## _cr15354

## Source details

**Canonical URL:** [_cr15354](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2015/_cr15354.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2015/_cr15354.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2015/_cr15354.pdf.json)

---

### Export diversification and product quality: findings and metrics
- Export-to-GDP ratio: declined from 56 percent in 2000 to 43 percent in 2014.
- Two largest export items in 2014: “Synth. perfume & flavour materials” (includes Coca-Cola concentrates) and “raw sugar, beet & cane”.
- Share of top 10 products: 68 percent as of 2014.
- Number of export products: increased since 2000, peaked in 2010, then declined moderately afterwards.
- Large portion of relatively small exports (exports less than $1 million): accounted for about three quarters of exports in 2014.
- Product composition shifts (2000 → 2014):
  - Textiles: 19% → 10%
  - Miscellaneous Edibles: 33% → 41%
  - Sugar: 10% → 21%
  - Woodpulp: 10% → 3.6%
  - Others: 23% → 19%
- Destination shifts (2000 → 2014):
  - E.U.: 8% → 13%
  - South Africa: 60% → 60%
  - U.S.: 9% → 4%
  - Others: 23% → 23%
- Theil index (export diversification):
  - Swaziland is among the more diversified exporters in Sub-Saharan Africa by product-level Theil index.
  - Export diversification has decreased since 2008 (Theil index increased).
  - Using the IMF (2014b) coefficient, if diversification returns to 2008 levels, output growth would increase by about 0.4 percentage points.
- Export competitiveness and quality:
  - Swaziland’s share in global non-mineral exports: percent-levels around 0.005–0.025 between 2000–2014.
  - Staff analysis finds the real exchange rate is moderately overvalued.
  - Quality index (based on unit values and Henn et al. (2013)): higher than the African average but has declined since 2006.
  - Sectors with scope to move up the quality ladder: food and beverages, and manufacturing.
- Selected indicators (comparative):
  - Secondary education enrollment, gross = 59.9
  - Quality of public institutions (max. 7) = 3.9
  - Extent of market dominance rank = 11
  - Private sector credit/GDP (2014) = 21.2 percent
- Banking concentration:
  - Three commercial banks comprise 85 percent of total banking system assets; the remaining 15 percent is owned by a state-owned development bank.

### Policy implications for diversification and upgrading
- Cross-country evidence (IMF 2014a, 2014b) recommends:
  - Improve infrastructure (transport and power) and trade networks to lower business costs.
  - Invest in human capital (health and education).
  - Encourage financial deepening through banking sector reforms.
  - Reduce barriers to entry for new products to boost diversification.
- Quality upgrading is associated with:
  - Skilled labor force.
  - Strong institutional quality.
  - Well-developed domestic financial markets.
- Summary recommendation:
  - Reverse deterioration in export diversification and product quality, with urgency given expiration of AGOA eligibility; focus on human capital, institutional quality, and financial deepening to improve diversification, attract investment, and support growth and employment.

### Macro-financial risks from SACU revenue fall: overview
- Key risk channel:
  - Fall in Southern African Customs Union (SACU) revenues could adversely impact fiscal and external balances and the financial sector.
  - Expected effects: increased sovereign spread, reduced private credit growth, government liquidity problems and arrears, borrower payment-capacity deterioration, worsening bank asset quality with possible feedback to the real economy.
- Historical precedent (2010–11 fiscal crisis):
  - NPLs increased starting immediately with the 2010–11 fiscal crisis.
  - Loan-to-deposit ratio dropped and stayed low due to decline in private sector credit growth.
  - In 2012, NPLs increased sharply owing to economic contraction and defaults in government-related sectors and SMEs.
  - Sovereign risk premium remained high when it was above 50 bps; changes in NPLs exhibited positive correlation with the sovereign spread.
- Structural vulnerability:
  - NPL ratio declined to 6.4 percent in 2015Q2, but average NPL in recent years was the highest among SACU countries and among the highest after the 2010–11 crisis in Sub-Saharan Africa.
  - Deterioration in asset quality since 2010–11 was widespread across all four banks; some banks had much higher NPLs due to large, volatile exposures to construction and small businesses.
  - Swaziland’s banking system is predominantly foreign-owned; three commercial banks (85 percent of assets) are subsidiaries of South African banks; remaining 15 percent is a state-owned development bank.

