(1) Qualitative information

As illustrated in Table 2, the banks and securities firms included in the sample significantly expanded the qualitative, summary discussion of their trading and derivatives activities over the 1993-1996 period. This trend can be observed for all of the disclosure items reviewed in Table 2.

In comparison with 1995, progress continued in 1996 although improvements were generally less pronounced than in previous years. In particular, there was an increase in the number of institutions discussing operating and legal risks, market value adjustments and reserves, and the accounting treatment for derivatives credit losses. Moreover, institutions provided more ample discussions of risks, objectives and accounting policies. These improvements are not always visible from the survey results, since the numbers do not show when the quality of an institution's discussions of risks, objectives or accounting policies has improved.

Increasingly, internationally active banks and securities firms provide a comprehensive overview of the business objectives of their trading and derivatives activities, the associated risks, and the methods used to manage these risks. In 1996, the most noteworthy development was the increase in the number of institutions describing operating and legal risks. As shown in Chart 1, 40 institutions (51 %) provided such information in 1996, as compared with 32 in 1995 (41 %) and 10 in 1993 (13 %). In other areas, progress made in previous years was sustained. Seventy-four institutions discussed objectives and strategies for trading activities and 75 for non-trading activities, as compared with 71 and 66, respectively, in 1995, and 38 and 37 in 1993. The number of institutions discussing how credit and market risk arises increased from 34 and 35, respectively, in 1993 to 66 and 68 in 1995, and 68 and 69 in 1996. Forty-nine institutions discussed how liquidity risk arises compared with 46 in 1995 and 19 in 1993.

Disclosures of valuation policies also continued to expand in 1996, as shown in Chart 2. Sixty-five institutions (82 %) discussed the methods and assumptions used in valuing financial instruments in 1996, compared with 58 in 1995 (73 %) and 26 in 1993 (33 %), and 27 institutions (34 %) provided a discussion of their market valuation adjustments or reserves, compared with 18 in 1995 (23 %) and 9 in 1993 (11 %). The number of institutions providing a general discussion of their accounting policies for derivative instruments increased from 63 in 1993 to 72 in 1995 and 1996. A significant number of institutions provided further detail on their accounting policies, for example, by distinguishing between accounting methods for different types of derivatives instruments (62) or by discussing hedge accounting criteria (49). In comparison to 1995, 1996 also saw an increase in the number of institutions discussing the accounting treatment for credit losses related to derivative instruments (27 in 1996, as compared with 19 in 1995 and 9 in 1993).

(2) Quantitative information

Table 3 presents an overview of disclosures about notional amounts and market values of instruments held for trading purposes (on- and off-balance-sheet) and derivatives held for non-trading purposes. These measures are indicative of an institution's involvement in derivative instruments. As Table 3 shows, disclosures of position indicators expanded considerably over the 1993-1996 period. There was a significant increase in the number of institutions disclosing information for almost all of the disclosure items reviewed in Table 3.

In comparison with 1995, the number of institutions disclosing information on notional amounts separately for over-the-counter and exchange-traded instruments increased in 1996, as shown in Chart 3. Fifty-five institutions (70 %) distinguished OTC from exchange-traded instruments in 1996, as compared with 44 in 1995 (56 %) and 13 in 1993 (16 %). All of the 67 banks and 12 securities firms provided information about the notional amounts of their derivatives holdings from 1994 and onwards. The number of institutions that separated trading from non-trading positions increased significantly, with a majority of institutions now providing this information. As regards market value data, there was expanded disclosure on the gross negative market value of derivatives, the market value of derivatives in the trading account, and the market values for different types of non-trading derivatives position in 1996. Thirty-seven institutions disclosed the gross negative market value for derivatives in 1996, as compared with 29 in 1995 and 13 in 1993. Fifty institutions disclosed the market values of derivatives in the trading account in 1996, compared with 44 in 1995 and 21 in 1993. The number of institutions disclosing information on the market values by type of derivative held outside of the trading account (for example for hedging purposes) increased from 5 in 1993 to 12 in 1995 and 22 in 1996.

(a) Credit risk

Over the 1993-1996 period, banks and securities firms materially expanded the quantitative information provided on credit risk, as illustrated in Table 4. In some cases, this information was provided separately for derivatives instruments; in other cases, cash and derivatives-related disclosures were combined.

