## Executive Summary — Stress Testing of Guernsey Financial Sector (_cr1104)

## Source details

**Canonical URL:** [Executive Summary — Stress Testing of Guernsey Financial Sector (_cr1104)](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1104.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1104.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1104.pdf.json)

---

### Purpose and approach
- Purpose: assess vulnerabilities in the Guernsey banking and insurance sectors through simulation of the effect of a potential double dip recession on solvency of Guernsey banks and insurance companies.
- Approach:
  - Combined bottom-up (BU) STs run by Guernsey banks/insurers and top-down (TD) STs by the mission with the GFSC.
  - Tests applied to end-2009 financial positions.
  - Risk types tested: interest rates, exchange rates, credit quality, asset price risk, operational risk, liquidity risk; for insurance, (life) insurance risks also assessed.
  - Single-factor sensitivity tests pooled into a multivariate scenario (global double dip stress scenario).

### Key contextual indicators (end-2009)
- Banking sector capital adequacy: CARs at around 19 percent (up from 15 percent at end 2008).
- Return on equity (ROE): about 15.8 percent in 2009 (compared with below 11 percent in 2008).
- Nonperforming loans (NPLs) to total gross loans in 2009: 0.09 percent.
- Sum of ten largest exposures to capital in 2009: 132.9.
- Liquid assets to total assets in 2009: 61.3 percent.
- Deposits as a percentage of assets in 2009: 86.5 percent.
- Cross-border loans to total loans in 2009: 90.3 percent; domestic loans 9.7 percent.
- Total foreign currency-denominated loans to total loans in 2009: 73.4 percent.
- Net open position in foreign exchange to capital in 2009: 1.5.

### Banking sector — Coverage and balance-sheet context
- Coverage:
  - STs covered 19 subsidiaries of foreign banks (out of 21 on the island); liquidity tests also included 5 foreign branches.
  - The 19 subsidiaries accounted for, in principle, the total assets of the Guernsey subsidiaries: 36 percent of the total assets of the banking sector.
- Ownership composition of the 19 subsidiaries sample (percent of sample assets as stated):
  - Seven subsidiaries belonged to U.K. banking groups: 40 percent.
  - Three subsidiaries belonged to EU banking groups: 4 percent.
  - Five subsidiaries belonged to Swiss banks: 18 percent.
  - One subsidiary belonged to a Bermuda based bank: 3 percent.
  - Two subsidiaries belonged to North America based banks: Canada 20 percent, United States 10 percent.
  - One subsidiary belonged to a South African group: 5 percent.
- Select aggregate balance-sheet figures (Table 2, end 2009):
  - Total assets: 135,803 (GBP millions); Percent of total assets: 100.0; Percent of GDP: 7,893.9.
  - Cash: 250.0 (GBP millions); Percent of total assets: 0.1; Percent of GDP: 1.5.
  - Loans to Banks: 75,605 (GBP millions); Percent of total assets: 55.7; Percent of GDP: 4,394.7.
    - Loans to Parent: 56,464 (GBP millions); Percent of total assets: 41.9; Percent of subgroup: 75.3; Percent of GDP: 3,282.1.
  - Marketable Assets: 24,353 (GBP millions); Percent of total assets: 17.2; Percent of GDP: 1,415.6.
  - Loans and Advances: 25,710 (GBP millions); Percent of total assets: 18.9; Percent of GDP: 1,494.5.
  - Investments: 6,791 (GBP millions); Percent of total assets: 5.0; Percent of GDP: 394.7.
  - Deposits due to: 114,751 (GBP millions); Percent of total assets: 84.5; Percent of GDP: 6,670.2.
  - Capital: 2,273 (GBP millions); Percent of total assets: 1.7; Percent of GDP: 132.1.
- For short-term liquidity assessment five large foreign branches were added; these branches represented 62 percent of Guernsey banking sector assets and 69 percent of the liquidity risk test sample.

### Banking sector — Methods and calibration
- Two approaches:
  - Top-Down (TD): mission tests using supervisory off-site data collected by the GFSC; aggregate consistent bank-specific results.
  - Bottom-Up (BU): banks ran standardized tests using their own frameworks; GFSC ensured consistency.
- Shocks applied instantaneously to end-2009 positions; outcomes expressed as post-shock CARs against GFSC minimum capital adequacy level: 10 percent (TD test includes Pillar 2 charge).
- Liquidity tests: ability to survive five days of deposit runs used as benchmark.
- Calibration principle: shocks represent “extreme but plausible” events, broadly consistent with recent FSAPs and observed crisis stress levels.

### Banking sector — Core solvency results (high-level)
- Core finding: credit risk identified as the most relevant risk.
- Sensitivity analysis:
  - Banks most vulnerable to credit risk (including parent exposure and large exposures) and FX rate risk.
- Scenario analysis:
  - Three banks would need recapitalization in a severe macroeconomic shock (double dip scenario).
- Liquidity risk:
  - Vulnerable if daily withdrawals of all deposits (maturity up to one month) exceed 15 percent per day for 5 days (56 percent cumulative).

### Market risk — shocks and outcomes
- BU market risk shocks (Box 1) preserved verbatim:
  - Interest rate risk I1–I12: parallel shifts of pound sterling, US dollar, and euro yield curves by ±200 and ±300 basis points.
  - Foreign exchange shocks F1–F8: pound and dollar depreciations/appreciations by 20 percent and 30 percent against all other currencies.
  - Asset price shocks A1–A3: A1 equity prices decline by 35 percent; A2 rated securities downgraded by two notches; A3 rated securities downgraded by four notches.
- Treatment in tests:
  - Interest rate shock effects simulated on income; losses (gains) directly deducted (added) from capital.
  - FX and asset price effects simulated on risk-weighted assets.
- Outcomes:
  - Limited exposure to market risk overall.
  - FX rate risk was the most notable market risk exposure.
  - Interest rate risk limited in the tests.
  - Asset price extreme shock A3 would not threaten system but is relatively conservative.

### FX risk — exposures and outcomes
- FX exposures:
  - 75 percent of total loans are foreign-currency denominated.
  - 79 percent of liabilities are denominated in foreign currency.
  - Off-balance sheet Forex contracts amount to 26 percent of the assets.
- BU findings (F1, F3, F6, F8):
  - In most severe shocks (F3, F8: 30 percent moves), capital adequacy of up to four banks falls below 10 percent; recapitalization needs amount to almost 3 percent of GDP.
  - For 20 percent instantaneous FX changes, effects are considerably lower with hardly any recapitalization needs.
- TD tests:
  - Based on open currency positions with linear RWAs increase assumption; simulate less severe stress than BU.
- Overall assessment: system appears robust against FX risks given shock severities used, but evolution should be closely monitored.

### Interest rate risk — tests and implications
- Tests examined ability to absorb sudden yield-curve changes.
- BU and TD tests did not show substantial capital adequacy movements from interest-rate changes.
- TD referenced Economic Value-at-Risk based on a 200 bps shock; 300 bps assumed twice as severe; impacts put directly against capital without profit as first line of defense.
- Caveat: interest rates were exceptionally low at testing time; regular tests recommended.

### Credit risk — design and outcomes
- Credit risk accounts for about 94 percent of Pillar 1 minimum capital requirements and 92 percent of total capital requirements excluding Pillar 2 concentration charges.
- Lending to the real sector: 6 percent of assets but almost 400 percent in terms of GDP.
- Seven credit scenarios run (C1, C2, C3, C4, C5, C6, C7; concentration tests C3, C7).
  - Banks asked to increase PDs by 100 percent and 300 percent with recovery rates of 50 percent.
- Outcomes:
  - BU tests do not suggest credit risk is a major threat in baseline.
  - TD tests with PDs +300 percent (C2): recapitalization needs amount to 2.2 percent of GDP with the most affected institution losing its entire capital base.
- Caveats:
  - Credit risk is low probability–high impact.
  - Basel II Standardized Approach may delay visibility of risk for unrated exposures.
  - Loan Loss Provisions were not part of study; 2009 provisions decreased while NPLs increased — provisioning should be monitored.

