Enstar Group Limited
P.O. Box HM 2267
Windsor Place, 3rd Floor, 18 Queen Street
Hamilton HM JX, Bermuda
December 15, 2011
Via EDGAR
Frank Wyman
Staff Accountant
Division of Corporation Finance
United States Securities and Exchange Commission
100 F. Street, N.E.
Washington, D.C. 20549
| Re: | Enstar Group Limited |
Form 10-K for Fiscal Year Ended December 31, 2010
Filed March 7, 2011
File No. 001-33289
Dear Mr. Wyman:
Enstar Group Limited (the “Company” or “we”) has carefully considered the comments you have raised in telephone calls with us over the last several weeks relating to our October 12, 2011 response to your September 29, 2011 comment letter. Attached are proposed revisions to our disclosure regarding our run-off strategy and related reserving methodologies that we intend to incorporate into future filings, as appropriate.
If you have any questions about the attached, please do not hesitate to contact me at (441) 278-1445.
| Sincerely, |
| /s/ Richard J. Harris |
| Richard J. Harris |
| Chief Financial Officer |
| cc: | Joel Parker (Securities and Exchange Commission) |
Don Abbott (Securities and Exchange Commission)
Robert C. Juelke, Esq. (Drinker Biddle & Reath LLP)
John Johnston (Deloitte & Touche Ltd., Hamilton, Bermuda)
Exhibit A
In future filings, the following underlined disclosure would amend or supplement the disclosure included in Item 1 – Business, Item 7 – Management’s Discussion and Analysis of Financial Condition and Results of Operations, and the “Loss and Loss Adjustment Expense” footnote to the Consolidated Financial Statements included in the Company’s Form 10-K for the fiscal year ended December 31, 2010. The Company may seek to eliminate the redundancy created by including the same disclosure in three separate sections of its future filings, but the salient points will be included in such filings.
Year Ended December 31, 2010
The net reduction in ultimate loss and loss adjustment expense liabilities for the year ended December 31, 2010 was $311.8 million, excluding the impact of foreign exchange rate movements of $3.8 million and including both net reduction in ultimate loss and loss adjustment expense liabilities of $19.0 million relating to companies and portfolios acquired during the year and premium and commission adjustments triggered by incurred losses of $16.5 million.
The net reduction in ultimate loss and loss adjustment expense liabilities for the year ended December 31, 2010 of $311.8 million was attributable to a reduction in estimates of net ultimate losses of $278.1 million, a reduction in aggregate provisions for bad debts of $49.6 million and a reduction in estimates of unallocated loss adjustment expense liabilities of $39.7 million, relating to 2010 run-off activity, partially offset by the amortization, over the estimated payout period, of fair value adjustments relating to companies acquired amounting to $55.4 million.
The reduction in estimates of net ultimate losses of $278.1 million comprised net incurred favorable loss development of $41.1 million and reductions in IBNR reserves of $236.9 million. The decrease in the aggregate estimate of IBNR loss reserves of $236.9 million (compared to $318.2 million during the year ended December 31, 2009) was comprised of $67.8 million relating to asbestos liabilities (compared to $158.4 million in 2009), $4.2 million relating to environmental liabilities (compared to $17.0 million in 2009) and $164.9 million relating to all other remaining liabilities (compared to $142.8 million in 2009). The aggregate reduction in IBNR of $236.9 million was a result of the application, on a basis consistent with the assumptions applied in the prior period, of our actuarial methodologies to revised historical loss development data following 90 commutations to estimate loss reserves required to cover liabilities for unpaid losses and loss adjustment expenses relating to non-commuted exposures. The prior period estimate of aggregate net IBNR liabilities was reduced as a result of the combined impact on all classes of business of loss development activity during 2010, including commutations and the favorable trend of loss development related to non-commuted policies compared to prior forecasts. The lower reduction in asbestos IBNR reserves during 2010 was primarily due to reduced commutations of asbestos related exposures compared to the prior year. Total net loss reserves acquired from January 1, 2008 to December 31, 2010 amounted to $3,197.3 million, of which $2,634.5 million, or 82.4% related to all other losses. This increase in all other loss reserves provided the basis for a greater reduction in all other IBNR reserves. The net incurred favorable loss development of $41.1 million, resulting from settlement of net advised case and LAE reserves of $336.1 million for net paid losses of $295.0 million, related to the settlement of non-commuted losses in the year and approximately 90 commutations of assumed and ceded exposures. Net incurred liabilities settled by way of commutation during the year ended December 31, 2010 amounted to $109.7 million compared to the net reduction in advised case reserves during the same period of $336.1 million. Commutations provide an opportunity for us to exit exposures to entire policies with insureds and reinsureds at a discount to the previous estimated ultimate liability. As a result of exiting all exposures to such policies, all advised case reserves and IBNR liabilities relating to that insured or reinsured are eliminated. This often results in a net gain irrespective of whether the settlement exceeds the advised case reserves. We adopt a disciplined approach to the review and settlement of non-commuted claims through claims adjusting and the inspection of underlying policyholder records such that settlements of assumed exposures may often be achieved below the level of the originally advised loss, and settlements of ceded receivables may often be achieved at levels above carried balances. Of the 90 commutations completed during 2010, three related to our top ten insured and/or reinsured exposures, including one commutation completed shortly after December 31, 2009 whereby the related reduction in IBNR reserves was recorded in the reduction in net ultimate losses for the year ended December 31, 2009, and one related to the commutation of one of our largest ceded reinsurance assets. The remaining 86 commutations, of which approximately 43% were completed during the three months ended December 31, 2010, were of a smaller size, consistent with our approach of targeting significant numbers of cedant and reinsurer relationships, as well as targeting significant individual cedant and reinsurer relationships. The combination of the claims settlement activity in 2010, including commutations (but excluding the impact of the commutation that was completed subsequent to the year ended December 31, 2009) and the actuarial estimation of IBNR reserves required for the remaining non-commuted exposures (which took into account the favorable trend of loss development in 2010 related to such exposures compared to prior forecasts), resulted in our management concluding that the loss development activity that occurred subsequent to the prior reporting period provided sufficient new information to warrant a reduction in IBNR reserves of $236.9 million in 2010.
A-1
The reduction in aggregate provisions for bad debt of $49.6 million was a result of the collection, primarily during the three months ended December 31, 2010, of certain reinsurance receivables against which bad debt provisions had been provided in earlier periods.
Year Ended December 31, 2009
The net reduction in ultimate loss and loss adjustment expense liabilities for the year ended December 31, 2009 was $259.6 million, excluding the impact of adverse foreign exchange rate movements of $73.5 million and including both net reduction in ultimate loss and loss adjustment expense liabilities of $4.8 million relating to companies acquired during the year and premium and commission adjustments of $5.5 million triggered by incurred losses.
The net reduction in ultimate loss and loss adjustment expense liabilities for the year ended December 31, 2009 of $259.6 million was attributable to a reduction in estimates of net ultimate losses of $274.8 million, a reduction in aggregate provisions for bad debts of $11.7 million and a reduction in estimates of loss adjustment expense liabilities of $50.4 million, relating to 2009 run-off activity, partially offset by the amortization, over the estimated payout period, of fair value adjustments relating to companies acquired amounting to $77.3 million.
The reduction in estimates of net ultimate losses of $274.8 million comprised net incurred loss development of $43.3 million and reductions in IBNR reserves of $318.2 million. The decrease in the estimate of IBNR loss reserves of $318.2 million (compared to $210.4 million for the year ended December 31, 2008) was comprised of $158.4 million relating to asbestos liabilities (compared to $101.5 million in 2008), $17.0 million relating to environmental liabilities (compared to $10.8 million in 2008) and $142.8 million relating to all other remaining liabilities (compared to $98.1 million in 2008). The reduction in IBNR is a result of the application, on a basis consistent with the assumptions applied in the prior period, of our actuarial methodologies to loss data to estimate loss reserves required to cover liabilities for unpaid losses and loss adjustment expenses. The prior period estimate of net IBNR liabilities was reduced as a result of the combined impact of loss development activity during 2009, including commutations and the favorable trend of loss development related to non-commuted policies compared to prior forecasts. The larger reduction in asbestos IBNR reserves in 2009 was primarily due to the further commutation of asbestos reserves relating to one of our insurance entities that had benefited from a substantial stop loss protection until December 18, 2008. The larger reduction in all other IBNR reserves in 2009 was primarily due to incurred loss development in one of our entities acquired during 2008 that was in line with incurred loss development expected by our external actuaries and was, therefore, offset by a corresponding reduction in IBNR reserves, as well as the completion of a commutation by the same entity subsequent to the year-end referred to below. The net incurred loss development of $43.3 million resulting from settlement of net advised case and LAE reserves of $214.1 million for net paid losses of $257.4 million, related to the settlement of non-commuted losses in the year and approximately 79 commutations of assumed and ceded exposures. Of the 79 commutations completed during 2009, two related to our top ten insured and/or reinsured exposures. The remaining 77 were of a smaller size, consistent with our approach of targeting significant numbers of cedant and reinsurer relationships, as well as targeting significant individual cedant and reinsurer relationships. Approximately 76% of commutations completed in 2009 related to commutations completed during the three months ended December 31, 2009. Net incurred liabilities settled by way of commutation during the year ended December 31, 2009 amounted to $81.9 million compared to the net reduction in advised case reserves during the same period of $214.1 million. Subsequent to the year end, one of our insurance entities completed a commutation of another of one of our top ten reinsured exposures. The combination of the claims settlement activity in 2009, including commutations, and the actuarial estimation of IBNR reserves required for the remaining non-commuted exposures (which took into account the favorable trend of loss development in 2009 related to such exposures compared to prior forecasts as well as the impact of the commutation that was completed subsequent to the year-end), resulted in our management concluding that the loss development activity that occurred subsequent to the prior reporting period provided sufficient new information to warrant a reduction in IBNR reserves of $318.2 million in 2009.
The reduction in aggregate provisions for bad debt of $11.7 million was a result of the collection, primarily during the three months ended March 31, 2009, of certain reinsurance receivables against which bad debt provisions had been provided in earlier periods.
Year Ended December 31, 2008
The net reduction in ultimate loss and loss adjustment expense liabilities for the year ended December 31, 2008 was $242.1 million, excluding the impacts of favorable foreign exchange rate movements of $36.1 million (relating to companies acquired
A-2
in 2007 and earlier) and including both net reduction in ultimate loss and loss adjustment expense liabilities of $149.4 million relating to companies acquired during the year and premium and commission adjustments of $0.1 million triggered by incurred losses.
