29 February 2012

VoxEU: Basel regulation needs to be rethought in the age of derivatives (Parts 1 & 2)

These columns by Paul E Atkinson and Adrian Blundell-Wignall aim to bring banking regulation back to the centre of attention. They argue that the Basel III regulations currently being proposed are already desperately out of date. Part 2 outlines how to put this right.

PART 1

Helpful and welcome as recent moves have been, a more fundamental rethink of the Basel framework for determining minimum capital requirements for banks is needed.

Basel’s main flaws

The central problem is that banks have almost unlimited scope to arbitrage the system by reallocating portfolios away from assets with high risk weights to assets with low risk weights, not least by trading derivatives. Given their powerful incentive to save on capital costs, they use this scope abundantly. New capital charges (e.g. the surcharge for Globally Systemically Important Banks) and stricter calibration (i.e. higher required ratios based on ‘risk weighted assets’) just encourage more of the same.

Bank responses to Basel incentives lead to three major problems:

• Capital charges are ‘portfolio invariant’, i.e. while they depend on the borrower’s characteristics and the economic environment they are not influenced by whatever else is in the portfolio. There is no role for diversification in determining minimum capital requirements. This contradicts everything we have ever learned about managing portfolio risk.
• Nevertheless, the risk weights, which act as a system of regulatory taxes and subsidies, create a bias against diversification by encouraging concentration in asset classes favoured by the regulatory system. Favoured classes have been residential real estate, sovereign debt, and interbank claims, ie system interconnectedness. It is not coincidental that these have been at the heart of the crises we have been living through since 2007.
• Since minimum capital requirements can be arbitraged downward with little effective limit, the system allows far too much leverage.

Basel and derivatives

The distortions caused by the system are often obscured by its complexity and opacity, especially as regards derivatives. Consider, as an example, a Basel III innovation to deal with unexpected counterparty credit risk losses on derivatives: the charge for Credit Valuation Adjustment (CVA), ie unrealised losses marked-to-market which the Basel Committee estimates accounted for two thirds of the total during the crisis (ie double actual defaults). The base for calculating this charge allows gross derivative positions vis-à-vis each counterparty to be netted out, i.e. bilaterally, and the charge is additive across counterparties. This may be superficially reasonable but has several unhelpful consequences.

• For large universal banks, active in highly concentrated derivative markets, it seriously underestimates the underlying exposures. Bilateral netting allows the overwhelming bulk of derivative exposures, often of the order of 90 per cent or more, to be ignored for purposes of calculating the CVA charge. But many, if not most, positions may be valued very differently if the state of the world changes. Large losses last year by the MS/MUFG joint venture in Japan following an unexpected 45 basis point move in long-term interest rates there appear to be a case in point.
• Making the CVA charge additive across netted bilateral positions rewards counterparty concentration, limiting competition in derivatives trading, since concentration effectively increases the pools within which netting is permitted.
• All this operates to minimise the CVA charge.

Implications

The overall result is a vast, poorly diversified, highly interconnected banking system supported by far too small a capital base. It has little resilience or capacity to cope with adjustment so local problems too easily become systemic. Collapsing subprime real estate prices in California and Florida and excessive supply of Greek government bonds may be major problems for borrowers and lenders directly affected, but neither by themselves could create a global crisis with a resilient, well-capitalised banking system.

The Basel system should clearly be replaced with one whose parameters cannot be arbitraged by portfolio reallocation and derivative activity. 

Part 1 - full article

---

PART 2:

How can the system be put on a sounder foundation? By revising and shifting the balance of the three ‘Pillars’ of the Basel system:

• drastically simplify Pillar 1 (rules) while making the rules more effective;
• be realistic about Pillar 2 (supervision); and
• rely more on Pillar 3 (market discipline).

Limit leverage and separate low-risk from high-risk activities

As regards Pillar 1, the risk-weight system for calculating capital charges should be replaced by a simple leverage ratio, i.e. an upper limit on how many times a bank’s equity (net of any goodwill) can be leveraged. Essential to making this effective is to include gross derivative positions consistent with IFRS accounting as part of the asset base requiring equity backing, i.e. without the netting permitted under GAAP reporting and used for purposes of calculating capital charges under existing Basel rules.

