition: (i) Governor of the BCSTP, (ii) Deputy Governor, (iii) Administrators and Directors responsible for liquidity management, markets, economic studies, statistics, and supervision, and (iv) three senior advisors from the BCSTP with expertise in finance and economics.
  - Tasks: set policy reference rate and lending facility rate; provide recommendations on financial stability, fiscal policy, and government debt management.

### Interbank money market and liquidity forecasting
- Interbank Money Market (IMM):
  - Regulation effective since 2015 allows fund exchanges for up to one year, bilaterally, collateralized or non-collateralized.
  - Eligible collateral includes foreign currency deposits, T-bills, and BCSTP CDs.
  - Market currently inactive due to insufficient trust among banks and underdeveloped registration and settlement infrastructure.
  - Banks hesitant to engage in interbank trading and therefore maintain higher excess reserves.
- Liquidity forecasting:
  - Framework initiated in 2011 based on a one-day forecasting model and attempts to extend to one month.
  - Progress impeded by resource limitations and difficulties in obtaining accurate data.
  - Treasury Directorate at the MEF has sought to improve predictability of cash flows, but payments contingent on fund availability complicate forecasting.

### Liquidity situation and recent monetary developments
- Historical liquidity dynamics:
  - Persistent structural excess liquidity before 2017; excess reserves peaked at STN 783 million by early 2017.
  - From 2017 to 2019, excess reserves decreased to STN 494 million by end-2019 due to decreased external support.
  - From 2020 to mid-2022, excess reserves increased significantly due to external support during the pandemic.
- Decline in excess liquidity since mid-2022 driven by:
  - Fiscal consolidation, improved public financial management, Value-Added Tax (VAT) introduction.
  - BCSTP tighter liquidity management, increased reserve requirements, enhanced monitoring, and cessation of BCSTP credit to the government.
- Interest rate and inflation developments:
  - Weighted average CD interest rates rose from 1.3 percent in July 2022 to 4.6 percent by December 2024.
  - T-bill interest rates climbed from approximately 4 percent in mid-2022 to 6.2 percent in April 2024.
  - MEF issued two-year T-bonds offering interest rates ranging from 4 to 6 percent.
  - Short-term deposit interest rates rose from 2.3 percent in mid-2022 to 2.6 percent by December 2024.
  - Six-month moving average of core inflation fell from approximately 12 percent during mid-2022 to end-2023 to 3.3 percent by December 2024.
- Early 2025 developments:
  - Excess reserves increased in Q4 2024 and early 2025 despite sizable CD issuance; stock of CDs rose to STN 335 million by mid-February 2025—50 percent higher than excess reserves in January 2025.
  - Increase attributed to foreign exchange inflows from external disbursements and BCSTP regulation requiring banks to sell 25 percent of their foreign exchange revenues to the BCSTP.
- Distributional issues:
  - Excess liquidity unevenly distributed; one commercial bank holds around two-thirds of the total liquidity surplus.
  - Uneven distribution hampers effective liquidity absorption because not all excess liquidity is readily available for absorption; banks retain portions for transactional and precautionary needs amid lack of functioning standing credit facility, robust interbank market, acceptable collateral, and trust among banks.

*Source: BCSTP and IMF staff calculations, as presented in the chapter.*

### 25.      The BCSTP's clear objectives and effective monetary policy framework are crucial for

### The BCSTP's clear objectives and effective monetary policy framework are crucial for maintaining price stability

### Monetary policy objectives and current challenges
- The BCSTP has established clear objectives aligned with maintaining price stability, including a target for core inflation of 3.1 percent.
- Interest rates set via BCSTP CDs can shape monetary policy and act as a benchmark for the banking system.
- Key issues to address: unabsorbed surplus liquidity, low intermediation, and lack of trust among market participants.

