## 1bgrea2019003 - Preface (draft SREP manual review and mission report)

## Source details

**Canonical URL:** [1bgrea2019003 - Preface (draft SREP manual review and mission report)](https://www.imf.org/-/media/files/publications/cr/2019/1bgrea2019003.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2019/1bgrea2019003.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2019/1bgrea2019003.pdf.json)

---

### Mission, purpose, and scope
- MCM TA mission visited Sofia during the period July 16–20, 2018 to assist the Bulgarian National Bank (BNB).
- Advisor: Mr. Filip Gijsel, short-term Bank Supervision advisor for the Monetary and Capital Markets Division (MCM) of the IMF.
- Purpose: review the formalization and implementation of a comprehensive SREP that includes an explicit and detailed supervisory Pillar 2 capital requirement; review the draft SREP manual received on July 4, 2018 and being finalized in the BSD.
- The manual is intended to form an integral part of the Banking Supervisory Process Manual (approved February 9, 2017 by the GC) and its approval by the GC is expected in the coming months.
- Review takes account of IMF and World Bank missions in 2015, 2016, and 2017 on the observance of the Basel Core Principles.

### Executive-summary findings and context
- The BNB SREP is based on EU-wide common procedures, processes and methodologies; the draft SREP Manual includes recently proposed changes to the Guidelines issued by the European Banking Authority (EBA).
- Focus: implementation into appropriate and proportionate supervisory assessments and into focused actions and planning.
- During the SREP process in 2016 and 2017 the BNB did not identify the need to impose additional capital requirements as per Pillar 2; consequently no Total Supervisory Capital Requirement (TSCR) and Overall Capital Requirement (OCR) were communicated to banks.
- The assessment was communicated in all EU-colleges but was not for all banks summarized in a standardized SREP framework.
- Presence of banking groups under direct supervision by the Single Supervisory Mechanism (SSM) leads SSM processes to have a predominant place in BSD activities.
- Bulgarian Government requested to enter into close cooperation with the ECB on supervision (article 7 of Council Regulation 1024/2013 of October 15, 2013); cooperation with the SSM will lead to harmonized procedures and methodologies, and a common decision process for most administrative acts, creating additional workload in the transition period.
- BNB categorization: supervised institutions grouped into four categories by size, structure, scope of activities and complexity; out of the 11 institutions in the first category (identified by the BNB GC as O-SIIs and counting globally for 84 percent of total assets), seven have a working SSM college, led by the ECB.
- Capital positions globally are adequate and soundness indicators improved partly due to the 2016 Asset Quality Review (AQR), but nonperforming loans (NPLs) remain high with notable differences between banks.
- Loan Loss Provision coverage is lowering, partly due to higher reliance on collateral and in some cases due to banking strategies to reduce NPLs by sales and write-offs.
- Credit risk remains the primary risk driver and is notably higher than in most other EU countries.
- Precise outcome of introduction of IFRS 9 on provisions is yet to be analyzed.
- Persistent presence of “high NPL Banks,” predominantly locally owned, requires attention to upcoming EU regulation and EBA guidelines on management of nonperforming and forborne exposures and statutory prudential backstops.

### Key supervisory observations and methodological issues
- The manual is conceived to outline concrete actions expected from all participants and to guarantee consistent quality of work; it references Bulgarian laws and BNB regulations and can refer publicly to EBA Guidelines where appropriate.
- New EBA draft Guidelines (October 2017) include more explicit guidance, recent changes in EU regulation, and a specific chapter on supervisory stress testing; these are expected to be applied in the 2019 SREP cycle. The manual has been adapted to include these changes.
- Unsound banking practices or regulatory breaches cannot be compensated by complementary capital charges under common standards; loans resulting from such practices should be isolated in the SREP process and be subject to adapted additional capital requirement for operational (reputational) risk and credit risk as an interim measure while deficiencies are addressed.
- Proportionality is included in the supervisory approach, but BSD should limit descriptive quarterly reporting and formally apply the multi-year supervisory cycle based on a minimum engagement as defined by the EBA (maximum three years), to increase efficiency and focus supervisory efforts.
- BSD has developed a methodology for the combined risk assessment and subsequent definition of an additional capital requirement for credit risk. The method is based on:
  - the level of NPLs,
  - their aging,
  - their coverage,
  - the relative importance of single large exposures,
  - the risk level of sector concentration,
  - and gives important weighting to the quality of risk management and internal control.
- The approach is close to methods chosen in other countries and fits the profile of the Bulgarian banking sector; it will soon be presented to the GC for immediate implementation during the current SREP cycle.
- Past onsite inspections and other supervisory work have not yet cleaned up inconsistencies and deficiencies in the Internal Capital Adequacy Assessment Process (ICAAP) developed by most banks—reported ICAAP numbers, especially for local banks, have not yet achieved reliability sufficient for easy use in SREP calculations, while global ICAAP levels are lower than expected.
- New supervisory expectations on stress testing offer an opportunity to plan well-prepared discussions with institutions and to include the effectiveness of ICAAP numbers in the review scope.
- Individual outcomes of top-down stress tests carried out by the Macroprudential Supervision and Financial Stability Directorate can contribute to SREP by assessing quality of internal control and capacity to produce complete and reliable data.
- When approved, the draft rules of order of the GC will specify its scope of responsibilities in supervisory matters and criteria used in assessment of motions presented by the Deputy Governor; BSD should use the SREP manual submission for approval to inform the GC on SREP consistency, key drivers and supervisory anchor points, and to give a global preliminary overview of expected decisions and sector-wide capital impact.
- Proposals for supervisory measures should be substantiated with:
  - comparison to peers or to prudent supervisory anchor points,
  - expected outcome of the measure,
  - expected date when situation will be submitted for evaluation,
  - possible next steps if measure is not effective.
- Communication of the SREP to banks currently mentions the amount of the Pillar 2 capital add-on but not the internal benchmarking or anchor points; common supervisory practice is to complement this with an overview of main findings, motivation for individual risk assessments that led to scores 3 and 4, and formal communication of the overall score.
- Yearly approval by the GC of the BSD action plan provides an opportunity to communicate overall outcome: main findings and scores, most important supervisory expectations, and global supervisory priorities.

