Office of the Superintendent of Financial Institutions
1. What are Small and Medium Sized Deposit-Taking Institutions' (SMSBs) capital buffers intended to guard against? What are the implications of using these buffers and how does OSFI monitor their use?
Buffers are built-up during normal times to provide an institution with additional flexibility in times of stress. Capital buffers can be used to absorb unexpected losses while continuing to provide financial services. As part of on-going supervisory activities, OSFI has frequent discussions with SMSBs and monitors their capital and liquidity positions as well as their exposures. SMSBs that plan to use Pillar II buffers by operating below their internal capital targets should have prior discussions with their Lead Supervisor, including a credible plan that demonstrates how and when it intends to restore its buffers. OSFI's review of planned buffer usage will consider the specific circumstances of each institution, including whether a SMSB's expected use of the capacity is prudent, and capital conservation actions have been incorporated as appropriate.
2. What sorts of risks or factors should an SMSB consider when establishing its internal capital targets?
A SMSB's internal target is the sum of the minimum requirements, Pillar I and Pillar II buffers. OSFI expects SMSBs to set their internal targets at levels that are adequate to support the nature and level of their risks. This includes consideration of material risks unique to an SMSB's operations, risks that are not sufficiently captured in Pillar I capital requirements and a SMSB's stress testing results. For example, a SMSB should consider its exposure to various aspects of credit risk (including concentrations), market risk, operational risk, and interest rate risk in the banking book, amongst others.
3. How will OSFI define senior management for purposes of the capital distribution constraints required for a breach of the capital conservation buffer?
OSFI will rely on each institution to determine what the most appropriate definition is for their own institution. However, OSFI would refer Canadian incorporated institutions to the definition of Senior Management included in the OSFI Corporate Governance Guideline for further clarifications.
1. Is paragraph 66 of CAR Chapter 2 applicable only to credit protection bought by way of insurance similar to mortgage insurance? Does this paragraph exclude credit derivatives?
Mortgage insurance purchased from private mortgage insurers in Canada is subject to the rules in paragraph 274 of Chapter 4 or 147 of Chapter 5 of the CAR Guideline. All other forms of credit protection, including credit derivatives, are included in the scope of paragraph 66 of CAR Chapter 2.
2. How does OSFI define a “regular” or “special / irregular” dividend?
In contrast with a regular dividend – defined as a cash payment made by an institution to its shareholders at specific times of the year, following a consistent payment pattern – “special” or “irregular” dividends should be non-recurring, limited to a specific business objective and not for distributing capital to a broad group of shareholders.
3. Are institutions required to notify OSFI of their capital distribution plans?
As part of OSFI’s ongoing supervisory dialogue with institutions, it is expected that institutions maintain and update capital plans that reflect their planned capital distributions. The current requirement prescribed in the legislation for 15 days notification prior to the payment of any dividend (i.e. the dividend has already been declared) remains in place.
1. Should operational errors that result in a gain be included in the loss data set under the Standardized Approach for Operational Risk?
Under paragraph 27 (e) of CAR Chapter 3, only operational events that each result in a net loss greater than CAD $30,000 over the previous ten years should be included in the loss data set. Therefore, loss events that result in a gain should not be included. However, institutions should report and collect information on these events for internal risk management purposes.
2. Are institutions required to adjust timing losses in their historical operational loss dataset so that it aligns with the more detailed definition and thresholds set out in paragraph 31 (e) of CAR Chapter 3?
For timing losses prior to fiscal Q2 2023, institutions are not required to make any adjustments to their loss data. As such, institutions may use the same definition as they used under the Advanced Measurement Approach (AMA), if applicable, consistent with their own policies at the time the losses occurred. Starting in Q2 2023, loss data must include all timing losses with the definition and thresholds consistent with paragraph 31 (e) of CAR Chapter 3 (e.g., no netting of previous revenue overstatement as recoveries against timing losses).
3. What is a
financial accounting period referred to in paragraph 31 of CAR Chapter 3?
financial accounting period refers to one fiscal quarter.
