Q: Are capital requirements for G‐SIBs an effective way of reducing systemic risk?
A: Higher capital requirements are necessary, but not sufficient, to reduce systemic risk.
Since the financial crisis banks have been required to hold significantly higher levels of capital to protect against the risk of firm failure.   Minimum capital requirements have been increased and global systemically important banks (G-SIBs) are generally required to hold a higher proportion of capital than other firms.  Firms must not only hold more capital than before, they are also required to hold a higher quality of capital.  A larger proportion of bank capital must be made up of common equity and some instruments that previously contributed towards capital are being phased out.  The consequence of this is that banks are more resilient; they are better able to withstand stress and less likely to fail than they were in the past.
The revised capital framework also addresses risks to the system as a whole: the requirement for a countercyclical buffer seeks to guard against the cyclical build-up of risk and means that banks may be required to hold additional capital specifically for the purpose of reducing systemic risk.  This is overseen by dedicated macroprudential authorities.   In the UK, for example, the Financial Policy Committee is explicitly charged with identifying, monitoring and taking action to remove or reduce systemic risks and it can direct the regulator to adjust specific marcoprudential tools for this purpose.
Moreover a non-risk-based leverage ratio will, when fully introduced, serve as a backstop to the risk-based capital requirements and help contain system-wide build-up of leverage.
In addition to enhanced capital requirements banks, as well as other financial firms, are expected to be better run. They must meet higher standards of governance and individuals are being held accountable for their decisions and actions to a much greater extent than was previously the case.
These reforms represent significant progress and we should not underestimate the scale of what has been achieved.  There has been a substantial amount of international work, not least through the Financial Stability Board (FSB), the Basel Committee on Banking Supervision (BCBS) and the European process to achieve a consensus on some fundamental and often difficult changes to the regulatory framework and this has taken persistence and determination. Individual countries have worked equally hard to implement – and in some cases build on and refine – international standards and rules in their domestic regimes.
Individual firms are demonstrably less likely to fail than they were in the past and authorities now have explicit mandates to address risks arising in the system as a whole as well as vulnerabilities in individual banks.
Nevertheless banks should be allowed to, and will continue to, fail.  The international standard setters have acknowledged this and the UK goes as far as to be explicit that it does not run a zero-failure regulatory regime^[1]http://www.bankofengland.co.uk/publications/Documents/praapproach/bankingappr1406.pdf.  It is accepted that banks will continue to fail from time to time and this is generally considered an ordinary and desirable feature of a market economy in which there is a healthy competition for business.
Rather than avoiding failure altogether the goal is that, if a firm does fail, it should do so in an orderly fashion: without excessive disruption to the financial system, without avoidable interruption to the critical economic functions that it provides and while ensuring that losses arising from failure are borne by the shareholders and creditors of the failed firm rather than the general public.
Achieving this will mean that financial stability is maintained – the widespread disruption that characterised the crisis in 2007/8 will be avoided.  Moreover, if it is feasible and credible that a firm can be resolved, the implicit state guarantee from which the largest banks have benefited in the past will be removed.  Risk will be appropriately priced and market discipline improved, further reducing the probability of a crisis.  Finally, orderly resolution can ensure that firms with inefficient or obsolete business models can exit the market and can make room for more efficient new challengers.
It is therefore vital that authorities have effective resolution regimes – and the FSB has set out the parameters for these in its Key Attributes^[2]http://www.financialstabilityboard.org/2014/10/r_141015/.  In Europe the Bank Recovery Resolution Directive, which is now in force and is being implemented across the EU, ensures that all Member States have appropriate tools and legal frameworks to deal with weak and failed banks.
What this means in relation to capital is that we need to focus not only on going concern regulatory capital requirements aimed at avoiding failure but also on requirements for gone concern loss absorbency – that is requirements for liabilities that can credibly and feasibly be used to absorb losses and recapitalise an institution in a resolution.  G-SIBs in particular must have sufficient total loss absorbing capacity – both going and gone concern capital – so as to be able to absorb losses prior to a failure, and to enable the authorities to effect a resolution following a failure.  Although there are a number of resolution tools available, the most likely approach for a G-SIB and other large banks is the application of the bail-in tool, where losses are absorbed by liabilities that are written down or converted into equity but the firm, or a successor entity, remains open for business.  Authorities would convert a sufficient amount of liabilities into equity to ensure that the firms can continue to meet solvency requirements and maintain market confidence.  This means that the firm must have the capacity not only to absorb pre-resolution losses, but also to meet recapitalisation needs. Following this initial stabilisation phase the G-SIB would be restructured and/or wound down in an orderly fashion.
