1. Exchange rates are volatile and open short or long currency positions can lead to sizeable losses. Capital thus needs to be held to cover such possibilities. This section describes proposed capital requirements on open foreign exchange positions. It does not propose supervisory requirements for defining and limiting undue concentrations of currency positions or counterparty exposures. which national supervisors are nonetheless encouraged to address.
2. There are three distinct tasks in the process of setting capital requirements on banks' open currency positions. The first is to measure the exposure in a single currency position. The second is to measure the risks inherent in a bank's mix of long and short positions in different currencies. The third is to determine an appropriate capital charge.
I. Measuring the exposure in a single currency
3. The bank's net open currency position in each currency would be calculated by summing the following elements:
- the net spot position (i.e. all asset items less all liability items, including accrued interest, denominated in the Currency
- in question);
- the net forward position (i.e. all amounts to be received less all amounts to be paid under forward foreign exchange transactions, including currency futures and the principal on currency swaps not included in the spot position);
- guarantees (and similar instruments) that are certain to be called and are likely to be irrecoverable;
- net future income/expenses not yet accrued but already fully hedged (at the discretion of the reporting institution);
- the net delta (or delta-based) equivalent of the total book of foreign currency options; 26
- depending on particular accounting conventions in different countries, any other item representing a profit or loss in foreign currencies.
4. Trading by banks in gold and precious metals is often regarded as an extension of foreign exchange trading. However, the price volatility of precious metals is significantly greater than most exchange rates and closer to that of other commodities. If specific capital requirements were to be developed to cover commodity positions, some or all precious metals could be included. For the time being and in the absence of such requirements it is proposed that precious metal positions be included within the scope of foreign exchange capital requirements. Where this is done, long positions presently subject to the existing credit risk related requirements in the Basle Accord would cease to be so .
5. Positions in composite currencies, such as the ECU, would need to be separately reported but, for measuring banks' open positions, may be either treated as a currency in their own right or split into their component parts on a consistent basis.
6. Four aspects call for more specific comment: the treatment of interest, other income and expenses; the measurement of forward currency positions; the treatment of currency options; and the treatment of "structural" positions.
(a) The treatment of interest. other income and expense
7. Interest accrued (i.e. earned but not yet received) would be included as a position. Accrued expenses would also be included. Unearned but expected future interest ant anticipated expenses would usually be excluded unless the amounts are certain and banks have taken the opportunity to hedge them. If reporting institutions include future income/expenses they should do so on a consistent basis, and not be permitted to select only those expected future flows which reduce their position.
8. Forward currency positions would normally be valued at current spot market exchange rates. Using forward exchange rates would result in the measured positions reflecting current interest rate differentials, which would normally be taken into account in measuring a bank's interest rate exposure.
9. For some banks an alternative method would be to discount forward positions to net present value. This recognises that cash flows occurring at different future dates have different values if measured in terms of their present value to the bank. Where this is the basis of the bank's normal management accounting, using net present values would be an acceptable way of measuring currency positions for supervisory purposes.
(c) The treatment of Currency options
10. The most difficult instruments to fold into a risk measurement system are options. The main complication is that the price of an option does not move in a one-for-one relationship with the spot rate of the deliverable currency, since an option's value is a complex function of the spot rate of the underlying currency, its volatility, interest rate differentials, the strike price and the option's remaining term.
11. Nevertheless, the purchase of an option can be a useful means of hedging other positions, since it provides the holder with a form of insurance, the cost of which is represented by the option premium. For institutions which do not trade a portfolio of options but which find it convenient to use purchase options as hedges, it is proposed that both the options and the hedged positions would be carved out of the foreign exchange book ant that the following treatment would be applied:
- a long position (spot or forward) hedged by a long put; or a short position hedged by a long call, the capital charge would be 8% of the position being hedged less the amount by which the option is in the money (if any); 27
- where a long call/long put is held outright (i.e. not held as a hedge), the capital charge would be the lesser of (l) 8% of the market value of the underlying and (2) the market value of the option.
12. For institutions which are actively trading options and will therefore wish to manage their risk on a portfolio basis, a measure in general use is the net delta (or delta-based) equivalent of the total book of foreign currency options. However, delta hedging is effective only for small changes in spot rates and there are other variables, notably volatility, not captured in the delta. It is proposed that the net delta value will be used as the measure of exposure for major option players, the deltas being calculated according to an exchange model or internal pricing model approved by the supervisor. In addition, managements would be expected to apply additional internal safeguards to deal with the other option risks, for instance by simulating the performance of portfolios over ranges of possible levels of spot rates, volatilities ant interest rate differentials.
