## _cr16198

## Source details

**Canonical URL:** [_cr16198](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16198.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16198.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16198.pdf.json)

---

### Executive Summary: Key Findings
- FSAP stress testing conducted in an environment of low profitability and low credit growth, creating vulnerabilities due to low margins to absorb costs or losses.
- Baseline GDP growth projected: 3.2 percent.
- Three one-year macroeconomic scenarios:
  - Baseline: IMF staff projections.
  - Adverse scenario 1: Moderate contraction driven mainly by reduced external demand from a protracted euro area slowdown and deterioration in Russia and Ukraine; includes risk of reduced FDI and asset price declines.
  - Adverse scenario 2: Severe scenario where adverse global financial market conditions compound the moderate scenario, leading to capital outflows and elevated funding costs.
- Tests and scope:
  - Solvency, liquidity, and contagion top-down stress tests for 12 banks active by end-2014.
  - Authorities did not conduct a bottom-up stress test based on banks’ internal models.
- Provisioning concerns:
  - Potentially inadequate loan-loss provisioning; strong indications of inadequate provisioning on NPLs in four banks.
  - Parallel solvency tests used adjusted capital adequacy for these banks; for three banks assessment based on qualified external auditor opinions by end-2014.
- Capital shortfalls:
  - Adverse scenarios generated capital shortfalls of 1 percent and 3.5 percent of GDP under the moderate and severe scenarios, respectively, to meet a CAR of 10 percent.
- Liquidity and funding:
  - Short-term liquidity resilience is high; long-term liquidity tests reveal structural issues—large stocks of NPLs or low-quality loans could threaten long-term funding viability in some banks.
  - Low profitability makes banks sensitive to funding cost volatility; modest funding shocks could lead to negative profits and undermine capital adequacy.
- Contagion:
  - Large foreign exposures alongside limited domestic interconnectedness; nearly half of external interbank claims are to Austria, Germany, and Hungary.
  - Contagion analysis indicates systemically important home countries for Montenegro: France, Germany, the United Kingdom, and the United States; contagion could lead to severe capital impairment of up to 10 Montenegro banks and loss of up to 83 percent of the total banking sector’s regulatory capital, primarily via second- and third-round credit shocks.
- Recommendation highlight:
  - Conduct an Asset Quality Review (AQR) by external auditors for all banks, focusing on collateral valuation and provisioning.
  - Strengthen CBM stress-testing practices by testing more extreme events.

### Objectives, Data, and Methodology
- Objective: Test the banking system’s resilience to extreme but plausible shocks; stress tests illustrate tail events and inform policy actions.
- Data and perimeter:
  - Bank-by-bank supervisory data as of December 31, 2014.
  - 12 banks active by end-2014; spillover and contagion assessment covered banks, insurance companies, investment funds, leasing companies, microcredit organizations, and other financial institutions.
- Methods:
  - FSAP-developed top-down solvency, liquidity, and contagion models.
  - Panel-data logit-transformed NPL model with bank fixed effects and up to eight-quarter lags.
  - Due to weak statistical relation between macro variables and NPLs, global rules-of-thumb sensitivities were used for adverse credit-loss projections.
  - One-year stress-test horizon.

### Banking Sector Structure, Asset Quality, and Profitability (end-2014; select metrics)
- Number of licensed banks: 12.
- Total banking sector assets: EUR 3.1 billion.
- Banking sector assets as percent of GDP: 92 percent.
- Share of assets in subsidiaries of foreign-owned banks: 79 percent.
- Profitability:
  - Aggregate return to total assets: 0.8 percent.
  - Aggregate net interest and fee income margin to total assets: 3.6 percent.
  - Overhead expenses as percent of total net interest and fee margin: 74.7 percent.
  - Number of banks making losses: 2.
- Asset composition:
  - Cash and central bank reserves: 16 percent of total assets.
  - Net loans: 70 percent of aggregate balance sheet.
  - Debt security holdings and other assets: 14 percent of total assets.
  - Domestic sovereign bonds and real estate: 7.4 percent of total assets.
  - Equity instrument holdings: 0.4 percent of total assets.
- External positions:
  - External claims: 16 percent of total assets.
  - External liabilities: 21 percent of total assets.
  - Borrowings from parent banks and development funds: 7.9 percent of total assets.
- Nonperforming loans and provisioning:
  - Aggregate NPLs: 16.8 percent of total assets.
  - Aggregate loans past due more than 90 days: 13.9 percent.
  - Aggregate provisioning (regulatory reserves considered): 87.5 percent of loans past due by more than 90 days.
  - IFRS-based provisioning: 47.8 percent of total provisions to loans past due by more than 90 days.
  - Dispersion of NPL ratios: largest NPL ratio 36 percent; lowest 7.9 percent of total gross loans.
  - One bank: NPL ratio 6.3 percent and ratio of loans past due by more than 90 days to total loans 23.8 percent.
- Capital adequacy:
  - System aggregate CAR: 16.2 percent.
  - Four banks have CARs below 15 percent; the two lowest ratios belong to domestically owned banks.
- Regulatory treatment difference vs IFRS: Montenegro deducts additional regulatory reserves from equity to calculate own funds; treatment of collateral differs between IFRS and domestic rules.

### Stress Test Scenarios (macroeconomic and calibration)
- Three scenarios (one-year horizon):
  - Baseline: GDP growth 3.2 percent (IMF staff projections as of August 2014).
  - Moderate adverse (Adverse 1):
    - real GDP growth: -2.5 percent
    - gross credit growth: minus 3 percentage points of GDP
    - funding cost rise: 1.6 percent
    - general interest rates decrease: 60 basis points
    - domestic-sovereign, foreign-sovereign, corporate bond credit spreads increase by 4.4, 1.3, and 4.7 percentage points, respectively
    - stock prices fall by 22 percent
    - real estate prices fall by 10 percent
  - Severe adverse (Adverse 2):
    - real GDP growth: -5.2 percent
    - gross credit growth: -6 percentage points of GDP
    - funding cost rise: 3.1 percent
    - general interest rates decrease: 120 basis points
    - domestic-sovereign, foreign-sovereign, corporate bond credit spreads increase by 8.7, 2.6, and 9.5 percentage points, respectively
    - stock prices fall by 43 percent
    - real estate prices fall by 20 percent
- Credit-loss calibration:
  - Due to weak model relations, adverse scenario credit-loss projections used sensitivities informed by emerging-market experience and an assumed loss-given-default (LGD) of 60 percent for computing NPL stocks implied by credit losses.
  - Reference average credit loss rates (Hardy and Schmieder (2013)): 2.2 percent (moderate), 3.4 percent (medium), 7.4 percent (severe), 14 percent (extreme) of credit outstanding; chosen sensitivities lie in moderate/medium and severe ranges.

### Solvency Stress Tests: Adjustments and Results
- Adjustments for provisioning shortfalls:
  - IMF staff increased NPL stocks and provisions in affected banks where past-due loans exceeded reported NPLs and where external auditors issued qualified opinions.
  - Total adjustments: EUR 35 million (1 percent of 2014 GDP) across four banks.
  - Provisioning adjustments reduced system aggregate CAR from 16.2 percent to 14.7 percent.
- Key solvency results (preserve reported exact figures):
  - Provisioning adjustments immediately cause capital needs before stress:
    - Capital shortfall = 0.7 percent of 2014 GDP.
    - Of this, 0.2 percentage points are needed to bring undercapitalized banks to solvency.
  - Severe adverse scenario (with provisioning adjustments):
    - Capital shortfalls range between 2.8 percent and 3.5 percent of 2014 nominal GDP.
    - After provisioning adjustments, capital needs to bring banks to 10 percent CAR = 3.5 percent of 2014 GDP, of which 1.1 percent of GDP is needed to bring undercapitalized banks to solvency.
    - Without provisioning adjustments, capital needs to meet 10 percent CAR = 2.8 percent of GDP, of which 0.3 percent of GDP needed to bring undercapitalized banks to solvency.
  - Moderate scenario:
    - With provisioning adjustments, capital needs to bring banks to 10 percent CAR = 1 percent of GDP, of which 0.4 percent of GDP to bring undercapitalized banks to solvency.
    - Without provisioning adjustments, 0.2 percent of GDP needed to meet 10 percent CAR; all banks would be solvent.
  - Baseline scenario:
    - Accumulation of profits increases system aggregate CAR overall; one bank experiences reduced capital adequacy despite minor NPL rise due to negative profits.
- Table highlights (selected figures as reported):
  - Before stress (12/31/2014, without provisioning adjustments): Regulatory capital 3159.2; RWAs 1,947 (56.9); Total assets 3,136 (91.6); CAR 16.2.
  - After Adverse 2 (12/31/2015, without provisioning adjustments): Regulatory capital 802.3; RWAs 1,590 (46.4); Total assets 2,290 (66.9); CAR 5.0; Capital shortfall to meet 10 percent CAR 952.8 (in millions of EUR).
  - With provisioning adjustments, before stress (12/31/2014): Regulatory capital 2808.2; RWAs 1,913 (55.8); CAR 14.7.
  - With provisioning adjustments, after Adverse 2 (12/31/2015): Regulatory capital 471.4; RWAs 1,562 (45.6); CAR 3.0; Capital shortfall to meet 10 percent CAR 1223.5 (in millions of EUR).

