## cr18322

## Source details

**Canonical URL:** [cr18322](https://www.imf.org/-/media/files/publications/cr/2018/cr18322.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2018/cr18322.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2018/cr18322.pdf.json)

---

### Selected Indicators, 1964-2017 — Introduction and Background
- Purpose:
  - Compare past economic developments in the islands (São Tomé and Príncipe, Cabo Verde, Mauritius, Seychelles).
  - Draw lessons for STP’s relative under-performance and recommend policies to foster future growth.
- Key contextual points:
  - STP’s GDP per capita has essentially stagnated since independence, while it has tripled or quadrupled in peers.
  - All four islands evolved into plantation economies (except Cabo Verde’s arid climate).
  - Independence dates: Mauritius 1968; Cabo Verde and STP 1975; Seychelles 1976.
  - Political/institutional inheritance differences: Mauritius retained ties to British Commonwealth and democratic values; Cabo Verde and STP transitioned with little advance preparation for self-rule.
- Role of government at independence:
  - Protectionist inward-oriented policies and strong government role were common; Mauritius remained more market-oriented while Cabo Verde, Seychelles, and STP adopted single-party socialist regimes.

### Conditions at Independence — STP’s initial endowments and institutional weaknesses
- Common constraints: small size, remoteness, underdevelopment, limited resources; agriculture-based economies with few export goods and high import dependence.
- STP-specific features:
  - Cocoa plantations: accounted for 90 percent of farmed land and 2/3 of employment.
  - Departure of most Portuguese at independence removed skilled workers, managers and financial capital; infrastructure rudimentary.
- Initial conditions ranking (1–4, 4 most favorable for growth) — selected indicators and STP’s positions:
  - Size: STP = 2.
  - Remoteness: STP = 3.
  - Natural resources: STP = 1.
  - Income per capita: STP = 4.
  - Life expectancy: STP = 1.
  - Human capital: STP = 1.
  - Physical and financial capital: STP = 1.
  - Democratic values: STP = 1.
- Institutional indicators (post-1996): STP scores lowest among the four islands on governance indices cited.

### Developments Since Independence — Mismanagement, macro volatility, and weak investment
- Post-independence policy actions and consequences:
  - Nationalization, centralized planning, price controls, subsidies, nepotism, inexperienced appointees.
  - Plantation production dropped by almost half due to mismanagement.
  - SOEs registered large losses; commerce and services contracted as Portuguese firms exited.
  - Lax fiscal policy financed by external and domestic credit (including central bank financing) led to rising inflation, deficits, and debt.
- 1990s and 2000s:
  - Political regime shift to multi-party democracy in 1991; persistent low capacity, clientelism, poor governance.
  - Political instability: 17 changes in government between 1991 and 2014 (compared to 3 in Cabo Verde, 5 in Mauritius, and 3 in Seychelles).
  - Income per capita declined in STP during 1990–2000 despite substantial external assistance.
  - Private investment sparse in the 1990s; tourism activity absent/measurable in that period.
- Fiscal and revenue features:
  - Tax revenue meagre and cyclical; dependence on external aid; constrained social spending and capital accumulation.

### Key Findings and Implications (Selected)
- Structural constraints: low human capital, weak physical and financial capital, poor governance, limited natural-resource advantages relative to peers.
- Institutional development and political stability are critical for sustained growth — Mauritius, Seychelles, and Cabo Verde retained technical know-how and achieved higher private investment.
- Policy-relevant observations:
  - Nationalization + weak institutions + nepotism → production declines and SOE losses.
  - Lax fiscal/monetary policy and connected lending → macroeconomic volatility (rising inflation, debt, currency depreciation).
  - Political instability (17 changes in government, 1991–2014) undermined reform ownership and deterred private investment.
  - Dependence on external aid and cyclical revenues constrained social spending and human capital accumulation.

*Prepared by Gabriel Srour; July 9, 2018 — Democratic Republic of São Tomé and Príncipe, International Monetary Fund.*

### Reforms since the 2000s — Macroeconomic outcomes, fiscal structure, and recommendations
- Reform chronology and outcomes:
  - Reforms in early 2000s after crisis and fraud cases; HIPC initiative spurred action but implementation slow.
  - Exchange rate stabilized; inflation reduced from 1990s peak (but rose again after global financial crisis).
  - Growth surged mid-2000s driven by government expansion financed by oil exploration bonuses and FDI in tourism and banking.
  - HIPC-related debt forgiveness cut debt to about 60 percent of GDP in 2007.
  - Anti-corruption/governance measures delayed (new anti-corruption law approved in 2012).
- Macroeconomic indicators and vulnerabilities:
  - GDP growth: steady around 4 percent.
  - Inflation: declined to about 5 percent, anchored by the dobra peg to the euro since 2010.
  - Domestic primary deficit: brought close to 2 percent of GDP.
  - Public debt: high at 70 percent of GDP.
  - Cross-arrears in the economy: 60 percent of GDP in unconsolidated terms.
  - Vulnerabilities: fiscal slippages around elections; large contingent liabilities from SOEs; weak public financial management and pervasive cross-arrears.
- Revenue, fiscal structure, and comparative 2016 metrics (Percentage of GDP):
  - Total revenue and grants: STP 27.7; Cabo Verde 27.0; Mauritius 21.1; Seychelles 37.9
  - Total revenue: STP 14.4; Cabo Verde 24.2; Mauritius 20.5; Seychelles 36.6
  - Tax revenue: STP 12.1; Cabo Verde 19.8; Mauritius 18.9; Seychelles 32.5
  - Consumption taxes: STP 1.2; Cabo Verde 9.1; Mauritius 6.8; Seychelles 10.5
  - Grants: STP 13.3; Cabo Verde 2.8; Mauritius 0.7; Seychelles 1.3
  - Total expenditure and net lending: STP 31.6; Cabo Verde 30.1; Mauritius 24.7; Seychelles 38.2
  - Primary balance: STP -3.5; Cabo Verde -0.5; Mauritius -1.1; Seychelles 3.4
  - Overall fiscal balance: STP -4.2; Cabo Verde -3.1; Mauritius -3.6; Seychelles -0.4
  - Notes: For STP, 13.9 percent of GDP in capital expenditure are financed by grants or concessional loans; for STP includes HIPC-related social expenditures equal to 0.24.
- Revenue performance:
  - Historical tax revenue average around 13.5 percent of GDP.
  - Current tax revenue around 12 percent of GDP as oil prices are in a down cycle.
  - Comparison thresholds: Sub-Saharan Africa median 15 percent; recent literature minimum threshold close to 13 percent to support growth and development.
  - Consumption tax shortfall: STP’s consumption tax revenue falls between 5½ and 9 percent of GDP below the other three islands.
- Private sector, tourism, and external capital:
  - Tourism receipts: rose from $11 million in 2010 to about $66 million in 2017.
  - Per capita tourism receipts: from $60 to slightly over $300 (2010–2017).
  - In absolute terms 2017 tourism receipts: Cabo Verde and Seychelles close to $400 million; Mauritius close to $1 billion.
  - Per capita comparison: Seychelles almost $4,500 in 2017.
  - Infrastructure plans: airport and seaport improvements planned; financial sustainability should be prerequisite given potential fiscal liabilities.

Recommendations and policy priorities:
- Consolidate macroeconomic stability:
  - Strengthen fiscal discipline at central government; contain vulnerability from public debt and SOE losses.
  - Contain fiscal deficit at sustainable level, instill financial discipline, clear arrears, reduce public debt.
  - Anchor fiscal policy around a medium- to long-term fiscal deficit target to restrain political interference.
- Strengthen revenue mobilization:
  - Implement planned transition to a VAT regime in 2019.
  - Expected VAT gain: about 2 percent of GDP in additional revenue over the medium term.
  - VAT gain alone insufficient; improve tax administration, broaden tax base, improve tax collection.
  - Accelerate strategy to clear cross-arrears (about 60 percent of GDP in 2017 unconsolidated; tax arrears 3.5 percent of GDP).
  - Improve government services to encourage tax compliance and reduce informal economy.
- Expand market access and attract private investment:
  - Continue efforts to expand market and attract private investment.
  - Implement holistic tourism development strategy and better integrate private sector and foreign capital and knowhow.
  - Ensure financial sustainability before undertaking large infrastructure projects.

Conclusion and prospects:
- Determinants of underperformance: country-specific characteristics and weak institutions led to inefficient resource use, delayed reforms, and macro imbalances.
- Signs of opportunity: improved macroeconomic conditions, strengthened institutions, rising human capital, important infrastructure projects, greater policy continuity, and global eco-tourism trends.
- Synthesis: fiscal discipline, revenue mobilization, and an active private sector—particularly in tourism—are key to tapping STP’s growth potential; improved public financial management and multi-annual fiscal frameworks recommended.