The net reduction in ultimate loss and loss adjustment expense liabilities for 2008 of $242.1 million was attributable to a reduction in estimates of net ultimate losses of $161.4 million, a reduction in aggregate provisions for bad debt of $36.1 million (excluding $3.1 million relating to one of our entities that benefited from substantial stop loss reinsurance protection discussed below) and a reduction in estimates of loss adjustment expense liabilities of $69.1 million, relating to 2008 run-off activity, partially offset by the amortization, over the estimated payout period, of fair value adjustments relating to companies acquired amounting to $24.5 million.
The reduction in estimates of net ultimate losses of $161.4 million comprised the following:
(i) A reduction in estimates of net ultimate losses of $21.7 million in one of our insurance entities that benefited from substantial stop loss reinsurance protection. Net incurred loss development relating to this entity of $21.6 million was offset by reductions in IBNR reserves of $94.8 million and reductions in provisions for bad debt of $3.1 million, resulting in a net reduction in estimates of ultimate losses of $76.3 million. The entity in question benefited, until December 18, 2008, from substantial stop loss reinsurance protection whereby $54.6 million of the net reduction in ultimate losses of $76.3 million was ceded to a single AA- rated reinsurer such that we retained a reduction in estimates of net ultimate losses relating to this entity of $21.7 million. On December 18, 2008, we commuted the stop loss reinsurance protection with the reinsurer for the receipt of $190.0 million payable by the reinsurer to us over four years together with interest compounded at 3.5% per annum. The commutation resulted in no significant financial impact to us. The decrease in the estimate of IBNR loss reserves of $94.8 million for this one insurance entity was comprised of $77.7 million relating to asbestos liabilities, $9.0 million relating to environmental liabilities and $8.1 million relating to all other remaining liabilities. The reduction in IBNR is a result of the application, on a basis consistent with the assumptions applied in the prior period, of our actuarial methodologies to loss data to estimate loss reserves required to cover liabilities for unpaid losses and loss adjustment expenses. The prior period estimate of net IBNR liabilities was reduced as a result of the combined impact of loss development activity during 2008, which was comprised of the settlement of certain advised case reserves below their prior period carried amounts, commutations completed and the trend of loss development relating to non-commuted policies compared to prior forecasts. The net incurred loss development relating to this entity of $21.6 million, whereby advised net case reserves of $25.0 million were settled for net paid losses of $46.6 million, primarily related to six commutations of assumed and ceded liabilities completed during 2008. As a result of exiting all exposures to such policies, all advised case reserves and IBNR liabilities relating to that insured or reinsured were eliminated. This often results in a net gain irrespective of whether the settlement exceeds the advised case reserves. Of the six commutations completed for this entity, of which the three largest were completed during the three months ended December 31, 2008, one was among its top ten assumed exposures. The remaining five commutations were of a smaller size, consistent with our approach of targeting significant numbers of cedant and reinsurer relationships as well as targeting significant individual cedant and reinsurer relationships. The combination of the claims settlement activity in 2008, including commutations, combined with the actuarial estimation of IBNR reserves required for the remaining non-commuted exposures (which took into account the favorable trend of loss development in 2008 related to such exposures compared to prior forecasts), resulted in our management concluding that the loss development activity that occurred subsequent to the prior reporting period provided sufficient new information to warrant a reduction in IBNR reserves of $94.8 million for this one insurance entity in 2008.
(ii) A reduction in estimates of net ultimate losses of $139.7 million in our other insurance and reinsurance entities comprised net favorable incurred loss development of $24.1 million and reductions in IBNR reserves of $115.6 million. The decrease in the estimate of IBNR loss reserves of $115.6 million was comprised of $23.8 million relating to asbestos liabilities, $1.8 million relating to environmental liabilities and $90.0 million relating to all other remaining liabilities. The reduction in IBNR is a result of the application, on a basis consistent with the assumptions applied in the prior period, of our actuarial methodologies to loss data to estimate loss reserves required to cover liabilities for unpaid losses and loss adjustment expenses. The prior period estimate of net IBNR liabilities was reduced as a result of the combined impact of favorable loss development activity during 2008, which was comprised of the settlement of advised case reserves below their prior period carried amounts, commutations completed and the favorable trend of loss development related to non-commuted policies compared to prior forecasts. The net favorable incurred loss development in our remaining insurance and reinsurance entities of $24.1 million, whereby net advised case and LAE reserves of $123.5 million were settled for net paid losses of $99.4 million, primarily related to the settlement of non-commuted losses in the year below carried reserves and approximately 59 commutations of assumed and ceded exposures at less than case and LAE reserves. Of the 59 commutations completed during 2008 for our other reinsurance and insurance companies, two (both of which were completed during the three months ended December 31, 2008) were among our top ten insured and/or reinsured exposures. The remaining 57 were of a smaller size, consistent with our approach of targeting significant numbers of cedant and reinsurer relationships, as well as targeting significant individual cedant and reinsurer relationships.
A-3
The aggregate reduction in IBNR reserves during the year ended December 31, 2008 amounted to $210.4 million and was comprised of $101.5 million relating to asbestos IBNR reserves, $10.8 million relating to pollution IBNR reserves and $98.1 million relating to all other loss IBNR reserves. The aggregate reduction in IBNR reserves during the year ended December 31, 2007 amounted to $39.4 million and was comprised of $23.0 million relating to asbestos IBNR reserves, $4.7 million relating to pollution IBNR reserves and $11.7 million relating to all other loss IBNR reserves. The larger reduction in asbestos IBNR reserves of $78.5 million during the year ended December 31, 2008 compared to 2007 was primarily due to commutations concluded during the year relating to one of our insurance entities that had benefited from a substantial stop loss reinsurance protection until December 18, 2008. The larger reduction in all other IBNR reserves of $86.4 million during the year ended December 31, 2008 compared to 2007 was primarily due to the acquisition of a significant amount of all other loss reserves during 2008, which developed favorably during the year. Total net loss reserves acquired during the year ended December 31, 2008 amounted to $1,781.3 million, of which $1,449.0 million related to all other loss reserves.
Approximately 82% of commutations completed in 2008 related to commutations completed during the three months ended December 31, 2008. Net incurred liabilities settled by way of commutation during the year ended December 31, 2008 amounted to $39.2 million compared to the net reduction in non-commuted advised case reserves during the same period of $147.6 million. The combination of the claims settlement activity in 2008, including commutations, with the actuarial estimation of IBNR reserves required for the remaining noncommuted exposures (which took into account the favorable trend of loss development in 2008 related to such exposures compared to prior forecasts), resulted in our management concluding that the loss development activity that occurred subsequent to the prior reporting period provided sufficient new information to warrant a reduction in IBNR reserves of $115.6 million for our remaining insurance and reinsurance entities in 2008.
One of our reinsurance companies had retrocessional arrangements providing for full reinsurance of all risks assumed. During the year, this entity commuted its largest assumed liability and related retrocessional protection whereby the subsidiary paid net losses of $222.0 million and reduced net IBNR by the same amount, resulting in no gain or loss to us.
The reduction in aggregate provisions for bad debt of $36.1 million (excluding $3.1 million relating to one of our entities that benefited from substantial stop loss reinsurance protection discussed above) was comprised of: (1) $13.7 million as a result of the collection, primarily during the three months ended December 31, 2008, of certain reinsurance receivables against which bad debt provisions had been provided in earlier periods, (2) $8.5 million as a result of the revision of estimates of bad debt provisions following the receipt of new information during the three months ended December 31, 2008 and (3) $13.9 million as a result of reduced exposures to reinsurers with bad debt provisions following the commutation of assumed liabilities.
A-4
Exhibit B
In future filings, the following highlighted disclosure would amend or supplement the disclosure included in the “Critical Accounting Policies” section of Item 7 – Management’s Discussion and Analysis of Financial Condition and Results of Operations and the “Significant Accounting Policies” and “Acquisitions” footnotes to the Consolidated Financial Statements included in the Company’s Form 10-K for the fiscal year ended December 31, 2010. The Company may seek to eliminate the redundancy created by including similar disclosure in three separate sections of its future filings, but the salient points will be included in such filings.
Management’s Discussion and Analysis of Financial Condition and Results of Operations
Critical Accounting Policies
The Company’s primary objective in running off the operations of an acquired company is to effect an orderly and efficient settlement of all liabilities and assets and, in so doing, to strive to achieve savings in the settlement of such amounts in relation to the values implied by the purchase price of the transaction. The Company’s run-off process is led by disciplined management and includes the adjustment and settlement of valid claims, commutations of exposures, disciplined collection of reinsurance receivables, achievement of early finality of the acquired run-off by way of solvent scheme of arrangement (if available) and imposition of strong financial and operational governance over acquired companies.
The most significant liability and asset of an acquired company are typically the liability for loss and loss adjustment expenses and the asset related to any reinsurance recoverable on these liabilities that may be contractually due to the acquired entity. The market for acquisition of run-off companies is not sufficiently active and transparent to enable the Company to identify reliable, market exit values for acquired assets and liabilities. Accordingly, consistent with provisions of U.S. GAAP, the Company has developed internal models that it believes allow it to determine fair values that are reasonable proxies for market exit values. The Company is familiar with the major participants in the acquisition run-off market and believes that the key assumptions it makes in valuing acquired assets and liabilities are consistent with the kinds of assumptions made by such market participants. Furthermore, in the Company’s negotiation of purchase price with sellers, it is frequently clear to the Company that other bidders in the market are using models and assumptions similar in nature to the Company’s during the competitive bid process. The majority of acquisitions are completed following a public tender process whereby the seller invites market participants to provide bids for the target acquisition.
The Company accounts for acquisitions using the purchase method of accounting, which requires that the acquirer record the assets and liabilities acquired at their estimated fair value. The fair values of each of the reinsurance assets and liabilities acquired are derived from probability-weighted ranges of the associated projected cash flows, based on actuarially prepared information and management’s run-off strategy. The Company’s run-off strategy, as well as that of other run-off market participants, is expected to be different from the seller’s as generally sellers are not specialized in running off insurance and reinsurance liabilities whereas the Company and other market participants do specialize in such run-offs.
The key assumptions used by the Company and, it believes, by other run-off market participants in the fair valuation of acquired companies are (i) the projected payout, timing and amounts of claims liabilities; (ii) the related projected timing and amount of reinsurance collections; (iii) a risk-free discount rate, which is applied to determine the present value of the future cash flows; (iv) the estimated unallocated loss adjustment expenses to be incurred over the life of the run-off; (v) the impact that any accelerated run-off strategy may have on the adequacy of acquired bad debt provisions; and (vi) an appropriate risk margin.