### Funding Cost and Liquidity Assumptions and Results
- Deposit outflow assumptions:
  - Severe scenario gross outflow assumptions:
    - Nonresident deposits gross outflow: 46 percent (aggregate outflow between March 2009 and March 2010).
    - Domestic deposits gross outflow: 23 percent (outflows between June 2008 and June 2009).
  - Moderate scenario assumes half the severe outflows for nonresident deposits.
  - Scenarios assume 50 percent of gross deposit outflows are replaced by funding at a similar cost.
- Translated net deposit outflows:
  - Moderate scenario: net deposit outflows of 6 percent of domestic deposits and 12 percent of nonresident deposits.
  - Severe scenario: net deposit outflows of 12 percent of domestic deposits and 23 percent of nonresident deposits.
- Deposit rate spread shocks (aggregate deposit rate spreads above three-month Euribor):
  - Severe adverse: increase by two standard deviations = 3.2 percentage points.
  - Moderate: increase by one standard deviation = 1.6 percentage points.
- Liquid asset haircuts and uncovered funding costs:
  - Haircuts: 15 percent (moderate) and 25 percent (adverse).
  - If liquid assets do not cover outflows, domestic banks face increased funding costs of 2 percentage points (moderate) and 4 percentage points (adverse) above current cost.
  - Foreign-owned banks assumed to fund uncovered outflows at half the cost using parent bank funding.
- Liquidity stress-test results (exact reported figures):
  - Aggregate reported short-term liquidity ratio (LCR, standard): 863; excluding domestic sovereign bonds: 629.
  - LCR assuming severe adverse scenario outflows: Total 525; excluding domestic sovereign bonds 382.
  - System aggregate long-term liquidity ratio (NSFR, standard): 124; two banks below 100 percent.
  - NSFR after severe adverse deposit outflows: Total 108; seven banks below 100 percent.
  - Implied liquidity shortfall (NSFR standard): 10; under severe scenario: 95.

### Sensitivity Analyses (selected)
- Concentration and single-name defaults:
  - Default of largest borrower in each bank with no recovery: capital needs = 0.3 percent of GDP to bring undercapitalized banks’ CARs to 10 percent (text summary).
  - Default of five largest borrowers in each bank with no recovery: capital needs = 3.2 percent of GDP to bring banks to 10 percent CAR (text summary).
  - Table 12 concentration results (selected reported cells, recovery rates):
    - Assumed recovery rate 70 percent: CAR percent after default of largest borrower = 15.3; capital shortfall to meet 12 percent CAR = 519 (in millions of EUR).
    - Assumed recovery rate 35 percent: CAR percent after default of largest borrower = 14.4; capital shortfall to meet 12 percent CAR = 1375 (in millions of EUR).
- Market risk sensitivities (aggregate impacts):
  - 10 percent haircut on domestic sovereign exposures → losses = 1.5 percent of GDP; aggregate CAR falls from 16.2 percent to 13.5 percent; total capital shortfall = 0.4 percent of GDP.
  - 20 percent haircut → losses = 3.1 percent of GDP; aggregate CAR falls to 10.8 percent; total capital shortfall = 1.5 percent of GDP.
  - 35 percent real estate price decline → losses = 0.9 percent of GDP; aggregate CAR declines from 16.2 percent to 14.5 percent.
  - 500 basis point interest rate rise → loss = 0.2 percent of GDP; aggregate CAR declines from 16.2 percent to 15.8 percent.
  - 35 percent equity price fall → losses = 0.1 percent of GDP; aggregate CAR declines from 16.2 percent to 15.9 percent.
- Operational risk:
  - Tripled operational losses (adverse) amount to 0.21 percent of total assets.
  - Historical highest annualized operational loss 2010–2014: 0.15 percent of total assets (excluding one outlier).
  - Two-standard-deviation operational loss → losses = 0.02 percent of GDP; aggregate CAR reduces from 16.2 percent to 16.1 percent.

### Contagion Risks and Cross-Border Spillovers
- Cross-border exposures and composition:
  - Banks’ foreign claims and liabilities account for over 115 percent and 50 percent of total regulatory capital, respectively, as of end-2014.
  - About 47 percent of total interbank external claim is to Austria, Germany, and Hungary.
- Network analysis framework:
  - Espinosa-Vega and Solé (2010) methodology; credit shock (LGD 100 percent on interbank exposures) and funding shock (50 percent haircut on asset fire sales; 65 percent roll-over ratio on interbank debt).
  - Bank failure defined as capital-to-RWA falling below CAR requirement of 10 percent.
- Spillover results:
  - Credit shock dominates funding shock in impact on Montenegro banks.
  - Large contagion risks identified from France, Germany, the United Kingdom, and the United States.
  - Under combined credit and funding shocks:
    - Severe capital impairment in 7 and 10 Montenegro banks (text reports two variants).
    - Up to 83 percent of the overall banking system’s regulatory capital could be lost.
  - Large home-country shocks to foreign-owned banks tended to produce confined failures rather than system-wide collapse for most home countries, except for France.

### Main Policy Recommendations (summary)
- Short-term:
  - Conduct Asset Quality Review (AQR) of all banks with focus on the adequacy of loan loss provisioning and collateral valuation. — Timeline: Short-term.
  - Introduce more severe scenarios in CBM stress tests to assess the system’s resilience to extreme but plausible shocks. — Timeline: Short-term.
  - Increase transparency around banking system capital adequacy by publishing information about regulatory reserves and own funds as memorandum items in the official quarterly bank-specific balance sheets. — Timeline: Short-term.
- Medium-term:
  - Develop stress testing techniques at CBM to make bank-specific stress tests taking into account all relevant credit, market and liquidity risks. — Timeline: Medium-term.
  - Regular compilation of bank-specific LCR in euro. — Timeline: Medium-term.
  - Regular compilation of bank-specific NSFR in euro. — Timeline: Medium-term.

### Key Risks from the Preliminary Risk Assessment Matrix (RAM)
- Principal threats (staff assessments):
  - Protracted growth slowdown in the Euro area and neighboring countries — Staff assessment: High; Impact: High.
  - Spillovers from a deterioration of global financial market conditions — Staff assessment: High; Impact: High.
  - Geopolitical and macroeconomic events associated with Russia/Ukraine tensions and Greece’s negotiations — Staff assessment: Medium; Impact: Medium.
- Expected channels:
  - Reduced external demand, lower FDI, tourism declines, sovereign funding-cost increases, deposit outflows, higher funding costs for banks, and credit contraction with adverse feedbacks on asset quality and solvency.

*Source: IMF Staff.*

### EXECUTIVE SUMMARY ___________________________________________________________________________ 5

### _cr16198 - EXECUTIVE SUMMARY ___________________________________________________________________________ 5

### Executive Summary: Key Findings
- The FSAP stress testing exercise was conducted in an environment of low profitability and low credit growth, creating vulnerabilities due to low margins to absorb costs or losses.
- Baseline GDP growth projected: 3.2 percent.
- Three one-year macroeconomic scenarios were considered:
  - Baseline: IMF staff projections.
  - Adverse scenario 1: Moderate contraction driven mainly by reduced external demand from a protracted euro area slowdown and deterioration in Russia and Ukraine; includes risk of reduced FDI and asset price declines.
  - Adverse scenario 2: Severe scenario where adverse global financial market conditions compound the moderate scenario, leading to capital outflows and elevated funding costs.
- Solvency, liquidity, and contagion top-down stress tests were conducted for 12 banks active by end-2014. Authorities chose not to conduct a bottom-up stress test based on banks’ internal models.
- Solvency tests incorporated potentially inadequate loan-loss provisioning; strong indications of inadequate provisioning on NPLs in four banks. Parallel solvency tests used adjusted capital adequacy for these banks; for three banks assessment was based on qualified external auditor opinions by end-2014.
- Adverse scenarios generated capital shortfalls of 1 percent and 3.5 percent of GDP under the moderate and severe scenarios, respectively, to meet a CAR of 10 percent.
- Low profitability makes banks sensitive to funding cost volatility; modest funding shocks could lead to negative profits and undermine capital adequacy.
- Short-term liquidity resilience is high, but long-term liquidity tests reveal structural issues—large stocks of NPLs or low-quality loans could threaten long-term funding viability in some banks.
- Concentration risks may cause sudden, material impacts on capital adequacy; large single exposures in a small economy suggest need for additional capital buffers.
- Large foreign exposures exist alongside limited domestic interconnectedness; nearly half of external interbank claims are to Austria, Germany, and Hungary.
- Contagion analysis indicates systemically important home countries for Montenegro: France, Germany, the United Kingdom, and the United States. Contagion from these countries could lead to severe capital impairment of up to 10 Montenegro banks and loss of up to 83 percent of the total banking sector’s regulatory capital, primarily via second- and third-round credit shocks.
- Recommendation highlight: Conduct an Asset Quality Review (AQR) performed by external auditors for all banks, focusing on collateral valuation and provisioning; strengthen CBM stress-testing practices by testing more extreme events.

