Primary tabs

British Bankers' Association (BBA)

The BBA is pleased to respond to this consultation.

Key messages

We note that the EBA sets out in its executive summary “the geographical location of credit risk exposures should reflect the ultimate risk of an institution’s loan portfolios”.

Note: Whereas we agree with this general statement, the EBA Consultation Paper then uses the term ‘ultimate risk’ in a very broad sense (with a choice between obligor basis and guarantor basis), which confusingly diverges from the use of the term ‘ultimate risk’ in the existing reporting of country exposures to the central banks and onto the BIS (where it refers only to ‘guarantor basis’, i.e. after country risk transfers).

We would recommend the EBA to align its terminology with existing BIS terminology, and use the term ‘ultimate risk’ only to refer to the ‘guarantor basis’.

Credit Risk

• For retail exposures the country of residence of the obligor is acceptable but in the case of guarantees from or collateral elsewhere banks should be permitted to assign the exposure to the country of residence of the guarantor or location of that collateral.

• For all other exposures and in contrast to what is proposed in the EBA paper, we suggest that the firms be allowed to use their existing country risk management methodology, which may include (1) looking beyond the obligor’s country of residence (head office) to the country of its main assets (look-through principle) and (2) the application of country risk transfers based on guarantees or insurance from, or collateral in, third countries. (i.e. guarantor principle).

• The same should apply to specialised lending such as project finance.

• Institutions recommend that the threshold should be increased to 5%, but that they should be permitted to identify exposures below the 2% threshold if they choose to do so.

Trading Book exposures

• The terminology used for this section is too general. It should be re-defined to specifically refer to Specific risk for those on the standardised approach, and Incremental Default and Migration Risk (IRC), for those on an internal model approach.

• The text needs to be improved to require those institutions that do not use internal model approach to use of the specific risk charge of the Standardised Approach.

• We do not agree that institutions that use the internal model approach should be also required to have a parallel process to calculate the specific risk under the Standardised Approach as the only basis of determining the allocation of exposures

• We set out an alternative approach that permits institutions to recalculate modelled specific risk by country and to apportion the diversification benefits to each country exposure.

• We recommend that the threshold for excluding the identification of geographical specific risk is increased to a maximum of 15%, (in order to introduce a degree of proportionality with Credit Risk) with the proviso that institutions should be permitted to use a lower threshold should they wish to do so.

Securitisation

• We also draw the EBA attention to fact that exposures that have passed the significant risk transfer test (SRT) should be excluded from the scope, as should exposures that are already deducted from capital (which should also apply to other credit risk exposures). The rationale is that there will be no risk weighted assets associated with these assets even though they will be still reported in the balance sheet.

• We agree that look-through approach is appropriate.



• However, for securitisations that embrace more than one country, institutions should be permitted the option to apportion the exposures to each country instead of allocating the entire risk to only one country in order to avoid unintended consequences.

• A threshold of 20% of total credit, specific and securitisation RWAs should be established permitted to obviate the need to identify the geographical location, although institutions should be permitted to apply lower thresholds. Exposures below the threshold would be treated as domestic.


Cost-benefit Analysis

It is difficult to develop a meaningful cost benefit analysis at this stage of the choice between obligor, guarantor and or other techniques such as “look-through”.

While we understand the theoretical justification for a CCB, we think that an overly onerous and prescriptive calculation methodology creates significant costs, that are ultimately born by customers, as there is no risk management improvement stemming from it that is not already incorporated into existing country risk methodologies. We would therefore like to re-iterate that a CCB calculation that is based on firms existing country risk methodologies is to be encouraged. Any deviations from this do not seem to add any benefit.

As such, we consider the methods should allow the use of practical expedients while the system and methods are under development where these would lead to a materially similar outcome. In that way, firms will be able to leverage existing information systems and approaches as well as retain alignment with CO-REP and other reporting as these processes develop.
Location of exposures

We agree that using the definition of the “place of obligor” will for many exposures be appropriate, specifically for all classes of Retail exposures.

However, we don’t agree with using the obligor principle in all cases, as explained in our answers to sub-questions (a) and (b) below.

Scope

CRD Article 140 (4) setting out the calculation of institution-specific countercyclical capital buffer rates states that “relevant credit exposures shall include all those exposure classes, other than those referred to in points (a) to (f) of Article 112 of Regulation (EU) No 575/2013”.