To this end, the FSB has agreed in broad terms, and is in the process of finalising, a common international minimum standard for total loss absorbing capacity (TLAC) for G-SIBs.  In Europe the equivalent standard is a minimum requirement for own funds and eligible liabilities (MREL), which applies to all banks and not just to G-SIBs.  However both standards essentially aim to achieve the same thing and the expectation is that in Europe TLAC for GSIBs will be given effect through MREL.  MREL will be set on a firm-by-firm basis and can be set in a way that is consistent with the global minimum requirement for G-SIBs.
Q: What is TLAC?
A: TLAC is the FSB’s proposal for a common international minimum standard for total loss absorbing capacity for G-SIBs.
The FSB TLAC proposal is publicly available^[3]http://www.financialstabilityboard.org/wp-content/uploads/TLAC-Condoc-6-Nov-2014-FINAL.pdf – indeed the FSB have actively sought views on it through an open consultation process – but it is worth recalling the basic principles that underpin it.
First: firms must have sufficient loss absorbing and recapitalisation capacity available in resolution to allow resolution authorities to effect an orderly resolution and recapitalise the firm.  An orderly resolution is one that minimises the impact on financial stability, ensures the continuity of critical functions that the firm provides and avoids exposing taxpayers to loss. It needs to be credible – to a high degree of confidence – that this can be achieved.
Second: resolution authorities should determine a firm-specific Minimum TLAC requirement for each G-SIB that: a) is at least equal to a common Pillar 1 TLAC floor agreed by the FSB (see below); b) makes prudent assumptions about losses incurred prior to resolution and realised during the prudent valuation that informs resolution actions and c)  ensures that the entity (or entities) emerging from resolution will meet conditions for authorisation – including any consolidated capital requirements – and will be sufficiently well capitalised to command market confidence.
Third: given that G-SIBs operate in multiple jurisdictions, and to avoid disruptive fragmentation in the event of failure and facilitate cooperation between home and host authorities, host authorities must have confidence that there is sufficient loss absorbing and recapitalisation capacity available to subsidiaries in their jurisdictions with legal certainty about how losses and loss absorption will be allocated within a group at the point of resolution.
Fourth: exposing TLAC-eligible instruments to loss should not give rise to systemic risk or disruption to the provision of critical functions.  In particular authorities should place appropriate prudential restrictions on G-SIBs’ and other internationally active banks’ holdings of liabilities eligible to meet the TLAC requirement.
Fifth: liabilities that qualify as TLAC should be stable, long term debt claims that cannot be called at short or no notice, or equity capital.  This is necessary in order to provide comfort that TLAC liabilities will be available at the point of resolution.
Sixth: a breach or likely breach of TLAC should be treated as severely as a breach or likely breach of minimum capital requirements and addressed swiftly, again to ensure that sufficient loss absorbing capacity is available in resolution.  However regulatory capital buffers must be usable without entry into resolution.
Seventh: There must be clarity – to holders of TLAC and more broadly – about the order in which losses will be absorbed in resolution, which should be aligned with the insolvency creditor hierarchy.  This is also necessary to ensure that exposing TLAC-eligible instruments is legally enforceable and does not give rise to valid compensation claims.
Q: What does this mean that G-SIBs will have to do?
A: The proposed FSB standard sets out requirements in relation to the quantity and quality of TLAC that G-SIBs must hold, as well as in relation to the distribution of TLAC within a group and the disclosure of TLAC holdings. 
Quantity: TLAC will be calibrated as the higher of between 16% and 20% of risk-weighted assets or twice any Basel leverage requirement.  Existing Basel capital buffers continue to apply – they ‘sit on top’ of TLAC so that they remain usable. This means that banks that experience losses would initially only breach buffer requirements, which is associated with limited but well-defined consequences. Including buffers G-SIBs will, under the current proposal, therefore have to hold TLAC equivalent to 19.5% – 25% of RWAs.  The TLAC standard is a Pillar 1 minimum requirement but authorities will continue to set additional firm specific requirements and are expected to do so with a view to ensuring that the principles underpinning TLAC are met.
Quality and composition: The TLAC requirement can be satisfied by all regulatory capital instruments, as well as unsecured and uninsured liabilities with a residual maturity of more than one year that are readily loss-absorbing in resolution.  In order for debt liabilities to count towards TLAC they must be within the scope of statutory bail-in tools and be capable of being readily converted into equity.  This means that they must be subordinated to liabilities that are explicitly excluded from TLAC or bail-in (see below on subordination).  The key here is that TLAC must be easily usable in resolution in a manner which supports the principles outlined above.