### Introduction: Objectives and Methodology
- Objective: Test the banking system’s resilience to extreme but plausible shocks; stress tests illustrate tail events and help identify policy actions.
- Data and scope:
  - Bank-by-bank supervisory data as of December 31, 2014.
  - Top-down solvency, liquidity, and contagion risk stress tests for 12 banks active by end-2014.
  - Spillover and contagion assessment covered banks, insurance companies, investment funds, leasing companies, microcredit organizations, and other financial institutions.
- Method: FSAP-developed models used; CBM chose not to run bottom-up tests based on banks’ models. Authorities shared internal satellite models to facilitate scenario and model design.

### Banks’ Balance Sheets and Profits: Sector Structure and Risks
- By end-2014:
  - Number of licensed banks: 12.
  - Total banking sector assets: EUR 3.1 billion.
  - Banking sector assets as percent of total financial system assets: 88 percent.
  - Banking sector assets as percent of GDP: 92 percent.
  - Share of banking system assets in subsidiaries of foreign-owned banks: 79 percent.
- Profitability:
  - Aggregate return to total assets: 0.8 percent.
  - Aggregate net interest and fee income margin to total assets: 3.6 percent.
  - Overhead expenses as percent of total net interest and fee margin: 74.7 percent.
  - Number of banks making losses: 2.
- Asset composition:
  - Cash and central bank reserves: 16 percent of total assets.
  - Net loans: 70 percent of aggregate balance sheet.
  - Debt security holdings and other assets: 14 percent of total assets.
  - Domestic sovereign bonds and real estate: 7.4 percent of total assets.
  - Equity instrument holdings: 0.4 percent of total assets.
  - No net FX exposure to a single currency exceeds 0.1 percent of total assets (reported data).
- External positions:
  - External claims: 16 percent of total assets.
  - External liabilities: 21 percent of total assets.
  - Borrowings from parent banks and development funds: 7.9 percent of total assets.
- Nonperforming loans and provisioning:
  - Aggregate NPLs: 16.8 percent of total assets.
  - Aggregate loans past due more than 90 days: 13.9 percent.
  - Aggregate provisioning (regulatory reserves considered): 87.5 percent of loans past due by more than 90 days.
  - IFRS-based provisioning: 47.8 percent of total provisions to loans past due by more than 90 days.
  - Dispersion of NPL ratios: largest NPL ratio 36 percent; lowest 7.9 percent of total gross loans.
  - One bank: NPL ratio 6.3 percent and ratio of loans past due by more than 90 days to total loans 23.8 percent.
- Capital adequacy:
  - System aggregate CAR: 16.2 percent.
  - Four banks have CARs below 15 percent, with the two lowest ratios belonging to domestically owned banks.
- Regulatory treatment:
  - Domestic legislation deducts additional regulatory reserves from equity to calculate own funds for CAR.
  - Main difference vs IFRS: treatment of collateral—IFRS deducts expected collateral value from exposure before provisioning; Montenegro rules multiply full exposure by provisioning rate depending on collateral and borrower credit quality.
- Concerns:
  - Low IFRS provisioning and reliance on collateral valuation raise concerns about adequacy of asset valuation across the system.

### Solvency Stress Test: Scenarios, Adjustments, and Results (summary)
- Scenarios:
  - Baseline: GDP growth 3.2 percent.
  - Adverse 1: Moderate contraction from external demand reduction; includes reduced FDI and asset price declines.
  - Adverse 2: Severe, adds global financial market deterioration with capital outflows and elevated funding costs.
- Accounting for provisioning weaknesses:
  - Strong indications of inadequate provisioning on NPLs in four banks; parallel tests used adjusted capital adequacy.
  - For three banks, mission’s assessment based on qualified external auditor opinions by end-2014.
- Results:
  - Moderate and adverse scenarios generate capital shortfalls of 1 percent and 3.5 percent of GDP, respectively, to meet a CAR of 10 percent.
  - Low profitability amplifies sensitivity to funding cost shocks; funding shocks can produce negative profits and erode capital adequacy.
- Sensitivity analyses conducted on concentration risk, market risk, and operational risk (details in main report).

### Liquidity Stress Test: Findings
- Short-term liquidity resilience: High.
- Long-term liquidity: Structural issues apparent; large stock of NPLs or low-quality loans could threaten stability of long-term funding in some banks.
- If loan-book quality is not improved, attracting funding at viable cost in a stressed scenario may be difficult for some banks.

### Contagion Risks and Spillovers
- Domestic interconnectedness: Limited domestic interbank transactions and limited exposure to nonbank financial institutions.
- Cross-border interconnectedness: High, reflecting foreign ownership and investment strategies.
  - Nearly half of external interbank claims are to Austria, Germany, and Hungary.
- Contagion analysis (Espinosa-Vega and Solé, 2010-based network model):
  - Banking systems of France, Germany, the United Kingdom, and the United States are systemically important for Montenegro.
  - Contagion from these countries could:
    - Cause severe capital impairment of up to 10 Montenegro banks.
    - Result in loss of up to 83 percent of the total banking sector’s regulatory capital.
  - Such impacts arise mainly from credit shocks and are predominantly second- and third-round effects.
  - Large home-country shocks to foreign-owned banks tended to produce confined failures rather than system-wide collapse due to limited domestic interbank exposures.

### Main Policy Recommendations (Table 1 summary)
- Conduct Asset Quality Review (AQR) of all banks with focus on the adequacy of loan loss provisioning and collateral valuation. — Timeline: Short-term.
- Introduce more severe scenarios in CBM stress tests to assess the system’s resilience to extreme but plausible shocks. — Timeline: Short-term.
- Increase transparency around banking system capital adequacy by publishing information about regulatory reserves and own funds as memorandum items in the official quarterly bank-specific balance sheets. — Timeline: Short-term.
- Develop stress testing techniques at CBM to make bank-specific stress tests taking into account all relevant credit, market and liquidity risks. — Timeline: Medium-term.
- Regular compilation of bank-specific LCR in euro. — Timeline: Medium-term.
- Regular compilation of bank-specific NSFR in euro. — Timeline: Medium-term.

*Source: IMF Staff.*

### 13.      Montenegro’s financial sector faces both external and domestic risks. The identified risks

### _cr16198 - 13.      Montenegro’s financial sector faces both external and domestic risks. The identified risks

### Identified risks (organized by RAM)
- A protracted slowdown or re-emergence of financial stress in the euro area (RAM Risk #1).
  - 40 percent of total exports go to the EU.
  - Trade, construction and tourism represent about 15 percent, 9 percent, and 8 percent of GDP, respectively.
  - These sectors represent 20 percent, 5 percent, and 4 percent of total banking system loans, respectively.
- Elevated costs of sovereign debt or uncertainty around sovereign debt sustainability (RAM Risks #1 and #2).
  - Public debt rose from 20 percent of GDP in 2008 to more than 60 percent of GDP in 2014.
  - Uncertainty could raise funding costs for the government and for Montenegro as a whole, reducing credit demand and impairing borrowers’ balance sheets.
- A disorderly withdrawal of unconventional monetary policy in the United States (RAM Risk #2).
  - Gross external debt of 164 percent of GDP in only the private sector.
  - Could trigger global capital flows, risk aversion, volatility, and euro depreciation.
- Economic deterioration or a change of risk sentiment in Russia and Ukraine (RAM Risk #3).
  - Could cause reductions in FDI, trade and tourism consumption.

### Banking sector vulnerabilities (likely transmission channels)
- Credit risk and low profitability.
  - NPLs at 16.8 percent of total loans by end-2014.
  - Aggregate banking system return on assets of 0.8 percent in 2014.
  - Occurrences of negative credit growth historically.
- Limited capacity of liquidity support facility due to euroization.
  - CBM has an emergency liquidity assistance (ELA) framework but limited capacity due to inability to create money.
  - Liquidity is an important structural vulnerability, especially for domestic banks without supporting parent banks.
- Significant use of nonresident deposits.
  - Average percentage of nonresident deposits to total deposits in the banking sector was 19.6 percent by end-June 2015.
  - Ratio was 28 percent for domestically owned banks by end-June 2015.
- Sovereign exposures.
  - Gross credit exposures to the public sector average 17 percent total assets by end-June 2015.
  - Equivalent to an average of 165 percent of own funds by end-June 2015.
- Asset quality and provisioning concerns.
  - Aggregate ratio of loan-loss reserves to total loans past due by more than 90 days was 87.5 percent by end-2014.
  - This ratio was only 14 percent for one bank by end-2014.
  - Several banks report IFRS provisions to total loans past due by more than 90 days of below 50 percent, or even below 40 percent.
  - Concerns that collateral may be overvalued in parts of the financial sector.

### Adjustments to reported capital adequacy
- Two solvency stress tests: one based on non-adjusted CARs and one adjusted for estimated provisioning shortfalls.
- Adjustments rationale and method:
  - Past due loans larger than NPLs indicated inadequate loan-loss provisioning in two banks; NPL stocks in these banks were increased to the amount of past-due loans.
  - Loan-loss provisions adjusted by the difference between a provisioning ratio of 75 percent (category D) and 7 percent (category B).
  - IMF staff considered qualified opinions by external auditors for three banks by end-2014 regarding inadequate provisioning on certain assets.
  - Adjustments made by increasing loan-loss provisions and reducing equity, regulatory capital, and risk-weighted assets (RWA) by the sum of estimated provisioning shortfalls.
- Total adjustments:
  - EUR 35 million (1 percent of 2014 GDP) in four banks.