### Open Market Operations (OMOs)
- Recommendations:
  - Introduce 12-month CD options and utilize T-bills for monetary policy purposes to absorb structural excess liquidity and achieve the excess reserves target of STN 275 million.
  - Offer limited quantities of CDs once excess reserves dip below the established target or core inflation falls below the 3.1 percent target.
  - Continue CD auctions with one-month maturities or initiate auctions with shorter maturities of 7 or 14 days to assess interest rate signals.
  - Promote transparency, diversify counterparties (including smaller institutions and non-bank entities), and leverage technology (electronic trading platforms) to foster a more competitive and inclusive market.
  - Issue longer-term instruments to provide additional flexibility in managing market conditions.
  - Establish regular market assessments, improved communication channels, and incentives for smaller institutions to participate in OMOs.
  - Implement monitoring and reporting frameworks for OMO outcomes to build trust and inclusivity.

### Required Reserves (RR)
- Findings:
  - Current practice uses banks’ liabilities with a two-month lag as the RR base, which fails to reflect recent balance sheet positions.
- Recommendations:
  - Calculate the RR base using the average of the last month’s data for a more accurate reflection of current banking conditions and liquidity needs.
  - Consider remunerating RR at or near the targeted policy rate to reduce the spread between deposit and lending rates if the objective is to correct distortions; if the objective is to drain liquidity cheaply, do not (fully) remunerate RR.
  - Recognize trade-offs: most central banks do not remunerate RR, reflecting an emphasis on draining liquidity cheaply.

### Standing facilities
- Standing lending facility:
  - Make the facility permanently available instead of limiting it to five days.
  - Lift access restrictions, allowing eligible institutions unlimited borrowing subject to acceptable collateral.
  - Set the lending facility rate above the BCSTP CD rate to act as a penalty, discourage unnecessary borrowing, and establish a ceiling for interbank rates.
  - Distinguish clearly between the standing lending facility (monetary policy tool for qualifying banks) and lender of last resort function (loans to solvent institutions that have exhausted other funding options), with different parameters for maturity, collateral policy, and pricing.
- Standing deposit facility:
  - Operationalize the facility by linking its interest rates to the BCSTP CD rate.
  - Ensure the deposit facility rate remains below the BCSTP CD rate to create a corridor and encourage banks to lend excess reserves in the interbank market rather than deposit them with the BCSTP.

### Emergency Liquidity Assistance (ELA)
- Findings:
  - Existing ELA framework lacks clear distinctions between short-term liquidity requirements, temporary needs, and resolution-related needs.
- Recommendations:
  - Establish well-defined categories for liquidity needs to clarify appropriate support avenues.
  - Standardize rates applied to each facility within the ELA framework to enhance transparency and predictability.

### Interbank Money Market (IMM)
- Recommendations:
  - Strengthen the legal and regulatory framework to ensure enforceable contracts and reliable information sharing.
  - Invest in technology and infrastructure, including electronic trading and settlement systems.
  - Make CDs tradable within the interbank market, supported by a robust collateral framework to build participant confidence.
  - These measures aim to create a more dynamic IMM, promote interbank trading, and reduce banks’ need to hold high excess reserves.

### Liquidity forecasting
- Findings:
  - Advances in liquidity concepts exist, but data collection challenges persist; improving data quality requires collaboration with data providers.
- Recommendations:
  - Initiate liquidity forecasting, starting with one-month horizons and extending to one year once validated.
  - Improve data collection systems for government transactions, foreign exchange operations, and banking sector activities.
  - Enhance coordination between the MEF and the BCSTP to share forecasts of government expenditures and revenues.
  - Use simple estimation methods and available high-frequency data to understand intramonth liquidity when current month excess reserves may differ from lagged data.
  - Establish daily routines for monitoring bank reserves and implement medium- to long-term liquidity assessments.
  - Invest in IT systems and human resources to automate forecasting and expand projections to include maturities of issued CDs, T-bills, and Treasury Bonds (T-bonds).