*Source: cr18322 - 15.       Reforms implemented since the 2000s improved macroeconomic stability and boosted economic performance*

### Inflation Dynamics — Drivers, empirical results, and policy implications
- Background:
  - Annual inflation 2010–2017: 8 percent (euro area 1.2 percent).
  - Exchange rate peg adopted in January 2010; redenomination on January 1, 2018 (24.50 new dobras per euro).
  - CPI compiled from Agua Grande district; CPI basket: 423 products across 12 categories.
  - Food weight in CPI: 73 percent.
  - Annual inflation spiked to 7.7 percent at end-2017 due to increased import taxes on alcoholic drinks and meat and temporary local food shortages from unusual rainfall.
- Descriptive analysis:
  - Food sub-index main contributor to overall inflation (January 2015–December 2017).
  - Month-to-month CPI changes volatile but usually positive; only 4 negative month-to-month rates during January 2015–December 2017.
  - Monthly inflation rarely exceeded 1 percent; 4 of 5 such cases recorded in 2017.
  - Within food: cereals, vegetables, and fish have largest weights; vegetables and fish particularly important recently due to rainfall.
  - Since mid-2017 locally-produced food (fish, fruits, vegetables) key driver of food inflation; previously imported food was primary driver.
- Time-series regression setup:
  - Data: quarterly 2002 Q4–2017 Q4; decompositions focus on January 2015–December 2017 and December 2016–December 2017.
  - Independent variables: international food price index, local rainfall (percentage deviation from long-run quarterly mean since 1901), M2 growth, domestic primary spending, tourist arrivals growth, international oil price index, NEER, dummy for exchange regime switch (peg).
- Key regression results and interpretations:
  - Peg effect:
    - Peg lowered food inflation by 7.2 percentage points (Peg = -7.240* in Table 2).
    - Peg lowered non-food inflation by 8.7 percentage points (Peg = -8.705*** in Table 3).
  - Food inflation regressions (selected coefficients):
    - IntFoodPrice: 0.240**, 0.560**, 0.287*** across pooled/pre-change/post-change.
    - Rainfall: 0.008, -0.002, 0.016** across pooled/pre-change/post-change.
    - M2: -0.000**, -0.000*, 0.088*** across pooled/pre-change/post-change.
    - NEER: -0.004, 0.808, -0.564*** across pooled/pre-change/post-change.
    - Constants: 18.325***, 24.499***, 6.962***.
  - Non-food inflation regressions (selected coefficients):
    - Oil: 0.018, 0.050, -0.018 across pooled/pre-change/post-change.
    - M2: -0.000, -0.000*, 0.068** across pooled/pre-change/post-change.
    - NEER: 0.049, 0.248, -0.270 across pooled/pre-change/post-change.
    - Tourist arrivals (post-change): 0.029*** in column (3).
    - Constants: 14.779***, 16.322***, 2.731***.
  - Interpretations:
    - Peg has statistically significant dampening effect on both food and non-food inflation.
    - Post-peg period: M2 growth highly significant and positively correlated with food and non-food inflation — evidence of demand-pull linked to fiscal spending.
    - Food inflation significantly affected by international food prices across periods and by local rainfall post-peg — showing dependence on imports and agricultural weather vulnerability.
    - Non-food inflation post-peg positively correlated with tourist arrivals growth — demand pressures on local services and supply bottlenecks.
- Graphical confirmations:
  - Figure 3a: food inflation broadly follows M2 growth post-peg.
  - Figure 3b: food inflation follows international food price growth.
  - Figure 3c: non-food inflation follows M2 growth and tourist arrivals growth.
- Policy implications and recommendations:
  - Support the peg with coherent macro policies; strengthen reserves buffer to support the peg.
  - Continue fiscal consolidation to manage domestic demand; M2 growth linked to public expenditure growth.
  - Increase local supply, enhance productivity, and improve infrastructure:
    - Utilize more local products and decrease dependence on food imports (e.g., rely more on breadfruit and green banana instead of imported rice; enhance local processing).
    - Improve agricultural productivity (greenhouses, sustainable agriculture, weather resilience). Ongoing joint efforts with partners including China and the United Nations noted.
    - Build infrastructure and increase capacity of local services to support tourism; implement tourism development strategy launched January 2018 with World Bank support to alleviate supply-side bottlenecks over time.

*Source: IMF staff calculations.*

### SOEs’ Financial Performance and Risk Spillovers
- Background and core issues:
  - SOEs long loss-making due to administrated prices below cost recovery, elevated operating costs, outdated facilities, large technical/commercial losses, elevated personnel costs.
  - Resulting effects: delayed investments, worsening operations, further losses.
- SOE ownership structure:
  - Four fully state-owned firms: EMAE (electricity and water), ENAPORT (port), ENASA (airport services), Correios (postal services).
  - Four enterprises with minority government stakes (16–49 percent): telecommunications (CST), banking (BISTP), air transportation (STP Air), oil import/distribution (ENCO).
- EMAE specifics:
  - Registered losses since 2011; average retail tariff USD 0.21/KWh; average cost of service USD 0.32/KWh.
  - Energy transmission and distribution losses: 43 percent of energy dispatched (SSA average: 15 percent).
  - Reliance on inefficient thermal turbines and costly fuel imports; large domestic arrears to suppliers.
  - World Bank Power Sector Recovery Project supports rehabilitation, meter installation, and purchasing lower-cost electricity from private operators.
- SOE financial statement highlights (end-2017 or earlier, EUR million):
  - EMAE: Revenue 13.1; Profit -5.9; Assets 45.6; Debt (incl. arrears) 35.2; Arrears (gross) 30.8; Equity 10.4; Debt-to-Equity 3.4.
  - ENAPORT: Revenue 4.2; Profit 0.1; Assets 3.3; Debt (incl. arrears) 1.5; Arrears (gross) 0.4; Equity 1.8; Debt-to-Equity 0.8.
  - ENASA: Revenue 3.4; Profit 0.1; Assets 7.6; Debt (incl. arrears) 3.8; Arrears (gross) 3.3; Equity 3.8; Debt-to-Equity 1.0.
  - Correios: Revenue 0.2; Profit n/a; Assets 1.0; Debt (incl. arrears) n/a; Arrears (gross) n/a; Equity -0.4; Debt-to-Equity n/a.
  - CST (Telecom): Revenue 9.3; Profit 1.1; Assets 18.1; Debt (incl. arrears) 5.5; Arrears (gross) 0.0; Equity 12.6; Debt-to-Equity 0.4.
  - BISTP (Bank): Revenue 9.1; Profit 2.0; Assets 83.5; Debt (incl. arrears) 70.8; Arrears (gross) 0.0; Equity 12.7; Debt-to-Equity 5.6.
  - STP Air*: Revenue 10.5; Profit 0.7; Assets 3.0; Debt (incl. arrears) n/a; Arrears (gross) 0.6; Equity 1.9; Debt-to-Equity n/a.
  - ENCO*: Revenue 40.4; Profit 1.7; Assets 159.6; Debt (incl. arrears) n/a; Arrears (gross) 139.6; Equity 16.2; Debt-to-Equity n/a.
  - Notes: * end-2015; ** end-2016; # end-2017 preliminary figures.
- Key SOE performance indicators (end-2017 or earlier):
  - EMAE*:
    - Operational Cost-to-Revenue: 177.5%
    - Profit Margin: -61.0%
    - Working Capital: -45.3 EUR million
    - Working Capital-to-Equity Capital: -272.8%
    - Cash-to-Short-Term Liabilities: 0.1%
    - Debt-to-EBITDA: undefined
    - Note: * end-2016.
  - ENAPORT:
    - Operational Cost-to-Revenue: 93.6%
    - Profit Margin: 2.2%
    - Working Capital: 0.1 EUR million
    - Working Capital-to-Equity Capital: 6.8%
    - Cash-to-Short-Term Liabilities: 4.9%
    - Debt-to-EBITDA: 4.9%
  - ENASA**:
    - Operational Cost-to-Revenue: 98.0%
    - Profit Margin: 0.2%
    - Working Capital: -1.4 EUR million
    - Working Capital-to-Equity Capital: -36.3%
    - Cash-to-Short-Term Liabilities: 9.9%
    - Debt-to-EBITDA: 15.2%
    - Note: ** end-2017 preliminary figures.
- Key quantitative findings:
  - Operational cost-to-revenue ratios: 94% to 177% (88% to 160% excluding depreciation and amortization).
  - Profit margins: range from 2.2% to -61%.
  - EMAE working capital deficit: STD 1.1 billion at end-2016; nominal capital STD 0.4 billion at end-2016 (would be STD -934 million without capitalizing investment subsidies).
  - Cash-to-short-term liabilities: EMAE 0.1%; ENAPORT 4.9%; ENASA 9.9%.
  - Debt-to-EBITDA critical threshold noted as five; ENAPORT: 4.9; ENASA: 15.2; EMAE: negative/undefined.
  - For EMAE’s debt of STD 1.7 billion to become sustainable, earnings (EBITDA) would have to increase by STD 440 million.
  - ENAPORT plans EUR 2 million local bank loan; ENASA plans EUR 1.5 million.
- Minority SOEs:
  - CST and BISTP profitable with moderate leverage; market shares about 85% (CST) and 50% (BISTP).
  - STP Airways profitable in 2015 but financially weak; government did not participate in recent capital increase.
  - ENCO profitable on accounting basis but burdened by large public sector arrears (ENCO arreas gross 139.6 EUR million).
- Structure of arrears and spillovers:
  - Intricate web of net arrears among SOEs, government entities, private firms, foreign firms, and banks.
  - ENCO built large arrears to its parent company due to public institution arrears.
  - Collection/enforcement weaknesses: EMAE and ENAPORT unable to collect bills from many clients.
  - Repayment plans: ENAPORT and ENASA agreed repayment plans of 5 and 25 years respectively; prior plans stalled as new arrears accumulated.
  - Penalty interest accumulated: about 40% (ENAPORT) and close to 100% (ENASA) of underlying arrears; penalties forgiven only after full clearance.
  - Risks to compliance: low liquidity and prospective bank loans may make adherence to schedules difficult (example: ENASA’s agreed annual repayment would deplete cash-at-hand within three years).
  - Fiscal contingent liabilities concentrated with a few creditors (e.g., ENCO) could undermine government viability and raise future debt premia.
  - Tax arrears equivalent to almost 7 percent of GDP have deprived the state of revenue; contributed to government arrears to suppliers.
  - Current bank exposures to SOEs relatively small; anecdotal evidence of supplier liquidity issues leading to defaults on bank loans.
- Recommendations and short-term measures to address arrears:
  - Identify areas for cost saving in SOEs (e.g., staff costs) and collect more arrears from customers.
  - Explore netting bilateral arrears between SOEs and government where claims considered valid by both parties.
  - Consider forgiving late-payment penalties in line with full scheduled repayment of main arrears per year to ease clearance burden.
  - Ensure SOE borrowing from banks aligns with debt servicing capacity from SOE cash flows to avoid NPLs or additional arrears.
- Conclusion heading present; no substantive text in excerpt.