### Table of Key Recommendations (select entries)
- Loan loss provisions and capital charges for loans granted as a result of unsound banking practices:
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: Onsite Directorate/Offsite Directorate
- Formally apply the multi-year supervisory cycle (maximum three years) and specify to GC that SREP is captured in this cycle:
  - Time Frame: June 2019
  - Implementing Department/Directorate: BSD
- Clarify interaction between directorates and contributions to assessment of governance and internal control:
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: BSD
- Invite relevant banks to test reliability and feasibility of business models, comparing BMA to budgets, past accounts, capital planning and ICAAP:
  - Time Frame: Sept 2019
  - Implementing Department/Directorate: Offsite directorate
- Continue strengthening the supervisory toolkit focused on NPLs and collateral value:
  - Time Frame: Dec 2019
  - Implementing Department/Directorate: BSD
- Implement use of a predefined set of Key Risk Indicators (KRI) on credit risk in yearly SREP:
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: Offsite directorate/Onsite
- Complete analysis of new methodology for additional capital requirements for credit risk and present outcome to GC:
  - Time Frame: Immediately
  - Implementing Department/Directorate: BSD
- Introduce assessment of operational risk:
  - Time Frame: Sept 2019
  - Implementing Department/Directorate: Offsite directorate/Onsite
- Prioritize meetings with all banks on effectiveness of ICAAP outcome:
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: Offsite / Onsite Directorate
- Use individual outcome of top-down stress tests as input on quality of internal control in the SREP:
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: Financial Stability Directorate
- Justify proportionality by systematic peer review and horizontal analysis; complement information to GC with peer comparisons and anchor points:
  - Time Frame: Sept 2019
  - Implementing Department/Directorate: BSD
- Always substantiate motivation and expected effectiveness of proposed decisions:
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: BSD
- Combine yearly GC approval of BSD action plan with information on overall outcome of review process:
  - Time Frame: Sept 2019
  - Implementing Department/Directorate: BSD
- Complete individual communication of the SREP outcome to banks with overview of main findings and weaknesses, and communication of final score:
  - Time Frame: Dec 2018
  - Implementing Department/Directorate: Offsite Directorate

### Capital and ICAAP — key findings
- All total capital ratios are comfortably above the OCR ratios, with a notable exception for one larger local bank.
- In eight out of nine larger banks the internal capital need specified in the ICAAP is far below the OCR.
- The sum of ICAAP estimates in that group is only 56 percent of the available capital.
- On the level of the global market, this ratio drops to 53 percent.
- The banking sector’s self-assessment of capital is very different from—and significantly less prudent than—the assessment by the authorities.

### SREP process and proportionality — observations and recommendations
- Observations:
  - BNB categorizes supervised institutions by systemic risk and size into four categories; category one covers other systemically important institutions (O-SIIs).
  - Principle of proportionality is repeatedly mentioned but not explicitly embedded in procedures.
  - All banks (except branches) are monitored quarterly; all individual SREP elements updated yearly and overall assessment produced yearly.
  - Granularity and proportionality not defined beyond broad references; scores may be outdated for smaller institutions where inspections occur at long intervals.
- Recommendation: adopt supervisory cycle of maximum three years with specifics:
  - All institutions included in yearly BSD action plan with onsite and offsite analysis based on size, activity and internal control quality.
  - KRIs for all institutions monitored quarterly with thresholds and triggers provoking immediate supervisory attention.
  - Documented summary of the SREP assessment and an institution-specific action plan produced yearly.
  - Category one institutions subject to yearly supervisory cycle.
  - Update all individual SREP elements at least every two years for category two and every three years for category three; these in-depth assessments drive two- to three-year action plans.
  - In intermediate years, formal decision on capital and liquidity may simply state BNB updated information and has no reason to review earlier assessments.
  - Yearly assessment letters for category 2 and 3 should indicate timing of subsequent full assessment.

### Interaction between directorates — recommended cooperation lines
- For each score of an individual risk, include a short motivation of the judgment and proposed supervisory measures.
- Offsite directorate should prepare final assessment and timely collect information from other directorates; it creates the internal calendar.
- Onsite directorate has final responsibility for scores related to areas subject to onsite analysis during the current supervisory cycle and contributes to related scores.
- Other directorates informed about final date of overall assessment and contribute where appropriate.
- Formal meetings with internal and external auditors of the institution are recommended contributions to the overall assessment.

### Business Model Analysis (BMA) — observations and recommendations
- Observations:
  - Most banks share a similar business model focused on local lending to retail, SME, and corporate markets; interest rate margin is main income source.
  - Some banks carry out specialized credit activities in subsidiaries; these should be considered in BMA.
  - BNB uses KRI to calculate recurrent profitability and efficiency and to score viability of business models; offsite directorate has evaluated business models for most group one and two banks.
  - Current reviews provide insight but reliability not tested; framework needs upgrading and alignment with accounting, budgeting, ICAAP and ILAAP.
  - Scores on viability and sustainability of business plans are not generally available.
- Recommendations to banks:
  - Indicate how and when they test reliability and feasibility of business models by comparing BMA numbers to budgets, past accounts, and ICAAP/ILAAP.
  - Discuss outcomes within management and supervisory board and transmit minutes to supervisor.
  - High NPL banks should review impact of upcoming EU regulation on high NPL and prudential backstops on business plan and internal structure.
  - Invite banks to integrate expected changes of ownership in their approach.

### Governance — observations and recommendations
- Observations:
  - Manual’s chapter on internal governance and institution-wide controls modeled on Draft EBA Guidelines; manual adapted to reflect EBA steps.
  - Methodology is new to BSD; governance and internal control assessment historically carried out by onsite directorate with CAMEL scoring.
  - Updating scores often requires new inspection, so many smaller institutions have outdated scores.
  - Conversion of CAMEL to new methodology showed overall results aligned with competent supervisors’ assessments: close to half of credit institutions obtain a global score 3 and 18 percent a score 4.
  - Ongoing offsite work (recovery plans, internal audit planning, regulatory compliance, fit and proper assessments) links to governance assessment.
- Recommendation:
  - Further clarify contributions of different directorates to overall governance and control assessment.
  - For each of the ten steps in internal governance assessment, indicate which supervisory activities normally contribute explicitly to supervisory judgment.
  - Ensure accurate, permanently updated insight into strengths and weaknesses with adequate formal follow-up.

### Credit risk — overview, prudential toolkit, and KRI monitoring
- Credit risk is the main risk driver in the Bulgarian banking system.
- 2016 AQR:
  - Led to large value adjustments and significant enhancements in internal control and risk management.
- Post-AQR:
  - NPL levels lowered overall but large differences remain; some banks remain “high NPL-banks.”
- Reporting improvements provide better insight into portfolio quality.
- Outstanding policy focus areas: underprovisioning and fair collateral valuation.
- Prudential toolkit — not concluded/implemented items:
  - Tool based on internal PD estimate for Bulgarian corporates not concluded due to insufficient historic data.
  - Comparison of banks’ NPL write-downs to a supervisory benchmark not realized.
  - Tools to identify underestimations and analyze IFRS 9 impact, enhanced knowledge-building, and guidance/adapted prudential metrics remained unimplemented.
- Partly implemented / ongoing:
  - Some tools included in 2018 SREP cycle methodical and sector-wide assessment leading to a Pillar 2 add-on.
  - Specific capital add-on for banks with low NPL coverage introduced; no comprehensive automated benchmarking tool yet.
  - System-wide periodic assessment of collateral valuation practices vs. BNB metrics missing; a specific capital add-on will be applied to high NPL-banks with low coverage ratios for loans more than one year past due.
- BSD Policy Directorate impact study:
  - Assessed introduction of EU regulation on prudential backstops (annual increase for capital coverage of NPLs and gradual lowering of collateral value) via a static sector-wide simulation to measure capital impact.