The consequence is the article refers only to the classifications listed in accordance with the Standardised Approach.

It is clear that the following exposures are excluded:
(a) central governments or central institutions;
(b) regional governments or local authorities;
(c) public sector entities;
(d) multilateral development institutions;
(e) international organisations;
(f) institutions;

The consequence is that all of the following are included
(g) corporates;
(h) retail
(i) secured by mortgages on immovable property;
(j) in default;
(k) associated with particularly high risk;
(l) in the form of covered bonds;
(n) institutions and corporates with a short-term credit
(o) in the form of units or shares in collective investment undertakings ('CIUs');
(p) equity exposures;
(q) other items.

We believe that n) institutions with a short-term credit have been included by mistake. We believe it should be excluded.

We question why (q) other items should be included as normally these do not represent credit exposures.
IRB approach

It is noted that the classification of exposures for IRB is slightly different to Standardised Approach and is set out in CRR Article 147: Methodology to assign exposure to exposures classes.

Each exposure shall be assigned to one of the following exposure classes:
(a) central governments and central institutions;
(b) institutions

These classifications cover the list of (a) to (f) in Article 112 set out above. For the avoidance of doubt, we believe that the text should explicitly make reference to excluding these IRB exposures and thus aligning the exclusions in the Standardised and IRB approaches. We believe it would be prudent to explicitly state that the following exposures set out in Article 147 are included:

(c) corporates;
(d) retail

We believe that (g) other non credit-obligation assets should be excluded from the scope. There is no justification for including a CCB on assets, such as offices, computers and other such assets, as an example.

Equity exposures

We note that the equity instruments are proposed to be in scope both in the Standardised and IRB Credit Risk as well as for Specific Risk in the Trading Book.

Standardised Credit Risk
Article 112 of Regulation (EU) No 575/2013”.
(p) equity exposures;
IRB Credit Risk
Article 147
(e) equity;
Specific Risk in the Trading Book
Part Three, Title IV, Chapter 2
Section 3 Equities
Article 342 Specific risk of equity instruments

Equity exposures are not normally classified as credit exposures and we would request that they are excluded from the scope.

Counterparty Credit Risk

We suggest that for counterparty credit risk for OTC derivatives that the geographical location of the underlying instrument is key for counterparty risk rather than the location of the obligor.

Definitions
We note that in Article 1 there are number of terms, natural person, legal person, ordinarily resident, registered office and actual centre of administration. Internal processes and systems will need to be put in place to identify Risk Weighted Assets differentiated by the many different nuances of location. We believe that it would beneficial if the EBA could cross-reference these terms to existing EU legislation.


a) If not, could you provide specific examples where this principle would not work in practice
Non-Retail exposures are managed on a relationship basis. Therefore using the definition proposed by the EBA will not result in the risk (i.e. loss) in all cases being linked to the country of risk.

The three main examples of this are where:
1. The business (i.e. risk) is conducted in a country different and or countries to the definition as proposed by the EBA, and
2. The use of credit risk mitigation (CRM) techniques change the country of risk.
3. Lending to subsidiaries is based upon the assessment of support provided by the parent

Examples include

• Lending by a British bank to an investment firm residing in Luxembourg with assets largely in the UK should be assigned to the UK, since the risk for the bank is associated with the UK economy rather than with Luxembourg.

• Lending to a Dutch multinational corporation or real estate company with a broad international distribution of assets/investments should be assigned to the Netherlands, as it cannot be meaningfully assigned to any other country and it is difficult to locate or operationally not feasible to report all underlying exposures geographically

• Lending by a French institution to an obligor in Spain guaranteed by a German guarantor. The risk for the bank of losses depends on the German guarantor rather than on the counterparty in Spain (unless the German guarantor would have a weaker rating than the Spanish counterparty).

b) explain why an alternative option, for instance the guarantor principle, would work better
The purpose of the CCB ('which is to protect the banking system against potential future losses associated with excess credit growth' - see page 13 of the EBA's Consultative Paper) would seem to basically coincide with the purpose of country risk management (which is to protect the bank against potential future losses associated with country events).

Therefore we would think the country assignment principle for the CCB should be the same as for country risk management. This implies we think banks should be permitted to use their existing country risk management methodology, which may include looking beyond the obligor’s country of residence (head office) to the country of its main assets (look-through principle).