*Prepared by Torsten Wezel; sources: SOEs’ financial statements; World Bank; São Tomean authorities; IMF staff calculations.*

### Banking System, Central Credit Registry (CRC) Data, and NPL Dynamics
- System-wide NPLs and CRC:
  - NPL ratio of banking system tripled between 2013 and end-June 2017 to 32%, peaking at about 36% in 2016.
  - “Unweighted” NPL ratio (share of number of loans non-performing) substantially lower: less than 10% of GDP.
  - High NPL ratio in value terms driven by poor performance of limited number of large loans.
  - Central credit registry (CRC) operational March 2011; coverage increased from about 50% initially to 96% recently.
  - CRC counts: 269 loans in 2011; 6,787 in 2013; 10,864 in 2017.
- CRC sample and coverage (March 2011–June 2017):
  - 469,823 observations from up to 12 banks.
  - Related to 29,238 loans and 13,039 borrowers (around 7% of total population or 14% of adult population).
- CRC data fields include: bank name, loan origination date, maturity date, loan amount at origination, carrying amount, currency, interest rate, loan status by month, rate type, borrower type, sector/occupation, district, indirect measure of borrower income; no collateral value currently.
- Credit concentration and distribution:
  - Demand-side: top 5% firms received 54% of total corporate credit.
  - Largest individual borrowers received 29% of retail credit.
  - Supply-side: largest bank extended almost half of loans; other two largest banks around 15% each.
  - Geographic concentration: capital district 59.4% of loans.
  - Examples of concentration risk: at one small bank largest borrower accounts for 40% of loan portfolio; at three others that share is 20%; largest 10 borrowers at largest banks account for over 30% of loan volume while representing <1% of borrower count.
- Credit composition, sizes, currencies, and rates:
  - By count/value: individual loans 97% by number; firms represent 71% of total loan amount.
  - Median domestic-currency loan (individuals): STD 35,000 (about USD 1,750); 10% above STD 130,000 (about USD 6,500).
  - Median firm loan: STD 500,000 (about USD 25,000); 10% of firm loans exceed STD 4.9 million (about USD 245,000).
  - GDP per capita: USD 1,772 (end-2017).
  - Median USD loan value: about USD 15,000; median EUR loan value: EUR 30,000.
  - Currency composition: New Dobras (STD) 95.6%; EUR 2.6%; USD 1.8%.
  - Only 5% of borrowers have access to foreign-currency credit.
  - Interest rates up to 54%; average STD loan rate 20%; average foreign currency loan rate 11%.
  - Average maturity: STD loans 2.7 years; foreign currency loans 5.0 years.
  - About 80% of loans fixed rate.
  - Largest and third largest bank extend loans on average twice as large as other two large banks.
  - Two largest banks charge average interest slightly above 20%; third largest bank charges 6%.
  - Largest bank supplies 72% of foreign currency loans.
- Borrower relationships and loan counts:
  - 80% of borrowers have loans from one bank; 16% from two banks; 4% from three or more banks.
  - Loan counts per borrower: 50% have only one loan; 25% have two loans; 14% have three loans; 1.5% have ten or more loans.
  - Four largest borrowers (>30 loans) borrow exclusively from one bank; all four are enterprises.
- Loan classification and NPL definition (BCSTP 2007 categories):
  - Category I Regular: on-time payments.
  - Category II Under supervision: 30–89 days past due.
  - Category III Below standard: 90–179 days past due.
  - Category IV Doubtful: 180–359 days past due.
  - Category V Loss: >360 days past due.
  - Category VI Written off.
  - Standard NPL definition: borrower late by 90 or more days (categories III–V considered non-performing).
- Descriptive NPL statistics (sample period averages):
  - 7% of loans became non-performing.
  - 20% of loans had delayed payment at least 30 days.
  - Loans that become non-performing do so after 15 months on average.
  - 11% of borrowers had at least one loan classified as non-performing.
  - By borrower type: 39% of firms had at least one NPL; 10% of individual borrowers had at least one NPL.
- As of end-June 2017:
  - NPLs represented 32% of total loan amount.
  - Top 10 delinquent borrowers (~1% of borrowers with NPLs) account for 43% of total NPL amount; top 3 represent 22%.
  - Among top 10 with NPLs, nine are firms, mainly wholesale trading sector.
  - One active bank holds 32% of loans classified as loss (unweighted number); a liquidated bank holds another 40%.
- Characteristics of NPLs vs performing loans:
  - NPLs have higher interest rates, longer maturities, greater likelihood of local currency denomination, larger loan amounts, and more often fixed rate.
  - Differences statistically significant for most characteristics.
- Borrower behavior post-default:
  - 45% of borrowers delayed >90 days are able to contract a new loan afterwards.
  - Of new loans post-delay: 41% with same bank; 59% with a different bank.
  - Multiple-bank borrowers: 10% of their loans were NPLs vs 6% for single-bank borrowers.
- Vintage and default trends:
  - Default incidence within first 12 months rose from 0.2% (2012 vintage) to 3.5% (2015 vintage).
  - When weighting by months in default within first year, default rate drops because average loan is around 4 months in default before performing again.
  - BCSTP reported 12-month default rate at 6.7% in December 2017 (outside sample).
  - Lifetime default rates peaked around 10% in 2014 but trended lower subsequently.
- Transition dynamics (monthly transition matrix, average 2011–17, in percent):
  - cat I → cat I: 97.8; cat I → cat II: 1.9; cat I → cat III: 0.1; cat I → cat IV: 0.1; cat I → cat V: 0.1; cat I → cat VI: 0.0
  - cat II → cat I: 41.6; cat II → cat II: 53.9; cat II → cat III: 3.8; cat II → cat IV: 0.4; cat II → cat V: 0.3; cat II → cat VI: 0.0
  - cat III → cat I: 4.6; cat III → cat II: 1.7; cat III → cat III: 85.0; cat III → cat IV: 7.0; cat III → cat V: 1.6; cat III → cat VI: 0.0
  - cat IV → cat I: 2.1; cat IV → cat II: 0.4; cat IV → cat III: 1.2; cat IV → cat IV: 90.5; cat IV → cat V: 5.1; cat IV → cat VI: 0.7
  - cat V → cat I: 3.3; cat V → cat II: 0.1; cat V → cat III: 0.2; cat V → cat IV: 0.3; cat V → cat V: 96.2; cat V → cat VI: 0.0
  - cat VI → cat I: 1.7; cat VI → cat II: 0.0; cat VI → cat III: 0.0; cat VI → cat IV: 0.0; cat VI → cat V: 0.0; cat VI → cat VI: 98.3
- Transition-matrix implications:
  - 2% of “regular” loans re-classified as “under supervision” monthly on average.
  - A loan “under supervision” has 41.6% chance of becoming “regular” next month.
  - Once non-performing (category III or worse), probability of returning to regular status falls to at most 5%.
  - A “below standard” loan has a 9% probability of worsening to “doubtful” or “loss” next month.
- Recent trends in loan-classification transitions (2015–first half 2017):
  - Increased cat II → cat I: "54.8 percent versus 40.0 percent".
  - Higher share of early NPLs (III and IV) migrating to loss (V): "a combined rate of 17.9 percent compared to 11.5 percent in 2015".
  - Diminished re-classification of loss loans to regular in 2017: "0.8 percent versus 5.7 percent in 2016".
  - No cases in 2017 of written-off loans returned to category I compared with "7.6 percent" in 2015.
- Amount-weighted transitions (Jan–Jun 2017, percent):
  - cat II → cat I: "6.0"
  - cat II own retention: "90.7"
  - cat III → cat I: "0.2"
  - cat III → cat V: "16.5"
  - cat IV → cat V: "22.4"
  - cat V → cat V: "99.4"
  - Interpretation: small loans (likely consumer) drive many temporary past-due cases; larger loans account for amount-weighted persistence.
- Write-off practices and classification accuracy:
  - Low transition rate V → VI consistent with banks not writing off loans; only one institution wrote off loans during sample.
  - Loans labeled loss typically not written off for "9 months (median: 5 months)"; "10 percent of category V loans labeled loss for at least 22 months"; one bank has a loan labeled loss for over five years.
  - Correct reclassification based on days past-due occurred in "86 percent" of cases overall; for four largest banks range 20% to 96%.
- Recommendations to improve CRC and supervision:
  - Collect more borrower information (income, wealth, gender) and collateral value estimates.
  - Standardize sector classification.
  - Oblige banks to write off NPLs after a defined period (e.g., two years in category V).
  - Reclassify restructured loans as performing only after a compliance period (e.g., "12 months").
  - Continue striving for timely and correct loan classification consistent with BCSTP system.

