Banking StandardsWritten evidence from HSBC (V002)
Preamble
It is important at the outset to recognise that banks act as principal in virtually all transactions and therefore all market price movements accrue to the stated capital position of the bank either through the profit and loss account or directly to reserves in defined circumstances. Accordingly, it is not possible to define proprietary trading by reference to the beneficiary of market price movements as the impact of all such movements accrue to equity holders yet only a subset of activities would be regarded as trading and only a subset of these would be regarded as proprietary trading by most analysts. Proprietary trading definition is both complex and subject to considerable interpretation challenges.
We hope the comments below are helpful to the Commission in its deliberations on this important topic. The comments follow the pattern of questions set out in the Commission’s request for an additional submission.
Stability
1. What prudential concerns does proprietary trading within banks give rise to? To what extent are any concerns being addressed through existing measures and proposals, including ring-fencing?
The prudential risks of proprietary trading arise in three main areas:
(a)
(b)
(c)
A further risk increasingly being recognised arises from the possibility of market disruption from algorithmic or high frequency trading where the impact of programming errors or mis-keyed data input can cause a huge volume of automated trading to occur with potential to disadvantage other market participants who deal in the period of market disruption. While algorithmic and high frequency trading can improve pricing and transparency for all market participants automated trading systems do not have behavioural standards—they simply do what they have been programmed to do.
Valuation and calibration concerns
The run-up to the financial crisis saw increased and widespread acceptance of and reliance on internal models by both management and regulators. These proved defective in many ways; they were highly pro-cyclical, they took inadequate account of leverage and liquidity risks and, because there had not recently been crises, most did not reflect adequately the potential scale of tail risks inherently accepted. The complexity of the models however created a false sense of precision of risk metrics which allowed a build-up of directional and tail risk. As a result many risk models failed to be valid at precisely the point where they were most needed—with the realisation of tail events.
To address these issues the completeness and adequacy of trading book risk models has now been extensively re-assessed. Pro-cyclicality remains, but with the introduction of various incremental models for tail risk, capital standards for bank trading books have been strengthened by a multiple. Wider reforms in the capital markets, in particular a substantial increase in trading book capital requirements, will also reduce the quantum of trading assets and the associated capital risks carried by banks. In large part this is because higher capital and liquidity standards have rendered arbitrage trading activity that relied on leverage uneconomic.
Even with more comprehensive models, improved calibration and strengthened capital levels, there remains some residual danger that these have not been calibrated for all possible risks. For less sophisticated banks, for example, there is a residual risk that standardised capital requirements have been wrongly calibrated, and that their proprietary trading operations, albeit smaller, could create a risk of unexpected capital loss. Valuation, particularly in illiquid market conditions will remain an issue and be most difficult for the more complex instruments or structured products, in particular so called Level 3 assets, as defined for accounting purposes, where the inputs to valuation are not observable and the valuation has therefore a greater element of judgment.
Contagion
The creation of central counterparties for standardised derivatives and the severe capital penalties for using OTC derivatives will be an important structural mitigant against contagion from proprietary position-taking through derivatives. In addition, for the six largest banks in the UK, the ring-fencing proposals together with recovery and resolution planning including bail-in debt will protect systemically important retail and commercial banking activities from exposure to risks from inaccurate risk calibration or valuation of proprietary transactions. However, the current ring-fencing proposals do not address:
(a)
(b)
The proposals from the High Level Expert Group chaired by Erkki Liikanen have set out a different approach to addressing the risks arising from trading activities within banks by considering both the absolute size of these operations and their size relative to the remainder of the firm. Once a materiality threshold has been reached on both bases, a separate subsidiary would be required. This structural requirement is premised on an assessment of the extent to which risks from trading activity could damage the capacity of the bank to support its other activities. It also reflects a judgment that trading risks, of an appropriate size and properly capitalised, within a much larger balance sheet should give rise to no more prudential concerns than an equivalent risk profile arising from any other asset classes such as UK mortgages or SME loans which could also be subject to material fluctuations in value.
By combining proprietary trading activity with other markets activity, neither the UK nor the Liikanen proposals address the risks that contagion from proprietary trading could disrupt other important markets activities which support customers. In the UK proposals, relationships with financial institutions and market-making activities are required to be excluded from the ring-fenced bank and so like the Liikanen proposals they are specifically included alongside proprietary trading activities within a separate subsidiary.