### Credit concentration risk — parent and large-exposure channels
- Upstreaming of deposits to parents makes concentration and parent spillovers significant.
- Test C7 (10 percent of banks’ claims on parent fail with no recovery):
  - BU: capital shortfalls for about half the institutions: 9 banks.
  - TD: capital shortfalls for about half the institutions: 13 banks.
  - Recapitalization needs in TD test amount to almost the island’s GDP.
- Default of three largest non-parent counterparts (C3) with 50 percent recovery:
  - BU: recapitalization needs up to 14 percent of GDP; 3 banks require recapitalization.
  - TD: 6 banks require recapitalization with needs of 22 percent of GDP.
- Recommendations:
  - Monitor parent solvency using solvency FSIs (ratings, CDS spreads, KMV EDFs) and liquidity positions.
  - Maintain cooperation with home supervisors.
  - Run large-exposure tests regularly using conservative exposure definitions.
  - Ensure authorities have off-site collateral and exposure data for meaningful tests.

### Operational risk — design and outcomes
- Operational risk constitutes 5 percent of total Pillar 1 capital requirements (Standardized/Basic Indicator Approaches).
- TD increases simulated:
  - O1: Increase by 50 percent (RWAs increase by 7.5 percent; capital adequacy decline 7 percent in relative terms).
  - O2: Increase by 100 percent (RWAs increase by 10 percent; capital adequacy decline 13 percent in relative terms).
  - O3: Increase by 200 percent (RWAs increase by 20 percent; capital adequacy decline 17 percent in relative terms).
- Impact limited given high solvency levels.
- Recommendation: operational risk STs should be based on economic models (ICAAP) to capture bank-specific circumstances.

### Combined scenario (I4 & F3 & A2 & C2 & O2)
- Composite scenario combined worst single-factor interest and FX shocks with severe asset price, credit, and operational shocks.
- Bottom-up combined result:
  - CAR 15.4 percent; Min. -64.5 percent; Max. 73.1 percent; 3 banks breaching min. CAR; Recapitalization needs 66.3 (GBP m) and 3.9 percent of GDP.
- Observation: three banks fall short under medium-to-high severity combined scenario; recapitalization needs only slightly higher than sum of shocks given limited number of banks entering recapitalization zone.
- Recommendation: run scenario tests to capture simultaneous multi-factor shocks.

### Additional TD credit-risk assumptions and results
- Additional TD assumptions:
  - A1: No profits available as first line of defense.
  - A2: Average NPLs equal to max(system 2009 average 0.4 Percent, bank-specific level).
  - A3: NPLs assumed to be 2 Percent for all banks baseline.
  - A4: RWAs increase by 0.5 times relative PD increase (rule of thumb).
- Key results:
  - If A1 + A2 + A4 and PDs +100 percent: six banks need recapitalization; recapitalization needs 6 percent of GDP.
  - If PDs +300 percent under same assumptions: 15 out of 19 banks would be hit; recapitalization needs 28 percent of GDP.
- Sector-specific shocks:
  - Domestic loans or mortgage portfolio shocks could trigger recapitalization needs around GBP 20 millions.
- Recommendation: include additional assumptions (A1–A4) selectively for future STs; use A4 to approximate IRB-type model effects.

### Liquidity risk — design, shocks, and findings
- Purpose: examine days banks can withstand daily deposit withdrawals with no external funding.
- Test simulates severe run with liquidation restricted to most liquid assets (cash, government bonds, interbank loans up to eight days, marketable securities).
- TD liquidity assumptions:
  - Simulated withdrawal: 30 percent of remaining deposits withdrawn during five consecutive days => 83 percent withdrawn in a business week.
  - Two interbank-access scenarios:
    - L1: interbank market accessible (except funding >1 month); no intra-group funding.
    - L2: interbank market essentially closed (except up to 8 days); no intra-group funding.
  - Two liquid-asset scenarios:
    - (i) All marketable securities liquid.
    - (ii) Only government bonds liquid; other marketable securities illiquid.
- Findings:
  - Under L1, scenario (i): seven subsidiaries and two branches become illiquid during one business week.
  - Under L1, scenario (ii): 13 illiquid subsidiaries and 4 illiquid branches.
  - Under L2: most banks face considerable liquidity stress after 4 days of a deposit run.
  - If withdrawal is 15 percent daily: vast majority survive a business week under L1, scenario (i); 15 percent per day considered critical threshold for system liquidity (conditional on no parental support).
- Policy recommendations:
  - GFSC ensure regular reporting of liquidity positions.
  - GFSC run TD liquidity analysis regularly (daily/weekly if needed; otherwise monthly), supplemented by BU analysis.
  - Ensure data availability to support analysis.
  - Monitor parental support assumptions carefully.
  - Implement best practice liquidity frameworks in absence of central bank lender-of-last-resort.
  - Deposit guarantee system introduced is appropriate to avoid runs.

### Insurance sector — coverage, method, and scope
- Objective: align insurance STs with banking tests where possible and add insurance-specific risks.
- Two-step BU exercise:
  - 2008: initial STs for all insurance companies except pure captives (134 firms).
  - Early 2010: re-assessment for 22 firms most sensitive in 2008 (6 life insurers, 1 life reinsurer, 2 general reinsurers, 7 commercial insurers, 1 mortgage indemnity insurer, 5 PCC cells) using end-2009 data.
- TD tests not run due to insufficient off-site data and heterogeneous business models.
- Time horizon longer than banking STs reflecting insurance business duration and behavioral considerations.

### Insurance sector — shock types and Box 7 tests
- Insurance tests parallel bank shocks for market risk and add insurance-specific shocks:
  - A4: Residential and commercial property prices decline by 20 percent.
  - M1: Mortality permanent increase 25 percent (insured lives only).
  - M2: Morbidity permanent increase 25 percent.
  - M3: Longevity permanent reduction in annuitant mortality 25 percent (annuitants only).
  - M4: Persistency: 30 Percent of policyholders surrender contracts where surrender right exists.
- Valuation caveats: tests based on statutory minimum capital requirements; liabilities not market-based — baseline distortions possible.
- Recommendation: move to OSCA-based BU tests for risk-based solvency; collect FSIs and data to support transition to a risk-based regime.

### Insurance sector — results and selected numeric outcomes (end-2009; sample of 15 firms)
- Sample focus: 15 firms out of 22 with coverage ratios < 200 percent.
- Pre-shock Excess of Asset over Liabilities:
  - Simple average = 32.0; Weighted average = 30.6; Min. = 0.7; Max. = 173.0.
- Interest Rate Risk:
  - I1 (200 bp upward): # of firms < 0 perc. = 0; Simple average = 32.2; Weighted average = 30.6; Min. = 0.9; Max. = 173.0.
  - I2 (200 bp downward): # of firms < 0 perc. = 1; Simple average = 30.9; Weighted average = 30.6; Min. = -3.0; Max. = 173.0.
  - I4 (300 bp downward): # of firms < 0 perc. = 2; Simple average = 30.4; Weighted average = 30.3; Min. = -9.2; Max. = 173.0.
- FX Risk:
  - F1 (20% pound appreciation): # of firms < 0 perc. = 0; Simple average = 32.2; Weighted average = 32.0; Min. = 0.8; Max. = 179.6.
  - F2 (20% pound depreciation): # of firms < 0 perc. = 0; Simple average = 32.1; Weighted average = 29.6; Min. = 0.9; Max. = 173.0.
  - F3 (30% pound appreciation): # of firms < 0 perc. = 0; Simple average = 35.2; Weighted average = 32.7; Min. = 0.8; Max. = 183.4.
- Asset Price Risk:
  - A1 (35% shares decline): # of firms < 0 perc. = 1; Simple average = 34.2; Weighted average = 30.8; Min. = -5.9; Max. = 173.0.
  - A3 (rated securities downgraded by four categories): # of firms < 0 perc. = 2; Simple average = 32.6; Weighted average = 28.1; Min. = -7.4; Max. = 173.0.
- Insurance shocks (M1–M4): no instance where >0 firms fell below 0 percent except as noted in combined scenarios.
- Scenario Analysis:
  - I4 & F2 & A2 & M1: # of firms < 0 perc. = 3; Simple average = 24.5; Weighted average = 28.5; Min. = -31.1; Max. = 173.0.
  - I4 & F2 & A2 & M1 & Buffer for other risks (10 percent of OSCA econ. capital): # of firms < 0 perc. = 3; Simple average = 23.7; Weighted average = 27.3; Min. = -31.9; Max. = 173.0.
- OSCA comparison:
  - Economic capital (OSCA) exceeds statutory minimum levels by 200 percent for 2009 stress test sample.
  - Total OSCA capital requirements were 47 percent of total available capital for the 2009 test sample of 15 firms with highest risk.