### Stress-testing approach and methodology
- Three-stage stress testing exercise:
  1. Use a VAR model to explore macro-financial links from SACU revenue shock to macro variables.
  2. Map macroeconomic shocks to banks’ asset quality (NPL ratio) using cross-country panel regression, accounting for feedback of banks’ credit risk on the real economy.
  3. Conduct stress testing based on calibrated NPL shock and the stress-testing model developed in IMF (2015), grounded in Čihák (2007).
- VAR specification:
  - Reduced-form VAR with quarterly series and one-quarter lag.
  - Endogenous variables included:
    - Growth in electricity consumption (in millions kilowatt hours) as proxy for real GDP (quarter-on-quarter).
    - Growth in credit to private sector (quarter-on-quarter).
    - De-trended sovereign spread (Swazi 3-month T-bill rate minus South African T-bill rate).
    - Current gross official reserves at the Central Bank of Swaziland (trend removed).
  - Exogenous variable: deviation of actual SACU disbursements from trend (z_t), computed by applying HP filter to actual and projected SACU disbursement; HP smoothing parameter = 1,600.
- Linkages modeled:
  - SACU shock directly worsens fiscal position, potentially crowds out private credit and increases sovereign risk premium (raising real interest rate).
  - Indirect effects via deterioration of bank asset quality.
  - Feedback: rising NPLs can reduce private investment through deleveraging and lower economic growth.

### VAR estimates and macro-financial transmission
- Sample period: Swaziland 2009Q3-2014Q4; quarterly VAR with one-quarter lag; HP filter smoothing parameter = 1,600 for trends.
- Key empirical linkages:
  - A fall in SACU disbursements significantly reduces GOR, leading to a hike in sovereign spread.
  - Large negative coefficient of international reserve on the hike of sovereign spread indicates high elasticity of Swaziland’s spread to SACU shock.
  - Impulse response: sovereign spread peaks with a 4 quarter lag after the SACU revenue shock and remains elevated over the medium-term.
  - Higher sovereign spread increases real interest rates, which reduces private credit growth (with a lag of one quarter) and lowers consumption/growth.
- Model diagnostics (R-squared for endogenous variables reported): 0.43284; 0.058147; 0.794076; 0.941532.

### Projections and scenario analysis (2015–2018)
- Scenarios:
  - Scenario 1 (benchmark): SACU shock of the same magnitude as projected in the staff report.
  - Scenario 2 (alternative): Prolonged SACU shock that persistently stays lower than the historical trend in SACU disbursement.
- Projected outcomes:
  - Output growth could be as low as -2 percent in 2016–17.
  - Annual growth of electricity consumption projected to about -2 percent during the shock period under adverse scenarios.
  - Figure 2 (charts referenced) highlights a large and immediate spike in sovereign spread and a sharp slowdown in growth due to sluggish credit intermediation.
- Caveats:
  - VAR model is backward-looking and does not account for future policy reactions.
  - Electricity consumption is used as a quarterly proxy for GDP due to lack of quarterly GDP data.

### Mapping macro shocks to banks’ asset quality
- Cross-country satellite panel model:
  - Fixed effects for CMA countries (Lesotho, Swaziland, Namibia); sample period 2009-2014 quarterly.
  - Dependent variable: change in aggregate NPL ratios.
  - Regressors: changes in sovereign spread and SACU revenues; lag of private sector credit growth; crisis dummy for NPL spikes >30 percentage points; country and quarter dummies.
- Key coefficient estimates:
  - A spike in the country’s spread by 100 bps beyond the historical trend has immediate impact in higher NPL ratio by 10.8 percent.
  - Private sector credit growth appears to increase the NPL ratio (effect not statistically significant).
- Bank-level satellite regressions:
  - Four individual banks in Swaziland; electricity consumption growth included.
  - Bank-level results confirm that higher sovereign risk premium and stagnant real economy activities contribute to higher NPL ratios.
- Feedback effects:
  - VAR with NPLs endogenous confirms: lower growth increases NPLs; increase in NPLs leads to temporary reduction in private credit growth and contraction in economic growth.

### NPL projections and implications
- Under prolonged SACU revenue shock (Scenario 2):
  - NPL ratio projected to rise rapidly when the SACU crisis materializes in 2016 and peak in the same year.
  - Calibrated credit shock used to examine adverse impact on capital adequacy when NPL ratio peaks in each scenario.