But these markets activities are also systemically important. They are considered to be Critical Economic Functions by the FSA for resolution purposes and, in the case of market-making, the creation of a deep and liquid trading market is essential to the financing of the real economy through securitised funding.
We note that both the US and proposed French approaches to structural reform have recognised this by drawing the line for ring-fencing simply to separate all the activities which are established to support customers from any residual proprietary trading activities. This recognises, amongst other things, the importance of acceptable market-making to facilitate client funding and investing activity.
Standards
2. To what extent does the presence of proprietary trading activity alongside client-facing activity in banks create the potential for conflicts of interest and affect remuneration practices, culture and banking standards?
As noted in the preamble, whether in proprietary activities or in market making or other financial markets activity, banks trade as principal and therefore there are inherent conflicts which are addressed through formally regulating conduct—best execution standards for example—and through firms’ own conduct rules. The key structural protection arises from the firm’s own conduct rules bolstered by regulation around transparency, conduct, concepts of fairness (including addressing information asymmetry) and most importantly classification of which customers are regarded as sophisticated enough to deal on a principal basis with banks. Essentially the inherent conflicts are regarded as acceptable and manageable if there is clear disclosure of the nature of the potential conflicts and clients are regarded as sufficiently sophisticated and well informed to deal with the bank on a principal basis, understand all aspects of the transactions being undertaken and have alternative suppliers for these services.
There is clearly the potential for conflict when firms have both proprietary trading activities (which have no client relationship) and client-related activities, and the two operations have underlying positions which are in conflict; the most obvious example in the recent crisis was where some firms were positioning for a collapse of the US housing market while continuing to service and promote client demand for exposure to that market. Although firms have clear “chinese walls”—internal separation—to reduce potential conflicts of interest, there will inevitably be instances when a bank will profit from a proprietary transaction positioned on the other side of client facing activity.
It should be understood that, by definition, market activity requires two sides to a transaction and therefore differing views on market movements. As a consequence, banks and their counterparties will always have both same way and offsetting exposures at any point in time. The inherent conflicts have less behavioural implications when they arise from market making or risk management activity as opposed to proprietary trading; essentially this is because the firm is not seeking in these activities to benefit from directional moves and so has no incentive to attempt to influence pricing. Remuneration from market-making and the firm’s own risk management activities is also not structured around trading profits as would be the case in respect of proprietary trading.
Some of these issues were highlighted in the report of the US Senate’s Permanent Subcommittee on Investigations into Wall Street and the Financial Crisis: Anatomy of a Financial Collapse published on 13 April 2011. We would expect that firms which engage in proprietary trading and client-focused will have modified their activities in the light of these findings.
However, given that we do not undertake proprietary trading activities, HSBC does not feel it is in a good position to comment further.
3. How adequately do current conduct rules manage any potential for conflicts and harm to standards? To what extent could stronger conduct rules address any problems?
There is no reason why the current rules should be viewed as inadequate in principle, yet it is clear from recent experience there is little confidence that in practice they are being adequately monitored and enforced. In reality, no matter how strong the conduct rules are, there will always be concerns over their application particularly where a proprietary trading function is seen to benefit at the expense of clients.
4. Are there stronger grounds for concern about proprietary trading when it is conducted in a group which contains a ring-fenced bank than if it is conducted by a standalone wholesale bank? If so, why?
In terms of standards of conduct, it is difficult to see a material difference between proprietary trading conducted in a group which contains a ring-fenced bank (which will necessarily be in a separate subsidiary from the ring-fenced activities) and through a stand-alone wholesale bank. Again we note the US and proposed French solution is simply to prohibit proprietary trading absolutely in banks of any shape.
Practical Considerations
5. How, if at all, would prohibiting proprietary trading by banking groups affect the case for implementing a ring-fence? Would there be any practical benefits from implementing both such measures at the same time?
If the UK’s ring-fence proposal had been based solely upon the perceived risks to critical activities arising from proprietary trading, then the prohibition of such trading would have been the most appropriate parameter for the ring-fence. This would have been in line with the US approach and what is now being contemplated in France.
