Forum Documents
   Equity at risk
   IFCI-ARTHUR ANDERSEN Risk Disclosure Sur...
   Rethinking the quality of risk managemen...
















 

Forum Documents

IFCI-ARTHUR ANDERSEN Risk Disclosure Survey

EXECUTIVE SUMMARY

Both regulators and financial institutions have recognised the benefits of enhanced public disclosure of risk information. Since 1995, the Basel Committee and IOSCO have surveyed the risk disclosure practices of some 72 internationally active financial firms on an annual basis. The 1999 Basel Committee paper on a new approach to regulatory capital makes enhanced risk disclosure the key element of its third pillar of market discipline, complementing the pillars of capital adequacy and supervisory review.

However, despite the increasingly public focus on risk disclosure, little attention to date has been paid either to the information requirements of the end-users of these disclosures, or to the quality of the risk disclosure itself. This study on risk management disclosure by financial institutions in their annual reports seeks to fill that gap. Our primary conclusion is that, despite the fact that enhanced risk disclosure has been an objective of regulators, senior management, stakeholders and counterparties for some time, the actual quality of risk disclosure by the financial industry is disappointingly poor. Indeed, our study found that risk disclosure needed enhancement even among those firms which might be regarded as the leaders in this field.

The background to this study is an IFCI Research Paper, which was published in June 1999. IFCI's objective in undertaking this research was to develop, for the first time, a standardised framework of risk disclosure information designed by the users of accounts rather than by regulators, the accounting profession or the firms themselves which could be recommended to financial firms. Professor Gibson chaired an IFCI sponsored study group comprising representatives from the financial analyst community, rating agencies, fund managers and audit firms. We wanted to generate a standardised framework to assess the quality of risk management disclosure which was meaningful to the users of accounts. The focus is on the quality of risk management disclosure for market and credit risks based on five core principles and a series of tables (see Approach section) that provide an appropriate level of meaningful qualitative and quantitative information about these risks.

In line with current regulatory thinking, the study group distinguished between the information that was relevant to a firm's trading activities and information that was relevant to its non-trading activities. This distinction was intended to:

  • encourage firms to disclose risk information in a manner that is consistent with its underlying business structure and product mix;

  • discourage the notion that risk disclosure is solely, or even primarily, about a firm's derivatives trading activities;

  • encourage firms to disclose meaningful information about the risks in its non-trading activities as well as in relation to its trading activities.

The study assesses the extent to which the quality of risk disclosure in annual reports meets the standards proposed in Professor Gibson's original research paper. IFCI, in collaboration with Arthur Andersen, reviewed the 1998 annual reports of a sample of 42 internationally active financial firms to check their consistency with the recommended risk management disclosure framework. The review focused both on the type and quality of disclosure, assessing the latter in relation to the five core principles and with respect to the minimal set of quantitative and qualitative criteria.

Based on the IFCI risk disclosure framework set out in Professor Gibson's paper, the study group established a template for scoring the reports; a maximum score of 100 would be awarded if a firm had provided good quality information in relation to each of the items identified in the template. On this basis in relation both to trading and non-trading disclosure, not one firm whose annual report was reviewed scored above 50. Moreover, of the 42 firms reviewed, the large majority scored below 25 in relation to both trading and non- trading risk disclosure.

The study highlights a number of issues concerning current risk disclosure practices. In particular:

  • that a standardised framework for the disclosure of risk management information is valuable, particularly in facilitating comparability between different firms. Comparability of this type between different firms is a necessary precondition to the 'market discipline and efficiency' that is sought by financial regulators such as the Basel Committee;
  • while disclosure practices can be improved in almost all areas, there is particular scope for improvement with respect to non-trading activities and the credit risk associated with trading activities;
  • specific improvements are needed in relation to:

  1. the consistency of disclosed VaR numbers and the detail of VaR methodologies;
  2. the approach which firms are taking to stress testing and its applicability to both trading and non-trading activities;
  3. discussion of internal credit rating methodologies and loss given default estimates.