### KRI-based SREP first-step and quarterly monitoring
- Quarterly monitoring:
  - All credit institutions and branches monitored quarterly by Key Risk Indicators (KRIs); standardized quarterly report provides overview of financial situation and main risk indicator evolution.
  - Draft SREP manual: quarterly monitoring used to identify material changes; full report produced only when material changes identified.
- Yearly SREP first-step KRI set:
  - Consumer loans and growth of the portfolio
  - Retail mortgage loans and global loan to value ratio of this portfolio
  - SME loans and NPL ratio of this portfolio
  - Corporate loans portfolio and concentration on sector and individual counterparties
  - Complementary KRIs:
    - Level and evolution of the coverage ratio
    - NPL level and part of NPL-loans more than 360 days due in total NPL portfolio
    - Value and evolution of collateral value of foreclosed assets
  - 2018 SREP will include additional information on impact of IFRS 9 on provisions on transition date and relative importance to total provisions.
  - Outliers vs. market averages or supervisory benchmarks can trigger offsite analysis and focused onsite inspections.

### Methodical assessment on NPLs, coverage, concentration and capital add-ons
- BSD methodology inputs and anchors:
  - NPL level, NPL aging, coverage, single large exposures, sector concentration based on HHI index (from Finrep).
  - Sequential comparison to prudently chosen anchor points:
    - Level of NPL
    - Level of coverage ratio
    - For banks exceeding both NPL-level and coverage anchors: additional benchmarking to prudent levels of:
      - Part of NPLs longer than one-year overdue in total NPL
      - Coverage ratio for this specific category
    - Sum of large exposures versus total eligible capital
    - Sector concentration based on HHI index
- Add-on mechanics:
  - When anchor points are exceeded, a step-by-step add-on to risk-weighted assets for credit risk is created, evolving from 0 percent to 1.50 percent in 6 steps, counting each for 0.25 percent.
  - For banks with high NPLs and low coverage, matrix outcome is combined with assessment of credit risk management and internal control:
    - Scores 1 and 2 (strong/adequate) = no add-on.
    - Score 3 (“medium to high” risk / weak risk management) = adds 1.25 percent to the add-on.
    - Score 4 (“high risk” / inadequate) = adds 2.75 percent to the add-on.
  - Credit risk generally represents 85 percent of total capital charges; total capital add-on for banks with high risk volumes but sound risk management can be expected to be around 1.3 percent.
- Preliminary impact results (method in finalization):
  - Gradual add-on to risk weighted assets is 1 percent or higher for 6 banks.
  - 9 banks are impacted with 0.5 percent or 0.75 percent when quality of credit risk management is neutralized.
  - For three banks with material weaknesses in risk management/internal control, the Pillar 2 capital add-on for credit risk will be above 2 percent.
  - For five banks, impact varies between 1 percent and 2 percent.
  - Further methodological work and sensitivity analysis planned before final proposal to the GC and SREP manual submission.

### Interest rate risk in the banking book, operational and market risk
- Interest rate risk in the banking book:
  - Draft SREP manual includes qualitative assessment and a quantitative indicator based on change in economic value of capital and change in net interest income from an unfavorable interest-rate change.
  - Actual introduction and calibration require adapted reporting from January 1, 2019.
- Operational risk:
  - Gross income and asset size used as reliable indicators for operational risk given similar business models.
  - Only two banks (both subsidiaries) use the advanced measurement approach for regulatory capital.
  - Operational risk assessments mainly onsite; offsite directorate holds information on important losses.
  - Reputation risk and ICT risk monitoring highlighted as important; lack of ICT audits in internal audit reports may indicate insufficient ICT risk follow-up.
- Market risk:
  - Market risk level relatively low; impact in global assessment limited.
  - General assessment framework available, including benchmarking exposure to market risk relative to total risk exposure and contribution of return on financial assets to total profit.

### ICAAP, ILAAP, supervisory review, and stress testing
- ICAAP and ILAAP supervisory review:
  - Executed as part of onsite missions or overall assessments; onsite analyses mostly cover larger banks.
  - Inspections frequently identify deficiencies in ICAAP effectiveness and lack of evidence that ICAAP capital is used in decision-making and management.
  - Market-wide ICAAP estimate stands at 53 percent of available capital.
    - Seven banks present ICAAP outcomes identical or lower than Pillar 1 calculations for the leading credit risk indicator.
    - Local banks’ reported ICAAP numbers not yet sufficiently reliable for direct use in SREP calculations.
  - Periodic assessment of stress-testing capacity and analysis of institutions’ stress tests expected to complement ICAAP information and improve ICAAP quality.
- Stress testing in supervision:
  - Draft SREP manual includes supervisory stress testing per EBA Guidelines; development deferred until 2019 SREP cycle.
  - 2016 AQR provided meaningful information on vulnerabilities, data quality, and internal structures.
  - Top-down macroeconomic stress tests by Macroprudential Supervision and Financial Stability Directorate contribute to SREP via outcomes and quality scores.
  - Initial supervisory micro stress testing focus: review and quality assessment of institutions’ internal stress testing.
  - Supervisory bottom-up stress testing planned from 2019 SREP cycle.

