## 1. Sovereign Spreads: Actual vs Model-Fitted

## Source details

**Canonical URL:** [1. Sovereign Spreads: Actual vs Model-Fitted](https://www.imf.org/-/media/files/publications/cr/2025/english/1bhsea2025002-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2025/english/1bhsea2025002-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2025/english/1bhsea2025002-print-pdf.pdf.json)

---

### A. Introduction
- Spreads on Bahamian sovereign bonds issued on international capital markets increased sharply following the onset of the COVID-19 pandemic, rising by as much as 1,200 basis points—more than five times their pre-pandemic levels.
- Causes cited: substantial rise in debt from pandemic-related expenditures and the impact of a major hurricane that occurred just prior to the pandemic; severe impact on tourism from the global downturn in travel.
- Wider spreads imply higher government borrowing costs and a larger share of fiscal resources devoted to servicing public debt.

### B. Sample and Stylized Facts
- Sample: unbalanced panel of 73 countries classified as emerging and developing economies by the IMF, monthly frequency January 2005 to April 2024.
- Dependent variable: hard-currency sovereign bond spread.
- Explanatory variables: domestic covariates (GDP growth, inflation, government debt, current account balance, reserves, governance index) and global covariates (global growth, MOVE), plus dummies (EMBIG inclusion, IMF program, commodity exporter) and interactions (sovereign rating grades × VIX).
- Data handling: variables not at monthly frequency converted using quarter or year values; Dickey-Fuller and Perron-P tests used for stationarity; governance index, government debt, external debt, dollar index, and global growth detrended using a linear trend.
- Bahamian position in sample:
  - Sovereign spreads near sample median pre-COVID, above 75th percentile during pandemic.
  - One of the largest declines in output during the pandemic, followed by rapid recovery.
  - Public debt increased steadily since 2005; large upward jump in debt-to-GDP in 2020; now among top 25 percent of countries in sample.
  - External debt below sample median.
  - Sovereign credit rating deteriorated over sample period; lowered twice since 2020 and currently standing at a single-B rating.
  - Current account balance tends to be very negative; relatively high level of currency reserves throughout the pandemic.
  - Governance index relatively high, close to 75th percentile, though declining over sample period.

### C. Empirical Model
- Baseline panel regression specification estimates sovereign spread as function of:
  - Domestic covariates X_it (GDP growth, inflation, government debt, current account balance, reserves, governance index) with coefficients β.
  - Global covariates Z_it (global growth, MOVE index) with coefficients γ.
  - Dummy variables D_it (including EMBIG inclusion, IMF program, commodity exporter) with coefficients δ.
  - Interaction terms: sovereign rating grade dummies s_irg_k interacted with VIX, coefficients ω_k.
  - Intercept α and error term ε_it.
- Rating variable: median of Moody’s, S&P, Fitch; values AAA/AA (combined), A, BBB, BB, B, CCC, CCX (CCX denotes default).
- Sample excludes observations with spreads above 3000 basis points.
- Estimation period: January 2005 to April 2024.

### D. Key Empirical Results and Findings
- Model fit and dynamics:
  - Model fits Bahamian sovereign spread dynamics well up to 2020.
  - In 2020 both actual and model-implied spreads surged by over 400 basis points.
  - From 2021 a gap emerges: model-implied spreads decline more than actual; residual substantial in 2023 but narrowed to 70 basis points by end of sample (April 2024).
- Rating changes:
  - Median rating dropped from “BBB” to “BB” in early 2020, and from “BB” to “B” in late 2022.
  - Fitted spread exhibits discrete jumps at downgrades; spread increases can precede official rating downgrades by several months.
- Cross-country alignment:
  - By 2024 Bahamian spreads aligned with peer economies and model-implied spread.
- Quantitative associations and signs (statistically significant at least at the 10 percent level for most variables):
  - A higher domestic GDP growth rate associated with around 13 basis points decline in sovereign spread.
  - A one percentage point higher global growth rate associated with a 28 basis points lower sovereign spread.
  - Higher inflation and higher government debt associated with higher spreads.
  - Higher current account balance, higher governance index, higher reserves, and being a commodity exporter associated with lower spreads.
  - MOVE index positively related to sovereign spreads.
- Role of global risk aversion and rating interaction (at median VIX of 17):
  - “A” rated country faces about 19 basis points higher spread than “AAA”/“AA”.
  - “BB” rated country faces about 66 basis points higher spread.
  - “B” rated country (like The Bahamas) faces about 269 basis points higher spread.
  - “CCX” (default) associated with 663 basis points higher spread compared to “AAA”/“AA”.
- EMBIG inclusion:
  - Inclusion in JP Morgan’s EMBIG index associated with lower sovereign spreads.
  - The Bahamas would have had spreads compressed by 56 basis points compared to similar fundamentals if included in EMBIG.

### E. Decomposition of Recent Change (January 2023 to March 2024)
- Total model-implied decline in fitted spread: about 140 basis points.
- Major contributors to decline:
  - Interaction term between sovereign rating and VIX: reduction of over 100 basis points (entirely attributable to decline in VIX since Bahamas rating unchanged).
  - Decline in MOVE index: about 20 basis points reduction.
  - Increase in global growth outlook: about 40 basis points reduction.
- Offsetting domestic factors:
  - Moderation in Bahamian GDP growth and slight decline in reserves imply an increase in fitted spread by about 30 basis points.
- Net depicted in Figure 3 (bar decomposition; basis points scale).

### F. Projected Spreads Scenarios to end-2029
- Two illustrative scenarios using staff baseline projections for domestic variables and staff baseline projections and historical averages for global variables:
  - Scenario 1 — Unchanged sovereign rating:
    - Fitted spread declines to about 430 basis points by end of 2029.
    - Reduction of 65 basis points compared to March 2024.
    - Large part of projected reduction (45 basis points) driven by assumed decline in MOVE index to historical averages.
  - Scenario 2 — One-notch rating upgrade:
    - Fitted spread declines by an additional 175 basis points relative to unchanged-rating scenario.
    - Implied spread about 260 basis points by 2029, close to pre-pandemic level.
    - Interpretation: rating upgrades often linked to improvements in macro fundamentals and fiscal situation; appropriate domestic policy mix can lead to substantial drops in spreads and funding costs.