### Key policy recommendations (banking and insurance)
- Banking:
  - Calibrate TD FX stress tests using a rule of thumb based on BU outcomes.
  - Closely monitor evolution of FX risk and marketable securities exposures in times of market uncertainty.
  - Continue regular interest-rate-risk testing given low-rate environment.
  - Improve bank-internal stress test sensitivity (ICAAP focus) to identify credit risk earlier.
  - Monitor provisioning trends carefully as provisioning fell in 2009 despite rising NPLs.
  - Monitor parent-bank solvency and liquidity using FSIs and cooperate with home supervisors.
  - Run regular, conservative large-exposure and concentration tests; ensure off-site data availability for meaningful tests.
  - Ensure GFSC receives regular liquidity reports; run TD liquidity risk analysis regularly; consider best practice liquidity frameworks and the deposit guarantee system.
  - Operational risk STs should be based on economic models (ICAAP).
- Insurance:
  - Develop OSCA-based BU tests to move to risk-based solvency tests.
  - Standardize solvency measures, methods, and scenario definitions to allow TD monitoring.
  - Collect FSIs and relevant data to support transition to Solvency II or equivalent.
  - Monitor operational risk, reinsurance risk, expense risk, and new business risk in future STs; simulate reinsurance losses at maximum historical levels or double worst historical provision levels.
  - Consider explicit capital buffer for operational and non-covered risks (reference used: 10 percent of OSCA economic capital).

*Source: Executive Summary and stress-testing chapter, _cr1104 (IMF staff note; GFSC data and staff estimates).*

### Executive Summary ......................................................................................................

### Executive Summary

### Purpose and approach
- The purpose of stress testing (ST) was to assess vulnerabilities in the Guernsey banking and insurance sectors through simulation of the effect of a potential double dip recession on solvency of Guernsey banks and insurance companies.
- STs were carried out in a manner consistent with those applied in Financial Sector Assessment Program (FSAP) Updates for the other two Crown Dependencies in 2008, but reflected actual experience of institutions and markets in the financial crisis; in practice the Guernsey tests were slightly more conservative than in the other Crown Dependencies.
- Tests were applied against end-2009 financial positions, reflecting recovery in capital adequacy since the peak of the crisis (bank capital adequacy ratios (CARs) at around 19 percent, up from 15 percent at end 2008).
- The ST program combined bottom-up (BU) STs run by Guernsey banks/insurers and top-down (TD) STs carried out by the mission in collaboration with the Guernsey Financial Services Commission (GFSC).
- Risk types tested: interest rates, exchange rates, credit quality, asset price risk, operational risk, liquidity risk; for insurance, (life) insurance risks were also assessed.
- Single-factor sensitivity tests were pooled into a multivariate scenario intended to capture an external macroeconomic shock (a global double dip stress scenario), reflecting an extreme yet plausible joint shock from a major linked economy (particularly the United Kingdom (UK)).

### Key contextual indicators (selected from end-2009 data and historical comparisons)
- Banking sector capital adequacy: CARs at around 19 percent (up from 15 percent at end 2008).
- Return on equity (ROE): about 15.8 percent in 2009 (compared with below 11 percent in 2008).
- Nonperforming loans (NPLs) to total gross loans in 2009: 0.09 percent.
- Sum of ten largest exposures to capital in 2009: 132.9.
- Liquid assets to total assets in 2009: 61.3 percent.
- Deposits as a percentage of assets in 2009: 86.5 percent.
- Cross-border loans to total loans in 2009: 90.3 percent; domestic loans 9.7 percent.
- Total foreign currency-denominated loans to total loans in 2009: 73.4 percent.
- Net open position in foreign exchange to capital in 2009: 1.5.

### Stress tests for the Guernsey banks — Method
- Tests comprised both BU STs (based on STs run by Guernsey banks) and TD STs run by the mission with the GFSC.
- STs assessed sensitivity to single-factor shocks and also pooled shocks into a multivariate scenario (double dip).
- Tests used institutions’ financial positions at end 2009 to assess potential risks during 2010.

### Stress tests for the Guernsey banks — Findings: Sensitivity and scenario analysis
- Sensitivity Analysis: Guernsey banks are most vulnerable with respect to credit risk (resulting from parent exposure and large exposures) as well as foreign exchange (FX) rate risk.
- Scenario Analysis: Three banks would need recapitalization in case of a severe macroeconomic shock (e.g., a double dip scenario).
- Liquidity Risk: Banks are vulnerable if daily withdrawals of all deposits (with a maturity of up to one month) exceed 15 percent per day (relative to the remaining balance) for 5 days (i.e., 56 percent in cumulative terms).

### Stress tests for the Guernsey banks — Conclusions and recommendations
- Conclusions:
  - The banking sector is relatively robust due to the high level of capitalization, but prudence is warranted given vulnerability to low frequency/high impact events.
- Recommendations:
  - Future STs should be risk-based (i.e., Risk-Weighted Assets should be calculated based on Value-at-Risk methods).
  - Off-site data to run TD stress tests must be improved, especially to monitor large exposures.
  - Macro-prudential analysis (including STs) should be run by the authorities on a regular basis.
  - In putting in place the above measures, the authorities should consider the appropriate supervisory response to the risks faced by banks in a pro-active manner.
  - Appropriate pro-active measures should be taken to reduce risk where deemed appropriate.

### Stress tests for Guernsey insurers — Method
- BU STs were run based on standardized STs carried out by all relevant Guernsey insurers (those of systemic importance and/or with the riskiest balance sheet structures).

### Stress tests for Guernsey insurers — Findings: Sensitivity and scenario analysis
- Sensitivity Analysis: Guernsey insurers are most vulnerable to changes in interest rates and asset prices. Operational risk is also significant, although not explicitly tested for (but reflected in insurers’ economic capital calculations).
- Scenario Analysis: Up to three Guernsey insurers could be adversely affected if a severe macroeconomic shock (double dip) were to materialize. Appropriate pro-active measures should be taken to reduce risk.

### Stress tests for Guernsey insurers — Conclusions and recommendations
- Conclusions:
  - The Guernsey insurance sector appears relatively robust due to a high level of capitalization, but extreme shocks could trigger stress and contagion.
- Recommendations:
  - Future STs should be risk-based (i.e., capital requirements should be based on economic measures rather than statutory ones).
  - Data to run on- and off-site STs should be improved as a high priority.
  - Insurers should be encouraged to further develop internal STs and the GFSC should apply pro-active risk mitigation measures as deemed appropriate.

