2.2 Actividades económicas
2.2.2 Sistema Económico
Although banks had previously applied statistical methodology in their assessment of market risk arising from different areas of the trading book, the Basel Committee agreed that this approach was not effective enough.
445
Slaughter & May LLP, New Basel Capital Accord Guide (n422) 23. 446
ibid, 22-23. 447
Excluding derivatives. 448
For all off-balance sheet exposures including exposures of less than 1 year, a conversion factor of 75% is usually applicable. However, for short term exposures such as trade credits, a credit conversion factor of 20% is applicable.
The fluctuations in the market prices of equity and debt instruments as well as in ‘foreign exchange and commodity positions’449 that banks experienced after the introduction of the Basel Capital Accord, prompted the Basel Committee to make amendments to the Basel Capital Accord.
This process was initiated when the BCBS issued a proposal450 in April 1993 aimed at addressing banks’ exposure to market risks. These markets had been identified by the BCBS as comprising of risks emanating from the trading book due to shifts in market prices of equity and debt instruments, off- balance sheet positions and contracts, as well as foreign exchange risk.
Thus after welcoming participants’ suggestions and opinions, the BCBS in 1995 accepted that it was logical and prudent to extend the ambit of Basel 1 to include market risks. This paved the way for the BCBS to introduce the Basel Committee’s Market Risk Amendment to the Capital Accord (the ‘Amendment’) in 1996.
The rationale for this amendment was to provide in explicit terms, a requirement for banks that were engaged to a significant degree in some form of business trading to compute a capital charge using the standardised approach or the internal models approach.
This was deemed a necessity by the Basel Committee to address the risk of loss a bank potentially faced following a depreciation in value in the market prices of its equity and debt holdings.
The Market Risk Amendment also established a distinction between financial instruments that were kept by banks on a short-term basis with the intention of gaining profit following a resale. Such assets, constituting the bank’s trading book were distinct from assets in the banking book, including off-balance sheet items.
Thus there are two methods that can be employed under the Market Risk Amendment, to calculate the capital charge applicable for assets within the trading book as well as off-balance sheet items, i.e. the standardised approach as already discussed and the Internal Models-based approach.
449
McKnight, (n439) 333. 450
4.8.1 Internal Models-based approach.
Under the framework for the internal models-based approach, banks implementing this process were required to conform to qualitative and quantitative standards. The qualitative standard criteria required senior management within the bank to set up an independent risk management unit whose task it was to be closely involved with the daily risk management procedures of the bank.
These processes were expected to be rigorous and robust and there had to be a system in place to ensure independent and regular assessment of the risk measurement processes with a view to ensure compliance.
The quantitative standard on the other hand involved the daily computation of the ‘value-at-risk’ which involved using a ‘99th
percentile, one tailed confidence interval’ measured over a period of 10 days during which the instrument or item is expected to be held in the trading book or on-off balance sheet. Thus a bank using the internal models-based approach was expected to conform to these standards and the capital charge set aside was based on data obtained from a ‘historical observation period’ of a minimum of 1 year. Applying such data, the capital charge was calculated using the greater value of either the previous day’s ‘value-at-risk’ measurement or the average figure representing the daily ‘value-at-risk’ figures for each of the previous 60 business days which is then multiplied by a (factor)451.
4.8.2 Market Risk Value-at-Risk Model (VaR)
The rationale behind the introduction of the VaR model stems from the fact that exposures in the trading book were not expected to remain in the trading book on a long-term basis as the ultimate intention of the bank was to re-sell thus raking in some profit. The fact that a bank may hold these instruments on a short-term basis and not until maturity implies that the valuation of these instruments should not be conducted on a mark-to-market basis to ensure
451
This factor is expected to be provided by the national banking supervisor in that jurisdiction and is based on their individual assessment of the risk management framework existing in each bank.
effective management of market risk452. It is the nature of such valuation which necessitated the introduction of the VaR model.
The VaR model was indeed more sophisticated than the standardised methodology previously applied and with it came a number of parameters within which it could be applied.
Following the introduction of the VaR model, the Basel Committee encouraged banks to implement their own version of value-at-risk (VaR) within the parameters set by the Committee. The purpose of the VaR model was to ensure a consistent outcome in the calculation (by a bank) of the likelihood of loss that a bank could suffer in the entire trading book.
The likelihood of consistency in outcome was perceived to be characterised by the VaR models’ ability to gauge price fluctuations of instruments in the respective markets and to measure the extent to which these market prices differed with other market prices. There are two types of VaR models introduced under the Market Risk amendment of the Basel Capital Accord, i.e. the ‘Variance/Covariance analysis’ approach and the ‘Historical Simulation’ approach.
4.8.2.1 Variance/Covariance analysis approach
Under this approach, historical data accumulated for price fluctuations within the market as well as price correlations resulting from market comparisons are inserted into a statistical formula which provides an estimation of the value of likely loss(es).
An assumption is created that the normal distribution of price changes (fluctuations) allows a confidence level to be calculated for the bank i.e. an attempt to predict the value-at-risk figure over a fixed period of time with a degree of near certainty (e.g. 95% or 99% likelihood that this confidence level would not be exceeded).
Finally this approach provides a formula for calculating the confidence level by multiplying a standard deviation derived from previous price fluctuations453 with a scale factor.
452 P Jackson, ‘Risk Measurement and Capital requirement for banks’, Bank of England Quarterly Bulletin May 1995 pp177-184.
4.8.2.2 Historical Simulation approach
This approach is based on future potential losses which a bank expects to incur through the analysis of data relating to previous price454 fluctuations. This data analysis is used to assess the volatility in prices on the assumption that the instruments would be kept for the entire duration of two years.
This method ensures the attainment of a 99% confidence level without supporting the assumption that price fluctuations are uniformly distributed, i.e. it indicates that loss was never exceeded in 99% of the time.
A distinguishing feature455 between the two types of VaR approaches is the assumption of normal distribution of price changes under the variance/co- variance approach. The fact that the historical simulation approach does not make such assumptions ensures that this approach often results in an arguably effective determination of the confidence level by ensuring that price volatility is not miscalculated. This is mainly because price volatility can be very extreme and as a result an assumption that volatility gives rise to a normal distribution is likely to give inaccurate and unreliable results.
An advantage which the historical simulation approach has over the variance/co-variance analysis approach is that the former is able to accurately capture the entire risk within the trading book including risks from the option portfolio456.
Although banks can implement either of the two approaches in their quest to set aside capital for market risk, the UK banks have been known to favour the historical simulation analysis approach. It is submitted that the analysis of the responses provided by the commercial banks that participated in this research suggests that most of the commercial banks from both jurisdictions commonly
453
Measured in percentages. 454
Usually data over two years is utilised. 455
Another difference is in the calculation of confidence interval, where a statistical method is used for the variance/co-variance method as opposed to the observation method used for the historical simulation approach.
456
used the standardised approach as opposed to the Internal Models – based approach or the Market Risk Value-at-Risk (Model) VaR457
This may be attributed to the fact that the risk management systems and framework in place within these banks are not sophisticated enough to support such updated risk management/measurement techniques.