The probability-weighted projected cash flows of the acquired company are based on projected claims payouts provided by the seller predominantly in the form of the seller’s most recent independent actuarial reserve report. In the absence of the seller’s actuarial reserve report, the Company’s independent actuaries will determine the estimated claims payout.
With respect to the Company’s U.K., Bermudian and Australian insurance and reinsurance subsidiaries, the Company is able to pursue strategies to achieve complete finality and conclude the run-off of a company by promoting solvent schemes of arrangement. Solvent schemes of arrangement are a popular means of achieving financial certainty and finality for insurance and reinsurance companies incorporated or managed in the U.K., Bermuda and Australia by making a one-time full and final settlement of an insurance and reinsurance company’s liabilities to policyholders. On acquisition of a U.K., Bermudian or Australian company, the claims payout projection is weighted according to management’s estimated probability of being able to complete a solvent scheme of arrangement. To the extent that solvent schemes of arrangement are not available to an acquired company, no weighting is applied to the projected claims payout.
B-1
On acquisition, the Company makes a provision for unallocated loss adjustment expense liabilities. This provision considers the adequacy of the provision maintained and recorded by the seller in light of the Company’s run-off strategy and estimated unallocated loss adjustment expenses to be incurred over the life of the acquired run-off as projected by the seller’s actuaries or, in their absence, the Company’s actuaries. To the extent that the Company’s estimate of the total unallocated loss adjustment expense provision is different from the seller’s, an adjustment will be made. While it is the objective of the Company to accelerate the run-off by completing commutations of assumed and ceded business (which would have the effect of shortening the life, and therefore the cost, of the run-off), the success of this strategy is far from certain. As a result, the estimates of unallocated loss adjustment expenses are based on running off the liabilities and assets over the actuarially projected life of the run-off, which the Company considers to be a prudent basis. In those domiciles where solvent schemes of arrangement are available, management’s estimates of the total unallocated loss adjustment expenses are probability-weighted in accordance with the estimated time that a solvent scheme of arrangement could be completed, which has the effect of reducing the period of the run-off and the related unallocated loss adjustment expenses. For those acquisitions in domiciles where solvent schemes of arrangement are not available, the unallocated loss adjustment expenses are estimated over the projected life of the run-off.
The Company believes that providing for unallocated loss adjustment expenses based on the Company’s run-off strategy is appropriate in determining the fair value of the assets and liabilities acquired in an acquisition of a run-off company. The Company believes that other participants in the run-off acquisition marketplace factor into the price to pay for an acquisition the estimated cost of running off the acquired company based on how that participant expects to manage the assets and liabilities.
The difference between the carrying value of reserves acquired at the date of acquisition and the fair value is the Fair Value Adjustment, or FVA. The FVA is amortized over the estimated payout period and adjusted for accelerations on commutation settlements or any other new information or subsequent change in circumstances after the date of acquisition. To the extent the actual payout experience after the acquisition is materially faster or slower than anticipated at the time of the acquisition, there is an adjustment to the estimated ultimate loss reserves, or there are changes in bad debt provisions or in estimates of future run-off costs following accelerated payouts, then the amortization of the FVA is accelerated or decelerated, as the case may be, to reflect such changes.
Loss and Loss Adjustment Expenses
Our primary objective in running off the operations of acquired companies and portfolios of insurance and reinsurance business in run-off is to increase book value by settling loss reserves below their acquired fair value. The earnings created in each acquired company or portfolio of insurance and reinsurance business, together with the related decrease in loss reserves, leads to a reduction in the capital required for each company, thereby providing the ability to distribute both earnings and excess capital to the parent company.
To the extent that the nature of the acquired loss reserves are conducive to commutation, our aim is to settle the majority of the acquired loss reserves within a timeframe of approximately 5 to 7 years from the date of acquisition. To the extent that acquired reserves are not conducive to commutation, we will instead adopt a disciplined claims management approach to pay only valid claims on a timely basis and endeavor to reduce the level of acquired loss adjustment expense provisions by withdrawing, where appropriate, from existing litigation and otherwise streamlining claims handling procedures.
By adopting either of the above run-off strategies, we would expect that over the targeted life of the run-off, acquired ultimate loss reserves would settle below their recorded value, resulting in reductions in ultimate loss and loss adjustment expense liabilities. There can be no assurance, however, that we will successfully implement our strategy. At the beginning of each year we prepare our projections of potential ultimate reserve releases for the year based on the assumptions that each company achieves its commutation targets and that non-commuted reserve development is better than that expected by our external actuaries. For each commutation target, a settlement objective is estimated. Probability of success weightings and assumptions of IBNR to case reserve ratios are applied to provide an indication of potential ultimate reserve savings that may be achieved from the target commutations.
Commutations of blocks of policies, along with disciplined claims management, have the potential to produce favorable claims development compared to established reserves. For each newly-acquired company, we determine a commutation strategy that broadly identifies commutation targets using the following criteria:
1. Previous commutations completed by existing portfolio companies with policyholders of the newly-acquired company;
2. Nature of liabilities;
3. Size of incurred loss reserves;
B-2
4. Recent loss development history; and
5. Targets for claims audits.
Once commutation targets are identified they are prioritized into target years of completion. At the beginning of each year, the approach to commutation negotiations is determined by the commutation team, including claims and exposure analysis and broker account reconciliations. On completion of this analysis, settlement parameters are set around incurred liabilities. Commutation discussions can take many months or even years to come to fruition. Commutation targets not completed in a particular year are re-prioritized for the following year.
Every commutation, irrespective of value, requires the approval of our chief financial officer or one of our two joint chief operating officers. For each commutation settled within the guideline settlement parameters, there is an expectation that there will be a favorable impact on the IBNR reserve when the annual actuarial review is completed. However, if a significant commutation is completed during the year loss reserves will be adjusted in the corresponding quarter to reflect our best estimate of the impact.
The following table provides a breakdown of gross loss and loss adjustment expense reserves by type of exposure as of December 31, 2010 and 2009:
| 2010 | 2009 | |||||||||||||||||||||||
| OLR | IBNR | Total | OLR | IBNR | Total | |||||||||||||||||||
| (in thousands of U.S. dollars) | ||||||||||||||||||||||||
| Asbestos |
$ | 221,567 | $ | 492,772 | $ | 714,339 | $ | 191,238 | $ | 470,113 | $ | 661,351 | ||||||||||||
| Environmental |
62,592 | 48,281 | 110,873 | 46,252 | 43,369 | 89,621 | ||||||||||||||||||
| All other |
1,567,454 | 720,360 | 2,287,814 | 1,065,160 | 530,444 | 1,595,604 | ||||||||||||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||||||||
| Total |
$ | 1,851,613 | $ | 1,261,413 | $ | 3,113,026 | $ | 1,302,650 | $ | 1,043,926 | $ | 2,346,576 | ||||||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||||||||
| Unallocated loss adjustment expenses |
178,249 | 132,560 | ||||||||||||||||||||||
|
|
|
|
|
|||||||||||||||||||||
| Total |
$ | 3,291,275 | $ | 2,479,136 | ||||||||||||||||||||
|
|
|
|
|
|||||||||||||||||||||
The following table provides a breakdown of loss and loss adjustment expense reserves (net of reinsurance balances recoverable) by type of exposure as of December 31, 2010 and 2009:
| 2010 | 2009 | |||||||||||||||
| Total | % of Total |
Total | % of Total |
|||||||||||||
| (in thousands of U.S. dollars) | ||||||||||||||||
| Asbestos |
$ | 640,063 | 23.2 | % | $ | 588,411 | 27.6 | % | ||||||||
| Environmental |
96,109 | 3.5 | 79,221 | 3.7 | ||||||||||||
| All other |
1,851,414 | 66.9 | 1,331,216 | 62.5 | ||||||||||||
| Unallocated loss adjustment expenses |
178,249 | 6.4 | 132,560 | 6.2 | ||||||||||||
| Total |
$ | 2,765,835 | 100 | % | $ | 2,131,408 | 100 | % | ||||||||
|
|
|
|
|
|
|
|
|
|||||||||
Our “All other” exposure category consists of a mix of general casualty (approximately 40% of “All other” net reserves), marine and aviation (approximately 11% of “All other” net reserves), workers compensation/personal accident (approximately 16% of “All other” net reserves) and other miscellaneous exposures, which are generally long-tailed in nature.
B-3
As of December 31, 2010, the IBNR reserves (net of reinsurance balances receivable) accounted for $1,119.9 million, or 40.5%, of our total net loss reserves. The reserve for IBNR (net of reinsurance balance receivable) accounted for $953.1 million, or 44.7%, of our total net loss reserves at December 31, 2009.
Annual Loss and Loss Adjustment Reviews
Because a significant amount of time can lapse between the assumption of risk, the occurrence of a loss event, the reporting of the event to an insurance or reinsurance company and the ultimate payment of the claim on the loss event, the liability for unpaid losses and loss adjustment expenses is based largely upon estimates. Our management must use considerable judgment in the process of developing these estimates. The liability for unpaid losses and loss adjustment expenses for property and casualty business includes amounts determined from loss reports on individual cases and amounts for IBNR reserves. Such reserves, including IBNR reserves, are estimated by management based upon loss reports received from ceding companies, supplemented by our own estimates of losses for which no ceding company loss reports have yet been received.
Loss advices or reports from ceding companies are generally provided via the placing broker and comprise treaty statements, individual claims files, electronic messages and large loss advices or cash calls. Large loss advices and cash calls are provided to us as soon as practicable after an individual loss or claim is made or settled by the insured. The remaining broker advices are issued monthly, quarterly or annually depending on the provisions of the individual policies or the ceding company’s practice. For certain direct insurance policies where the claims are managed by Third Party Administrators (TPA’s) and Managing General Agents (MGA’s), loss bordereaux are received either monthly or quarterly depending on the arrangement with the TPA and MGA.