But this was not the case and the UK ring-fence draws a distinction between retail activities, where customers are less sophisticated in their ability to assess the position and stability of their bank and have fewer options for their financial needs, and wholesale activities, where customers do have the capacity to understand these risks and manage or switch their providers accordingly. The UK approach also is designed to offer a distinct resolution approach for the retail ring-fenced bank versus the non ring-fenced bank, with the former to be restructured using bail-in of unsecured uninsured creditors but leaving the option of liquidation open for the latter (although the practicalities of that distinction are questionable).
As an effective ring-fence should protect the retail activities, having gone down the ring-fence route, it is difficult to see what additional practical benefits would arise from a total prohibition on proprietary trading, unless there are now concerns about the systemic importance of the wholesale activities to be contained in the non ring-fenced bank. Any conclusion that this is the case would logically re-open a discussion about the positioning of the ring-fence so that all economically-critical activities, whether wholesale or retail, could be protected from proprietary trading. This is a discussion which is taking place in Europe but it was not the approach adopted by the Independent Commission on Banking (“ICB”).
If there is to be further consideration of the nature of the ring-fence, we believe that the Commission should also reflect on the increasing trend towards geographic subsidiarisation which has developed since the ICB took its evidence and produced its first report. We see a significant trend towards this in recent proposals from the US on the regulation of Foreign Banking Organisations and in the actions of our own FSA to restrict the authorisation of branches for non-EU banks. This will significantly change the nature of banks in the UK—at the time of the ICB’s interim report, we estimated that drawing a perimeter to exclude non-EEA assets would reduce the scale of the UK domestic banking sector by some 30% and so offer potentially considerable protection to UK financial stability from risks arising from non EEA operations.
In essence we are now seeing four distinct types of ring-fencing emerging:
the UK ICB proposals based on ring-fencing critical activities, drawing a distinction between retail and wholesale;
the US approach based on prohibiting defined proprietary trading;
the proposed French approach defining acceptable market making and risk management activities and prohibiting everything else; and
geographic ring-fencing in a variety of formats.
6. What powers does the regulator already have that could be used to prohibit banks from conducting proprietary trading? What are the constraints on any such powers?
We believe that if the supervisor has concerns about the risks involved in any proprietary trading operations, it has the ability and authority already effectively to prohibit these activities in an individual bank if so required. At the simplest level, this could be achieved by increasing the capital requirements under the bank’s Individual Capital Guidance (Basel Pillar 2) on the basis of protection against prudential risks, to levels which would effectively make the targeted proprietary trading activities non-viable. Additionally, supervisors have wide powers to constrain activities based upon their assessment of the capabilities and capacity of individual firms to control and manage the underlying risks. Finally, in the current environment, reaching out to the Board of the bank concerned would undoubtedly, at least in an HSBC context, be effective in curtailing activity of concern to the supervisory authorities; a Board would not lightly go against a regulatory concern over a trading operation.
7. What are the main challenges in defining proprietary trading, and how could these best be addressed?
The most useful definition which goes to the heart of where there are public policy issues is that “proprietary trading” is specific risk positions taken by banks solely in pursuit of their own profits and not undertaken to support clients or to hedge the bank’s own risk profile including prospective risks. Proprietary trading would typically be identified as activity undertaken by separately identified trading desks with identified capital, risk limits and specific proprietary trading goals in terms of profitability.
The main challenge is that there can be a lack of distinction between positions which are undertaken to support client activity or hedging and “proprietary trading”; this makes it difficult to set out clear technical and legal distinctions. Again as set out in the preamble, all banks trade as principal and therefore the price risks from market positions are identical irrespective of the purpose of the trade. The largest exposures to capital within banks typically arise from structural interest rate and foreign exchange positions rather than any trading activities and these risks are inherent in the core business model of lending long at market interest rates funded by short term deposits priced at administered interest rates, as well as deploying a portion of the capital base in markets other than the domestic base of the bank.
Even the simplest core banking activity exposes a bank to market risk as principal. Consider a bank making a fixed rate mortgage offer: it will hedge its interest rate exposure risk on those loans through interest rate swaps or by raising fixed rate funding from the market. It can do this in two ways:
(a)
(b)
Banks will choose which approach to use depending on their own circumstances (for example, do they have an existing source of funds or is additional funding required) but both have risks—there is a danger that the mortgage loans cannot be originated or the funding risks cannot be hedged. Most practitioners would not regard this as proprietary trading but, in both cases, since any gains or losses which occur during the transition period are taken by the bank to the profit and loss account, it might fall under some definitions of “proprietary trading”.
