Global Rating Methodologies Global Rating …

1 7 World Trade Center at 250 Greenwich Street New York, New York 10007 November 28, 2007 Re: NAIC Hearing on Use of Catastrophe Modeling by Rating Agencies Dear Ms. Simpson: We appreciate the opportunity to comment on the use of catastrophe modeling by Moody s Investors Service in analyzing property & casualty insurers and reinsurers. Moody s provides insurance financial strength ratings to property and casualty insurers and reinsurers and those ratings express our opinions of the ability of insurance companies to repay punctually senior policyholder claims and obligations. These ratings reflect our opinion of long-term relative risk and are forward-looking in nature because they apply to liabilities that may pay out over long periods of time.

2 Moody s analytical approach includes significant qualitative analysis in addition to quantitative analysis, and incorporates the opinions and judgments of experienced analysts. Global Rating Methodologies As outlined in our recently released Global Rating methodology for Property & Casualty Insurers (September 2006) and Moody s Global Rating methodology for Reinsurers (September 2007), Moody s reviews seven key Rating factors underlying (re)insurers business and financial profile, discusses why each factor is important to our ratings , what the relevant financial metrics are in analyzing these factors and how we interpret those metrics. Moody s has long considered catastrophe risk to be the most significant and volatile risk to capital over the short term.

3 Our analysis assesses a company s risk appetite and its ability to monitor and manage its risk exposures and also considers its reliance on reinsurance as a risk management tool. We evaluate catastrophe risk, both gross and net, relative to earnings and capitalization. We incorporate the views of the company s third-party vendor models, internal surveys, relative market share analysis, and stress case scenarios. Please refer to Exhibit 1 and Exhibit 2 for additional information on the Global Rating Methodologies . Current Trends in Insurance Catastrophe Risk Management A special comment titled Current Trends in Insurance Catastrophe Risk Management (January 2007) outlines our views on catastrophe risk management. The Appendix of that publication provides additional insight into how Moody s evaluates property catastrophe risk.

4 Please refer to Exhibit 3 for the special comment. The catastrophe risk profile of individual P&C companies can vary widely, depending on differences in risk-appetite, line of business focus, financial flexibility and capital adequacy, modeling sophistication/rigor, reliance on reinsurance (or other risk transfer mechanisms) and regulation. The principal components of Moody s analytical framework for assessing catastrophe risk for North American P&C companies include (1) Moody s Annual Catastrophe Risk Survey, (2) Moody s Risk z Page 2 November 28, 2007 Adjusted Capital Model (MRAC) for primary companies and (3) Moody s Probability of Default and Expected Loss Modeling for reinsurers and structured vehicles. Moody s Annual Catastrophe Risk Survey Moody s annual catastrophe risk survey includes qualitative and quantitative components.

5 The qualitative portion of the survey asks issuers to respond to a series of questions relating to (1) risk aggregation management, (2) model usage, (3) input data, (4) terrorism exposure, and (5) catastrophe reinsurance protection. Moody s is particularly interested in how catastrophe risk management fits into a firm s overall enterprise risk management, including risk committees (top down/bottom up) approaches. The quantitative portion of our survey asks rated issuers to provide aggregate loss Exceedance Probability (EP) profiles (ground, up, gross and net of reinsurance) at various return periods by requested peril. Moody s also requests that companies provide their entire aggregate EP distribution over all modeled perils. This facilitates a number of analyses that consider the entire tail of the EP curve, instead of simply focusing on losses at specific return periods.

6 However, Moody s is mindful that an aggregate-all-perils (AAP) EP curve may not completely capture a firm s catastrophe exposure (given certain perils and other factors are difficult or not conducive to modeling). Moody s Risk Adjusted Capital Model Moody s P&C Risk Adjusted Capital model (MRAC) is a tool used to quantify the multiple sources of risk including catastrophe risk within primary P&C insurance groups, and to gauge the adequacy of the group s resources for covering those risks. The basic data input source is the NAIC Annual Statutory Statement. The model employs Monte Carlo simulation to develop an estimated probability distribution for the aggregate performance of the group (gains or losses) over a one-year time horizon, allowing for comparison to the group s adjusted book capital at different probability thresholds.

7 For a given company, the model calculates exceedence curves for each of the seven catastrophes, by analyzing gross written premiums by line and by state, using a market share approach. Exceedance curves for the various hurricane zones have been adjusted to contemplate increased frequency and severity assumptions post-Katrina including greater loadings for demand surge and storm surge similar to the adjustments made to the leading vendor models. In addition, the three zonal earthquake models contemplate increased loadings for fire-following earthquake. Updating the model for revised curves did not result in any Rating changes. The particulars of the MRAC model including a summary of industry exceedance curve profiles by peril-zone are detailed in the MRAC Rating methodology (August 2006).

8 Please refer to Exhibit 4 for additional information. Probability of Default and Expected Loss Modeling for Monoline Reinsurers and Structured Vehicles Moody s uses annual aggregate all perils curves provided by respondents to our Catastrophe Risk Survey to stochastically model the impact of catastrophe risk on a company s earnings, and thereby compute estimates of the probability of default and expected loss to policyholders and/or bondholders over a one-year period. These results (incorporating stress testing based on individual analysts judgment) are then compared to Moody s idealized cumulative default rates and expected credit losses to generate implied insurance financial strength ratings and debt ratings . This type of model is focused on assessing the contribution of catastrophe risk to the income statement volatility, so this approach is most useful for portfolios with a preponderant exposure to catastrophes, such as monoline catastrophe reinsurers, and structured vehicles including catastrophe bonds and reinsurance sidecars.

9 Reinsurance Sidecars: Moody s Five Principles The evaluation of catastrophe risk is the primary Rating driver in the analysis of reinsurance sidecars. In a special comment titled Reinsurance Sidecars: Moody s Five Principles (March 2007), we discuss our approach to analyzing sidecars which utilizes output from a catastrophe model as a starting point. The original model provides estimated insured losses based on (1) estimated damages (vulnerability function) from (2) estimated hazards (Mother Nature) on (3) estimated exposures (company data). This layering of risk on risk suggests that model outputs are highly uncertain. Given risk on risk, our z Page 3 November 28, 2007 approach as outlined in the special comment focuses on Composition, Calibration, Conservatism, and Comparison.

10 We typically ask a series of questions to the issuer to evaluate their risk management and cat modeling practices, and subsequently make adjustments to the aggregate curves based on our judgment. Please refer to Exhibit 5 for additional information. Thank you for considering our comments. We would be pleased to discuss any questions you may have at your convenience. Sincerely, Ted Collins Group Managing Director Global Insurance Rating MethodologyMoody's Global Rating methodology forProperty and Casualty InsurersSummaryGlobally, Moody's rates just over 400 property and casualty insurance companies, with approximately $100 billion ofrated securities and over $500 billion of insurance reserve liabilities as of December 31, 2005. These ratings reflectMoody's opinions of these institutions' creditworthiness, which considers both business and financial fundamentals foreach rated company.