Response to consultation Paper on draft RTS on the treatment of non-trading book positions subject to foreign-exchange risk or commodity risk
Go back
While it makes sense to use a fair value basis for positions in the trading book as these positions are held with trading intent (or to hedge positions held with trading intent), it does not necessarily make sense to use fair value for those positions in the banking book, which are non-trading positions under accrual accounting and where these managed under accrual accounting.
Therefore, we are generally in favour of using the accounting value as a basis for computing the own funds requirements for foreign exchange risk stemming from non-trading book positions. However, most banks take actions to mitigate or offset these risks, including FX risks that arise from their business-as-usual activities using hedging strategies usually involving derivatives. In order to represent and measure those risks correctly (and to avoid creating mismatches and duplications due only to different accounting valuation), banks should be allowed to match the risk components hedged under different accounting valuation regimes.
For example, when using derivatives to hedge foreign currency denominated cash flows of an amortised item, by establishing a cash flow hedge relationship, the FX exposure will be completely closed. However, asymmetries in the account value by currency will appear. Furthermore, the article 33 of the CRR indicates that valuations of cash flow hedges are not considered subject to capital requirements when hedging amortised cost items.
As stated in Articles 1 and 2 of this Consultation Paper, institutions may use the last available (minimum quarterly) accounting or fair value of non-trading position as basis to calculate their market risk. Indeed, the valuation for most non-trading positions might not be available on a daily basis. What is also important to consider that different institutions may apply different capitalisation methods for these types of risk, i.e. either standardised approach or internal model. In the view of some AFME members, in case of the former, the daily computation of own funds for foreign exchange risk by valuing only the FX component might be an excessive requirement, especially in terms of required infrastructure.
Therefore, AFME members would suggest amending the proposed Article 3 of this Consultation Paper to allow the flexibility to the institution, which apply the standardised approach, to compute own funds for foreign exchange risk less frequent than daily.
The Associations disagree with the EBA’s proposed approach to treat non-monetary items measured at historical cost that are subject to the risk of diminution of value due to movements in the exchange rate between a foreign currency and the reporting currency as risk positions subject to capitalisation, where it overrides the Level 1 text which requires the valuation of assets and liabilities to be effected in accordance with the applicable accounting framework (cf. CRR article 24).
Items at historical cost are not managed with a trading intent nor with the purpose of being sold. They are subject to diminution in value if and only if there is an indication of reduced value due to the difference between the net realisable value or recoverable amount[1] and the carrying amount (cf. IAS 21 The Effects of Changes in Foreign Exchange Rates, para. 25). While a foreign exchange depreciation may cause a diminution in value, it is not the only factor which is considered to determine whether there is an indication of impairment.
Accordingly, and as any diminution of value is recognised in the accounts and group capital position regardless of this proposal, we see no legal reason to consider that these instruments that are denominated in foreign currency for the determination of the net open position and the determination of own funds requirements for foreign exchange risk. The potential FX risk on historical cost items should be captured possibly within Pillar 2 but should not lead to a Pillar 1 capital requirements.
As such, this question is also closely linked with the scoping of the structural FX provisions. As stated in the AFME response to the structural FX consultation, we believe that investments in subsidiaries and capital allocation to branches as well as associates and joint ventures and any positions stemming from such investments should not form part of the Net Open Position (NOP) under article 352 (1) CRR subject to Pillar 1 capital charge. This is as such positions do not affect P&L and are of a structural nature. We note that elsewhere in the CRR or the Basel framework there is no example of a Pillar 1 capital charge that would not relate to an impact on profit and loss accounts (P&L). The proposal to subject net investment (Type A translation risk) to a Pillar 1 capital requirement (notably at solo level) by requiring to capitalise these non-monetary items that are measured at historical cost as denominated in foreign currency would go beyond the existing regulatory requirements and against the spirit of the Basel standards.