We log all claims advices in our internal ‘Claims Tracking System’ upon receipt from brokers and cedants. Each advice is then assigned to the appropriate internal claims adjuster. Our professional claims adjusters and lawyers have many years of experience specializing in each class of business that we manage and also have established authority and internal referral levels. Individual large claims are reviewed and approved by senior management. Every item in the Claims Tracking System is monitored and tracked from the date of receipt of documents to review by adjusters and management and subsequent recording by our internal operations team. All loss reports are processed within three months of receipt with any items not processed during this period identified and flagged for review by senior management. The accuracy and completeness of the loss reports is assessed during the claims adjusting process. We also track where additional information is required for certain claims so that the exact status of all claims received can be monitored to ensure that additional requests and queries are tracked and acted upon. By carrying out additional onsite audits for larger exposures (by cedant for reinsurance or by policyholder for direct claims), we are able to test the accuracy of the figures in the actual underlying files and loss advices.
Where we provide reinsurance or retrocession reinsurance protection, the process of claim advice from the direct insurer to the reinsurers and/or retrocessionaires naturally involves more levels of communication, which inevitably creates delays or lags in the receipt of loss advice by the reinsurers/retrocessionaires relative to the date of first advice to the direct insurer. Certain types of exposure, typically latent health exposures such as asbestos-related claims, have inherently long reporting delays, in some cases many years, from the date a loss occurred to the manifestation and reporting of a claim and ultimately until the final settlement of the claim. For Asbestos and Environmental exposures, our actuaries apply explicit time lag assumptions in their reserving methodologies. This time lag varies by portfolio from one to five years depending on the relative mix of domicile, percentages of product mix of insurance, reinsurance and retrocessional reinsurance, primary insurance, excess reinsurance, reinsurance of direct and reinsurance of reinsurance within any given exposure category. Exposure portfolios written from a non-US domicile are assumed to have a greater time lag than portfolios written from a US-domicile. Portfolios with a larger proportion of reinsurance exposures are assumed to have a greater time-lag than portfolios with a larger proportion of insurance exposures.
An industry-wide weakness in cedant reporting affects the adequacy and accuracy of reserving for advised claims. We attempt to mitigate this inherent weakness as follows:
(i) We closely monitor cedant loss reporting and, for those cedants identified as providing inadequate, untimely or unusual reporting of losses, we conduct, in accordance with the provisions of the insurance and reinsurance contracts, detailed claims audits at the insured’s or reinsured’s premises. Such claims audits have the benefit of validating advised claims, determining whether the cedant’s loss reserving practices and reporting are adequate and identifying potential loss reserving issues of which our actuaries need to be made aware. Any required adjustments to advised claims reserves reported by cedants identified during the claims audits will be recorded as an adjustment to the advised case reserve.
B-4
(ii) Onsite claims audits are often supplemented by further reviews by our internal and external legal advisors to determine the reasonableness of advised case reserves and, if considered necessary, an adjustment to the reported case reserve will be recorded.
(iii) Our actuaries project expected paid and incurred loss development for each class of business, which is monitored on a quarterly basis. Should actual paid and incurred development differ significantly from the expected paid and incurred development, we will investigate the cause and, in conjunction with our actuaries, consider whether any adjustment to ultimate loss reserves is required.
Our actuaries consider the quality of ceding company data as part of their ongoing evaluation of the liability for ultimate losses and loss adjustment expenses, and the methodologies they select for estimating ultimate losses inherently compensate for potential weaknesses in this data, including weaknesses in loss reports provided by cedants.
We strive to apply the highest standards of discipline and professionalism to our claims adjusting, processing and settlement and disputes with cedants are rare. However, we are from time to time involved in various disputes and legal proceedings in the ordinary course of our claims adjusting process. The majority of the losses ceded to us are from the subscription insurance market (where there are often many insurers and reinsurers underwriting each policy), and we often are involved in disputes commenced by other co-insurers who act in unison with any litigation or dispute resolution controlled by the lead underwriter. Coverage disputes arise when the insured/reinsured and insurer/reinsurer cannot reach agreement as to the interpretation of the policy and/or application of the policy to a claim. Most insurance and reinsurance policies contain dispute resolution clauses requiring arbitration or mediation. In the absence of a contractual dispute resolution process, civil litigation would be commenced. We aim to reach a commercially acceptable resolution to any dispute, using arbitration or litigation as a last resort. We regularly monitor and provide internal reports on all disputes involving arbitration and litigation and engage external legal counsel to provide professional advice and assist with case management.
In establishing reserves, management includes amounts for IBNR reserves based on independent actuarial estimates of ultimate losses. Our independent actuaries employ generally accepted actuarial methodologies to estimate ultimate losses and loss adjustment expenses and those estimates are reviewed by our management.
Nearly all of our unpaid claims liabilities are considered to have a longtail claims payout. Gross loss reserves relate primarily to casualty exposures, including latent claims, of which approximately 25.1% relate to asbestos and environmental, or A&E, exposures.
Within the annual loss reserve studies produced by our external actuaries, exposures for each subsidiary are separated into homogeneous reserving categories for the purpose of estimating IBNR. Each reserving category contains either direct insurance or assumed reinsurance reserves and groups relatively similar types of risks and exposures (for example, asbestos, environmental, casualty, property) and lines of business written (for example, marine, aviation, non-marine). Based on the exposure characteristics and the nature of available data for each individual reserving category, a number of methodologies are applied. Recorded reserves for each category are selected from the indications produced by the various methodologies after consideration of exposure characteristics, data limitations and strengths and weaknesses of each method applied. This approach to estimating IBNR has been consistently adopted in the annual loss reserve studies for each period presented.
We review the external actuaries’ reports for consistency and appropriateness of methodology and assumptions, including assumptions of industry benchmarks, and discuss any concerns or changes with them. Our chief actuary and chief financial officer then consider the reasonableness of the reduction (or increase) in ultimate loss reserves that would result by amending loss reserves to the level recommended by our external actuaries, in light of actual loss development during the year using the following reports produced internally on a quarterly basis for each company:
1. Gross, ceded and net incurred loss report – This report provides, for each reporting period, the total (including commuted policies) gross, ceded and net incurred loss development for each company and a commentary on each company’s loss development prepared by our chief actuary. The report highlights the causes of any unusual or significant loss development activity (including commutations) and raises any concerns regarding the quality of the underlying reserve data.
2. Actual versus expected gross incurred loss development report – This report provides a summary, and commentary thereon, of each company’s (excluding companies or portfolios of business acquired in the current year) non-commuted incurred gross losses compared to the estimate of the development of non-commuted incurred gross losses provided by our external actuaries at the beginning of the year as part of the prior year’s reserving process.
B-5
3. Commutations summary schedule – This schedule summarizes all commutations completed during the year for all companies, and identifies the policyholder with which we commuted, the incurred losses settled by the commutation (comprising outstanding unpaid losses and case reserves) and the amount of the commutation settlement.
4. Analysis of paid, incurred and ultimate losses – This analysis for each company, and in the aggregate, provides a summary of the gross, ceded and net paid and incurred losses and the impact of applying our external actuaries’ recommended loss reserves. This report, reviewed in conjunction with the previous reports, provides an analytical tool to review each company’s incurred loss or gain and reduction in IBNR reserves to assess whether the ultimate reduction in loss reserves appears reasonable in light of known developments within each company.
The above reports provide our chief actuary and chief financial officer with the relevant information to determine whether loss development (including commutations) during the year has, for each company, been sufficiently favorable so as to warrant a reserve reduction of the level that would result by applying our external actuaries’ recommended reserve levels. It is not possible to quantify how much of any reserve release specifically relates to commutations or favorable development of non-commuted claims as the revised historical loss development used by the actuaries to estimate required reserves is a combination of both the elimination of historical loss development relating to commuted policies and non-commuted loss development. It is not practicable to determine the loss reserves that would be required relating to commuted policies as this would require an additional actuarial review each year for each company based on loss development statistics including the historical loss development for commuted policies.
Should the conclusions of the chief actuary and chief financial officer differ from those implied by our external actuaries, the chief actuary will engage in further discussions with the external actuaries to understand the rationale behind their reserve recommendations.
When establishing loss reserves we have an expectation that, in the absence of commutations and significant favorable or unfavorable non-commuted loss development compared to expectations, loss reserves will not exceed the high, or be less than the low, end of the following ranges of gross loss and loss adjustment expense reserves implied by the various methodologies used by each of our insurance subsidiaries as of December 31, 2010 were:
| Low | Selected | High | ||||||||||
| (in thousands of U.S. dollars) | ||||||||||||
| Asbestos |
$ | 612,272 | $ | 714,339 | $ | 784,486 | ||||||
| Environmental |
97,139 | 110,873 | 123,848 | |||||||||
| All other |
2,087,565 | 2,287,814 | 2,578,426 | |||||||||
| Unallocated loss adjustment expenses |
178,249 | 178,249 | 3,665,009 | |||||||||
|
|
|
|
|
|
|
|||||||
| Total |
$ | 2,975,225 | $ | 3,291,275 | $ | 3,113,026 | ||||||
|
|
|
|
|
|
|
|||||||
Latent Claims
Our loss reserves are related largely to casualty exposures including latent exposures relating primarily to A&E. In establishing the reserves for unpaid claims, management considers facts currently known and the current state of the law and coverage litigation. Liabilities are recognized for known claims (including the cost of related litigation) when sufficient information has been developed to indicate the involvement of a specific insurance policy, and management can reasonably estimate its liability. In addition, reserves are established to cover loss development related to both known and unasserted claims.
The estimation of unpaid claim liabilities is subject to a high degree of uncertainty for a number of reasons. First, unpaid claim liabilities for property and casualty exposures in general are impacted by changes in the legal environment, jury awards, medical cost trends and general inflation. Moreover, for latent exposures in particular, developed case law and adequate claim history do not exist. There is significant coverage litigation related to these exposures, which creates further uncertainty in the estimation of the liabilities. As a result, for these types of exposures, it is especially unclear whether past claim experience will be representative of future claim experience.
Ultimate values for such claims cannot be estimated using reserving techniques that extrapolate losses to an ultimate basis using loss development factors, and the uncertainties surrounding the estimation of unpaid claim liabilities are not likely to be resolved
B-6
in the near future. There can be no assurance that the reserves established by us will be adequate or will not be adversely affected by the development of other latent exposures.
Our asbestos claims are primarily products liability claims submitted by a variety of insureds who operated in different parts of the asbestos distribution chain. While most such claims arise from asbestos mining and primary asbestos manufacturers, we have also been receiving claims from tertiary defendants such as smaller manufacturers, and the industry has seen an emerging trend of non-products claims arising from premises exposures. Unlike products claims, primary policies generally do not contain aggregate policy limits for premises claims, which, accordingly, remain at the primary layer and, thus, rarely impact excess insurance policies. As the vast majority of our policies are excess policies, this trend has had only a marginal effect on our asbestos exposures thus far.