### Decision process, GC role, and communication to banks
- Governing Council (GC) intervention:
  - Recent Banking Act change: supervisory measures are issued by the BNB GC on a motion by the Deputy Governor rather than by the Deputy Governor alone.
  - Draft rules of order for the GC specify responsibilities and criteria for assessment of motions to avoid duplication of legal work.
  - GC issues administrative acts that impose specific Pillar 2 capital or liquidity requirements or other supervisory measures.
  - Annual SREP letters that create new obligations (Pillar 2 add-ons or specific liquidity requirements) will be considered administrative acts; letters that do not create new obligations will not trigger GC intervention.
- Proportionate decision taking and focused planning:
  - Manual provides legal references and motivation for supervisory requirements and systematic association of possible actions to regulatory breaches, deficiencies or weaknesses.
  - For banks globally rated 3, supervisory measures are taken in the form of a decision by the GC.
  - For banks globally rated 4, it is imperative to take supervisory measures.
  - Recommendation: when informing the GC about an institution, compare relevant KRIs to peers and prudent anchor points and indicate how similar deficiencies at other institutions have been addressed.
  - Motions to GC should include:
    - Summary of findings,
    - Motivation and anchor points,
    - Identification of supervisory measure and expected outcome,
    - Timing for effectiveness evaluation,
    - Possible next steps if measure is not effective.
- Communication to banks — individual decisions:
  - Manual does not contain an example letter; EBA draft guidelines contain examples.
  - Typical communication content:
    - Amount of the Pillar 2 capital add-on (typically mentioned) but not the specific internal benchmarking or anchor points used.
    - Overview of main findings during review.
    - Motivation for individual risk assessments leading to scores 3 and 4.
    - Formal communication of the global score.
  - Recommendation: inform bank of way and date effectiveness of measure will be evaluated and indicate possible consequences if expected outcome not realized.
- Communication to banks — global process and methodology:
  - Pursuant to Article 102 of the LCI the BNB has legal obligation to disclose general criteria and methodologies used in the SREP; BNB considers this provision fulfilled, stating supervisory approach and methodologies are fully compliant with EBA Guidelines.
  - Trend toward openness: authorities increasingly share common findings, global outcome of risk scores and additional requirements to capital and liquidity, and main supervisory expectations.
  - Yearly GC approval of BSD action plan offers opportunity to communicate overall outcome, main findings and scores, and global supervisory priorities; extracts can be made public or used with other authorities.

*Source: 1bgrea2019003 - Preface (draft SREP manual review and mission report).*

### Preface.................................................................................................................

### Preface

### Mission, purpose, and scope
- MCM TA mission visited Sofia during the period July 16–20, 2018 to assist the Bulgarian National Bank (BNB).
- Advisor: Mr. Filip Gijsel, short-term Bank Supervision advisor for the Monetary and Capital Markets Division (MCM) of the IMF.
- Purpose: review the formalization and implementation of a comprehensive SREP that includes an explicit and detailed supervisory Pillar 2 capital requirement; review the draft SREP manual that is currently being finalized in the BSD.
- The manual is intended to form an integral part of the Banking Supervisory Process Manual (approved February 9, 2017 by the GC) and its approval by the GC is expected in the coming months.
- The review relies on the version of the manual received on July 4, 2018 and takes account of IMF and World Bank missions in 2015, 2016, and 2017 on the observance of the Basel Core Principles.

### Executive-summary findings and context
- The BNB SREP is based on EU-wide common procedures, processes and methodologies; the draft SREP Manual includes recently proposed changes to the Guidelines issued by the European Banking Authority (EBA).
- The report focuses on implementation into appropriate and proportionate supervisory assessments and into focused actions and planning.
- During the SREP process in 2016 and 2017 the BNB did not identify the need to impose additional capital requirements as per Pillar 2; consequently no Total Supervisory Capital Requirement (TSCR) and Overall Capital Requirement (OCR) were communicated to banks.
- The assessment was communicated in all EU-colleges but was not for all banks summarized in a standardized SREP framework.
- The prominent presence on the Bulgarian market of banking groups under direct supervision by the Single Supervisory Mechanism (SSM) causes SSM processes for the SREP to already have a predominant place in BSD activities.
- The Bulgarian Government requested to enter into close cooperation with the ECB on the supervision of banks (article 7 of Council Regulation 1024/2013 of October 15, 2013); cooperation with the SSM will lead to harmonized procedures and methodologies, and a common decision process for most administrative acts, creating additional workload in the transition period.
- The BNB has categorized supervised institutions into four groups based on size, structure, scope of activities and complexity; out of the 11 institutions in the first category (identified by the BNB GC as O-SIIs and counting globally for 84 percent of total assets), seven have a working SSM college, led by the ECB.
- Capital positions globally are adequate and soundness indicators have improved partly due to the 2016 Asset Quality Review (AQR), but nonperforming loans (NPLs) remain high in Bulgaria with notable differences between banks.
- Loan Loss Provision coverage is lowering, partly due to higher reliance on collateral and in a limited number of cases due to banking strategies to reduce NPLs by sales and write-offs.
- Credit risk remains the primary risk driver in the Bulgarian financial sector and is notably higher than in most other EU countries.
- The precise outcome of the introduction of IFRS 9 on provisions is yet to be analyzed.
- Persistent presence of “high NPL Banks,” predominantly owned by local shareholders, requires attention to the impact of upcoming EU regulation and EBA guidelines on management of nonperforming and forborne exposures and statutory prudential backstops.

### Key supervisory observations and methodological issues
- The manual is conceived to outline concrete actions expected from all participants and to guarantee consistent quality of work; it references Bulgarian laws and BNB regulations and can refer publicly to EBA Guidelines where appropriate.
- New EBA draft Guidelines (October 2017) include more explicit guidance, recent changes in EU regulation, and a specific chapter on supervisory stress testing; these are expected to be applied in the 2019 SREP cycle. The manual has been adapted to include these changes.
- Unsound banking practices or regulatory breaches cannot be compensated by complementary capital charges under common standards. Loans resulting from such practices should be isolated in the SREP process and, even when performing or not breaching concentration rules, be subject to adapted additional capital requirement for operational (reputational) risk and credit risk as an interim measure while deficiencies are addressed.
- Proportionality is included in the supervisory approach, but BSD should limit descriptive quarterly reporting and formally apply the multi-year supervisory cycle based on a minimum engagement as defined by the EBA (maximum three years), to increase efficiency and focus supervisory efforts.
- BSD has developed a methodology for the combined risk assessment and subsequent definition of an additional capital requirement for credit risk. The method is based on:
  - the level of NPLs,
  - their aging,
  - their coverage,
  - the relative importance of single large exposures,
  - the risk level of sector concentration,
  - and gives important weighting to the quality of risk management and internal control.
- The approach is close to methods chosen in other countries and fits the profile of the Bulgarian banking sector; it will soon be presented to the GC for immediate implementation during the current SREP cycle.
- Past onsite inspections and other supervisory work have not yet cleaned up inconsistencies and deficiencies in the Internal Capital Adequacy Assessment Process (ICAAP) developed by most banks—reported ICAAP numbers, especially for local banks, have not yet achieved reliability sufficient for easy use in SREP calculations, while global ICAAP levels are lower than expected.
- New supervisory expectations on stress testing offer an opportunity to plan well-prepared discussions with institutions and to include the effectiveness of ICAAP numbers in the review scope.
- Individual outcomes of top-down stress tests carried out by the Macroprudential Supervision and Financial Stability Directorate can contribute to SREP by assessing quality of internal control and capacity to produce complete and reliable data.
- When approved, the draft rules of order of the GC will specify its scope of responsibilities in supervisory matters and criteria used in assessment of motions presented by the Deputy Governor; BSD should use the SREP manual submission for approval to inform the GC on SREP consistency, key drivers and supervisory anchor points, and to give a global preliminary overview of expected decisions and sector-wide capital impact.
- Proposals for supervisory measures should be substantiated with:
  - comparison to peers or to prudent supervisory anchor points,
  - expected outcome of the measure,
  - expected date when situation will be submitted for evaluation,
  - possible next steps if measure is not effective.
- Communication of the SREP to banks currently mentions the amount of the Pillar 2 capital add-on but not the internal benchmarking or anchor points; common supervisory practice is to complement this with an overview of main findings, motivation for individual risk assessments that led to scores 3 and 4, and formal communication of the overall score.
- The yearly approval by the GC of the BSD action plan provides an opportunity to communicate the overall outcome of the review process: main findings and scores, most important supervisory expectations, and global supervisory priorities.