### BCSTP recapitalization, communication, and capacity development
- Recapitalization:
  - Recapitalize the BCSTP (via government capital injection or issuing government bonds, and retaining a higher portion of profits) to strengthen its balance sheet, credibility, and capacity to absorb liquidity and implement monetary policy.
- Communication:
  - Enhance communication with financial institutions, the government, and the public through regular reports, press releases, public forums, and meetings with commercial banks to explain policy rationale and build trust.
- Capacity development:
  - Continue strengthening analytical tools and skills for monetary policy implementation and liquidity management.
  - The IMF stands ready to provide ongoing technical assistance and training to the authorities of São Tomé and Príncipe.

### Conclusion: expected outcomes of recommended reforms
- Implementing the recommendations (diverse CD options, enhanced RR system, improved standing facilities, liquidity forecasting, communication improvements, and BCSTP recapitalization) can:
  - Lead to more effective implementation of monetary policy and strengthen the monetary transmission mechanism.
  - Help reduce inflation and promote economic growth and development in São Tomé e Príncipe.
  - Increase transparency and trust among market participants, improving liquidity management and monetary policy effectiveness.
- The IMF is prepared to continue providing technical assistance to address the country’s specific challenges.

*Source: IMF staff paper excerpts on monetary policy and liquidity management for São Tomé and Príncipe.*

### 4.      As a full member of the United Nations Framework Convention on Climate Change

### 4.      As a full member of the United Nations Framework Convention on Climate Change 

### GHG targets and emissions
- São Tomé and Príncipe’s 2021 NDC target: reduce greenhouse gas emissions (GHG) by 27 percent by 2030 compared to a business-as-usual scenario.  
- Estimated total cost of the 2021 NDC: around US$150 million.  
- Emissions profile through 2018 (latest available data): energy sector largest emitter, followed by agriculture and waste sectors.  
- “Forest and other land uses” sector acts as a CO2 sequester, making São Tomé and Príncipe on balance a net carbon sink through 2018.  
- 2015 GHG reduction target was nearly doubled by the 2021 NDC, with a focus on forestry, energy, transport, and industrial sectors.

### Mitigation and adaptation measures
- Mitigation approaches: promote renewable energy, improve energy efficiency, implement sustainable land management practices.  
- Adaptation emphasis in 2021 NDC: coastal zones, agriculture, water resources. Key adaptation actions include:
  - improving disaster risk management;
  - developing early warning systems;
  - enhancing climate resilience across infrastructure and social systems.  
- First biannual update report (BUR) in 2022: provided details on mitigation measures, national MRV system, and the country’s capacity and financial needs to report emissions and support climate-resilient development.  
- Energy sector action plan: PADRES (decarbonization and resilience of the energy sector) — outlines steps to move away from fuel-based electricity production towards renewables.

### Institutional framework and planning
- Institutional bodies and instruments:
  - General Directorate for the Environment (DGA) under the Ministry of Infrastructure, Natural Resources and Environment (created 2007);
  - National Institute of Meteorology — records and disseminates climate data;
  - National Council for Disaster Preparedness and Response (CONPREC) (established 2011);
  - National Committee for Climate Change (created 2012).
- Plans and laws: Contingency Plan for Natural Disasters (2016-2020), National Strategy for Disaster Risk Management (last updated 2016), Multisectoral Investment Plan for coastal zone management. A draft disaster risk management law is under consideration (would introduce an early warning system, assign responsibilities, and have a color-coded alert system).
- Ongoing effort to expand disaster preparedness and response plans to local level with World Bank support.

### International support and projects
- National Adaptation Plan (NAP) development: launched June 2022 in collaboration with UNEP; a 42-month project funded by the Green Climate Fund (over US$2.9 million). Work on the NAP is ongoing.  
- World Bank support through West Africa Coastal Areas (WACA) Management Program: US$15 million over 5 years, mainly for physical investment including relocation to safe expansion zones, and strengthening early warning systems.