Distribution and internal TLAC: How TLAC is distributed around a group will depend on how the group would be resolved.  However, losses may arise in different parts of the group, financial resources are not fungible in resolution and, ex-post, the group may not have incentives to voluntarily recapitalise a failed subsidiary.  While TLAC would only be issued externally form the legal entity that would formally enter resolution, losses may arise elsewhere in the group. The TLAC standard requires banks to maintain ‘internal TLAC’ – certain intra-group liabilities – that allow losses to be passed to the ‘resolution entity’ from wherever they arise.  This provides a pre-defined way to channel losses to the resolution entity and provides host supervisors with confidence that losses arising in their jurisdictions will be absorbed.  It also provides clarity on the creditor hierarchy and ensures that a complex group does not have to be resolved on an entity-by-entity basis.
The FSB proposal requires G-SIBs to pre-position TLAC on the balance sheet of all material subsidiaries to ensure that losses can be absorbed by the legal entity that would be put into resolution.  The amount required to be pre-positioned is 75-90% of the TLAC requirement that would be applicable to the material subsidiary if it were itself a resolution entity.
Disclosure: G-SIBs must disclose, at legal entity level, a) the amount, maturity and composition of TLAC maintained by each resolution entity and at each material subsidiary and b) the liabilities of each resolution entity that are pari passu or junior to TLAC – that is liabilities that sit at the same level as, or below, TLAC liabilities in the creditor hierarchy.
Disclosure of the creditor hierarchy for each legal entity allows investors to better assess the risks to which they are exposed by providing clarity on the order in which losses will be allocated both at the legal entity level and within the group.  This should reduce uncertainty enhance market discipline and minimise the shock caused by any surprises in a bail-in.
Subordination: TLAC liabilities must be subordinated to liabilities that are excluded from TLAC, on which it may not be possible to readily impose losses in resolution.  This means that TLAC liabilities will be exposed to loss before liabilities that are excluded from TLAC.  The aim here is to avoid having to depart from the insolvency creditor hierarchy in resolution, which may give rise to legal risks and valid compensation claims on the grounds that resolution would treat some creditors worse than an insolvency would (we refer to this as the   ‘No Creditor Worse Off than in Insolvency’ safeguard).
TLAC does not need to be subordinated to liabilities that are ineligible for – but not excluded from –TLAC, for example liabilities that do not meet the maturity requirement.  This means that ineligible liabilities may be exposed to loss before, at the same time, or after TLAC liabilities – depending on where they fall in the creditor hierarchy.
Subordination increases clarity on the order in which losses will be allocated in resolution.  But it is important to be crystal clear: liabilities that do not count towards TLAC – either because they are explicitly excluded, or because they are ineligible to count, may still be exposed to loss in accordance with the creditor hierarchy.
There are three routes to subordination:

1. Contractual: subordination is specified in the terms of the TLAC liability’s contract. This is relatively straightforward to arrange and can be done by the parties to the contract, without intervention from public authorities.
2. Statutory: subordination is specified in law. This requires national governments to set out the terms of the subordination, and the liabilities to which it applies, in law.  The EU Bank Recovery and Resolution Directive, for example, specifies that deposits covered by EU deposit guarantee schemes are ‘super-preferred’ in the creditor hierarchy.  Similarly, for the purpose of TLAC, individual governments could specify that certain liabilities are generally subordinated to others, for example operating liabilities.
3. Structural: subordination is achieved through the structure of the bank. For example, TLAC liabilities issued by a ‘clean’ holding company or intermediate holding company (that is, does not have operational activities and does not issue liabilities that are excluded from TLAC) will be subordinated.  In a resolution losses will flow up to the holding company and be absorbed by the liabilities issued from it.  This is perhaps the most straightforward form of subordination in the long term but may take some time to achieve and can involve substantial and complex changes to how firms are organised.

Subordination can therefore be achieved in a variety of ways. The method used is less important than the end objective and may change over time.
Q: How, in practical terms, does TLAC relate to the resolution of a G-SIB? 
A: TLAC makes it feasible and credible to resolve a G-SIB.
The FSB TLAC agreement will provide the parameters within which TLAC is set.  But it is important to remember that the authorities’ resolution plan for the firm will drive the detail.  That said, bail-in is the only feasible resolution option for a G-SIB.  It is not credible to think that a G-SIB could be dismantled over a resolution weekend without a destabilising disruption to critical functions.  Finding a private sector purchaser capable of taking on the business – in whole or in parts – is likely to be even more difficult.
