SecProbe.io

CORRESP
1
filename1.htm

Correspondence Letter

 ACE Limited

 ACE Global Headquarters

17 Woodbourne Avenue

 Hamilton HM 08

 Bermuda

 PO Box 1015

 Hamilton HM DX

 Bermuda

 441 295-5200 main

 441 296-0087
fax

 www.acelimited.com

 December 23, 2005

 Jim B. Rosenberg

 Senior Assistant Chief Accountant

 Division of Corporate Finance

 United States Securities & Exchange

     Commission

 450 Fifth Street, N.W.

 Washington, DC. 20549

 USA

 RE: ACE Limited

 Form 10K/A for the
fiscal year ended December 31, 2004

 Filed August 12, 2005

 File No. 1-11778

 Dear Mr. Rosenberg,

 Thank you for your
letter dated December 12, 2005, setting forth comments of the staff of the Division of Corporation Finance (the “Staff”) on the Form 10-K/A for the year ended December 31, 2004 (the “2004 Form 10-K/A”) of ACE Limited
(the “Company” or “ACE”).

 We have considered the
Staff’s comments on the Form 10-K/A carefully and our responses to the Staff’s comments are set forth below. To facilitate the Staff’s review, the responses set forth below correspond to the paragraphs of your letter.

 Where we indicate that the Company will provide additional disclosure in future filings, such
disclosure will be included (unless specifically noted otherwise) in our Annual Report on Form 10-K for the year ended December 31, 2005 (the “2005 10-K”). After making the initial disclosures in response to your comments, comparable
disclosures will be made in subsequent Form 10-Ks and, to the extent such information is required to be included in a Form 10-Q, in subsequent Form 10-Qs.

 Form 10K/A – December 31, 2004

 Item 7. Management’s Discussion and Analysis of Financial Condition and Results of Operations

 Critical Accounting Estimates

1)      Comment:
We note on page 59 you include the reasonably likely changes in specific factors used to establish your loss reserves. We would recommend that you include discussion of reasonably
likely changes in your loss reserves in the ‘Critical Accounting Estimate’ section of MD&A.

Response:

We agree and will include the page 59 disclosure quantifying the uncertainty in loss reserve estimates within the Critical Accounting Estimate section of MD&A beginning with our
2005 10-K.

 Reinsurance Recoverables, page 36

2)      Comment:
It would appear that the default factor is the significant assumption used to determine the estimate of uncollectible reinsurance. Please explain and quantify for us, in
disclosure-type format the impact that reasonably likely changes in the default factor would have on reported results, financial position and liquidity. It would appear based on your determination of the estimate, as disclosed in the bullets on
pages 36-37, the proposed disclosure should differentiate between the different ways you determine the default factors (i.e. reinsurers with credit ratings, reinsurers with no credit ratings and reinsurers that are insolvent).

Response:

We agree with your observation and intend to clarify and further quantify the specific judgments involved in choosing default factors. Within our existing disclosure of reinsurance
recoverables in the 10-K/A for the year ended December 31, 2004, we provide the following quantitative disclosure regarding the sensitivity of our estimate for the bad debt provision as well as the related impact on reported results, financial
position, and liquidity:

“Given the various considerations used to develop default factors, we cannot precisely quantify the effect a specific industry event may have on the bad debt provision.
However, based on the composition (particularly the average credit quality) of the reinsurance recoverable balance at December 31, 2004, we estimate that a ratings downgrade of all rated reinsurers could increase our bad debt provision by as
much as 15 percent, assuming no other changes relevant to the calculation. While a rating downgrade would result in an increase in our provision for bad debt and a charge to earnings in that period, a downgrade in and of itself does not imply that
we will not ultimately collect all of the ceded reinsurance recoverable from the reinsurers in question.”

We intend to update this disclosure for the 2005 10-K. In addition, and pursuant to your comment, we will clarify the nature of judgments used to determine default factors for
various types of recoverables. Further, we will illustrate the sensitivity of default factors by disclosing, in tabular form, the reinsurance recoverable balances and the bad debt provision for each type of reinsurance recoverable. Due to the lack
of a clear alternative to base a numerical sensitivity analysis for default factors, we cannot precisely quantify a range of the bad debt provision. However, we believe the amended disclosure will provide readers with an improved understanding of
the judgments used to develop default factors and potential quantitative effects such judgments have on the financial statements.