### Macroeconomic scenarios used in stress tests
- Three scenarios: baseline and two adverse (moderate and severe).
- Baseline scenario:
  - Based on IMF staff projections as of August 2014.
  - Projects positive growth of 3.2 percent in 2015, driven mostly by investments in infrastructural projects.
  - Total credit projected to decrease by almost 3 percentage points of GDP (no credit growth in combination with nominal GDP growth).
- Adverse scenarios methodology:
  - Moderate adverse scenario: projections use one historical standard deviation of macro variables (with occasional expert judgment adjustments).
  - Severe adverse scenario: projections use two historical standard deviations.
- GDP growth projections:
  - Moderate adverse scenario: economic contraction of -2.5 percent (mainly due to euro area slowdown and deterioration in Russia and Ukraine).
    - Moderate scenario mimics 2012 trajectory when real GDP in the EU contracted by -0.5 percent.
    - Includes more than 50 percent decrease in growth of number of foreign tourists’ arrivals and nights in Montenegro tourism sector.
  - Severe adverse scenario: GDP growth of -5.2 percent (reflecting two standard deviations from historical mean; compare to -5.7 percent Montenegro GDP growth in 2009 and -4.4 percent in the EU in 2009).
    - Includes reduced confidence in sovereign finances and increased funding costs.
- Credit growth projections:
  - Gross credit projected to decline by 3 percent of GDP in the moderate adverse scenario.
  - Gross credit projected to decline by 6 percent of GDP in the severe adverse scenario.
  - Historical reference: credit dropped by 6 percentage points from end-September 2008 to end-September 2009.
- Inflation, interest rates, and exchange rates assumptions:
  - Declining inflation leads to assumed decrease in interest rates due to further expansionary monetary policy.
  - Three-month Euribor used as the measure of the “risk-free” interest rate.
  - Projected depreciation of the euro consistent with relatively larger contraction in Europe than in the United States.
  - Exchange rate projections based on one and two standard deviations of an equally weighted portfolio of exchange rates of the euro to the U.S. dollar, British pound, and Swiss franc.
  - Standard deviation projections of inflation and interest rates adjusted to less significant shocks given the low inflation environment.
- Asset prices:
  - Credit spreads on bonds assumed to rise.
  - Equity and real estate prices assumed to fall due to downturn and reduced FDI.
  - Sovereign bond credit spread increases intended to reflect reduced confidence in sovereign finances.
  - Real estate price projections adjusted using estimates provided by authorities and banks during the FSAP mission.

### Macrofinancial transmission — credit risk modeling and projections
- Credit losses modeled as a function of macroeconomic variables using bank-specific and panel data estimations of NPL ratios.
  - NPL ratios adjusted to incorporate NPLs written off or sold during the historical period.
  - Models used macro and financial parameters with up to eight-quarter time lags.
  - Logit transformation applied to ensure projected NPL ratios between 0 percent and 100 percent:
    - X = ln(p/(1-p)) where p is NPL ratio.
  - Estimation model included bank fixed effects and macro variables (see original model formulation).
- Estimation findings:
  - Weak relations between NPL ratios and explanatory variables.
  - Significant variables in panel model: fourth-quarter lagged quarterly change in equity prices, a proxy of annual real GDP growth, and the ratio of foreign direct investment to nominal GDP.
  - Autocorrelation present (Durbin-Watson statistic of 0.34); first quarter lag of NPL ratios introduced.
  - First lag of NPL ratios explained more than 89 percent of NPL developments in adjusted R-squared; explanatory macro variables explained 1.5 percent.
- Data limitations noted:
  - Quarterly real GDP available only since 2011; proxy used based on industrial production and tourism and construction output.
  - Montenegro law allows exclusion of de facto nonperforming loans from NPLs if adequately collateralized; affects dependent variable choice.
  - Projections of credit losses therefore based on estimated models of NPLs rather than past-due loans.
- Alternative method for adverse scenario projections:
  - Because of weak model relations, adverse scenario credit-loss projections used experience and findings from similar countries (Hardy and Schmieder (2013) sensitivities).
  - Assumed loss-given default of 60 percent used to compute stock of NPLs implied by credit losses.
  - Reference average credit loss rates in emerging markets per Hardy and Schmieder (2013): 2.2 percent (moderate), 3.4 percent (medium), 7.4 percent (severe), 14 percent (extreme) of credit outstanding.
  - Chosen sensitivities for the two adverse scenarios are in the range of estimated sensitivities for emerging markets in moderate/medium and severe scenarios.
- Comparison and allocation:
  - Moderate scenario NPL projections broadly in line with CBM’s adverse scenario projections.
  - Severe scenario NPL projections overshot CBM projections but align with recent FSAPs in neighboring countries.
  - System aggregate credit losses allocated to banks using a percentage key based on each bank’s share of total loans past due by more than 90 days and average annual loan portfolio growth during the most recent four quarters.
  - Adjustment increases allocation to banks with recent credit expansion.

### Funding cost and liquidity risk assumptions in scenarios
- Deposit outflow assumptions:
  - Severe adverse scenario gross outflow assumption for nonresident deposits is 46 percent (reflects aggregate outflow between March 2009 and March 2010).
  - Severe adverse scenario assumes gross outflows of domestic deposits of 23 percent (reflects outflows between June 2008 and June 2009).
  - Moderate scenario assumes half of the severe scenario outflows for nonresident deposits (no net nonresident outflows occurred in 2012).
  - Scenarios assume 50 percent of gross deposit outflows are replaced by funding at a similar cost (e.g., from parent banks or development funds).
- Translated net deposit outflows:
  - Moderate scenario: net deposit outflows of 6 percent of domestic deposits and 12 percent of nonresident deposits.
  - Severe scenario: net deposit outflows of 12 percent of domestic deposits and 23 percent of nonresident deposits.

*Source: IMF staff FSAP chapter on Montenegro’s financial sector risks and stress testing (excerpts as provided).*

### 27.      Funding cost projections are affected by assumed deposit outflows. Due to reduced

### _cr16198 - 27.      Funding cost projections are affected by assumed deposit outflows. Due to reduced

### Funding cost projections and deposit outflow assumptions
- Banks are assumed to increase deposit rates to retain depositors amid reduced confidence and capital outflows, but hikes are constrained by profitability pressures.
- Deposit rate hike scenarios:
  - Severe adverse scenario: aggregate deposit rate spreads above three-month Euribor increase by two standard deviations = 3.2 percentage points.
  - Moderate scenario: aggregate deposit rate spreads increase by one standard deviation = 1.6 percentage points.
  - Historical comparators:
    - Post-2008 spread increase in Montenegro: 3 to 4 percentage points (corresponds to severe adverse projection).
    - Reference year 2012 spread increase: about 1.5 percentage points (corresponds to moderate projection).

- Liquid asset financing and uncovered outflows:
  - Liquid asset haircuts assumed: 15 percent (moderate scenario) and 25 percent (adverse scenario).
  - If liquid assets do not cover deposit outflows, domestic banks face increased funding costs of:
    - 2 percentage points above current cost (moderate scenario).
    - 4 percentage points above current cost (adverse scenario).
  - These shocks correspond to a 50 percent and 100 percent increase of estimates of the country risk premium.
  - Foreign-owned banks are assumed to fund uncovered outflows at half the cost using parent bank funding.

### Other modelling assumptions
- Net interest income projection method:
  - Derived from previous reporting period’s income statement as percentage of total net loans and total liabilities.
  - Percentages multiplied by projected net loans and liabilities adjusted for credit growth, loan-loss provisions, and deposit outflows.
- Operational risk losses:
  - Assumed to double (moderate) and triple (adverse).
  - Tripled losses amount to 0.21 percent of total assets.
  - Historical highest annualized loss in period 2010–2014: 0.15 percent of total assets (excluding one outlier).
- Behavioral assumptions for stress tests:
  - Weighting of assets for regulatory capital adequacy remains constant.
  - Banks do not raise capital.
  - Loan book composition unchanged except for credit growth and NPL ratio changes.
  - Managerial portfolio optimization and loan write-offs not assumed.
  - In adverse scenarios equity, commodity, and foreign exchange positions contract by nominal GDP growth.

### Results (solvency stress tests and baseline)
- Provisioning adjustments immediately cause capital needs before stress:
  - Capital shortfall = 0.7 percent of 2014 GDP.
  - Of this, 0.2 percentage points are needed to bring undercapitalized banks to solvency.
  - Provisioning adjustments to four banks reduce system aggregate CAR from 16.2 percent to 14.7 percent.

- Severe adverse scenario (solvency stress):
  - Capital shortfalls range between 2.8 percent and 3.5 percent of 2014 nominal GDP.
  - After provisioning adjustments, capital needs to bring banks to 10 percent CAR = 3.5 percent of 2014 GDP, of which 1.1 percent of GDP is needed to bring undercapitalized banks to solvency.
  - Without provisioning adjustments, capital needs to meet 10 percent CAR = 2.8 percent of GDP, of which 0.3 percent of GDP is needed to bring undercapitalized banks to solvency.