*International Monetary Fund staff report content; Source: cr18322 - IMF.*

### 1. Selected Indicators, 1964-2017 _____________________________________________________ 10

### 1. Selected Indicators, 1964-2017

### A. Introduction
- Economic performance in São Tomé and Príncipe (STP) has lagged behind other small island-states in Africa: Cabo Verde, Mauritius, and Seychelles.
- STP's GDP per capita has essentially stagnated since independence, whereas it has tripled or quadrupled in the other islands.
- Purpose of the case study:
  - Compare past economic developments in the islands.
  - Draw lessons for STP’s relative under-performance.
  - Recommend policies to foster future growth.
- Organization of the paper:
  - Section 2: brief historical background.
  - Sections 3 and 4: conditions at independence and developments since then.
  - Section 5: lessons from the past.
  - Section 6: conclusion.

### B. Background
- Historical commonalities:
  - All four islands were settled by European countries between the fifteenth and eighteenth century and evolved into plantation economies (except Cabo Verde’s arid climate).
  - Mauritius and Seychelles were transferred from French to British control in the early nineteenth century; French plantation owners stayed and dominated the economy.
  - Cabo Verde and STP were ruled by Portugal uninterrupted from the fifteenth century until independence.
- Independence dates:
  - Mauritius: 1968.
  - Cabo Verde and STP: 1975.
  - Seychelles: 1976.
- Political trajectories and institutional inheritance:
  - Mauritius retained ties to the British Commonwealth and democratic values, facilitating establishment of a middle class with vested interests in private property.
  - Cabo Verde and STP transitioned with little advance preparation for self-rule and no market foundations.
- Role of government:
  - All islands initially had governments playing a central role; protectionist inward-oriented policies were common, later prompting re-orientation and liberalization.
  - Cabo Verde, Seychelles, and STP adopted single-party socialist regimes with centralized planning and administered prices; Mauritius remained more market-oriented.

### C. Conditions at Independence
- Common constraints across the islands at independence:
  - Small size, remoteness, underdevelopment, limited resources.
  - Economies were essentially agriculture-based with few export goods and high import dependence.
  - In STP, large cocoa plantations accounted for 90 percent of farmed land and 2/3 of employment.
- Specific challenges for STP:
  - Micro size and remoteness amplified diseconomies of scale and higher production and trade costs.
  - Small communal environment fostered political and institutional systems based on personal connections and favors, undermining rule of law and good governance.
- Initial conditions ranking (Table 1: islands ranked 1–4, 4 most favorable for growth) — selected indicators and STP's relative positions:
  - Size: STP = 2.
  - Remoteness: STP = 3.
  - Natural resources: STP = 1.
  - Income per capita: STP = 4 (lowest starting level, implying higher potential to grow).
  - Life expectancy: STP = 1.
  - Human capital: STP = 1.
  - Physical and financial capital: STP = 1.
  - Democratic values: STP = 1.
- Natural resource and capital disadvantages:
  - STP had attractive environment, rich soil and fish stock, but was less favored for tourism than Seychelles and lacked Mauritius’s ethnic diversity that aided external trade and investment.
  - Departure of most Portuguese at independence removed skilled workers, managers and financial capital; infrastructure was rudimentary.
  - By contrast:
    - Mauritius and Seychelles retained technical know-how and institutional capacity.
    - Seychelles benefited from an international airport built in 1971.
    - Cabo Verde had a deep-water port, international airport, skilled mid-level administrators, and a diaspora whose remittances accounted for about 15-20 percent of GDP.
- Institutional weaknesses:
  - Weak or absent institutions at independence due to lack of transition preparation.
  - Measures from 1996 onward (Worldwide Governance Indicator, Ibrahim Index of African Governance) give STP the lowest score among the four islands.

### D. Developments Since Independence
- Immediate post-independence period:
  - Nationalization of much of the economy, including former Portuguese-owned plantations.
  - Centralized planning with price controls and subsidies.
  - Widespread nepotism and appointment of inexperienced individuals to key positions.
  - Plantation production dropped by almost half due to mismanagement.
  - State-owned enterprises (SOEs) registered large losses; commerce and services contracted as Portuguese firms exited.
  - Lax fiscal policy financed by external and domestic credit (including central bank financing) and connected bank lending led to rising inflation, deficits, and debt, forcing eventual political and economic re-orientation.
- 1990s experience:
  - Political regime shift to multi-party democracy in 1991, but persistent low capacity, clientelism, and poor governance.
  - Political instability: 17 changes in government between 1991 and 2014 (compared to 3 in Cabo Verde, 5 in Mauritius, and 3 in Seychelles).
  - Governance indicators regularly placed STP significantly below the other three islands.
  - Stop-and-go adjustment policies and lack of ownership of reforms and donor-financed public investment projects.
  - Expansionary fiscal and monetary policies, particularly around elections and in anticipation of oil-related revenue prospects.
  - Consequences:
    - Inflation soared and public debt spiked in the mid-1990s following a sharp depreciation of the dobra.
    - Despite substantial external financial assistance, income per capita during 1990-2000 declined in STP.
- Private investment and social spending:
  - Private investment was sparse during the 1990s, including in tourism (absence of measurable tourism activity in that period).
  - In contrast, Cabo Verde and Seychelles saw substantial private investment in tourism; Mauritius saw substantial private investment in textile manufacturing.
  - Tax revenue in STP remained meagre and cyclical, with increases tied to expansionary fiscal policies or international oil price increases (raising import tax revenue).
  - Low and cyclical tax revenue perpetuated dependence on external aid, fiscal imbalances, and limited domestic spending on priority social programs, hindering poverty reduction and human and physical capital accumulation.

### E. Key Findings and Implications
- Structural constraints and initial conditions were important contributors to STP’s under-performance:
  - Low human capital, weak physical and financial capital, poor governance, and limited natural-resource advantages relative to peers.
- Institutional development and political stability matter for sustained growth:
  - Islands that retained technical know-how, institutional capacity, and experienced more stable governance (Mauritius, Seychelles, Cabo Verde to varying degrees) achieved higher private investment and stronger income per capita growth.
- Policy-relevant observations from STP’s trajectory:
  - Nationalization and centralized planning, combined with weak institutions and nepotism, led to sharp production declines and SOE losses.
  - Lax fiscal and monetary policies, connected lending, and anticipation-driven spending (e.g., oil prospects) produced macroeconomic volatility: rising inflation, debt, and currency depreciation.
  - Persistent political instability (17 changes in government, 1991–2014) undermined reform ownership and implementation, deterring private investment.
  - Dependence on external aid and cyclical revenues constrained social spending and human capital accumulation, reinforcing low growth.

*Prepared by Gabriel Srour; July 9, 2018 — Democratic Republic of São Tomé and Príncipe, International Monetary Fund.*

### 15.       Reforms implemented since the 2000s improved macroeconomic stability and

### 15.       Reforms implemented since the 2000s improved macroeconomic stability and boosted economic performance

### Background and reform chronology
- Faced with an economic crisis and cases of fraud, and spurred by the HIPC initiative, the authorities took some steps towards reforms in the early 2000s, albeit implementation was slow.
- Exchange rate stabilized and inflation brought down from its peak in the 90s (although it reemerged in the wake of the global financial crisis).
- Growth surged in the mid-2000s, driven by government expansion financed by oil exploration bonuses and foreign direct investment in the tourism and banking sectors.
- Investors were attracted by prospects of finding oil and expected HIPC-related debt forgiveness, which eventually cut debt to about 60 percent of GDP in 2007.
- Implementation of some anticorruption and governance measures was delayed (for instance, a new anti-corruption law was not approved until 2012).

### Macroeconomic outcomes and vulnerabilities
- Macroeconomic conditions improved in the 2000s and continued to strengthen in recent years.
- Key indicators:
  - GDP growth has been steady around 4 percent.
  - Inflation declined to about 5 percent, anchored by the dobra peg to the euro since 2010.
  - Domestic primary deficit has been brought close to 2 percent of GDP.
  - Public debt is high: at 70 percent of GDP.
  - Large cross-arrears in the economy: 60 percent of GDP in unconsolidated terms.
- Vulnerabilities:
  - Macroeconomic stability remains vulnerable to fiscal slippages at the central government, particularly in the runup to elections, and to large contingent liabilities from SOEs.
  - Weak public financial management and pervasive cross-arrears exacerbate fiscal risks.