This however should not imply that banks should face an obligation of exactly tracking where individual companies applies the loans they have been given by banks.





Further notes

We note that the second scenario is catered for within the EBA draft technical standards on supervisory reporting requirements published on 26th July 2013 Annex II – Reporting on own funds and own funds requirements, section 3.4.(Credit and counterparty credit risks and free deliveries: Information with geographical breakdown (CR GB)) paragraphs 80 & 81.

80. The term ‘residence of the obligor’ refers to the country of incorporation of the obligor. This concept can be applied on an immediate-obligor basis and on an ultimate-risk basis. Hence, CRM techniques can change the allocation of an exposure to a country.
81. Data regarding ‘original exposure pre conversion factors’ shall be reported referring to the country of residence of the immediate obligor. Data regarding ‘exposure value’ and ‘Risk weighted exposure amounts’ shall be reported as of the country of residence of the ultimate obligor.

We draw the EBA attention to the fact that the reference to 'country of incorporation' in the first sentence of paragraph 80 above is not appropriate. The country of residence, as already used by the BIS and national regulators for country exposure reporting and as implemented in bank systems, often differs from the country of incorporation. Thus we think institutions are best placed to determine where the country of incorporation and the country of residency differ and which one is the better reflection of the real risk.

We therefore believe that the determination of the country of location should be consistent with the institutions’ internal country risk management approaches.

Other suggestions

Institutions should be permitted and encouraged to use a “look-through” approach to identify circumstances in which the risk is different to using only the prescriptive definition of “place of obligor”. This will ensure that the geographical location of credit risk exposures will reflect the ultimate risk of an institution’s portfolios.

A further refinement should permit a bank to apportion the exposure of a corporate customer across more than one country. This should though not be a prescriptive requirement.

In general, if the approach is to align risk with a country of risk then institutions should be encouraged and permitted to develop risk-sensitive approaches to assign, allocate and apportion exposures to the countries in which the risk exists in line with the country risk methodologies.

To do otherwise would result in a weighted average buffer rate that would not be equitable.

We draw the EBA attention also to exposure covered by ocean-going vessels. The Bank of England has specified that their country of residence should be the country of registration (flag) - which often is Liberia, Panama, Marshall Islands etcetera.

However, since the banks don’t run risks on those countries in this example nor do those countries run a resulting risk of excessive credit booms, we propose that again such exposures be assigned in line with firms existing practices for country risk management.
The same should be true for exposures to aircraft.
We believe that it is too prescriptive.

We feel that the place of income can in many instances be just as elusive as any other definition and may not add anything to the objectiveness of the CCB calculation.

The notion of place of income is only useful in a limited number of instances, such as a property development in a specific country. Even then though if the lessee’s are overseas firms, it is again unclear where the country of income is. Leaving that aside we would contend that in practice there is no justification for diverging from the notion of place of residency. The notion of place of residency means where the economic activity is taking place. Particularly we note that for immovable assets it is generally the law of the country in which the asset is located that determines obligations.
In line with the recommendations on the approach to Credit Risk other techniques including “look through” and credit mitigation techniques should be permitted.

Furthermore, when a bank uses the Standardised Approach there is no specific category for Specialised Lending. Therefore to introduce a sub-categorisation of the Standardised Approach for Specialised Lending introduces an inconsistency within CoRep.

We believe that there is no need to identify separately this classification of exposures, especially in view of the small percentage of total exposures that this category represents. .

We therefore believe that the same approach as set out in “Annex II – Reporting on own funds and own funds requirements, section 3.4.(Credit and counterparty credit risks and free deliveries: Information with geographical breakdown (CR GB)) paragraphs 80 & 81” should be adopted allowing for credit risk mitigation techniques to be included in the determination of the country of risk

To that the geographical location of credit risk exposures will reflect the ultimate risk of an institution’s portfolios we remain of the view that the EBA should rely on the banks own country risk management methodology.
Yes, we believe that CRM techniques should be permitted to be used to determine the location of the exposure. This will ensure that the geographical location of credit risk exposures will reflect the ultimate risk of an institution’s portfolios and is in line with firms’ internal management.