- Moderate scenario:
  - With provisioning adjustments, capital needs to bring banks to 10 percent CAR = 1 percent of GDP, of which 0.4 percent of GDP to bring undercapitalized banks to solvency.
  - Without provisioning adjustments, 0.2 percent of GDP needed to meet 10 percent CAR; all banks would be solvent.

- Baseline scenario:
  - Accumulation of profits increases system aggregate CAR overall.
  - Exception: one bank experiences reduced capital adequacy despite a minor 0.8 percentage point increase in the NPL ratio due to negative profits, highlighting low profit margin risk.

- Main drivers of results (contributions to aggregate CAR change):
  - Credit losses:
    - System can offset credit losses with net income in moderate scenario.
    - Credit losses drive CAR decline in severe adverse scenario and reduce net loans and interest income.
  - Funding risk:
    - Increased funding costs contribute about -1.8 percentage points (moderate) and -3.6 percentage points (adverse) to aggregate CAR change.
    - Profit margins were 0.5 percent of assets by end-June 2015; funding risk is material especially for banks without parent funding or other liquidity support.
    - Under assumed deposit outflows in adverse scenario, all except three domestic banks cover outflows by selling liquid assets. For the three banks, uncovered outflows amount to no more than 6 percent of total liabilities, causing at most a 0.4 percentage point effect on their aggregate funding cost.
  - Asset prices:
    - Total contribution of increased interest rate spreads and declining stock and real estate prices = approximately -1.2 percentage points (moderate) and -2.8 percentage points (adverse).

### Sensitivity analyses of solvency
A. Credit concentration risk
- Sector concentration:
  - Wholesale and retail sector constitutes 17 percent of total loans; limited sectoral concentration at aggregate level.
- Single borrower concentration:
  - Average largest exposure remains below 25 percent of regulatory capital, but several banks have exposures above the threshold.
  - Reported large exposures reflect net exposure after credit mitigation; several single gross exposures approach or exceed twice the regulatory limit of 25 percent.
- Impact of defaults:
  - Default of largest borrower in each bank with no recovery: capital needs = 0.3 percent of GDP to bring undercapitalized banks’ CARs to 10 percent.
  - Default of five largest borrowers in each bank with no recovery: capital needs = 3.2 percent of GDP to bring banks to 10 percent CAR, of which 0.3 percentage points needed to bring undercapitalized banks to solvency.
- Public sector exposures:
  - In a handful of banks, public sector exposures exceed 15 percent of total assets at an average of 36 percent.
  - Public sector exposures include sovereign bonds and loans to the public sector.

B. Market risks (targeted sensitivity tests)
- Testing assumptions:
  - RWA assumed constant; only regulatory capital affected.
- Haircuts on domestic sovereign exposures:
  - 10 percent haircut → losses = 1.5 percent of GDP; aggregate CAR falls from 16.2 percent to 13.5 percent; total capital shortfall = 0.4 percent of GDP.
  - 20 percent haircut → losses = 3.1 percent of GDP; aggregate CAR falls to 10.8 percent; total capital shortfall = 1.5 percent of GDP.
- Real estate price decline:
  - 35 percent drop in real estate values → losses = 0.9 percent of GDP; aggregate CAR declines from 16.2 percent to 14.5 percent.
- Interest rate hike:
  - 500 basis point interest rate rise (one-year effect) → loss = 0.2 percent of GDP; aggregate CAR declines from 16.2 percent to 15.8 percent.
  - Note: assumption that lending rates and funding costs increase equally may not hold; banks may not fully pass funding cost increases to borrowers.
- Equity price decline:
  - 35 percent fall in equity prices → losses = 0.1 percent of GDP; aggregate CAR declines from 16.2 percent to 15.9 percent.
- Euro appreciation:
  - 30 percent appreciation → losses would not impact aggregate CAR by more than a few basis points given very small open foreign exchange positions.

C. Operational risk
- Historical aggregate operational risk losses average 0.01 percent of total assets since 2010 (with a spike in 2013).
- Assuming operational risk losses of two standard deviations above the mean → losses = 0.02 percent of GDP; aggregate CAR reduces from 16.2 percent to 16.1 percent.
- Operational loss event types include internal and external fraud; business disruptions and system failures; damage to physical assets; execution, delivery, and process management failures.

### Liquidity stress test
Measures and assumptions
- Short-term liquidity ratio informed by LCR (next 30 days potential net outflows and capacity to cover using high-quality liquid assets).
- Long-term liquidity ratio informed by NSFR (capacity to fund assets with stable funding over one year).
- Both measures weight assets and liabilities to reflect liquidity risk and align with Basel III (BCBS, 2013 and BCBS, 2014).
- External liabilities = 21 percent of total assets; reliance on external funding makes banks vulnerable to funding difficulties when global market conditions deteriorate.
- Complementary calculations: ratios computed with assumed deposit outflows (severe adverse scenario) and after assuming no liquidity in domestic sovereign bonds.
- Basel III requirement: both LCR and NSFR should equal or exceed 100 percent.

Results
- Short-term liquidity:
  - Aggregate reported short-term liquidity ratio = 863 percent; no banks below 100 percent threshold.
  - Excluding domestic sovereign bonds from HQLA, aggregate ratio = 629 percent of projected net outflows.
  - Adjusted for deposit outflows (share falling due in 30 days), none of the banks’ short-term ratios fall below 100 percent.
  - Drivers: large stock of high quality liquid assets and large expected inflows from financial institutions within 30 days.
  - Note on adjustment: initial assumed average run-off rate of deposits = 7.5 percent.
- Long-term liquidity:
  - System aggregate long-term liquidity ratio = 123 percent; two banks below 100 percent threshold.
  - Drivers of low ratios: low asset quality and higher portion of unstable funding.
  - After reducing available stable funding by deposit outflows assumed under severe adverse scenario:
    - Aggregate long-term ratio drops to 108 percent.
    - Seven banks fall below the 100 percent threshold.
  - Excluding domestic sovereign bonds from HQLA had marginal effect and did not cause additional banks to fail the 100 percent threshold.

### Contagion risks and spillover analysis
- Domestic interconnectedness:
  - Limited domestic interconnectedness; gross claims and liabilities of banking sector to nonbanks averaged about 2 percent of total assets.
  - Domestic interbank transaction volumes very small = 0.2 percent of banking sector’s total assets.
  - No interbank market; instruments to support interbank transactions are underdeveloped.
  - Surplus liquidity lessens incentives for interbank transactions.
- Insurance sector exposures:
  - About 22 percent of the insurance industry’s assets are deposits at commercial banks.
  - Most insurance companies spread deposits across many banks; only a few small insurers concentrate deposits in few banks.

*Source: IMF/CBM staff calculations and data as presented in the provided content.*

### 51.      The banking sector has high cross-border interbank exposures, which reflects foreign

### _cr16198 - 51.      The banking sector has high cross-border interbank exposures, which reflects foreign

### Cross-border exposures and composition
- Banks’ foreign claims and liabilities account for over 115 percent and 50 percent of total regulatory capital, respectively, as of end-2014.
- Exposures to foreign banks are mainly in the form of claims.
- About 47 percent of total interbank external claim is to Austria, Germany, and Hungary.
- Exposures to Austria and Hungary reflect close linkages with the parent banks.
- Montenegro banks invest in foreign government securities and deposit part of their excess liquidity in EU banks, particularly in Germany.
- Some banks have invested in higher-yield sovereign bonds in the region, although the exposures are still limited.

### Network analysis framework and data
- Analytical framework: methodology proposed by Espinosa-Vega and Solé (2010), simulating hypothetical failure of a country’s banking system and tracking spillover effects on other countries.
- Two separate shocks considered:
  - Credit shock: impact of a banking system defaulting on its liabilities to foreign banks.
  - Funding shock: impact of a banking system deleveraging by withdrawing funding from foreign banks, triggering fire sales of the latter’s assets.
- Data: bilateral exposures among the 12 banks and the banking systems of 35 foreign countries provided by the central bank; exposures among foreign banking systems from (restricted) BIS locational statistics.

### Shock assumptions and failure definition
- Credit shock assumption: loss-given default of 100 percent on interbank exposures.
- Funding shock assumptions: haircut of 50 percent on asset fire sales and 65 percent on the roll-over ratio of interbank debt.
- Bank ‘failure’ definition: undercapitalization defined as capital-to-RWA falling below the CAR requirement (i.e., 10 percent in Montenegro).
- Adjustments for outflows were made assuming an initially assumed average weight of 90 percent of deposits.

### Spillover analysis results
- The credit shock dominates the funding shock in impact on Montenegro banks.
- The credit shock has a significantly larger impact than the funding shock, consistent with Montenegro banks’ exposures to foreign banks being mainly in the form of claims.
- Large contagion risks identified from: France, Germany, the United Kingdom, and the United States.
- Under the combined credit and funding shock analysis:
  - The impact of the two credit and funding shocks would lead to severe capital impairment in 7 and 10 Montenegro banks, respectively.
  - Up to 83 percent of the overall banking system’s regulatory capital would be lost.
- A large shock to most home countries of foreign-owned banks caused a confined rather than system-wide failure, except for France.
  - For Austria, Hungary, and Slovenia, where the impact caused bank failures, severe capital impairment was confined to the subsidiaries and/or other Montenegro banks with relatively large exposures to these countries.
  - Confined impact was partly due to diversified funding and portfolio structures and limited domestic interbank transactions.

