There is a strong case for the RBI to review the calculation of potential future exposure on derivatives.
The conceptual foundation of market risk and Basle II capital ratios has been to bring regulatory capital nearer “economic capital”: capital needed to absorb the likely losses in the bank’s business, without impairing its ability to pay off deposit liabilities. What the experience of Northern Rock in the UK, IKB in Germany and other financial institutions/banks suggests, however, is that a strong capital ratio is no guarantee that the bank can continue to function smoothly. (How right Walter Bagehot, a highly respected and influential 19th century economist and journalist, was when arguing that “good bankers don’t need any capital, and there is not enough capital in the world to support bad bankers”!) In these cases, drying up of market liquidity and confidence forced the hands of the lender of last resort (or the state), who had to step in directly or indirectly to protect one bank failure becoming a systemic risk to the whole financial backbone of the economy. No wonder so many pundits are putting forth proposals for reform of the whole regulatory approach: among them Charles Goodhart and Avinash Persaud; Satyajit Das; Martin Wolf; Lawrence Summers; John Kay; George Soros; IIF; the Financial Stability Forum; BIS itself; etc. The Swiss National Bank has recently gone beyond Basle II and prescribed a gearing ratio. And, regulators are having to step in to help rescue even unregulated entities (Bear Stearns, for example) which pose systemic risks as counterparty to derivative contracts in trillions of dollars.
Before the current credit crisis became news, I had argued (see “Proprietary trading: risks and rewards,” Business Standard, June 8, 2007), on empirical analysis, that given the scale of trading profits of major banks, the reported “value at risk” measure seems to be grossly underestimating the market risk (and hence the capital therefore): so much profits just cannot be earned by taking so little risk! The subsequent write- off of $500 bn or so by banks through MTM valuation of the trading books clearly substantiates the argument then advanced.
If the earlier argument was about market risk capital, I find the same weakness in the prescription of credit risk capital for off balance sheet exposures of banks through derivatives. The Reserve Bank has recently come out with revised guidelines for calculating the “current exposure” on which capital is needed to be earmarked: current exposure is the sum of the (negative — for the counterparty) MTM value of the contract and the “potential future exposure”. It is the calculation of the latter that seems grossly inadequate, as it is merely a small percentage of the notional principal — regardless of exchange rate volatility, other variables which would influence the values, etc. To be sure, what the RBI is doing is similar to what the BIS has recommended (and what the Fed and FSA are following). The Basle II numbers are supposed to have been drawn up on the basis of simulation analysis of the maximum loss at 95% confidence level. Nevertheless, they do not seem realistic. The potential future exposure should be an estimate, calculated at a given degree of confidence, of the maximum adverse movement in the value of the contract if held for a given period: in other words the PFE needs to be calculated in the same manner as the value at risk on the contract, using recognised volatility models like EWMA, GARCH, etc.
I have been able to find support for this argument from two different authorities. The Monetary Authority of Singapore prescribes the following principles:
“Potential future exposure (PFE) is a measure for pre-settlement risk arising from a financial instrument as a result of market changes. Both simulation analysis as well as analytical tools may be employed to measure PFE. The method used to measure pre-settlement risk should be commensurate with the volume and complexity of an institution’s treasury and financial derivatives activities. The assumptions used to calculate PFE should be reasonable and consistent. The time horizon used can vary depending upon the contract residual maturity, collateral protection and the institution’s ability to terminate its credit exposure. A time horizon equal to the tenure of the contract may be inappropriate in the case of collateralised exposures. In such cases, the horizon should reflect the time required for the institution to terminate the contract and liquidate existing collateral when an obligor fails to meet a collateral call” (Paragraph 4.4 from “Credit Risk”, February 2006).
While discussing the estimate of credit losses (“potential future exposure”) on derivatives, John Hull, in his “Risk Management and Financial Institutions”, gives the example of a one-year forward contract and uses exchange rate volatility to calculate the present value of the cost of default on the contract. And this does seem to be logical. Under the existing model used by regulators, the potential future exposure depends purely on whether it is an interest or exchange rate contract, and its residual maturity. This means, for example, that the PFE on a, say, 13 months INR: USD principal only swap is identical to that on a bought, three-year maturity INR:JPY option, or a five-year INR:CHF swap! This seems a manifestly illogical proposition.
Theoretical arguments apart, the recent experience in the Indian currency derivatives market also evidences the gross inadequacy of the RBI measures of the potential future credit exposure even on plain vanilla products. For example, the maximum credit exposure on a ten-year cross currency swap (receive 6% USD, pay 9% GBP) could be as high as almost a quarter of the notional principal, even with future spot equalling the forward rate at inception! The PFE for the swap at inception is just 10% in terms of the regulator’s model! In the case of complex products, the situation is of course much worse: I recently came across a case where the credit exposure has come to 150% of the notional principal! Since the transaction was back-to back for the bank, there was no market risk.
The pernicious effect of low regulatory capital requirements is that it tempts banks into marketing such products as the return on equity is so high, overlooking, in the process, the credit risk! To may mind, there is a strong case for the RBI to review the calculation of potential future exposure on derivatives and hence the capital charge.
