Insurer Financial Results: 2002

Executive Summary
Analysis of insurer financial results for 2002 reveals an industry recovering from the terrorist attack on September 11, 2001, and the cumulative effects of anemic premium growth during the 1990s. Though the insurance industry returned to profitability in 2002, further analysis reveals that the industry's recovery remains far from complete. Insurers' overall rate of return remained low compared with historical norms and compared with the rates of return earned by firms in other industries. Insurers' surplus, or statutory net worth, continued to decline.

Net Income and Return on Net Worth
In 2002, the U.S. property/casualty insurance industry's net income after taxes rose to $2.9 billion — a $9.9-billion positive swing from the industry's first recorded net loss after taxes in 2001.[1]1. This study defines the U.S. property/casualty industry as private property/casualty insurance companies domiciled in the United States, including excess and surplus insurers and domestic insurers owned by foreign parents. The data in this report excludes foreign subsidiaries of U.S. insurance groups. All figures are represented net of reinsurance, unless otherwise noted. The industry's GAAP[2]2. GAAP stands for Generally Accepted Accounting Principles, the accounting basis used by most industries. Unless otherwise stated, the figures in this report are based on Statutory Accounting Principles (SAP), the accounting basis used by insurers when preparing the Annual Statements they submit to regulators. rate of return on average net worth (RONW) rose to 2.2% in 2002 from -1.2% the year before. (See Table 1.)

Image

Improvement in underwriting results drove the increases in the industry's net income after taxes and its GAAP rate of return. In 2002, net losses on underwriting fell 42.0% to $30.5 billion from a record $52.6 billion in 2001. But deterioration in investment results offset much of the improvement in underwriting results. Last year, net investment income dropped 2.8% to $36.7 billion from $37.7 billion the year before. And insurers posted $1.1 billion in realized capital losses on investments in 2002 — a $7.7-billion adverse swing from insurers' $6.6 billion in realized capital gains on investments in 2001. Other factors limiting the improvement in insurers' net income and rate of return last year included $0.9 billion in miscellaneous other losses and $1.2 billion in federal income taxes. The industry's $0.9 billion in miscellaneous other losses in 2002 contrasts with its $1.1 billion in miscellaneous other income in 2001, while the industry's $1.2 billion in federal income taxes in 2002 compares with its $0.2 billion in income tax recoveries the year before.

The industry's 2.2% GAAP rate of return in 2002 compares with an average of 6.8% for the ten years ending 2002, an average of 9.4% for the ten years ending 1992, and an average of 12.6% for the ten years ending 1982.

Underwriting Results
The industry's net loss on underwriting after dividends to policyholders declined by $22.1 billion in 2002. The industry's statutory combined ratio — a key measure of losses and expenses per dollar of premium — improved by 8.6 percentage points to 107.2% last year from 115.9% in 2001.[3]3. Throughout this report, figures may not balance due to rounding. So far this decade, the combined ratio has averaged 111.1% — 3.4 percentage points more than the 107.7% average combined ratio for the 1990s and 1.8 percentage points more than the 109.3% average combined ratio for the 1980s. During the 1960s and 1970s, the average combined ratio was less than 101.0%.

The improvement in underwriting results last year reflects the excess of growth in premiums over growth in losses and other costs. Growth in written premiums accelerated to 14.1% in 2002 from 8.0% in 2001 and a record-low 1.8% in 1998. The 14.1% increase in written premiums last year was the largest increase since 1986, when written premiums rose 22.2%. So far this decade, written premiums have risen an average of 9.1% per year — more than the 3.2% average annual increase during the 1990s and the 8.7% average annual increase in the 1980s but still short of the 12.1% average annual increase during the 1970s.

Data from ISO MarketWatchTM about rates on renewals indicates that broad-based firming in commercial insurance markets has contributed to the acceleration in premium growth, with commercial markets beginning to firm in mid-1999 and increases in rates on renewals gathering momentum ever since. In particular, rates on renewals for the ISO MarketWatch basket of commercial lines[4]4. SO MarketWatchTM tracks rate changes on renewals for commercial auto liability, commercial auto physical damage, products liability, premises/operations liability, businessowners, commercial fire, and commercial allied lines by class and geographic location. fell an average of 2.3% in 1998 but rose an average of 0.1% in 1999, 5.2% in 2000, 8.9% in 2001, and 12.0% in 2002.