### Revenue, fiscal structure, and comparative performance
- Revenue performance and thresholds:
  - Tax revenue in STP remained on average around 13.5 percent of GDP historically.
  - Tax revenue is currently around 12 percent of GDP as oil prices are in a down cycle.
  - This is significantly lower than the median (15 percent) in Sub-Saharan Africa and lower than the minimum threshold (close to 13 percent) identified in recent empirical literature as necessary to support growth and development.
- Consumption tax shortfall:
  - STP’s consumption tax revenue falls between 5½ and 9 percent of GDP below the other three islands.
- Selected 2016 fiscal metrics (Percentage of GDP) from the comparative table:
  - Total revenue and grants: STP 27.7; Cabo Verde 27.0; Mauritius 21.1; Seychelles 37.9
  - Total revenue: STP 14.4; Cabo Verde 24.2; Mauritius 20.5; Seychelles 36.6
  - Tax revenue: STP 12.1; Cabo Verde 19.8; Mauritius 18.9; Seychelles 32.5
  - Consumption taxes: STP 1.2; Cabo Verde 9.1; Mauritius 6.8; Seychelles 10.5
  - Grants: STP 13.3; Cabo Verde 2.8; Mauritius 0.7; Seychelles 1.3
  - Total expenditure and net lending: STP 31.6; Cabo Verde 30.1; Mauritius 24.7; Seychelles 38.2
  - Primary balance: STP -3.5; Cabo Verde -0.5; Mauritius -1.1; Seychelles 3.4
  - Overall fiscal balance: STP -4.2; Cabo Verde -3.1; Mauritius -3.6; Seychelles -0.4
  - Notes: For STP, 13.9 percent of GDP in capital expenditure are financed by grants or concessional loans; for STP includes HIPC-related social expenditures equal to 0.24.

### Private sector, tourism, and external capital
- Role of private and foreign investment:
  - Private sector and particularly foreign capital and knowhow are key to developing the economy.
  - Cabo Verde and Seychelles promoted tourism by improving infrastructure and attracting FDI while protecting local agriculture and fishing.
- Tourism performance in STP:
  - Tourism receipts rose almost fivefold since 2010 from $11 million to about $66 million in 2017 (or from $60 per capita to slightly over $300).
  - In absolute terms STP tourism receipts remain significantly behind:
    - Cabo Verde and Seychelles close to $400 million in 2017.
    - Mauritius close to $1 billion in 2017.
  - In per capita terms STP pales compared to Seychelles’ almost $4,500.
- Infrastructure plans:
  - Authorities plan to improve airport and seaport to alleviate logistical barriers and reduce trade costs, but financial sustainability should be a prerequisite given potentially large fiscal liabilities.

### Recommendations and policy priorities
- Consolidate macroeconomic stability:
  - Strengthen fiscal discipline at the central government and contain vulnerability from public debt and losses from SOEs.
  - Contain the fiscal deficit at a sustainable level, instill financial discipline, clear arrears, and reduce public debt.
  - Anchor fiscal policy around a medium- to long-term fiscal deficit target to restrain political interference.
- Strengthen revenue mobilization:
  - Implement the planned transition to a VAT regime in 2019.
  - Expected VAT gain is about 2 percent of GDP in additional revenue over the medium term.
  - Recognize that the VAT gain alone will not bridge the revenue gap; improve tax administration, broaden the tax base, and improve tax collection.
  - Accelerate strategy to clear cross-arrears (about 60 percent of GDP in 2017 unconsolidated; tax arrears 3.5 percent of GDP).
  - Improve government services to encourage tax compliance and reduce the informal economy.
- Expand market access and attract private investment:
  - Continue efforts to expand STP’s market and attract private investment.
  - Implement a holistic strategy to develop the tourism sector and better integrate private sector and foreign capital and knowhow.
  - Ensure financial sustainability before undertaking large infrastructure projects.

### Conclusion and prospects
- Determinants of underperformance:
  - Country-specific characteristics (human capital, natural resources) and weak institutions contributed to STP’s relative underperformance since independence.
  - Political instability and weak institutions caused inefficient use of resources, delayed reforms, and large macroeconomic imbalances in the 90s and 2000s.
- Signs of opportunity:
  - Macroeconomic conditions have improved; institutions have been strengthened and human capital has been rising.
  - A number of important infrastructure projects are ongoing or close to launch.
  - Increased policy continuity in recent years and the rise in popularity of eco-tourism globally provide an opportunity to expand the tourism sector.
- Synthesis:
  - Past experience across the four island-states suggests fiscal discipline, revenue mobilization, and a more active private sector—particularly in tourism—are key to tapping STP’s growth potential.
  - Continued strengthening of public financial management, including implementing a multi-annual fiscal framework and maintaining a medium- to long-term fiscal deficit target, would help anchor fiscal policy and restrain political interference.
  - Improved tax administration and tax compliance, alongside VAT introduction, are needed to mobilize revenue to finance priority social programs and infrastructure.

*Source: cr18322 - 15.       Reforms implemented since the 2000s improved macroeconomic stability and boosted economic performance*

### 4.      However, inflation in São Tomé and Príncipe has been relatively high compared with

### 4.      However, inflation in São Tomé and Príncipe has been relatively high compared with

### Background and context
- Annual inflation in São Tomé and Príncipe between 2010 and 2017 was 8 percent, almost 7 times as high as that in the euro area during the same period (1.2 percent).
- The country adopted an exchange rate peg in January 2010.
- The consumer price index (CPI) is compiled using data from the Agua Grande district. The CPI basket was expanded, rebased and reweighted as of December 2015 with IMF technical assistance.
- CPI composition: 423 products across 12 categories. The relative weight order begins with: food and non-alcoholic beverages (hereafter called ‘food’); housing, water, electricity, gas etc.; transportation; clothing and footwear; alcoholic beverages and tobacco; furnishing and household equipment; health; recreation and culture; communication; restaurants and hotels; miscellaneous goods and services; and education.
- Food accounts for 73 percent of the consumption basket.
- The local currency was redenominated on January 1, 2018 by removing three zeros. The currency is now valued at 24.50 new dobras per euro, maintaining the peg adopted in 2010.
- Annual inflation spiked to 7.7 percent at end-2017, driven by increased import taxes on alcoholic drinks and meat and a temporary shortage of locally produced foods caused by unusual rainfall patterns.

### Descriptive analysis — headline and food inflation decomposition
- Food sub-index has been the biggest contributor to overall inflation (January 2015–December 2017).
- Month-to-month CPI changes are volatile but usually positive; only 4 instances of negative month-to-month rates during January 2015 to December 2017.
- Monthly inflation rarely exceeded 1 percent; 4 of the 5 such cases were recorded during 2017.
- Within the food sub-index, the product categories with the largest weights—and largest drivers of food inflation—are cereals, vegetables and fish; vegetables and fish were particularly important in recent periods because of unusual rainfall.
- Source decomposition: locally-produced food (mainly fish, fresh fruits and vegetables) has been a key driver of food inflation since mid-2017; previously imported food was the primary driver.

### Time-series analysis — drivers of medium-to-long-term inflation (food and non-food)
- Data: quarterly from 2002 Q4 to 2017 Q4; decompositions focus on January 2015–December 2017 (overall) and December 2016–December 2017 (food).
- Independent variables considered include: international food price index, local rainfall (percentage deviation from long-run quarterly mean since 1901), M2 growth, domestic primary spending, tourist arrivals growth, international oil price index, nominal effective exchange rate (NEER), and a dummy for the exchange regime switch (peg).

Key regression results (Tables 2 and 3):
- Peg effect:
  - Peg lowered food inflation by 7.2 percentage points (Peg = -7.240* in Table 2).
  - Peg lowered non-food inflation by 8.7 percentage points (Peg = -8.705*** in Table 3).
- Food inflation regressions (Table 2; observations: 60 pooled, 29 pre-change, 20 post-change):
  - IntFoodPrice coefficients: 0.240**, 0.560**, 0.287*** across (1) pooled, (2) pre-change, (3) post-change.
  - Rainfall coefficients: 0.008, -0.002, 0.016** across (1) pooled, (2) pre-change, (3) post-change.
  - M2 coefficients: -0.000**, -0.000*, 0.088*** across (1) pooled, (2) pre-change, (3) post-change.
  - NEER coefficients: -0.004, 0.808, -0.564*** across (1) pooled, (2) pre-change, (3) post-change.
  - Constant terms: 18.325***, 24.499***, 6.962*** across (1) pooled, (2) pre-change, (3) post-change.
- Non-food inflation regressions (Table 3; observations: 64 pooled, 29 pre-change, 23 post-change):
  - Oil coefficients: 0.018, 0.050, -0.018 across (1) pooled, (2) pre-change, (3) post-change.
  - M2 coefficients: -0.000, -0.000*, 0.068** across (1) pooled, (2) pre-change, (3) post-change.
  - NEER coefficients: 0.049, 0.248, -0.270 across (1) pooled, (2) pre-change, (3) post-change.
  - Tourist arrivals (post-change only): 0.029*** (in column (3)).
  - Constant terms: 14.779***, 16.322***, 2.731*** across (1) pooled, (2) pre-change, (3) post-change.

Interpretation of empirical findings:
- The exchange rate peg has a statistically significant dampening effect on both food and non-food inflation.
- In the post-peg period, M2 growth is highly significant and positively correlated with both food and non-food inflation, indicating a demand-pulling effect linked to fiscal spending.
- Food inflation is significantly affected by international food prices across periods and by local rainfall in the post-peg period, indicating heavy dependence on food imports and vulnerability of low-productivity agriculture to weather shocks.
- Non-food inflation in the post-peg period is positively correlated with growth in tourist arrivals, indicating demand pressures on local services and supply-side bottlenecks (e.g., hotel room shortages).