When considering the results of this survey, one should recognise that no disclosure template can meet all the requirements of a dynamic market: management must have the flexibility to disclose information about their firm's risk profile in a way that best reflects the particular characteristics of their business activities. Consequently, it would be unrealistic to expect firms to comply fully with one of a number of possible frameworks for risk disclosure. Nevertheless, the strong message from this study is that the quality of risk disclosure across the financial services industry lags well behind where it could be, even on a relatively simple basis of providing understandable data. Indeed, many firms currently disclose little if any meaningful risk information which would allow users of accounts to understand the firm's risk profile and risk management practices. It is clear that before any substantial progress in risk disclosure can be made, management needs to give risk disclosure equal importance in their annual reports to financial disclosure. Our research suggests that much remains to be done by all firms, and particularly by those (and they form the majority of those surveyed) which currently fall below the 25 mark.

APPROACH

In order to analyse the quality of financial risk management disclosure from both a quantitative and qualitative perspective, IFCI developed a disclosure framework-scoring matrix to provide a consistent benchmark to analyse the 42 sample financial firms. A summary of the framework can be found in Figure 1.

Figure 1: IFCI Disclosure Framework Matrix

Trading Activities

    1. Overview of risk management policy
    2. Market risk
    3. 1.2.1. VaR information

      1.2.2. Backtesting information

      1.2.3. Stress testing information

    4. Credit risk

1.3.1. Overview of credit risk in trading

1.3.2. General description of trading position (nominal amounts)

1.3.3. Current and potential credit risk exposures

1.3.4. Internal credit risk modelling

1.3.5. Integrating market and credit risk management

      1. Concentration risks
      2. Other relevant information

Non Trading Activities

    1. Overall assessment of the disclosure of non-trading activities
    2. Market risk
    3. 2.2.1. VaR for non trading activities

      2.2.2. Asset and liability management

      2.2.3. Hedging policy

    4. Credit Risk

2.3.1. Overview of non-trading activities and credit risk policy

2.3.2. Counterparty data analysis

2.3.3. Concentration risks

2.3.4. Cross border analysis

Each element of the matrix was given a weighting and maximum score with the total score of 100 for trading and non-trading book disclosures respectively.

The weightings of the matrix and the actual scoring process were based on the study group's views of the importance of both quantitative and qualitative risk disclosure. This philosophy is driven by the five principles of risk management disclosure set out in Professor Gibson's paper:

  1. the disclosure of risk management practices should enhance the confidence of market participants in the willingness of the firm to provide information about its specific and aggregate risk exposures;
  2. that a firm's risk management disclosure policy should be complete;
  3. that a firm's risk management disclosure policy should incorporate consensus risk assessment practices, focusing on materiality and symmetry;
  4. disclosure practices should be standardised in order to facilitate risk management comparisons across firms and products/market segments;
  5. disclosure quality should be analysed across both functions and risk factors, based on the risk management process of the firm.

This survey and its assessment framework will differ from the disclosure requirements set out by regulatory bodies in a number of ways. The framework is more forward looking with respect to future risk exposure as opposed to retrospectively analysing the disclosure of risk over the accounting period covered by the annual report. Additionally, regulatory bodies have only recently become concerned with elements in the scoring matrix such as the use of internal credit ratings. Therefore, regulatory disclosure requirements may lag the disclosure principles which are addressed in this survey.

Forty-two financial firms which form part of the sample of firms included in the annual IOSCO survey were selected for inclusion in this study. An analysis of the geographical distribution of the firms surveyed is outlined in Figure 2 below.

Figure 2: Geographical Distribution of Financial Firms Analysed

The 1998 annual report for each of the 42 selected financial firms were scored against the matrix.

SURVEY RESULTS

The results of this survey have been separated between trading and non-trading (banking book) disclosure, and are reported by quartile measures. It is important to note that as no firm scored more than 50% for either trading or non-trading book disclosure, the quartiles are divisions of 0-50% scores.