These countrywide composite numbers mask significant differences by line, location, and class of business. For example, by state, ISO MarketWatch data shows that rate increases on renewals in 2002 for commercial auto liability ranged from 2.2% in Hawaii to 19.3% in Nebraska. For products liability, rate increases on renewals ranged from 5.0% in Tennessee to 16.9% in Arizona. By class, ISO MarketWatch data shows that rates on renewals for premises and operations liability coverage for firms manufacturing certain types of machinery fell 0.8%, while rates on renewals for some meat, fish, and poultry processors jumped 30.6%. For businessowners coverage, rates on renewals for dental laboratories rose 4.1%, as rates on renewals for stores selling floor coverings jumped 29.8%.

Data from ISO Market ProfilerTM shows that commercial lines premium growth during the past five years has also varied significantly by standard industrial classification (SIC) code.[5]5. For eleven commercial lines of insurance, ISO Market ProfilerTM provides market segmentation information, including insurance premium and loss statistics and business demographics for more than 900 industries at the county, state, and countrywide levels. On average during the past five years, premiums rose most rapidly for "new economy" industries. For example, from 1998 to 2002, commercial lines premiums for security and commodity brokers rose an estimated 10.2% per year, while premiums for firms providing business services increased an estimated 10.0% per year. Those increases contrast with declines in premiums for some "old economy" industries. For example, premiums for forestry businesses fell an estimated 5.0% per year from 1998 to 2002. Premiums for metal-mining firms dropped an estimated 4.7% per year, and premiums for firms producing leather and leather products declined an estimated 2.6% per year. Differences in experience by line, location, and class of business likely contributed to differences in the results of individual insurers.

Growth in earned premiums rose to 11.8% last year from 6.0% in 2001 and a record-low 1.8% in 1999 as a consequence of the acceleration in written premium growth since 1999. The 11.8% increase in earned premiums in 2002 was the largest increase since 1987, when earned premiums grew 13.6%.

At the same time as growth in premiums accelerated to levels not seen since the latter 1980s, growth in loss and loss adjustment expenses (LLAE) fell to its lowest level in five years. LLAE rose just 2.6% to $282.5 billion in 2002 from $275.4 billion the year before, when LLAE grew 15.3%. The last year in which LLAE grew less than 2.6% was 1997, when LLAE declined 4.2%.

The slowing in the growth of LLAE reflects a sharp decrease in the catastrophe losses included in insurers' financial results. Such catastrophe losses declined to $5.5 billion in 2002 from $17.1 billion in 2001, as overall catastrophe losses receded from levels inflated by the terrorist attack on September 11, 2001.[6]6. ISO's Property Claim Services® (PCSTM) unit defines catastrophes as events resulting in $25 million or more in insured losses and affecting significant numbers of insureds and insurers. The $5.5 billion in catastrophe losses included in insurers' financial results for 2002 is the net result of $5.9 billion in losses from catastrophes that occurred in 2002 and downward revisions to losses from catastrophes in prior periods. The $17.1 billion in catastrophe losses included in insurers' financial results for 2001 includes $9.3 billion in net losses after reinsurance from the terrorist attack on September 11, 2001, based on ISO's analysis of the information included in U.S. insurers' statutory financial statements for the year. Other catastrophes in 2001 caused $7.8 billion in losses. As of June 30, 2003, PCS estimated that property and business interruption losses from the terrorist attack on September 11, 2001, will ultimately amount to $20.3 billion for all insurers and reinsurers worldwide. For all lines combined, ISO estimates that losses from the attack may ultimately range from $30 billion to $50 billion. Using the midpoint of that range, $40 billion, and excluding losses covered by foreign insurers and reinsurers, U.S. insurers and reinsurers may ultimately bear $20 billion in losses from the attack.