There is more than one way to effect a bail-in but, however it is applied, the bail in tool allows the losses of a failed firm to be absorbed and the firm (or its successor) to be recapitalised by writing down and/or converting into equity the claims of shareholders and uninsured and unsecured creditors in a manner that respects the hierarchy of claims in insolvency.
Effectively bail-in protects a firm’s critical functions.  It buys the time to stabilise the firm before an orderly reorganisation which may include winding down or selling parts of the failing firm.  The orderly reorganisation point is important and the FSB is explicit that the underlying causes of the firm’s failure must be addressed.
Moreover it is not enough simply to absorb losses and recapitalise the failed firm – the firm must be a recapitalised to a level that ensures that the firm complies with post-resolution conditions for authorisation and sustains market confidence.   One proxy for market confidence is access to market funding – but of course it is difficult to say with certainty what level of recapitalisation needs to be achieved before market funding is available.  What we can say is that following a bail-in it is likely that a firm will be required to hold capital in excess of regulatory capital requirements in order to restore access market funding.
The proposed TLAC framework makes it feasible and credible to conduct a bail-in on a G-SIB.  It ensures that firms have sufficient loss absorbing and recapitalisation capacity beyond going concern capital requirements available in the right place and in the right form at the point of resolution.  The proposed framework also ensures that TLAC is usable, both legally and practically.
Of course the process for setting TLAC is one part – though an essential part – of the resolution planning and resolvability assessment process which – as its name suggests – considers firm resolution and resolvability in its entirety, including:

1. the options available for reorganising the firm’s critical functions in resolution and whether they are to continue within the firm, to be transferred elsewhere in the market or to be wound down;
2. whether the firm should make ex ante changes to the way they are organised so as to remove impediments to resolution and guarantee that options to separate critical functions in resolution are credible.

Decisions on these wider resolvability issues sit alongside the TLAC framework and allow resolution authorities to ensure that the loss absorption and recapitalisation resources that a firm holds align with the resolution strategy for preserving its critical functions.  The authorities will specify not only how much TLAC firms must hold, but also where it should be held within an inevitably complex group, and the form in which it must be held.  The process involves significant cooperation between the home and host authorities which have a shared interest in planning for the resolution of the firm and significant dialogue with the firm itself.
Q: What are the costs and benefits of TLAC and are criticisms of TLAC justified?
A: The FSB is currently looking at the projected costs and benefits of TLAC, and the results of this will inform the final TLAC standard, but the outlook is promising. 
It is no surprise that there has been a vigorous debate about the costs and consequences of TLAC.  It marks a major change in the regulatory framework and brings firms’ liability structures into sharp focus.
It is nonetheless important to answer some of the criticisms that have been levelled at the proposed framework.  One is that the TLAC standard implies that the Basel reforms are inadequate –that it would be more straightforward to increase the Basel capital requirements than to design a new framework.  The TLAC standard in fact reinforces and complements the capital requirements agreed by the BCBS in the Basel III package.   Basel III requires banks to hold regulatory capital to absorb losses arising from financial and economic stress, whatever the source.  TLAC seeks to ensure that G-SIBs can fail, and that in the event of such a failure, firms have sufficient additional loss absorbing capacity, available in the right place and in the right form, to allow the firm or its successor entity to be recapitalised back to Basel III levels without disruption to the critical economic functions that the firm provides.
A second criticism is that TLAC concentrates risk and that banks will simply hold each other’s TLAC eligible liabilities.  This is not the case: under rules being finalised, GSIB holdings of other GSIB’s TLAC will be deducted from their own TLAC or regulatory capital.   This is designed to prevent or discourage other banks from holding TLAC-eligible debt and will limit the contagion effects of imposing losses on TLAC in a resolution. The treatment of TLAC holdings by other banks remains under review by the BCBS.
Critics have also suggested that the prospect of bail-in will lead to a ‘buyers’ strike’ – meaning that there will be a limited uptake of the TLAC-eligible liabilities issued by firms.   This view overlooks the benefits of the clarity that the TLAC proposals provide as well as their effects on the pricing of risk.  The TLAC framework provides ex ante clarity on the liabilities that will be exposed to loss in resolution, and on the order in which they will be exposed to that loss (i.e. the creditor hierarchy).  This in turn ensures that the risk that holders of TLAC liabilities are exposed to is properly priced.  And the simple fact is that there is no current evidence to support the notion that there will be a buyers’ strike.  UK G-SIBs have recently been able to issue TLAC-eligible liabilities at prices that were not materially higher than the price of their existing wholesale funding (see below).