We believe that if the proposals as set out above in answer to questions 1 and 2 also include also the obligation to:

1. Identify the country of risk (using internal country risk methodology) and

2. Take into account relevant CRM techniques, then such an approach would achieve the EBA objective.

For regular comprehensive guarantees and insurance: since risk lies with the guarantor / insurer (unless they have a weaker rating than the direct obligor/debtor)

For cash or gold collateral by an institution country risk lies with the country where the branch or subsidiary holds the cash collateral

For non-cash collateral such as residential or commercial mortgages: In principle no, unless it is obvious that the risk of losses does indeed lie with the country of location of the collateral.

For the commercial real estate portfolio the income principle might become more relevant as the exposure is generally designated by that class of assets where repayment is dependent on the cash generated by the assets.

There are a number of examples in which it can be argued that the risk could be associated with either the country of residence of the obligor or the country of risk. We suggest that the EBA acknowledges there is no absolute right or wrong allocation, and that firms will be allowed to use some judgement for a sensible allocation.

Examples:

• Lending by an a UK institution to UK residents covered by mortgages on Spanish residential real estate could be reported on the UK, since the UK residents will usually have more assets or income sources from which to service their debt.

• Lending by a UK institution to a French real estate company that fully or largely focusses on real estate investments in Finland, the risk for the institution of losses depends on the Finnish real estate market (which may be affected by excessive growth) rather than on the direct counterparty in France.
Threshold

Article 2.4 – Credit Exposures

Notwithstanding paragraphs 2 and 3 of this Article, if the aggregate amount of an institution’s foreign credit exposures does not exceed 2% of their aggregate of credit, trading book and securitization risk weighted exposures, these credit exposures shall be deemed to be located at the place of the institution.

As the text is written this implies that all RWAs, including exposures to Sovereign, Institutions and General Risk etc. are in scope. Please can this be reaffirmed?

Or in fact is the intent that the threshold only applies to the total of the RWAs that are in scope for the determination of the CCB rate.

It would be helpful for this matter to be made explicitly clear in the text.

We believe that the practical approach that institutions will adopt is to determine the percentage of domestic exposures of the total of RWAs (for the three asset classes that are in-scope). If that is less than 100% minus the EBA threshold, then this will cause a need to identify all foreign exposures.


A consensus is that a threshold of 5% would be more appropriate

However, some institutions may choose to put into place (and may already have done so) processes that identify exposures to each and every country. Thus the % should not be seen as floor and institutions should be permitted to impose upon themselves a lower threshold down to 0% and in line with internal country risk methodologies.


Burden on reporting

The argument for a higher threshold is that it would alleviate the burden of identifying the foreign exposures. The counter-argument is that all exposures below a threshold must be counted as domestic thus distorting the weighted rate.

This places a burden on the need to identify the RWA for every country, without any minimum floor value and or a percentage that could be ignored on the basis of immateriality.
Please refer to our comments on the cost benefit analysis.







Should guidance be provided on the re-calculation frequency?

Institutions understand that a firm should always have enough capital.

But once a CCB is calculated institutions would not expect to see material changes to the weighted average within each quarter.

We note that regulators are required to report the CCB percentage on a quarterly basis.

Many institutions are required to publically report own funds (capital ratios) on a quarterly basis. We consider that the frequency for recalculation of the CCB rate on a quarterly is sufficient. We would not support monthly re-calculation.

Any more frequent calculation would add significant costs that are not justified by the benefits of the CCB. We would accept that firms should implement policies that the CCB is not arbitraged in between reporting dates.



Other matters - Current EBA proposals are 10% threshold for reporting

Annex II – Reporting on own funds and own funds requirements, section 3.4. (Credit and counterparty credit risks and free deliveries: Information with geographical breakdown (CR GB)) paragraph 79 states

79. Institutions fulfilling the threshold set in Article 5 (a) (4) shall submit information regarding the domestic country as well as any non-domestic country. The threshold is only applicable to Table 1 and Table 2.

We note that the EBA FINAL draft Implementing Technical Standards on supervisory reporting under Regulation (EU) No 575/2013 Article 5 (a) (4) sets out the “Format and frequency of reporting on own funds and own funds requirements” in which it is stated as follows:

(4) the information on the geographical distribution of exposures by country as specified in template 9 of Annex I, according to the instructions in Part II point 3.4 of Annex II, where non-domestic original exposures in all ‘non-domestic’ countries in all exposures classes, as reported in row 850 of template 4 of Annex I, are equal or higher than 10% of total domestic and non domestic original exposures as reported in row 860 of template 4 of Annex I.