Noncatastrophe loss and loss adjustment expenses rose 7.2% to $277.0 billion in 2002 from $258.3 billion the year before. Factors contributing to the growth in noncatastrophe losses last year include an increase in newly recognized environmental and asbestos (E&A) losses from policies written long ago and strengthening of reserves for other losses. ISO estimates that newly incurred E&A losses on past policies rose to $7.8 billion in 2002 from $3.5 billion in 2001. ISO also estimates that insurers strengthened reserves for other noncatastrophe losses by $4.7 billion in 2002, after weakening reserves for such losses by $4.5 billion in 2001.

In recent years, swings in catastrophe losses, E&A losses, and possible changes in the adequacy of reserves for other losses have all had a significant effect on insurers' reported financial results. The industry's combined ratio based on reported results improved to 107.2% last year from 115.9% in 2001. But, adjusted for catastrophes, E&A losses, and changes in reserve adequacy, the combined ratio improved to 103.9% in 2002 from 112.5% the year before. The industry's GAAP RONW for 2002 was 2.2% based on reported results but 4.5% based on results adjusted for catastrophes, E&A losses, and changes in reserve adequacy. That is, the net effect of unusually low catastrophe losses, E&A losses, and changes in reserve adequacy in 2002 was to add 3.3 percentage points to the industry's combined ratio and to cut the industry's GAAP RONW by 2.3 percentage points.

Investment Income and Capital Gains
Partially negating the improvement in underwriting results last year, the industry's investment results deteriorated as a consequence of declines in interest rates and stock prices. In particular, net investment income fell 2.8% last year to $36.7 billion, as the yield on insurers' average holdings of cash and invested assets dropped to 4.6% from 4.8% in 2001. Investment income declined in 2002 despite a 2.3% increase in insurers' average holdings of cash and invested assets to $803.8 billion from $785.5 billion the year before.

Based on records since 1959, the industry's investment income has declined in just six years, and all six declines have occurred in the eleven years since 1992. Reflecting those declines, the average annual rate of growth in the industry's investment income since 1990 is just 1.3% per year — far less than the 12.8% average annual rate of increase during the 1980s and the 18.7% average annual rate of increase during the 1970s.

Compounding the weakness in investment income, the industry realized $1.1 billion in capital losses in 2002. Combining the $1.1 billion in realized capital losses last year with $20.6 billion in unrealized capital losses, insurers' overall capital losses ballooned to $21.7 billion in 2002 — almost twice the $11.4 billion in overall capital losses in 2001 and more than nine times the $2.3 billion in capital losses in 2000.

The industry's realized capital loss last year was the first since 1974, and the industry's overall capital losses on investments in each of the past three years reflect declines in stock prices. In 2002, insurers' holdings of common stocks depreciated 17.0%, as the Standard & Poor's (S&P) 500 dropped 23.4%. Since 2000, insurers' holdings of common stocks have depreciated an average of 10.4% per year, while the S&P 500 has fallen an average of 15.7% per year. This recent experience contrasts with the industry's $133.2 billion in capital gains from 1995 to 1999. During that period, insurers' holdings of common stocks appreciated 17.6% per year, as the S&P 500 rose 26.2% per year.

Surplus and Leverage
Despite the nascent recovery in the industry's net income after taxes last year and a record amount of new funds (capital) paid in, the property/casualty insurance industry's surplus declined 1.5% to $285.2 billion at year-end 2002 from $289.6 billion at year-end 2001. Last year, deductions from surplus totaled $27.3 billion and included $20.6 billion in unrealized capital losses and $6.7 billion in dividends to shareholders. Partially offsetting those deductions, additions to surplus totaled $22.9 billion. The additions to surplus included $2.9 billion in net income after taxes, $17.3 billion in new funds paid in, and $2.7 billion in miscellaneous other additions to surplus. (See Table 2.)

Image

The 1.5% decline in surplus in 2002 followed declines of 8.7% in 2001 and 5.1% in 2000. Based on data since 1959, this is the first time surplus has declined for three consecutive years.