4. Please clarify if any loss should be reported in the loss data set under the Standardized Approach for Operational Risk in the following scenario: Funds were inadvertently wired to an incorrect third-party on the last day of the fiscal quarter. On the same day, the error was caught, and the funds were re-sent to the correct third-party with the offsetting amount being posted to a suspense account of the General Ledger pending the return of funds. A request was made to return the funds from the initial third-party, and the funds were returned on the following day (i.e., in the next fiscal quarter) by the third-party who had received funds erroneously.
Under paragraph 31(d) of CAR Chapter 3, institutions should include in the loss data set “losses stemming from operational risk events with a definitive financial impact, which are temporarily booked in transitory and/or suspense accounts and are not yet reflected in the P&L ("pending losses").” If, at the end of the fiscal quarter, the institution has reason to expect that the funds will be returned in the following fiscal quarter, then there is no definitive financial impact. Therefore, in this example no loss would be reported even if there is a balance in the institution’s suspense account.
5. What are some examples of uncollected revenue and how are these losses different from other types of operational losses in the Standardized Approach for Operational Risk?
Paragraph 31 of CAR Chapter 3 details the types of losses that must be included in the loss data set under the Standardized Approach for Operational Risk which generally result in a negative impact recorded in an institution’s profit and loss (P&L) statement. The one exception is losses from uncollected revenue in paragraph 31(b)(iii), which do not impact the P&L. Uncollected revenue must be included as a loss when it can be quantified based on the contractual obligations of the institution's client or customer. Examples of loss events resulting in uncollected revenue include situations where:
Example 1: An institution decides not to charge a client the full fee amount in compensation for an error caused by the institution. The amount of compensation (i.e., the reduction in the fees charged to the client) should be reported as a loss if the net amount of the loss is greater than $30,000.
Example 2: An institution charges a lower interest rate to a customer than contractually required due to an operational error. The institution decides not to claim the additional amount owed by the customer, which is considered uncollected revenue. For example, the contractual interest rate was 10% and the institution charged the client 7% due to an operational error. The institution decided not to ask the client to repay the difference in interest payments of 3%. This amount should be reported as a loss if the net amount of the loss is greater than CAD $30,000.
Example 3: An institution overpaid sales commission to its sales team due to an internal accounting error. Management decided not to claw back the sales commission. The amount of overpaid sales commission due to the error should be reported as a loss if the net amount of the loss is greater than CAD $30,000.
6. Can gains and losses related to the same loss event be netted against one another when determining whether the loss meets the CAD $30,000 threshold for inclusion in the loss data set, as specified in paragraph 31(e) of the CAR Guideline, Chapter 3?
If a loss event is not considered as a timing loss per paragraph 31(e) of CAR Chapter 3, gains can be netted against losses if they are related to the same loss event (i.e., have the same root cause) when determining if the amount of the loss event exceeds the CAD $30,000 threshold. For example, if a model error impacted ten trades in a financial quarter and the error resulted in five of the trades having gains totalling $10 million and five of the trades having losses totalling $17 million, the net amount of the operational loss to be reported in that quarter would be $7 million.
On the other hand, if the loss event results in a timing loss, where there is a negative impact to the institution’s profit & loss (P&L) to correct gains in previous financial periods, the amount of the timing loss is determined by calculating the net negative impact to the P&L to correct the previous revenue overstatement or expense understatement. However, for a timing loss, the net amount of the gain reflected in the P&L in fiscal periods prior to the correction of the financial statements cannot be netted against the negative impact to the P&L in the quarter in which the timing loss is reported.
For example, if the same operational error five years ago resulted in revenue being overstated by $20 million in one business unit and revenue being understated by $15 million in another unit over the past five years, the amount of the timing loss would be $5 million. In this case, the net $5 million revenue overstatement over the past five years cannot be reported as a recovery or otherwise netted against the $5 million timing loss.
7. If an accounting error results in a re-statement of an institution’s balance sheet but does not impact the income statement, would this be an accounting error and therefore reported as a timing loss?
Unless there is an impact to the institution’s profit and loss due to the accounting error, there would be no loss to report in the loss data set for capital purposes.
8. Under footnote 21 in paragraph 31(e) of CAR Chapter 3, institutions can set the threshold for timing losses that are accounting errors at a level below the external auditors’ materiality level. Such a threshold is usually determined based on a certain percentage of the institution’s profit before tax and thus will vary from year to year. Are institutions supposed to adjust the reportable threshold for accounting errors year over year?