Asbestos reform efforts have been underway at both the federal and state level to address the cost and scope of asbestos claims to the American economy. While congressional efforts to create a federal trust fund that would replace the tort system for asbestos claims failed, several states, including Texas and Florida, have passed reforms based on “medical criteria” requiring certain levels of medically documented injury before a lawsuit can be filed, generally resulting in a drop of case filings in those states adopting this reform measure.
Asbestos claims primarily fall into two general categories: impaired and unimpaired bodily injury claims. Property damage claims represent only a small fraction of asbestos claims. Impaired claims primarily include individuals suffering from mesothelioma or a cancer such as lung cancer. Unimpaired claims include asbestosis and those whose lung regions contain pleural plaques.
Unlike traditional property and casualty insurers that either have large numbers of individual claims arising from personal lines such as auto, or small numbers of high value claims as in medical malpractice insurance lines, our primary exposures arise from A&E claims that do not follow a consistent pattern. For instance, we may encounter a small insured with one large environmental claim due to significant groundwater contamination, while a Fortune 500 company may submit numerous claims for relatively small values. Moreover, there is no set pattern for the life of an environmental or asbestos claim. Some of these claims may resolve within two years whereas others have remained unresolved for nearly two decades. Therefore, our open and closed claims data do not follow any identifiable or discernible pattern.
Furthermore, because of the reinsurance nature of the claims we manage, we focus on the activities at the reinsured level rather than at the individual claims level. The counterparties with whom we typically interact are generally insurers or large industrial concerns and not individual claimants. Claims do not follow any consistent pattern. They arise from many insureds or locations and in a broad range of circumstances. An insured may present one large claim or hundreds or thousands of small claims. Plaintiffs’ counsel frequently aggregate thousands of claims within one lawsuit. The deductibles to which claims are subject vary from policy to policy and year to year. Often claims data is only available to reinsurers, such as us, on an aggregated basis. Accordingly, we have not found claim count information or average reserve amounts to be reliable indicators of exposure for our reserve estimation process or for management of our liabilities. We have found data accumulation and claims management more effective and meaningful at the reinsured level rather than at the underlying claim level. As a result, we have designed our reserving methodologies to be independent of claim count information. As the level of exposures to a reinsured can vary substantially, we focus on the aggregate exposures and pursue commutations and policy buy-backs with the larger reinsureds.
We employ approximately 27 full time equivalent employees, including a U.S. attorney, actuaries, and experienced claims-handlers, to directly administer our A&E liabilities. We have established a provision for future expenses of $47.3 million, which reflects the total anticipated costs to administer these claims to expiration.
Our future environmental loss development may be influenced by other factors including:
| • | Existence of currently undiscovered polluted sites eligible for clean-up under the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) and related legislation. |
| • | Costs imposed due to joint and several liability if not all potentially reliable parties (PRPs) are capable of paying their share. |
| • | Success of legal challenges to certain policy terms such as the “absolute” pollution exclusion. |
B-7
| • | Potential future reforms and amendments to CERCLA, particularly as the resources of Superfund – the funding vehicle, established as part of CERCLA, to provide financing for cleanup of polluted sites where no PRP can be identified – become exhausted. |
The influence of each of these factors is not easily quantifiable and, as with asbestos-related exposures, our historical environmental loss development is of limited value in determining future environmental loss development using traditional actuarial reserving techniques.
There have been recent positive developments concerning lead paint liability, an area previously viewed as an emerging trend in latent claim activity with the potential to adversely affect reserves. After a series of successful defense efforts by defendant lead pigment manufacturers in lead paint litigation, in 2005, a Rhode Island trial court ruled in favor of the government in a nuisance claim against the defendant manufacturers. Since the Rhode Island decision, other government entities have employed the same theory for recovery against these manufacturers. In 2008, the Rhode Island Supreme Court reversed the sole legal liability loss experienced by lead pigment manufacturers in lead paint litigation. The court rejected public nuisance as a viable theory of liability for use by the government against the defendants and thus invalidated the entire claim against the lead pigment manufacturers. Subsequent to the Rhode Island Supreme Court decision at least one other government entity, an Ohio municipality, voluntarily dropped its lead paint suit. Thereafter, the State of Ohio, voluntarily dismissed its pending action against lead pigment manufacturers. Other state supreme courts equally rejected the public nuisance theory of liability, whereas no highest state court has ever adopted this theory as an acceptable cause of action.
We believe that lead paint claims now pose a lower risk to adverse reserve adjustment than previously thought, as the only trial court decision against lead pigment manufacturers to date was reversed on the basis that public nuisance is an improper liability theory by which a plaintiff may seek recovery against the lead pigment manufacturers. Even if adverse rulings under alternative theories succeed or if other states ultimately permit recovery under a public nuisance theory, it is questionable whether insureds have coverage under their policies under which they seek indemnity. Insureds have yet to meet policy terms and conditions to establish coverage for lead paint public nuisance claims, as opposed to traditional bodily injury and property damage claims. Still, there is the potential for significant impact to excess insurers should plaintiffs prevail in successive nuisance claims pending in other jurisdictions and coverage is established.
Our independent, external actuaries use industry benchmarking methodologies to estimate appropriate IBNR reserves for our A&E exposures. These methods are based on comparisons of our loss experience on A&E exposures relative to industry loss experience on A&E exposures. Estimates of IBNR are derived separately for each of our relevant subsidiaries and, for some subsidiaries, separately for distinct portfolios of exposure. The discussion that follows describes, in greater detail, the primary actuarial methodologies used by our independent actuaries to estimate IBNR for A&E exposures.
In addition to the specific considerations for each method described below, many general factors are considered in the application of the methods and the interpretation of results for each portfolio of exposures. These factors include the mix of product types (e.g., primary insurance versus reinsurance of primary versus reinsurance of reinsurance), the average attachment point of coverages (e.g., first-dollar primary versus umbrella over primary versus high-excess), payment and reporting lags related to the international domicile of our subsidiaries, payment and reporting pattern acceleration due to large “wholesale” settlements (e.g., policy buy-backs and commutations) pursued by us, lists of individual risks remaining and general trends within the legal and tort environments.
1. Paid Survival Ratio Method. In this method, our expected annual average payment amount is multiplied by an expected future number of payment years to get an indicated reserve. Our historical calendar year payments are examined to determine an expected future annual average payment amount. This amount is multiplied by an expected number of future payment years to estimate a reserve. Trends in calendar year payment activity are considered when selecting an expected future annual average payment amount. Accepted industry benchmarks are used in determining an expected number of future payment years. Each year, annual payments data is updated, trends in payments are re-evaluated and changes to benchmark future payment years are reviewed. This method has advantages of ease of application and simplicity of assumptions. A potential disadvantage of the method is that results could be misleading for portfolios of high excess exposures where significant payment activity has not yet begun.
2. Paid Market Share Method. In this method, our estimated market share is applied to the industry estimated unpaid losses. The ratio of our historical calendar year payments to industry historical calendar year payments is examined to estimate our market share. This ratio is then applied to the estimate of industry unpaid losses. Each year, calendar year payment data is updated (for both us and industry), estimates of industry unpaid losses are reviewed and the selection of our estimated market share is revisited.
B-8
This method has the advantage that trends in calendar year market share can be incorporated into the selection of company share of remaining market payments. A potential disadvantage of this method is that it is particularly sensitive to assumptions regarding the time-lag between industry payments and our payments.
3. Reserve-to-Paid Method. In this method, the ratio of estimated industry reserves to industry paid-to-date losses is multiplied by our paid-to-date losses to estimate our reserves. Specific considerations in the application of this method include the completeness of our paid-to-date loss information, the potential acceleration or deceleration in our payments (relative to the industry) due to our claims handling practices, and the impact of large individual settlements. Each year, paid-to-date loss information is updated (for both us and the industry) and updates to industry estimated reserves are reviewed. This method has the advantage of relying purely on paid loss data and so is not influenced by subjectivity of case reserve loss estimates. A potential disadvantage is that the application to our portfolios which do not have complete inception-to-date paid loss history could produce misleading results. To address this potential disadvantage, a variation of the method is also considered by multiplying the ratio of estimated industry reserves to industry losses paid during a recent period of time (e.g., 5 years) times our paid losses during that period.
4. IBNR: Case Ratio Method. In this method, the ratio of estimated industry IBNR reserves to industry case reserves is multiplied by our case reserves to estimate our IBNR reserves. Specific considerations in the application of this method include the presence of policies reserved at policy limits, changes in overall industry case reserve adequacy and recent loss reporting history for us. Each year, our case reserves are updated, industry reserves are updated and the applicability of the industry IBNR:Case Ratio is reviewed. This method has the advantage that it incorporates the most recent estimates of amounts needed to settle open cases included in current case reserves. A potential disadvantage is that results could be misleading where our case reserve adequacy differs significantly from overall industry case reserve adequacy.
5. Ultimate-to-Incurred Method. In this method, the ratio of estimated industry ultimate losses to industry incurred-to-date losses is applied to our incurred-to-date losses to estimate our IBNR reserves. Specific considerations in the application of this method include the completeness of our incurred-to-date loss information, the potential acceleration or deceleration in our incurred losses (relative to the industry) due to our claims handling practices and the impact of large individual settlements. Each year incurred-to-date loss information is updated (for both us and the industry) and updates to industry estimated ultimate losses are reviewed. This method has the advantage that it incorporates both paid and case reserve information in projecting ultimate losses. A potential disadvantage is that results could be misleading where cumulative paid loss data is incomplete or where our case reserve adequacy differs significantly from overall industry case reserve adequacy.
Under the Paid Survival Ratio Method, the Paid Market Share Method and the Reserve-to-Paid Method, we first determine the estimated total reserve and then deduct the reported outstanding case reserves to arrive at an estimated IBNR reserve. The IBNR:Case Ratio Method first determines an estimated IBNR reserve which is then added to the advised outstanding case reserves to arrive at an estimated total loss reserve. The Ultimate-to-Incurred Method first determines an estimate of the ultimate losses to be paid and then deducts paid-to-date losses to arrive at an estimated total loss reserve and then deducts outstanding case reserves to arrive at the estimated IBNR reserve.