Turning to the costs associated with TLAC, the FSB is currently looking at the projected costs of TLAC in great detail, in advance of the standard being finalised.  Early indications from the market suggest that the cost of TLAC will be manageable.
For banks with holding company structures, restructuring existing wholesale debt to become TLAC-eligible (by migrating it to the holding company) is likely to increase funding spreads by around 50bps, based on current yields.  The expectation is that this will have very limited effects on the average cost of credit to the real economy.
For banks that do not have holding company structures, and therefore have to issue contractually subordinated debt, current yields would suggest that the costs may be somewhat higher. However the expectation is that the pricing of existing debt instruments will change as significant new layers of subordinated debt reduce the riskiness of existing senior and subordinated debt. Since the different forms of subordination are economically equivalent, the long-run impact should be comparable to that for banks with holding company structures.  That is to say, it should be relatively benign.
Although some observers worry that banks’ traditional ability to transform illiquid and risky assets into liquid and safe liabilities (such as demand deposits or short‐term wholesale funding) may be affected by the requirement for TLAC, this is not borne out by currently available evidence.  In practice most G-SIBs will be able to satisfy the TLAC requirement by restructuring existing long-term wholesale debt to become TLAC-eligible. G-SIBs have significant amounts of non-deposit liabilities that can be converted into TLAC without restricting a bank’s ability to engage in maturity transformation.
As for the benefits of TLAC, these are more difficult to quantify since they depend largely on the counterfactual of how a future G-SIB failure would be managed in the absence of TLAC.
However, comparing a bail-in to a bail-out counterfactual (which was the way in which G-SIB failures have been historically handled), there are two key benefits of bail-in.
First, TLAC insulates sovereign balance sheets and ensures that, instead of being absorbed by governments, losses are borne by holders of bank debt.  Insulating the sovereign balance sheet helps to ensure that the level of public debt remains manageable, and there is no need for extensive fiscal consolidation in the aftermath of a bail-out, and it may leave governments with more scope to provide targeted fiscal stimuli in a crisis.
Moreover, although some critics fear that imposing losses on holders of bank debt may give rise to a bail-in ‘shock’, existing evidence suggests that the impact of exposing individuals to financial wealth shocks is limited, since holders of financial wealth tend to be able to bear the loss without significant changes to their spending patterns.
Second, and perhaps more important, is that the existence of credible resolution framework, and institution-specific resolution plans  which include adequate levels of TLAC, remove the perceived state guarantee from which G-SIBs have previously benefitted and make banks’ funding costs more sensitive to risk and therefore appropriately priced,
This:

1. reduces riskiness in the system as a whole: there is convincing evidence^[4]See for example. Acharya, Anginer and Warburton (2014), Afonso, Santos and Traina (2014) and Marques, Correa and Spriza (2013) that perceived government guarantees incentivise banks to take larger risks; when these are removed, and bail-in is credible, risk is more accurately priced. This makes funding risky activities more costly – so fewer are undertaken;
2. reduces the probability of failure in individual firms: as outlined above, firms take fewer risks – on an individual as well as an aggregate basis.
3. smoothes the credit cycle: banks expand their balance sheets less aggressively in a boom, potentially smoothing the credit cycle
4. reallocates funding to real economy lending: while higher funding costs may affect all business lines of a bank, this effect will be strongest for the most risky businesses. This should lead to a re-allocation of funding towards safer real-economy lending.

So although the cost benefit analysis is not yet complete – and of course the FSB is still finalising its proposed framework – the emerging evidence supports the view that the costs of TLAC will be manageable.  Conversely the benefits – of financial stability, of properly priced risk and of freeing up sovereign balance sheets are significant.  Requiring firms to hold TLAC represents a major step forward in the effort to solve too big to fail.

---

References
Acharya, V., Anginer, D. and Warburton, A. J., 2014. The End of Market Discipline? Investor Expectations of Implicit State Guarantees. NYU Working Paper.
Afonso, G., Santos, J. and Traina, J., 2014. Do “Too Big To Fail” Banks Take on More Risk?. Working paper, Federal Reserve Bank of New York.
Brandao, M. L., Correa, R. and Sapriza, H., 2013. International Evidence on Government Support and Risk Taking in the Banking Sector. IMF Working Paper.