13. National supervisors would be ready to consider more advanced approaches for capturing the risks in option positions and related spot or forward positions which sophisticated trading institutions may wish to propose. The criteria for the acceptance of such methods as alternatives to the delta measure would be that they should provide a more reliable estimate of possible loss according to parameters set by the regulators, and that they allow supervisory oversight in a cost-effective manner. For example, these parameters could be designed to cover the highest loss in the bank's foreign exchange options (and related spot or forward positions) portfolio assuming certain changes (on a currency pair by currency pair basis) in foreign exchange rates and implied volatilities. Comments and suggestions on what such parameters should be would be welcomed.
(d) The treatment of structural positions
14. A matched currency position will protect a bank against loss from movements in exchange rates. but will not necessarily protect its capital adequacy ratio. If a bank has its capital denominated in its domestic currency and has a portfolio of foreign currency assets and liabilities that is completely matched, its capital/asset ratio will fall if the domestic currency depreciates. By running a short position in the domestic currency the bank can protect its capital adequacy ratio, although the position would lead to a loss if the domestic currency were to appreciate.
15. It is proposed that supervisors should be free to allow banks to protect their capital adequacy ratio in this way. Thus, any positions which a bank has deliberately taken in order to hedge partially or totally against the adverse effect of the exchange rate on its capital ratio could be excluded from the calculation of net open currency positions, subject to each of the following conditions being met:
- such positions would need to be of a "structural", i.e. of a non-dealing, nature (the precise definition to be set by national supervisors according to national accounting standards and practices);
- the national supervisor would need to be satisfied that the "structural" position excluded does no more than protect the bank's capital adequacy ratio;
- any exclusion of the position would need to be applied consistently, with the treatment of the hedge remaining the same for the life of the assets or other items.
16. The same treatment could be applied to positions related to items that are deducted from a bank's capital when calculating its capital base.
II. Measuring the foreign exchange risk in a portfolio of foreign currency positions and setting the capital requirement
17. Any risk measurement framework involves a trade-off between simplicity ant accuracy. To take account of correlation5 between exchange rate relationships and their differing volatilities would require a complex measure, which might be unduly burdensome. However, a simpler measure based on an assumption about the composition of an average portfolio of currency assets and liabilities may require a conservative bias if portfolios of higher than average risk are not to be undercharged.
18. It is proposed to resolve this by allowing a choice between two alternative measures at supervisory discretion; a "shorthand" method; and a more complex "simulation" method, in which capital requirements would be determined by simulating the likely outcome from holding the bank's actual portfolio. The simulation method would only be available to banks which can satisfy their national supervisors of the adequacy of their measurement and control systems and of their access to the necessary data. It is intended that the two methods should deliver broadly equivalent capital requirements for a well diversified portfolio of foreign exchange positions of average riskiness.
A. The shorthand method
19. Under this method the nominal amount 28 of the net position in each currency and of the net position in each precious metal would be converted at spot rates into the reporting currency. 29 The net open position would be measured by aggregating:
- the sum of the short positions or the sum of the long positions, whichever is the greater; 30 plus
- the total of each net position (short or long) in any precious metal, regardless of sign.
Annex 9 shows a simplified example of the shorthand measure.
20. The shorthand method assumes some, but not perfect, correlation between the movements of different exchange rates. Cross currency position taking would neither be unduly penalised nor given undue preference. However, the method would not recognise cross-hedging in the case of precious metals, the price volatility of which seems to justify a more conservative treatment.
21. The Committee proposes a capital requirement of 8% of the net open position as calculated above. This would ensure an adequate level of protection against losses for most portfolios, and be broadly in line with that proposed for interest rate and equity position risks in average portfolios.
22. The Committee sees merit in permitting a "de minimis" exemption for banks with negligible business in foreign currency ant which do not take foreign exchange positions for their own account. Accordingly, national supervisors would have discretion to exempt a bank from capital requirements on its foreign exchange positions provided that:
- its foreign currency business (defined as the greater of the sum of its gross longs and the sum of its gross shorts, including precious metals) does not exceed 100% of its capital as defined in the 1988 Accord; and
- its net open position as defined in paragraph 18 above does not exceed 2% of its capital. 31 Comments are invited as to whether the increase in a bank's capital charge which would result from its crossing of one of the two thresholds is significant enough to raise problems and, if so, on possible ways to alleviate them.
B. The simulation method
23. The simulation method is conceptually simple: the actual exchange rates experienced in a defined past observation period would be used to revalue the bank's present foreign exchange 32 positions and - from those revaluations - to calculate "simulated" profits and/or losses which would have arisen if those positions had remained fixed for a defined holding period. The capital requirement would be set in relation to the worst or near to the worst simulated loss which would have arisen during that period.
24. In practice, however, it is a demanding exercise. A continuous run of exchange rates over the period for all the currencies in which significant positions are held would be needed. The consistency, of data used by different banks would be important and each national supervisor would need to monitor closely the data used in the simulation process. This could be achieved either by defining a common and precise specification of the rates to be applied, leaving the banks with the responsibility of collating the actual rates, or by providing them with exchange rates data series.