*Source: Executive Summary, _cr1104 - Executive Summary*

### 12.      The remainder of this note explains the stress testing methodology and results in

### 12. The remainder of this note explains the stress testing methodology and results in more detail

### II. BANKING SECTOR — A. Coverage
- The STs covered 19 subsidiaries of foreign banks (out of 21 on the island), except liquidity tests which included in addition 5 foreign branches (out of 21).
- The 19 subsidiaries included account, in principle, for the total assets of the Guernsey subsidiaries (i.e., 36 percent of the total assets of the banking sector).
- Ownership composition of the 19 subsidiaries sample:
  - Seven subsidiaries belonged to U.K. banking groups and accounted for 40 percent of the total assets of the subsidiaries included in the solvency test sample.
  - Three subsidiaries belonged to EU banking groups (two based in Cyprus and one in the Netherlands) and accounted for 4 percent of the solvency test sample.
  - Five subsidiaries belonged to Swiss banks and represented 18 percent of the sample.
  - One subsidiary belonged to a Bermuda based bank and represented 3 percent of the sample.
  - Two subsidiaries belonged to North America based banks: one based in Canada (20 percent of the sample) and one in the United States (10 percent of the sample).
  - One subsidiary belonged to a South African group and represented 5 percent of the sample.
- Table 2 (Balance Sheet of Banking System, end 2009) key aggregate figures:
  - Total assets: 135,803 (GBP millions); Percent of total assets: 100.0; Percent of GDP: 7,893.9.
  - Cash: 250.0 (GBP millions); Percent of total assets: 0.1; Percent of GDP: 1.5.
  - Loans to Banks: 75,605 (GBP millions); Percent of total assets: 55.7; Percent of GDP: 4,394.7.
    - Loans to Parent: 56,464 (GBP millions); Percent of total assets: 41.9; Percent of subgroup: 75.3; Percent of GDP: 3,282.1.
    - Loans to fellow banking subsidiaries: 16,619 (GBP millions); Percent of total assets: 12.3; Percent of subgroup: 22.3; Percent of GDP: 966.0.
    - Loans to other banks: 2,522 (GBP millions); Percent of total assets: 1.9; Percent of subgroup: 3.4; Percent of GDP: 146.6.
  - Marketable Assets: 24,353 (GBP millions); Percent of total assets: 17.2; Percent of subgroup: 100.0; Percent of GDP: 1,415.6.
  - Loans and Advances: 25,710 (GBP millions); Percent of total assets: 18.9; Percent of subgroup: 100.0; Percent of GDP: 1,494.5.
    - Other non-banking financial institutions: 18,128 (GBP millions); Percent of total assets: 13.5; Percent of subgroup: 71.1; Percent of GDP: 1,053.7.
    - Sovereigns: 30.0 (GBP millions); Percent of total assets: 0.0; Percent of subgroup: 0.0; Percent of GDP: 0.1.
    - Public sector entreprises: 690.1 (GBP millions); Percent of total assets: 0.3; Percent of subgroup: 4.0; Percent of GDP: 34.0.
    - Corporate Lending: 2,167 (GBP millions); Percent of total assets: 1.6; Percent of subgroup: 8.5; Percent of GDP: 125.9.
    - Retail Lending: 2,541 (GBP millions); Percent of total assets: 1.9; Percent of subgroup: 10.0; Percent of GDP: 147.7.
    - Residential Mortgages: 2,730 (GBP millions); Percent of total assets: 2.0; Percent of subgroup: 10.7; Percent of GDP: 158.7.
    - Capital Connected Lending: 730.1 (GBP millions); Percent of total assets: 0.1; Percent of subgroup: 0.3; Percent of GDP: 4.2.
  - Investments: 6,791 (GBP millions); Percent of total assets: 5.0; Percent of subgroup: 100.0; Percent of GDP: 394.7.
    - Non Marketable debt - Other non-banking financial institutions: 5,987 (GBP millions); Percent of total assets: 4.4; Percent of subgroup: 88.9; Percent of GDP: 348.0.
    - Non Marketable debt - Corporate: 155.0 (GBP millions); Percent of total assets: 0.1; Percent of subgroup: 2.3; Percent of GDP: 9.0.
    - Capital Investments in Subsidiaries and other associated companies: 210.0 (GBP millions); Percent of total assets: 0.0; Percent of subgroup: 0.3; Percent of GDP: 1.2.
    - Equity - Corporate: 274.0 (GBP millions); Percent of total assets: 0.2; Percent of subgroup: 4.1; Percent of GDP: 15.9.
  - Other Financial: 3,144 (GBP millions); Percent of total assets: 2.3; Percent of subgroup: 100.0; Percent of GDP: 182.8.
    - All past due assets: 760.1 (GBP millions); Percent of total assets: 0.1; Percent of subgroup: 2.4; Percent of GDP: 4.4.
  - Other: 175.0 (GBP millions); Percent of total assets: 0.1; Percent of subgroup: 100.0; Percent of GDP: 10.2.
    - Intangible assets including goodwill: 870.1 (GBP millions); Percent of total assets: 0.1; Percent of subgroup: 49.9; Percent of GDP: 5.1.
  - Liabilities:
    - Deposits due to: 114,751 (GBP millions); Percent of total assets: 84.5; Percent of subgroup: 100.0; Percent of GDP: 6,670.2.
      - Parent/Holding Company or Group: 65,508 (GBP millions); Percent of total assets: 48.2; Percent of subgroup: 57.1; Percent of GDP: 3,807.9.
      - Retail Accounts: 9,288 (GBP millions); Percent of total assets: 6.8; Percent of subgroup: 8.1; Percent of GDP: 539.9.
      - All Other Depositors: 39,955 (GBP millions); Percent of total assets: 29.4; Percent of subgroup: 34.8; Percent of GDP: 2,322.5.
    - CDs and Other Debt: 14,312 (GBP millions); Percent of total assets: 10.5; Percent of subgroup: 100.0; Percent of GDP: 831.9.
    - Creditors & Accruals etc: 4,467 (GBP millions); Percent of total assets: 3.3; Percent of subgroup: 100.0; Percent of GDP: 259.6.
    - Capital: 2,273 (GBP millions); Percent of total assets: 1.7; Percent of subgroup: 100.0; Percent of GDP: 132.1.
  - Sources: GFSC and staff estimates.
  - Memo: Percent of end 2008 GDP (Source: Guernsey, Policy Council).

- For short-term liquidity vulnerability assessment five large foreign branches were added; these branches represented 62 percent of Guernsey banking sector assets and 69 percent of the liquidity risk test sample.
- Table 3 (Financial System Structure, 2003-09) high-level figures:
  - Total financial system assets (2009 column): 1,409 (GBP billion); Percent of total assets: 312.3; Percent of GDP: 18,153 (GBP millions).
  - Banking Sector (2009): Number 44; Assets 135.8 (GBP billion); Percent of total assets: 43.5; Percent of GDP: 7,894.
  - Insurance sector (2009): Number 396; Assets 21.0 (GBP billion); Percent of total assets: 6.7; Percent of GDP: 1,223.
  - Other financial institutions (2009): Number 969; Assets 155.5 (GBP billion); Percent of total assets: 49.8; Percent of GDP: 9,037.
  - Memo items include number of employees in the financial sector: 6,975; Total employment: 32,332; Financial sector contribution to GDP (GBP m): 590; Total GDP (GBP millions): 1,685.
  - Source: Guernsey Financial Services Commission.

### II. BANKING SECTOR — B. Method
- Two stress testing approaches used:
  - Top-Down (TD) approach:
    - Mission performed tests for individual institutions based on reported supervisory (off-site) data collected by the GFSC.
    - Comprised same tests applied to each bank separately and results aggregated to system level.
    - Advantage: aggregation based on consistent bank-specific test results.
  - Bottom-Up (BU) approach:
    - Individual institutions conducted standardized tests based on their own stress testing calculation framework and internal data.
    - GFSC helped ensure consistency across institutions.
    - Many models based on parent banks’ risk management frameworks; in some cases tests run by parent banks.
    - Mission engaged with selected banks to understand data, assumptions, and methods and to ensure consistency between TD and BU tests.
- Test implementation and timing:
  - Shocks assumed to occur instantaneously; impact estimated individually assuming all other variables remain constant.
  - Impact estimated by calculating effect on end 2009 positions.
  - Credit risk figures are of annual frequency and profit is assumed to be a first line of defense; tests provide an outlook of what could happen to banks’ solvency during 2010.
- Caveats and limitations:
  - Tests do not model behavioral feedbacks or international/parental reactions; such reactions could mitigate or worsen effects (e.g., raising capital vs. “fire sales”).
  - No single agreed model exists to simulate market participants’ reactions; expert judgment required.
  - An explicit test was added to simulate loss of parental exposure.
  - Banks use simplified risk management models for regulatory purposes (Basel II Standardized Approach), so certain risk sensitivities may not be captured.
  - Mission applied additional (Basel II Internal-Rating based type) sensitivity analyses to avoid underestimating risks.