Within the annual loss reserve studies produced by our external actuaries, exposures for each subsidiary are separated into homogeneous reserving categories for the purpose of estimating IBNR. Each reserving category contains either direct insurance or assumed reinsurance reserves and groups relatively similar types of risks and exposures (e.g., asbestos, environmental, casualty and property) and lines of business written (e.g., marine, aviation and non-marine). Based on the exposure characteristics and the nature of available data for each individual reserving category, a number of methodologies are applied. Recorded reserves for each category are selected from the indications produced by the various methodologies after consideration of exposure characteristics, data limitations, and strengths and weaknesses of each method applied. This approach to estimating IBNR has been consistently adopted in the annual loss reserve studies for each period presented.
As of December 31, 2010, we had 35 separate insurance and/or reinsurance subsidiaries whose reserves are categorized into approximately 276 reserve categories in total, including 40 distinct asbestos reserving categories and 27 distinct environmental reserving categories.
To the extent that data availability allows, the five methodologies described above are applied for each of the 40 asbestos reserving categories and each of the 27 environmental reserving categories. As is common in actuarial practice, no one methodology is exclusively or consistently relied upon when selecting a recorded reserve. Consistent reliance on a single methodology to select a recorded reserve would be inappropriate in light of the dynamic nature of both the A&E liabilities in general, and our actual exposure portfolios in particular.
B-9
In selecting a recorded reserve, management considers the range of results produced by the methods, and the strengths and weaknesses of the methods in relation to the data available and the specific characteristics of the portfolio under consideration. Trends in both our data and industry data are also considered in the reserve selection process. Recent trends or changes in the relevant tort and legal environments are also considered when assessing methodology results and selecting an appropriate recorded reserve amount for each portfolio.
The following key assumptions were used to estimate A&E reserves at December 31, 2010:
1. $65 Billion Ultimate Industry Asbestos Losses – This level of industry-wide losses and its comparison to industry-wide paid, incurred and outstanding case reserves is the base benchmarking assumption applied to Paid Market Share, Reserve-to-Paid, IBNR:Case Ratio and the Ultimate-to-Incurred asbestos reserving methodologies.
2. $35 Billion Ultimate Industry Environmental Losses – This level of industry-wide losses and its comparison to industry-wide paid, incurred and outstanding case reserves is the base benchmarking assumption applied to Paid Market Share, Reserve-to-Paid, IBNR:Case Ratio and the Ultimate-to-Incurred environmental reserving methodologies.
3. Loss Reporting Lag – Our subsidiaries assumed a mix of insurance and reinsurance exposures generally through the London market. As the available industry benchmark loss information, as supplied by our independent consulting actuaries, is compiled largely from U.S. direct insurance company experience, our loss reporting is expected to lag relative to available industry benchmark information. This time-lag used by each of our insurance subsidiaries varies from 1 to 5 years depending on the relative mix of domicile, percentages of product mix of insurance, reinsurance and retrocessional reinsurance, primary insurance, excess insurance, reinsurance of direct, and reinsurance of reinsurance within any given exposure category. Exposure portfolios written from a non-U.S. domicile are assumed to have a greater time-lag than portfolios written from a U.S. domicile. Portfolios with a larger proportion of reinsurance exposures are assumed to have a greater time-lag than portfolios with a larger proportion of insurance exposures.
The assumptions above as to Ultimate Industry Asbestos and Environmental losses have not changed from the immediately preceding period. For our company as a whole, the average selected lag for asbestos has increased slightly from 2.8 years to 2.9 years and the average selected lag for environmental has decreased slightly from 2.5 years to 2.4 years. The changes to the selected lags arose largely as a result of the acquisition of new portfolios of A&E exposures.
The following tables provide a summary of the impact of changes in industry ultimate losses, from the selected $65 billion for asbestos and $35 billion for environmental, and changes in the time-lag, from the selected averages of 2.9 years for asbestos and 2.4 years for environmental, for us behind industry development that it is assumed relates to our insurance and reinsurance companies. Please note that the table below demonstrates sensitivity to changes to key assumptions using methodologies selected for determining loss and allocated loss adjustment expenses, or ALAE, at December 31, 2010 and differs from the table on page 73, which demonstrates the range of outcomes produced by the various methodologies.
| Sensitivity to Industry Asbestos Ultimate Loss Assumption |
Asbestos Loss Reserves |
|||
| (in thousands of U.S. dollars) |
||||
| Asbestos — $70 billion |
$ | 830,228 | ||
| Asbestos — $65 billion (selected) |
714,339 | |||
| Asbestos — $60 billion |
598,450 | |||
| Sensitivity to Industry Environmental Ultimate Loss Assumption |
Environmental Loss Reserves |
|||
| (in thousands of U.S. dollars) |
||||
| Environmental — $40 billion |
$ | 170,227 | ||
B-10
| Environmental — $35 billion (selected) |
110,873 | |||
| Environmental — $30 billion |
51,519 |
| Sensitivity to Time-Lag Assumption* |
Asbestos Loss Reserves |
Environmental Loss Reserves |
||||||
| (in thousands of U.S. dollars) | ||||||||
| Selected average of 2.9 years asbestos, 2.4 years environmental |
$ | 714,339 | $ | 110,873 | ||||
| Increase all portfolio lags by six months |
787,964 | 114,922 | ||||||
| Decrease all portfolio lags by six months |
630,826 | 106,046 | ||||||
| * | Using $65 billion/$35 billion Asbestos/Environmental Industry Ultimate Loss assumptions. |
Industry publications have, since 2001, indicated that the range of ultimate industry losses is estimated to be between approximately $55 billion and $65 billion for asbestos losses. One commonly-referenced benchmark estimate has recently increased its estimate of ultimate industry asbestos losses from $65 billion to $75 billion. One of the reasons cited for the increase in estimated industry ultimate asbestos losses is a shift of losses away from products liability claims to non-products claims. In considering the impact of this issue, it is important to understand how asbestos claims attach to policies issued by the insurance industry in general and the policies issued by the companies owned by us in particular.
Historically, asbestos claims have been presented as “products liability” claims brought against manufacturers and distributors of asbestos-containing products. For a given manufacturer, distributor, or other entity involved in asbestos litigation, multiple claims are filed by numerous individuals. There is typically an allocation of the settlement costs for asbestos claims over time based on exposure to asbestos by the injured claimants. Many asbestos claims will aggregate within each individual policy period to exhaust the annual aggregate policy limits which exist within policies sold to cover products liability claims.
Beginning in the mid-1990’s, a trend began to emerge whereby certain policyholders began to assert that their asbestos claims should not fall within the “products liability” section of their policies and, therefore, should not be subject to the aggregate limits of products liability claims. Instead, the policyholder would assert that each individual bodily injury claim should be treated as a separate occurrence under the “premises/operations” section of their policies. Under such presentation, individual claim or occurrence limits apply separately to each claim and there is no aggregate limit for the amount of “premises” or “non-products” claims within a particular policy.
Our exposure to asbestos losses arises largely from direct excess policies and assumed reinsurance policies written through the London market. With respect to direct excess policies, our companies typically participated on policies whereby liability would only attach in excess of primary and umbrella policy limits. As non-products asbestos losses are not aggregated and are generally confined to the limits of the primary and other lower layer insurance policies, we believe we have very little exposure to non-products asbestos losses through direct insurance policies issued by our owned subsidiary companies. To date, we have seen no material reporting of non-products asbestos claims on direct insurance policies. The trend of asbestos losses shifting from products to non-products is not a new phenomenon. As our insurance entities have not received any material reporting of non-products claims to date and their direct insurance exposures are generally in excess of the layers of insurance impacted by non-products asbestos losses, we do not expect any material future liability in respect of non-products asbestos claims.
Losses with respect to assumed reinsurance exposures to non-products asbestos claims are unlikely to be aggregated and are generally confined to the limits of the primary and other lower layer insurance policies. There is limited ability for such claims to exceed retained levels. Our assumed reinsurance portfolio with respect to asbestos exposures is largely excess of loss in nature and, therefore, not especially subject to non-products asbestos liabilities. To date, we have seen no material reporting of non-products asbestos claims on assumed reinsurance policies.
As stated above, the trend of asbestos losses shifting from products to non-products is not a new phenomenon. As our assumed reinsurance entities have not received any material reporting of non-products claims to date and their assumed reinsurance exposures generally cover layers of insurance not impacted by non-products asbestos losses, management does not expect any material future liability in respect of non-products asbestos claims.
B-11
Other reasons cited for the increase in estimated industry ultimate asbestos losses include the ongoing uncertainty surrounding insurance coverage of asbestos claims and the ongoing reporting of significant numbers and values of malignant mesothelioma claims. As we do not view these issues as new information any impact has already been factored into our actuarial reserving methodologies with no need for any change in assumptions.
Furthermore, in recent years, the overall asbestos loss development trend within our portfolio has been favorable. Our asbestos exposures are reviewed by independent actuaries on an annual basis as part of the overall annual loss reserve review. Actual loss reporting for asbestos claims in recent years has been below actuarial estimated expectations.
Having considered the recent increase in one commonly-referenced benchmark estimate of ultimate net asbestos losses in the context of our portfolio of loss exposures and actual asbestos loss reporting in recent years for us in particular, as well as for the insurance industry generally, we believe there is no need to increase the $65 billion asbestos ultimate industry loss assumption.
Guidance from industry publications is more varied in respect of estimates of ultimate industry environmental losses. Consistent with an industry published estimate, we believe the reasonable range for ultimate industry environmental losses is between $30 billion and $40 billion. We have selected the midpoint of this range as the basis for our environmental loss reserving based on advice supplied by our independent consulting actuaries. Another industry publication has recently reduced its estimate of ultimate industry environmental losses from $56 billion to $42 billion. Based on our own loss experience, including successful settlement activity by us, the decline in new claims notified in recent years, improvements in environmental clean-up technology and the reduced industry estimate, we believe that $35 billion remains a reasonable basis for inclusion in our methodologies for reserving for environmental losses.
Our current estimate of the time lag that relates to our insurance and reinsurance subsidiaries compared to the industry is considered reasonable given the analysis performed by our internal and external actuaries to date.
Over time, additional information regarding such exposure characteristics may be developed for any given portfolio. This additional information could cause a shift in the lag assumed.
All Other (Non-latent) Reserves
For our “All Other” (non-latent) loss exposure, a range of traditional loss development extrapolation techniques is applied by our independent actuaries and us. These methods assume that cohorts, or groups, of losses from similar exposures will increase over time in a predictable manner. Historical paid, incurred, and outstanding loss development experience is examined for earlier years to make inferences about how later years’ losses will develop. The application and consideration of multiple methods is consistent with the Actuarial Standards of Practice.
