Update on Basel III Implementation ahead of final rules release in January 2022

Ability to use the Internal Models Method (IMM) for measuring Counterparty Credit Risk (CCR) exposure

Chapter 1, paragraph 35 states that the IMM cannot be used directly or indirectly in the calculation of the output floor.

Stakeholders indicated that the investment benefit of an IT solution for institutions approved to use the IMM, but who are currently lacking systems to support the Advanced Credit Valuation Adjustment approach (A-CVA), would be temporary in nature given the need to be replaced by a revised approach in 2024.

Stakeholders requested permission to use IMM exposures at default (EADs) and maturities in the calculation of the Standardized Credit Valuation Adjustment approach (S-CVA), and to continue to use the A-CVA for purposes of calculating the Basel III output floor in 2023.

For the purposes of calculating the Basel III output floor in 2023, institutions may use IMM EADs and maturities in the S-CVA calculation and may continue to use the A-CVA. These treatments will remain in place until the revised CVA framework is implemented in Q1 2024. The investment benefit is not regarded as sufficient to justify new systems investments given the limited (less than one year) implementation delay of the revised CVA framework.

Proposed CET1 deduction of prepaid portfolio insurance

Stakeholders requested that OSFI maintain the current 100% risk-weight applicable to Other Assets.

OSFI is maintaining the 100% risk-weight for prepaid portfolio insurance but is introducing new amortization expectations.

Internal loss data requirements

Chapter 3, paragraph 27(a) requires that institutions must have ten years of high-quality internal loss data to meet the minimum loss data standards.

Stakeholders requested the ability to use five years of internal loss data on a transitional basis.

OSFI is maintaining the requirement for ten years of high-quality internal loss data to meet its minimum standards in the CAR Guideline. OSFI regards ten years of data as more prudent and more likely to be reflective of a full cycle.

Treatment of “timing losses”— operational risk events that result in the temporary distortion of an institution’s financial accounts (e.g. revenue overstatement, accounting errors and mark-to-market errors)

Stakeholders requested a number of adjustments or clarifications with respect to the treatment of timing losses including:

• clarification of timing loss definition;
• ability to use prior overstatement of revenues as a recovery against these losses; and
• higher materiality threshold than $30K for timing losses.

OSFI is maintaining the requirement that prior overstatement of revenues cannot be used as a recovery against timing losses. The following changes will be incorporated into the text:

• Removal of “and give rise to legal risk” from definition of timing losses.
• Institutions may set a threshold higher than $30K for timing losses that are accounting errors and don’t involve payments to third parties or mark-to-market valuation errors. The threshold for timing losses that are accounting errors must be below the level used by external auditors when determining summary of material misstatements.

Methodology to address merged or acquired businesses that do not have ten years of high-quality loss data

Stakeholders noted that the proposed requirement to estimate the loss data for the merged or acquired business (Chapter 3, paragraph 40) may pose implementation challenges.

OSFI will incorporate changes to the CAR and Basel Capital Adequacy Reporting (BCAR) instructions to address implementation challenges. As a result, institutions:

• may use 125% of previous year’s Adjusted Gross Income as a proxy for BI for merged or acquired businesses;
• may use the Internal Loss Method (ILM) of the institution for the previous quarter to determine the methodology to use for estimating loss data; and,
• are not required to make any pre-acquisition adjustments to BI or loss data for asset purchases.

Threshold for percentage of institution that must have ten years of actual loss data to meet loss data standards

Stakeholders requested that OSFI reconsider the 5% threshold (Chapter 3, paragraph 27(d)) for parts of the institution that may use any estimated loss data, above which would lead to an institution’s ILM being floored at one (1).

The following changes will be made to the text pertaining to this item:

• Threshold has been increased to 10%.
• Threshold is now calculated as % of total loss data that is estimated (i.e. calculation now excludes any years where actual loss data is available).
• Ability to be above the 10% threshold on a temporary basis without automatic ILM adjustment. Institutions must come below the threshold in a timely manner.

Loss events due to uncollected revenue

Stakeholders requested that OSFI exclude loss events stemming from uncollected revenue from the loss data because institutions cannot use the accounting date to determine which period the losses should be included in.