### Stress-testing: methodology, scenarios, and results
- Framework:
  - Balance sheet approach based on a modified Čihák framework; bank-level balance sheet and income statement data as of end-December 2014.
  - Basel I regulatory requirement: at least 4 percent Tier 1 capital and 8 percent total regulatory capital to RWA.
- Scenarios:
  - Scenario 1: Only a credit shock (increase in NPLs) calibrated from the benchmark VAR model.
  - Scenario 2: Larger credit shock due to prolonged SACU shock; plus interest rate shock (increase in interest rates by 10 percentage points as assumed in IMF (2015)).
  - Scenario 3: Larger credit shock for vulnerable banks set at their historical maximum; loss-given-default for largest exposure set at 100 percent; for three South African subsidiary banks, losses for subsidiaries capped at 50 percent of regulatory capital in case of default of largest exposure.
- System-level stress-test outcomes (selected figures):
  - Initial Tier 1 capital-to-RWA ratio: 19.1%
  - Subtractions due to shocks:
    - Aggregate NPL shock:
      - Scenario 1: -1.6%
      - Scenario 2: -3.3%
      - Scenario 3: -4.4%
    - P/L effect of interest rate shock (Scenario 2 and 3): 0.2%
    - Default of largest exposure (Scenario 3): -7.1%
  - Tier 1 capital-to-RWA ratio under stress:
    - Scenario 1: 17.4%
    - Scenario 2: 15.9%
    - Scenario 3: 7.6%
- Interpretation:
  - Scenario 1: losses of 1.6 percent of Tier 1 regulatory capital from loan quality deterioration would leave capital well above regulatory requirement.
  - Scenario 2: losses of 3.3 percent of regulatory capital; net-interest margin effect small but marginally positive.
  - Scenario 3: concentration risk could produce losses of 7.1 percent of regulatory capital and depress system Tier 1 capital-to-RWA to 7.6 percent.
  - Large variability across individual banks; under extreme scenarios capital adequacy could potentially be negative for some banks.

### Policy implications and recommendations
- Fiscal policy:
  - Fiscal consolidation is critical to mitigate adverse impacts of the SACU revenue shock, contain a rise in risk premium, mitigate crowding-out of private credit, and avoid government cash management problems.
- Financial sector policy:
  - Strengthen supervision and regulation of the financial sector.
  - Mitigate risks stemming from high concentration risk in bank exposures.
- Overall:
  - While banks are currently well capitalized, prolonged or compounded SACU shocks (with interest rate increases and defaults of large borrowers) could materially weaken asset quality and capital adequacy, affecting financial intermediation.

### Box 2 — Satellite Model: estimate on the macroeconomic shock on credit quality
- Preferred fixed-effect specification includes:
  - Detrended sovereign spread (above South African T-bill)
  - Private sector credit growth (first-differenced)
  - Crisis dummy (spike in NPL ratios >30 percentage points)
  - Country and quarter fixed effects
- Empirical finding: a spike in the country’s spread by 10 bps beyond the trend increases the NPL ratio by about 1.1 percent.
- Panel satellite model (Box Table 2) — selected coefficients and statistics:
  - D NPL: 0.108*** (0.011)
  - L.D Credit to private sector: 0.123 (0.134)
  - Crisis dummy: 0.549** (0.080)
  - Constant: -0.035* (0.009)
  - Observations: 63
  - Number of groups: 3
  - R-square (within): 0.625
  - R-square (between): 0.858
- Bank-level satellite model (Box Table 3) — selected coefficients and statistics:
  - D NPL: 0.009** (0.002)
  - D electricity consumption: -0.695*** (0.093)
  - L.D Credit to private sector: 0.371 (0.571)
  - Crisis dummy: 0.306** (0.082)
  - Observations: 86
  - Number of groups: 4
  - R-square (within): 0.271
  - R-square (between): 0.553
- Stylized dynamics and calibration:
  - NPLs spiked in 2010-11 and again in 2012; credit intermediation declined and was slow during crises; detrended sovereign spread positively correlated with change in NPLs.
  - Stress-test macro assumptions:
    - Baseline: reduction in SACU as in staff’s baseline.
    - Scenario 2: prolonged SACU decline from trend.
    - Spread (detrended) projected to rise and peak in 2017 under baseline; remains elevated under scenario 2.
    - GOR detrended expected to drop; reduction larger under scenario 2.
    - Growth (electricity consumption) slower after SACU shock; private credit growth slower immediately after the shock; NPL ratio spikes after the prospective SACU crisis (prolonged under scenario 2).