### Table of Key Recommendations (select entries)
- Principle: Loan loss provisions and capital charges for loans granted as a result of unsound banking practices cannot be created and judged on the basis of common standards. These loans should be isolated in the SREP process and, even when they are performing or not risking to create breaches to risk concentration rules, be subject to adapted additional capital requirement for operational (reputational) risk and credit risk.
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: Onsite Directorate/Offsite Directorate
- Organization: Formally apply the multi-year supervisory cycle, based on a minimum engagement, and specify to GC that SREP is captured in this cycle of maximum three years.
  - Time Frame: June 2019
  - Implementing Department/Directorate: BSD
- Organization: Further specify the interaction between directorates and clarify the contributions to the overall assessment of governance and internal control.
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: BSD
- Business Model Analysis: Invite relevant banks to indicate how and when they test the reliability and feasibility of their business models, by comparing the BMA numbers to data available in budgets and past accounts, and to the data that result from capital planning and ICAAP.
  - Time Frame: Sept 2019
  - Implementing Department/Directorate: Offsite directorate
- Credit risk: Continue the strengthening of the supervisory toolkit, focused on NPL’s and collateral value.
  - Time Frame: Dec 2019
  - Implementing Department/Directorate: BSD
- Credit risk: Implement the use of a predefined set of Key Risk Indicators (KRI) on credit risk in yearly SREP.
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: Offsite directorate/Onsite
- Credit risk: Complete the analysis on the global outcome of the new methodology for the definition of additional capital requirements for credit risk and present the outcome and an overview the expected SREP decisions to the GC.
  - Time Frame: Immediately
  - Implementing Department/Directorate: BSD
- Operational risk: Introduce assessment of operational risk.
  - Time Frame: Sept 2019
  - Implementing Department/Directorate: Offsite directorate/Onsite
- ICAAP: Give high priority to the start of a series of well-prepared meetings with all banks on the effectiveness of their ICAAP outcome.
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: Offsite / Onsite Directorate
- Stress testing: Use individual outcome of top-down stress tests as input on quality of internal control in the SREP.
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: Financial Stability Directorate
- Decision taking: Justify proportionality by systematic peer review and horizontal analysis and complement information to GC with comparison to peer and to anchor points.
  - Time Frame: Sept 2019
  - Implementing Department/Directorate: BSD
- Decision taking: Always substantiate motivation and expected effectiveness of proposed decisions.
  - Time Frame: Sept 2018
  - Implementing Department/Directorate: BSD
- Communication: Combine the yearly approval by the GC of the BSD action plan, with an information on the overall outcome of the review process, the main findings and scores, the most important supervisory expectations and global supervisory priorities.
  - Time Frame: Sept 2019
  - Implementing Department/Directorate: BSD
- Communication: Complete individual communication of the SREP outcome to banks with overview of main findings and weaknesses, and communication of final score.
  - Time Frame: Dec 2018
  - Implementing Department/Directorate: Offsite Directorate

### Introductory observations (I. Introduction, Scope, and Summary of Work Done)
- The SREP manual aligns with Directive 2013/36/EU and EBA Guidelines; national regulations with regard to the SREP are harmonized in all EU countries.
- EBA convergence is promoted through common Standards and Guidelines and assessed mainly through bilateral visits and calls; the last such assessment took place in the second semester 2017.
- The manual is not intended to be a complete copy of EBA Guidelines; public communication can refer directly to EBA Guidelines.
- The BNB was informed during the mission that the Bulgarian Government requested close cooperation with the ECB on supervision; the timeline and further course of this request are not clearly established.
- Footnote: All numbers on December 31, 2017, unless otherwise indicated.

*Source: 1bgrea2019003 - Preface (draft SREP manual review and mission report).*

### 9. With a notable exception for one larger local bank, all total capital ratios are

### 1bgrea2019003 - 9. With a notable exception for one larger local bank, all total capital ratios are

### Key findings on capital and ICAAP
- All total capital ratios are comfortably above the OCR ratios, with a notable exception for one larger local bank.
- In eight out of nine larger banks the internal capital need specified in the ICAAP is far below the OCR.
- The sum of ICAAP estimates in that group is only 56 percent of the available capital.
- On the level of the global market, this ratio drops to 53 percent.
- The banking sector’s self-assessment of capital is therefore very different from—and significantly less prudent—than the assessment by the authorities.

### Scope of the review (paragraphs 10–12)
- Focus: SREP implementation to ensure draft rules and procedures lead to appropriate and proportionate supervisory assessments and focused actions and planning.
- Timing consideration: Recent changes in the EBA guidelines, increased attention to business model analysis (BMA) and the extension of guidelines to supervisory stress testing will only come into effect in the 2019 SREP cycle, leaving room to develop or adapt supervisory tools and benchmarks.
- Specific supervisory attention agreed with BSD management:
  - Micro stress testing in the ICAAP and SREP process.
  - Supervisory benchmarks, especially in relation to credit risk.
  - Supervisory reactions and an efficient and proportionate decision process.
  - Communication of SREP outcomes to the banks.
- Mission aims:
  - Ensure current process complies with EBA standards and the manual, considering Bulgarian sector characteristics.
  - Encourage banks to develop an internal process to determine optimal capital level given risk profile and capacity, submitting the report to the supervisor.
  - Enable supervisors to review bank submissions and develop a process to determine and apply any additional capital requirements.
  - Tailor supervision of individual banks based on evaluation of each bank’s internal capital adequacy assessment.

### Summary of work done (paragraphs 13–14)
- Reviewed the draft SREP implementation guidelines prepared by the BNB, based on EBA standards and guidelines.
- Discussed manual content with onsite inspection and offsite directorate Directors; followed a “walk through” of steps leading to final SREP capital and liquidity assessment.
- Organized in-depth analysis and discussions on the assessment system and process, SREP elements, implementation and decision processes, planning and proportionality, and communication to banks.
- The report is “exception based”: it does not provide an overview of the manual content nor deal with recommendations outside SREP scope, but gives context to understand motivation and possible impact of recommendations.