The level used by the institution's external auditor for determining the summary of material misstatements is the maximum amount that institutions can use for their threshold for the inclusion of timing losses that are accounting errors. Institutions should set the threshold below the level set by the external auditors and low enough such that the threshold will not have to be adjusted from one year to the next.
9. If the total loss amount from a loss event is known upfront but the losses are amortized over a number of periods in the institution’s financial statements, how should the loss be reported for the purposes the loss data set under the Standardized Approach for Operational Risk?
Under paragraph 33 of CAR Chapter 3, institutions must use the date of accounting for building the loss data set. As such, losses should be reported in the same financial accounting period (i.e., fiscal quarter) in which they are amortized in the institution’s profit and loss (P&L) statement. In some cases, the cumulative losses recognized in the financial statements may be less than CAD $30,000 due to amortization. In these instances, the loss should be introduced into the loss data set in the first fiscal period in which the cumulative losses reach CAD $30,000. At that time, the losses should be reflected in the financial accounting period in which the loss was originally recognized.
1. If a covered bond does not meet the requirements in paragraphs 48 to 51 of CAR Chapter 4, how would the exposure be classified?
Consistent with paragraph 55 of CAR Chapter 4, covered bonds that do not meet the criteria set out in paragraphs 48 to 51 should be treated as exposures to the issuing institution and should be reported in the BCAR schedule corresponding to the issuing institution.
2. Paragraph 62 of CAR Chapter 4 provides two options for risk weighting unrated bank exposures: 1) separate risk weights for “investment grade” and “non-investment grade” exposures, or 2) a flat 100% risk weight if the institution chooses not to identify its unrated corporate exposures as "investment grade" and "non-investment grade" according to paragraph 63 of CAR Chapter 4. Can an IRB institution choose to apply the differentiated (“investment grade”) risk weights on a pre-floor basis, while applying a flat 100% risk weight for the same exposures for the purposes of the output floor?
No. An institution may not choose to apply a flat 100% risk weight for unrated corporate exposures for the output floor, while identifying “investment grade” exposures for pre-floor risk-weighted assets (RWA). Paragraph 62 allows institutions to choose to identify whether their unrated corporate exposures meet the “investment grade” definition. This is a single decision – institutions may not choose to identify only a portion of their unrated corporate exposures. Similarly, an institution that chooses to identify their “investment grade” unrated exposures must do so consistently for both pre-floor RWA and for the purposes of the output floor.
3. Regarding the “accrued interest over previous 12 months” for credit cards in paragraph 84 of CAR Chapter 4, is this interest actually charged (and not accrued) to clients for late payments or payments less than the full amount due? Is the $50 threshold the sum of all interest charged over a 12-month period or the interest that exceeds $50 for any one-month invoicing cycle?
Interest is understood to have accrued once there is an amount that can be charged to clients (e.g., for late payments or payments less than the full amount due). Utilization against the $50 threshold should consider the sum of all interest charged over a 12-month period.
4. Regarding the “no drawdown over the previous 12 months” for overdraft/lines of credit in paragraph 84 of CAR Chapter 4, please confirm whether this refers to: any act of drawing down a line over the previous 12 months, or the existence of a drawn amount over the previous 12 months, even if the actual draw occurred more than 12 months ago but has not yet been repaid. In addition, would this apply to customers who draw on their overdraft / line of credit but then repay them in full prior to the end of each month (or the next statement date)?
Drawdown for overdrafts and lines of credit over the previous 12 months refers to the existence of any drawn amount during the previous 12-month period, regardless of when the initial draw occurred or when it was repaid.
5. Does paragraph 89 (iii) of CAR Chapter 4 exclude all mortgages with a Loan-to-Value (LTV) above 80% from the real estate asset class?
Mortgages issued in Canada with LTVs above 80% are subject to mortgage insurance and should be treated according to paragraphs 272 to 274 of CAR Chapter 4. Mortgages issued outside Canada and not subject to mandatory mortgage insurance are subject to the applicable table for real-estate (i.e., General or Income-producing Residential Real Estate, or General or Income-producing Commercial Real Estate).