The declines in surplus in 2002 and 2001 would have been larger if not for the National Association of Insurance Commissioners' Codification of Statutory Accounting Principles. Effective January 1, 2001, those changes in statutory accounting require insurers to post deferred tax liabilities and deferred tax assets. For many insurers, unrealized capital losses on investments in 2002 and 2001 led to reductions in net deferred tax liabilities or increases in net deferred tax assets. ISO estimates that the major elements of the Codification of Statutory Accounting Principles raised surplus by $9.8 billion as of year-end 2002 and by $4.0 billion as of year-end 2001.

Leverage ratios, such as the premium-to-surplus ratio and the LLAE-reserves-to-surplus ratio, provide a simple measure of how much risk each dollar of surplus supports. The industry's premium-to-surplus ratio rose to 1.29 last year from 1.12 in 2001. The industry's LLAE-reserves-to-surplus ratio also rose in 2002, increasing to 1.38 from 1.28 in 2001.

The industry's premium-to-surplus ratio has risen for four consecutive years, increasing by half from a record-low 0.84 in 1998. But the long-term trend in premium-to-surplus ratios is downward. Even with recent increases, the 1.29 premium-to-surplus ratio for 2002 is down by more than half from a peak of 2.75 in 1974.

The LLAE reserves-to-surplus ratio also peaked in the mid-1970s but then remained relatively stable for almost two decades, falling from 2.13 in 1974 to 2.00 in 1992. Thereafter, the LLAE reserves-to-surplus ratio fell by nearly half to 1.08 in 1999 and then rose by nearly a third to 1.38 last year.

The declines in both leverage ratios over the years suggest that the industry's financial condition has improved. But interpretation of leverage ratios is not straightforward. Much of the decline in leverage ratios during the past decade may reflect declining prices in insurance markets during much of the 1990s and deteriorating reserve adequacy through 2001, rather than improvement in the industry's financial condition.

ISO estimates that the industry's reserves for losses other than E&A were deficient by between $59 billion and $68 billion as of year-end 2002. Using an estimate near the middle of that range and allowing for taxes, the premium-to-surplus ratio at year-end 2002 would have been 1.53 if the industry's loss reserves had been adequate. The LLAE-reserves-to-surplus ratio at year-end 2002 would have been 1.90 if reserves had been adequate. At 1.53, the premium-to-surplus ratio would have been at its highest level since the 1.58 in 1990. At 1.90, the LLAE-reserves-to-surplus ratio would have been at its highest level since the 2.00 in 1992.

Declines in the LLAE-reserves-to-surplus ratio may also reflect changes in insurers' mix of business. The percentage of premiums attributable to the long-tailed lines declined from 55.7% in 1992 to 48.3% in 2001, the latest year for which complete data was available.[7]7. For purposes of this analysis, the long-tailed lines of insurance comprise commercial auto liability, general liability (including products liability), medical malpractice, personal auto liability, workers' compensation, and reinsurance. Because claims for those lines of business take a relatively long period of time to settle, insurers hold proportionately more reserves for those lines.

Operating Cash Flow
Operating cash flow[8]8. Operating cash flow is the sum of net premiums collected, net investment income received, other underwriting income received, and other income received, minus net loss and loss adjustment expenses paid, other underwriting expenses paid, policyholder dividends paid, and taxes paid. (OCF) indicates the rate at which basic underwriting and investment operations generate cash to fund new investments, dividend payments to shareholders, or other activities. ISO estimates that significant improvement in underwriting cash flow in 2002 caused the industry's overall OCF to rise to $37.8 billion from $15.4 billion in 2001.[9]9. ISO estimated the industry's operating cash flow for 2002 based on the historical relationships between items on insurers' income and cash flow statements. OCF ratios measure operating cash flows relative to net written premiums, and the OCF ratio for the insurance industry more than doubled in 2002, rising to 10.3% from 4.8% in 2001. The industry's OCF ratio for 2002 was more than four times its 2.7% OCF ratio for 1999. But, even with recent increases, the OCF ratio remains less than half of the 24.2% it was in 1986.

