William P. Hannon
399 Park Avenue - 3rd Floor
New York, NY 10043
FAX - 212-793-4508

July 18, 2002

Jonathan G. Katz
Securities and Exchange Commission
450 Fifth Street, NW
Washington, D.C. 20549-0609

Re: Application of Critical Accounting Policies - File No. S7-16-02

Dear Mr. Katz:

Citigroup Inc. appreciates the opportunity to comment on the Securities and Exchange Commission's proposed rule on disclosure in Management's Discussion and Analysis of the application of critical accounting policies. Citigroup is committed to achieving the highest quality disclosures in its public filings, and we fully support the Commission's goal of improving the transparency and relevance of financial statements.

We agree with the overall objective of the proposed rule, and believe that a new section entitled "Application of Critical (or Significant) Accounting Policies" within MD&A will help users better understand a company's critical accounting estimates, as well as the initial adoption of accounting policies that have a material impact on a company's financial statements. However, as explained more fully below, we do not believe the manner and extent of disclosure the Commission proposes will be effective in providing greater transparency to investors, particularly the quantitative disclosure requirements of the proposal.

Critical Estimates

We understand the Commission's objective of better informing investors of a company's use of estimates in critical accounting areas. While we support that objective, we believe the need for clear and comprehensive disclosure must always be carefully balanced against the growing trend toward "disclosure overload," whereby excessive information is provided to readers in a manner that confuses or obfuscates the financial statements.

In our view, the proposal should focus on requiring more qualitative disclosures surrounding critical accounting estimates-that is, information that would help investors gain a better understanding of the estimation process surrounding a critical estimate, including a discussion of the key variables involved, and the subjective assessments made. However, we do not support the proposal's requirement regarding quantification of the sensitivities surrounding the estimate. Requiring registrants to quantify reasonably possible changes to numerous subjective input variables (or alternatively, to disclose the upper and lower ends of a range of reasonably possible estimates) does not, in our view, provide greater transparency to investors. To the contrary, we believe such a presentation would detract investors from the "best estimates" set by management through reasoned judgment and careful analysis, and would leave them uncertain about the overall credibility of the financial statements.

We find the examples in the proposal to be somewhat simplistic and not reflective of real-life complexities, at least not those in our business. Consider, for example, the disclosures that would be required for a diversified global company such as ours in an area like the allowance for credit losses. The allowance for credit losses is, of necessity, determined based on a variety of different input assumptions unique to each business and geographic region. Accordingly, it cannot be discussed adequately on an overall aggregate basis. Rather, to provide any kind of relevant disclosure in accordance with the proposal's requirements, a company such as ours would need to address the individual components of the allowance-by business and by region-and discuss the unique assumptions of each, followed by a quantification of sensitivities. Multiply that already extensive disclosure by the proposal's requirement to discuss changes to the estimates over the past three years, and an unwieldy set of disclosures results. We seriously question whether such voluminous information is relevant to investors and truly assists them in understanding a company's financial position.

Consider two additional examples dealing with investments and insurance claims. Many financial services companies hold investments, such as derivatives and certain debt instruments, that are required to be carried at fair values that must be estimated through use of sophisticated valuation models. Such models require input for many complex assumptions, including discount rates, yield curves, prepayment speeds, optionality, volatilities, cash flows, credit risk, currency risk and economic conditions, to name a few. Quantitatively presenting the proposed sensitivity disclosures for each of the assumptions and each of the many investments valued would not be meaningful to users.

Yet another example is that of claim liability estimates in the insurance industry. Workers' compensation claims, for instance, are based on numerous assumptions, including re-hospitalization rates, future nursing home costs, future drug treatment costs, inflation, and mortality rates for workers' compensation claimants suffering from lifetime injuries. Consider all of the many diverse claim types within the insurance industry and the many assumptions necessary for developing claim estimates, and the complexities of the quantitative disclosure requirements become readily apparent.

We are further concerned that the proposed disclosure requirements could result in presentation of proprietary information that is not in the best interests of companies to divulge. For example, presenting quantitative detail surrounding possible outcomes of a contentious lawsuit could jeopardize the company's defense and negotiations with plaintiffs.

In developing the sensitivity disclosures, registrants also would be faced with the burdensome task of describing how the estimates changed over a three-year period. We find this requirement particularly problematic for companies to implement upon transition to the proposed rule. Collection, summarization and presentation of three-year old quantitative data that has not previously been subject to financial reporting disclosure requirements would be prone to inaccuracies and be quite costly and burdensome for companies.

Notwithstanding our concerns regarding the proposal's quantitative requirements, we fully support clearly articulated qualitative disclosures that identify critical accounting estimates, describe the estimation process and key variables, and disclose the nature of the inherent sensitivities.

Discussions with Audit Committees

The proposal also would require companies to state, within MD&A, whether or not senior management discussed the development, selection and disclosure of the critical accounting estimates with the company's Audit Committee. We strongly believe in engaging our Audit Committee in robust and meaningful discussions regarding accounting matters and financial disclosure issues and, in fact, routinely do so. However, we disagree with this element of the proposal because it focuses on only a single area where an Audit Committee is actively involved in discussions with management, to the exclusion of many other aspects of financial reporting. To focus on only one area where discussions take place between senior management and the Audit Committee assigns more emphasis to the area of estimates than to other equally critical disclosures that affect investors' decisions-such as liquidity, market and credit risks, capital adequacy, and other risks facing companies.

We believe that the Audit Committee disclosures in registrants' proxy statements currently mandated by the Commission's rule requiring an annual Audit Committee Report (Item 306) are sufficient. The Audit Committee Report and Charter inform investors of the Audit Committee's role and the extent of its dependence on management and internal and external audit to provide them with appropriate information about the company's significant activities and the associated accounting treatment. The Audit Committee disclosure rules were designed to improve disclosure related to the functioning of corporate Audit Committees and to enhance the reliability and credibility of financial statements of public companies. We do not believe that further detail concerning discussions between senior management and the Audit Committee regarding estimates, or whether such discussions took place, merits separate disclosure in MD&A.

Selection of Acceptable Accounting Policies Upon Initial Adoption when Alternatives Exist

We fully support a thorough and clear description of any new accounting policy a company initially elects to apply. However, we do not believe that describing yet another existing acceptable alternative that the company did not elect to apply enhances an investor's understanding of the financial statements. Considering the complexity of accounting standards today, as well as "disclosure overload," the reader is not well served by discussing other accounting policy alternatives that are not reflected in a company's financial statements.


Citigroup is committed to high quality financial statement disclosures, and we fully support enhancing those disclosures to more fully meet investors' needs. We commend the Commission for moving in that direction by first requiring better disclosure of critical accounting policies, and now attempting to improve disclosures surrounding critical estimates. However, as discussed above, we respectfully believe that excessive quantitative disclosures surrounding estimates will not serve investors' needs, and instead may even detract from them.


William P. Hannon