### Robustness checks and regression-specification findings
- Baseline model extended to specification adding country fixed effects and time fixed effects; eight additional specifications estimated.
- Qualitative robustness:
  - Coefficients generally stable across specifications, including with country or time fixed effects.
  - EMBIG dummy more negative in specifications with country fixed effects (5, 7, 9) and when sovereign credit ratings excluded (8, 9).
  - IMF program dummy loses significance in some specifications with country fixed effects (5, 9) and changes sign when sovereign ratings excluded (8).
  - Excluding sovereign ratings (specs 8, 9) increases public debt coefficient magnitude (by as much as a factor of ten) while adjusted R-squared declines.
- Key regression coefficient patterns (selected exact values preserved):
  - GDP Growth coefficients across columns: (1) -13.23***  (se 1.22); (2) -14.25***  (se 1.26); (3) -17.12***  (se 1.72); (4) -13.58***  (se 1.22); (5) -8.71**    (se 3.99); (6) -13.05***  (se 3.58); (7) -11.29***  (se 3.85); (8) -18.13***  (se 1.51); (9) -13.67**   (se 5.99).
  - Inflation coefficients: (1) 7.03***  (se 0.47); (2) 6.28***  (se 0.55); (3) 6.90***  (se 0.65); (4) 6.78***  (se 0.46); (5) 6.18***  (se 1.42); (6) 6.88***  (se 1.39); (7) 3.55***  (se 1.33); (8) 10.57*** (se 0.67); (9) 8.86***  (se 2.23).
  - CA Balance coefficients: (1) -2.70***  (se 0.44); (2) -2.93***  (se 0.42); (3) -2.82***  (se 0.44); (4) -2.53***  (se 0.42); (5) -1.32     (se 1.81); (6) -2.42**   (se 1.22); (7) -1.64     (se 1.97); (8) -2.43***  (se 0.44); (9) 1.94      (se 1.49).
  - Governance coefficients: (1) -2.94***  (se 0.19); (2) -2.50***  (se 0.23); (3) -2.31***  (se 0.26); (4) -2.92***  (se 0.22); (5) -0.70     (se 4.07); (6) -3.28***  (se 0.87); (7) -0.59     (se 4.25); (8) -8.25***  (se 0.19); (9) -4.20     (se 4.63).
  - Debt coefficients: (1) 0.35*    (se 0.18); (2) 0.10     (se 0.17); (3) 0.13     (se 0.19); (4) 0.35**   (se 0.17); (5) 4.76***  (se 1.31); (6) 0.52     (se 0.70); (7) 5.93***  (se 1.40); (8) 2.69***  (se 0.19); (9) 5.87***  (se 1.70).
  - Reserves coefficients: (1) -4.12***  (se 0.32); (2) -4.14***  (se 0.27); (3) -3.91***  (se 0.30); (4) -4.08***  (se 0.30); (5) -3.19*    (se 1.86); (6) -4.53***  (se 1.36); (7) -3.41*    (se 1.87); (8) -6.37***  (se 0.31); (9) -3.60*    (se 1.92).
  - Comex coefficients: (1) -62.33***  (8.80); (2) -71.72***  (8.38); (3) -66.71***  (8.48); (4) -56.80***  (8.99); (6) -64.95**   (25.51); (8) -63.05***  (7.55).
  - IMF dummy coefficients: (1) -32.76***  (7.91); (2) -37.75***  (7.56); (3) -29.73***  (8.53); (5) -47.30     (33.14); (6) -29.57     (25.83); (7) -40.69     (30.72); (8) 57.60***   (7.15); (9) -16.56     (32.04).
  - EMBIG dummy coefficients: (1) -55.80***  (5.70); (2) -54.13***  (6.47); (3) -36.07***  (7.14); (5) -98.44**   (43.30); (6) -47.66     (42.17); (7) -141.80*** (40.50); (8) -142.45*** (7.36); (9) -125.07*** (33.12).
- Observations: 11,916 in all specifications.
- Adjusted R2 by specification: (1) 0.60; (2) 0.61; (3) 0.55; (4) 0.60; (5) 0.38; (6) 0.58; (7) 0.30; (8) 0.45; (9) 0.12.
- Note: *, **, *** indicate statistical significance at the 10%, 5%, and 1% levels; standard errors are robust standard errors.

### Data, variable definitions, and sample statistics
- Key variable definitions and sources preserved exactly:
  - Spread: Sovereign bond spread; Unit: Basis points; Source: Bloomberg Finance L.P., JPMorgan, IMF Spread Monitor.
  - GDP Growth: Real GDP growth; Unit: Percent, yoy; Source: IMF, Consensus Economics.
  - Inflation: CPI inflation rate; Unit: Percent, yoy; Source: IMF, Consensus Economics.
  - CA Balance: Current account balance (share of GDP); Unit: Percent; Source: IMF, National sources.
  - Governance: Governance index; Source: World Bank Worldwide Governance Indicators; Simple average of World Bank regulatory quality and government effectiveness.
  - Debt, Ext. Debt, Reserves, Sovereign credit rating, Comex, EMBIG, IMF, Dollar, Global Growth, VIX, MOVE: sources as listed in the source content.
- Summary statistics — Domestic covariates (observations, mean, standard deviation, minimum, median, maximum):
  - Spread: 12639 obs; Mean 476.6; Standard Deviation 660.2; Minimum -3.9; Median 305.4; Maximum 12481.9
  - GDP Growth: 12639 obs; Mean 3.6; Standard Deviation 3.0; Minimum -23.5; Median 3.7; Maximum 21.6
  - Inflation: 12639 obs; Mean 6.1; Standard Deviation 8.3; Minimum -2.5; Median 4.3; Maximum 201.9
  - Debt: 12611 obs; Mean 52.8; Standard Deviation 26.3; Minimum 0.0; Median 50.0; Maximum 158.1
  - CA Balance: 12639 obs; Mean -1.6; Standard Deviation 7.5; Minimum -46.2; Median -2.0; Maximum 37.8
  - Ext. Debt: 12619 obs; Mean 54.4; Standard Deviation 38.8; Minimum 0.0; Median 42.9; Maximum 243.1
  - Reserves: 12639 obs; Mean 17.6; Standard Deviation 9.9; Minimum 0.0; Median 15.5; Maximum 81.8
  - Governance: 12639 obs; Mean 47.3; Standard Deviation 18.6; Minimum 4.2; Median 48.1; Maximum 88.4
- Summary statistics — Global covariates (observations, mean, standard deviation, minimum, median, maximum):
  - Dollar: 232 obs; Mean 89.1; Standard Deviation 9.3; Minimum 72.4; Median 89.8; Maximum 110.8
  - Global growth: 232 obs; Mean 4.4; Standard Deviation 1.0; Minimum 1.4; Median 4.5; Maximum 6.5
  - VIX: 232 obs; Mean 19.4; Standard Deviation 7.6; Minimum 10.1; Median 17.0; Maximum 51.9
  - MOVE: 232 obs; Mean 84.3; Standard Deviation 30.5; Minimum 42.7; Median 75.2; Maximum 209.9
- EMBIG and IMF dummies:
  - EMBIG: 12639 observations; Obs. with d=1: 11721; Countries with at least one d=1: 68; Countries with at least one d=0: 15; Total nr of countries: 73
  - IMF: 12639 observations; Obs. with d=1: 2641; Countries with at least one d=1: 40; Countries with at least one d=0: 73; Total nr of countries: 73

### Climate-change risks, channels, and quantitative approach (Annex)
- Key contextual facts:
  - 80 percent of The Bahamas' land area is less than 1.5 meters above sea level.
  - Tourism contributes about 50 percent to GDP.
  - Hurricane Dorian (2019) caused $3.4 billion in damage to physical assets, amounting to 25 percent of The Bahamas' GDP.
  - Over the past decade, natural disasters have averaged 3.2 percent of GDP in annual damages.
- Climate projections and impacts:
  - Acevedo (2016) projects hurricane-related damages in The Bahamas will increase by 31 percent under 3°C warming and by 42 percent under 4.3°C.
  - Sea levels projected to rise by nearly 0.5 meters by the end of the century; this would place about 41 percent of the land and 22 percent of the population below sea level.
  - Without adaptation, a 0.5 m sea level rise could cause the complete loss of up to half of sandy beaches located near hotel infrastructure (Spencer et al., 2022).
- Exposure heterogeneity:
  - Smaller islands (Exuma, Andros, Cat Island, and other smaller family islands) contribute 5.1 percent of national income and 7 percent of the population, but may experience disproportionately higher income losses and have lower adaptation capacity.