6. How is “primary residence of the borrower” defined? Can a borrower have more than one primary residence (e.g., a cottage which they occupy)?
For the purposes of paragraph 89 (i) of CAR Chapter 4, the primary residence of the borrower for capital purposes is the borrower’s principal residence as applicable under tax rules. Only one property can be designated as a primary residence for each borrower. The same interpretation of a primary residence is to be applied for the purposes of paragraph 102 of CAR Chapter 4.
7. For the purposes of calculating the Loan-to-Value (LTV) in paragraph 92 of CAR Chapter 4, what is included in the “outstanding loan amounts” and “undrawn committed amounts” of a mortgage loan?
Outstanding loan amounts refer to the amount still owing to the lender from the borrower. The undrawn committed amount refers to the difference between the authorized amount and the outstanding balance (i.e., notional undrawn amounts, prior to the application of credit conversion factors).
8. Can an institution’s second line of defense fulfill the requirement in paragraph 92 of CAR Chapter 4 for the valuation of real estate exposures to be done independently of the loan decision process?
The intent of the requirement in paragraph 92 – that valuation of a property be assessed independently – is to ensure that the valuation is performed using sufficiently conservative and prudent methods, and in a way consistent with guidance provided in the CAR Guideline and OSFI's Guideline B-20:
Residential Mortgage Insurance Underwriting Practices and Procedures. This validation may be performed internally at the originating institution, provided it is done independently of the institution's mortgage acquisition, loan processing and loan decision process.
9. How should residential real estate exposures that are principal residences be treated under the CAR Guideline and reported under BCAR if they are held by wholesale borrowers?
Real estate exposures held by wholesale borrowers, even those secured by principal residences, are not eligible for the treatment of residential real estate and should be treated as Commercial Real Estate using the treatments set out in section 4.1.12 of CAR Chapter 4.
Such exposures should be reported in BCAR Schedules 40.230 or 40.240 depending on whether or not repayment is materially dependent on cash flows generated by property, respectively (and in BCAR Schedules 50.260 or 50.270 for IRB, if applicable).
10. Regarding paragraph 94 of CAR Chapter 4, is a residential real estate exposure that is more than 90 days past due no longer considered a residential real estate exposure?
Residential real estate exposures that are more than 90 days past due are to be treated according to paragraph 143 of CAR Chapter 4.
11. Paragraph 96 of CAR Chapter 4 states that residential real estate exposures that do not meet OSFI’s expectations related to Guideline B-20, are subject to either the risk weights outlined in Table 11 of CAR Chapter 4 or to a 0.22 correlation (R) factor in paragraph 78 of CAR Chapter 5. Moreover, footnote 44 states that “this treatment applies to all exposures secured by the same residential real estate collateral where one or more of the exposures do not meet OSFI’s expectations related to Guideline B-20.” Does the treatment extend to exposures secured by the same collateral as an exposure from another institution that does not meet OSFI’s expectations?
OSFI does not require institutions to monitor or assess whether other liens on a property from third-party lenders meet OSFI’s expectations related to Guideline B-20. Institutions lending in a junior position relative to a third-party lender will apply a 1.25 multiplier if the exposure is not in the lowest Loan-to-Value (LTV) bucket, as noted in footnote 39 of CAR Chapter 4, paragraph 92.
12. Regarding paragraphs 98 and 108 of CAR Chapter 4, is there a limit to the number of properties under a single obligor that can be treated as general residential real estate?
Consistent with paragraph 101 of CAR Chapter 4, a residential real estate exposure (RRE) can be treated as general RRE (subject to the risk weights in Table 10) as long as less than 50% of the income used by an institution in their determination of the borrower’s ability to service the loan is from cash flows generated by the residential property. Income generated from other real estate properties should not be included in a borrower’s income for this purpose.
13. What can be considered equity for the purposes of Loan-to-Value (LTV) calculations of a land acquisition, development and construction (ADC) exposure?
For the purposes of assessing the LTV of an ADC exposure, institutions may count the following as equity:
Note that land lift should not be reflected in LTV calculations for land acquisition loans.
14. What risk weight should be applied to mezzanine structures in a land acquisition, development and construction (ADC) exposure?