### Graphical confirmations
- Figure 3a: food inflation broadly follows the trend of M2 growth in the post-peg period.
- Figure 3b: food inflation follows the trend of international food price growth.
- Figure 3c: non-food inflation follows the trends of M2 growth and growth in tourist arrivals.

### Policy implications and recommendations (Conclusion)
- Support the pegged exchange rate system with coherent macroeconomic policies:
  - Strengthen the reserves buffer to support the peg.
  - This can be achieved through a package of policies, including closer monitoring of externally-financed projects to avoid delays in grant disbursements, and the demand-side and supply-side policies below.
- Continue fiscal consolidation to manage domestic demand:
  - Inflation is significantly impacted by M2 growth, which is closely associated with growth of public expenditure; continued fiscal consolidation would ease domestic demand pressure and help build international reserves.
- Increase local supply, enhance productivity, and improve infrastructure:
  - Utilize more local products and decrease dependence on food imports (e.g., adjust consumption habits to rely more on indigenous staples such as breadfruit and green banana instead of imported rice; enhance processing of local primary agricultural products).
  - Improve agricultural productivity (e.g., build greenhouses, promote sustainable agriculture, make agriculture more robust to weather shocks). The ongoing joint effort by authorities with partners including China and the United Nations is noted as positive.
  - Build infrastructure and increase capacity of local services to support tourism demand. Implementation of the tourism development strategy launched in January 2018 with World Bank support should help eliminate tourism supply-side bottlenecks over time.

*Source: IMF staff calculations.*

### References

### References

### Sources cited
- Barnichon, R. and S. J. Peiris, 2007, “Sources of Inflation in Sub-Saharan Africa,” IMF Working Paper 07/32 (Washington: International Monetary Fund).
- Caceres, C., M. Poplawski-Ribeiro, and D. Tartari, 2011, “Inflation Dynamics in the CEMAC Region,” IMF Working Paper 11/232 (Washington: International Monetary Fund).
- International Monetary Fund, 2018, São Tomé and Príncipe – Report on Consumer Price Index Mission (January 8-19) (Washington: International Monetary Fund).
- Mangani, R., 2012, “The Effects of Monetary Policy on Prices in Malawi,” AERC Research Paper 252 (Nairobi: Africa Economic Research Consortium).
- Nguyen, A. D. M., J. Dridi, F. D. Unsal, and O. H. Williams, 2015, “On the Drivers of Inflation in Sub-Saharan Africa,” IMF Working Paper No. 15/189 (Washington: International Monetary Fund).
- Wu, D. F., 2017, “Understanding Inflation in Malawi: A Quantitative Investigation,” IMF Working Paper No. 17/48 (Washington: International Monetary Fund).

### IMF chapter: The Financial Performance of State Owned Enterprises and Risk Spillovers

#### A. Background
- Key concerns:
  - SOEs in São Tomé and Príncipe have long been loss-making due to low administrated sales prices and elevated operating costs.
  - Causes: prices set by the government below cost recovery levels without adequate financial compensation; outdated production facilities (particularly for electricity and water); large technical and commercial losses; elevated personnel costs.
  - Result: delays in investments, worsening operations, and further losses (World Bank, 2017).
- SOE ownership structure:
  - Four fully state-owned firms: electricity and water (EMAE), port (ENAPORT), airport services (ENASA), postal services (Correios).
  - Four enterprises with minority government stakes (16 to 49 percent) in telecommunications, banking, air transportation, and oil import and distribution.
- EMAE specifics and challenges:
  - Registered losses since 2011; operating costs outpacing commercial revenues despite some government subsidies.
  - Average retail tariff: USD 0.21/KWh.
  - Average cost of service: USD 0.32/KWh.
  - Energy transmission and distribution losses: 43 percent of energy dispatched (SSA average: 15 percent).
  - Heavy reliance on inefficient thermal turbines and costly fuel imports.
  - Resulted in large domestic arrears to suppliers.
- Reform and investment support:
  - World Bank Power Sector Recovery Project supports structural improvement at EMAE (World Bank, 2017).
  - Central reform elements: Least Cost Development Plan and Financial Management Improvement Plan.
  - Specific measures: rehabilitating hydroelectric plant, large-scale installation of meters, purchasing lower-cost electricity from private operators (solar and natural gas investments).

#### B. Financial Performance of SOEs
- Overall assessment approach:
  - Performance assessed via indicators covering profitability, solvency, liquidity, and leverage.
  - Consolidated structure of arrears between SOEs, government entities, and private sector analyzed; spillover risks evaluated.

- Table: São Tomé and Príncipe: SOE Financial Statement Information (end-2017 or earlier, in EUR million)
  - Fully State-owned
    - EMAE (Energy): Revenue 13.1; Profit -5.9; Assets 45.6; Debt (incl. arrears) 35.2; Arrears (gross) 30.8; Equity 10.4; Debt-to-Equity 3.4.
    - ENAPORT (Port): Revenue 4.2; Profit 0.1; Assets 3.3; Debt (incl. arrears) 1.5; Arrears (gross) 0.4; Equity 1.8; Debt-to-Equity 0.8.
    - ENASA (Airport): Revenue 3.4; Profit 0.1; Assets 7.6; Debt (incl. arrears) 3.8; Arrears (gross) 3.3; Equity 3.8; Debt-to-Equity 1.0.
    - Correios (Postal): Revenue 0.2; Profit n/a; Assets 1.0; Debt (incl. arrears) n/a; Arrears (gross) n/a; Equity -0.4; Debt-to-Equity n/a.
  - Minority stakes
    - CST (Telecom): Revenue 9.3; Profit 1.1; Assets 18.1; Debt (incl. arrears) 5.5; Arrears (gross) 0.0; Equity 12.6; Debt-to-Equity 0.4.
    - BISTP (Bank): Revenue 9.1; Profit 2.0; Assets 83.5; Debt (incl. arrears) 70.8; Arrears (gross) 0.0; Equity 12.7; Debt-to-Equity 5.6.
    - STP Air (Airline)*: Revenue 10.5; Profit 0.7; Assets 3.0; Debt (incl. arrears) n/a; Arrears (gross) 0.6; Equity 1.9; Debt-to-Equity n/a.
    - ENCO (Oil firm)*: Revenue 40.4; Profit 1.7; Assets 159.6; Debt (incl. arrears) n/a; Arrears (gross) 139.6; Equity 16.2; Debt-to-Equity n/a.
  - Notes: * end-2015; ** end-2016; # end-2017 preliminary figures. Source: SOEs’ financial statements; World Bank.

- Performance variation:
  - ENAPORT: relatively healthy overall.
  - ENASA: turnaround with small profits since 2016, but remains financially vulnerable.
  - EMAE: operational and financial deficiencies comparable to a firm at pre-insolvency.
  - Correios: loss-making but very small operation.

- Table: São Tomé and Príncipe: Key SOE Performance Indicators (end-2017 or earlier)
  - EMAE* (Energy):
    - Operational Cost-to-Revenue (in %): 177.5
    - Profit Margin (in %): -61.0
    - Working Capital (EUR million): -45.3
    - Working Capital-to-Equity Capital (in %): -272.8
    - Cash-to-Short-Term Liabilities (in %): 0.1
    - Debt-to-EBITDA (in %): undefined
    - Note: * end-2016.
  - ENAPORT (Port):
    - Operational Cost-to-Revenue (in %): 93.6
    - Profit Margin (in %): 2.2
    - Working Capital (EUR million): 0.1
    - Working Capital-to-Equity Capital (in %): 6.8
    - Cash-to-Short-Term Liabilities (in %): 4.9
    - Debt-to-EBITDA (in %): 4.9
  - ENASA** (Airport):
    - Operational Cost-to-Revenue (in %): 98.0
    - Profit Margin (in %): 0.2
    - Working Capital (EUR million): -1.4
    - Working Capital-to-Equity Capital (in %): -36.3
    - Cash-to-Short-Term Liabilities (in %): 9.9
    - Debt-to-EBITDA (in %): 15.2
    - Note: ** end-2017 preliminary figures.
  - Source: SOEs’ financial statements.

- Key findings on financial indicators:
  - Operational cost-to-revenue ratios of the three large SOEs range from 94 to 177 percent (88 to 160 percent, if excluding depreciation and amortization).
  - Profit margin ranges from 2.2 to -61 percent.
  - EMAE’s working capital deficit: STD 1.1 billion at end-2016; nominal capital STD 0.4 billion at end-2016 (would be STD -934 million without capitalizing investment subsidies).
  - Cash-to-short-term liabilities: 0.1 percent (EMAE), 4.9 percent (ENAPORT), 9.9 percent (ENASA).
  - Debt-to-EBITDA critical threshold noted as five; ENAPORT: 4.9; ENASA: 15.2; EMAE: negative/undefined.
  - For EMAE’s debt of STD 1.7 billion to become sustainable, earnings (EBITDA) would have to increase by STD 440 million.
  - ENAPORT and ENASA intend to obtain local bank loans: ENAPORT EUR 2 million; ENASA EUR 1.5 million.