More generally, banks do not perfectly match the risks which they accumulate from their dealings with customers and from the risk mitigation contracts which they place in the market; partly this is infeasible and partly this is a management decision to run an inventory of open risk to facilitate market-making and to profit from knowledge gained from market flow information. For example, if there is a pattern of investment managers moving into Japanese equities, a foreign exchange desk is likely to hold a long Yen position in anticipation of the demand for settlement currency, in exactly the same way as retailers adjust their summer stock depending on projected weather patterns; the difference is the bank is marking its open currency (stock) position to market daily as opposed to the retailer marking down unsold stock in the end of season sale. The point being highlighted is that banks routinely have positions held as principal from which they may gain or lose but a practitioner would regard only a very limited and bespoke portion of these as “proprietary trading”.
The most difficult areas in attempting to identify proprietary trading are how to distinguish this from both client-driven activities such as market-making and risk management operations of the bank, hedging its own exposures including structural exposures. Essentially there are two possible routes to go down; first a very precise definition of activity that is regarded as proprietary trading—for example investment in hedge funds—with everything else not so regarded; or alternatively a definition of what is not proprietary trading based on purpose and intent of the underlying transactions with everything else regarded as proprietary trading.
In seeking to define proprietary trading, regulators and commentators have used a number of dimensions to separate the mainstream activities which enable banks to support customers and manage their own risks from “proprietary trading”. These include:
direct links to specific client activities;
links to overall bank risks arising from multi-client activities (ie hedging);
expected and actual investment durations;
relative and absolute scale of risk positions;
organisation (separation, desks); and
profitability (volatility of daily/weekly profits).
We believe that these dimensions can be used to create a good working definition of proprietary trading and, indeed, these will underlie the French approach to structural separation. Within this, there will be a definition of acceptable market-making and this has already been developed in a number of areas including the EU Short Selling Regulation which contains an exemption for market-making.
For the purposes of this response, a working definition of proprietary trading might be: “transacting in contracts, assets and securities not directly related to:
(a)
(b)
(c)
(d)
(e)
This is consistent with the underlying approach of the Volcker Rule in the US and the French proposals on separation published on 19 December 2012.
Overall Assessment
8. Given the factors above, how would you assess the case for:
(a) Including a prohibition on groups containing a ring-fenced bank from engaging in proprietary trading within the Banking Reform Bill
(b) Implementing or creating reserve powers for a proprietary trading ban through some other form
(c) Requiring that future reviews of the operation of the ring-fence have an explicit mandate to consider and report on the case for such a prohibition
(d) Using the Banking Reform Bill to give the regulator a reserve power to impose such a prohibition on individual banks if it concluded that it was necessary
We are concerned that adding at this stage in the process a prohibition on proprietary trading in groups containing a ring fence bank could lengthen the implementation timetable for the current ring-fencing proposals because of definitional challenges. This may not be warranted given that ring-fenced banks will already be prohibited from undertaking these activities.
We note also that the Commission has suggested that regulators should have a reserve power to require the full separation of the non-ring fenced bank and as this would necessarily include any proprietary trading activities within that bank it would be possible to distance the ring-fence bank further if thought desirable in the future.
If there is concern that the non ring-fenced bank has exposure to proprietary trading risks that could critically damage the important real economy activities that will reside in the non ring-fenced bank then there would be a case to prohibit such activities. In such circumstances we believe the French approach of defining what is not proprietary trading is the better solution to address the definitional challenge.
However, given that the ring-fence is intended to prevent contagion from non-core activities, any requirement for a prospective review to consider the possible case for a prohibition of proprietary trading should be part and parcel of a more general review of the effectiveness of ring-fencing, the appropriateness of the location of the ring fence, and the sustainability of critical business activities and infrastructure across both the ring-fenced and non ring-fenced banks.
Finally, we believe that if the supervisor has concerns about the risks involved in any proprietary trading operations, it can already with existing powers effectively prohibit these activities in an individual bank if so required.
15 January 2013