*Source: IMF staff calculations and Central Bank of Montenegro data as presented in the referenced IMF document.*

### 57.      Large spillover impact from systemically important countries on the Montenegro

### _cr16198 - 57.      Large spillover impact from systemically important countries on the Montenegro

### Spillovers and contagion dynamics
- Banking-system shock propagation is primarily second- or third-round for Montenegro.
- Direct exposures to France, the United Kingdom, and the United States were relatively small, but their indirect impact was large due to interconnectedness with countries that have direct exposures to Montenegro.
- Bank failures in France, Germany, and the United Kingdom caused Belgium and Luxembourg to be hit in the first round.
- The close linkage between the U.K. and the U.S. banking systems produced a U.K. banking-system failure in the first round from a U.S. banking-system shock.
- Figure 5 (sequence of successive failures) shows contagion spreading from failures in four systemically important countries, ultimately triggering Montenegrin bank failures in the second and third rounds.

### Solvency stress test results (aggregate and contributions)
- Aggregate CAR (without provisioning adjustment): Initial (2014) and scenarios shown for Baseline (2015), Adverse 1 (2015), Adverse 2 (2015) with aggregate CAR series plotted in percent of risk-weighted assets.
- Aggregate CAR (with provisioning adjustment): Initial (2014) and scenario series plotted in percent of risk-weighted assets.
- Contributions to CAR (without and with provisioning adjustment) reported by component for Baseline, Adverse 1, Adverse 2:
  - Components listed: Change in RWA; Total other net income; Real estate price risk; Equity price risk; Foreign exchange risk; Interest rate risk; Funding risk; Credit spread risk on bond holdings; Loan loss provisions.
- Table 10: Summary Solvency Stress Test Results (Without Provisioning Adjustments) (in millions of EUR, unless stated otherwise)
  - Before stress, 12/31/2014:
    - Regulatory capital 3159.2 (3159.2 in % of GDP column also shown)
    - Tier 1 capital 2707.9
    - Tier 2 capital 451.3
    - Risk-weighted assets (RWAs) 1,947 (56.9)
    - Total assets 3,136 (91.6)
    - Capital Adequacy Ratio (CAR), percent 16.2
    - Capital shortfall to meet 12 percent CAR 40.1
    - Capital shortfall to meet 10 percent CAR 0.0
  - Stress test losses (selected):
    - Credit risk -40.1 (Baseline), -531.5 (Adverse 1), -1855.4 (Adverse 2)
    - Loan loss provisions -40.1, -431.3, -1664.8
    - Credit spread risk on bond holdings 0.0, -100.3, -190.6
    - Market and liquidity risk 10.0, -411.2, -752.2
    - Funding risk 0.0, -320.9, -561.6 (of which caused by deposit outflows 0.0, 0.0, 10.0)
    - Equity price risk 0.0, -30.1, -60.2
    - Real estate price risk 10.0, -90.3, -180.5
    - Other net income 421.2, 391.1, 250.7 (of which are operational risk losses -20.1, -40.1, -60.2)
  - After stress, 12/31/2015:
    - Regulatory capital 35410.3 (Baseline), 2607.6 (Adverse 1), 802.3 (Adverse 2)
    - Tier 1 capital 3099.0, 2146.3, 341.0
    - RWAs 1,957 (57.1), 1,808 (52.8), 1,590 (46.4)
    - Total assets 3,157 (92.2), 2,755 (80.4), 2,290 (66.9)
    - CAR, percent 18.1 (Baseline), 14.4 (Adverse 1), 5.0 (Adverse 2)
    - Capital shortfall to meet 12 percent CAR 30.1 (Baseline), 170.5 (Adverse 1), 1223.6 (Adverse 2)
    - Capital shortfall to meet 10 percent CAR 0.0, 50.2, 952.8
    - Capital shortfall to meet 0 percent CAR 0.0, 0.0, 100.3
- Table 11: Summary Solvency Stress Test Results (With Provisioning Adjustments) (in millions of EUR)
  - Before stress, 12/31/2014:
    - Regulatory capital 2808.2
    - Tier 1 capital 2356.9
    - Tier 2 capital 451.3
    - RWAs 1,913 (55.8)
    - Total assets 3,136 (91.6)
    - CAR, percent 14.7
    - Capital shortfall to meet 12 percent CAR 280.8
    - Capital shortfall to meet 10 percent CAR 240.7
    - Capital shortfall to meet 0 percent CAR 80.2
  - Stress test losses (selected) mirror Table 10 values with provisioning adjustments applied.
  - After stress, 12/31/2015:
    - Regulatory capital 3209.3 (Baseline), 2256.6 (Adverse 1), 471.4 (Adverse 2)
    - Tier 1 capital 2748.0, 1805.3, 20.1
    - RWAs 1,922 (56.1), 1,776 (51.9), 1,562 (45.6)
    - Total assets 3,121 (91.1), 2,728 (79.6), 2,269 (66.3)
    - CAR, percent 16.6 (Baseline), 12.7 (Adverse 1), 3.0 (Adverse 2)
    - Capital shortfall to meet 12 percent CAR 270.8, 431.2, 1494.3
    - Capital shortfall to meet 10 percent CAR 240.7, 331.0, 1223.5
    - Capital shortfall to meet 0 percent CAR 70.2, 150.4, 371.1

### Liquidity stress test results
- Figure 7 and Table 13 summarize liquidity stress test metrics: Aggregate short-term liquidity ratio (LCR) and aggregate long-term liquidity ratio (NSFR), distributions across banks, and distributions under outflow assumptions.
- LCR (standard) and variants reported:
  - LCR, standard total 863
  - LCR, standard with domestic sovereign bond Level 2 classification 828
  - LCR, standard excluding domestic sovereign bonds 629
  - LCR, standard excluding domestic sovereign bonds and required reserves 474
- LCR, assuming outflows in severe adverse scenario:
  - Total 525; with domestic sovereign bond Level 2 classification 503; excluding domestic sovereign bonds 382; excluding domestic sovereign bonds and required reserves 288
- NSFR (standard):
  - Total 124; with domestic sovereign bond Level 2 classification 123; excluding domestic sovereign bonds 121
  - Implied liquidity shortfall 10 (standard)
  - Number of banks below 100 percent: 2 (standard)
- NSFR, assuming outflows of deposits in severe adverse scenario:
  - Total 108; with domestic sovereign bond Level 2 classification 107; excluding domestic sovereign bonds 105
  - Implied liquidity shortfall 95 (scenario)
  - Number of banks below 100 percent: 7 (scenario)

### Banking sector structure and key metrics (as of end-2014 and June 2015 data)
- Table 2 (As of June, 2015) — Share of total banking system assets:
  - Domestic-owned 21.4 percent
    - Prva Banka 8.7 (Montenegro)
    - Atlas Banka 8.0 (Montenegro)
    - Universal Capital Bank 2.3 (Montenegro)
    - Invest Banka Montenegro 1.4 (Montenegro)
    - Lovcen Banka 1.0 (Montenegro)
  - Foreign-owned 78.6 percent
    - Crnogorska Komercijalna Banka 17.4 (Hungary)
    - NLB Montenegro Banka 15.4 (Slovenia)
    - Erste Bank 11.5 (Austria)
    - Societe Generale Banka Montenegro 11.7 (France)
    - Hipotekarna Banka 11.3 (Italy)
    - Hypo Alpe Adria Bank 7.4 (Austria)
    - Komercijalna Budva Banka 3.6 (Serbia)
    - Zapad Bank 0.3 (Ukraine)
- Table 3 (As of end-2014) — Banking sector summary (reported data, mln EUR unless stated):
  - Number of banks 12
  - Assets 3,136 (100.0 percent of banking system assets; 87.6 percent of financial system assets; 91.6 percent of GDP)
  - Cash and central bank reserves 499 (15.9; 13.9; 14.6)
  - Customer loans 2,367 (75.5; 61.3; 69.1)
  - Loan impairments -173 (-5.5; -4.8; -5.0)
  - Net loans 2,195 (70.0; 61.3; 64.1)
  - Securities 243 (7.7; 6.8; 7.1)
  - Other assets 236 (7.5; 6.6; 6.9)
  - Total deposits from customers 2,308 (73.6; 64.5; 67.4)
    - of which protected by deposit insurance 1,064 (33.9; 29.7; 31.1)
  - Loans from parent banks and development funds 249 (7.9; 6.9; 7.3)
  - Equity capital 444 (14.2; 12.4; 13.0)
    - Regulatory reserves, total -291 (-9.3; -8.1; -8.5)
    - Own funds, total 315 (10.0; 8.8; 9.2)
  - Capital adequacy ratio (CAR), percent 16.2
- Asset quality (end-2014):
  - Total NPLs 402 (12.8 percent of banking system assets; 11.2 percent of financial system assets; 11.7 percent of GDP)
  - NPLs to total gross loans, percent 16.8
  - Total past due loans, > 90 days 332 (10.6; 9.3; 9.7)
  - Past due loans (>90 days) to total gross loans, percent 13.9
  - IFRS provisions to total NPLs, percent 39.5
  - IFRS provisions to total past due loans, percent 47.8
  - Regulatory reserves to total NPLs, percent 72.3
  - Regulatory reserves to total past due loans, percent 87.5
- Risk exposures (end-2014):
  - External claims 503 (16.0; 14.0; 14.7)
  - External liabilities 645 (20.6; 18.0; 18.8)
  - Own sovereign bond holdings 140 (4.5; 3.9; 4.1)
  - Foreign sovereign bonds holdings 83 (2.6; 2.3; 2.4)
  - Corporate bond holdings 16 (0.5; 0.5; 0.5)
  - Net open foreign exchange position 20.1 (0.1; 0.1; 0.1)
  - Net open position in equity instruments 140.4 (0.4; 0.4; 0.4)
  - Net open position in real estate assets 92 (2.9; 2.6; 2.7)
- Profitability (end-2014):
  - Return on assets, percent 0.8
  - Return on equity, percent 5.4
  - Net interest and fee income to average assets 4.6
  - Overhead expenses to net interest and fee income 74.7