### II. BANKING SECTOR — C. Metrics and Calibration of Shocks
- Outcome metric:
  - Shocks applied to banks’ business at end 2009; outcomes expressed in post-shock capital adequacy levels compared with GFSC minimum capital adequacy level: 10 percent.
  - TD test based on minimum capital requirements including the Pillar 2 charge set by the GFSC.
  - In case of breach, outcome shows recapitalization needs in absolute amounts (in million GBPs) and as a percentage of GDP.
- Calibration principles:
  - Shocks calibrated to represent “extreme but plausible” events, broadly corresponding to calibrations used in recent FSAPs for comparable jurisdictions and the 2008 tests for Jersey and the IOM, with some more severe stress levels observed during the crisis.
- Liquidity risk calibration:
  - Tests assessed how many days institutions would withstand a run on deposits assuming capital markets not available.
  - Point of reference: survive five days with liquidity position at each point in time.

### II. BANKING SECTOR — D. Results of Solvency Tests
- Presentation of results (Table 4 structure described):
  - Table shows shocks (first column) with information on pre-shock capitalization level.
  - Columns 2–7: BU stress test results: asset-weighted average CAR after stress (column 2); minimum and maximum post-shock CAR for a single bank (columns 3–4); number of banks breaching GFSC minimum CAR (10 percent, column 5); recapitalization needs in absolute terms (column 6) and relative to GDP (column 7).
  - Recapitalization needs calculated as capital required to bring bank’s CAR back to 10 percent minimum.
  - Next six columns show same information for TD tests.
- Comparison BU vs TD:
  - BU and TD tests produce similar overall results and identify same most relevant risks.
  - TD results range tends to be wider for some risk types (interest rate risk, asset price risk, credit risk) and narrower for FX tests.
  - Differences stem from more detailed information and different methods in BU and TD exercises.
- Conservatism and offsets in test design:
  - Tests defined to reflect the crisis, presenting a conservative picture of potential shock impacts.
  - Offsetting less-conservative assumptions included:
    - Profits were assumed to be a first line of defense in general; but TD stress tests simulated a shock assuming no profits available as first line of defense.
    - For credit risk, some banks had not observed any losses in 2009; this was taken as baseline for test 1 and 2 (C1, C2), so no simulated losses for those banks in those tests.
    - Follow-up analyses relaxed some simplifying assumptions, moving to a Basel II Internal Rating-based approach for credit risk.
- Core findings:
  - The tests identify credit risk as the most relevant risk.
  - An explicit test was performed to simulate the loss of parental exposure given reliance on parents and potential non-support in extreme circumstances.

### Market risk (interest rate, foreign exchange rate, and asset price risks)
- General observations:
  - Market risk treated as instantaneous shocks on balance sheet and capital adequacy ratios.
  - Guernsey banks do not have trading books; therefore limited vulnerabilities to market risk.
- BU market risk tests covered:
  - Interest rate risk, foreign exchange rate risk, and asset price risk.
  - Shocks applied to yield curves and exchange rates and to asset prices:
    - Interest rate shocks: I1–I12 include parallel shifts of pound sterling, US dollar, and euro yield curves by ±200 and ±300 basis points.
    - Foreign exchange shocks: F1–F8 include sterling and dollar depreciations/appreciations by 20 percent and 30 percent against all other currencies.
    - Asset price shocks: A1 (equity prices decline by 35 percent), A2 (rated securities downgraded by two notches), A3 (rated securities downgraded by four notches).
- Box 1 (Market Risk Shocks) specification (verbatim):
  - Interest rate risk:
    - I1: Parallel upward shift of the pound sterling yield curve by 200 basis points.
    - I2: Parallel downward shift of the pound sterling yield curve by 200 basis points.
    - I3: Parallel upward shift of the pound sterling yield curve by 300 basis points.
    - I4: Parallel downward shift of the pound sterling yield curve by 300 basis points.
    - I5: Parallel upward shift of the dollar yield curve by 200 basis points.
    - I6: Parallel downward shift of the dollar yield curve by 200 basis points.
    - I7: Parallel upward shift of the dollar yield curve by 300 basis points.
    - I8: Parallel downward shift of the dollar yield curve by 300 basis points.
    - I9: Parallel upward shift of the euro yield curve by 200 basis points.
    - I10: Parallel downward shift of the euro yield curve by 200 basis points.
    - I11: Parallel upward shift of the euro yield curve by 300 basis points.
    - I12: Parallel downward shift of the euro yield curve by 300 basis points.
  - Foreign Exchange Rate Risk:
    - F1: The pound depreciates by 20 percent against all other currencies.
    - F2: The pound appreciates by 20 percent against all other currencies.
    - F3: The pound depreciates by 30 percent against all other currencies.
    - F4: The pound appreciates by 30 percent against all other currencies.
    - F5: The dollar depreciates by 20 percent against all other currencies.
    - F6: The dollar appreciates by 20 percent against all other currencies.
    - F7: The dollar depreciates by 30 percent against all other currencies.
    - F8: The dollar appreciates by 30 percent against all other currencies.
  - Asset Price Risk:
    - A1: Prices of all shares listed on foreign stock markets decline by 35 percent.
    - A2: Rated securities are downgraded by two notches.
    - A3: Rated securities are downgraded by four notches.
- Treatment in tests:
  - Interest rate risk: effects simulated on income; losses (gains) directly deducted (added) from capital — conservative treatment.
  - FX rate risk and asset price risk: effects simulated on risk-weighted assets.
- Market risk test outcomes:
  - Guernsey banking system exhibits limited exposure to market risk shocks.
  - Most notable effect observed for FX rate risk; interest rate risk limited.
  - Asset price risk shocks, particularly A3 (downgrade of all rated securities by 4 notches), are relatively the most conservative (least plausible) but would not threaten the system.

*Source: IMF staff note (stress testing methodology and results, banking sector sections as presented).*

### 29.      The banks are highly exposed to FX risks in gross terms. This is because

### _cr1104 - 29.      The banks are highly exposed to FX risks in gross terms. This is because

### FX risk — exposures and stress-test outcomes
- 75 percent of the total loans are foreign-currency denominated.
- 79 percent of the liabilities are denominated in foreign currency.
- Off-balance sheet Forex contracts amount to 26 percent of the assets (used to hedge FX risk, mostly via group treasury).
- Because the majority but not all banks’ capital is denominated in pound sterling, a depreciation of the sterling and an appreciation of major foreign currencies (notably US dollars) negatively affect capital and solvency.
- Bottom-up (BU) stress tests by Guernsey banks (F1, F3, F6, F8) findings:
  - In the most severe shock scenarios (F3, F8), where the pound sterling (F3) and the US dollar (F8) move by 30 percent, the capital adequacy of up to four banks falls below 10 percent and recapitalization needs amount to almost 3 percent of GDP.
  - For instantaneous FX rate changes of 20 percent, the effect is considerably lower, with hardly any recapitalization needs.
- Top-down (TD) tests based on open currency positions and the assumption that risk (RWAs) increases linearly with actual FX positions:
  - Identify the same most relevant shocks but simulate less severe stress than BU tests.
  - Recommendation: authorities should take this into account when designing TD FX stress events and use a rule of thumb based on BU test outcomes.
- Overall assessment: given the severity of shocks reflecting market uncertainties, the system appears robust against FX risks, but the evolution of risk should be closely monitored.

### Interest rate risk — tests and implications
- Tests examined banks’ ability to absorb sudden yield-curve changes (monetary policy shifts or spread widening).
- Both BU and TD stress tests did not show substantial movements of capital adequacy from interest-rate changes, indicating repricing risk is limited.
- TD STs referenced Economic Value-at-Risk calculated based on a 200 bps shock by each bank; for the 300 bps shock it was assumed severity is twice as high; impacts were put directly against capital without assuming profit as first line of defense.
- Caveat: interest rates were exceptionally low at the time of testing; regular tests are warranted to investigate unexpected drops in rates later.