It should also be noted that these results were generated by analysing bank disclosure against a best practice framework and not international regulatory or local statutory requirements.

ANALYSIS OF RESULTS FOR TRADING BOOK DISCLOSURE

The results for trading book disclosure outlined in Figure 3 below suggest that US firms have better trading risk disclosure practices than their European and Japanese counterparts.

Figure 3: Trading Book Disclosure Results

Quartile

Average Score

Total Number

US Firms

European Firms

Japanese Firms

0-12.5%

8.1%

15

4

10

1

12.6%-25%

19%

17

6

10

1

25.1%-37.5%

28%

5

4

0

1

37.6%-50%

41.5%

5

2

2

1

Furthermore, these results appear to support the hypothesis that financial firms in countries with a strong shareholder culture disclose more information than those financial firms from countries where ownership is more closely held or dominated by family-industry-banking relationships (see Figure 4).

Figure 4: Trading Book Disclosure Scores

While the industry and the regulators have generally adopted value at risk (VaR) as one of the primary measures of market risk, it was surprising to find that very few of the financial firms analysed disclosed the major assumptions underpinning their disclosed VaR measures. Other issues identified regarding the reporting of market risk for the trading book are:

  • while the annual reports disclose the firm's VaR, they fail to explain that this measure is calculated on the assumption of normal market conditions and that stress testing is important to supplement the VaR measure;

  • very few firms did in fact provide any disclosure relating to stress testing. Where firms had provided some level of disclosure, in general, very little was given relating to the scenarios actually applied and the motivation for selecting those particular scenarios;

  • around 50% of the firms surveyed did not provide any details on backtesting.

The survey found that more emphasis was given to the disclosure of market risks than credit risk for trading activities:

  • Exposures are generally only reported for derivatives in the trading book and not for other products such as fixed income securities.

  • Credit risk concentration in the trading book is generally ignored with very little information being disclosed regarding industry, sector and geographic concentrations.

  • The survey found limited disclosure of credit risk by internal or external ratings categories.

ANALYSIS OF RESULTS FOR NON-TRADING BOOK DISCLOSURE

The results of the analysis suggest that the disclosure of risk information regarding non-trading book activities are poorer than trading book disclosure. Only one of the firms analysed fell into the 37.6%-50% quartile for non-trading book disclosure compared to five firms for trading book disclosure.

Figure 5: Non Trading Book Disclosure Results

Quartile

Average Score

Total Number

US Firms

European Firms

Japanese Firms

0-12.5%

5.3%

17

5

10

2

12.6%-25%

19%

15

7

6

2

25.1%-37.5%

31.1%

7

3

4

0

37.6%-50%

 

1

0

1

0

These results suggest that firms have recently concentrated more on trading book disclosure at the expense of the non-trading book disclosure. This may well be a response to the changes in the BIS rules relating to market risk in the trading book.

The recent BIS paper on credit risk identified concentration of credit risk in the non-trading book as one of the key risks which might bring down a bank Therefore it is an appropriate time for banks to think about improving their disclosure for non-trading book risks. Additional findings relating to credit risk disclosure for the non trading-book include:

  • little information is disclosed regarding the expected and unexpected losses calculated using a credit portfolio model;

  • there was a lack of counterparty analysis with respect to the internal or external ratings of counterparties;

  • very few firms provided sufficient information on cross border credit risk.

Figure 6: Non Trading Book Disclosure Scores

Unsurprisingly, the study found that the disclosure on market risk was poorer than that for credit risk in the non-trading book, reflecting the relative balance of market against credit risk in the banking book.

The study found that a large number of banks are not currently using VaR to manage the market risk within their banking books. Given the BIS pronouncements on the management of interest rate risk in the banking book, and that banking book loans are increasingly found to have more embedded optionality, such disclosure needs to be improved.