### Loan book composition and NPLs (end-2014)
- Gross loans total 2,396 (100.0); Total NPLs 402 (16.8 percent of total loans)
- Household loans, residents 880 (36.7 percent of total loans), Non-performing loans% of total loans 23.8
- Corporate loans, residents 989 (41.3 percent of total loans), Non-performing loans% of total loans 41.3
  - Agriculture, forestry and fishing 28 (1.2; NPL% 20.1)
  - Manufacturing 109 (4.5; NPL% 42.2)
  - Wholesale and retail trade; repair of motor vehicles and motorcycles 407 (17.0; NPL% 205.0)
  - Construction 111 (4.6; NPL% 431.8)
  - Transport and warehousing 49 (2.1; NPL% 90.8)
  - Accommodation, food, arts, recreation and other services 73 (3.1; NPL% 50.6)
  - Public administration and defence 13 (0.5; NPL% 120.5)
  - Real estate 24 (1.0; NPL% 50.2)
  - Financial and insurance sector 16 (0.7; NPL% 20.1)
  - Professional, scientific and technical activities 40 (1.7; NPL% 120.5)
- Loans to non-residents 434 (18.1; NPL% 20.5)
- Other loans 94 (3.9; NPL% 40.6)
- Total 2,396 (100.0; Total NPLs 402; NPL% 16.8)

### Macroeconomic projections and scenario assumptions
- Table 5: Macroeconomic series (baseline and projections, selected):
  - Real GDP growth (%) by year:
    - 2010 2.5
    - 2011 3.2
    - 2012 -2.5
    - 2013 3.3
    - 2014 1.5
    - 2015 Proj. 3.2 (Baseline) and projections for 2015 shown as -2.5 and -5.2 for adverse paths in table columns
  - CPI inflation (%) 0.7 (2010), 2.8 (2011), 5.1 (2012), 0.3-0.3 (2013-2015 entries include hedged/negative values)
  - Risk-free interest rate 1.0 (2010), 1.4 (2011), 0.2 (2012), 0.3 (2013), 0.1/0.1/-0.5/-1.1 entries for projections
  - Exchange rate 1/ (EUR expressed as equally-weighted average price of USD, GBP and CHF) 0.90 (2010), 0.93 (2011), 0.94 (2012), 0.91 (2013), 0.98 (2014), 1.11/1.13/1.16 (2015 projections)
  - Stock price index and Real estate price index series shown with base 100 and subsequent values (e.g., stock price index 100,64,68,68,78,78,61,44; real estate price index 100,105,91,97,81,82,73,65)
- Table 6: Balance Sheet and Income Statement Projections (selected lines and scenario assumptions)
  - Non-performing loans, percent of gross loans: 21.0 (2010), 15.5 (2011), 17.6 (2012), 18.4 (2013), 16.8 (2014), 16.9 (2015 Baseline), 19.9 (Adverse 1), 27.4 (Adverse 2)
  - Credit losses, percent of loans net of regulatory reserves: -0.4 (2010), -2.3 (2011), 1.9 (2012), 3.1 (2013), -0.8/-0.2/2.0/7.8 entries for projections
  - Gross credit, percent of GDP series and growth rates of various asset positions and open positions provided for Baseline and Adverse scenarios.
  - Funding cost, spread above Euribor 3M, percent: 2.3 (2010), 1.9 (2011), 3.0 (2012), 2.3 (2013), 1.8/0.0/1.6/3.1 entries for projections
  - Net outflows of deposits, percent of initial total deposits (scenario assumptions): 0.0 (Baseline), 7.3 (Adverse 1), 14.7 (Adverse 2)
    - Net outflows of non-resident deposits, percent of initial deposits: 0.0, 11.5, 23.0
  - Haircut on liquid assets sold to cover deposit outflows, percent: 0.0, 15.0, 25.0
  - Spread of domestic sovereign bonds above Euribor 3M, percent: 0.4, 0.5, 3.4, 0.2, -0.8, -0.8, 3.6, 7.9 (table entries by year/scenario)
  - Spread of foreign sovereign bonds above Euribor 3M, percent /1: 4.0, 4.8, 3.9, 3.4, 2.0, 2.0, 3.3, 4.6

### Empirical model linking macro variables to NPLs
- Table 8: Panel data model estimation output (Dependent Variable: LOGIT_NPL_RTO_; Panel Least Squares; sample 2005Q1–2014Q4; total panel observations 419)
  - Coefficients (with Std. Error, t-Statistic, Prob.):
    - LOGIT_NPL_RTO_(-1) coefficient 0.884376 (Std. Err 0.018327; t-Statistic 48.255460; Prob. 0.0000)
    - LOG_D_EQT_INX_(-4) coefficient -0.271751 (Std. Err 0.094578; t-Statistic -2.873290; Prob. 0.0043)
    - LOG_D_REAL_GDP_ coefficient -0.574935 (Std. Err 0.235812; t-Statistic -2.438109; Prob. 0.0152)
    - LOG_FDI_ coefficient -0.094513 (Std. Err 0.040536; t-Statistic -2.331583; Prob. 0.0202)
    - C (constant) -0.490345 (Std. Err 0.148041; t-Statistic -3.312234; Prob. 0.0010)
  - Model diagnostics:
    - Cross-section fixed effects (dummy variables)
    - R-squared 0.909064; Adjusted R-squared 0.905913
    - S.E. of regression 0.417602; F-statistic 288.4781; Prob(F-statistic) 0.000000
    - Durbin-Watson stat 1.895259
  - Variable definitions:
    - LOGIT_NPL_RTO is the logit-transformed NPL ratio.
    - LOG_D_EQT_INX is the quarter-on-quarter logarithmic change in the equity index.
    - LOG_D_REAL_GDP_ is the year-on-year logarithmic change in the proxy measure of real GDP.
    - LOG_FDI_ is the log-transformed ratio of FDI to GDP.

### Concentration risk and single-name default scenarios
- Table 12: Results of Concentration Risk Test (in millions of EUR)
  - Before losses:
    - CAR, percent 16.2
    - Capital shortfall to meet 12 percent CAR 444
    - Capital shortfall to meet 10 percent CAR 0
  - Assumed recovery rate of 70 percent:
    - CAR, percent 15.3 (default of largest borrower), 13.2 (largest 5), 11.6 (largest 10)
    - Capital shortfall to meet 12 percent CAR 519 (largest borrower), 43 (largest 5), 3 (largest 10)
    - Capital shortfall to meet 10 percent CAR 17 (largest borrower), 20 (largest 5), 0 (largest 10)
  - Assumed recovery rate of 35 percent:
    - CAR, percent 14.4, 9.4, 5.7
    - Capital shortfall to meet 12 percent CAR 1375, 137, 13
    - Capital shortfall to meet 10 percent CAR 446, 108, 0
  - Assumed recovery rate of 0 percent:
    - CAR, percent 13.2, 5.3, -1.1
    - Capital shortfall to meet 12 percent CAR 2114, 02, 236 (table entries as presented)
    - Capital shortfall to meet 10 percent CAR 1011, 02, 08 (table entries as presented)
    - Capital shortfall to meet 0 percent CAR 0, 1086, (table entries as presented)

*Source: CBM and IMF staff calculations, as presented in _cr16198 (IMF).*

### Appendix I. Preliminary Risk Assessment Matrix

### Appendix I. Preliminary Risk Assessment Matrix

### Overall Level of Concern and Expected Impact
- Overall likelihood of severe realization of threat in the next 1–3 years: Staff assessments provided for each threat (High, Medium).
- Expected impact on financial stability if threat is realized: Staff assessments provided for each threat (High, Medium).

Key threats and staff assessments
- 1. Protracted growth slowdown in the Euro area and neighboring countries  
  - Staff assessment: High (G-RAM, June 2015)  
  - Staff assessment of impact: High  
  - Expected channels and impacts:  
    - Adverse impact on external demand and internal demand combined with further “lowflation” imported from the euro area.  
    - Credit contraction due to low demand and tightened supply resulting from banks’ risk aversion and foreign subsidiaries’ reduced presence in the Montenegro market.  
    - Reduced income among banks’ borrowers, leading to further credit quality deterioration.