### Quantification strategy and modeling
- Dual-pronged approach:
  - (i) intensification of hurricanes (recurring disasters) modeled with a Markov-switching dynamic general equilibrium model that switches between disaster and non-disaster states; hurricanes destroy a share of private and public capital stocks, lowering output while capital is rebuilt.
  - (ii) permanent land inundation (slow onset sea level rise) losses estimated via a two-sector output accounting framework incorporating natural capital; uses Spencer et al. (2022) sandy beach erosion projections and assumes exposed physical assets proportional to exposed population.
- Key modeling assumptions preserved:
  - Hurricanes inflict damages amounting to 7.3 percent of GDP per event.
  - Damages increase in line with projections in Acevedo (2016).
  - Probability of hurricane occurrence matches historical patterns and is assumed to remain constant with climate change.
  - Resilient investments cost the government 25 percent more than standard investments.
  - Illustrative resilience assumption: protective measures ensure 80 percent of the exposed private physical assets are safeguarded against permanent inundation; similar share of government infrastructure is made resilient; protection of natural capital assumes 80 percent preservation.
  - Staff estimate: prevailing natural disaster profile reduces current potential output level by about 8 percent, relative to a counterfactual without tropical storms.
- The analysis translates risk reduction achieved through adaptation into avoided output losses but does not evaluate cost-effectiveness of specific adaptation strategies.

### Adaptation scenarios (illustrative) and projected potential output losses
- Scenario 1: No adaptation — substantial capital losses from permanent inundation and repeated hurricanes.
- Scenario 2: Protection of physical assets — public investment in safeguarding people, private assets, and government infrastructure; long-term increase in the level of GDP of around 5.5-6.8 percent once physical assets are secured.
- Scenario 3: Protection of physical assets and natural capital — further reduction in long-run traded output losses compared to Scenario 2.
- By 2100, combined with losses from permanent inundation, additional potential output losses could be:
  - between 8.3 percent of GDP under intermediate GHG emission scenario (SSP2-4.5)
  - and 11.1 percent of GDP under severe global warming scenario (SSP5-8.5).
- Under Scenario 2 (physical protection), traded output losses remain sizeable because natural capital remains vulnerable.
- Under Scenario 3 (including natural capital preservation), potential output losses could be reduced from 2.8-4.3 percent of GDP under Scenario 2 to 1.2-2.3 percent of GDP under Scenario 3.
- Export-oriented traded sector particularly vulnerable due to reliance on natural capital.

### Calibration table — selected parameters and initial values (exactly preserved)
- Potential GDP growth rate: 1.5
- Private investment to GDP: 23
- Private consumption to GDP: 59
- Public investment to GDP: 2
- Public consumption to GDP: 20
- Net exports to GDP: -4
- Depreciation rate of private capital: 5
- Depreciation rate of non-resilient public capital: 7.5
- Depreciation rate of resilient public capital: 3
- Public investment efficiency: 50
- Share of traded goods in total output: 40
- Share of domestic demand for traded output: 10
- Share of imported capital goods in total investment: 50
- Share of imported goods in domestic consumption: 27
- Annualized (implicit) interest rate on domestic debt: 4
- Share of hotel rooms in New Providence and Grand Bahama: 70
- Share of natural capital resistant to climate change on New Providence and Grand Bahama: 70
- Share of natural capital resistant to climate change on other islands: 50
- Share of exposed physical assets resilient to climate change on New Providence and Grand Bahama: 70
- Share of exposed physical assets resilient to climate change on other islands: 60
- Share of exposed public capital resilient to SLR on New Providence and Grand Bahama: 75
- Share of public capital resilient to SLR on other islands: 30
- Share of physical capital located in New Providence and Grand Bahama: 69.4
- Traded sector: Labor share in production: 42.7
- Traded sector: Capital share in production: 18.3
- Traded sector: Elasticity of output to public capital: 0.15
- Traded sector: Natural capital share in the production function: 39
- Non-traded sector: Elasticity of output to public capital: 0.15
- Non-traded sector: Labor share in production: 60
- Non-traded sector: Capital share in production: 40

### Electricity sector — challenges and 2024 reform plan
- Challenges:
  - Renewable Energy (RE) participation: about 2 percent of the electricity mix.
  - Average solar irradiation: 5.3-kilowatt hours per square meter per day (kWh/m2/day).
  - Electricity price in The Bahamas: US$0.28 kWh (above Caribbean average US$0.24 kWh).
  - Generation replication across ~30 inhabited islands increases costs.
  - More than 80 percent of firms experienced power-outs in 2019/20; outages averaged 3 times per month and created a loss of annual sales of 2.5 percent.
  - T&D technical losses: 12.3 percent; commercial losses (unbilled): 9.6 percent.
  - Energy intensity (2021): 2.87 Mega Joule (MJ) per 2017 USD PPP GDP — an increase of 42 percent over the last 15 years.
  - Institutional: MET created in 2024 and will need to build capacity rapidly.
- 2024 Electricity Reform Plan components and projected impacts:
  - GOBH updated NDCs (2022) targets:
    - 30 percent reduction of GHG emissions compared to Business-as-Usual by 2030.
    - 35 percent of vehicle purchases electric and 15 percent hybrid by 2030.
    - At least 30 percent of renewables in the electricity mix by 2030.
  - Reform aims: repair T&D, partially replace heavy-oil with LNG, expand RE production.
  - MET anticipates reforms will reduce energy production cost to 22.5 cents per kWh, potentially saving up to $170 million annually, and will not require direct financing from the central government.
  - Reform completion target: December 2025.
  - Key components:
    - 30 percent renewable energy target by 2030 and 120 MW of solar energy: 50 MW utility-scale in Family Islands; 70 MW in New Providence (25 MW used to charge existing batteries). Independent power producers required to have at least 50 percent Bahamian ownership.
    - Integrating LNG: convert diesel systems to LNG and build combined cycle LNG plant in New Providence via PPA.
    - Modernization of T&D in New Providence: Estimated $130 million needed; create SPV Bahamas Grid Company Ltd with BPL owning 40 percent of SPV shares. SPV will allocate .0025 cents per Bahamian dollar of its revenue to pay down legacy debt and another .0025 cents to a hurricane fund for New Providence disasters.
    - Rate adjustments (implemented July 2024): fuel charge tariff reduced by 2.5 cents for first 800 kWh and increased by 1.5 cents for usage above 800 kWh; should benefit 58 percent of residential consumers and 56 percent of regular commercial consumers; no base rate charged for first 200 kWh for residential consumers; commercial base tariff changes specified.
- Projected macroeconomic benefits:
  - Reform could narrow the current account deficit, reduce vulnerability to commodity price shocks, boost growth, and significantly reduce CO2 emissions.
  - IMF RE-BOP tool estimate: successful implementation could lift real GDP growth by 0.5 percentage points by 2035.
- PPP fiscal risk mitigation measures by authorities for announced reforms:
  - Ministry of Finance plays a central role in negotiations.
  - Government is not granting any guarantees to private sector participants.
  - Some PPAs expected to include key performance indicators with penalties.

### Box — Bahamas Power and Light (BPL) (selected findings)
- BPL operates 29 plants with total generating capacity of 532 Megawatts (MW) and serves 115,000 customers.
- BPL needs an estimated USD 500 million to refurbish and replace aging infrastructure.
- Fuel constitutes over half of BPL’s operating expenses.
- BPL has been operating at a loss since 2022; net profit margin was -20 percent and -24 percent in FY2022 and FY2023, respectively.
- Average collection period: 135 days (increase of 45 percent since 2017); receivables from government customers about USD 89 million.
- Family Islands: average age of generation units 23 years vs 10 years in New Providence; opportunities exist for hybrid solar/battery/diesel systems.