- Minority state-owned enterprises performance:
  - CST (telecom) and BISTP (bank) are profitable with moderate leverage; market shares about 85 percent (CST) and 50 percent (BISTP).
  - STP Airways: profit in 2015 but financially weak; government did not participate in a recent capital increase.
  - ENCO (oil supplier): appears profitable in accounting terms but burdened by large public sector arrears.

#### C. Structure of Arrears and Risk Spillovers
- Arrears network:
  - An intricate web of net arrears exists among large SOEs, central government, other government entities, partially state-owned enterprises, private firms, foreign firms, and banks.
  - ENCO has built large arrears to its parent company (an Angolan SOE) due to public institution arrears.
  - Central government alternates between arrears and claims with SOEs and other entities (claims reflecting back taxes).
- Collection and enforcement issues:
  - EMAE and ENAPORT unable to collect bills from many clients; weak collection/enforcement mechanisms reduce cash inflows and exacerbate arrears.
- Arrears clearance plans:
  - ENAPORT and ENASA agreed repayment plans of 5 and 25 years, respectively, to fully clear arrears (including penalty interest).
  - Prior repayment plans stalled as new arrears accumulated while clearing old ones.
  - Repayment of government arrears to ENCO is running ahead of schedule due to favorable domestic-import price differentials used to reduce debt from past oil price subsidies.
  - Penalty interest accumulated: about 40 percent (ENAPORT) and close to 100 percent (ENASA) of the underlying arrears; penalties forgiven only after full clearance at completion of repayment plan.
- Risks to compliance:
  - SOEs’ low liquidity and potential additional bank loan servicing may make adhering to repayment schedules difficult.
  - Example: ENASA’s agreed annual repayment would, other things equal, deplete its cash-at-hand within three years.
- Fiscal contingent liabilities and macrofinancial linkages:
  - Large SOE arrears constitute significant contingent fiscal liabilities, effectively guaranteed by the government (implicitly or explicitly).
  - These contingent liabilities are concentrated with a few creditors (e.g., ENCO) and could undermine government financial viability and raise future debt risk premia.
  - Tax arrears equivalent to almost 7 percent of GDP have deprived the state of revenue for social programs and infrastructure, and have contributed to government arrears to suppliers.
  - Current bank exposures to SOEs are relatively small; no significant past-due loans reported. Anecdotal evidence indicates cases of non-payment by government leading to liquidity issues at suppliers who then defaulted on bank loans.

#### D. Conclusion and Recommendations
- (Heading present in source; no substantive text provided in this excerpt.)

*Prepared by Torsten Wezel; sources: SOEs’ financial statements; World Bank; São Tomean authorities; IMF staff calculations.*

### 23.      The SOEs’ poor operational and financial performance has contributed to an intricate

### 23.      The SOEs’ poor operational and financial performance has contributed to an intricate

### SOEs: operational performance, arrears, and macrofinancial risks
- The large SOEs suffer from low profitability and liquidity as well as excess leverage, which threaten solvency in individual cases.
- SOEs have had, and in the case of EMAE continue to have, difficulties clearing their large stock of arrears.
- Recently-agreed repayment plans for some SOEs may prove challenging given weak cash flow and the prospect of servicing additional bank loans to be taken out for investment purposes.
- Macrofinancial linkages appear to be muted at present, but could become an issue as SOEs increase their bank debt and debt service without immediate efficiency gains.

### SOE reform strategy and short-term measures to address arrears
- Medium-term reform plans and effective structural measures are the only solution to the SOE issue.
- Pending deeper reforms, recommended short-term measures aimed at addressing the arrears situation:
  - With medium-term investment projects in train, SOEs should attempt to identify areas for cost saving (e.g. staff cost, including training expenses) and collect more arrears from customers.
  - The potential for netting out bilateral arrears between SOEs and government should be explored, presupposing that the claims are considered valid by both parties.
  - To ease somewhat the burden of arrears clearance, the central government could consider forgiving late-payment penalties in line with full scheduled repayment of the main arrears per year.
  - Borrowing by SOEs from banks should be in line with debt servicing capacity from SOEs’ cash flows to avoid NPLs with banks or accumulation of additional arrears elsewhere.

*International Monetary Fund staff report content.*

### Banking system context: NPLs and the Central Credit Registry (CRC)
- The banking system is characterized by persistent high levels of non-performing loans (NPLs).
  - The NPL ratio of the banking system tripled between 2013 and end-June 2017 to 32 percent, peaking at about 36 percent in 2016.
  - The “unweighted” NPL ratio (share of number of loans that are non-performing) is substantially lower, at less than 10 percent of GDP.
  - The high NPL ratio in value terms is driven by the poor performance of a limited number of large loans.
- A central credit registry (CRC) was made operational in March 2011 and its coverage increased from about 50 percent in the initial stages to 96 percent recently.
  - The CRC covered 269 loans in 2011, 6,787 in 2013, and 10,864 in 2017.
- CRC microdata permit identification of trends within banks’ credit portfolios and derivation of indicators for financial stability analysis and banking supervision.

### CRC data sample and coverage
- Sample period: March 2011 to June 2017.
- Sample size and scope:
  - 469,823 observations from up to 12 banks (including banks closed or taken over).
  - Related to 29,238 loans and 13,039 borrowers (around seven percent of the total population or 14 percent of the adult population).
- Reporting improvements: banks used to report data with a certain lag which will be reduced significantly through a new online reporting system.

### Information collected by the CRC
- Data fields include: bank name, date of loan origination, maturity date, loan amount at origination, carrying amount, currency, interest rate, status of the loan in given month (category of loan classification), type of loan rate (fixed vs. variable rate), type of borrower (household vs. firm), sector/occupation of the borrower, district of the borrower, and an indirect measure of borrower income (share of disposable income in total debt).
- Currently no information on collateral value is provided.

### Credit market concentration and distribution
- Demand-side concentration:
  - The largest firms (top 5 percent) received 54 percent of total corporate credit.
  - The largest individual borrowers received 29 percent of retail credit.
- Supply-side concentration:
  - The largest bank extended almost half of the loans in the sample; the other two largest banks around 15 percent each.
- Geographic concentration:
  - Loans concentrated in the district of the capital (59.4 percent).
- Implications of concentration:
  - At one small bank the largest borrower accounts for 40 percent of the loan portfolio; at three others that share amounts to 20 percent of total loans.
  - Even at the largest banks the largest 10 borrowers, representing less than 1 percent of the total numbers of borrowers, account for over 30 percent of the loan volume.

### Credit composition, sizes, currencies, and rates
- By borrower type and loan amounts:
  - Loans to individuals account for 97 percent in terms of number of loans, while loans to firms represent 71 percent of total loan amount.
  - Median domestic currency loan value for individuals: STD 35,000 (about USD 1,750); 10 percent of individual loans above STD 130,000 (about USD 6,500).
  - Median loan value for firms: STD 500,000 (about USD 25,000); 10 percent of firm loans exceed STD 4.9 million (about USD 245,000).
  - São Tomé and Príncipe’s GDP per capita is USD 1,772 (end-2017).
  - Median loan value for USD loans: about USD15,000; for EUR loans: EUR 30,000.
- Currency composition:
  - New Dobras, STD—95.6 percent; EUR—2.6 percent; USD—1.8 percent.
  - Only 5 percent of borrowers have access to foreign-currency credit.
- Interest rates and maturities:
  - Interest rates may be as high as 54 percent.
  - Maturities range from 1 day to almost 37 years.
  - Average interest rate for STD loans: 20 percent; for foreign currency loans: 11 percent.
  - Average maturity: STD loans 2.7 years; foreign currency loans 5.0 years.
  - About 80 percent of the loans have a fixed rate.
- Variation across largest lenders:
  - The largest and third largest bank extend loans on average twice as large as the other two large banks.
  - The two largest banks charge an average interest rate slightly above the sample average of 20 percent; the third largest bank charges 6 percent.
  - The largest bank supplies 72 percent of foreign currency loans.

### Borrower relationships and loan counts
- Bank relationships:
  - 80 percent of borrowers have loans only from one bank.
  - 16 percent have credit with two banks.
  - 4 percent have credit with three or more banks.
- Loan counts per borrower:
  - Around 50 percent of borrowers have only one loan in the dataset.
  - 25 percent have two loans.
  - 14 percent have three loans.
  - 1.5 percent of borrowers have ten or more loans.
  - The four largest borrowers with over 30 loans borrow exclusively from one bank; all four are enterprises.

### Loan classification and NPL definitions
- Loan classification follows BCSTP (2007) categories principally based on days past-due and financial condition/collateral status:
  - Category I Regular: The borrower has made all scheduled payments on time.
  - Category II Under supervision: The borrower is late in making its scheduled payment by 30-89 days.
  - Category III Below standard: The borrower is late in making its scheduled payment by 90-179 days.
  - Category IV Doubtful: The borrower is late in making its scheduled payment by 180-359 days.
  - Category V Loss: The borrower is more than 360 days late in making its scheduled payment.
  - Category VI Written off: The loan was written off from the bank’s balance sheet.
- Standard NPL definition: borrower late by 90 or more days; categories III to V are considered non-performing.