The latest available data suggests that a significant proportion of the insurance industry continued to endure negative operating cash flows. When OCF is negative, an insurer may have to liquidate assets to meet financial obligations, potentially reducing investment earnings and capital gains. In 2001, the latest year for which complete data was available, insurers accounting for 34.7% of industry premium had negative operating cash flows. Though the market share of insurers with negative OCF did fall from its peak of 37.2% in 2000, it remained more than four times the 7.7% it was in 1996. And, at 34.7% in 2001, the market share of insurers with negative OCF was about one and a half times the 22.5% it was in 1984 — the bottom of the worst underwriting cycle in the industry's history.

Comparisons with Other Industries
The Fortune 500 consists of the 500 largest industrial and service corporations in the United States. ISO estimates that the median GAAP RONW for Fortune 500 companies in 2002 was 10.7% — two and a half times the 4.0% RONW earned by large insurers[10]10. This study classifies an insurer as "large" if the insurer accounts for more than 0.5% of the industry's net written premium in a given year. In 2002, thirty-eight large insurers each wrote more than $1.8 billion in premiums. and nearly five times the 2.2% RONW earned by the property/casualty industry overall.

Longer-term data indicates that insurers have a history of being less profitable than the Fortune 500. From 1983 to 2002, the median RONW for the Fortune 500 averaged an estimated 13.6% — 5.2 percentage points more than the 8.4% average RONW for large insurers and 5.5 percentage points greater than the 8.1% average RONW for the entire property/casualty industry.

Analyses in this report also compare insurers' profitability with that of firms in a broad array of other industries, using COMPUSTAT® data obtained from Standard & Poor's Institutional Market Services. During the ten years from 1992 to 2001 (the latest year for which complete COMPUSTAT data was available), the RONW for 294 noninsurance industries averaged 10.3% — 3.3 percentage points more than the insurance industry's 7.0% average RONW for the period. In a ranking from most to least profitable during that ten-year period, insurance ranked 204 out of 295 industries. The results were similar when ISO lengthened the analysis to include the ten years from 1982 to 1991.

Performance of Insurance Stocks
Insurers' capacity to provide financial protection to policyholders depends on insurers' ability to attract and retain the capital necessary to bear risk. For stock insurers, the ability to attract and retain capital depends, in turn, on their ability to offer attractive rates of return to investors.

In this report, ISO compares the performance of the S&P Property & Casualty Insurance Index with the performance of the S&P Life & Health Insurance Index, the S&P Financials Index, and the S&P 500 — a widely used measure of the stock market overall — to assess how investors in property/ casualty insurance stocks have fared.[11]11. As of year-end 2002, the S&P Property & Casualty Insurance Index was based on data for eleven insurers that collectively accounted for 19.2% of total industry net written premium in 2002 and 28.4% of total stock company net written premium in 2001 (the latest year for which separate data on stock company premiums was available). ISO also compares the recent performance of the Goldman Sachs (GS) index for the property/casualty industry overall and the GS indexes for various sectors of the property/casualty industry to assess how investors in different segments of the industry fared last year.

In 2002, investors in the S&P property/casualty index earned a total rate of return (capital gains plus dividends) of -11.0%. That is, $100 invested in the stock of the insurers in the S&P property/casualty index on December 31, 2001, would have dwindled to about $89 on December 31, 2002. Investors in the S&P life/health index earned a total rate of return of -16.2% in 2002.

The S&P financial index includes the stocks in the S&P insurance indexes, as well as the stocks of companies in a broad array of other financial service industries. The S&P 500 includes all the stocks in the S&P financial index, as well as stocks of companies in a wide spectrum of nonfinancial industries, such as manufacturing and retail trade. In 2002, investors in the S&P financial index experienced a total rate of return of -14.6%, while investors in the S&P 500 experienced a total rate of return of -22.1%.

On balance, even though investors in the S&P property/casualty index lost money last year, they fared better than investors in the S&P life/health index, S&P financial index, and the S&P 500. The relative strength in the performance of the S&P property/casualty index last year may reflect the property/casualty industry's return to profitability and investor expectations for results in 2003 and beyond.