Investments in subsidiaries and capital allocations of branches as they are obviously not ‘deliberately taken in order to hedge against adverse effect of the exchange rate on its ratios’ should not be subject to any capital charge (whether at consolidated level or at solo level). As such, they are already exempted from capital charge by nature. A different approach would be inconsistent with the regulatory framework and would lead to inconsistent results. As an illustration, such an approach would lead to a Pillar 1 capital charge for having appropriately capitalized a subsidiary (i.e. to the extent of their risk).
In its draft Guidelines, the EBA tentatively clarifies that non-monetary items held at historical costs are in the scope of positions to be included in the calculation of the open position and that in the context of the structural FX treatment, they are not taken into consideration when comparing the value of the net open position stemming from positions that are eligible to be structural against the threshold set by the guidelines.
However, the calculation of the net open position is based on the valuation of assets, liabilities and off-balance sheet items, according to the most commonly applicable accounting framework, i.e. International Financial Reporting Standards for European groups or national GAAP (cf. CRR s. 24, art. 111 and art. 166). According to the framework, items at historic costs are denominated in the functional currency of the entity in which they are accounted for per IFRS. Accordingly, based on the level 1 text, those items do not affect the net open position.
The Associations therefore disagree with the proposed methodology for capturing the foreign-exchange risk stemming from non-monetary items at historical cost under the standardised approach as it would lead to an undue Pillar 1 capital charge to investment in subsidiaries at solo level (if no waiver granted).
In this respect a causal relation between a sharp FX movement alone and the impairment (i.e. credit loss) of a non-monetary item appears to be a fairly implausible “in vitro” scenario and quite difficult to relate to real-world risks. If we refer to participation in a foreign subsidiary, an impairment could derive from a general deterioration of the economic conditions of the hosting country, that might be well associated with a depreciation of its currency. Disentangling the domestic effects from the FX ones, establishing a cause-consequence relation between the two in order to identify those cases in which FX depreciation alone eventually leads to an impairment is unlikely.
We consequently believe that capitalising impairment risk as FX conceptually overlaps with credit risk capitalisation and Pillar 2 stresses and add-ons. Therefore, assessment of capital adequacy for regional/domestic shocks is better suited for stress tests that may lead to Pillar 2 capital requirements rather than trying to capitalise for the standalone FX risk under the Pillar 1 requirements.
We do not see the same issues arising in the trading book.
To reiterate the point we believe investments in subsidiaries and capital allocation to branches as well as associates and joint ventures and any positions stemming from such investments should neither form part of the Net Open Position (NOP) nor being subject to Pillar I requirements under IMA as they do not affect P&L and are of a structural nature. More in general no Pillar 1 capital charge should be required on items that would not relate to an impact on profit and loss accounts (P&L).
We also reiterate that any attempt to model impairment risk as a consequence of pure FX movement appears extremely artificial and disconnected from actual economic risk and likely events that underpin the situation. Such an FX driven significant diminution in value is more likely to stem from wider economic impacts. A risk of creating an overlap with credit risk on the balance sheet is very concrete.
The implementation would make it very difficult for banks that use official accounting systems to produce actual and hypothetical P&L for banking book purposes. Sterilizing portfolio changes could be very complicated and require a significant overhaul of bank specific systems and practices and could require decoupling the calculation of the banking book backtesting-relevant P&L from other P&L calculations.
If such an approach is deemed appropriate based on impact assessment, on proportionality grounds we recommend allowing the use of a sensitivity-based approaches for the calculation of FX impacts in the Banking Book.
We also highlight that some of the requirements contemplated for valuation of banking book positions in the draft RTS go beyond the valuation requirements of available for sale portfolios, which in our view is an overreach and unlikely to provide meaningful safety and soundness benefits while significantly increasing operational efforts.
In reiterating that we do not support the inclusion in a market risk internal model of any approach to capture impairment risk deriving from changes in FX, it would appear unlikely that any modelling approach could yield an FX exposure higher than the one suggested by the RTS for the SA (Art 1 §5).