GENERAL OBSERVATIONS

In analysing the 42 sets of annual reports, the research team also identified a number of general observations regarding the quality and coverage of risk disclosure:

  • a number of firms surveyed still provide generic information relating to definitions of options and swaps etc. While such disclosure may have been appropriate in the past for investor education purposes, it is no longer required; this effort would be more effectively directed towards more useful disclosure items;

  • a number of firms surveyed do not clearly distinguish between trading and banking book;

  • there is a lack of depth in many sets of accounts. Frequently, information is often confusing or in a poor format, with no reference to key pieces of information set out in the notes to the accounts (e.g. derivatives disclosures on fair values and notional, or loan disclosures on concentrations, etc.);

  • there is a general lack of a good quality, comprehensive overview of policies in the accounts. There is a lack of clear policy statements generally they omit this information, although a number of the of accounts do give details of the roles and responsibilities of the various committees and key individuals. There is rarely any inclusion of information regarding any significant changes to the trading books and also, no attempt is ever made to give a profit forecast. Additionally, no attempt is made to identify concentrations across assets in terms of geography and industry and discuss these concentrations;

  • one of the most important policy disclosures that a bank could state in terms of derivatives exposures (especially in illiquid or volatile trades) is the use of collateral and the policy framework supporting it. Virtually none of the banks have given any disclosures in this area or quantified its impact in the accounts (collateral impact is generally mixed in with netting agreements when net exposures are given);

  • economic performance has a very mixed set of approaches. Some banks have identified the major economic and market factors that have affected the business (e.g. LTCM, Russia etc), but the monetary impact of these on the business has not generally been disclosed.

CONCLUSION

Regulators view improved disclosure as one of the three key pillars of regulation. The provision of meaningful information should enable investors, depositors and creditors to execute market discipline and ensure more efficient allocation of capital.

This in turn should mean that firms manage their businesses to maximise shareholder value, by ensuring that the risk adjusted return profile is optimised.

Despite regulatory supervisors promoting the improvement of risk disclosure, the results of this survey suggests that firms are not providing sufficient levels or quality of disclosure in relation to market and credit risk.

If stakeholders are to be able to exert market discipline, they need to have access to a high degree of transparency in risk disclosure which is comparable across different firms. However, this study has found that firms are not yet providing sufficient disclosure in a comparable manner to stakeholders. Two of the key recommendations from this survey are:

  • there needs to be an standardisation of the way that risk disclosure is presented by firms in order to facilitate comparability both through time and between firms possible;

  • that the disclosure of risk information by firms could be considerably improved.

This study will be carried out annually and will serve as a backtesting exercise to assess the level of improvements firms make to their disclosure over time and also compared with their peer group.

ABOUT IFCI

IFCI: International Financial Risk Institute is a Geneva based, industry-led think-tank for risk management, education, information and research. IFCI operates as a non-profit foundation.

ABOUT ARTHUR ANDERSEN

Arthur Andersen's vision is to be the partner for success in the new economy. The firm helps clients find new ways to create, manage and measure value in the rapidly changing global economy. With world-class skills in assurance, tax, consulting and corporate finance, Arthur Andersen has more than 77,000 people in over 80 countries that are united by a single worldwide operating structure that fosters inventiveness, knowledge sharing and a focus on client success. Since its beginning in 1913, Arthur Andersen has realized 86 years of uninterrupted growth, with 1999 revenues over $7 billion. Arthur Andersen is a business unit of Andersen Worldwide.

Relevant Regulatory Papers on Disclosure

(1) Sound Practices for Loan Accounting and Disclosure (1999) , Basle Committee for Banking Supervision.

(2) Best Practices for Credit Risk Disclosure (1999), Basle Committee for Banking Supervision.

(3) Enhancing Bank Transparency (1998), Basle Committee for Banking Supervision.

(4) Recommendations for Public Disclosure and Derivatives Activities of Banks and Securities Firms , (1999) The Basle Committee and IOSCO.

(5) Core Principles for Effective Banking Supervision. (1997), Basle Committee for Banking Supervision.

(6) Principles for the Management of Credit Risk (1999) . Basle Committee for Banking Supervision.

Risk Library * Forum Documents