### Fiscal risk mitigation for PPPs and housing affordability (policy options)
- PPP mitigation approaches:
  - Leading role for Ministry of Finance in gateway process and approvals.
  - Clearly define risk-sharing; public sector should take on risks it can control; consider PPP laws or standardized contracts.
  - Ensure accountability of executing agencies with sufficient autonomy.
  - Clearly define payment authority to permit timely and legal payments to contractors.
  - Impose limits on scale and scope of PPP program (annual limits, constraints on new commitments/guarantees, ceilings on outstanding commitments).
  - Alternative: treat PPPs as budgetary expenditure with commensurate impact on public sector debt to improve transparency.
- Electricity reform PPP measures already in place:
  - Ministry of Finance central role; no government guarantees; PPAs with KPI-based penalties expected.
- Housing affordability — background and recommended options:
  - Population grew 13 percent between 2010 and 2022; New Providence up 20 percent.
  - Residential construction completions: 607 in 2023 vs peak 1,705 in 2006.
  - Between 2012 and 2022, implied price of Real Estate, Owner Occupied and Actual Rents activities increased by 14 percent vs 2 percent for employee compensation per capita.
  - Mortgage approvals H1 2024: 54.3 percent.
  - Short-term rentals: 7,265 listings as of August 2024 (up from 4,076 in January 2018); short-term rentals accounted for less than 5 of total dwelling units at end-2022; average daily rate rose 30 percent between 2018 and 2023.
- Recommended options to improve housing affordability:
  - Increase public spending on affordable public housing.
  - Incentivize private investment combined with additional public investment.
  - Ease access to credit to support homeownership.
  - Ensure short-term rental growth does not crowd out residential housing investment.

*Source: IMF staff calculations and analysis (chapter "Determinants of Sovereign Spreads in The Bahamas").*

### 1. Sovereign Spreads: Actual vs Model-Fitted____________________________________________ 4

### 1. Sovereign Spreads: Actual vs Model-Fitted____________________________________________ 4

### A. Introduction
- Spreads on Bahamian sovereign bonds issued on international capital markets increased sharply following the onset of the COVID-19 pandemic, rising by as much as 1,200 basis points—more than five times their pre-pandemic levels.
- Causes cited: substantial rise in debt from pandemic-related expenditures and the impact of a major hurricane that occurred just prior to the pandemic; severe impact on tourism from the global downturn in travel.
- Wider spreads imply higher government borrowing costs and a larger share of fiscal resources devoted to servicing public debt.

### B. Sample and Stylized Facts
- Sample: unbalanced panel of 73 countries classified as emerging and developing economies by the IMF, monthly frequency January 2005 to April 2024.
- Dependent variable: hard-currency sovereign bond spread.
- Explanatory variables: domestic covariates (GDP growth, inflation, government debt, current account balance, reserves, governance index) and global covariates (global growth, MOVE), plus dummies (EMBIG inclusion, IMF program, commodity exporter) and interactions (sovereign rating grades × VIX).
- Data handling: variables not at monthly frequency converted using quarter or year values; Dickey-Fuller and Perron-P tests used for stationarity; governance index, government debt, external debt, dollar index, and global growth detrended using a linear trend.
- Bahamian position in sample:
  - Sovereign spreads near sample median pre-COVID, above 75th percentile during pandemic.
  - One of the largest declines in output during the pandemic, followed by rapid recovery.
  - Public debt increased steadily since 2005; large upward jump in debt-to-GDP in 2020; now among top 25 percent of countries in sample.
  - External debt below sample median.
  - Sovereign credit rating deteriorated over sample period; lowered twice since 2020 and currently standing at a single-B rating.
  - Current account balance tends to be very negative; relatively high level of currency reserves throughout the pandemic.
  - Governance index relatively high, close to 75th percentile, though declining over sample period.

### C. Empirical Model
- Baseline panel regression specification (equation reproduced in source) estimating sovereign spread as function of:
  - Domestic covariates X_it with coefficients β.
  - Global covariates Z_it with coefficients γ.
  - Dummy variables D_it with coefficients δ (including EMBIG inclusion, IMF program, commodity exporter).
  - Interaction terms: sovereign rating grade dummies s_irg_k interacted with VIX, coefficients ω_k.
  - Intercept α and error term ε_it.
- Domestic covariates: GDP growth, inflation, government debt, current account balance, reserves, governance index.
- Global covariates: global growth, MOVE index.
- Rating variable constructed as median of Moody’s, S&P, Fitch; takes values: AAA/AA (combined), A, BBB, BB, B, CCC, CCX (CCX denotes default).
- Sample excludes observations with spreads above 3000 basis points (more than four times the standard deviation above the median).
- Estimation period: January 2005 to April 2024.

### D. Key Empirical Results and Findings
- Model fit:
  - Model fits Bahamian sovereign spread dynamics well up to 2020.
  - In 2020 both actual and model-implied spreads surged by over 400 basis points.
  - From 2021 a gap emerges: model-implied spreads decline more than actual, residual substantial in 2023 but narrowed to 70 basis points by end of sample (April 2024).
- Rating changes:
  - Median rating dropped from “BBB” to “BB” in early 2020, and from “BB” to “B” in late 2022.
  - Fitted spread exhibits discrete jumps at downgrades; spread increases can precede official rating downgrades by several months.
- Cross-country alignment:
  - By 2024 Bahamian spreads aligned with peer economies and model-implied spread.
  - Scatterplots for single-B rated countries (March 2023 and March 2024) show residuals generally small; most countries line up along 45-degree line.
- Statistical significance and directions:
  - Most variables significant at least at the 10 percent level; all coefficients have expected sign (Table 3, column 1).
  - Quantitative associations:
    - A higher domestic GDP growth rate associated with around 13 basis points decline in sovereign spread.
    - A one percentage point higher global growth rate associated with a 28 basis points lower sovereign spread.
  - Sign relationships:
    - Higher inflation and higher government debt associated with higher spreads.
    - Higher current account balance, higher governance index, higher reserves, and being a commodity exporter associated with lower spreads.
    - MOVE index positively related to sovereign spreads.
- Role of global risk aversion and rating interaction:
  - Global risk aversion (VIX) has asymmetric effect: lower rated countries experience disproportionately higher widening when global risk sentiment deteriorates.
  - At median VIX of 17:
    - “A” rated country faces about 19 basis points higher spread than “AAA”/“AA”.
    - “BB” rated country faces about 66 basis points higher spread.
    - “B” rated country (like The Bahamas) faces about 269 basis points higher spread.
    - “CCX” (default) associated with 663 basis points higher spread compared to “AAA”/“AA”.
- EMBIG inclusion:
  - Inclusion in JP Morgan’s EMBIG index associated with lower sovereign spreads.
  - The Bahamas would have had spreads compressed by 56 basis points compared to similar fundamentals if included in EMBIG.

### E. Decomposition of Recent Change (January 2023 to March 2024)
- Total model-implied decline in fitted spread: about 140 basis points.
- Major contributors:
  - Interaction term between sovereign rating and VIX: reduction of over 100 basis points (entirely attributable to decline in VIX since Bahamas rating unchanged).
  - Decline in MOVE index: about 20 basis points reduction.
  - Increase in global growth outlook: about 40 basis points reduction.
- Offsetting domestic factors:
  - Moderation in Bahamian GDP growth and slight decline in reserves imply an increase in fitted spread by about 30 basis points.
- Net depicted in Figure 3 (bar decomposition; basis points scale).

### F. Projected Spreads Scenarios to end-2029
- Two illustrative scenarios constructed using staff baseline projections for domestic variables and staff baseline projections and historical averages for global variables.
- Scenario 1 — Unchanged sovereign rating:
  - Fitted spread declines to about 430 basis points by end of 2029.
  - Reduction of 65 basis points compared to March 2024.
  - Large part of projected reduction (45 basis points) driven by assumed decline in MOVE index to historical averages.
- Scenario 2 — One-notch rating upgrade:
  - Fitted spread declines by an additional 175 basis points relative to unchanged-rating scenario.
  - Implied spread about 260 basis points by 2029, close to pre-pandemic level.
  - Interpretation: rating upgrades often linked to improvements in macro fundamentals and fiscal situation; appropriate domestic policy mix can lead to substantial drops in spreads and funding costs.

