When the structural FX Risk is calculated on a consolidated basis, some trading book positions, such as FX Swap included in the hedging strategy may be considered of a structural nature. The hedging strategy implies that subsidiaries may directly manage their FX exposure in a currency different than their reporting currency, through hedging operations on the market or against the parent company. Such hedging operations employ trading book instruments like FX Swap or Cross Currency Swaps (CCS), which should be considered as structural.
The formula for the sensitivity range requires the specification of a target sensitivity, defined by the bank, and employs a predetermined value (5%); we believe that such a formula is too restrictive for two main reasons: first, it is not clear with which frequency such values need to be updated and monitored, secondly, regarding the 5% value, we believe it should be determined separately for each currency as every currency will have different levels of volatility, cost of hedging and sufficient market liquidity to execute hedges.
The main issue resulting from excluding the eligible net open position within the internal model perimeter is procedural rather than methodological. Indeed, we see a misalignment in terms of the reference period for computing:
- the sensitivity of the ratio to movements of FX to be hedged together with the maximum open position and;
- the capital charge generated by the exempted structural FX-positions
In the first case, the level is set relying on the current ratio at the reporting date (stand-alone value in T). In the second case, the hedges taken to reduce the volatility for that specific ratio (assessed in T) will be required to be (potentially) capitalised at the end of the following quarter. It is worth noting, moreover, the capital charge for these positions is determined by daily PL data within the quarter.
As a result, this asymmetric mechanism might lead to:
- a non-effective hedge (structural FX positions refer to prior capital ratio)
- the potential exemption of a position higher than the maximum open position (over-hedges will be measured only at the following reporting date)
The above reinforces the argument that the standard methodology would be more reliable for the application of the structural FX (S-FX) provision under Article 352(2).
In our opinion, a mathematical model to measure Structural FX risk would result in a too rigid framework.
Firstly, the EBA’s Guidelines MaxOp formula implies that the optimal position is proportional to the RWAs denominated in the currency for which the waiver is requested. The consequence of this is that the larger the RWAs on a specific currency, the larger would be the resulting waiver that can be granted by the Supervisor. This means that, given a subsidiary operating in a foreign currency, assets with lower risks would be penalised, e.g. assets to which we assign lower weights for capital charge calculation would translate in a lower denominator, thus a smaller optimal position.
The Max Op formula is interpreted as a cap against which the net open position is compared. We believe that this leads to an asymmetry: on the one hand, the net open position includes all assets and liabilities in currency, reported as stated in art. 352 CRR; on the other hand, we consider only RWAs in currency (i.e., only assets multiplied by their weighting factors). This leads to the fact that FX positions, to be exempted, need higher RWAs to immunise the ratio. As an example, consider the case of a swap derivative: this is included in the net open position with its notional, while its contribution to RWAs in currency is reduced to its risk weight.
Secondly, we believe that for the sensitivity formula, as well as for the definition of the sensitivity range (which implies the selection of a sensitivity target for each currency) proposed by EBA’s Guidelines, a grace period would be desirable so that the formulas can be subject to a monitoring process to grasp the correct calibration of such values and their volatility based on historical series.
In particular, as stated in the answer Q12, the formula for the sensitivity range requires the specification of a target sensitivity, defined by the bank, and employs a predetermined value (5%); we believe that such a formula is too restrictive for two main reasons: first, it is not clear with which frequency such values need to be updated and monitored, secondly, regarding the 5% value, we believe it should be determined separately for each currency as every currency will have different levels of volatility, cost of hedging and sufficient market liquidity to execute hedges.
Thirdly the formulas presented in the guidelines provide a quantitative definition of the capital ratio sensitivity concerning a specific FX rate. To this aim, they require some simplifying assumptions and a consistent effort to collect all data. The application mechanism of the obtained values (in particular the MaxOp, against which we compare our Net open position) seems to be excessively rigid, meaning that any change of the quantities due to FX rates fluctuations could imply a change in the optimal position of the bank.
From a mathematical point of view, many of the formulas are derived as approximations. In fact, in the equation of the optimal position, the right-hand side of the equation depends on itself from the optimal position, which makes the equation recursive. It should be clarified if the quantities on the right-hand side of the equation are to be considered as the current values.
In the specifics of the formulas, regarding the definition of Capital Ratio MaxOp, we consider the formulas as mathematically coherent only if the denominator of the following formula may be interpreted the total RWAs of the bank (balance sheet value) excluding the RWA FX for the specific currency; in other words, it can be decomposed in one part which depends on a generic FX rates and another part which does not:
CR_MaxOP=CET1/(RWA_(NoFX_FC ) ).