### DIGNAD model: purpose and channels
- Model: Debt, Investment, Growth, and Natural Disasters (DIGNAD) — dynamic general equilibrium model to analyze impact of investing in climate-resilient capital on economic growth and debt sustainability under various scenarios.  
- Five channels through which natural disasters affect the economy:
  1. damages to public capital;
  2. damages to private capital;
  3. temporary productivity loss;
  4. reduced public investment efficiency;
  5. loss in creditworthiness.  
- The model evaluates debt sustainability risks, effects of ex-ante policies (climate-resilient infrastructure vs non-resilient; fiscal buffers), accounts for public investment inefficiency and absorptive capacity constraints, and captures linkages among public investment, growth, and debt.

### Calibration and initial values (table of calibrated parameters and initial values)
- Grants to GDP ratio: 7.8  
- Domestic public debt to GDP ratio: 42.8  
- Concessional public debt to GDP ratio: 59.2  
- External commercial public debt to GDP ratio: 12.3  
- Return to standard infrastructure: 30.0  
- Return to climate-resilient infrastructure: 40.0  
- Public standard infrastructure investment to GDP ratio: 6.0  
- Depreciation rate of standard public infrastructure: 7.5  
- Depreciation rate of resilient public infrastructure: 3.0  
- Public investment efficiency: 50.0  
- Initial calibration aligns steady state with projected values for 2025-28 (about 6 percent of GDP public capital infrastructure investment).  
- Initial domestic public debt calibrated using 2023 value of 42.8 percent of GDP.  
- In absence of specific local information, parameters of the average LIC in Marto, Papageorgiou and Klyuev (2018) were used.

### Scenario design and simulated disaster
- Baseline starting point: state of the economy in 2024.  
- Simulated natural disaster: occurs in 2029, modeled as one-time flooding with a one-off 6 percent point GDP loss.  
- Three simulation sets, each with three scenarios:

Simulation set 1 — role and scale of adaptation investment:
- Scenario 1.1 (Baseline): no resilient infrastructure investment pre-disaster. Post-disaster recovery (2029 onward) rebuilds damaged infrastructure to pre-disaster level in about five years. Recovery financed equally with domestic debt and external commercial borrowing.
- Scenario 1.2 (Resilient infrastructure): authorities invest 2 percent of GDP per year in climate-resilient public infrastructure in 2025-28. Financing equally split between grants and concessional loans. Disaster and recovery assumptions same as baseline.
- Scenario 1.3 (Resilience scaling-up): invest 4 percent of GDP per year in 2025-28 in resilient infrastructure. Additional public investment fully financed with grants, overall financing mix ¾ grants and ¼ concessional loans.

Simulation set 2 — resilient vs standard infrastructure and public investment management (PIM):
- Scenario 2.1 (Standard infrastructure): public investment in standard infrastructure of 2 percent of GDP per year in 2025-28, financing equally split between grants and concessional loans; disaster in 2029.
- Scenario 2.2 (Resilient infrastructure): same as scenario 1.2, replacing standard with climate-resilient infrastructure.
- Scenario 2.3 (Resilient + PIM reform): scenario 2.2 plus an increase in public investment efficiency by 10 percentage points.

Simulation set 3 — ex-post vs ex-ante financial preparedness:
- Scenario 3.1 (Ex-post disaster management): disaster in 2029, recovery requires investment spending of 2 percent of GDP per year in 2029-2032, financed equally by grants and concessional loans.
- Scenario 3.2 (Ex-ante contingency fund): authorities invest 2 percent of GDP per year during 2025-28 into a disaster contingency fund, funding split equally between grants and concessional loans; after disaster hits in 2029, drawdown at 2 percent of GDP per year in 2029-32.
- Scenario 3.3 (Resilient infrastructure): same resource amount invested ex-ante in climate-resilient physical infrastructure rather than a financial contingency fund (same as scenario 1.2).