*Source: IMF staff calculations and analysis (chapter "Determinants of Sovereign Spreads in The Bahamas").*

### 11.      The reported findings are generally robust to different model specifications. For

### 11.      The reported findings are generally robust to different model specifications. For

### Robustness checks and regression-specification findings
- Baseline model is extended to the specification in equation (2), which adds country fixed effects (푠푠푖) and time fixed effects (휏휏푖), controls X and Z, dummy D, interactions ∑휔푘 g푖푖푘 x VIXX, and error term 휖푖푖.
- Eight additional specifications are estimated (Table 3). Key qualitative findings:
  - Coefficients are generally stable across specifications, including when country or time fixed effects are added.
  - The dummy for inclusion in the EMBI is notably more negative in specifications with country fixed effects (5, 7, 9), and when sovereign credit ratings are excluded as covariates (8, 9). Interpretation: limited within-country variation of EMBIG inclusion may drive this result.
  - The dummy for an active IMF program loses significance in some specifications that include country fixed effects (5, 9), and changes sign in the specification excluding sovereign ratings (8). Interpretation: countries with an active IMF program tend to have lower sovereign ratings, and lower sovereign ratings are associated with higher sovereign spreads.
  - For specifications excluding sovereign ratings (8, 9), the coefficient on public debt is generally larger, by as much as a factor of ten, while adjusted R-squared declines.

### Key regression coefficient patterns (from Table 3)
- GDP Growth coefficients (all preserved exactly):
  - (1) -13.23***  (se 1.22)
  - (2) -14.25***  (se 1.26)
  - (3) -17.12***  (se 1.72)
  - (4) -13.58***  (se 1.22)
  - (5) -8.71**    (se 3.99)
  - (6) -13.05***  (se 3.58)
  - (7) -11.29***  (se 3.85)
  - (8) -18.13***  (se 1.51)
  - (9) -13.67**   (se 5.99)
- Inflation coefficients:
  - (1) 7.03***  (se 0.47)
  - (2) 6.28***  (se 0.55)
  - (3) 6.90***  (se 0.65)
  - (4) 6.78***  (se 0.46)
  - (5) 6.18***  (se 1.42)
  - (6) 6.88***  (se 1.39)
  - (7) 3.55***  (se 1.33)
  - (8) 10.57*** (se 0.67)
  - (9) 8.86***  (se 2.23)
- CA Balance coefficients:
  - (1) -2.70***  (se 0.44)
  - (2) -2.93***  (se 0.42)
  - (3) -2.82***  (se 0.44)
  - (4) -2.53***  (se 0.42)
  - (5) -1.32     (se 1.81)
  - (6) -2.42**   (se 1.22)
  - (7) -1.64     (se 1.97)
  - (8) -2.43***  (se 0.44)
  - (9) 1.94      (se 1.49)
- Governance coefficients:
  - (1) -2.94***  (se 0.19)
  - (2) -2.50***  (se 0.23)
  - (3) -2.31***  (se 0.26)
  - (4) -2.92***  (se 0.22)
  - (5) -0.70     (se 4.07)
  - (6) -3.28***  (se 0.87)
  - (7) -0.59     (se 4.25)
  - (8) -8.25***  (se 0.19)
  - (9) -4.20     (se 4.63)
- Debt (public debt share of GDP) coefficients:
  - (1) 0.35*    (se 0.18)
  - (2) 0.10     (se 0.17)
  - (3) 0.13     (se 0.19)
  - (4) 0.35**   (se 0.17)
  - (5) 4.76***  (se 1.31)
  - (6) 0.52     (se 0.70)
  - (7) 5.93***  (se 1.40)
  - (8) 2.69***  (se 0.19)
  - (9) 5.87***  (se 1.70)
- Reserves coefficients:
  - (1) -4.12***  (se 0.32)
  - (2) -4.14***  (se 0.27)
  - (3) -3.91***  (se 0.30)
  - (4) -4.08***  (se 0.30)
  - (5) -3.19*    (se 1.86)
  - (6) -4.53***  (se 1.36)
  - (7) -3.41*    (se 1.87)
  - (8) -6.37***  (se 0.31)
  - (9) -3.60*    (se 1.92)
- Comex (commodity exporter dummy) coefficients (with se):
  - (1) -62.33***  (8.80)
  - (2) -71.72***  (8.38)
  - (3) -66.71***  (8.48)
  - (4) -56.80***  (8.99)
  - (6) -64.95**   (25.51)
  - (8) -63.05***  (7.55)
- IMF dummy coefficients:
  - (1) -32.76***  (7.91)
  - (2) -37.75***  (7.56)
  - (3) -29.73***  (8.53)
  - (5) -47.30     (33.14)
  - (6) -29.57     (25.83)
  - (7) -40.69     (30.72)
  - (8) 57.60***   (7.15)
  - (9) -16.56     (32.04)
- EMBIG dummy coefficients:
  - (1) -55.80***  (5.70)
  - (2) -54.13***  (6.47)
  - (3) -36.07***  (7.14)
  - (5) -98.44**   (43.30)
  - (6) -47.66     (42.17)
  - (7) -141.80*** (40.50)
  - (8) -142.45*** (7.36)
  - (9) -125.07*** (33.12)
- Global covariates and interactions of ratings with VIX show statistically significant coefficients for MOVE, VIX, and rating×VIX interactions (A, BB, B, CCC, CCX with VIX), with magnitudes preserved exactly in Table 3.
- Country FE and Year FE inclusion across columns:
  - Country FE: Y in columns (5), (7), (9)
  - Year FE: Y in columns (6), (7), (9)
- Observations: 11,916 in all columns.
- Adjusted R2 by specification:
  - (1) 0.60
  - (2) 0.61
  - (3) 0.55
  - (4) 0.60
  - (5) 0.38
  - (6) 0.58
  - (7) 0.30
  - (8) 0.45
  - (9) 0.12
- Note: *, **, *** indicate statistical significance at the 10%, 5%, and 1% levels; standard errors (se) are robust standard errors.