In accordance with IAS 36 (Impairment of Assets), the carrying amount of an asset for which there is an indication of a diminution of value is the lower of its carrying value before considering possible impairment losses and its recoverable amount. When such an asset is non-monetary and is measured in a foreign currency, the carrying amount is determined by comparing: a) the cost or carrying amount, as appropriate, translated at the exchange rate at the date when that amount was determined (i.e. the rate at the date of the transaction for an item measured in terms of historical cost); and b) the net realisable value or recoverable amount, as appropriate, translated at the exchange rate at the date when that value was determined (i.e. the closing rate at the end of the reporting period). Based on these definitions, it appears to be quite clear that FX rates can be at most an element in the determination of the loss from an impairment, not certainly the cause.
For the last few years, the EBA has been trying to harmonise the definition of default, impairment and non-performing exposures among financial reporting, regulatory reporting and internal risk management practices (EBA/GL/2016/07 and EBA/GL/2018/06). In the same vain, the model should use same definition of impairment as the accounting standards.
Q1. Do you agree with the approach in relation to the use of the accounting value and alternatively the fair value as a basis for computing the own funds requirements for foreign exchange risk, or do you think that institutions should be requested to use e.g. only the accounting value? Please elaborate.
We welcome EBA proposal to allow the flexibility of choosing the valuation approach between accounting and full fair value for the purposes of capitalising FX risk of non-traded balance sheet positions. A minor but important suggestion we have is to highlight directly in the document the reference to FRTB MAR11.4, which explicitly states that no FX risk capital requirements should be applied to positions related to the items that are deducted from a bank’s capital when calculating its capital base, or to positions of structural nature as per EBA-CP-2019-11.While it makes sense to use a fair value basis for positions in the trading book as these positions are held with trading intent (or to hedge positions held with trading intent), it does not necessarily make sense to use fair value for those positions in the banking book, which are non-trading positions under accrual accounting and where these managed under accrual accounting.
Therefore, we are generally in favour of using the accounting value as a basis for computing the own funds requirements for foreign exchange risk stemming from non-trading book positions. However, most banks take actions to mitigate or offset these risks, including FX risks that arise from their business-as-usual activities using hedging strategies usually involving derivatives. In order to represent and measure those risks correctly (and to avoid creating mismatches and duplications due only to different accounting valuation), banks should be allowed to match the risk components hedged under different accounting valuation regimes.
For example, when using derivatives to hedge foreign currency denominated cash flows of an amortised item, by establishing a cash flow hedge relationship, the FX exposure will be completely closed. However, asymmetries in the account value by currency will appear. Furthermore, the article 33 of the CRR indicates that valuations of cash flow hedges are not considered subject to capital requirements when hedging amortised cost items.
Q2. Do you agree that institutions should be requested to update on a daily basis only the foreignexchange risk component of banking book instruments? Please elaborate.
The industry fully agrees with EBA that the measurement of foreign currency risk from non-trading positions is not just a good risk management practice but also a regulatory requirement to underpin associated exposures with the capital irrespective of their non-trading nature.As stated in Articles 1 and 2 of this Consultation Paper, institutions may use the last available (minimum quarterly) accounting or fair value of non-trading position as basis to calculate their market risk. Indeed, the valuation for most non-trading positions might not be available on a daily basis. What is also important to consider that different institutions may apply different capitalisation methods for these types of risk, i.e. either standardised approach or internal model. In the view of some AFME members, in case of the former, the daily computation of own funds for foreign exchange risk by valuing only the FX component might be an excessive requirement, especially in terms of required infrastructure.
Therefore, AFME members would suggest amending the proposed Article 3 of this Consultation Paper to allow the flexibility to the institution, which apply the standardised approach, to compute own funds for foreign exchange risk less frequent than daily.
Q3. Could you please describe the current risk-management practices that institutions use for managing the foreign-exchange risk stemming from banking book positions, e.g. whether the accounting or the fair-value is used as a basis for determining the exposure in a currency, the frequency at which banking book positions are fully revalued, the frequency at which the foreignexchange component is updated?
