It is clear from this analysis that there is a real challenge in defining an effective regulatory capital system that will provide effective protection against international systemic risk, particularly if it is to operate in a manner that is consistent across all systemically significant firms. In particular, the system will need to address the following three issues

• First, the approach will requires a common definition of risk that is sufficiently generic to enable it to be related to capital in a broadly consistent manner across different firms with very different business profiles.
• Secondly, the regulatory benchmark of what constitutes 'sufficient' or 'adequate' capital to protect against the particular risks which a firm faces needs to be clearly set out and obtain a sufficient level of credibility and plausibility to be broadly acceptable to regulated firms.
• Thirdly, in view of the discussion in the previous section about the relative importance of capital adequacy and liquidity management in protecting against systemic risk, any effective regulatory capital system will also need to recognise the importance of the interaction between capital management and liquidity management at financial firms.

- A common definition of risk

In considering the first of these issues, it is interesting to note that the Basel Committee has not itself attempted any definition of risk when outlining the case for a 'risk based' approach to setting capital adequacy standards. Nevertheless, since all risks ultimately impact the bottom line of the firm, it is relatively straightforward to define the risk that a financial firm faces. In the most general terms risk is the possibility that a firm's actual income will turn out to be lower than the expected value of its earnings. In most circumstances this shortfall may be disappointment to the shareholders, but should not be a serious concern to regulators since the firm will still be trading profitably. But the regulators are likely to be concerned about those cases where the shortfall could be sufficiently large to result in a loss which depletes the firm's equity and retained earnings or in the extreme case to exceed the shareholders' funds and result in the insolvency of the firm.

This provides us with the following generalised definition of risk

Risk_t = P_t (Earnings_t+1< E_t(Earnings_t+1))

where:

Risk_t is the measure of risk at time t
P_t is a measure of probability at time t
Earnings_t+1 is the actual observed earnings at time t+1
E_t(Earnings_t+1) is the expected value, measured at time t, of earnings at time t+1

from which we can derive those cases which we believe are of particular concern to regulators as follows

RRisk_t = P_t (Earnings_t+1< 0)

where:

RRisk_t is the measure of regulatory risk at time t

Since for the vast majority of firms we can assume that expected earnings are positive, we can see that RRisk_t is simply a sub set of our general definition of risk. Consequently anything which a firm does to reduce its exposure to risk will also necessarily reduce its exposure to the regulatory measure of risk.

The adoption of a common definition of risk of this type underlines the following two important points for the design of an effective regulatory capital system:

• in order to reduce regulatory risk, the approach to setting regulatory capital has to be as consistent as possible with the way that firms manage risk internally and in particular has to avoid perverse incentives; and

• any effective measure of risk has to incorporate assumptions about future earnings

- Defining an adequate level for regulatory capital

If we accept that a common regulatory capital requirement is an appropriate mechanism to help to protect against the risk of loss at a firm which might lead to a systemic problem, the next question that needs to be addressed is how should a regulator determine the appropriate quantum of capital that the firm needs in order to protect against this risk. In considering this issue it is important to recognise that in such a situation, capital fulfils two functions:

• it minimises the likelihood that the firm will itself be the immediate cause of a systemic problem; and

• it ensures that a firm which is faced with a systemic problem (whether or not of its own making) is well positioned to survive any crisis that may arise

It is clear from this that a minimum level of regulatory capital is appropriate in order to ensure that the firm is adequately protected against any reasonable scenario that might give rise to an unexpected or stress type loss. (As explained in Section 3 above, we assume here that expected losses are already provided for and do not require additional capital cover.) But in addition the regulatory system should seek to ensure that a firm which does encounter such a loss still has sufficient resources to survive the crisis without posing a threat to the system or a cost to the public sector during the workout period.

The need for a capital buffer of this type can be most clearly understood if we consider of the role which the central bank plays in providing liquidity support during a crisis. This issue is relevant because, in the event that it proves necessary for a central bank to take steps to provide liquidity to a firm during a systemic crisis, there needs to be a high level of assurance that the lending will remain solely for the purposes of liquidity and will not be used to underwrite the solvency of the firm. This in turn means that any firm to which the liquidity is provided by the central bank must be sufficiently solvent to be able to provide an adequate level of margin against any collateral which is provided to the central bank. Consequently, when establishing common minimum capital standards, regulators need to determine not only how to measure the amount of capital which a firm requires in order to survive a shock, but also what additional safety margin the firm should be required to hold over and above this amount in order to enable the firm to continue to fund its business following the crisis, whether from the market or from the central bank.

From this analysis, we can also see that the appropriate time horizon for setting regulatory capital requirements is comparatively short. The key question here is for how long should the regulators seek to maintain the solvency of a firm once it is clear that the firm is indeed facing a crisis and thus is unable to adjust its capital structure or risk profile adequately to address the problem. We would suggest that the correct time horizon here is probably no more than three years - based on a one year horizon for the crisis to develop and a further two years to allow any work-out to be organised and executed. This is significant because many firms' own capital measures use much longer time horizons of up to forty years. While a long horizon of this type may be appropriate when a Board of Directors is determining an appropriate capital level for their firm in order to maintain strategic flexibility and to meet shareholder and creditor concerns about the probability of insolvency over the longer term, we believe that this is not an appropriate horizon for regulators. The regulator's concern is how quickly could a serious problem emerge and how long would it take to in order to take quick corrective action.

(This difference in time horizon provides one explanation as to why we would expect a minimum regulatory capital requirement to be generally lower than the level of capitalisation voluntarily chosen by well managed and well capitalised firms)

- The interaction between capital and liquidity management

The foregoing analysis shows clearly the importance of the link between capital and liquidity management which we referred to in Section 3. Indeed, the primary consideration in determining whether a firm is able to avoid a systemic event is the extent to which the firm is itself managed in a way that will enable it to survive a liquidity problem, whether specific to the firm itself (as for example in the case of Salomons, Barings, Drexel or LTCM) or a more general market wide systemic liquidity crisis. Consequently, we believe that any regulatory system which is focused on protecting against systemic risk must concern itself at least as much with the adequacy of a firm's liquidity management arrangements as with the adequacy of a firm's capital. Indeed, in view of the central importance of effective liquidity management to the maintenance of systemic safety it is somewhat surprising that this issue does not feature explicitly in the Basel Committee proposals. Nevertheless, it fits naturally into the approach which the Committee proposes, since effective liquidity management procedures form an integral part of good risk management practices at firms and should therefore be subject to review under the supervisory pillar and should also be subject to effective market discipline under pillar three.

At the same time, however, it is important to recognise that a liquidity problem affecting a firm can itself affect the adequacy of the capital which a firm has available to survive a liquidity problem. This issue arises because, at a time of market stress, a firm's ability to raise funds to finance its business may be dependent on the market price of certain of its assets. If the price of these assets is uncertain, or dependent on the activities of the firm itself, which may be the case if the firm has a particular concentration of holdings in a particular asset class (as was the case for example with both Drexel and LTCM), then the need to liquidate the assets may precipitate a large fall in the value of those assets. This in turn will mean that the firm is able to raise significantly less funding for the rest of its business than it anticipated and will suffer a substantial depletion in its capital at the same time. Consequently, any measure of capital adequacy which seeks to assess the ability of a firm to survive a market stress event needs to build in a measure of liquidity adjustment which considers the potential impact on the firm of a forced sale of assets with particular attention being paid to any large or illiquid assets within the trading book.

(For further discussion of this issue see 'Improving Counterparty Risk Management Practices' - Counterparty Risk Management Policy Group June 1999)