### Descriptive NPL statistics from CRC data
- On average during the sample period:
  - Seven percent of the loans became non-performing.
  - 20 percent of loans had a delayed payment of at least 30 days.
  - Loans that become non-performing do so after 15 months on average.
  - Around 11 percent of the borrowers in the sample have had at least one loan classified as non-performing.
- By borrower type:
  - 39 percent of the firms in the sample had at least one non-performing loan.
  - 10 percent among individual borrowers had at least one non-performing loan.
- As of end-June 2017:
  - NPLs represented 32 percent of the total loan amount.
  - Top 10 delinquent borrowers (about 1 percent of borrowers with NPLs) account for 43 percent of total NPL amount; the top 3 borrowers represent 22 percent.
  - Among top 10 borrowers with NPLs, nine are firms, mainly from the wholesale trading sector.
  - One active bank holds 32 percent of loans classified as a loss (unweighted number); a liquidated bank holds another 40 percent.

### Characteristics of NPLs vs performing loans
- NPLs tend to have:
  - A higher interest rate.
  - A longer maturity.
  - Greater likelihood of being denominated in local currency.
  - Larger loan amounts.
  - Tend to have a fixed rate.
- Statistical significance: differences between NPL and non-NPL loans are statistically significant for most characteristics, with the exception of one item noted as "58 59 no" in the original table; loan amount and having a fixed rate show larger magnitudes of difference.

### Borrower behavior post-default and multi-bank relationships
- 45 percent of borrowers who had a delay in loan repayment of more than 90 days are able to contract a new loan afterwards.
  - Of these new loans, 41 percent are with the same bank where the delay occurred and 59 percent are with a different bank.
- Loans from multiple-bank borrowers:
  - 10 percent of the loans from multiple-bank borrowers were NPL, compared to 6 percent of loans from single-bank borrowers.
  - Having more than one loan does not seem to influence the probability that a borrower will have an NPL loan, suggesting some borrowers exploited banks’ incomplete information about credit histories.

### Default rates, vintage analysis, and trends
- Default rates of new loans increased during 2012-15:
  - Vintage analysis: incidence of newly-granted loans turning non-performing within first 12 months rose from 0.2 percent for the 2012 vintage to 3.5 percent in the 2015 vintage.
  - When weighting the rate by number of months a loan stayed in default during that 12-month period, the default rate drops significantly because on average a loan is only around 4 months in default within the first year before performing again.
- BCSTP reported the 12-month default rate picked up toward the end of 2017 (outside sample), reaching 6.7 percent in December 2017.
- Lifetime default rates peaked at around 10 percent in 2014 but have trended lower in recent years because loans are still relatively new to default.

### Comprehensive vintage analysis (default rates by month and origin year, in percent)
- Default rate after month:
  - Columns: 1, 3, 6, 12, 18, 24, 32, 36, 48
- 2012: 0.4, 0.4, 0.4, 0.7, 1.4, 1.8, 3.5, 4.2, 5.7
- 2013: 0.8, 0.8, 1.0, 1.9, 3.7, 5.1, 6.5, 7.2
- 2014: 0.5, 1.3, 2.0, 3.6, 5.3, 6.1, 8.3
- 2015: 1.7, 1.7, 2.6, 4.2, 5.8
- 2016: 0.3, 0.5, 0.9

### Transition dynamics within loan classification system (monthly transition matrix, average 2011-17, in percent)
- Matrix entries are sample frequencies/probabilities that a loan will change classification next month or remain unchanged, given current classification.
- cat I - regular: cat I 97.8, cat II 1.9, cat III 0.1, cat IV 0.1, cat V 0.1, cat VI 0.0
- cat II - under supervision: cat I 41.6, cat II 53.9, cat III 3.8, cat IV 0.4, cat V 0.3, cat VI 0.0
- cat III – below standard: cat I 4.6, cat II 1.7, cat III 85.0, cat IV 7.0, cat V 1.6, cat VI 0.0
- cat IV - doubtful: cat I 2.1, cat II 0.4, cat III 1.2, cat IV 90.5, cat V 5.1, cat VI 0.7
- cat V - loss: cat I 3.3, cat II 0.1, cat III 0.2, cat IV 0.3, cat V 96.2, cat VI 0.0
- cat VI - written off: cat I 1.7, cat II 0.0, cat III 0.0, cat IV 0.0, cat V 0.0, cat VI 98.3

- Key implications from transition matrix:
  - On average, 2 percent of “regular” (category I) loans were re-classified as “under supervision” (30-89 days).
  - A loan classified as “under supervision” has a 41.6 percent chance of becoming “regular” again next month.
  - Once a loan is classified as non-performing (“below standard” or worse), the probability of returning to regular status falls considerably (at most 5 percent).
  - A loan classified as “below standard” has a 9 percent probability (sum of worsening transitions to IV or V) of worsening to “doubtful” or “loss” in the next month.

*International Monetary Fund staff report content.*

### 21.      A comparison of transition matrices at different years reveals recent trends in loan

### A comparison of transition matrices at different years reveals recent trends in loan classifications

### Recent trends in loan-classification transitions (2015–first half 2017)
- Increased transition rate for category II watchlist loans back to category I: "54.8 percent versus 40.0 percent".
- Higher share of loans in the “early” NPL categories III and IV migrating to the loss category V: "a combined rate of 17.9 percent compared to 11.5 percent in 2015".
- Diminished ad-hoc re-classification of loss loans as regular category I loans in 2017 compared to previous years: "0.8 percent versus 5.7 percent in 2016".
- No cases in 2017 of written-off loans being returned to category I, compared with a transition rate of "7.6 percent" in 2015.

### Selected transition-matrix highlights (as reported)
- Table 5 (Average of Monthly Transitions, January-December 2015, in percent) key entries:
  - cat II → cat I: "40.0"
  - cat III → cat V and cat IV → cat V combined indicated in analysis as 11.5 percent (system-wide comparison).
  - cat VI → cat I: "7.6"
- Table 6 (Average of Monthly Transitions, January-December 2016, in percent) key entries:
  - cat II → cat I: "49.9"
  - cat V → cat I: "5.7"
- Table 7 (Average of Monthly Transitions, January-July 2017, in percent) key entries:
  - cat II → cat I: "54.8"
  - cat III → cat V: "12.5"
  - cat IV → cat V: "25.3"
  - cat V → cat I: "0.8"
  - cat VI → cat VI: "100.0"

### Amount-weighted transition analysis (January-June 2017)
- When transitions are weighted by carrying value, category II and III loans returning to category I show much lower rates, implying small loans are more likely to be temporarily non-performing.
- Table 8 (Amount-Weighted, Average of Monthly Transitions, January-June 2017, in percent) key entries:
  - cat II → cat I: "6.0"
  - cat II (own retention): "90.7"
  - cat III → cat I: "0.2"
  - cat III → cat V: "16.5"
  - cat IV → cat V: "22.4"
  - cat V → cat V: "99.4"
  - Interpretation: many small loans—probably mostly consumer loans—become temporarily past-due before arrears are cleared; delays sometimes due to late salary payments to borrowers.

### Write-off practices and persistence of loss loans
- Low transition rate from category V to VI consistent with banks not writing off loans; only one financial institution (a bank being liquidated) wrote off loans during the sample period.
- Loans labeled as a “loss” (over 365-day payment delay) are typically not written off for "9 months (median: 5 months)".
- "Ten percent of the category V loans are labeled as a loss for at least 22 months", with "80 percent of these loans concentrated in one bank".
- That one bank has a loan labeled as a loss for "over five years".

### Loan-classification timeliness and accuracy
- Correct reclassification based solely on days past due occurred in "86 percent" of cases.
- For the remainder, loans worsened by more than one category within a month (inconsistent with the 30-day interval downgrade schedule).
- One in every five loans that have a category worse than “regular” seems to be misclassified by staying longer than appropriate in a given category (i.e., about "20 percent").
- Among the four largest banks, classification correctness based solely on days past-due ranges from "20 to 96 percent".

### Conclusions (summary of paper's main conclusions)
- High credit concentration creates vulnerabilities as banks depend on the performance of a few clients.
- Enterprises are mainly responsible for excessive levels of NPLs; households’ contribution is small.
- Banks are reluctant to write off NPLs, possibly due to protracted or partial recovery of underlying collateral.
- Vintage analysis: default rates increased through 2015 before receding somewhat.
- Transition matrices: once a loan becomes non-performing, it tends to deteriorate further; however, recent years show more recoveries during early past-due periods and more consistent loan classification.

### Recommendations: improvements to the credit registry
- Make available more information on the borrowers, such as income, wealth, and gender.
- Collect information on the collateral underlying a loan and its estimated value, which could be used for macroprudential purposes.
- Classify borrowers/loans according to principal economic sectors (e.g. agriculture, manufacturing, construction trade, services etc.). (The credit registry already includes sector information, but the classification is too granular and not standardized.)

### Recommendations: regulation and supervision enhancements
- Oblige banks to write-off non-performing loans from their balance sheets after a certain period (e.g., after two years in category V).
- Reclassify restructured loans as performing only after a certain period of compliant debt service (e.g., "12 months"), instead of immediately.
- Continue to strive for a timely and correct classification of loans so that such action is fully consistent with the BCSTP’s loan classification system.

*Source: cr18322 - 21. A comparison of transition matrices at different years reveals recent trends in loan classifications (IMF).*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2018/cr18322.pdf_