### Marketable securities and asset-price risk
- Marketable securities held by banks at the time of the test constituted 17 percent of total assets (considerably higher than Jersey and IOM banks at the time of their last FSAPs, each below 4 percent).
- BU and TD asset-price tests:
  - A severe drop in equity prices does not trigger severe system stress.
  - A downgrade of all rated securities by two categories originates some recapitalization needs for two banks.
  - An extreme downgrade of ratings by four categories would require recapitalization for up to four banks, but:
    - (i) occurrence is very unlikely; and
    - (ii) recapitalization needs remain below 5 percent of GDP.
- Given the noteworthy extent of marketable securities held, this risk should be regularly monitored during market uncertainty.

### Credit risk — overall findings
- Credit risk constitutes about 94 percent of Pillar 1 minimum capital requirements and 92 percent of total capital requirements without including Pillar 2 concentration risk charges.
- Although historical asset quality is high (low credit losses), the absolute amount of credit relative to economy size warrants careful monitoring: lending to the real sector is 6 percent of assets but constitutes almost 400 percent in terms of GDP.
- Seven credit scenarios were run (Box 2: C1, C2, C4, C5, C6; concentration tests C3, C7). Key test designs:
  - Banks asked to increase PDs by 100 percent and 300 percent with recovery rates of 50 percent.
  - TD tests used NPLs observed in 2009 as baseline (broad NPL definition: watch list, substandard, actual losses).
  - Pillar 2 capital charges required by the GFSC were added to required capital buffer.
- Test outcomes:
  - BU tests do not suggest credit risk is a major threat.
  - TD tests show risk where PDs increase by 300 percent (C2): recapitalization needs amount to 2.2 percent of GDP with the institution most affected losing its entire capital base.
- Interpretation caveats:
  - Credit risk is a low probability–high impact risk.
  - If banks use Basel II Standardized Approach, increasing risk may only be visible once it materializes for unrated exposures; BU industry tests did not identify risk for at least one bank when PDs rose by 300 percent.
  - Some TD tests may understate risk because certain effects (e.g., RWA increases from rating downgrades) were not considered in the initial step; follow-up steps took these into account.
  - Loan Loss Provisions were not part of this study; 2009 provisions were decreasing while NPLs were increasing — provisioning should be carefully monitored going forward.

### Credit concentration risk — parent and large-exposure channels
- Upstreaming of deposits to parent banks makes concentration risk and spillovers from parent banks particularly significant.
- Test C7 (assume 10 percent of banks’ claims on their parent fail with no recovery) results:
  - Both BU and TD tests simulate capital shortfalls for about half the institutions (BU: 9 banks; TD: 13 banks).
  - Recapitalization needs amount to almost the island’s GDP in the TD test.
- Default of three largest non-parent counterparts (C3) tested with 50 percent recovery:
  - BU: several institutions vulnerable; recapitalization needs up to 14 percent of GDP and 3 banks require recapitalization.
  - TD: 6 banks require recapitalization with recapitalization needs of 22 percent of GDP.
  - Differences between BU and TD outcomes may reflect exposure calculation differences, definitions of groups, and lack of off-site collateral data.
- Recommendations:
  - Monitor solvency of parent banks rigorously using solvency FSIs (ratings, CDS spreads, KMV EDFs) and liquidity positions.
  - Maintain well-functioning cooperation with home supervisors.
  - Run large-exposure tests regularly based on conservative definitions of exposure and connected borrowers.
  - Ensure banks’ data are readily available to authorities to run meaningful off-site tests (e.g., default of the 3 riskiest exposures among the largest 10).

### Operational risk and scenario analysis
- Operational risk RWAs shock scenarios (O1, O2, O3) were included in the exercise with increasing RWA assumptions (50%, 100%, 200% increases); corresponding TD capital impacts were reported.
- Composite scenario (I4 & F3 & A2 & C2 & O2) results:
  - Bottom-up: CAR 15.4 percent; Min. -64.5 percent; Max. 73.1 percent; 3 banks breaching min. CAR; Recapitalization needs 66.3 in GBP m and 3.9 percent of GDP.

### Key policy recommendations (from the text)
- Calibrate TD FX stress tests using a rule of thumb based on BU test outcomes to enrich TD framework.
- Closely monitor evolution of FX risk and marketable securities exposures in times of market uncertainty.
- Continue regular interest-rate-risk testing given low-rate environment may change.
- Work with banks to improve sensitivity of bank-internal stress tests (ICAAP focus) to identify credit risk earlier.
- Monitor provisioning trends carefully as provisioning fell in 2009 despite rising NPLs.
- Monitor parent-bank solvency and liquidity using FSIs and cooperation with home supervisors.
- Run regular, conservative large-exposure and concentration tests and ensure authorities have access to the data needed for meaningful off-site testing.

*Sources: GFSC, and IMF staff estimates.*

### 43.      The tests also simulated a potential stress of operational risk in the system. As

### _cr1104 - 43.      The tests also simulated a potential stress of operational risk in the system. As

### Operational risk stress tests — design and key metrics
- Operational Risk constitutes 5 percent of total Pillar 1 capital requirements (based on the use of the Standardized type Basel II approaches — Basic Indicator Approach and Standardized Approach).
- The Standardized approaches do not reflect bank-specific operational risk and tend to decrease in periods of stress (with lower levels of income).
- Authorities require banks to apply Pillar 2 capital add-ons reflecting the outcome of their ICAAP processes.

- Top-Down (TD) tests simulated increases in Pillar 1 Operational Risk capital requirements:
  - O1: Increase by 50 percent.
  - O2: Increase by 100 percent.
  - O3: Increase by 200 percent.
- These correspond to simulated increases of risk-weighted assets by 7.5 percent, 10 percent and 20 percent respectively, producing declines of capital adequacy by 7, 13 and 17 percent in relative terms.
- Impact observed: limited, given generally high solvency level of the Guernsey banks.

- Recommendation: operational risk stress tests should be based on economic models (for example those developed for ICAAP) to capture bank-specific circumstances; Operational Risk might then trigger risks for specific banks in scenario-type tests.

*Source: IMF staff estimates and text.*

### Combined scenario analysis (combined risk-factor stress)
- A combined stress scenario combined five single risk shocks: I4, F3, A2, C2, and O2.
  - Single shocks combined are: the worst single-factor shocks for interest rate risk and FX risk, and severe (but not worst-case) shocks for asset prices, credit risk and operational risk.
- Under a combined stress of all risk factors the Guernsey banking system shows resilience, but:
  - Three banks fall short of capital under a scenario of medium to high severity (compared to at most one bank in any single-risk test).
- Given the assumption that single shocks have a linear effect on the combined result, recapitalization needs are only slightly higher than the sum of the shocks (as not many banks reach the recapitalization zone).
- Recommendation: authorities should consider running scenario tests in the future to obtain a comprehensive view on system solvency, since macroeconomic (or financial market) shocks usually materialize at the same time and affect multiple risk factors.

### Additional Top-Down tests for credit risk — assumptions and outcomes
- Four additional TD assumptions were tested to capture potential underestimation of credit risk:
  - Assumption A1: No profits available as a first line of defense.
  - Assumption A2: Average NPLs are assumed to be equal to the maximum of (i) the NPLs of the 2009 average for the system (0.4 Percent) and (ii) the bank-specific level.
  - Assumption A3: NPLs are assumed to be 2 Percent for all banks in the baseline case.
  - Assumption A4: RWAs are assumed to increase by 0.5 times of the relative increase of the PDs. (Rule of thumb: If PDs increase by 50 percent in relative terms, RWAs are assumed to increase by 25 percent in relative terms.)
- Key additional credit-risk test results:
  - If A1, A2, and A4 become true simultaneously and all PDs increase by 100 percent:
    - Six banks would need recapitalization (while no bank would need recapitalization without these assumptions).
    - Recapitalization needs would be at 6 percent of GDP.
  - If PDs increase by 300 percent under the same combined assumptions:
    - Most banks (15 out of 19) would be hit.
    - Recapitalization needs would amount to 28 percent of GDP.
- Sector-specific shocks:
  - Banks are relatively robust against sector-specific shocks, which could trigger recapitalization needs of around GBP 20 millions in case of shocks affecting domestic loans (C4) or mortgage portfolios (C5).
- Recommendations:
  - Authorities should take at least some of the additional assumptions into account for future STs, depending on stress level appropriateness at test time.
  - Assumption A4 will inform how capital adequacy would change under an IRB-type credit risk model and could be part of ‘standard’ tests.
  - Consider specific stress tests for high-risk exposure types (such as structured credit).
- Baseline numeric context (as shown in tables for pre-shock capital and certain scenarios):
  - Pre-shock capital: 18.5 9.8 90.8 10.9 0.1 (as presented in tabular rows of the source).
  - Example entries (preserving source formatting): C2 (Increase of all probabilities of default (PDs) on loans by 300%): 17.9 -0.8 90.8 24 0.0 2.3 (as shown in table rows).
  - A1 + A2 + A4 combined scenarios show entries such as: C2: 19.3 -0.5 93.2 138.2 2.2 (as shown in table rows).
  - Note: tables contain multiple row entries and columns with figures in GBP m and Percent of GDP (full tabular detail retained in the source).