OSFI will include a clarification in the text such that losses from uncollected revenue that can be quantified based on the contractual obligations of the institution’s client or customer must be included in the loss data. Institutions can use either the date in which the revenue should have been collected, or the date in which the decision was made not to collect the revenue, instead of the accounting date, to determine the period in which these losses should be included.

Coefficient for Simplified Standardized Approach (SSA)

Stakeholders requested that OSFI change the coefficient from 15% to 12%.

OSFI is maintaining the 15% coefficient as further analysis has confirmed that the Business Indicator (used in the SA) is generally greater than or equal to Adjusted Gross Income (used in the SSA), and the amount of the difference between these measures varies greatly between institutions. It is therefore appropriate to have a higher coefficient under the SSA than the SA, and OSFI believes that 15% is an appropriate level, as it is consistent with both the coefficient used in the current Basic Indicator Approach used by most SMSBs, and the same as the marginal coefficient under the SA for BI over $1.5 billion.

Treatment of uninsured residential real estate with loan-to-values (LTVs) in the 70%-80% band

Paragraphs 93 to 100 of Chapter 4 outline that institutions should assign a 30% risk weight to general residential real estate (GRRE) exposures with an LTV between 60% and 80%, in line with the international Basel III reform package rules. Further, the RW for income producing residential real estate (IPRRE) exposures in the 60% to 80% LTV band is 45%, as risk weights for IPRRE exposures are calibrated at roughly 50% higher than GRRE exposures for a given LTV.

The revisions to the risk weights noted for this item were not included in OSFI’s March 2020 public consultation draft CAR Guideline.

OSFI will introduce a new 70%-80% LTV bucket in order to preserve the risk sensitivity of the credit risk framework, and to better reflect the structure of the Canadian housing market. For GRRE exposures in this LTV bucket, the current 35% risk weight will be maintained, whereas for IPRRE exposures, a 50% risk weight will be implemented.

Risk weight multiplier to certain exposures with currency mismatch

Chapter 4, paragraph 120 outlines that institutions should apply a 1.5 times multiplier to the applicable risk weights for unhedged retail and residential real estate exposures where the lending currency differs from the currency of the borrower’s source of income.

Stakeholders noted that such exposures are immaterial for many institutions, and institutions’ current systems do not capture the currency of the borrower’s source of income.

Application of the 1.5 multiplier for currency mismatch will be limited to residential real estate exposures. The 1.5 multiplier will need to be applied to all applicable exposures with currency mismatch upon implementation of the new rules in institutions’ fiscal Q2 2023.

Maximum aggregate exposure to retail borrowers

The maximum aggregate exposure to a retail borrower, including all small business loans guaranteed by that borrower, must be less than CAD$1.5 million. This applies to exposures under both the standardized approach (SA) and internal ratings based (IRB) approach.

Stakeholders noted that certain institutions’ systems do not aggregate exposures by the same individuals, since the breaching of these thresholds is quite rare and doing so would be quite onerous in certain cases.

Institutions will be required to demonstrate to OSFI that the amount of exposures that breach the retail thresholds are immaterial on at least an annual basis. If the threshold is breached, a discussion with OSFI will be triggered.

Prepayment rates used in the calculation of Tranche Maturity

OSFI introduced a new methodology for applying a prepayment rate greater than zero in the calculation of Tranche Maturity in Chapter 6, paragraph 26.

Stakeholders noted that:

• the methodology may be interpreted to give conservative results if calculated on a vintage basis;
• it was unclear whether unavailable approaches may be excluded from the calculation (or whether they should be assumed to be zero); and,
• data from previous transactions from the same originator are more relevant and suitable than data from all other transactions of the same asset class in the same country.

OSFI agrees with the stakeholder comments and will modify the paragraph related to the calculation of Tranche Maturity:

• to provide an alternative vintage-based calculation;
• to clarify that when an approach is not available, it would not be considered in the calculation of prepayment rates; and
• to require average data from the same originator to be used rather than average data from the same country, if five years of originator data are available.

Specific Right-Way Risk

OSFI proposed updates to the treatment of specific right-way risk in Chapter 7, paragraphs 66 to 69.

Stakeholders requested early adoption immediately upon issuance of the final rules rather than waiting for 2023 implementation.