*Source: IMF staff calculations and analysis as presented in the provided content unit.*

### REFERENCES _______________________________________________________________________ 7

### REFERENCES _______________________________________________________________________ 7

### Export diversification and product quality: findings and metrics
- Swaziland’s exports:
  - Export-to-GDP ratio declined from 56 percent in 2000 to 43 percent in 2014.
  - Two largest export items in 2014: “Synth. perfume & flavour materials” (includes Coca-Cola concentrates) and “raw sugar, beet & cane”.
  - Share of top 10 products stood at 68 percent as of 2014.
  - The number of export products increased since 2000, peaked in 2010, then declined moderately afterwards.
  - Large portion of relatively small exports (exports less than $1 million) accounted for about three quarters of exports in 2014.
- Product composition shifts (2000 → 2014):
  - Textiles: 19% → 10%
  - Miscellaneous Edibles: 33% → 41%
  - Sugar: 10% → 21%
  - Woodpulp: 10% → 3.6%
  - Others: 23% → 19%
- Destination shifts (2000 → 2014):
  - E.U.: 8% → 13%
  - South Africa: 60% → 60%
  - U.S.: 9% → 4%
  - Others: 23% → 23%
- Theil index (export diversification):
  - Swaziland is among the more diversified exporters in Sub-Saharan Africa by product-level Theil index.
  - Export diversification has decreased since 2008 (Theil index increased).
  - Using the IMF (2014b) coefficient, if diversification returns to 2008 levels, output growth would increase by about 0.4 percentage points.
- Export competitiveness and quality:
  - Swaziland’s share in global non-mineral exports declined (series shows percent-levels around 0.005–0.025 between 2000–2014).
  - Staff analysis finds the real exchange rate is moderately overvalued (see annex 1 of the staff report).
  - Quality index (based on unit values and Henn et al. (2013)) is higher than the African average but has declined since 2006.
  - Sectors with scope to move up the quality ladder: food and beverages, and manufacturing.
- Selected indicators (comparative):
  - Swaziland: Secondary education enrollment, gross = 59.9; Quality of public institutions (max. 7) = 3.9; Extent of market dominance rank = 11; Private sector credit/GDP (2014) = 21.2 percent.
  - Three commercial banks comprise 85 percent of total banking system assets; the remaining 15 percent is owned by a state-owned development bank.

### Policy implications for diversification and upgrading
- Cross-country evidence (IMF 2014a, 2014b) identifies effective policies:
  - Improve infrastructure (transport and power) and trade networks to lower business costs.
  - Invest in human capital (health and education).
  - Encourage financial deepening through banking sector reforms.
  - Reduce barriers to entry for new products to boost diversification.
- Quality upgrading is associated with:
  - Skilled labor force.
  - Strong institutional quality.
  - Well-developed domestic financial markets.
- Summary recommendation:
  - Reverse deterioration in export diversification and product quality, with urgency given expiration of AGOA eligibility; focus on human capital, institutional quality, and financial deepening to improve diversification, attract investment, and support growth and employment.

### Macro-financial risks from SACU revenue fall: overview
- Key risk channel:
  - Fall in Southern African Customs Union (SACU) revenues could have serious macroeconomic consequences: adverse impacts on fiscal and external balances and on the financial sector.
  - Expected effects include increased sovereign spread, reduced private credit growth, government liquidity problems and arrears, borrower payment-capacity deterioration, and worsening bank asset quality with possible feedback to the real economy.
- Historical precedent (2010–11 fiscal crisis):
  - Nonperforming loans (NPLs) increased starting immediately with the 2010–11 fiscal crisis.
  - Loan-to-deposit ratio dropped and stayed low due to decline in private sector credit growth.
  - In 2012, NPLs increased sharply owing to economic contraction and defaults in government-related sectors (e.g., construction) and SMEs.
  - Sovereign risk premium remained high during the period when it was above 50 bps; changes in NPLs exhibited positive correlation with the sovereign spread.
- Structural vulnerability:
  - NPL ratio declined to 6.4 percent in 2015Q2, but average NPL in recent years was the highest among SACU countries and among the highest after the 2010–11 crisis in Sub-Saharan Africa.
  - Deterioration in asset quality since 2010–11 was wide-spread across all four banks; some banks had much higher NPLs due to large, volatile exposures to construction and small businesses.
  - Swaziland’s banking system is predominantly foreign-owned; three commercial banks (85 percent of assets) are subsidiaries of South African banks and operate similar business models; the remaining 15 percent is a state-owned development bank.