Generally, FX risk for banking and trading book positions are measures according to accounting rules. This means: monetary assets / liabilities are revaluated daily using fixing rate through profit and loss, non-monetary assets/liabilities are revaluated at historical costs or at the revaluation date. Additionally, an entity shall assess at the end of each reporting period whether there is any indication that a non – monetary asset may be impaired. If any such indication exists, the entity shall estimate the recoverable amount of the non-monetary asset, which includes fair value decrease and foreign-exchange impact.Q4. Do you agree with the proposed methodology for capturing the foreign-exchange risk stemming from non-monetary items at historical cost under the standardised approach? Do you have any other proposal for capturing the foreign-exchange risk stemming from non-monetary items at historical cost that would be prudentially sound while fitting within the standardised approach framework? Please elaborate.
The regulatory framework set out by the CRR does not provide a clear definition of NMI and refers to Accounting Standards for a general description. This could be misleading as there are deviations across the accounting standards, jurisdictions and legal entity structures as well as different interpretations among banks. Therefore, EBA should consider those deviations by providing a better definition of NMI or clarifying and detailing NMI for prudential purposes to ensure a common ground (level playing field) is established. We do believe that giving a detailed description of NMI can become more productive and meaningful when it is comparable across jurisdictions and avoiding any unnecessary regulatory fragmentation.The Associations disagree with the EBA’s proposed approach to treat non-monetary items measured at historical cost that are subject to the risk of diminution of value due to movements in the exchange rate between a foreign currency and the reporting currency as risk positions subject to capitalisation, where it overrides the Level 1 text which requires the valuation of assets and liabilities to be effected in accordance with the applicable accounting framework (cf. CRR article 24).
Items at historical cost are not managed with a trading intent nor with the purpose of being sold. They are subject to diminution in value if and only if there is an indication of reduced value due to the difference between the net realisable value or recoverable amount[1] and the carrying amount (cf. IAS 21 The Effects of Changes in Foreign Exchange Rates, para. 25). While a foreign exchange depreciation may cause a diminution in value, it is not the only factor which is considered to determine whether there is an indication of impairment.
Accordingly, and as any diminution of value is recognised in the accounts and group capital position regardless of this proposal, we see no legal reason to consider that these instruments that are denominated in foreign currency for the determination of the net open position and the determination of own funds requirements for foreign exchange risk. The potential FX risk on historical cost items should be captured possibly within Pillar 2 but should not lead to a Pillar 1 capital requirements.
As such, this question is also closely linked with the scoping of the structural FX provisions. As stated in the AFME response to the structural FX consultation, we believe that investments in subsidiaries and capital allocation to branches as well as associates and joint ventures and any positions stemming from such investments should not form part of the Net Open Position (NOP) under article 352 (1) CRR subject to Pillar 1 capital charge. This is as such positions do not affect P&L and are of a structural nature. We note that elsewhere in the CRR or the Basel framework there is no example of a Pillar 1 capital charge that would not relate to an impact on profit and loss accounts (P&L). The proposal to subject net investment (Type A translation risk) to a Pillar 1 capital requirement (notably at solo level) by requiring to capitalise these non-monetary items that are measured at historical cost as denominated in foreign currency would go beyond the existing regulatory requirements and against the spirit of the Basel standards.
Investments in subsidiaries and capital allocations of branches as they are obviously not ‘deliberately taken in order to hedge against adverse effect of the exchange rate on its ratios’ should not be subject to any capital charge (whether at consolidated level or at solo level). As such, they are already exempted from capital charge by nature. A different approach would be inconsistent with the regulatory framework and would lead to inconsistent results. As an illustration, such an approach would lead to a Pillar 1 capital charge for having appropriately capitalized a subsidiary (i.e. to the extent of their risk).
In its draft Guidelines, the EBA tentatively clarifies that non-monetary items held at historical costs are in the scope of positions to be included in the calculation of the open position and that in the context of the structural FX treatment, they are not taken into consideration when comparing the value of the net open position stemming from positions that are eligible to be structural against the threshold set by the guidelines.