OSFI’s requirements for specific right-way risk will be implemented as planned in Q2 2023. There is not appropriate justification to early adopt these particular rules, while not doing so for any other rules.

Supervisory Risk Weights

Chapter 8 outlines the ability for AIRB banks to use internal ratings for counterparties which do not have agency ratings. It also provides increased granularity of risk weights for financial counterparties.

Stakeholders noted that:

• guidance already afforded the use of internal ratings on a case by case basis with OSFI approval. In practice, receiving such approvals on a case by case basis would be burdensome for industry and OSFI.
• increased granularity in risk weights for financial counterparties will improve the representation of underlying CVA risk. There are several public pension plans using derivatives as part of their liability driven strategies which would be impacted.

Regarding item (i), OSFI agrees with the stakeholder comments and will modify Chapter 8 text to permit the use of previously approved internal ratings for IRB and internal ratings mappings to external ratings, without needing additional OSFI approval.

Regarding item (ii), OSFI will maintain the existing risk weights for financial counterparties. OSFI will, however, consider collecting data with the additional granularity for financial counterparties as part of future regulatory return revisions.

Boundary – Arms’ length transactions

OSFI permits a limited exemption to the boundary restrictions for arm’s length transactions related to certain liquid assets managed by institutions’ Treasury areas.

Stakeholders requested additional exceptions for arm’s length transactions between the trading book and banking book. OSFI’s guidance already made an exception for CAD securities qualifying as Level 1 and Level 2A high quality liquid assets (HQLA), to support market liquidity in Canada where the number of broker dealers is limited. However, stakeholders requested the exception be extended to:

• foreign currency Level 1 and 2A HQLA; and,
• Bankers Acceptances (BAs).

OSFI will permit additional exemptions from the boundary restrictions for the following:

• Non CAD-denominated Level 1 and Level 2A HQLA issued by Canadian entities, to support market liquidity in Canadian issued debt, regardless of the currency; and,
• Primary issuance of the institution’s own-name BAs purchased by Treasury from its dealer, to help manage wholesale funding limits and short term liquidity needs.

Policy for Internal Risk Transfers (IRTs) and grandfathering of IRTs

Chapter 9 outlines that internal risk transfers are only allowed under the banking book / trading book boundary restriction where there is an exact match.

Stakeholders requested clarity and flexibility on the exact match definition for multiple transactions and the application of the Residual Risk Adds-On (RRAO).

Stakeholders also requested that IRTs executed prior to the implementation of the revised market risk rules be permitted to be grandfathered.

Additional clarity will be included in Chapter 9 for the exact match definition for multiple transactions and the application of the RRAO charge.

OSFI will permit grandfathering of IRTs for interest rate risk transactions that have been executed prior to the implementation of the revised market risk rules in Q1 2024.

Internal Models Approach (IMA) coverage threshold

OSFI proposed a requirement that institutions applying to use internal models to determine market risk capital requirements meet an internal models coverage threshold of 80% initially, and of at least 70% on an on-going basis (Chapter 9, paragraph 266).

Stakeholders requested that OSFI lower the coverage requirement significantly to preserve incentives to adopt the IMA and to reduce uncertainty for IMA approval status and initial application. A lower threshold would also align better with international standards.

A revised minimum application coverage threshold of 60% at initial application and a minimum level of 50% thereafter was proposed by stakeholders.

Institutions applying for the use of internal models will be required to meet an internal models coverage threshold of 50% at all times. To simplify the framework, OSFI will not implement a separate higher application threshold but will need to be satisfied through the approval process that the ongoing 50% threshold can be maintained.

Default Risk Charge (DRC) mismatch between maturities of actively managed derivatives and their hedges

Chapter 9, paragraph 229 outlines the DRC treatment and governance related to a large cap equity hedging a total return swap. Any mismatch applied between long and short positions is capped at 40 days for the purpose of the DRC under both the Standardized Approach (SA) and the IMA.

Stakeholders requested that the treatment related to the maturity mismatch cap at 40 days be extended to also apply to a bond forward hedging a Level 1 HQLA. Bond forwards have grown in popularity and are now important hedging instruments for asset managers, pension plans, small banks, and provincial treasuries. This change would be consistent with other amendments made in Chapter 9 concerning the treatment of Level 1 HQLA.