### SREP principles and prudential stance (paragraphs 15–19)
- The SREP is a key supervisory process providing a condensed overview of supervisory activities, concerns and future actions.
- Even where overall ratios are adequate, high global risk scores remain recurrent and some individual scores require enhanced supervisory scrutiny.
- Unsound banking practices or regulatory breaches cannot be compensated by complementary capital charges; EBA guidelines explicitly allow setting additional capital requirements to cover risks from deficiencies in control and governance as an interim measure.
- Recent Bulgarian public concerns: press publications on excessive loans to parties related to shareholders and privileged clients, opaque structures, and shell companies with unidentified ultimate beneficiary ownership.
- Loans granted without well-developed internal procedures, unclear counterparties, unreliable ultimate owner financial information, or uncertain project profitability lead to diminished transparency and additional credit and operational risk; deterioration of such loans leads to long, complex, costly restructurings and lower realization value of collateral.
- Loan loss provisions and capital charges for such loans cannot be created and judged on common standards; these loans should be isolated in the SREP process and subject to adapted additional capital requirements for operational (reputational) risk and credit risk.
- Singular treatment justified for all material loans granted on unsound basis, even when performing or not breaching concentration rules; specific provisions and capital requirements can be imposed to cover operational risk associated with claims or money laundering enquiries resulting from weak internal control or management fraud.
- For banks where the 2016 AQR revealed significant documentation weaknesses in certain loan/guarantee categories, this “questionable” pre-AQR portfolio should be isolated and treated more stringently in the SREP.

### Process and proportionality (paragraphs 20–25) — observations and recommendations
- Observations:
  - BNB categorizes supervised institutions by systemic risk: category one covers other systemically important institutions; categories two and three cover other institutions incorporated in Bulgaria based on size; category four includes branches of foreign banks (lowest supervisory intensity).
  - Principle of proportionality is repeatedly mentioned but not explicitly included in procedures.
  - All banks (except branches) are monitored quarterly; all individual SREP elements are updated yearly and overall assessment produced yearly.
  - Granularity and proportionality are not defined beyond references to size, systemic importance, nature, scale, and complexity.
  - RAS system of BSD provides lines to focus supervisory scrutiny and avoids mechanical yearly “checking all the boxes”.
  - Different SREP elements vary in relevance across institutions; appropriate degrees of depth and granularity should be applied per category.
  - Scores may be outdated for smaller institutions where inspections occur at long intervals.
- Recommendation: Adopt a supervisory cycle of maximum three years to capture the SREP process globally, with specifics:
  - All institutions included in yearly BSD action plan with onsite and offsite analysis based on size, activity nature, and internal control quality; updates to relevant SREP element assessments and possible impacts on next year’s supervisory action plan.
  - KRI’s for all institutions monitored quarterly; thresholds and triggers provoke immediate supervisory attention and measures for identified weaknesses or unexpected developments.
  - A documented summary of the SREP assessment and an institution-specific action plan produced yearly.
  - Institutions in category one subject to yearly supervisory cycle.
  - Update of all individual SREP elements, as a minimum, every two years for category two institutions, and every three years for category three institutions; these complete in-depth assessments drive institution-specific action plans covering two to three years.
  - In intermediate years, the formal decision on capital and liquidity requirements and communication to banks limited to the statement that BNB updated its information and has no reason to review earlier assessments.
  - When institutions in category 2 and 3 receive the yearly assessment letter, it should indicate that, barring intervening elements, a subsequent full assessment will be communicated in two or three years.

### Interaction between directorates (paragraphs 26–27)
- Context:
  - Many Bulgarian banks are part of EU banking Groups whose home supervisor is the SSM; therefore SREP calendar for these banks is largely determined by the SSM, with specific deadlines for Joints Supervisory Teams and BNB contributions.
  - High demands on efficiency of interaction between BSD directorates, especially for internal governance and control assessments where multiple directorates contribute.
- Recommended cooperation lines to enhance efficiency:
  - For each score of an individual risk, include a short motivation of the judgment and proposed supervisory measures.
  - Offsite directorate should prepare the final assessment and timely collect information from other directorates; this directorate creates the internal calendar.
  - Onsite directorate has final responsibility for scores directly related to areas subject to onsite analysis during the current supervisory cycle and can contribute to related scores.
  - Other directorates informed about the final date of the overall assessment and contribute when appropriate.
  - Formal meetings with internal and external auditors of the concerned institution are recommended contributions to the overall assessment.

### Business Model Analysis (BMA) (paragraphs 28–32)
- Observations:
  - Most Bulgarian banks share a similar business model focused on local lending to retail, SME, and corporate markets; the relative importance of these markets varies. Interest rate margin is the main income source.
  - Some banks carry out specialized credit activities in subsidiaries; this should be considered in BMA.
  - BNB uses a set of KRI to calculate recurrent profitability and efficiency and to score viability of business models for all banks; offsite directorate conducts surveys on subsidiaries’ business models in preparation for annual college meetings and has evaluated business models for most group one and two banks based on policy documents and reports.
  - Current reviews provide insight but reliability has not been tested, rarely giving grounds for well-motivated assessments and supervisory reactions.
  - BMA was included in onsite inspections in a limited number of banks, revealing that the framework needs upgrading and should be aligned to accounting, budgeting, ICAAP and ILAAP.
  - Scores on viability and sustainability of business plans are not generally available.
  - No indications that business models need fundamental changes to be sustainable, but little evidence business models offer all guarantees expected in supervisory guidelines.
- Recommendations for banks:
  - Invite relevant banks to indicate how and when they test reliability and feasibility of their business models by comparing BMA numbers to budgets, past accounts, and capital planning/ICAAP data.
  - Expect outcomes of these tests to be discussed within management and supervisory board, and request minutes be transmitted to the supervisor.
  - High NPL banks should review the impact of upcoming EU regulation on high NPL and prudential backstops on their business plan and internal structure.
  - Banks can be invited to integrate expected changes of ownership in their approach.
  - This information will facilitate SREP scoring and indicate areas needing further supervisory attention.