Given the size and interconnectedness of the large universal banks, their main activities should be separated from each other in some way to limit the exposure of the entire equity base to heavy losses in any single activity, including trading derivatives. In particular, traditional commercial banking focused on domestic markets needs to be insulated from capital market activities where prices are determined in global markets. In the United States, the Glass-Steagall Act (until its repeal in 1999) and the recently passed Volcker rule have aimed to achieve this but at the cost of restricting competition. A better approach is to allow banks to engage in activities as they like but each within a separate subsidiary structure. Each subsidiary should have its own equity base and the full group should be structured as a non-operating holding company (NOHC), along Australian lines or as proposed in the United Kingdom by the Vickers Report, with firm firewalls that subject transfers of cash or other assets between subsidiaries to regulatory approval. This would allow explicit guarantees such as deposit insurance and government support mechanisms to be confined to certain activities, such as retail banking, without extending them to others, such as derivatives trading. This cannot entirely eliminate a perceived, or implicit, guarantee for such activities. But it would work to eliminate cross-subsidisation by other activities and increase the risk that a subsidiary at least might be allowed to fail. This would force it to stand on its own merits and probably raise the cost of capital to a level appropriately reflecting the risk involved in the activity.

Banks are likely to argue against the NOHC proposal on the grounds that it would require inefficient use of capital. Since more capital is desirable, this should be regarded as a powerful argument in its favour.

In contrast, the new Basel III rules relating to liquidity management serve little purpose. A sound capital adequacy framework and resolution regimes to ensure that problems are addressed while banks are still solvent should suffice to allow bank management to take responsibility for its own liquidity issues.

Be realistic about supervision

As regards Pillar 2, it is important to recognise the limits to what supervisors can achieve. They can enforce sensible rules, insist on honest accounting and transparent reporting, and attack fraud and corruption. They should aim to perform these tasks well. But we should not pretend that they can run banks. The constraints are overwhelming: their resources are too limited to process the information needed to second-guess management of mega-banks; they operate as bureaucrats within a political environment which they must respect; salary differences with the people they are supervising are often stark and invite capture; they are detached from the markets banks are operating in; and while banks will provide whatever information they are asked for, supervisors do not necessarily know what to ask for. Finally, their track record is poor – see for example Northern Rock and the SEC’s Consolidated Supervised Entities programme.

Reduce bank interconnectedness to allow greater market discipline

The effectiveness of Pillar 3 requires a credible willingness to allow large creditors, who provide a much larger fraction of banks’ total funding than do shareholders, to lose money more frequently when they make mistakes. The key to this is a reduction in the interconnectedness of the major universal banks and other large players, which results from reliance on wholesale funding and large-scale derivatives trading. This must be brought down to a point where problems do not contaminate under-capitalised counterparties and thereby become systemic. The leverage ratio suggested above would be a good first step in this direction since it would eliminate the existing privileged regulatory status of interbank claims and derivative activity.

But this may not suffice. Since common sense dictates that the authorities will support the system in the event of problems, the large universal banks are widely perceived as too-big-to-fail and benefit from an ‘implicit guarantee’ that is difficult to disavow. As a result these banks enjoy a competitive advantage that works to entrench their position even further. An additional useful step would be to impose a fee, which would have to be agreed internationally, on interbank activity. This would work to increase the cost of such activity and thereby discourage it. In the case of non-derivative funding activities, this could be structured as deposit insurance, with the implicit guarantees becoming explicit and the fee being accumulated in a fund to cover future calls on the guarantee. In the case of derivative activity the fee, which could similarly be used for a fund to insure the financial system, would be tied to the potential full exposure involved in the derivative commitment where a practical basis for measuring this exposure can be devised. Otherwise it would have to be based on transaction values, i.e. would resemble a Tobin tax.

Part 2 - full article

© VoxEU.org