The amount of exposure an institution holds in a subordinated/mezzanine tranche of an ADC loan structure should be risk-weighted at 300%. The same institution cannot hold both the senior and subordinated/mezzanine tranche(s) in the same ADC loan structure.
15. What distinguishes a land acquisition loan from a construction loan?
The underlying property must be zoned for construction in order to be considered a construction loan for capital purposes and to qualify for the preferential risk weights available to construction projects in the ADC treatment set out in section 4.1.13 of CAR chapter 4, provided all other criteria for the preferential risk weights are also met. The preferential treatments for both land acquisition and construction loans are only available for residential real estate projects.
16. Please clarify the treatment of prepaid portfolio insurance (PPI), with examples, including where to apply the assumptions noted in paragraph 164 of CAR Chapter 4 that if not met will result in a CET1 capital deduction.
In paragraph 164, “expected life” refers to the first term of the underlying portfolio-insured mortgage or pool of mortgages assuming no renewals. The 100% risk-weight is only available for capitalized PPI relating to mortgages or NHA MBS with an initial term of five years or less. PPI eligible for this treatment must be amortized over the lesser of five years, or the term of the underlying mortgage(s) or NHA MBS assuming no renewal of the underlying mortgage(s). If the underlying mortgages do not meet these above expectations, this treatment does not apply and the PPI asset is to be deducted from CET1 capital as set out in paragraph 59 of CAR Chapter 2. A couple of illustrative examples are provided below.
Example 1: New portfolio insurance transaction, ‘first term’ of the underlying mortgage pool is seven years. The PPI asset associated with this transaction is not eligible for the 100% risk-weight, and the PPI must be fully deducted from CET1 capital as per paragraph 59 of CAR Chapter 2.
Example 2: New portfolio insurance transaction, ‘first term’ of underlying mortgage pool is four years. The PPI asset associated with this transaction would be eligible for the 100% risk-weight described in Chapter 4. The PPI exposure amount must be amortized for capital purposes on a straight-line basis over the first term of the mortgage pool (i.e., four years). At acquisition, 100% of the PPI exposure would be risk-weighted with the amount eligible for the 100% risk-weight, declining by 25% in each of the following four years. After two years, for example, only the unamortized portion (i.e., 50%) would be eligible for the 100% risk-weight. At the end of the four years, the entire value of the PPI must be amortized for capital purposes. Any differences between the amount capitalized amount remaining for capital purposes versus accounting purposes should be deducted from CET1 capital.
17. Do commercial mortgages insured by Canadian Mortgage and Housing Corporation (CMHC) follow the same treatment as in paragraphs 272 to 274 of CAR Chapter 4?
Paragraphs 272 to 274 of CAR Chapter 4 apply to residential mortgages insured under the
National Housing Act or equivalent provincial mortgage insurance programs. Institutions may still recognize the risk-mitigating effects of mortgage insurance on a commercial mortgage, however, as commercial mortgages would be exclusively insured by CMHC, and would not be subject to a deductible as set out in paragraphs 273-274 pursuant to
Protection of Residential Mortgage or Hypothecary Insurance Act, the protected portion would simply receive a risk weight of 0% in recognition of the fact that obligations incurred by CMHC are legal obligations of the Government of Canada.
1. In paragraph 18 (c) and 20 of CAR Chapter 5, please confirm that the high-volatility commercial real estate (HVCRE) classification is only applicable to commercial real estate project ADC loans in Canada (that is, ADC loans for the purpose of residential real estate projects for rent or sale are excluded – e.g., population covered in paragraphs 112 and 113 of CAR Chapter 4).
Loans financing ADC of any Canadian properties (including residential real estate exposures) where the criteria for a 100% risk weight as defined in Section 4.1.13 of CAR Chapter 4 are not met are considered HVCRE. In addition, any loans financing foreign properties designated as HVCRE by national supervisors in those jurisdictions will also be deemed as HVCRE in the CAR Guideline.
2. Is ‘substantial equity at risk’, as referenced in paragraph 18 (c) and 20 of CAR Chapter 5, defined the same as in paragraph 112 of CAR Chapter 4 (i.e., at least 25% of the real estate’s apprised as-completed value has been contributed by the borrower)?