When determining which loss development extrapolation methods to apply to each company and each class of exposure within each company, we and our independent actuaries consider the nature of the exposure for each specific subsidiary and reserving segment and the available loss development data, as well as the limitations of that data. In cases where company-specific loss development information is not available or reliable, we and our independent actuaries select methods that do not rely on historical data (such as incremental or run-off methods) and consider industry loss development information published by industry sources such as the Reinsurance Association of America. In determining which methods to apply, we and our independent actuaries also consider cause of loss coding information when available.
A brief summary of the methods that are considered most frequently in analyzing non-latent exposures is provided below. This summary discusses the strengths and weaknesses of each method, as well as the data requirements for each method, all of which are considered when selecting which methods to apply for each reserve segment.
1. Cumulative Reported and Paid Loss Development Methods. The Cumulative Reported (Case Incurred) Loss Development method relies on the assumption that, at any given state of maturity, ultimate losses can be predicted by multiplying cumulative reported losses (paid losses plus case reserves) by a cumulative development factor. The validity of the results of this method depends on the stability of claim reporting and settlement rates, as well as the consistency of case reserve levels. Case reserves do not have to be adequately stated for this method to be effective; they only need to have a fairly consistent level of adequacy at all stages of maturity. Historical “age-to-age” loss development factors (“LDFs”) are calculated to measure the relative development of an accident year from one maturity point to the next. Age-to-age LDFs are then selected based on these historical
B-12
factors. The selected age-to-age LDFs are used to project the ultimate losses. The Cumulative Paid Loss Development Method is mechanically identical to the Cumulative Reported Loss Development Method described above, but the paid method does not rely on case reserves or claim reporting patterns in making projections. The validity of the results from using a cumulative loss development approach can be affected by many conditions, such as internal claim department processing changes, a shift between single and multiple payments per claim, legal changes, or variations in a company’s mix of business from year to year. Typically, the most appropriate circumstances in which to apply a cumulative loss development method are those in which the exposure is mature, full loss development data is available, and the historical observed loss development is relatively stable.
2. Incremental Reported and Paid Loss Development Methods. Incremental incurred and paid analyses are performed in cases where cumulative data is not available. The concept of the incremental loss development methods is similar to the cumulative loss development methods described above, in that the pattern of historical paid or incurred losses is used to project the remaining future development. The difference between the cumulative and incremental methods is that the incremental methods rely on only incremental incurred or paid loss data from a given point in time forward, and do not require full loss history. These incremental loss development methods are therefore helpful when data limitations apply. While this versatility in the incremental methods is a strength, the methods are sensitive to fluctuations in loss development, so care must be taken in applying them.
3. IBNR-to-Case Outstanding Method. This method requires the estimation of consistent cumulative paid and reported (case) incurred loss development patterns and age-to-ultimate LDFs, either from data that is specific to the segment being analyzed or from applicable benchmark or industry data. These patterns imply a specific expected relationship between IBNR, including both development on known claims (bulk reserve) and losses on true late reported claims, and reported case incurred losses. The IBNR-to-Case Outstanding method can be used in a variety of situations. It is appropriate for loss development experience that is mature and possesses a very high ratio of paid losses to reported case incurred losses. The method also permits an evaluation of the difference in maturity between the business being reviewed and benchmark development patterns. Depending on the relationship of paid to incurred losses, an estimate of the relative maturity of the business being reviewed can be made and a subsequent estimate of ultimate losses driven by the implied IBNR to case outstanding ratio at the appropriate maturity can be made. This method is also useful where loss development data is incomplete and only the case outstanding amounts are determined to be reliable. This method is less reliable in situations where relative case reserve adequacy has been changing over time.
4. Bornhuetter-Ferguson Expected Loss Projection Reported and Paid Methods. The Bornhuetter-Ferguson Expected Loss Projection Method based on reported loss data relies on the assumption that remaining unreported losses are a function of the total expected losses rather than a function of currently reported losses. The expected losses used in this analysis are based on initial selected ultimate loss ratios by year. The expected losses are multiplied by the unreported percentage to produce expected unreported losses. The unreported percentage is calculated as one minus the reciprocal of the selected cumulative incurred LDFs. Finally, the expected unreported losses are added to the current reported losses to produce ultimate losses. The calculations underlying the Bornhuetter-Ferguson Expected Loss Projection Method based on paid loss data are similar to the Bornhuetter-Ferguson calculations based on reported losses, with the exception that paid losses and unpaid percentages replace reported losses and unreported percentages. The Bornhuetter-Ferguson method is most useful as an alternative to other models for immature years. For these immature years, the amounts reported or paid may be small and unstable and therefore not predictive of future development. Therefore, future development is assumed to follow an expected pattern that is supported by more stable historical data or by emerging trends. This method is also useful when changing reporting patterns or payment patterns distort historical development of losses. Similar to the loss development methods, the Bornhuetter-Ferguson method may be applied to loss and ALAE on a combined or separate basis. The Bornhuetter-Ferguson method may not be appropriate in circumstances where the liabilities being analyzed are very mature, as it is not sensitive to the remaining amount of case reserves outstanding, or the actual development to date.
5. Reserve Run-off Method. This method first projects the future values of case reserves for all underwriting years to future ages of development. This is done by selecting a run-off pattern of case reserves. The selected case run-off ratios are chosen based on the observed run-off ratios at each age of development. Once the ratios have been selected, they are used to project the future values of case reserves. A paid on reserve factor is selected in a similar way. The ratios of the observed amounts paid during each development period to the respective case reserves at the beginning of the periods are used to estimate how much will be paid on the case reserves during each development period. These paid on reserve factors are then applied to the case reserve amounts that were projected during the first phase of this method. A summation of the resulting paid amounts yields an estimate of the liability. The Reserve Run-off Method works well when the historical run-off patterns are reasonably stable and when case reserves ultimately show a decreasing trend. Another strength of this method is that it only requires case reserves at a given point in time and incremental paid and incurred losses after that point, meaning that it can be applied in cases where full loss history is not available. In cases of volatile data where there is a persistent increasing trend in case reserves, this method will fail to produce a reasonable estimate. In several cases, reliance upon this method was limited due to this weakness.
B-13
Our independent actuaries select the appropriate loss development extrapolation methods to apply to each company and each class of exposure, and then apply these methods to calculate an estimate of ultimate losses. Our management, which is responsible for the final estimate of ultimate losses, reviews the calculations of our independent actuaries, considers whether the appropriate method was applied, and adjusts the estimate of ultimate losses as it deems necessary. Historically, we have not deviated from the recommendations of our independent actuaries. Paid-to-date losses are then deducted from the estimate of ultimate losses to arrive at an estimated total loss reserve, and reported outstanding case reserves are then deducted from estimated total loss reserves to calculate the estimated IBNR reserve.
Net Reduction in Ultimate Loss and Loss Adjustment Expense Liabilities
The change in our estimated total loss reserves for both latent and all other exposures compared to that of the previous period, less net losses paid during the period, is recorded as a reduction in net ultimate losses on our statement of earnings for the period. Our estimated total loss reserve at December 31, 2011 was determined by estimating the ultimate losses and deducting paid-to-date losses. The estimated ultimate losses, for both latent and all other (non-latent) liabilities, were determined by the amount of advised case reserves and the application of the actuarial methodologies described above to estimate IBNR reserves. Future changes in our estimates of ultimate losses are likely to have a significant impact on future operating results. Our operating objective is to commute our loss exposures and manage non-commuted loss development in a disciplined manner such that future incurred loss development will be less than expected. A combination of future commutations and better-than-expected incurred loss development of non-commuted exposures could improve the trend of loss development and, after the application of actuarial methodologies to the improved trend, reduce the December 31, 2011 estimates of ultimate losses with a positive impact on our future results. However, it is not possible to project future commutation settlements or whether incurred loss development will be better than expected, and it is possible that ultimate loss reserves could increase based on the factors discussed herein.
Quarterly Reserve Reviews
In addition to an in-depth annual review, we also perform quarterly reserve reviews. This is done by examining quarterly paid and incurred loss development to determine whether it is consistent with reserves established during the preceding annual reserve review and with expected development. Loss development is reviewed separately for each major exposure type (e.g., asbestos, environmental, etc.), for each of our relevant subsidiaries, and for large “wholesale” commutation settlements versus “routine” paid and advised losses. This process is undertaken to determine whether loss development experience during a quarter warrants any change to held reserves.
Loss development is examined separately by exposure type because different exposures develop differently over time. For example, the expected reporting and payout of losses for a given amount of asbestos reserves can be expected to take place over a different time frame and in a different quarterly pattern from the same amount of environmental reserves.
In addition, loss development is examined separately for each of our relevant subsidiaries. Companies can differ in their exposure profile due to the mix of insurance versus reinsurance, the mix of primary versus excess insurance, the underwriting years of participation and other criteria. These differing profiles lead to different expectations for quarterly and annual loss development by company.
Our quarterly paid and incurred loss development is often driven by large, “wholesale” settlements – such as commutations and policy buy-backs – which settle many individual claims in a single transaction. This allows for monitoring of the potential profitability of large settlements which, in turn, can provide information about the adequacy of reserves on remaining exposures which have not yet been settled. For example, if it were found that large settlements were consistently leading to large negative, or favorable, incurred losses upon settlement, it might be an indication that reserves on remaining exposures are redundant. Conversely, if it were found that large settlements were consistently leading to large positive, or adverse, incurred losses upon settlement, it might be an indication – particularly if the size of the losses were increasing – that certain loss reserves on remaining exposures are deficient. Moreover, removing the loss development resulting from large settlements allows for a review of loss development related only to those contracts which remain exposed to losses. Were this not done, it is possible that savings on large wholesale settlements could mask significant underlying development on remaining exposures.
Once the data has been analyzed as described above, an in-depth review is performed on classes of exposure with significant loss development. Discussions are held with appropriate personnel, including individual company managers, claims handlers and
B-14
attorneys, to better understand the causes. If it were determined that development differs significantly from expectations, reserves would be adjusted.
Quarterly loss development is expected to be fairly erratic for the types of exposure insured and reinsured by us. Several quarters of low incurred loss development can be followed by spikes of relatively large incurred losses. This is characteristic of latent claims and other insurance losses which are reported and settled many years after the inception of the policy. Given the high degree of statistical uncertainty, and potential volatility, it would be unusual to adjust reserves on the basis of one, or even several, quarters of loss development activity. As a result, unless the incurred loss activity in any one quarter is of such significance that management is able to quantify the impact on the ultimate liability for loss and loss adjustment expenses, reductions or increases in loss and loss adjustment expense liabilities are carried out in the fourth quarter based on the annual reserve review described above.