In 1996, there was an expansion in the disclosure of current and potential credit exposure, as shown in Chart 4. Forty-two institutions (53 %) disclosed data on current credit exposure, taking into account the effects of netting, as compared with 36 in 1995 (46 %) and 22 in 1993 (28 %). Twenty-one institutions provided information about the potential credit exposure (27 %), a measure of how much current credit exposure could increase in the future as a result of movements in underlying rates or prices, as compared with 15 in 1995 (19 %) and only one in 1993 (1 %).

Furthermore, survey institutions continued to provide more information on the credit quality of their trading and derivatives portfolios in 1996. For example, 50 institutions disclosed information on counterparty credit quality, as compared with 41 in 1995 and just 6 in 1993. The most common type of disclosure on credit exposure was information on gross positive market values (without netting), counterparty credit quality, and risk-based capital credit-equivalent amounts. Additionally, the number of institutions disclosing information on credit concentrations grew to 48 in 1996, compared with 46 in 1995 and 11 in 1993.

As in the two earlier surveys, no institution provided data on the volatility of credit exposure of its derivatives holdings over the reporting period, and, as shown in Chart 5, just 11 (14 %) firms provided information on collateral and other credit enhancements, and only 14 firms (18 %) disclosed information on actual credit losses.

(b) Market risk

The number of institutions disclosing quantitative information on their exposure to market risk grew substantially over the 1993-1996 period, and the information provided amplified (Table 5). Also when comparing 1995 and 1996 annual reports, considerable progress is visible. Increasingly, the banks included in the survey are basing such disclosures on their internal value-at-risk methodologies¹². Value-at-risk is an estimate of potential trading losses over a given time horizon, measured at a certain level of statistical confidence. In 1996, 50 institutions provided such value-at-risk-based disclosures, as compared with 36 in 1995 and 4 in 1993. Increasingly, value-at-risk figures are given for a holding period of one day. As shown in Chart 6, data on daily value-at-risk was disclosed by 34 institutions in 1996 (43 %), as compared with 22 in 1995 (28 %) and 4 in 1993 (5 %).

In comparison with 1995, there was also a significant increase in 1996 in the number of institutions disclosing certain major assumptions underlying their value-at-risk estimates. This is an area where the Basle Committee/IOSCO November 1995 report identified the need for further improvements. In 1996 annual reports, 48 banks disclosed the confidence interval used, 47 the holding period, and 23 the method of aggregation across risk factors, as compared with 35, 33 and 14 in 1995, and 2, zero and zero in 1993. Chart 7 below illustrates examples of disclosures of the major assumptions underlying value-at-risk estimates presented in 1996 annual reports¹³. Chart 7:
Illustrative examples of disclosures of major assumptions
underlying value-at-risk estimates presented in 1996 annual reports

		Institution 1	Institution 2	Institution 3	Institution 4	Institution 5	Institution 6	Institution 7	Institution 8	Institution 9	Institution 10
Holding period		10 days	10 days	10 days	10 days	1 day	1 day	1 day	30 days	1 day	1 day
Confidence interval		99 %	99 %	99 %	99 %	97.7%	99 %	97.7%	99 %	98 %	95 %
Aggregation method		No corr.	Correlation	-	-	Correlation	Simulation	Correlation	Correlation	Correlation	No corr.
Average daily VaR		-	415	81	280	100	34	23	-	10	4.4
High (max.) daily VaR		80	471	121	-	118	47	30	1090	21	6.9
Low (min.) daily VaR		20	389	64	-	63	19	10	366	4	3.2

In addition to disclosing a point in time value-at-risk number for the end of the financial statement period, a number of banks also provided information on their value-at-risk exposures over the whole reporting period. For example, 27 banks disclosed the average value-at-risk number for the reporting period, as compared with 20 in 1995 and zero in 1993. Twenty-four banks disclosed the high and low value-at-risk numbers in 1996, compared with 17 in 1995 and zero in 1993. Moreover, 15 banks directly related daily value-at-risk estimates to actual changes in portfolio value, compared with 10 in 1995 and zero in 1993, one of the key recommendations of the 1994 Fisher Report. Institutions typically used graphical means to compare daily value-at-risk estimates with actual portfolio outcomes.

In 1996, disclosure of the results of scenario analyses expanded. Thirteen institutions disclosed such information, in comparison with 6 in 1995 and one in 1993.