The treatment will be extended to a bond forward that is hedging a Level 1 HQLA as defined in Chapter 2 of OSFI’s LAR Guideline. The mismatch applied between long and short positions will be capped at 40 days under both the SA and the IMA.

Treatment of index instruments and multi underlying options

Chapter 9, paragraph 145 requires that indices not meeting the requirements of major benchmarks must be looked through.

Stakeholders requested that OSFI allow a no look-through option for certain indices that cannot be decomposed, consistent with the current treatment for Equity Investment Funds (EIFs). Institutions generally trade and manage the risks associated with EIFs and indices in a consistent manner and the capital treatment should therefore align. Look-through would continue to be the preferred option whenever possible.

In case a look-through approach for such indices is not possible, institutions may treat these indices in the same manner as prescribed for EIFs that cannot be looked-through. The alternative to a look through approach for affected indices still results in a conservative capital charge in the “other bucket” with a 70% risk weight, but cost and operational burden may be reduced.

Exposure Measurement of Cash Pooling Arrangements

OSFI’s leverage ratio rules include an exposure measure for cash pooling arrangements.

Stakeholders indicated that cash pooling is not material for many institutions; however, adoption of this new leverage requirement is costly given institutions must have a methodology to track whether cash pooling is in place in order to prove they are meeting the requirement.

In the event that the cash pooling requirements are deemed to be disproportionately operationally burdensome, institutions may reflect the individual balances of the participating customer accounts separately in the leverage ratio exposure measure consistent with paragraph 33.

Frequency and timeliness of reporting

For non-DSIBs, OSFI requires that they have the capacity to produce a Net Cumulative Cash Flown (NCCF) calculation on a weekly basis and submit it to OSFI within three business days in stress conditions.

Stakeholders requested further guidance on the nature of the information that is critical to achieve sufficient accuracy, given challenges associated with meeting this requirement.

The requirement will not change. OSFI acknowledges in its guidance that intra-period reporting may not be subject to the same rigour and control infrastructure as that of month-end and quarter-end reporting. Institutions should nonetheless have a framework in place that outlines their process to report intra-period LCRs and NCCFs.

Application of the NSFR

OSFI requires Category I SMSBs with significant reliance on wholesale funding (defined as funding 40% or more of their total on-balance sheet assets with wholesale funding sources) to comply with the NSFR nine months after reaching or exceeding the 40% wholesale funding reliance threshold.

Stakeholders requested that OSFI extend the timeline for implementation of the NSFR once the wholesale funding threshold is exceeded.

After considering the request, OSFI still believes that nine months is a sufficient lead time especially given expectations that funding plans/strategies be in place several months in advance. Due to the nature of the calculation of the threshold, institutions will have sufficient lead time to implement the NSFR.

Treatment of undrawn committed credit facilities to non-financial corporates

Undrawn committed credit facilities were segmented using three criteria:

1. presence of a pricing incentive to draw under stress conditions;
2. credit rating (using external ratings); and,
3. the lender being the primary provider of banking services to the counterparty.

Outflow rates ranged between 10% and 30% for undrawn committed credit facilities.

Stakeholders requested that OSFI reconsider the segmentation criteria and that OSFI limit the NCCF outflow rates for undrawn credit/liquidity facilities to those prescribed in the LCR.

OSFI has modified the segmentation criteria and corresponding outflow rates. The segmentation criteria now includes a differentiation of counterparties – i.e., commercial vs corporates, where commercial clients will be defined as clients where the group has total consolidated annual revenue less than CAD$750 million, as reported in the financial statements calculated on an average of the three prior years (consistent with CAR Guideline, Chapter 5).  Revised outflow rates to be applied in the Comprehensive NCCF include:

• Commercial clients:
  • Lender is the primary provider of banking services: 5%
  • Otherwise: 10%
• Corporate client:
  • Lender is the primary provider of banking services: 5%
  • Otherwise: 15%

Under the Streamlined NCCF, all undrawn amounts of committed credit facilities to non-financial corporates will be subject to a 10% outflow rate.

Integration of forecasts over the projection period – expected lending and net non-interest expenses

OSFI requires that institutions forecast expected lending and net non-interest expenses over the projection period, as well as backtest projections to improve accuracy of estimates over time.