### Data, variable definitions, and sample statistics
- Key variable definitions and sources (preserved text):
  - Spread: Sovereign bond spread; Unit: Basis points; Source: Bloomberg Finance L.P., JPMorgan, IMF Spread Monitor; Reference rate based on US Treasuries or index provider default.
  - GDP Growth: Real GDP growth; Unit: Percent, yoy; Source: IMF, Consensus Economics.
  - Inflation: CPI inflation rate; Unit: Percent, yoy; Source: IMF, Consensus Economics.
  - CA Balance: Current account balance (share of GDP); Unit: Percent; Source: IMF, National sources.
  - Governance: Governance index; Source: World Bank Worldwide Governance Indicators; Simple average of World Bank regulatory quality and government effectiveness.
  - Debt: Public debt (share of GDP); Unit: Percent; Source: IMF.
  - Ext. Debt: External public debt (share of GDP); Unit: Percent; Source: Haver, BIS, IMF.
  - Reserves: Currency reserves (share of GDP); Unit: Percent; Source: Haver, IMF.
  - Sovereign credit rating: Rating scale AAA/AA, A, BBB, BB, B, CCC, CCX; Source: Bloomberg Finance L.P., S&P, Fitch, Moody’s; Median of S&P, Fitch, and Moody’s. Grades AAA and AA combined. CCX denotes rating grades below CCC.
  - Comex: Dummy for commodity exporter; Source: IMF.
  - EMBIG: Dummy for inclusion in JP Morgan EMBIG index; Source: Bloomberg Finance L. P.
  - IMF: Dummy for active IMF lending program; Source: IMF MONA; Includes SBA, SCF, EFF, ECF, and respective predecessor programs. Both completed and cancelled programs.
  - Dollar: Dollar index (DXY); Source: Bloomberg Finance L. P.
  - Global Growth: Global GDP growth; Unit: Percent, yoy; Source: IMF, Consensus Economics; Forward looking weighted average using current year and 1-year forward; uses consensus surveys for US and China (60/40).
  - VIX: Chicago Board Options Exchange's Volatility index; Source: Bloomberg Finance L. P.
  - MOVE: Merrill Lynch Option Volatility Estimate index; Source: Bloomberg Finance L. P.
- Summary statistics (Table 2) — Domestic covariates (observations, mean, standard deviation, minimum, median, maximum):
  - Spread: 12639 obs; Mean 476.6; Standard Deviation 660.2; Minimum -3.9; Median 305.4; Maximum 12481.9
  - GDP Growth: 12639 obs; Mean 3.6; Standard Deviation 3.0; Minimum -23.5; Median 3.7; Maximum 21.6
  - Inflation: 12639 obs; Mean 6.1; Standard Deviation 8.3; Minimum -2.5; Median 4.3; Maximum 201.9
  - Debt: 12611 obs; Mean 52.8; Standard Deviation 26.3; Minimum 0.0; Median 50.0; Maximum 158.1
  - CA Balance: 12639 obs; Mean -1.6; Standard Deviation 7.5; Minimum -46.2; Median -2.0; Maximum 37.8
  - Ext. Debt: 12619 obs; Mean 54.4; Standard Deviation 38.8; Minimum 0.0; Median 42.9; Maximum 243.1
  - Reserves: 12639 obs; Mean 17.6; Standard Deviation 9.9; Minimum 0.0; Median 15.5; Maximum 81.8
  - Governance: 12639 obs; Mean 47.3; Standard Deviation 18.6; Minimum 4.2; Median 48.1; Maximum 88.4
- Summary statistics — Global covariates (observations, mean, standard deviation, minimum, median, maximum):
  - Dollar: 232 obs; Mean 89.1; Standard Deviation 9.3; Minimum 72.4; Median 89.8; Maximum 110.8
  - Global growth: 232 obs; Mean 4.4; Standard Deviation 1.0; Minimum 1.4; Median 4.5; Maximum 6.5
  - VIX: 232 obs; Mean 19.4; Standard Deviation 7.6; Minimum 10.1; Median 17.0; Maximum 51.9
  - MOVE: 232 obs; Mean 84.3; Standard Deviation 30.5; Minimum 42.7; Median 75.2; Maximum 209.9
- EMBIG inclusion and IMF program dummies (Table 2):
  - EMBIG: 12639 observations; Obs. with d=1: 11721; Countries with at least one d=1: 68; Countries with at least one d=0: 15; Total nr of countries: 73
  - IMF: 12639 observations; Obs. with d=1: 2641; Countries with at least one d=1: 40; Countries with at least one d=0: 73; Total nr of countries: 73

### Climate-change risks, channels, and quantitative approach (Annex: Assessing climate change risks)
- Key contextual facts:
  - 80 percent of The Bahamas' land area is less than 1.5 meters above sea level.
  - Tourism contributes about 50 percent to GDP.
  - Hurricane Dorian (2019) caused $3.4 billion in damage to physical assets, amounting to 25 percent of The Bahamas' GDP.
  - Over the past decade, natural disasters have averaged 3.2 percent of GDP in annual damages.
- Climate projections and impacts:
  - Acevedo (2016) projects hurricane-related damages in The Bahamas will increase by 31 percent under 3°C warming and by 42 percent under 4.3°C.
  - Sea levels are projected to rise by nearly 0.5 meters by the end of the century; this would place about 41 percent of the land and 22 percent of the population below sea level.
  - Without adaptation, a 0.5 m sea level rise could cause the complete loss of up to half of sandy beaches located near hotel infrastructure (Spencer et al., 2022).
- Exposure heterogeneity across islands:
  - Smaller islands (Exuma, Andros, Cat Island, and other smaller family islands) contribute 5.1 percent of national income and 7 percent of the population, but may experience disproportionately higher income losses and have lower adaptation capacity compared to New Providence and Grand Bahama.
- Quantification strategy:
  - Dual-pronged approach: separate estimation of economic losses from (i) intensification of hurricanes (recurring disasters) and (ii) permanent land inundation (slow onset sea level rise); combined to produce total potential output losses.
  - Slow-onset (land inundation) losses estimated via two-sector output accounting framework that explicitly incorporates natural capital; uses Spencer et al. (2022) sandy beach erosion projections and assumes exposed physical assets proportional to exposed population.
    - Conservatively assumes half of projected sandy beach losses from Spencer et al. (2022).
    - Assumes 50 percent of exposed assets are destroyed due to sea level rise in less developed islands and 20 percent in more developed ones. Similar assumption for government infrastructure.
    - Traded sector relies on natural capital as a complement; natural capital deterioration in traded sector modeled solely via sandy beach erosion in this approach.
  - Hurricane impacts modeled with a Markov-switching dynamic general equilibrium model (Fernandez-Corugedo, Gonzalez, and Guerson, 2023) that switches between disaster and non-disaster states; hurricanes destroy a share of private and public capital stocks, lowering output while capital is rebuilt.
    - Production function distinguishes standard public capital (vulnerable to disasters) and resilient public capital (remains intact during disasters); adaptation increases share of resilient capital.
  - The modeling setup:
    - Nontraded output and traded output production functions include private capital, public capital, labor, and natural capital with parameters 훼퐾, 훼퐺, 훾, 휃, 휔 preserved from the text; natural capital is complementary in the traded sector (휃 negative specifies complementarity in text).
  - The analysis translates a given risk reduction achieved through adaptation into avoided output losses but does not evaluate cost-effectiveness of specific adaptation strategies.

*Source: IMF staff calculations.*

### 1.      Both frameworks are calibrated using recent data,  capturing salient features of The

### Both frameworks are calibrated using recent data, capturing salient features of The Bahamas’ economy.

### Calibration and key assumptions
- Hurricanes inflict damages amounting to 7.3 percent of GDP per event, reflecting the average losses from past major disasters.  
- Damages are assumed to increase in line with projections in Acevedo (2016).  
- The probability of hurricane occurrence matches historical patterns and is assumed to remain constant with climate change.  
- Resilient investments are estimated to cost the government 25 percent more than standard investments.  
- Illustrative resilience assumption: protective measures ensure 80 percent of the exposed private physical assets are safeguarded against permanent inundation; a similar share of government infrastructure is made resilient; protection of natural capital assumes 80 percent preservation from climate change impact.  
- Staff estimate: the prevailing natural disaster profile reduces current potential output level by about 8 percent, relative to a counterfactual without tropical storms.

### Adaptation scenarios (illustrative)
- Scenario 1: No adaptation.
  - Neither the public nor private sector takes preemptive measures.
  - Capital losses are substantial due to permanent inundation and repeated hurricanes.
- Scenario 2: Protection of physical assets.
  - Public sector invests in safeguarding people, private assets, and government infrastructure from permanent submersion via hard protection measures, nature-based solutions, and planned retreat.
  - Government ensures its infrastructure is resilient to natural disasters.
  - Public investment in adaptation yields a long-term increase in the level of GDP of around 5.5-6.8 percent once physical assets are secured.
- Scenario 3: Protection of physical assets and natural capital.
  - In addition to Scenario 2 measures, government invests in preserving natural capital (breakwater construction, coral reef and mangrove protection, beach nourishment).
  - Expected to reduce long-run traded output losses further compared to Scenario 2.