Historically, the major securities firms have not provided quantitative market risk disclosures of their trading and derivatives activities in their annual reports. As part of the Derivatives Policy Group's "Framework for Voluntary Oversight" on over-the-counter derivatives, released in March 1995, the securities firms that are major US derivatives dealers are providing to United States supervisors on a quarterly basis measures of "capital-at-risk", defined as the maximum loss expected to be exceeded with a probability of one per cent over a two-week period. In addition, these dealers provide supervisors with the results of a series of core risk factor stress tests of their over-the-counter derivatives portfolios.

Non-trading derivatives activities

In the 1993-1996 period, the most common form of disclosure by the surveyed banking institutions that used derivatives for non-trading purposes involved schedules of notional amounts, maturities and (for swaps) contractual rates paid and received. For the 1993-1996 period, the most prevalent means of conveying how derivatives are used to manage a bank's interest rate risk was a gap position schedule (used by 24 of the banks in 1996 as compared with 26 in 1995 and 23 in 1993¹⁴). Many banks publishing a gap schedule for interest rate risk cautioned that it represented only a point-in-time picture of risk and did not capture options risk and other dynamic characteristics of the balance sheet.

The number of banks that furnished quantitative information on their non-trading activities remained low. Sixteen banks provided a discussion of the effect on capital or earnings of a specified rate shock. A few of the banks providing information on their non-trading derivatives holdings described in varying detail whether the derivatives were linked to specific components of the balance sheet or were used to manage overall risk exposures.

(c) Earnings

Trading activities

As illustrated in Table 6, information on trading income has expanded since 1993. In 1996, 69 institutions (87 %) disclosed information on trading income, as compared with 64 in 1995 (81 %) and 48 in 1993 (61 %), as shown in Chart 8. While there was a significant increase in 1996 in the number of institutions providing some type of breakdown of their trading income, still only about half of the institutions (40 or 51 %) disclosed information by line of business or risk exposure and fewer than one-quarter (17 or 22 %) by instrument type, as shown in Chart 9. Twenty-one institutions provided information about trading income broken down between cash positions and derivative instruments in 1996, as compared with 18 in 1995 and 22 in 1993, while 33 institutions presented other information about trading income in 1996, as compared with 34 in 1995 and 29 in 1993.

Non-trading derivatives activities

With regard to derivatives held for non-trading purposes, 1996 saw an increase in the number of institutions disclosing details about how derivatives affect accrual-based accounting income and expense (historical cost accounting), as shown in Chart 10. Twenty institutions reported the effect that derivatives accounted for on an accrual basis had on revenue, compared with 10 in 1995 and 5 in 1993. Nine banks and 6 securities firms reported the overall effect on net interest margins of their non-trading derivatives activities. Twenty institutions disclosed deferred gains or losses on non-trading derivatives and 7 provided information on when the deferrals would be reflected in future earnings. Twenty-five banks and 3 securities firms disclosed the unrealised gains and losses associated with non-trading derivatives positions, compared with 21 and 3, respectively, in 1995 and 9 and 3 in 1993.

November 1997

Footnotes:

11 This Chart and Chart 4 give examples of credit risk disclosures made by surveyed institutions. There are many other types of disclosure listed in Table 4, such as current credit exposure, gross positive market value, counterparty credit quality, information on concentrations, and risk based credit equivalent (for banks). The majority of the institutions surveyed provided these types of disclosure about their trading and derivatives activities.

12 In many countries capital adequacy rules for market risk were introduced in 1996. In some of those countries, disclosure of the capital requirement for market risk and its component parts became common in 1996. This type of disclosure was not surveyed.

13 It should be noted that value-at-risk estimates are not necessarily comparable among different institutions unless major underlying assumptions and parameters are comprehensively disclosed, such as information on the portfolios covered by the model and model parameters (holding period, confidence level, observation period, aggregation method).

14 Gap schedules disclosed by banks organise financial assets and liabilities according to maturity in a number of time bands. The difference between assets and liabilities in each time interval ("gap" or net exposure) forms the basis for assessing interest rate risk. Derivatives of various maturities can be used to adjust the net exposure of each time interval to alter the overall interest rate risk of the institution. Historically, securities firms have not presented gap table disclosures in their annual reports.