Stakeholders requested that OSFI standardize methodologies for the NCCF calculation instead of utilizing institution-specific forecasts.

Further, stakeholders requested that OSFI remove net non-interest expenses from the NCCF calculation given their expected immateriality.

OSFI will standardize the methodology for expected lending, to be a function of maturities:

• For mortgages, the current methodology will remain unchanged, i.e. 0% of balances at maturity will be recognized as inflows.
• For loans to businesses and governments, institutions can recognize only 50% of balances at maturity as inflows.

Nonetheless, forecasted expected lending will be collected as a memo item for all DTIs on the NCCF returns.

For all DTIs, net non-interest expenses will be removed from NCCF calculation; however, this information will be collected as a memo item for SMSBs only.

Treatment of liquidity facilities backstopping ABCP

In addition to LCR requirements, under the NCCF, OSFI requires institutions to hold liquidity for amounts of outstanding asset-backed commercial paper (ABCP) issued by the special purpose entity (SPE) that matures after 30 days as well as the unutilized capacity liquidity facilities provided to SPEs that can be drawn only after 30 days.

Stakeholders requested that OSFI reconsider the treatment of liquidity facilities backstopping ABCP, noting that the products and facilities were structured with the LCR requirement in mind.

The NCCF will only assign outflow rates to paper maturing in the first 30 days, as well as to the unutilized capacity of facilities that can be drawn within 30 days.

Calibration of the NCCF metric

The overall calibration of the NCCF metric, and the requirement to meet a supervisory-communicated, institution-specific survival horizon.

Stakeholders provided several requests, specifically that OSFI:

• decrease the severity of the NCCF metric given the expected material impact of the public consultation proposal on NCCF survival horizon;
• consider other sources of inflows to offset material outflows;
• review the need for supervisory-communicated, institution-specific NCCF levels; and,
• review the appropriateness of current supervisory-communicated, institution-specific NCCF levels.

In addition to areas of recalibration noted above, applicable run-off rates will be adjusted downwards for all retail and small business deposits (see Annex 2).

No additional sources of inflows will be recognized.

Supervisory-communicated, institution-specific NCCF levels will remain and continue to be supervisory expectations.

Where supervisory-communicated, institution-specific NCCF levels are currently applied to institutions, OSFI will review these levels prior to 2023 implementation.

Category III Segmentation Criteria

Stakeholders requested that OSFI include more specific criteria that could permit an institution to be placed into Category II when Category III requirements may not be appropriate for institution’s activities.

OSFI incorporated additions to the text to specify that an SMSB will be in Category II if it meets any of the following criteria:

1. the institution has greater than $100 million in total loans;
2. it has interest rate or foreign exchange derivatives with a combined notional amount greater than 100% of total capital;
3. it has any other type of derivative exposure;
4. its exposure to certain other off-balance sheet items is greater than 100% of total capital.

An SMSB that otherwise meets the criteria for Category III will be permitted to make a one-time request to move to Category II prior to implementation if it believes the Category III requirements are not appropriate for its activities.

Template ENC: Asset Encumbrance

In Template ENC, stakeholders requested that OSFI allow D-SIBs to include off-balance figures as well on-balance sheet assets.

Stakeholders recommended that OSFI permit signposting to current asset encumbrance disclosures developed through the Enhanced Disclosure Task Force (EDTF) which typically reside in the Liquidity Risk sections of banks’ MD&As, and which offer enhanced detail over the Pillar III disclosure.

In the final Pillar 3 Disclosure Guideline for D-SIBs, Template ENC will accommodate off-balance sheet amounts.

D-SIBs may signpost to their current asset encumbrance disclosures developed through EDTF in place of Template ENC.

Table REMA and Templates REM1-3: Remuneration

Stakeholders requested that OSFI consider permitting signposting to the shareholder proxy circular in place of disclosing Table REMA and Templates REM1-3. The timing of the shareholder proxy circular’s release, subsequent to the reporting of the financial fiscal year ending October 31, would provide users with more relevant, clear and meaningful information.

D-SIBs will be expected to disclose Table REMA and Templates REM1-3 to facilitate comprehensive and comparable disclosure among the D-SIBs.

Regarding timing of disclosure, year-end figures are expected to be reported at Q2.