### Stress-testing approach and methodology
- Three-stage stress testing exercise to assess SACU revenue shock impact:
  1. Explore macro-financial links from SACU revenue shock to macroeconomic variables using a vector autoregression (VAR) model.
  2. Map macroeconomic shocks to banks’ asset quality (NPL ratio) using cross-country panel regression, accounting for feedback effects of banks’ credit risk on the real economy.
  3. Conduct stress testing based on calibrated NPL shock and the stress-testing model developed in recent MCM TA (IMF, 2015), grounded in the standard macro stress-testing framework (Čihák, 2007).
- VAR specification and variables:
  - Reduced-form VAR with quarterly series and one-quarter lag.
  - Endogenous variables included:
    - Growth in electricity consumption (in millions kilowatt hours) as proxy for real GDP (quarter-on-quarter).
    - Growth in credit to private sector (quarter-on-quarter).
    - De-trended sovereign spread (Swazi 3-month T-bill rate minus South African T-bill rate).
    - Current gross official reserves at the Central Bank of Swaziland (trend removed).
  - Exogenous variable:
    - Deviation of actual SACU disbursements from trend (z_t), computed by applying HP filter to actual and projected SACU disbursement in the baseline framework.
- Linkages modeled:
  - SACU shock directly worsens fiscal position, potentially crowds out private credit and increases sovereign risk premium (raising real interest rate).
  - Indirect effects via deterioration of bank asset quality.
  - Feedback: banking sector shocks (e.g., rising NPLs) can reduce private investment through deleveraging and thereby feedback to lower economic growth.

### Observed and referenced empirical sources
- Data sources explicitly cited in the analysis:
  - UN Comtrade (export product-level data, SITC Rev.1, 2000−14).
  - Central Bank of Swaziland.
  - IMF staff calculations and IMF (2014a, 2014b).
  - Henn et al. (2013) for export quality estimation.
  - Financial Stability Indicator and IMF staff for NPL cross-country comparisons.
- Referenced literature (selected):
  - Edwards et al., 2013, “Swaziland Economic Diversification Study,” DNA Economics.
  - Gigineishvili, Mauro, and Ke Wang, 2014, “How Solid Is Economic Growth in the East African Community,” IMF working paper 14/150.
  - Henn, Papageorgiou, and Spatafora, 2013, “Export Quality in Developing Countries,” IMF Working Paper 13/108.
  - IMF, 2014a and 2014b, Sustaining Long-run Growth and Macroeconomic Stability in Low-income Countries—Role of Structural Transformation and Diversification.
  - Jerzmanowski and Cuberes, 2009, “Democracy, Diversification and Growth Reversals,” The Economic Journal, Vol. 119, pp. 1270–1302.
  - World Bank, 2015, “Kingdom of Swaziland Growth Report.”
  - IMF (2015) MCM technical assistance stress-testing model; Čihák (2007) macro stress-testing framework.

*Prepared materials and analysis as presented in the IMF country materials for the Kingdom of Swaziland (November 24, 2015).*