For this purpose exposures shall be deemed to be domestic where they are exposures to counterparties located in the Member State where the institution is located. The entry and exit criteria of Article 4 shall apply.

We consider that the Co-Rep reporting threshold should remain in place and should not be impacted by the other CCB thresholds.

However, in light of the final decision taken with regard to the location, it is recommended that the EBA update the definitions in Co-Rep templates.
Scope of specific risk

We do believe that the approach could be made clearer.

We have read CRD Article 140.4 “Calculation of institution-specific countercyclical capital buffer rates” that confirms: Relevant credit exposures shall include all those exposure classes, other than those referred to in points (a) to (f) of Article 112 of Regulation (EU) No 575/2013 that are subject to d) where the exposure is held in the trading book, own funds requirements for specific risk under
• Part Three, Title IV, Chapter 2 of that Regulation
or
• Incremental default and migration risk under Part Three, Title IV, Chapter 5 of that Regulation;

We draw the EBA attention to the word “or” in the statement.

Requirement for institutions that only use the Standardised Approach

We recommend that the text is expanded to make it clear that the following exposures are in scope, namely the “standardised” approaches set out in Part Three, Title IV, Chapter 2 for those institutions that use the Standardised Approach either alone and on in combination with the use of internal models.

Section 2 Article 334 Debt instruments
Sub - Section 1
Article 335 Specific risk
Article 336 Own funds requirement for non-securitisation debt instruments
Article 337 Own funds requirement for securitisation instruments
Article 338 Own funds requirement for the correlation trading portfolio

Section 3 Equities
Article 342 Specific risk of equity instruments


Section 5 Specific risk own funds requirements for positions hedged by credit derivatives
Article 346 Allowance for hedges by credit derivatives
Article 347 Allowance for hedges by first and nth-to default credit derivatives

Section 6 Own funds requirements for CIUs
Article 350 Specific methods for CIUs


We note that equity exposures are not normally classified as credit exposures and we would request that they are excluded from the scope.

Internal Model Approach

We note that chapter 5 relates to the “Use of internal models to calculate own funds requirements” Articles 362 – onwards.

Allocating the total specific (SIR) and IRC capital charges

For institutions that model Specific and IRC, we do not agree that these institutions should be mandated to establish a process to replicate the calculation specific risk using the Standardised Approach. Indeed, some banks use IRC methodologies built up on an exposure basis and can therefore directly allocate IRC exposures to the relevant countries.

However for those banks that use a portfolio approach IRC, the Standardised approach is not a good measure of genuine exposure.

If institutions using internal models are required to calculate hypothetical standard rules based capital requirements, this information may not be readily available. Depending on existing infrastructure, this may require significant effort.

Not only do many institutions no longer calculate the SIR capital charge, but to require the calculation of the Standardised Approach would unhelpfully and confusingly result in a capital charge different to the charge calculated utilising the internal model approach.

Therefore for institutions that use the internal model approach, our preference is to permit those institutions to adopt the following approach

1. Use their existing IRC methodology directly where their models enables them to do so.
2. Run the internal model with only exposures to the in-scope asset classes (i.e. excluding exposures to Sovereign and Institutions) to calculate the capital charge
3. Re-run the internal model for each country in isolation to calculate a country specific charge.
4. Sum the country specific charges
5. Allocate the diversification benefit to each country proportionately.


The following sets out the approach in further detail, at least for those banks using similar methodologies.

Standalone IRC by geographic location might be used to pro-rate IRCRWAs to geographic locations. Most firms have to run IRC over various aggregation levels to reflect the IRC requirement by desk, entity level and for Group. Like VaR, the IRC at a higher level of the aggregation is less than the sum of the parts being aggregated. In other words, there is a diversification benefit such that IRC for an entity is less than the sum of IRC by desk within the entity, and IRC at Group level is less than the sum of IRC for each entity. Nevertheless, in order to decompose IRC by geographic region it should be possible for firms to use the same aggregation infrastructure to compute IRC by geographic location of the obligors. This simply requires the obligors to be grouped by geographic region and IRC run for each region and then also on an aggregated basis. The results of these runs can then be used to pro-rate aggregate IRC according to the obligor geography IRC:

Other pro-rate schemes could similarly be applied. For example, by region of ownership rather than region of obligor if required, which may correspond to the entity level aggregation that firms may already carry out.