Yes, the definitions are the same.
3. Assume a wholesale borrower has a loan to finance the construction of an office building
in Canada without substantial equity at risk. In such a case, should this loan be considered as high-volatility commercial real estate (HVCRE)?
Yes, in this case the loan should be considered as HVCRE.
4. Assume a wholesale borrower has a loan to finance the construction of a residential condo building for sale in Canada. As high-volatility commercial real estate (HVCRE) lending is to finance commercial real estate, is this loan in scope for HVCRE lending? If yes, assume this residential condo building has over 50% pre-sale – is this considered to be ‘certain cash flows’ and the loan should be treated as a corporate loan?
Yes, this loan is in scope for the HVCRE classification unless it has at least 50% of total contracts pre-sold. If at least 50% of total contracts are pre-sold, then the loan should be treated as a corporate loan under the IRB approach.
5. Assume a wholesale borrower has a loan to finance land acquisition in Canada. The land is prepared for residential condo construction, but the construction has not started yet.
As high-volatility commercial real estate (HVCRE) lending is to finance commercial real estate, is this loan in scope for HVCRE lending? If yes, according to paragraph 112 of CAR Chapter 4, assuming the land has more than 60% Loan-to-Value (LTV), should this loan be treated as a corporate loan?
Yes, this loan is in scope for the HVCRE classification unless it has an LTV of 60% or below. If the loan has an LTV of 60% or lower, then it should be treated as a corporate loan under the IRB approach.
6. Please confirm that under paragraph 20 of CAR Chapter 5, the requirement to adopt high-volatility commercial real estate (HVCRE) classification for exposures in foreign jurisdictions based on their national supervisor guidance can result in inconsistency from the CAR Guideline HVCRE definition (e.g., equity at risk is 15% and residential projects are included as HVCRE in the United States).
We understand that national supervisors in foreign jurisdiction may apply a different definition for HVCRE loans.
7. Does the reference to the definition of land acquisition, development and construction (ADC) in CAR Chapter 4 apply to Canadian properties or foreign properties? Paragraph 18 of CAR Chapter 5 seems to apply the definition to foreign properties while paragraph 20 is specific to Canadian properties.
As per paragraph 20 of CAR Chapter 5, the only Canadian exposures that are considered to be high-volatility commercial real estate (HVCRE) are exposures to financing ADC of properties where the criteria for a 100% risk weight as defined in Section 4.1.13 of CAR Chapter 4 are not met. Foreign properties where the national supervisor has designated them as HVCRE are also considered HVCRE.
8. Please clarify paragraph 25, bullet 2 of CAR Chapter 5, which refers to Section 4.1.9 of CAR Chapter 4. Should the reference be Section 4.1.10, as one-to-four units is only defined in Section 4.1.10, paragraph 89 (i)?
Yes, the reference will be changed.
9. Please clarify if paragraph 25, bullet 2 of CAR Chapter 5 can include second homes such as cottages that are not rentals. Specifically, that these would not be categorized as non-regulatory retail.
Yes, second homes qualify as regulatory retail under the IRB approach, where the loan is extended to an individual.
10. In the 2023 CAR Guideline, the “Other Retail” asset class has been inadvertently removed from Chapter 5 with the introduction of “Non-Regulatory Retail” asset class. However, the treatment of Other Retail and Non-Regulatory Retail is clear based on the OSFI’s cover letter and past discussions. Chapter 5 does not mention the treatment for Regulatory Retail exposures as defined under paragraph 25. Can the wording in CAR Guideline under Chapter 5 be updated as appropriate?
In addition, the Retail risk weight functions only contain Residential Mortgage, Qualifying Revolving Retail and Non-Regulatory Retail from paragraphs 79 to 81. Which risk weight formula should Regulatory Retail exposures use? For Non-Regulatory Retail exposures subject to the Corporate small-and-medium-sized entities (SME) risk weight function, please confirm that the firm size adjustment for SMEs (correlation factor) should be used with the Corporate RWA formula. Finally, for Non-Regulatory Retail exposures that are subject to the Corporate SME risk weight function, please confirm what LGD floors (unsecured and secured) should be used.