### Liquidity risk tests — design, shocks, and findings
- Purpose: examine how many days banks could withstand a daily withdrawal of deposits with no access to additional external funding (including from parent banks).
- Test simulates a severe name crisis with liquidation restricted to most liquid assets (cash, government bonds, interbank loans with maturity up to eight days, and marketable securities).
- TD liquidity tests assumptions:
  - Simulated withdrawal: 30 percent of remaining deposits withdrawn during five consecutive days, resulting in 83 percent of deposits withdrawn within a business week.
  - Two interbank-access scenarios:
    - L1: Interbank market remains accessible (except funding with maturity of more than one month); no additional intra-group funding available.
    - L2: Interbank market more or less closed (except for maturities up to 8 days); no additional intra-group funding available.
  - Two liquid-asset scenarios:
    - (i) Liquid assets comprise cash, government bonds, claims to other banks, and other marketable securities.
    - (ii) All marketable securities except government bonds become illiquid.
- Box 6 definitions reiterated:
  - L1: daily drawdown of liabilities on demand and due within one month by 30 percent; interbank market accessible (except >1 month) and no intra-group funding available; two scenarios for liquid assets ((i) and (ii)).
  - L2: same drawdown but interbank market essentially closed (except up to 8 days); two scenarios for liquid assets ((i) and (ii)).

- Findings:
  - Under L1 (interbank funding remains available):
    - If all marketable securities are liquid (scenario i): seven subsidiaries and two branches would become illiquid during one business week.
    - If less liquid marketable securities are illiquid (scenario ii): 13 illiquid subsidiaries and 4 illiquid branches.
  - Under L2 (interbank market more or less closed):
    - Most Guernsey banks would face considerable liquidity stress after 4 days of a deposit run.
  - If withdrawal simulated is 15 percent only:
    - Vast majority of banks would survive a business week without additional external funding under L1, scenario (i), considered the critical threshold for system liquidity (conditional on no parental support).

- Tabulated example results (number of illiquid banks by day, preserving presented figures):
  - Test L1 (interbank funding remains available, no additional intragroup funding):
    - Day 1: Subsidiaries 10 / Branches 1 / Total 10 3 3
    - Day 2: 10 11 23 5
    - Day 3: 12 3 7 4 11
    - Day 4: 13 25 10 4 14
    - Day 5: 7 29 13 4 17
  - Test L2 (interbank funding essentially closed, no additional intragroup funding):
    - Day 1: 10 11 33 6
    - Day 2: 12 36 4 10
    - Day 3: 7 29 12 4 16
    - Day 4: 11 21 31 5 4 19
    - Day 5: 12 31 51 6 5 21
  - Note: table entries appear in the source as two columns for (i) All Marketable Securities are Liquid (as is Cash) and (ii) Only Government Securities Remain Liquid (and Csh) with numbers of illiquid banks by day; full tabular detail is presented in the source.

- Policy recommendations:
  - GFSC should ensure banks report liquidity positions on a regular basis.
  - GFSC should run TD liquidity risk analysis on a regular basis (if needed on a daily or weekly basis, otherwise on a monthly basis), supplemented by bank-level (BU) analysis.
  - Data should be made available in an adequate format to support such analysis.
  - Monitor parental support assumptions carefully: strong parental support can mitigate local stress, but absence of support can exacerbate local stress.
  - Implement best practice liquidity frameworks in absence of a central bank lender of last resort (examples cited in source: current Basel Committee proposal once agreed and/or liquidity framework introduced by U.K. FSA).
  - Recently introduced deposit guarantee system is identified as an appropriate response to avoid bank runs triggered by liquidity risk.

### Insurance sector stress testing — coverage, method, and scope
- Objective: align insurance sector STs with banking sector tests where possible and add specific insurance risk stresses.
- Two-step Bottom-Up (BU) exercise:
  - 2008: initial STs for all insurance companies except ‘pure captives’ (134 firms).
  - Early 2010: re-assessment for a sample of 22 firms most sensitive in 2008 (6 life insurers, one life reinsurer, 2 general reinsurance firms, 7 commercial insurers, 1 mortgage indemnity insurer, and 5 PCC cells) using end-2009 data.
- Rationale for not running Top-Down (TD) tests for insurance:
  - (i) Available off-site data do not allow meaningful TD analysis at this stage.
  - (ii) The most important companies differ significantly in business models, making standardized TD tests challenging.
  - (iii) Tests show capitalization of the system is high and risks appear limited except for specific businesses.
- Time horizon and behavioral considerations:
  - Insurance tests assess vulnerabilities over a longer horizon than banking tests, reflecting long-term nature of insurance business and importance of assumptions about behavior of insurance companies and policyholders in stress.
- Sector context and trends:
  - Sector consists of 396 firms (including 300 captives), constituting 7 percent of total assets in the Guernsey financial sector.
  - Sector total assets grew by 50 percent during the last five years (growth mostly in life sector), while total gross premiums remained at a similar level during the same period.
  - Some life insurers were hit hard by the financial crisis, with 2008 aggregate ROE of life insurance sector becoming negative (mainly driven by one firm that faced losses).
  - Some recovery during 2009, but risks remain as reflected in stress test outcomes.

- Method notes:
  - 2008 tests used end-June 2008 financial reporting data (if available) or latest available figures otherwise.
  - 2010 re-assessment for the 22-firm sample used end-2009 data.

*Source: IMF staff estimates and the text of the stress-testing chapter provided.*

### 61.      In general, a shock can have two effects on the solvency position of an insurance

### _cr1104 - 61.      In general, a shock can have two effects on the solvency position of an insurance company

### Shock transmission and solvency definitions
- A shock can have two effects on the solvency position of an insurance company:
  - an instantaneous change in the value of the company’s assets (capital gains or losses); and
  - a progressive change in the return on the company’s assets (flow effect represented by asset income level changes).  
- Stress test (ST) outcome measurement approaches described:
  - changes in companies’ solvency margin (ratio of net admissible assets to total liabilities);  
  - comparison of pre- and post-shock solvency ratio (actual solvency margin relative to the required solvency margin);  
  - pre- and post-shock policy yields.  
- This study uses the excess of asset over liabilities (liabilities include the statutory minimum solvency requirements as defined by GFSC rules), comparing pre- and post-shock levels (aligned with definition (ii) above).

### Baseline and methodological caveats
- The STs are based on statutory minimum capital requirements; valuation of liabilities is not market-based (not an economic capital regime), so:
  - the baseline scenario may be distorted: liabilities tend to be overstated and capital requirements may be understated; the starting point could overstate or understate solvency.  
  - crude methods to determine shock impacts could underestimate risk.  
- The mission recommends interpreting ST results with these distortions in mind to avoid false conclusions on system stability.