In addition, we consider that stress testing might be used to identify the extent of exposure to each region. A firm might apply credit spread shifts to their credit portfolios by currency or geographic region, one region at a time, and pro-rate IRC according to the stress loss results. This method may require some alignment of the spread shifts used in the stress test, with the shifts used in the IRC model for migrations in order to ensure that the decomposition is performed reasonably accurately. To note, the IRC decomposition in this method would be aligned mainly to the migration part of the IRC and may not be able to capture multiple step downgrades and jump to default. However, it is likely still to be a reasonable decomposition method.
We note that Article 3.4 – Trading Book Exposures states

• Notwithstanding paragraphs 2 and 3 of this Article, institutions whose total trading book risk-weighted exposures does not exceed 2% of their total credit, trading book and securitisation risk-weighted exposures, shall not include the trading book exposures for the purposes of identifying the geographical location of their exposures.

Thus the requirement to determine whether or not the exposures are in scope is to look at total trading book RWAs inclusive of General Risk and specific risk exposures to institutions. This is inconsistent with other approaches.


We recommend that the threshold for excluding the identification of geographical specific risk should be applied only to the Specific Risk and IRC.

We also believe that the threshold should be increased so that firms whose own fund requirement for market risk is less than 5% of their total own funds may be entitled to ignore these risks for the sake of the CCB calculation.

The rationale for this to achieve a sensible proportionality that is not to labour intensive for those firms that are almost exclusively on banking book activities.
a) Can the approach proposed be implemented for re-securitisation exposures?

Given the fact that securitisation is a distinct asset class subject to a separate calculation of own funds requirements we agree that it may need a separate approach. Where a firm does apply look through for its internal country risk management purposes it should also do so for the CCB calculation. However we do not think it is necessary to introduce an overly costly specific look-through requirement. Institutions should have flexibility.

We also draw the EBA attention to fact that exposures that have passed the significant risk transfer test (SRT) should be excluded from the scope, as should exposures that are already deducted from capital (which should also apply to other credit risk exposures). The rationale is that there will be no risk weighted assets associated with these assets even though they will be still reported in the balance sheet,


b) Should other exposures such as CIUs also use the look-through approach? If yes, please justify the answer


We note that the RTS does not provide guidance on CIUs. We are, however, not in favour of implementing a mandatory look-through approach to these exposures.

In the event such an approach is adopted, we urge that the 5% granularity threshold introduced by the CEBS 2009 Guidance threshold be is adopted.

Most funds provide Net Asset Value and fund distribution data on a yearly cycle, geared to the annual review that most firms apply to these otherwise lower risk exposures. The internal processes and resources within firms are geared to analyse on this scale and frequency. Should a threshold be implemented, we would also urge recognition within the RTS that portfolio distribution limits within CIU investment mandates can be used to limit the data requested from vehicle managers, by exempting from analysis those underlying positions which are below any established threshold. More frequent data provision and review for calculating the CCB rate might thereby be more easily accommodated.

For the avoidance of doubt, we do not view covered bonds as falling under the securitisation requirements. Instead they should fall under the credit risk approach. Similarly, we do not consider that CIUs should be subject to a mandatory look through approach.
8 a. Would it be difficult to locate all underlying exposures geographically?


Whilst it is usual for the underlying risk within a securitisation to be primarily domiciled in one country, it is possible that some securitisations might have exposures to more than one country. Nevertheless we do see that it could become overly laborious to take a full look-through approach and feel the EBA should rely on firms internal approaches where ever possible.

We believe that the fall-back process should be to assign the risk to the geographical location with the highest proportion of underlying exposure.

However, this can result in unintended consequences. For example in a securitisation in which 40% of exposures are in country ABC that are not performing, it would be the assets in the three other countries each with 20% that would be required to pay the senior tranche first, even though the cyclical buffer would be applied 100% to country ABC. We believe that a further refinement should be to permit a bank to apportion the exposure of a securitisation across more than one country.