However, the calculation of the net open position is based on the valuation of assets, liabilities and off-balance sheet items, according to the most commonly applicable accounting framework, i.e. International Financial Reporting Standards for European groups or national GAAP (cf. CRR s. 24, art. 111 and art. 166). According to the framework, items at historic costs are denominated in the functional currency of the entity in which they are accounted for per IFRS. Accordingly, based on the level 1 text, those items do not affect the net open position.
The Associations therefore disagree with the proposed methodology for capturing the foreign-exchange risk stemming from non-monetary items at historical cost under the standardised approach as it would lead to an undue Pillar 1 capital charge to investment in subsidiaries at solo level (if no waiver granted).
Q5. How are you currently treating, from a prudential perspective, non-monetary items at historical cost that may be subject to an impairment due to a sharp movement in the foreign-exchange rate? In which currency are those items treated from an accounting perspective?
The industry suggests that the term diminution of value is more appropriate in the context of FX risk than the term “impairment”, as the latter has further connotations including provisioning under IFRS9. While a foreign exchange depreciation may cause an impairment of non-monetary items at historical cost, it is not the only factor which is considered to determine whether there is an indication of impairment and accordingly whether an impairment loss shall be recognised.In this respect a causal relation between a sharp FX movement alone and the impairment (i.e. credit loss) of a non-monetary item appears to be a fairly implausible “in vitro” scenario and quite difficult to relate to real-world risks. If we refer to participation in a foreign subsidiary, an impairment could derive from a general deterioration of the economic conditions of the hosting country, that might be well associated with a depreciation of its currency. Disentangling the domestic effects from the FX ones, establishing a cause-consequence relation between the two in order to identify those cases in which FX depreciation alone eventually leads to an impairment is unlikely.
We consequently believe that capitalising impairment risk as FX conceptually overlaps with credit risk capitalisation and Pillar 2 stresses and add-ons. Therefore, assessment of capital adequacy for regional/domestic shocks is better suited for stress tests that may lead to Pillar 2 capital requirements rather than trying to capitalise for the standalone FX risk under the Pillar 1 requirements.
Q6. Could you please provide an estimate of the materiality of non-monetary items that are held at historical cost for your institution (e.g. size of the non-monetary items at historical cost with respect to the institution’s balance sheet)? Please elaborate.
Institution specific question that the Associations cannot answer.Q7. Do you think there are any exceptional cases where institutions are not able to meet the requirement to daily fair-value commodity positions? Would these exceptional cases occur only for commodity positions held in the banking book or also for commodity positions held in the trading book?
It is quite possible for institutions to take ownership of commodities in the banking book in which no daily prices are available. This could for example be as a result of a restructuring or the taking on a commodity that had previously been provided as collateral and for which no liquid markets exist. On such assets valuation methodologies for capital requirements should be no different to the valuation requirements for accounting purposes.We do not see the same issues arising in the trading book.
Q8. Do you agree that, with respect to the valuation of foreign-exchange and commodity positions held in the banking book, the provisions applicable in the context of the alternative standardised approach (Article 1 paragraphs 1 and 2) should also apply in the context of the alternative internal model approach (Article 3 paragraphs 1 and 2)? Please elaborate.
The industry believes that banks should be allowed to use internal models for calculation, should they have sufficient data to do so. This is as the general principle that the measurement approach used to quantify a risk should not affect the scope of the relevant positions subject to that risk. In particular, all the considerations expressed before on the relevance of at-cost items in the context of FX-related Market Fisk are also applicable under the IMA.To reiterate the point we believe investments in subsidiaries and capital allocation to branches as well as associates and joint ventures and any positions stemming from such investments should neither form part of the Net Open Position (NOP) nor being subject to Pillar I requirements under IMA as they do not affect P&L and are of a structural nature. More in general no Pillar 1 capital charge should be required on items that would not relate to an impact on profit and loss accounts (P&L).