### Governance (paragraphs 33–36)
- Observations:
  - The manual’s chapter on internal governance and institution-wide controls is modeled on the Draft EBA Guidelines; manual recently adapted to reflect EBA steps.
  - Methodology is new to BSD; to date governance and internal control assessment carried out by onsite directorate when concluding inspections, with scoring based on CAMEL approach.
  - Because updating scores requires a new inspection, many smaller institutions have outdated scores.
  - A global conversion of CAMEL scores to the new methodology showed overall results did not deviate from competent supervisors’ assessments; overall score level is in line with expectations: close to half of credit institutions obtain a global score 3 and 18 percent a score 4.
  - BSD cannot rely solely on onsite inspectors to judge governance and control; ongoing offsite work (recovery plans, internal audit planning, regulatory compliance, fit and proper assessments) also links to governance assessment.
- Recommendation:
  - Further clarify contributions of different directorates to overall governance and control assessment.
  - For each of the ten steps in the assessment of internal governance, indicate which supervisory activities normally result in an explicit contribution to the supervisory judgment.
  - Ensure accurate, permanently updated insight into strengths and weaknesses of governance and control mechanisms with adequate and formal follow-up of implementation.

### Risk to Capital
- Heading present in source: "E. Risk to Capital" with subheading "Credit Risk Benchmarking of Credit Risk" — content beyond this heading is not included in the provided excerpt.

*Source: 1bgrea2019003 - 9. With a notable exception for one larger local bank, all total capital ratios are*

### 37. Credit risk is the main risk driver in the Bulgarian banking system, and reliable

### 37. Credit risk is the main risk driver in the Bulgarian banking system, and reliable

### Credit-risk overview and recent developments
- Credit risk is identified as the main risk driver in the Bulgarian banking system.  
- 2016 AQR results:
  - Led to large value adjustments.
  - Contributed to significant enhancements in internal control and risk management.
- Post-AQR:
  - NPL levels have lowered overall, but large differences remain between banks; some remain “high NPL-banks.”
- Reporting improvements:
  - New and more detailed reporting on non-performing and forborne loans provides better insight into credit portfolio quality.
- Key policy focus areas left outstanding:
  - Underprovisioning and fair collateral valuation remain priorities.
  - Several supervisory-suggested measures were not implemented (see Prudential toolkit & measures).

### Prudential toolkit: measures proposed, partially implemented, and studies
- Not concluded or implemented measures:
  - Development of a tool based on an internal PD estimate for Bulgarian corporates to capture excess idiosyncratic risk under Pillar 2— not concluded due to insufficient historic data.
  - Comparison of banks’ treatment of NPL write-downs to a supervisory benchmark—not realized.
  - Tools to identify possible underestimations and to analyze the impact of IFRS 9, enhanced knowledge-building among supervision officers/board members/external auditors, and introduction of guidance/adapted prudential metrics—all remained unimplemented.
- Partly implemented / ongoing:
  - Some suggested tools are partly included in the methodical and sector-wide assessment of credit risk used in the 2018 SREP cycle and leading to a Pillar 2 add-on.
  - A specific capital add-on for banks with a low coverage ratio for NPLs is introduced, but no comprehensive automated benchmarking tool yet.
  - A system-wide periodic assessment of collateral valuation practices vs. BNB metrics is still missing; however, a specific capital add-on will be applied to high NPL-banks with low coverage ratios for loans more than one year past due.
- BSD Policy Directorate impact study:
  - Assessed introduction of EU regulation on prudential backstops (annual increase for capital coverage of NPLs and gradual lowering of collateral value).
  - The static sector-wide analysis simulated the immediate and full introduction of valuation rules close to strict supervisory metrics and measured impact on capital across institutions.

### Supervisory monitoring, SREP process, and KRI-based assessments
- Quarterly monitoring and reporting:
  - All credit institutions and branches are monitored quarterly by Key Risk Indicators (KRIs).
  - A standardized quarterly report provides an overview of financial situation and evolution of main risk indicators.
  - Draft SREP manual: ongoing quarterly monitoring used to identify changes in financial conditions and risk profiles; full report produced only when material changes are identified.
- Yearly SREP first-step KRI set to identify major credit-risk carriers and gross risk volumes:
  - Consumer loans and growth of the portfolio
  - Retail mortgage loans and global loan to value ratio of this portfolio
  - SME loans and NPL ratio of this portfolio
  - Corporate loans portfolio and concentration on sector and on individual counterparties
  - Complementary KRIs:
    - Level and evolution of the coverage ratio
    - NPL level and part of the NPL-loans that are more than 360 days due in the total NPL portfolio
    - Value and evolution of collateral value of foreclosed assets
  - 2018 SREP will include additional information on the impact of IFRS 9 on provisions on the transition date and relative importance to total provisions.
  - Outliers vs. market averages or supervisory benchmarks can trigger offsite analysis and focused onsite inspections.

### Methodical assessment on NPLs, coverage, concentration and capital add-ons
- BSD methodology inputs and anchors:
  - Uses NPL level, NPL aging, coverage, single large exposures, sector concentration based on HHI index (from Finrep).
  - Sequential comparison of indicators to prudently chosen anchor points:
    - Level of NPL
    - Level of coverage ratio
    - For banks exceeding both NPL-level and coverage anchors: additional benchmarking to prudent levels of:
      - Part of NPLs longer than one-year overdue in total NPL
      - Coverage ratio for this specific category
    - Sum of large exposures versus total eligible capital
    - Sector concentration based on HHI index
- Add-on mechanics:
  - When anchor points are exceeded, a step-by-step add-on to risk-weighted assets for credit risk is created, evolving from 0 percent to 1.50 percent in 6 steps, counting each for 0.25 percent.
  - For banks with high NPLs and low coverage, the matrix outcome is combined with assessment of credit risk management and internal control:
    - Scores 1 and 2 (strong/adequate) = no add-on.
    - Score 3 (“medium to high” risk / weak risk management) = adds 1.25 percent to the add-on.
    - Score 4 (“high risk” / inadequate) = adds 2.75 percent to the add-on.
  - Credit risk generally represents 85 percent of total capital charges; total capital add-on for banks with high risk volumes but sound risk management can be expected to be around 1.3 percent.
- Preliminary impact results (method in finalization):
  - Gradual add-on to risk weighted assets is 1 percent or higher for 6 banks.
  - 9 banks are impacted with 0.5 percent or 0.75 percent when quality of credit risk management is neutralized.
  - For three banks with material weaknesses in risk management/internal control, the Pillar 2 capital add-on for credit risk will be above 2 percent.
  - For five banks, impact varies between 1 percent and 2 percent.
  - Further methodological work and sensitivity analysis planned before final proposal to the GC and SREP manual submission.

