A1. No, not yet adequate. Please see our answer to Q2.
A2. No, not yet adequate. We recognise the difficulty in establishing a quantitative methodology which can accurately assess the systemic importance of institutions across business models and geographies. However, we have concerns about a number of areas of the proposal.
Size and organisational structure
Overall, the twelve indicators appear highly correlated with size, measured for example in terms of the group-wide consolidated value of certain balances or the volume of payments.
In our view, it is simplistic, in the case of decentralised banking structures, to measure the impact in terms of the aggregate group-level numbers without due consideration being given to the benefits of the distribution and dispersal of those numbers across markets around the world.
Aggregate size is not the sole defining indicator of financial markets impact and more attention should to be paid to risk profile, organisational complexity and business mix.
Moreover, using a group-wide consolidated approach has other consequences of significant concern, as the frameworks for assessing G-SIIs and Domestic Systemically Important Banks (‘D-SIBs’) are not necessarily complementary. For example, our largest European operating entity, HSBC Bank plc, is significantly smaller than the other G-SIIs designated alongside the HSBC group in buckets 3 and 4. However, on a stand-alone basis, HSBC Bank plc would rank in bucket 1 of the G-SII list.
Notwithstanding this, HSBC Bank plc and our other entities are required to operate under the capital requirements and capital cost constraints of a 2.5% CET1 HSBC Group G-SII buffer, while seeking to compete in overseas markets against much larger local D-SIBs that are likely to be placed under less stringent capital requirements, given the degree of national discretion provided to member states to impose both global and domestic buffers.
We hope that such scope for inequality will be addressed in the EBA’s impending guidelines on the framework for domestic systemically important institutions, which we understand is due for publication by 1 January 2015. It is also an issue that we intend to raise with our own supervisor, the Prudential regulation Authority ahead of publishing its own proposals in this area later in 2014.
We understand that the G-SII assessment is attempting to quantify the adverse impact on counterparties, customers, markets and financial infrastructures of a G-SII’s failure, not the likelihood of failure. This has been equated to measuring a Loss Given Default as opposed to a Probability of Default.
In our view, this overlooks the important role of good risk management practice and governance, not only in preserving the soundness of the institution but also in limiting eventual adverse impacts, for example, in managing operational risks, ensuring operational transparency and accountability, conducting and properly evaluating stress scenario testing exercises and having the capacity actively to manage risk mitigating actions in the event of external or internal difficulties.
Provided that it is assessed under sound and consistent standards, the quality of risk management should be factored into the proposed scoring system, partially replacing weights assigned to quantitative factors. This would have the further benefit of incentivising improved risk management.
Importantly, in the universal application of the assessment approach across all banking groups there is no recognition of the substantial differences in the ways financial groups are organised or of the fundamental differences in the way such groups could and would be resolved in case of need.
To give an example, HSBC operates as a series of locally incorporated and regulated banks across the world, separately capitalised with their own pools of liquidity and funding. There are strict limits on the permitted financial interactions between these banks so that they largely operate on an arm’s-length basis from each other.
We believe this structure gives HSBC significant advantages in the event of resolution. Namely, in the event of failure of one (or more) of its operating banks, these could be resolved on a stand-alone basis, leaving other banks in the Group to continue to operate as normal under a ‘multiple point of entry’ approach.
As HSBC’s resolution strategy envisages allowing an individual bank to fail without affecting other parts of the HSBC Group, using group-wide consolidated metrics would not seem appropriate in determining the level of G-SII buffer. Therefore, since the G-SII methodology is currently structured to assess banking groups at group consolidated level, recognition and credit should at the least be given to the structure and quality of resolution arrangements within a banking group perhaps as an ‘overlay’ to the basic framework.
A3. Yes, in our view these timelines would be adequate, provided national supervisors do not accelerate the implementation of systemic capital buffers.
A4. Generally, they are. However, we strongly question, in respect of the data collection template definition of cross-jurisdictional liabilities, why only local currency exposures to obligors in the home EU jurisdiction of the group parent entity are excluded from the calculation.
By contrast, the template treats liabilities of a UK banking group’s Brazilian bank subsidiary in Brazilian currency to domestic Brazilian clients and counterparties as cross-jurisdictional. In our view, this imposes an illogical EU-centric classification of business activities; there appears no reason why non-EU ‘home country’ activity should be assessed for the weight of its market impact, but not such activity within the EU home country.
This inflates considerably the total ‘cross-jurisdictional’ liabilities of a group such as HSBC that operates with a holding company and subsidiary operating bank structure, with a punitive impact on the G-SII score. UK domiciled banks are similarly and further disadvantaged in respect of lending within the Eurozone area vis-à-vis their EU domiciled competitors. We would like to see the template instructions changed to reflect these issues.
A6. We do not entirely agree. We believe there is a real danger that the identification of G-SIIs will lead to market distortions. In particular, there is a risk of contraction of depositor preference in any crisis. This is likely to arise given the perceived advantages of and protection from criticism from depositing funds with a G-SII, particularly if the depositing party has fiduciary responsibilities for managing monies.
At the same time, during a crisis there is likely to be wariness amongst G-SII about re-circulating those funds into the inter-bank market, particularly to non G-SIIs – this would potentially increase the level of key indicators – with the result of potentially constraining the flow of funds to non G-SIIs at a critical time. Perceptions of such risks are likely to be self-fulfilling, pushing market liquidity provision back to the central banks that will be the recipients of deposit inflows from the G-SIIs.
One area of notable difference from the Basel Committee framework is disclosure – the granularity of the format in these proposals is more onerous, we believe, than that envisaged under other regulatory regimes, to the disadvantage of European headquartered G-SIIs. We think it important that rules for disclosure, as for assessment, be applied consistently, allowing better comparisons and the fostering of market discipline.
That said, as far as we are aware there is no evidence that the market sets any value upon this data set. If such interest were to emerge, we would prefer to see it articulated and developed through initiatives such as the Enhanced Disclosure Task Force, which we fully support.
Regarding the extension of the scope of disclosure requirements to institutions with exposures exceeding EUR200bn, we suggest that the proposals be clarified to spell out an exception to that rule, in order to reflect guidance we have received from both yourselves and the PRA. This is that, for a banking group headquartered in the EU, where G-SII reporting and disclosure take place at group consolidated level, an operating entity of that group also domiciled in the EU would be exempted from the G-SII reporting and disclosure requirements set out in these RTS and ITS, notwithstanding it may exceed the EUR 200bn exposures threshold and score at or above the 130bp threshold referred to in Article 5.3 of the draft RTS.
In the HSBC Group’s case, this would exempt HSBC Bank plc. We unfortunately cannot definitively conclude from the definition in Article 2.1 of the draft RTS that HSBC Bank plc would be excluded from the set of ‘Relevant entities’. Our rationale for believing they should be is that we presume such an operating entity will be covered in due course by D-SIB requirements.
As currently structured, we remain unconvinced that the proposed methodology and indicators provide an appropriate assessment of the risks that our group presents to the global financial system.
We believe that, in the majority of respects, HSBC is close to the place that the authorities’ reforms would like systemically important institutions to reach – with higher capital ratios, a better funding structure, stronger liquidity and a simpler business model, further supported in the case of HSBC, by virtue of operating through strongly capitalised and funded legal entities capable of being separated if required.
There are fundamental differences between the structure and business models of institutions. These have important consequences for resolvability which are yet to be reflected in these proposals. As the regulatory framework evolves, it is important that this analysis be undertaken, if there are to be proper incentives for banking groups to adopt or retain more resilient and resolvable structures, thereby reducing the impacts on the global financial system in the remote event that they were to get into severe difficulties.