### Projected potential output losses and sectoral impacts
- By 2100, combined with losses from permanent inundation, additional potential output losses could be:
  - between 8.3 percent of GDP under intermediate GHG emission scenario (SSP2-4.5)
  - and 11.1 percent of GDP under a severe global warming scenario (SSP5-8.5).
- Under Scenario 2 (physical protection), traded output losses remain sizeable because natural capital remains vulnerable.
- Under Scenario 3 (including natural capital preservation), potential output losses could be reduced from 2.8-4.3 percent of GDP under Scenario 2 to 1.2-2.3 percent of GDP under Scenario 3.
- The export-oriented traded sector is particularly vulnerable due to reliance on natural capital (sandy beaches, coral reefs, arable land, fish stocks).
- Comprehensive resilience-building could help stabilize output during climate shocks and reduce uncertainty in fiscal and foreign exchange income streams.

### Table 1 — Calibrated parameters and initial values (selected entries)
- Potential GDP growth rate: 1.5  
- Private investment to GDP: 23  
- Private consumption to GDP: 59  
- Public investment to GDP: 2  
- Public consumption to GDP: 20  
- Net exports to GDP: -4  
- Depreciation rate of private capital: 5  
- Depreciation rate of non-resilient public capital: 7.5  
- Depreciation rate of resilient public capital: 3  
- Public investment efficiency: 50  
- Share of traded goods in total output: 40  
- Share of domestic demand for traded output: 10  
- Share of imported capital goods in total investment: 50  
- Share of imported goods in domestic consumption: 27  
- Annualized (implicit) interest rate on domestic debt: 4  
- Share of hotel rooms in New Providence and Grand Bahama: 70  
- Share of natural capital resistant to climate change on New Providence and Grand Bahama: 70  
- Share of natural capital resistant to climate change on other islands: 50  
- Share of exposed physical assets resilient to climate change on New Providence and Grand Bahama: 70  
- Share of exposed physical assets resilient to climate change on other islands: 60  
- Share of exposed public capital resilient to SLR on New Providence and Grand Bahama: 75  
- Share of public capital resilient to SLR on other islands: 30  
- Share of physical capital located in New Providence and Grand Bahama: 69.4  
- Traded sector: Labor share in production: 42.7  
- Traded sector: Capital share in production: 18.3  
- Traded sector: Elasticity of output to public capital: 0.15  
- Traded sector: Natural capital share in the production function: 39  
- Non-traded sector: Elasticity of output to public capital: 0.15  
- Non-traded sector: Labor share in production: 60  
- Non-traded sector: Capital share in production: 40

### The electricity sector — challenges
- Low energy security clustered around: (i) sustainability, (ii) elevated costs, and (iii) aging infrastructure.  
- Renewable Energy (RE) participation: about 2 percent of the electricity mix.  
- Regional RE installed capacity averages in 2022: Jamaica 16 percent, Barbados 21 percent, Dominican Republic 29 percent, LAC 57 percent.  
- Average solar irradiation: 5.3-kilowatt hours per square meter per day (kWh/m2/day).  
- Electricity price in The Bahamas: US$0.28 kWh, above the Caribbean average of US$0.24 kWh.  
- Generation replication across ~30 inhabited islands increases costs and complicates cost recovery.  
- More than 80 percent of firms experienced power-outs in 2019/20 (compared to 67 percent in the Caribbean); outages averaged 3 times per month and created a loss of annual sales of 2.5 percent (vis-à-vis 1.7 percent in the Caribbean).  
- T&D technical losses: 12.3 percent of electricity produced; commercial losses (unbilled): 9.6 percent.  
- Energy intensity (2021): 2.87 Mega Joule (MJ) per 2017 USD PPP GDP — an increase of 42 percent over the last 15 years.  
- Institutional capacity: prior to 2024, sector largely managed by Ministry of Public Works; Utilities Regulation & Competition Authority (URCA) is independent regulator; Ministry of Energy and Transport (MET) created in 2024 and will need to build capacity rapidly.

### 2024 Electricity Reform Plan — components and projected impacts
- GOBH targets in updated NDCs (2022):
  - 30 percent reduction of GHG emissions compared to Business-as-Usual by 2030.
  - 35 percent of vehicle purchases electric and 15 percent hybrid by 2030.
  - At least 30 percent of renewables in the electricity mix by 2030.
- Reform aims: repair T&D, partially replace heavy-oil with LNG, expand RE production.
- BPL will enter PPAs with independent power producers and solar providers in New Providence and the Family Islands.
- MET anticipates reforms will reduce energy production cost to 22.5 cents per kWh, potentially saving up to $170 million annually, and will not require direct financing from the central government.
- Reform completion target: December 2025.
- Reform components:
  - 30 percent renewable energy target by 2030 and 120 MW of solar energy:
    - 50 MW utility-scale power plants in the Family Islands.
    - Expansion in New Providence targeting 70 MW solar generation (25 MW used to charge existing batteries).
    - Independent power producers required to have at least 50 percent Bahamian ownership.
  - Integrating LNG into energy production: convert diesel systems to LNG and build combined cycle LNG plant in New Providence via PPA.
  - Modernization of T&D in New Providence:
    - Estimated $130 million needed for upgrades.
    - Creation of Special Purpose Vehicle (SPV) Bahamas Grid Company Ltd; BPL will own 40 percent of SPV shares.
    - SPV will allocate .0025 cents per Bahamian dollar of its revenue to pay down legacy debt and another .0025 cents to a hurricane fund for New Providence disasters.
  - Rate adjustments (implemented July 2024):
    - Fuel charge tariff reduced by 2.5 cents for the first 800 kWh and increased by 1.5 cents for usage above 800 kWh.
    - Adjustment should benefit 58 percent of residential consumers and 56 percent of regular commercial consumers.
    - No base rate charged for the first 200 kWh used by residential consumers.
    - Commercial base tariff will rise from 8.7 to 10 cents per kWh for the first 900,000 units and from 6.2 to 9 cents per kWh for usage above 900,000 kWh.
- Projected macroeconomic benefits of reform over the medium-term:
  - Narrow the current account deficit.
  - Reduce vulnerability to commodity price shocks.
  - Boost growth.
  - Significantly reduce CO2 emissions.
- Implementation note: power projects should have a clear delineation of risk sharing between the private and public sector.

*Source: IMF staff calculations.*

### Box 1.   Bahamas Power and Light (BPL)

### Box 1.   Bahamas Power and Light (BPL)

### Background
- MET has overall responsibility for the energy sector in The Bahamas and oversees the government owned public utility BPL, which is a vertically integrated utility company.
- BPL operates 29 plants with a total generating capacity of 532 Megawatts (MW) and serves 115,000 customers.
- BPL was established by the Electricity Act of 2015 and has operational autonomy with its own Board of Directors.
- The electricity sector is regulated by the Utilities Regulation & Competition Authority (URCA).
- From a legislative perspective BPL is a fully owned subsidiary of Bahamas Electricity Corporation (BEC), whose sole shareholder is the Government of the Bahamas.

### Infrastructure, generation mix, and reliability
- BPL needs an estimated USD 500 million to refurbish and replace its aging generation, transmission, and distribution infrastructure.
- About 10 percent of substations require replacement, and another 10 percent need upgrades or repairs.
- Fuel constitutes over half of BPL’s operating expenses.
- In New Providence, the dispatch process does not fully adhere to a merit order due to challenges with handling heavy fuel oil and obligations with independent power producers.
- Benchmarking reliability indicators shows that BPL customers experience more frequent interruptions compared to other utilities, though these interruptions are shorter in duration.
- According to the URCA audit, reliability has improved since 2021 but still falls short.

### Financial performance and debt
- BPL has been operating at a loss since 2022 and has a large stock of debt.
- Net profit margin was -20 percent and -24 percent in FY2022 and FY2023, respectively.
- Since FY2017, BPL’s net profit margin has averaged -2 percent.
- URCA audit highlighted that BPL’s debt service coverage ratio has been negative for the last 2 years, reaching as low as -8.8 in 2022 and -2.5 in 2023.