 Page 2/7

Accordingly, in the 2005 10-K we will provide revised disclosure on default factors as follows:

“Generally, we use a default analysis to estimate uncollectible reinsurance. The primary components of the default analysis are reinsurance recoverable balances by reinsurer,
net of collateral, and default factors used to determine the portion of a reinsurer’s balance deemed uncollectible. The definition of collateral for this purpose requires some judgment and is generally limited to assets held in trust, letters
of credit, and liabilities held by us with the same legal entity for which we believe there is a right of offset. The determination of the default factor is principally based on the financial rating of the reinsurer and a default factor applicable
to the financial rating. Default factors require considerable judgment and are determined using the current rating, or rating equivalent, of each reinsurer as well as other key considerations and assumptions. The more significant considerations and
assumptions include, but are not necessarily limited to, the following:

•

For reinsurers that maintain a financial rating from a major ratings agency, and for which recoverable balances are considered representative of the larger population (i.e., default
probabilities are consistent with similarly rated reinsurers and payment durations conform to averages), the judgment exercised by management to determine the bad debt provision of each reinsurer is typically limited because the financial rating is
based on a published source and the default factor we apply is based on a default factor of a major ratings agency applicable to the particular rating class. Default factors applied for financial ratings of AAA, AA, A, BBB, BB, B, CCC, are .5%,
1.2%, 1.9%, 4.7%, 9.6%, 23.8%, and 49.7%, respectively. Because the model we use is predicated on the default factors of a major ratings agency, we do not generally consider alternative factors. However, when a recoverable is expected to be paid in
a brief period of time by a highly rated reinsurer, such as certain property catastrophe claims, a default factor may not be applied.

•

For balances recoverable from reinsurers that are both unrated by a major rating agency and for which management is unable to determine a credible rating equivalent based on a
parent, affiliate or peer company, we determine a rating equivalent based on an analysis of the reinsurer that considers an assessment of the creditworthiness of the particular entity, industry benchmarks, or other factors as considered appropriate.
We then apply the applicable default factor for that rating class.

•

For balances recoverable from reinsurers that are either insolvent or under regulatory supervision, we establish a default factor and resulting bad debt provision based on specific
facts and circumstances surrounding each company. Upon initial notification of an insolvency, we generally recognize expense for a substantial portion of all balances outstanding, net of collateral, through a combination of write-offs of recoverable
balances and increases to the bad debt provision. When regulatory action is taken on a reinsurer, we generally recognize a default factor by estimating an expected recovery on all balances outstanding, net of collateral. When sufficient information
becomes available from managers of the reinsurer, we adjust the bad debt provision by establishing a default factor pursuant to information received.

 Page 3/7

•

For recoverables arising from structured settlements, pools, and disputed balances, management determines a default factor based on specific facts and circumstances.

The following table summarizes reinsurance recoverables and the bad debt provision for each type of recoverable balance as of September 30, 2005 (in millions).

 Type

Total
Reinsurance
Recoverables

Recoverable,
net of usable
collateral

Bad Debt
Provision

 Reinsurers with Credit Ratings

$
10,365

$
9,964

$
179

 Reinsurers Not Rated

$
2,939

$
2,682

$
253

 Reinsurers under Supervision

$
435

$
410

$
256

 Other—Structured settlements, pools, disputed balances

$
2,702

$
1,789

$
91

 Total

$
16,441

$
14,845

$
779

For purposes of the upcoming 2005 10-K disclosure, please note that the table above will be updated to reflect balances at December 31, 2005.

 Consolidated Financial Statements

 Notes to Consolidated Financial Statements

 Note 3. Restatement of Previously Issued Financial Statements

 Reinsurance Contracts Purchased, page F-22

3)      Comment:
We refer to the multi-year aggregate excess of loss reinsurance contract purchased in 1997 by a company acquired by ACE in 1999. Please explain to us how the change in accounting
for this reinsurance contract resulted in an increase in goodwill. In your response please tell us how you accounted for the transaction prior to the restatement under SFAS 113 and how you accounted for the contract under SOP 98-7 (i.e. what risk
was not provided for in the contract that caused deposit accounting) for each year. It would also be helpful to show us how the $87 million increase in goodwill was calculated.