- 2. Spillovers from a deterioration of global financial market conditions  
  - Staff assessment: High (G-RAM, June 2015)  
  - Staff assessment of impact: High  
  - Expected channels and impacts:  
    - External financial market turmoil exposes domestic vulnerabilities affecting domestic economic output and financial stability. Importantly, the government has large external financing needs, averaging around 9 percent of GDP over 2016–2020. The stock of debt is large (nearly 70 percent of GDP) and expected to increase, absent policy measures. A sharp increase in international risk premiums and/or a loss of market access would have significant negative economic and financial sector spillovers. Triggers could include a loss of confidence in the authorities’ economic reform program, a tightening of U.S. monetary policy, and a general reassessment of risk amidst the increase in risk aversion affecting emerging markets.  
    - Banks raise deposit rates to avoid deposit outflows, further reducing their profitability.  
    - Banks suffer deposit outflows from withdrawals by external creditors, such as nonresident depositors and by domestic depositors, to repay external creditors and to avoid exposure to the domestic banking system. Besides the use of liquid assets, banks completely rely on market and foreign-parent bank funding at elevated interest rates, due to the lack of an LOLR mechanism.  
    - The combination of a reduction in external and internal demand, and increased funding costs, causes a significant reduction in banks’ borrowers’ credit quality.

- 3. Geopolitical and macroeconomic events associated with Russia/Ukraine tensions and Greece’s negotiations with creditors  
  - Staff assessment: Medium (G-RAM, June 2015)  
  - Staff assessment of impact: Medium  
  - Expected channels and impacts:  
    - Reduction in FDI inflows and growth prospects cause a reduction in the credit quality among banks’ borrowers.  
    - Real estate prices decline due to reduced demand from Russia and Ukraine, causing direct losses due to asset revaluation and indirect losses due to a reduction of credit quality.  
    - Direct financial links to Greece are negligible, but fallout through secondary channels could undermine banks’ borrowers’ credit quality.

### Banking Sector: Solvency Risk — Scope, Methods, and Assumptions
- Institutional perimeter and data  
  - All banks (12 banks) as of end-2014.  
  - Market share: 100 percent of total banking sector assets by end-2014; 99.7 percent of total banking sector assets by mid-2015.  
  - Bank-by-bank supervisory data as of end-2014. Data especially requested from banks.  
  - Consolidated basis stress test of domestic-owned banks, and foreign-owned bank subsidiaries active in Montenegro.

- Methodology and satellite models  
  - IMF solvency stress testing balance sheet framework.  
  - Panel data model (fixed effects) estimation of logit-transformed bank-specific NPL ratios as a function of macroeconomic variables using up to eight quarterly lags. Given insufficient statistical significance of the relation between NPL ratios and macroeconomic variables, global rules-of-thumb sensitivities of credit losses to GDP growth were used to estimate credit losses and project NPL ratios.  
  - Peer group analysis compared Adverse 1 and Adverse 2 projections to internal CBM models and FSAPs in neighboring countries.  
  - Expert judgment to estimate pre-impairment income sensitivities to macroeconomic events.

- Stress test horizon  
  - 1-year horizon was chosen because of (i) the significant uncertainty around macroeconomic projections in Montenegro, (ii) the lack of historical time series of macro variables, and (iii) the low reliability of some historical data.

### Scenario Design and Tail Shocks (Solvency)
- Baseline scenario: IMF macroeconomic projections as of August 2015.

- Moderate adverse scenario (protracted slowdown in the euro area; projections generally estimated as one standard deviation from the historical mean, with expert judgment adjustments):  
  - real GDP growth of -2.5 percent  
  - gross credit growth of minus 3 percentage points of GDP  
  - funding cost of banks rise by 1.6 percent  
  - general interest rates decrease by 60 basis points  
  - domestic-sovereign bond, foreign-sovereign bond and corporate bond credit spreads increase by 4.4, 1.3, and 4.7 percentage points, respectively  
  - stock prices fall by 22 percent  
  - real estate prices fall by 10 percent

- Severe adverse scenario (moderate scenario developments complemented by elevated uncertainty; projections generally estimated as two standard deviations from the historical mean, with expert judgment adjustments):  
  - real GDP growth of minus 5.2 percent  
  - gross credit growth of -6 percentage points of GDP  
  - funding cost of banks rises by 3.1 percent  
  - general interest rates decrease by 120 basis points  
  - domestic-sovereign bond, foreign-sovereign bond and corporate bond credit spreads increase by 8.7, 2.6, and 9.5 percentage points, respectively  
  - stock prices fall by 43 percent  
  - real estate prices fall by 20 percent

- Sensitivity analysis:  
  - Single-factor shocks: haircut on public sector loans and sovereign securities; interest rate hike; exchange rate; equity price decline; real estate price decline.  
  - Credit concentration risk based on local regulatory standards.

### Risks, Buffers, and Behavioral Assumptions (Solvency)
- Risks assessed: Credit risk (household, corporate, public sector loans; domestic-sovereign, foreign-sovereign, corporate bond exposures), market risk (interest rate, credit spread, FX, equity price, real estate price), funding risk, operational risk.  
- Funding risk assumptions:  
  - Moderate scenario: net deposit outflows of 6 percent of domestic deposits and 12 percent of nonresident deposits.  
  - Severe scenario: net outflows of 12 percent of domestic deposits and 23 percent of nonresident deposits.  
  - To cover outflows banks assumed to use cash balances and sell liquid assets at haircuts of 15 percent (moderate) and 25 percent (severe).  
  - If liquid assets did not cover outflows, domestic banks were assumed to meet increased funding cost of 2 and 4 percentage points above the current cost in the two scenarios, respectively (approximately a 50 percent and 100 percent increase of the country risk premium). Foreign-owned banks assumed to fund uncovered outflows at half the cost using parent bank funding.

- Operational risk: Losses set at twice and three times the 2014 level in the moderate and severe scenarios, respectively.

- Behavioral adjustments and balance sheet treatment:  
  - Evolution of total assets and liabilities reflected behavioral assumptions made to counter deposit outflows as described above.  
  - Evolution of RWAs based on constant balance sheet assumption, i.e. adjusting projected RWA by loan loss provisions at a 100 percent risk weight.  
  - No other management actions considered.  
  - Other net income items, dividends, and taxes based on a pre-determined rule of evolution in line with economic activity measured as gross credit growth.

- Regulatory and market-based standards:  
  - Basel II regulatory standard.  
  - Hurdle rates based on local regulatory minimum Capital Adequacy Ratio (CAR) of 10 percent.  
  - Output presentation: CAR and capital shortfall under the 10 percent and a 12 percent hurdle rate; system-wide sum of gross capital shortfall (without offsetting); contribution of each source of losses to aggregate capital shortfall.

### Banking Sector: Liquidity Risk — Framework and Assumptions
- Institutional perimeter and data:  
  - All banks (12 banks) as of end-2014.  
  - Market share: 100 percent of total banking sector assets by end-2014; 99.7 percent by mid-2015.  
  - Supervisory data; data especially requested from banks.

- Methodology:  
  - Short-term measure informed by the Basel III LCR setup.  
  - Long-term measure informed by the Basel III NSFR setup.

- Risks and buffers:  
  - Risks: Deposit outflows, market liquidity, maturity mismatches.  
  - Buffers: Counterbalancing capacity (HQLA, ASF).

- Tail shocks and calibration:  
  - Haircuts and run-off rates as defined in Basel III for LCR and NSFR.  
  - Additional deposit outflows as assumed in the severe adverse scenario (see Solvency section).  
  - Exclusion of required central bank reserves.  
  - Exclusion of government bonds from HQLA.

- Regulatory benchmarks and outputs:  
  - LCR proxy should exceed 100 percent (not a legal/regulatory requirement).  
  - NSFR proxy should exceed 100 percent (not a legal/regulatory requirement).  
  - Output presentation: Aggregate LCR and NSFR proxies; aggregate gross liquidity shortfalls (without offsetting); number of banks that fail.

### Banking Sector: Contagion Risk — Scope and Methods
- Institutional perimeter and scope:  
  - All banks (12)  
  - 4 insurance companies  
  - 6 investment funds  
  - 1 microcredit institution  
  - 46 other financial institutions  
  - Percentage of total sector assets: 100 percent  
  - Data and baseline date: Supervisory data, banks’ own data, publicly available data. Baseline date: Dec 31, 2014

- Methodology and tail shocks:  
  - Network analysis, using Espinosa-Vega and Solé (2010) methodology.  
  - Stress scenarios: severe stress in a bank or the banking system causing default on all liabilities to domestic institutions or foreign banks; joint credit and funding shock where default leads to a liquidity squeeze for institutions funded by the defaulting bank or system.

- Output presentation:  
  - Capital impairment to domestic banking system, number of failed banks, and remaining buffers (system and bank level).  
  - Capital impairment to domestic insurance sector, number of failed insurance companies, and remaining buffers (sector-wide and company level).

*Source: IMF staff.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16198.pdf_