### 6.      The VAR estimates and the impulse response functions show that a SACU revenue fall

### 6.      The VAR estimates and the impulse response functions show that a SACU revenue fall

### VAR estimates and macro-financial transmission
- Methodology and data:
  - VAR uses quarterly series for Swaziland (2009Q3-2014Q4) with endogenous variables: growth rate of electricity consumption (proxy for quarterly GDP growth), private sector credit growth, detrended sovereign spread (above South African T-bill rate), and gross international reserve (GOR); all lagged by one quarter. Detrended SACU disbursements is included as exogenous.
  - Trend components computed by applying the Hodrick-Prescott (HP) filter to the quarterly series with the smoothing parameter of 1,600.
- Key empirical linkages:
  - A fall in SACU disbursements (below its trend) significantly reduces GOR, leading to a hike in sovereign spread (above its trend).
  - A large negative coefficient of international reserve on the hike of sovereign spread suggests high elasticity of Swaziland’s spread in response to the SACU revenue shock.
  - Impulse response: sovereign spread peaks with a 4 quarter lag after the SACU revenue shock and remains elevated over the medium-term.
  - Higher sovereign spread increases real interest rates, which reduces private credit growth (with a lag of one quarter) and lowers consumption/growth.
- Quantitative model diagnostics (selected):
  - R-squared for endogenous variables reported as: 0.43284; 0.058147; 0.794076; 0.941532 (as shown in Box Table 1).

### Projections and scenario analysis (2015–2018)
- Scenarios:
  - Scenario 1 (benchmark): SACU shock of the same magnitude as projected in the staff report.
  - Scenario 2 (alternative): Prolonged SACU shock that persistently stays lower than the historical trend in SACU disbursement.
- Projected outcomes:
  - Output growth could be as low as -2 percent in 2016–17.
  - Annual growth of electricity consumption projected to about -2 percent during the shock period under adverse scenarios.
  - Charts (Figure 2) highlight a large and immediate spike in sovereign spread and a sharp slowdown in growth due to sluggish credit intermediation.
- Caveats:
  - VAR model is backward-looking and does not account for future policy reactions.
  - Electricity consumption is used as a quarterly proxy for GDP due to lack of quarterly GDP data.

### Mapping macro shocks to banks’ asset quality
- Approach:
  - Cross-country satellite panel model using fixed effects for CMA countries (Lesotho, Swaziland, Namibia); sample period 2009-2014 quarterly.
  - Dependent variable: change in aggregate NPL ratios for country i in year t.
  - Regressors include changes in sovereign spread and SACU revenues; a lag of private sector credit growth; a crisis dummy indicating periods when the NPL ratio grows by more than 30 percent; country and quarter dummies included.
  - Bank-level satellite regressions run for four individual banks in Swaziland (quarterly electricity consumption growth included for bank-level regressions).
- Key findings:
  - A spike in the country’s spread by 100 bps beyond the historical trend has immediate impact in higher NPL ratio by 10.8 percent.
  - Private sector credit growth appears to increase the NPL ratio (effect not statistically significant).
  - Bank-level results confirm that higher sovereign risk premium and stagnant real economy activities (captured by electricity consumption growth) contribute to higher NPL ratios.
- Feedback effects:
  - VAR with NPLs as an endogenous variable confirms that lower growth increases NPLs and, conversely, an increase in NPLs leads to temporary reduction in private credit growth and contraction in economic growth.

### NPL projections and implications
- Projections:
  - Under prolonged SACU revenue shock (Scenario 2), the NPL ratio is projected to rise rapidly when the SACU crisis materializes in 2016 and peak in the same year.
  - The calibrated credit shock is used to examine adverse impact on capital adequacy when NPL ratio peaks in each scenario.

### Stress testing: methodology and scenarios
- Framework:
  - Balance sheet approach based on a modified Čihák framework; bank-level balance sheet and income statement data as of end-December 2014.
  - Basel I regulatory requirement: at least 4 percent Tier 1 capital and 8 percent total regulatory capital to RWA.
- Scenarios summarized:
  - Scenario 1: Only a credit shock (increase in NPLs) calibrated from the benchmark VAR model (assuming SACU disbursements projected in the staff report).
  - Scenario 2: Larger credit shock due to prolonged SACU shock; plus interest rate shock (increase in interest rates by 10 percentage points as assumed in IMF (2015)).
  - Scenario 3: Larger credit shock for vulnerable banks set at their historical maximum; loss-given-default for largest exposure set at 100 percent; for three South African subsidiary banks, losses for subsidiaries capped at 50 percent of regulatory capital in case of default of largest exposure (reflecting partial off-balance risk sharing).
- Stress test hurdle outcomes (system-level, selected figures):
  - Initial Tier 1 capital-to-RWA ratio: 19.1%
  - Subtractions due to shocks:
    - Aggregate NPL shock:
      - Scenario 1: -1.6%
      - Scenario 2: -3.3%
      - Scenario 3: -4.4%
    - P/L effect of interest rate shock (Scenario 2 and 3): 0.2%
    - Default of largest exposure (Scenario 3): -7.1%
  - Tier 1 capital-to-RWA ratio under stress:
    - Scenario 1: 17.4%
    - Scenario 2: 15.9%
    - Scenario 3: 7.6%
- Interpretation:
  - Under Scenario 1, losses of 1.6 percent of Tier 1 regulatory capital from loan quality deterioration would leave capital well above regulatory requirement.
  - Under Scenario 2, losses of 3.3 percent of regulatory capital; net-interest margin effect small but marginally positive.
  - Scenario 3 shows concentration risk could produce losses of 7.1 percent of regulatory capital and depress system Tier 1 capital-to-RWA to 7.6 percent.
  - Large variability across individual banks; under extreme scenarios capital adequacy could potentially be negative for some banks.