In general, if the approach is to align risk with a country of risk then institutions should be encouraged and permitted to develop risk-sensitive approaches to assign, allocate and apportion exposures to the countries in which the risk exists in line with the approach proposed for Credit Risk in answer to earlier questions.

Other matters

The proposals exclude any threshold for Securitisation.

A threshold of 20% of total credit, specific and securitisation RWAs should be permitted to obviate the need to identify the geographical location, although institutions should be permitted to apply lower thresholds. Exposures below the threshold would be treated as domestic.

The threshold would introduce a degree of proportionality with Credit Risk

8 b) Should other exposures such as CIUs also use the look-through approach? If yes, please justify the answer

We agree that CIUs should be treated in the same way as Securitisation, taking into account the recommendations set out above.
It is difficult to develop a meaningful cost benefit analysis at this stage of the choice between obligor, guarantor and or other techniques such as “look-through”.

We believe that too prescriptive an approach may make its calculation very costly. As firms already have country risk management approaches the CCB could, under the EBA proposals, lead to a dual calculation with the CCB calculation not adding to an individual bank’s risk management ability.

We therefore think that the EBA has underestimated the cost of implementing its CCB proposals, if it does not take a flexible approach aligned with internal management of country risk.

As such, we consider the methods should allow the use of practical expedients while the system and methods are under development where these would lead to a materially similar outcome. In that way, firms will be able to leverage existing information systems and approaches as well as retain alignment with COREP and other reporting as these processes develop.


Proportionality

The calculation described in the CP (and in the CRD) requires that a buffer rate is calculated on a subset of positions; which is then multiplied by the firm’s total own funds (for all exposure classes and all different calculations of risk exposure).

This could give some unusual results and inappropriate additional capital burdens.

For example, in the extreme case where a firm holds only Sovereigns and Corporate to one country in a 99:1 split, the CCyB rate will be calculated only on the country of risk of the Corporate risk weighted asset but would be applied to the whole portfolio including the Sovereign RWAs, Operational Risk and any general Market Risk too.

Yet, according to the regulation, an institution that has only exposures to sovereigns and institutions will not have any counter-cyclical buffer.

However, should it have the smallest exposure with an in-scope exposure, it would be subjected to the whole process and the national buffer rate would be applied to the total RWA for the institution. This cliff effect should be examined and removed in the final paper.

It is recommended that that a threshold of x % of total risk weighted assets should be applied to ensure that entities are not unnecessarily overburdened with requirements for complex calculations that represent a proportionally important cost for the institution, if the prudential regulatory objectives are reasonably covered.

It has been suggested that “x” could be 5%.


Proportionality – Thresholds

The EBA has set out different approaches to thresholds for each of the asset classes.

Article 2.4 – Credit Exposures

Notwithstanding paragraphs 2 and 3 of this Article, if the aggregate amount of an institution’s foreign credit exposures does not exceed 2% of their aggregate of credit, trading book and securitization risk weighted exposures, these credit exposures shall be deemed to be located at the place of the institution.

Article 3.4 – Trading Book Exposures

Notwithstanding paragraphs 2 and 3 of this Article, institutions whose total trading book risk-weighted exposures does not exceed 2% of their total credit, trading book and securitisation risk-weighted exposures, shall not include the trading book exposures for the purposes of identifying the geographical location of their exposures.

Comments
The consequence is that there could be Specific Risk that is completely excluded whereas the Credit Risk is included.

We note that in Article 3 – Securitisation, there is no threshold. Thus every securitisation and CIU is required to be identified.

The reality is that in most institutions Credit Risk RWAs, Trading Book RWAs and Securitisation RWAs represent very differing percentages of total RWAs. Thus the imposition of the thresholds is not proportional.

In the response to each question, we have highlighted these points and made some suggestions for higher thresholds with the proviso that institutions should be permitted to identify all RWAs by location in order to ensure that domestic exposures are treated failrly and proportionally.





Co-rep reporting

We would welcome alignment with the Co-Rep templates although note the Co-Rep templates are only structured around Credit Risk under Part Three, Title II of the CRR.

It is not yet clear there is an equivalent template for Market Risk to capture Trading Book exposures or Securitisation exposures identified by geographical location.

It is therefore not correct to conclude that the proposals can be satisfied using information that will be provided for Co-Rep.
Simon Hills
B