25. Policy decisions would have to be taken on four parameters if the simulation method were to be used as a measure of risk:
- the period for which a bank is assumed to hold a position, and during which losses could accumulate;
- the period of observation of exchange rates necessary to be sure of capturing sufficient evidence of currency volatility;
- the level of confidence required in measuring the risk for the purpose of setting a capital requirement;
- the scaling factor used to set the toughness of a capital charge, depending on what is seen as an appropriate buffer against possible losses.
26. The choice of holding period depends on the speed with which banks can be expected to close out loss-making positions. Banks may argue that they are able to cover positions within a single day, but in periods of high exchange rate volatility and thin markets it may well not be possible to close sizeable positions so quickly. Moreover, banks may decide to run loss- making positions in the expectation that they will become profitable. Since exchange rates often move in one direction for some time, the longer the holding period the larger the losses could be. A holding period of two weeks (i.e. ten working days) is proposed. It would be rolled on daily, so that with each succeeding day the holding period would cease to include the earliest day included in the last holding period.
27. The next step is to determine the length of the observation period which adequately captures relationships between currencies. This depends on how confident one is that future currency volatility will mirror that exhibited in the recent rather than more distant past. A five-year period has been chosen as sufficiently long to avoid over-reliance on recent exchange rate movements while still being practical and not too burdensome to run. It is less necessary that the observation period rolls on daily, but at a minimum it should be updated at regular reporting dates. There would thus be some 1,300 observations for the value of each currency position (i.e. 5 working days x 52 weeks x 5 years).
28. It is then necessary to identify the level of confidence required in measuring the risk for the purposes of setting capital requirements. This could be the worst loss, or be expressed in terms of loss quantiles, for example the level which includes 95% of the hypothetical losses that would have arisen from the bank's current set of open positions. It is important not to confuse the required level of confidence in measuring the risk with the toughness of the capital requirement. They are quite independent.
29. Setting the required leve1 of confidence as the worst loss appears simple and prudential, appealing, but it has conceptual and practical drawbacks. From a conceptual point of view. in placing complete weight on the most extreme situation. it is very sensitive to the choice of observation period. In emphasising the importance of the abnormal event, it would also maximise the problem of the asymmetry of result that the simulation method could deliver depending on whether a short or a long position is held in one particular currency whose exchange rate has consistently appreciated or depreciated in the past. From a practical point of view, it has the disadvantage of putting a very high stress on data series quality. These disadvantages progressively lessen if the level of confidence required is lower, but that benefit is offset by the greater chance of missing an exchange rate movement which may recur in the future. The 95% quantile, which corresponds to the sixty-fifth largest loss out of 1,300 observations, is proposed.
30. The final step is to select the scaling factor which determines the toughness of the capital requirement. The first three parameters having been fixed, adding to the 95% quantile a scaling factor of somewhere between 2% and 4% of the overall net open position as measured under the shorthand method would deliver approximate equivalence in terms of toughness of the capital requirement for a portfolio of average riskiness between the shorthand and the simulation methods. The Committee is continuing to test bank portfolios to help it come to a final view, but at present inclines towards 3%. Setting the scaling factor in this way would mean that the minimum capital requirement would never be less than 3%.
25. Warrants would be treated in the same manner as options.
26. If the alternative treatments described in paragraphs 11 and 13 are used, the option together with related open and forward positions wou1d be treated separately.
27. e.g. if a German bank has a long forward position of US$ 100 million hedged by a corresponding put option with a strike price of 1.45 DM/dollar, with the current exchange rate being 1.40 DM/dollar the capital charge would be:
8% x DM 140 million - (DM 145 million - DM 140 million] - 11.2 - 5 = DM 6.2 million. 28. Or net present value in the case of those banks using net present value for measuring currency positions (see paragraph 9).
If the dollar appreciates to over 1.45 DM/dollar the option would be out of the money and the capital charge would just be 8I of the dollar forward position.
29. Where the bank is assessing its foreign exchange risk on a consolidated basis, it may be technically impractical in the case of some marginal operations to include the currency positions of a foreign branch or subsidiary of the bank. In such cases it is suggested that the internal 1imit in each currency be used as a proxy for the positions. Provided there is adequate ex post monitoring of actual positions against such limits, the li its would be added, without regard to sign, to the net option position in each currency. One country is in favour of allowing the substitution of limits only up to a maximum threshold expressed as a percentage of capital.
30. An alternative calculation, which produces an identical result, would be to include the reporting currency as a residual and to take the sum of all the short (or long) positions.
31. Two countries have reservations about expressing a de minimis exemption in this manner, and pending further discussion and more thorough consideration of the issue, are in favour of a free zone approach exempting 2% of a bank's capital base from the net open position.
32. Precious metals positions would also be eligible for treatment under the simulation method. Option positions in currencies or precious metals would be included on the basis of their net delta (or delta-based equivalent).