BIS Discusses Bank Capital & Deposit Insurance

The Bank for International Settlements has announced (via the Basel Committee on Banking Supervision):

that the level of capital in the banking system needs to be strengthened to raise its resilience to future episodes of economic and financial stress. This will be achieved by a combination of measures such as introducing standards to promote the build up of capital buffers that can be drawn down in periods of stress, strengthening the quality of bank capital, improving the risk coverage of the capital framework and introducing a non-risk based supplementary measure. Also, the regulatory minimum level of capital will be reviewed in 2010, taking into account the above and other relevant factors to arrive at a total level and quality of capital that is higher than the current Basel II framework. Strengthening the global capital framework in this manner will enhance confidence and lay the foundation for a more resilient banking system.

The Committee notes that current reactions in the market place regarding capital levels have been highly procyclical. It will not increase global minimum capital requirements during this period of economic and financial stress. Indeed, the Committee has earlier stated that capital buffers above the regulatory minimum are designed to absorb losses and support continued lending to the economy.

We can hope that OSFI signs on to the bit about increasing bank capital quality! The important issue being discussed is the “build up of capital buffers” – one of the issues in the current crisis is that bank capital as currently defined may be all very well and good in terms of protecting depositors when a bank is wound-up, but doesn’t help too much when a bank gets into trouble and needs to recapitalize. A system of surcharges based on asset growth would go a long way towards fixing this problem.

I will bet a nickel that the “non-risk based supplementary measure” is the leverage ratio (US nomenclature) / Assets-to-Capital Multiple (Canadian nomenclature).

I am very disappointed that there is no mention of the influence of bank size. A Megabank has so many layers of management that the Board’s Risk Committee – comprised, generally, of people who are appointed for their gender and/or connections, nothing to do with ability – has many, many layers of self-interested subordinates between it and the guts of the matter.

Their other announcement is a joint paper with the International Association of Deposit Insurers. The consultative document is open for comments – hear that, OSFI? Comments from affected parties! How revolutionary! – until May 15. I suspect that debate between Iceland and the UK will be highly entertaining, as briefly review on Guy Fawkes’ Day.

Update: I also note a recent speech by David Longworth, Deputy Governor of the Bank of Canada:

Now, the assumption that most market participants use the same risk-management systems based on short historical samples is very much an exaggeration. Some researchers, however, have argued that enough institutions follow very similar risk-management systems that the dynamics described above can happen, and indeed have happened, in the real world in response to sizable shocks.14 Moreover, in its Global Financial Stability Report issued in the second half of 2007, the International Monetary Fund concluded – based on simulations it carried out, which seemed realistic based on observed risk-management practices – that “seemingly prudent behavior by individual firms, reacting to similar market-risk systems, could serve to amplify market volatility in periods of stress beyond what would otherwise have occurred.”15 Observations and anecdotal information following the failure of Lehman Brothers suggest that this behaviour of firms was very important in amplifying price volatility in the autumn of 2008. Analysis of such behaviour strongly suggests the need for a macroprudential approach.

[Footnotes]14. See A. Persaud (previous footnote) and the Committee on the Global Financial System, “A Review of Financial Market Events in Autumn 1998” (CGFS Publications No.12, Bank for International Settlements, 1999). This latter text has a section (see page 14) on the over-reliance on quantitative tools.

15. International Monetary Fund, “Do Market Risk Management Techniques Amplify Systemic Risks?” in Global Financial Stability Report October 2007, 52-76.

Cliff risk due to similarity of trading techniques (that is, the “best practices” so beloved of bureaucracies) were last discussed on PrefBlog on March 12. It has been a worry for BoC for a long time.

Back to David Longworth:

Two main principles have been proposed. The first is that, in parallel with the probability of default on credit exposures on the banking book being calculated on a “through-the-cycle” basis, VaR for the trading book also be calculated on a through-the-cycle basis. One implication of this principle is that all historical data should be exploited to calculate the distribution of possible losses for a given asset or asset class. The second principle is that a “stress VaR” – a VaR calculated on the basis of assumed stress conditions – should be used, especially to consider the heightened correlation of losses across various assets or asset classes. It is well known that correlations among losses in categories of risky assets increase dramatically (sometimes approaching one), when the financial system is under great stress.

I have problems with the “through the cycle” approach. It throws out a lot of data; and should a firm become insolvent it doesn’t mean a lot to say ‘well, we’re solvent through the cycle, so trust me!’ There is a reason for cycles; recessions are nature’s way of telling us we’re doing something wrong. The problem should be attacked from the other end, focussing on capital.

The “stress VaR” is nothing more nor less than common sense.

One Response to “BIS Discusses Bank Capital & Deposit Insurance”

  1. […] PrefBlog Canadian Preferred Shares – Data and Discussion « BIS Discusses Bank Capital & Deposit Insurance […]

Leave a Reply

You must be logged in to post a comment.