Response:

 Prior to the restatement, the “multi-year aggregate excess of loss reinsurance contract” was accounted for as prospective reinsurance. As a result, ceded
premiums were recognized over the term of the

 Page 4/7

contract and ceded losses were adjusted each period to reflect the extent reported losses of the ceding companies exceeded attachment points for each year
covered under the contract subject to policy limits. In connection with the Company’s investigation into non-traditional products, we discovered letters between the acquired company and each of its reinsurers. Our review of these letters led us
to conclude that the original parties to the contract did not intend to transfer risk over the 7-year term of their original 1997 agreement. Accordingly, we concluded that the accounting for the contracts was in error, and we applied the deposit
model from SOP 98-7 that is consistent with contracts that have timing but no underwriting risk.

In applying deposit accounting, we concluded that the contract should have been reported as a deposit from its inception in 1997, consistent with the date of the letters, through
2005. Because this error preceded the date ACE acquired the companies that entered into the reinsurance contract, the restatement affected purchase accounting for the 1999 acquisition and the related determination of goodwill. Specifically, the
acquired balance sheet included an asset for reinsurance recoverables under the contract of $276.5 million that reflected the expected future loss recoveries under the contract; however, under deposit accounting, the acquired balance sheet should
have included a deposit asset of $165 million reflecting the amount of the deposit. The reduction of this asset serves to reduce the acquired net assets thereby increasing the excess of purchase price over acquired net assets. A reconciliation of
these items to the $87 million increase in goodwill at December 31, 2004 is as follows (in millions):

 Adjustments to Purchase Accounting

 Reversal of Reinsurance Recoverable

$
276.5

 Establishment of a Deposit Asset

(165.0
)

 Corresponding increase in deferred tax assets

(19.0
)

 Reduction of Surplus of Acquired Company/Increase in Goodwill

$
92.5

 Amortization of additional goodwill from July 1, 1999 through December 31, 2001

(5.5
)

 Increase in Goodwill at December 31, 2004 from change in accounting from reinsurance to deposit

$
87.0

 The change in accounting from
reinsurance to deposit accounting for tax purposes had the effect of restating the book-tax basis temporary differences relating to the multi-year aggregate excess of loss reinsurance contract as of the purchase date by $54 million, which represents
the estimated effect on tax basis discounted loss reserves. This change results in a deferred tax asset of $19 million.

 Note 7. Unpaid Losses and Loss Expenses

 Property and Casualty, page F-35

4)      Comment:
 We refer to the second paragraph below your loss reserve rollforward. You state in

 Page 5/7

your disclosure that you exclude the effects of changes in your assumed premium estimate related to prior accident years from your prior year loss reserve
development. Please explain to us the rationale for including this change in estimate in your current year loss reserve development. The change in the premium estimate would appear to be no different, for example, than a change in a case reserve
estimate established in a prior year based on information received in the current year. It would appear that you are changing the premium estimate related to prior accident years based on current information. In addition, to help us understand the
impact of this classification within your rollforward please quantify the impact this has had on roll-forwards presented for the three years ended December 31, 2004.

Response:

We exclude from prior year loss reserve development in the reserve rollforward the effects of losses arising from changes in the earned premium estimates for the following reasons:

•

Our disclosure of losses incurred related to prior years is a critical measure of the accuracy of our loss reserve estimates from prior years and the impact changes in those
estimates have on current year results. Readers of our financial statements use this disclosure to assess the accuracy of our loss reserve estimates, and potentially consider this amount when assessing the expected future profitability of our
company.

Because the related premium exposures were not known and not recognized in prior calendar years, it would have been inappropriate to have recognized related loss reserves in prior
calendar years. Accordingly, we believe the inclusion in prior year development of losses arising from changes in prior year estimates of assumed premium is inconsistent with the principal objective of the disclosure, a measure of the accuracy of
prior year loss reserve estimates.

•

It is reasonable to consider losses recognized in connection with an increase in premium estimates to be those of the current calendar year since the premium exposures and related
recorded earned premium giving rise to the changed loss estimate are first being reported in the current calendar year.

•

Through the loss reserve rollforward, a reader of our financial statements can gain insight into our loss ratio for current calendar year business by comparing current year losses
incurred within the rollforward to premiums earned for the current calendar year. Because changes in premium estimates are reflected in current calendar year premiums, the reflection of corresponding loss reserve increases in current year losses
incurred ensures consistent reflection of these premiums and related losses in this ratio.

 Pursuant to your request, for the years ended December 31, 2004, 2003, and 2002, losses arising from the developm