Regarding location of disclosure, OSFI will allow for flexibility on the location of Remuneration disclosures (either Q2 Pillar 3 Report or Shareholder Proxy Circular, with reference (link) from Q2 Pillar 3 Report).

Eight existing disclosure templates required to be updated to align with 2023 CAR Guideline implementation

Stakeholders requested that OSFI provide the D-SIBs with mock-ups of the eight existing disclosure templates reflecting updated rows and columns to ensure clarity on 2023 disclosure requirements.

The updated mocked-up versions of the eight templates requested can be found here.

Referencing OSFI’s “Financial Data for Banks” website

Stakeholders requested that OSFI permit Category II SMSBs to reference OSFI’s “Financial Data for Banks” website in place of completing Templates KM1 and LR2.

Although some of the information in Template KM1 can be found in other disclosures, OSFI expects SMSBs to disclose Template KM1 as it comprises a summary of key metrics all in one place, facilitating clearer disclosure for users. Much of the information in Template LR2 cannot be found in other disclosures. Accordingly, OSFI expects SMSBs to disclose Template LR2.

Frequency of existing disclosures increasing from annual to quarterly

For institutions that only provide Pillar 3 disclosures on an annual basis, stakeholders requested the option to disclose the templates for Key metrics (KM1), Composition of Capital (CC1) and Leverage Ratio (LR2) only on an annual basis.

OSFI expects SMSBs to implement quarterly reporting for Template KM1 and to follow the new quarterly frequency of reporting for Templates CC1 and LR2 because metrics can change significantly during the year and users deserve transparency of the measures.

• Where criteria outlined in Chapter 2, paragraph 59 are met
• Where criteria outlined in Chapter 2, paragraph 59 are not met

• 0.50%
• 1.00%

• 0.75%
• 0.75%

• Term deposits from non-financial corporates, sovereigns, central banks, multilateral development banks, and PSE customers
• All other non-small business customers

• 40% at maturity
• 100% at maturity

• Where the deposit is not fully covered by deposit insurance
• Where the deposit is fully covered by deposit insurance and:
  • Jurisdiction where the deposit is located permits a 3% run-off factor
  • Jurisdiction where the deposit is located does not permit a 3% run-off factor

• 2.5%
• 0.75%
• 1.25%

• 5%
• 3%
• 5%

Non-financial corporates, sovereigns, central banks, PSEs and MDBs – non-operational deposits:

• Where the deposit is not covered by an effective deposit insurance scheme or public guarantee
• Where the deposits is covered by an effective deposit insurance scheme or public guarantee

• 3%
• 3%

• 10%
• 5%

All other counterparties (including other FIs and other legal entities) – non-operational deposits

• Facilities extended to transactors
• Other uncommitted facilities
• Other committed facilities

• 0%
• 2%
• 5%

• Streamlined NCCF;
• Comprehensive NCCF, where:
• The customer is a corporate client:
  • Absence of an operational relationship
  • Presence of an operational relationship
• The customer is a commercial client:
  • Absence of an operational relationship
  • Presence of an operational relationship

• 10%
• 15%
• 5%
• 10%
• 5%

• Sovereigns, central banks, PSEs and multilateral development banks
• Deposit-taking institutions subject to prudential supervision
• Other financial institutions including securities firms, insurance companies, fiduciaries and beneficiaries
• Other legal entities (including SPEs, conduits and special purpose vehicles, and other entities not included in prior categories)

• 10%
• 40%
• 40%
• 100%

Uncommitted credit facilities to non-retail and non-small business customers

• Non-financial corporates, sovereigns and central banks, PSEs, and multilateral development banks
• Deposit-taking institutions subject to prudential supervision
• Other financial institutions including securities firms, insurance companies, fiduciaries, and beneficiaries
• For liquidity facilities backstopping ABCPs: for maturities within 30 days and unutilized capacity that can be drawn within 30 days;
• Other legal entities (including SPEs, conduits and special purpose vehicles, and other entities not included in prior categories)

• 30%
• 40%
• 100%
• 100%
• 100%

Uncommitted liquidity facilities to non-retail and non-small business customers

Trade finance instruments

Other guarantees and letters of credit unrelated to trade finance obligations