### Policy implications and recommendations
- Fiscal policy:
  - Fiscal consolidation is critical to mitigate adverse impacts of the SACU revenue shock, contain a rise in risk premium, mitigate crowding-out of private credit, and avoid government cash management problems.
- Financial sector policy:
  - Strengthen supervision and regulation of the financial sector.
  - Mitigate risks stemming from high concentration risk in bank exposures.
- Overall:
  - While banks are currently well capitalized, prolonged or compounded SACU shocks (with interest rate increases and defaults of large borrowers) could materially weaken asset quality and capital adequacy, affecting financial intermediation.

*Source: IMF staff calculations and analysis as presented in the provided content unit.*

### Box 2. Satellite Model: Estimate on the Macroeconomic Shock on Credit Quality

### Box 2. Satellite Model: Estimate on the Macroeconomic Shock on Credit Quality

### Preferred specification and approach
- Preferred fixed-effect specification includes:
  - Detrended sovereign spread (above South African T-bill)
  - Private sector credit growth (first-differenced credit to private sector indicates growth rates)
  - Crisis dummy (one if the country-quarter observation experienced a spike in NPL ratios by more than 30 percent points)
  - Country and quarter fixed effects (Hasuman test supports fixed effect model)
- Crisis dummy captures a spike in the NPL ratio at the onset of the fiscal crisis.
- Empirical finding summarized: a spike in the country’s spread by 10 bps beyond the trend increases the NPL ratio by about 1.1 percent.

### Box Table 2 — Panel Satellite Model: Macroeconomic Impacts of NPL Ratio (cross-country panel regression)
- Dependent variable: percentage point change in NPL ratio.
- Coefficients and statistics (as reported):
  - D NPL: 0.108*** (0.011)
  - L.D Credit to private sector: 0.123 (0.134)
  - (unnamed coefficient) 0.016 (0.042)
  - Crisis dummy: 0.549** (0.080)
  - Constant: -0.035* (0.009)
  - Observations: 63
  - Number of groups: 3
  - R-square (within): 0.625
  - R-squared (between): 0.858
  - Country and quarter fixed effect: Y
- Notes:
  - Robust standard errors reported.
  - SACU and Spread variables are detrended (remove respective historical trends).
  - “L” denotes one period lag.

### Bank-level satellite regression (bank-level panel, 2009–2014 quarterly)
- Financial stability unit compiled balance sheets and income statements for four individual banks; regression includes electricity consumption growth as an additional control.
- Key finding: increase in sovereign risk premium and electricity consumption significantly influence bank-level NPL ratio; growth rate in the real economy matters for change in NPLs.

### Box Table 3 — Bank-level Satellite Model: Macroeconomic Impacts of NPL Ratio
- Coefficients and statistics (as reported):
  - D NPL: 0.009** (0.002)
  - D electricity consumption: -0.695*** (0.093)
  - L.D Credit to private sector: 0.371 (0.571)
  - (unnamed coefficient): -0.002 (0.051)
  - Crisis dummy: 0.306** (0.082)
  - Constant: -0.079 (0.034)
  - Observations: 86
  - Number of groups: 4
  - R-square (within): 0.271
  - R-squared (between): 0.553
  - Bank and quarter fixed effect: Y