Paragraphs 25 and 70 of CAR Chapter 5 describe the correct treatment for Non-Regulatory Retail exposures of applying the Corporate SME function with the firm size set at $7.5 million. The LGD floors for these exposures should be consistent with LGDs for Non-Retail exposures. Paragraph 81 of CAR Chapter 5 describes the treatment for Other Regulatory Retail exposures, and the label to paragraph 81 will be revised accordingly.
11. Please clarify paragraph 46 of CAR Chapter 5, which allows institutions to use the Foundation Approach for their Specialized lending portfolios. While this is the original Basel II text, OSFI has not in the past allowed the use of the Foundation Approach. Can AIRB institutions move their Corporate Specialized lending exposures to the Foundation Approach on adoption of the 2023 CAR Guideline? Would the institution need to advise OSFI? Could both AIRB and FIRB approach be used for an institution’s specialized lending portfolios?
AIRB institutions are still expected to continue to use the AIRB approach for their Specialized Lending exposures. If an institution wishes to voluntarily move any AIRB portfolios to the FIRB approach, they must receive explicit approval from OSFI, as described in paragraph 58 of CAR Chapter 5.
12. In respect of Retail Margin Lending, for capital floor purposes and institutions that only use the Standardized Approach to Credit Risk, please confirm that credit risk mitigation is not permitted. Note that this paragraph does not exist in CAR Chapter 4.
Credit risk mitigation is not recognized for retail margin lending business under the Standardized Approach to Credit Risk.
13. In relation to OSFI’s response to stakeholder feedback in the cover letter published on January 31, 2022 with respect to guarantees (paragraphs 301 and 305 of CAR Chapter 5), institutions would like to clarify that banks are not trying to recover more than the dollar amount lost in the case of default. There is a greater chance of recovery from the presence of both collateral and a guarantor, and that this should be recognized. Please clarify the OSFI’s response and provide a calculation example of how this should be implemented for a transaction that is covered by collateral and guarantee.
OSFI understands that institutions are not trying to recover more than the dollar amount lost in the case of default. The clarification in the CAR Guideline is to show that recognizing both guarantees and collateral simultaneously on the entirety of the exposure overestimates the benefits of credit risk mitigation (CRM). For example, consider the situation where an institution has a $100 loan to Party A and Party A has pledged $90 worth of collateral (assuming no haircuts for simplicity) and the exposure is also guaranteed by Party B. In the event that Party A defaults, the institution will likely liquidate the collateral and recover $90. Further, even though the entire $100 of exposure is guaranteed by Party B, the institution will only be in a position to request $10 from Party B. Therefore, applying Party B’s probability of default (PD) and the secured loss given default (LGD) to the entire $100 exposure is overstating the CRM benefits of the transaction. The actual recovery value of the collateral is likely to be more volatile and uncertain than in the simplistic example above, making it difficult to allocate a certain portion of the losses to the collateral and a certain portion to the guarantee. For that reason and for simplicity, the framework only allows the recognition of either the collateral or the guarantee, whichever provides the greater capital benefit to the institution.
1. Which asset class should derivatives referencing bond indices be assigned to? Further, how should single-asset bond forwards be assigned?
Institutions must assign derivatives to an asset class based on its primary risk driver under paragraph 120 of CAR Chapter 7. For bond forwards, it is possible that the primary risk driver differs depending on the bond’s issuer. For bond indices, the same concept would apply. However, as there may be difficulty computing the duration for these instruments since they do not have a maturity date like individual bond, one potential solution would be to compute the weighted average maturity of the bonds in the bond index. Institutions may also reach out to OSFI on a bilateral basis if they have specific questions related to specific exposures.
2. Does collateral included only as initial margin impact the margin period of risk (MPOR)?
Initial margin collateral should not impact the MPOR. That is, if a netting set contains only liquid variation margin and some illiquid initial margin, the illiquidity clause to increase the MPOR of the netting set would not be triggered.
3. Do margin period of risk (MPOR) rules apply to all collateral or to inbound collateral only?
The increased MPOR clause related to illiquidity in paragraph 142 of CAR Chapter 7 only applies to inbound collateral. However, the increased MPOR clause related to disputes in this paragraph applies to both inbound and outbound collateral.