### Financial planning, management, and strategy
- BPL does not have forward-looking and automated financial planning and management processes.
- BPL has a 100-day strategy plan (as of November 2023), but it does not include details and specificity.
- Lack of detailed planning makes it difficult to project capital expenses accurately, develop a strategy to optimize asset utilization, and makes risk management and planning reactive rather than preventative.

### Family Islands operations
- BPL’s operation in the Family Islands is small, outdated, and relatively inefficient.
- Average age of generation units in the Family Islands is 23 years, more than double the average age of units in New Providence (10 years).
- Continuous generation has been difficult in these areas due to skill and capital shortages and the overburdened staff.

### Potential technical and cost-saving opportunities
- Exploring generation options on some Family Islands could help cut costs.
- Hybrid systems combining solar, battery storage, and diesel could potentially power entire islands more economically than using oil.
- BPL has opportunities to lower its generation costs but currently lacks a cost-effective generation and system plan.

### Metering, billing, and collections
- Effective metering and billing processes are in place, but BPL can improve in ensuring timely payment from customers.
- BPL’s average collection period is 135 days, an increase of 45 percent since 2017.
- Collection from government customers is much lower compared to other customer categories.
- BPL does not enforce disconnection and arrears control process on government customers, which has led to receivables from government customers of about USD 89 million.

*Prepared from the URCA audit and URCA 2023 Annual Report and 2024 Annual Plan as presented in the source content.*

### 19.      While fiscal risks cannot be completely eliminated, governments can put in place

### While fiscal risks cannot be completely eliminated, governments can put in place measures to minimize their occurrence and costs.

### Fiscal risk mitigation for PPPs
- Approaches to limiting the buildup of risks include:
  - A leading role for the Ministry of Finance as reviewer of the PPP proposal at various stages of a “gateway process”, including reviewing the project’s proposal before it has gathered political momentum and granting approvals where necessary.
  - Clearly defining the risk-sharing agreement. The risks to be borne by each party in the PPP should be clearly defined. As much as is practical, the public sector should take on only risks over which it has control or exerts some influence and has the capacity to manage that risk (e.g., Jamaica), with partial guarantees possible in cases where all partners choose to share risks. This is made easier with the introduction of PPP laws or standardized contracts.
  - Ensuring accountability for participating entities. Executing agencies should be accountable for their roles and the management of risks in the project cycle, provided they are given sufficient autonomy to manage those risks.
  - Clearly defining who has the authority to make payments to permit the government to make timely and legal payments to contractors.
  - Imposing limits on the scale and scope of the PPP program, including:
    - annual limits on the size of PPP payments,
    - constraints on the size of new commitments or guarantees made during the year,
    - ceilings on the size of outstanding PPP commitments.
  - An alternative approach: treat PPPs as budgetary expenditure (subject to the applicable accounting rules), with commensurate impacts on public sector debt, improving transparency and the government’s capacity to manage costs and risks.

### Authorities’ mitigation measures for the announced electricity sector reform
- Measures taken by the authorities to mitigate PPP-related fiscal risks for the announced reforms:
  - The Ministry of Finance plays a central role in the negotiations of the proposed reform.
  - The government is not granting any guarantees to private sector participants.
  - Some of the PPAs between the government and the private sector are expected to include key performance indicators for the private participant, allowing for penalties if performance commitments are not met.

### Expected economic and environmental impacts of the electricity sector reform
- Reform objectives and targets:
  - Shift the mix of electricity generation to 30 percent solar energy and around 50 percent LNG by 2030.
  - Modernize transmission and distribution of electricity.
  - Reduce the cost of electricity for low-income households.
- Estimated macroeconomic impact:
  - Using the IMF’s RE-BOP tool, staff estimates that successful implementation could lift real GDP growth by 0.5 percentage points by 2035.
- External balance and emissions:
  - After initial grid modernization investments are completed, the transition to renewable energy and LNG would reduce the demand for imported fossil fuels, narrow the current account deficit, and reduce The Bahamas’ CO2 emissions from the electricity sector over the long-term.

### Managing fiscal risks to maximize returns from reform
- Key principles to minimize fiscal risks from PPPs and the electricity reform:
  - Ensure a leading role for the Ministry of Finance at each project stage.
  - Clearly define roles and equitable sharing of risks between private and public entities.
  - Enforce limits on the size of guarantees granted to the private sector.
  - Use transparent and competitive selection of experienced private sector counterparts with strong financial buffers and the capacity to attract private capital to minimize costs and reduce the risk of unnecessary public sector intervention.

### Housing affordability in The Bahamas — Background and trends
- Population and household changes:
  - The population of The Bahamas expanded by 13 percent between 2010 and 2022.
  - Growth particularly strong in New Providence (20 percent), Acklins (20 percent), Berry Islands (24 percent) and Bimini (19 percent).
  - The average household in New Providence increased from 3.5 to 3.7 persons between 2010 and 2022.
- Housing supply trends:
  - An assessment in 2000 estimated that housing stock would have to rise by 2,378 units annually between 2000 and 2011.
  - Residential construction completions reached 607 completions in 2023, down from a peak of 1,705 in 2006.
- Price and wage dynamics:
  - Between 2012 and 2022, the implied price of Real Estate, Owner Occupied and Actual Rents activities increased by 14 percent, compared to just 2 percent for employee compensation per capita.
  - 58 percent of the poor live in privately-rented housing subject to changes in annual rental rates, compared to 34 percent across the country.
- Mortgage and credit constraints:
  - The number of new residential mortgage commitments for single dwellings and duplex and row dwellings have trended downward since 2008.
  - Mortgage applications recorded the lowest approval rate of all credit categories in H1 2024, standing at 54.3 percent.
  - Common reasons for mortgage denials included low credit scores, banks’ internal lending policy constraints, underemployment, prior delinquency, higher debt service ratios, insufficient working history in the current job, inability to verify income, and inadequate funds for a down payment.
- Short-term rental market growth:
  - As of August 2024, there were 7,265 listings of short-term rentals across The Bahamas—up from 4,076 at January 2018.
  - Growth strongest for properties classified as apartments, condos, or lofts, with listings more than doubling over the period.
  - Short-term rental properties accounted for less than 5 of the total stock of dwelling units in The Bahamas at the end of 2022.
  - Average daily rate on short-term rentals rose 30 percent between 2018 and 2023, compared with a 2 percent rise in the implied price of Real Estate, Owner Occupied and Actual Rents activities over the same period.
  - Hotel room capacity declined by 9 percent between 2018 and 2023.

### Policy options and existing initiatives for housing affordability
- Existing initiatives:
  - The Department of Housing administers the Guaranteed Loan Programme with financial institutions; Bahamas Mortgage Corporation (BMC) has traditionally been the primary user of this facility for homes costing up to $250,000.
  - The government removed the duty on building materials to support rebuilding in the Family Islands after natural disasters.
  - The government has constructed public rental units and senior citizen units across the country.
  - As of September 2024, 11 percent of outstanding residential mortgages for existing structures were held by the BMC.
- Fiscal stance and comparisons:
  - In 2021, central government spending on housing and community amenities declined to 0.02 percent of GDP, compared to 1.3 percent for an average of Caribbean peers, and down from 0.11 percent of GDP in 2008.
- Recommended options to improve housing affordability:
  - Increase public spending on affordable public housing to address demand pressures in New Providence and shortages in Grand Bahama and the Family Islands.
  - Incentivize private investment combined with additional public investment in affordable housing.
  - Ease access to credit for residents to support increased homeownership.
  - Ensure that initiatives encouraging growth in short-term rentals do not crowd out investments in residential housing.

*Source: 1bhsea2025002-print-pdf - 19.      While fiscal risks cannot be completely eliminated, governments can put in place*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2025/english/1bhsea2025002-print-pdf.pdf_