### Metrics and calibration of shocks
- OSCA (Own Solvency Capital Assessment) based BU (bottom-up) tests are recommended to move toward risk-based solvency tests and to ensure consistency across institutions.  
- Insurance sector shocks parallel bank shocks except for credit risk and operational risk:
  - credit risk exposure is relatively low and not material for STs here; operational risk is important and should be analyzed in future tests.  
  - an additional capital buffer under stress for operational and other uncovered insurance risks is considered via scenario analysis combining single risk factor tests. Operational risks were implicitly included when benchmarking with OSCA outcomes.  
- Market risk tests:
  - interest rate tests simulate parallel shifts of the pound sterling yield curve (I1-I4);  
  - foreign exchange tests use the same sterling pound tests as banks (F1-F4);  
  - asset price tests include A1–A3 and a property price fall of 20 percent (A4).  
- Insurance-specific tests include shocks on mortality, morbidity, longevity, and persistency (see Box 7). Reinsurance-specific explicit shock was not included because it is highly company-specific; recommendation: GFSC should simulate actual losses at maximum historical levels within a scenario or double the worst historical provision levels.

### Box 7 — Additional tests run for the insurance sector (definitions preserved)
- A4: Residential and commercial property prices decline by 20 percent.  
- M1: Mortality: a permanent increase in mortality rates of insured lives (i.e., mortality over normal prevailing rates) of 25 percent. This shock should only be applied to insured lives (i.e. not to annuitants).  
- M2: Morbidity: a permanent increase in morbidity rates (i.e., morbidity over normal prevailing rates) of 25 percent.  
- M3: Longevity: a permanent reduction in annuitant mortality of 25 percent across all current and deferred annuitants. This shock should only be applied to annuitants, i.e., not to insured lives.  
- M4: Persistency: 30 Percent of policyholders where there is a right to surrender, surrendered their contracts at the date of the investigation.

### Results — summary findings and key statistics
- Test sample: outcome displayed for 15 firms (out of 22) with coverage ratios of less than 200 percent (tests most meaningful for these firms).  
- Overall conclusion: Guernsey insurance sector exhibits resilience against severe shocks mainly due to:
  - comparatively low exposure to risk (considerable share of unit-linked business where risks are borne by policyholders and a low share of pure life insurance business); and
  - sound levels of capitalization.  
- Specific vulnerabilities identified:
  - exposure to a sudden decline of sterling interest rates: I2 (200 bp downward) resulted in # of firms < 0 perc. = 1; I4 (300 bp downward) resulted in # of firms < 0 perc. = 2.  
  - severe asset price shocks could trigger vulnerability: A1 (35% price decline of shares on foreign stock markets) -> # of firms < 0 perc. = 1; A3 (rated securities downgraded by four categories) -> # of firms < 0 perc. = 2.  
  - a downgrade of rated securities by four categories would make two firms undercapitalized.  
  - a past 30 percent pound sterling appreciation (F3) was important in 2008 but was not observed this time.  
- Scenario aggregation:
  - Combining worst single-factor shocks (I4, F2, A2, and M1) additively: 3 out of the 22 firms would run short of capital.  
  - Adding an additional buffer for other risks amounting to 10 percent of the economic capital requirements calculated in the OSCA exercises: no additional firm would run short of capital, but two additional firms would be close to falling short and the system overall would run the risk of going down.  
- OSCA comparison:
  - Companies identified as most vulnerable in IMF BU STs are also those with the lowest capitalization in OSCA terms.  
  - Economic capital requirements (OSCA figures) exceed statutory minimum capital requirement levels by 200 percent for the 2009 stress test sample.  
  - Total OSCA capital requirements were 47 percent of total available capital for the 2009 test sample of 15 firms with the highest risk.

### Selected numeric outcomes from Table 9 (end 2009 Data; "Excess of Asset over Liabilities" in percent)
- Actual number of firms in sample: 15 (results displayed for 15 firms out of 22 with coverage ratios < 200 percent).  
- Pre-shock level: Simple average = 32.0; Weighted average = 30.6; Min. = 0.7; Max. = 173.0.  
- Interest Rate Risk:
  - I1 (200 bp parallel upward shift): # of firms < 0 perc. = 0; Simple average = 32.2; Weighted average = 30.6; Min. = 0.9; Max. = 173.0.  
  - I2 (200 bp parallel downward shift): # of firms < 0 perc. = 1; Simple average = 30.9; Weighted average = 30.6; Min. = -3.0; Max. = 173.0.  
  - I3 (300 bp parallel upward shift): # of firms < 0 perc. = 0; Simple average = 32.0; Weighted average = 30.3; Min. = 1.0; Max. = 173.0.  
  - I4 (300 bp parallel downward shift): # of firms < 0 perc. = 2; Simple average = 30.4; Weighted average = 30.3; Min. = -9.2; Max. = 173.0.  
- FX Risk:
  - F1 (20% pound appreciation): # of firms < 0 perc. = 0; Simple average = 32.2; Weighted average = 32.0; Min. = 0.8; Max. = 179.6.  
  - F2 (20% pound depreciation): # of firms < 0 perc. = 0; Simple average = 32.1; Weighted average = 29.6; Min. = 0.9; Max. = 173.0.  
  - F3 (30% pound appreciation): # of firms < 0 perc. = 0; Simple average = 35.2; Weighted average = 32.7; Min. = 0.8; Max. = 183.4.  
  - F4 (30% pound depreciation): # of firms < 0 perc. = 0; Simple average = 34.9; Weighted average = 29.1; Min. = 0.9; Max. = 173.0.  
- Asset Price Risk:
  - A1 (35% price decline of shares on foreign stock markets): # of firms < 0 perc. = 1; Simple average = 34.2; Weighted average = 30.8; Min. = -5.9; Max. = 173.0.  
  - A2 (rated securities downgraded by two categories): # of firms < 0 perc. = 0; Simple average = 34.4; Weighted average = 29.9; Min. = 0.8; Max. = 173.0.  
  - A3 (rated securities downgraded by four categories): # of firms < 0 perc. = 2; Simple average = 32.6; Weighted average = 28.1; Min. = -7.4; Max. = 173.0.  
  - A4 (real estate prices decline by 20%): # of firms < 0 perc. = 0; Simple average = 37.7; Weighted average = 30.6; Min. = 1.9; Max. = 173.0.  
- Insurance Risk:
  - M1 (permanent 25% increase in mortality of insured lives): # of firms < 0 perc. = 0; Simple average = 34.8; Weighted average = 30.6; Min. = 0.9; Max. = 173.0.  
  - M2 (permanent 25% increase in morbidity): # of firms < 0 perc. = 0; Simple average = 37.7; Weighted average = 30.6; Min. = 1.9; Max. = 173.0.  
  - M3 (permanent 25% reduction in annuitant mortality): # of firms < 0 perc. = 0; Simple average = 37.3; Weighted average = 30.5; Min. = 1.9; Max. = 173.0.  
  - M4 (persistence: 30% surrender): # of firms < 0 perc. = 0; Simple average = 36.8; Weighted average = 34.2; Min. = 1.9; Max. = 173.0.  
- Scenario Analysis (I4 & F2 & A2 & M1):
  - # of firms < 0 perc. = 3; Simple average = 24.5; Weighted average = 28.5; Min. = -31.1; Max. = 173.0.  
- Scenario Analysis (I4 & F2 & A2 & M1 & Buffer for other risks):
  - # of firms < 0 perc. = 3; Simple average = 23.7; Weighted average = 27.3; Min. = -31.9; Max. = 173.0.

### Policy recommendations and next steps
- Develop OSCA-based bottom-up tests to arrive at risk-based solvency tests and ensure consistency across institutions.  
- Standardize tests: common solvency measure(s), comparable methods, and scenario definitions to allow economic (top-down) monitoring.  
- Collect pertinent data including FSIs to support transition to a risk-based solvency regime (Solvency II or an equivalent framework).  
- Monitor and analyze operational risk, reinsurance risk, expense risk, and new business risk in future STs; simulate reinsurance losses at maximum historical levels or double worst historical provision levels for reinsurance scenario analysis.  
- Consider adding an explicit capital buffer for operational and other non-covered risks (a reference of 10 percent of OSCA economic capital was used in scenario analysis).

*Source: IMF staff compilation from GFSC data and stress test exercise as presented in the `_cr1104` chapter (end-2009 data).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1104.pdf_