As described above, our management regularly reviews and updates reserve estimates using the most current information available and employing various actuarial methods. Adjustments resulting from changes in our estimates are recorded in the period when such adjustments are determined. The ultimate liability for loss and loss adjustment expenses is likely to differ from the original estimate due to a number of factors, primarily consisting of the overall claims activity occurring during any period, including the completion of commutations of assumed liabilities and ceded reinsurance receivables, policy buy-backs and general incurred claims activity.
B-15
Footnotes to Consolidated Financial Statements
2. Significant Accounting Policies – Loss and Loss Adjustment Expenses. The liability for loss and loss adjustment expenses includes an amount determined from loss reports and individual cases and an amount, based on historical loss experience and industry statistics, for losses incurred but not reported. These estimates are continually reviewed and are necessarily subject to the impact of future changes in such factors as claim severity and frequency. While management believes that the amount is adequate, the ultimate liability may be significantly in excess of, or less than, the amounts provided. Adjustments will be reflected as part of net increase or reduction in loss and loss adjustment expense liabilities in the periods in which they become known. Premium and commission adjustments may be triggered by incurred losses and any amounts are reflected in net loss and loss adjustment expense liabilities at the same time the related incurred loss is recognized.
Commutations provide an opportunity for the Company to exit exposures to entire policies with insured and reinsureds for an agreed upon payment, or payments, often at a discount to the previously estimated ultimate liability. As a result of exiting all exposures to such policies, all advised case reserves and IBNR liabilities relating to the insured or reinsured are eliminated. A commutation is recognized upon the execution of a Commutation Release Agreement. On completion of a commutation, all the related balances, including insurance and reinsurance balances payable and/or receivable, funds held by ceding companies, and losses and loss adjustment expenses (including fair value adjustments and estimated IBNR), are written off with corresponding gain or loss recorded in the net reduction of ultimate losses. A commutation may result in a net gain irrespective of whether the settlement exceeds the advised case reserves. Advised case reserves are those reserve estimates for a specific loss or losses reported to the Company by either the broker or insured or reinsured. IBNR liabilities (or reserves) are established by the Company at a class of business or exposure level for claims that have not yet been reported to the Company but can reasonably be expected to have occurred, as well as for the future development of reported claims. A commutation settlement is a negotiated settlement of both the advised case reserves and an estimate of the IBNR reserves that relate to the policies being commuted. For latent exposures with a long reporting tail, the estimated level of IBNR reserves may be significantly higher than the advised case reserves. In such an instance, the commutation settlement of a block of such policies may be greater than the advised case reserves but less than the aggregate of the advised case reserves plus the estimated related IBNR reserves, resulting in a total saving to the remaining liability.
To the extent possible, all prior historical loss development that relates to commuted exposures is eliminated to produce revised historical loss development for the remaining non-commuted exposures. The Company’s estimates of IBNR reserves are not determined at the policyholder level but at the aggregate class of business or exposure level. Therefore, the Company does not typically identify a specific amount of IBNR reserves settled with each commutation. Rather, on an annual basis in the fourth quarter, the Company’s actuaries apply their actuarial methodologies to the remaining aggregate exposures and revised historical loss development information to reassess their estimates of gross and net ultimate liabilities and required gross and net IBNR reserves. Should a commutation that the Company considers significant occur in one of the first three quarters, then the Company, in conjunction with its independent actuaries, would estimate the amount of IBNR that would be associated with the policies being commuted. If the financial impact (including release of IBNR) of the commutation is considered significant, the Company would adjust its estimate of ultimate loss and loss adjustment expense liabilities in the quarter that the commutation was concluded. The agreed commutation settlement is recorded in net losses paid.
To the extent that commuted policies are protected by reinsurance, then the Company will, on completion of a commutation with an insured or reinsured, negotiate with the reinsurers to contribute their share of the commutation settlement. Any amounts received from such reinsurers will be recorded in net losses paid and the impact of any savings or loss on reinsurance recoverable on unpaid losses will be implicitly included in the actuarial reassessment of net ultimate liabilities and net IBNR reserves.
Commutations of acquired companies’ exposures have the effect of accelerating the payout of claims compared to the probability-weighted ranges of actuarially projected cash flows that the Company applies when estimating the fair values of assets and liabilities at the time of acquisition. Any material acceleration of payout together with the impact of any material loss reserve savings in any period will also accelerate the amortization of fair value adjustments in that period.
The Company’s insurance and reinsurance subsidiaries establish provisions for loss adjustment expenses relating to run-off costs for the estimated duration of the run-off. These provisions are assessed at each reporting date and provisions relating to future periods are adjusted to reflect any changes in estimates, including the impact of any acceleration of the run-off period that may be caused by commutations, of the periodic run-off costs or the duration of the run-off. Provisions relating to the current period together with any adjustment to future run-off provisions are included in loss and loss adjustment expenses in the consolidated statements of earnings.
B-16
3. Acquisitions
The Company’s primary objective in running off the operations of an acquired company is to effect an orderly and efficient settlement of all liabilities and assets and, in so doing, to strive to achieve savings in the settlement of such amounts in relation to the values implied by the purchase price of the transaction. The Company’s run-off process is led by disciplined management and includes the adjustment and settlement of valid claims, commutations of exposures, disciplined collection of reinsurance receivables, achievement of early finality of the acquired run-off by way of solvent scheme of arrangement (if available) and imposition of strong financial and operational governance over acquired companies.
The most significant liability and asset of an acquired company are typically the liability for loss and loss adjustment expenses and the asset related to any reinsurance recoverable on these liabilities that may be contractually due to the acquired entity. The market for acquisition of run-off companies is not sufficiently active and transparent to enable the Company to identify reliable, market exit values for acquired assets and liabilities. Accordingly, consistent with provisions of U.S. GAAP, the Company has developed internal models that it believes allow it to determine fair values that are reasonable proxies for market exit values. The Company is familiar with the major participants in the acquisition run-off market and believes that the key assumptions it makes in valuing acquired assets and liabilities are consistent with the kinds of assumptions made by such market participants. Furthermore, in the Company’s negotiation of purchase price with sellers, it is frequently clear to the Company that other bidders in the market are using models and assumptions similar in nature to the Company’s during the competitive bid process. The majority of acquisitions are completed following a public tender process whereby the seller invites market participants to provide bids for the target acquisition.
The Company accounts for acquisitions using the purchase method of accounting, which requires that the acquirer record the assets and liabilities acquired at their estimated fair value. The fair values of each of the reinsurance assets and liabilities acquired are derived from probability-weighted ranges of the associated projected cash flows, based on actuarially prepared information and management’s run-off strategy. The Company’s run-off strategy, as well as that of other run-off market participants, is expected to be different from the seller’s as generally sellers are not specialized in running off insurance and reinsurance liabilities whereas the Company and other market participants do specialize in such run-offs.
The key assumptions used by the Company and, it believes, by other run-off market participants in the fair valuation of acquired companies are (i) the projected payout, timing and amounts of claims liabilities; (ii) the related projected timing and amount of reinsurance collections; (iii) a risk-free discount rate, which is applied to determine the present value of the future cash flows; (iv) the estimated unallocated loss adjustment expenses to be incurred over the life of the run-off; (v) the impact that any accelerated run-off strategy may have on the adequacy of acquired bad debt provisions; and (vi) an appropriate risk margin.
The probability-weighted projected cash flows of the acquired company are based on projected claims payouts provided by the seller predominantly in the form of the seller’s most recent independent actuarial reserve report. In the absence of the seller’s actuarial reserve report, the Company’s independent actuaries will determine the estimated claims payout.
With respect to the Company’s U.K., Bermudian and Australian insurance and reinsurance subsidiaries, the Company is able to pursue strategies to achieve complete finality and conclude the run-off of a company by promoting solvent schemes of arrangement. Solvent schemes of arrangement are a popular means of achieving financial certainty and finality for insurance and reinsurance companies incorporated or managed in the U.K., Bermuda and Australia by making a one-time full and final settlement of an insurance and reinsurance company’s liabilities to policyholders. On acquisition of a U.K., Bermudian or Australian company, the claims payout projection is weighted according to management’s estimated probability of being able to complete a solvent scheme of arrangement. To the extent that solvent schemes of arrangement are not available to an acquired company, no weighting is applied to the projected claims payout.
On acquisition, the Company makes a provision for unallocated loss adjustment expense liabilities. This provision considers the adequacy of the provision maintained and recorded by the seller in light of the Company’s run-off strategy and estimated unallocated loss adjustment expenses to be incurred over the life of the acquired run-off as projected by the seller’s actuaries or, in their absence, the Company’s actuaries. To the extent that the Company’s estimate of the total unallocated loss adjustment expense provision is different from the seller’s, an adjustment will be made. While it is the objective of the Company to accelerate the run-off by completing commutations of assumed and ceded business (which would have the effect of shortening the life, and therefore the cost, of the run-off), the success of this strategy is far from certain. As a result, the estimates of unallocated loss adjustment expenses are based on running off the liabilities and assets over the actuarially projected life of the run-off, which the Company considers to be a prudent basis. In those domiciles where solvent schemes of arrangement are available, management’s estimates of the total unallocated loss adjustment expenses are probability-weighted in accordance with the estimated time that a solvent scheme of
B-17
arrangement could be completed, which has the effect of reducing the period of the run-off and the related unallocated loss adjustment expenses. For those acquisitions in domiciles where solvent schemes of arrangement are not available, the unallocated loss adjustment expenses are estimated over the projected life of the run-off.
The Company believes that providing for unallocated loss adjustment expenses based on the Company’s run-off strategy is appropriate in determining the fair value of the assets and liabilities acquired in an acquisition of a run-off company. The Company believes that other participants in the run-off acquisition marketplace factor into the price to pay for an acquisition the estimated cost of running off the acquired company based on how that participant expects to manage the assets and liabilities.
The difference between the carrying value of reserves acquired at the date of acquisition and the fair value is the Fair Value Adjustment, or FVA. The FVA is amortized over the estimated payout period and adjusted for accelerations on commutation settlements or any other new information or subsequent change in circumstances after the date of acquisition. To the extent the actual payout experience after the acquisition is materially faster or slower than anticipated at the time of the acquisition, there is an adjustment to the estimated ultimate loss reserves, or there are changes in bad debt provisions or in estimates of future run-off costs following accelerated payouts, then the amortization of the FVA is accelerated or decelerated, as the case may be, to reflect such changes.
B-18