### Stylized dynamics and stress-test calibration
- Historical and projected dynamics highlighted:
  - NPLs spiked during the 2010-11 fiscal crisis and again in 2012 (increase in unsecured lending and riskier fiscal situation with risk premium remaining above 50 bps); NPLs declined thereafter.
  - Increases in NPLs coincided with drops in large exposure to capital ratio at onset of 2010-11 crisis and in 2012.
  - Credit intermediation declined and was slow during 2010-11 crisis; lower credit growth in 2012, with recent recovery.
  - Evolution of detrended sovereign spread is positively correlated with change in NPLs.
- Stress-test calibration (macroeconomic assumptions):
  - Baseline scenario: assumes a reduction in SACU as in staff’s baseline framework.
  - Scenario 2: assumes prolonged SACU decline from the trend.
  - Projected behavior (IMF staff calculation):
    - Spread (detrended) will continue to rise and peak in 2017 under the baseline; remains elevated under scenario 2.
    - Gross official reserve (GOR) detrended expected to drop as observed in previous crisis in scenario 1; reduction larger under scenario 2.
    - Growth (proxied by electricity consumption growth) will be slower after the SACU shock and recover in following years.
    - Private credit growth will be slower immediately after the SACU shock.
    - NPL ratio will spike after the prospective SACU crisis; shock is prolonged under scenario 2.

### VAR analysis (impulse response functions)
- VAR impulse responses estimated based on the VAR model specified in Eq. (1) with change in NPL ratio and crisis dummy additionally controlled (both defined in Eq. (2)).
- Responses reported (IMF staff calculation):
  - Response of D_electricity to D_private_credit
  - Response of D_private_credit to D_real_interest_rate
  - Response of D_private_credit to Spread (detrended)
  - Response of D_real_interest_rate to Spread (detrended)
  - Response of Spread (detrended) to Spread (detrended)
  - Response of Spread (detrended) to GOR (detrended)
  - With feedback effect: responses of D_electricity and D_private_credit to D_NPL, and responses of D_NPL to Spread (detrended), to D_NPL, and to D_electricity
  - Responses shown are to one S.D. innovations ± 2 S.E.

### Stress-testing scenarios (Table 1)
- Scenario 1. Decline in SACU revenues
  - Expected impact on the economy:
    - Lower SACU revenues worsen fiscal position and external balance, causing:
      - Reduced fiscal spending
      - Increase in government’s debt, arrears, and contingent liabilities
      - Higher risk premium
      - Subdued economic activity
  - Expected impact on banks:
    - Higher credit risk (an increase in NPLs)
  - Expected indirect impact:
    - Reduction in lending to the private sector which could trigger further deceleration in economic activities and credit intermediation

- Scenario 2. Prolonged decline in SACU revenues + Increase in Real Interest Rates
  - Expected impact on the economy:
    - Prolonged decline in SACU revenues worsen fiscal and external positions, causing:
      - Reduced fiscal spending
      - Increase in government’s debt, arrears, and contingent liabilities
      - Higher risk premium
      - Subdued economic activity
  - Expected impact on banks:
    - Higher credit risk (an increase in NPLs; effect stronger than scenario 1)
    - Higher interest rate risk (through changes in spread between interest earning assets and interest bearing liabilities)
  - Expected indirect impact:
    - An increase in real interest rate and the sharper reduction in lending to the private sector which could trigger further deceleration in economic activities and credit intermediation (effect stronger than in scenario 1)

- Scenario 3. Prolonged decline in SACU revenues + Concentration risk (a default of major borrowers)
  - Expected impact on the economy:
    - Prolonged decline in SACU revenues worsen fiscal and external positions, causing:
      - Reduced fiscal spending
      - Increase in government’s debt, arrears, and contingent liabilities
      - Higher risk premium
      - Subdued economic activity
      - Cut in subsidy to SOEs
  - Expected impact on banks:
    - Higher credit risk (an increase in NPLs; effect stronger than scenario 1 and 2 due to a default of large exposures)
    - Higher concentration risk (a default of the single largest exposure or other major borrowers)
  - Expected indirect impact:
    - An increase in real interest rate and the sharper reduction in lending to the private sector which could trigger further deceleration in economic activities and credit intermediation (effect much stronger than in scenario 1 and 2)

*Source: IMF staff calculations and Central Bank of Swaziland as presented in Box 2 of the provided material.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2015/_cr15354.pdf_