### Key simulation outcomes and macro-fiscal implications
- General mechanisms:
  - Pre-disaster scaling up of public investment crowds in private investment and increases output; effects larger for resilient than standard infrastructure due to higher returns and slower depreciation of resilient capital.
  - Natural disaster reduces output via permanent damages to public and private capital and temporary productivity losses.
  - A natural disaster increases fiscal deficits due to reconstruction costs; resilient infrastructure reduces damages and repair costs, containing the debt increase and enabling quicker stabilization of debt levels.
  - Higher public investment efficiency reduces recovery spending needs and yields more favorable debt dynamics.

- Quantitative outcomes highlighted:
  - GDP decline when disaster hits:
    - Baseline scenario: 6.3 percentage points relative to the steady state.
    - Standard infrastructure scenario: about 5 percentage points.
    - Resilient infrastructure scenario: 2.3 percentage points.
  - Private investment: more resilient under resilient infrastructure scenarios because climate-resilient capital raises returns to private investment and spurs growth.
  - Public debt: resilient infrastructure associated with lower public debt in the long-term compared to baseline and standard scenarios; higher public investment efficiency further improves debt dynamics.
  - Private consumption: decline following the shock is less pronounced under resilient infrastructure; resilient capital supports swifter recovery and rebound in private consumption that can exceed pre-disaster levels.

### Policy implications and priorities
- Investing ex-ante in climate-resilient infrastructure yields:
  - smaller immediate output losses when disasters occur;
  - lower reconstruction and fiscal costs post-disaster;
  - stronger crowding-in of private investment and higher long-run GDP stabilization.
- Increasing public investment efficiency amplifies benefits of resilient investment by delivering more infrastructure with the same resources and lowering post-disaster recovery costs.
- Trade-offs between financial insurance (contingency funds) and physical resilient investment: investing the same pre-disaster resources in resilient physical infrastructure tends to generate larger macroeconomic resilience benefits than holding equivalent resources in a contingency fund, according to simulated scenarios.

*Source: IMF staff calculations.*

### 17.      The DIGNAD simulation 3 (Figure 5) results imply that investing in resilient

### 17.      The DIGNAD simulation 3 (Figure 5) results imply that investing in resilient infrastructure is better than both ex-post disaster management and ex-ante disaster preparedness via a financial contingency fund.

### Key findings from the simulations
- The timing of financial inflows has no impact on GDP growth, private investment or private consumption, unless they are invested into real capital/infrastructure (monetary neutrality).
- The real impact of ex-post disaster management using grants and concessional loans is identical to the ex-ante use of similar financial resources to accumulate a contingency fund to finance the recovery.
- Using the same amount of money to build climate-resilient infrastructure leads to better outcomes for GDP growth, private investment, and private consumption in the short- and long-run.
- Climate-resilient infrastructure lowers disaster losses and damages and, thereby, repair costs, contributing to lower public debt over the long term.

### Implications for disaster finance and investment
- Financial instruments that merely time transfers (ex-ante contingency funds vs ex-post grants/concessional loans) do not change macroeconomic outcomes unless used to augment real capital/infrastructure.
- Prioritizing investments in resilient public capital yields superior macroeconomic and fiscal outcomes compared with non-capital financial preparedness measures.

### C. Conclusion and Policy Recommendations

### Overarching conclusion
- The DIGNAD simulations illustrate that climate-resilient public investment could increase São Tomé and Príncipe’s resilience to natural disasters.

### Policy recommendations
- Secure grants and concessional financing for resilient infrastructure and adaptation plans.
- Strengthen public investment management to ensure more efficient use of limited resources and maximize impact.
- Accelerate the preparation of a National Adaptation Plan to provide strategic direction.
- Invest in resilient infrastructure, install, and use early warning systems, and invest in climate and environmental protection to better prepare for natural disasters.

*Democratic Republic of São Tomé and Príncipe — International Monetary Fund content excerpt*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2025/english/1stpea2025003-source-pdf.pdf_