4. If institutions have the unilateral right to terminate contracts early at predefined dates, can the predefined dates be used in the calculation of duration instead of the final contractual maturity? Would the answer differ if the early termination date is also a cash settlement date? Does the answer differ if instead of the unilateral right to terminate, there is a mutual put option where both parties must agree to terminate early?
Based on the evidence we have collected from institutions, early termination rights are very rarely exercised due to the negative impact on client relationships. As such, the final contract maturity must still be used to calculate duration. This is the case for all early termination options, regardless of whether they are unilateral or mutual or if the termination date is also a cash settlement date.
5. Are central banks considered a source of specific wrong-way risk (SWWR)?
Given that central banks are assumed to present no default risk, counterparties (including central banks) subject to a zero percent risk weight under sections 4.1.1-4.1.3 of CAR chapter 4 are not considered a source of SWWR.
1. For paragraph 114 of CAR Chapter 9, regarding an institution’s choice to include instruments without optionality in the curvature risk calculation, please clarify the requirement to apply this approach consistently through time.
The choice to include instruments without optionality in the calculation of curvature risk rests with each institution; however, once this choice is made, there should be a consistent approach applied at each desk level over time. Should changes be necessary as a result of major methodology or infrastructure enhancements, OSFI should be notified.
1. Is the statement OSFI made in March 2020 that Liquidity Coverage Ratios (LCRs) can go below 100% a signal that institutions are in trouble?
OSFI expects that institutions' pools of liquid assets to be used during a period of financial stress, even if that means an institution falls below a 100% LCR level. This is consistent with previous statements OSFI has made over time where we have encouraged institutions to use their liquidity buffers as appropriate.
2. How does lowering the LCR facilitate lending?
By using the pool of high-quality liquid assets, and thus lowering its LCR ratio, an institution is able to convert securities / market instruments into cash (through sale or repo). The institution is then able to use that cash to cover outflows, which could include generating new loans to customers.
1. Regarding the “accrued interest over previous 12 months” for credit cards in paragraph 67 of the Liquidity Adequacy Requirements (LAR) Guideline, please clarify if this is interest actually charged (and not accrued) to clients for late payments or payments less than the full amount due. Is the $50 threshold the sum of all interest charged over a 12-month period or the interest that exceeds $50 for any one-month invoicing cycle?
Measurement/treatment of transactors in the LAR Guideline should be consistent with how products are measured/treated in paragraph 84 of Chapter 4 in OSFI’s Capital Adequacy Requirements (CAR) Guideline. In that, interest is understood to have accrued once there is an amount that can be charged to clients (e.g., for late payments or payments less than the full amount due). Utilization against the $50 threshold should consider the sum of all interest charged over a 12-month period.
2. Regarding the “no drawdown over the previous 12 months” for overdraft/lines of credit in paragraph 67 of the Liquidity Adequacy Requirements (LAR) Guideline, please confirm whether this refers to: any act of drawing down a line over the previous 12 months, or; the existence of a drawn amount over the previous 12 months, even if the actual draw occurred more than 12 months ago but has not yet been repaid. In addition, would this apply to customers who draw on their overdraft / line of credit but repay them in full prior to the end of each month (or the next statement date)?
Measurement/treatment of transactors in the LAR Guideline should be consistent with how products are measured/treated in paragraph 84 of Chapter 4 in OSFI’s Capital Adequacy Requirements (CAR) Guideline. In that, drawdown for overdrafts and lines of credit over the previous 12 months refers to the existence of any drawn amount during the previous 12-month period, regardless of when the initial draw occurred or when it was repaid.
1. Paragraph 13 of the Leverage Requirements Guideline states that the leverage ratio buffer will be set at 50% of a D-SIB’s higher-loss absorbency risk-weighted requirements. How does the new leverage ratio buffer interact with the Total Loss Absorbing Capacity (TLAC) leverage ratio?
The leverage ratio buffer is added to the 6.75% minimum TLAC leverage ratio applicable to D-SIBs. Given the calibration of the leverage ratio buffer, this results in a supervisory target TLAC leverage ratio of 7.25%.
2. Are institutions' authorized leverage ratios still in force, or only the 3% regulatory minimum?
Institutions are required to maintain leverage ratios above their authorized leverage ratios at all times.