Q9. Do you agree with the provision requiring institutions to model the risk that non-monetary items at historical cost are impaired due to changes in the relevant exchange rate or do you think that the RTS should be more prescribing in this respect? Please elaborate.
AFME, ISDA and our members do not agree with the provision, both based on materiality and proportionality reasons. We already noted that the term “impaired” is confusing in this context and a better way of looking at it would be diminution of value. We do see permanent diminution of value as a result of extreme exchange rate movements as not fitting into the objective of Pillar 1 and as a result it should be considered under Pillar 2. If this is not done the industry would be worried that an extension of requirements beyond what is required in CRR is taking place.We also reiterate that any attempt to model impairment risk as a consequence of pure FX movement appears extremely artificial and disconnected from actual economic risk and likely events that underpin the situation. Such an FX driven significant diminution in value is more likely to stem from wider economic impacts. A risk of creating an overlap with credit risk on the balance sheet is very concrete.
Q12. Do you agree with the definitions of hypothetical and actual changes in the portfolio’s value deriving from non-trading book positions that have been included in the proposed draft RTS?
The Associations disagree with the proposal because the notion of hypothetical portfolio is inappropriate in a non-trading book context. While we agree with the definition of Hypothetical and Actual P&L provided in the draft RTS in that they are conceptually coherent with how banking book risks should be modelled under IMA, we do not view the proposal as implementable.The implementation would make it very difficult for banks that use official accounting systems to produce actual and hypothetical P&L for banking book purposes. Sterilizing portfolio changes could be very complicated and require a significant overhaul of bank specific systems and practices and could require decoupling the calculation of the banking book backtesting-relevant P&L from other P&L calculations.
If such an approach is deemed appropriate based on impact assessment, on proportionality grounds we recommend allowing the use of a sensitivity-based approaches for the calculation of FX impacts in the Banking Book.
We also highlight that some of the requirements contemplated for valuation of banking book positions in the draft RTS go beyond the valuation requirements of available for sale portfolios, which in our view is an overreach and unlikely to provide meaningful safety and soundness benefits while significantly increasing operational efforts.
Q11. Do you think that the requirement to capture the impairment risk in the risk-measurement model for institutions using the internal model approach is less or more conservative than the requirement proposed for institutions using the standardised approach? Please elaborate.
The Associations are unclear about how such assessment of conservativism should be conducted as it clearly depends on how banks would implement the modelling requirement.In reiterating that we do not support the inclusion in a market risk internal model of any approach to capture impairment risk deriving from changes in FX, it would appear unlikely that any modelling approach could yield an FX exposure higher than the one suggested by the RTS for the SA (Art 1 §5).
Q10.How institutions would capture the risk of an impairment in their risk-measurement model? Would the definition of impairment used in the internal model be identical to the one proposed in the accounting standards? Please elaborate.
We do not support to include in the model but, if it’s included in the model, the internal model should follow the same definition of impairment or diminution in value used in the accounting standards to maintain consistency.In accordance with IAS 36 (Impairment of Assets), the carrying amount of an asset for which there is an indication of a diminution of value is the lower of its carrying value before considering possible impairment losses and its recoverable amount. When such an asset is non-monetary and is measured in a foreign currency, the carrying amount is determined by comparing: a) the cost or carrying amount, as appropriate, translated at the exchange rate at the date when that amount was determined (i.e. the rate at the date of the transaction for an item measured in terms of historical cost); and b) the net realisable value or recoverable amount, as appropriate, translated at the exchange rate at the date when that value was determined (i.e. the closing rate at the end of the reporting period). Based on these definitions, it appears to be quite clear that FX rates can be at most an element in the determination of the loss from an impairment, not certainly the cause.
For the last few years, the EBA has been trying to harmonise the definition of default, impairment and non-performing exposures among financial reporting, regulatory reporting and internal risk management practices (EBA/GL/2016/07 and EBA/GL/2018/06). In the same vain, the model should use same definition of impairment as the accounting standards.