### Interest rate risk in the banking book, operational and market risk
- Interest rate risk in the banking book:
  - Draft SREP manual includes qualitative assessment and a quantitative indicator based on change in economic value of capital and change in net interest income from an unfavorable interest-rate change.
  - Actual introduction and calibration require adapted reporting from January 1, 2019.
- Operational risk:
  - Gross income and asset size used as reliable indicators for operational risk given similar business models.
  - Only two banks (both subsidiaries) use the advanced measurement approach for regulatory capital.
  - Operational risk assessments have mainly been onsite; offsite directorate holds information on important losses.
  - Reputation risk and ICT risk monitoring highlighted as important components; lack of ICT audits in internal audit reports may indicate insufficient ICT risk follow-up.
- Market risk:
  - Market risk level is relatively low; its impact in global assessment is limited.
  - General assessment framework available, including benchmarking exposure to market risk relative to total risk exposure and contribution of return on financial assets to total profit.

### ICAAP, ILAAP, supervisory review, and stress testing
- ICAAP and ILAAP supervisory review:
  - Executed as part of onsite missions or overall assessments; onsite analyses mostly cover larger banks.
  - Inspections frequently identify deficiencies in ICAAP effectiveness and lack of evidence that ICAAP capital is used in decision-making and management.
  - Market-wide ICAAP estimate stands at 53 percent of available capital.
    - The banking sector’s self-assessment differs markedly from authorities’ expectations.
    - For the leading credit risk indicator, seven banks present outcomes identical or lower than Pillar 1 calculations.
    - Local banks’ reported ICAAP numbers have not yet achieved sufficient reliability for direct use in SREP calculations.
  - Periodic assessment of stress-testing capacity and analysis of institutions’ stress tests expected to complement ICAAP information and improve ICAAP quality through well-prepared discussions.
- Stress testing in supervision:
  - Draft SREP manual includes supervisory stress testing per EBA Guidelines; supervisory stress testing development deferred until the 2019 SREP cycle.
  - 2016 AQR provided meaningful information on vulnerabilities, data quality, and internal structures.
  - Top-down macroeconomic stress tests by Macroprudential Supervision and Financial Stability Directorate contribute to SREP via outcomes and quality scores.
  - Initial supervisory micro stress testing focus: review and quality assessment of institutions’ internal stress testing.
  - Supervisory bottom-up stress testing planned from 2019 SREP cycle; top-down results and local stress-testing scenario quality (credit risk and collateral valuation) are critical inputs.

### Decision process and governance
- Governing Council (GC) intervention in SREP decisions:
  - Recent Banking Act change: supervisory measures are issued by the BNB GC on a motion by the Deputy Governor rather than by the Deputy Governor alone.
  - Draft rules of order for the Governing body specify GC responsibilities and criteria for assessment of motions, aiming to avoid duplication of legal work.
  - GC issues administrative acts that impose specific Pillar 2 capital or liquidity requirements or other supervisory measures.
  - Annual SREP letters that create new obligations (Pillar 2 add-ons or specific liquidity requirements) will be considered administrative acts; letters that do not create new obligations will not trigger GC intervention.

*IMF mission chapter: supervisory assessment and policy measures on credit risk, prudential toolkit, SREP methodology, ICAAP/ILAAP, and stress testing.*

### 62. The GC also approves BNB regulations and guidelines, as well as the manuals of the

### 62. The GC also approves BNB regulations and guidelines, as well as the manuals of the

### Proportionate Decision Taking and Focused Planning
- The manual provides:
  - A complete overview of the legal references and motivation for supervisory requirements on capital and liquidity, and other supervisory measures and actions.
  - Systematic association of possible supervisory actions to regulatory breaches, deficiencies or observed weaknesses.
  - Guidance to supervisors on how to formulate findings and motivate proposed reactions.
- Supervisory decision-taking by global score:
  - For banks globally rated 3, supervisory measures are taken in the form of a decision by the GC.
  - For banks globally rated 4, it is imperative to take supervisory measures.
- Use of benchmarking and transversal analysis:
  - Proportionality in exercising supervisory powers is often based on benchmarking findings to prudently chosen supervisory anchor points.
  - Systematic peer comparisons and wide scale transversal analysis are inseparable from supervision of individual institutions to underpin supervisory judgment and motivate consistency.
  - Recommendation: when informing the GC about an individual institution, compare relevant KRI’s to peers and to prudent anchor points and indicate how similar deficiencies at other institutions have been addressed or are being addressed.
- The manual recommends that proposed decisions and motions for decision by the GC be substantiated by explicitly providing:
  - Summary of the findings,
  - Motivation and anchor points for supervisory judgment,
  - Identification of the supervisory measure and the expected outcome of the measure,
  - Expected moment measure will come into effect and the moment the situation will be submitted to evaluate this outcome,
  - Possible next steps if the measure is not effective at the expected moment.
- Communication of measures:
  - If supervisory measures are announced in the letter confirming the SREP outcome, it is recommended to inform the bank on the way and date the effectiveness of the measure will be evaluated and to indicate the possible consequences and additional measures or sanctions that can be taken into consideration if the expected outcome is not sufficiently realised.
- Notable contextual input:
  - "The information on the global expected capital strengthening impact of the SREP 2018 assessments for all institutions from January 1, 2019 onwards, is also a useful input to the discussions on the level of macro-economic buffers."

### Communication to Banks — Individual Decisions
- Observations on individual decision communications:
  - The manual does not contain an example of a letter with the decision on capital and liquidity.
  - An example of such communication can be found in the EBA draft guidelines.
  - Typical content of such a communication:
    - The amount of the Pillar2 capital add-on (typically mentioned) but not the specific method of internal benchmarking or the anchor points used.
    - A brief overview of the main findings during the review process.
    - A motivation for the individual risk assessments that have led to scores 3 and 4.
    - A formal communication of the global score.
  - The letter provides an overview of supervisory measures based on the SREP and indicates other supervisory expectations, expressing the supervisor's view on the quality of risk management and control and indicating supervisory actions in the short and medium term.

### Communication to Banks — Global Communication on Process, Outcome, and Methodology
- Legal disclosure obligation and BNB position:
  - Pursuant to Article 102 of the LCI the BNB has the legal obligation to disclose the general criteria and methodologies used in the SREP process.
  - The BNB considers this provision fulfilled, stating its supervisory approach and methodologies are fully compliant with the guidelines issued by the EBA.
- Trend toward openness:
  - Increasingly, authorities offer some openness on the most common findings and assessments during the SREP process, on the global outcome of the risk scores and the additional requirements to capital and liquidity, and on the main supervisory expectations for the near future.
- Communication opportunities:
  - The yearly approval by the GC of the BSD action plan offers an opportunity to communicate on:
    - The overall outcome of the review process,
    - The main findings and scores,
    - The most important supervisory expectations and global supervisory priorities.
  - Extracts from the information to the GC can be made public or used in relations with other authorities.

*Source: 1bgrea2019003 - 62. The GC also approves BNB regulations and guidelines, as well as the manuals of the*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2019/1bgrea2019003.pdf_