Based on the performance of the S&P property/casualty index over the past twenty years, investors in property/casualty stocks have not done as well as investors in financial services more generally or the S&P 500 overall. From year-end 1982 to year-end 2002, investors in the S&P property/casualty index earned a compound average annual total rate of return of 12.1%, while investors in the S&P financial index earned returns averaging 14.0% and investors in the S&P 500 earned returns averaging 12.7%.[12]12. Investors in the S&P life/health index earned a compound average total rate of return of 11.5% per year from year-end 1982 to year-end 2002.

But many insurers are smaller than the companies in the S&P 500. And in 2002, as the S&P 500 fell 23.4%, the S&P MidCap 400 declined 15.5%, and the S&P SmallCap 600 dropped 15.3%. That is, investments in middle-capitalization and small-capitalization stocks outperformed investments in large-capitalization stocks last year.

The Goldman Sachs (GS) property/casualty index is based on data for thirty-seven insurers, including some in the S&P MidCap 400 and the S&P SmallCap 600. Investors' preference last year for smaller companies may explain in part why the GS property/casualty index fell just 8.2% in 2002, while the S&P property/casualty index, which includes eleven large insurers, fell 12.5%.

The Goldman Sachs property/casualty index is subdivided into a personal property/casualty index, a commercial property/casualty index, and a reinsurance index. In 2002, the GS indexes for all three segments of the property/ casualty industry declined. But the GS reinsurance index fell just 3.6%, as the GS personal property/casualty index fell 4.9% and the GS commercial property/casualty index fell 12.5%.

The declines in the GS property/casualty indexes last year contrast with gains for the last three years as a whole. From year-end 1999 to year-end 2002, the GS reinsurance index increased 28.2% per year, as the GS personal property/casualty index rose 11.3% per year and the GS commercial property/casualty index rose 8.3% per year. Those gains contrast with declines in the S&P 500 averaging 15.7% per year.

Mergers, Acquisitions, and Market Concentration
Based on data compiled by Conning & Company, the number of mergers and acquisitions in the property/casualty insurance industry declined to 35 in 2002 from 41 in 2001 and a record 117 in 1998. The reported dollar value of property/casualty mergers and acquisitions fell to $0.5 billion last year from $1.2 billion in 2001 and a record $55.8 billion in 1998.[13]13. The number and value of the property/casualty mergers and acquisitions exclude the 1998 merger of Travelers Group, Inc. with Citicorp Inc., a noninsurer, and Citigroup Inc.'s acquisition of Associates First Capital Corporation in 2000. The 35 mergers and acquisitions in 2002 compare with an average of 59 per year from 1988 to 2002. The $0.5-billion reported value of mergers and acquisitions last year contrasts with an average of $10.8 billion per year from 1988 to 2002.

Factors that may have contributed to the decline in merger and acquisition activity last year include:

  • volatility and weakness in financial markets;
  • concern about the outlook for the economy;
  • concern about the industry's ongoing exposure to terrorism; and
  • questions about the success of past transactions.

Despite the 888 property/casualty mergers and acquisitions from 1988 to 2002, quantitative measures based on long-term data show that the industry has been consolidating slowly over time. Based on standards established by the U.S. Department of Justice, countrywide data indicates that the industry remains far from concentrated.

Other Analyses
Additional analyses in this study report on trends in the number of property/casualty insolvencies, compare reinsurers' results with those of the insurance industry overall, and examine results by line of business.

1. This study defines the U.S. property/casualty industry as private property/casualty insurance companies domiciled in the United States, including excess and surplus insurers and domestic insurers owned by foreign parents. The data in this report excludes foreign subsidiaries of U.S. insurance groups. All figures are represented net of reinsurance, unless otherwise noted.

2. GAAP stands for Generally Accepted Accounting Principles, the accounting basis used by most industries. Unless otherwise stated, the figures in this report are based on Statutory Accounting Principles (SAP), the accounting basis used by insurers when preparing the Annual Statements they submit to regulators.

3. Throughout this report, figures may not balance due to rounding.

4. ISO MarketWatchTM tracks rate changes on renewals for commercial auto liability, commercial auto physical damage, products liability, premises/operations liability, businessowners, commercial fire, and commercial allied lines by class and geographic location.

5. For eleven commercial lines of insurance, ISO Market ProfilerTM provides market segmentation information, including insurance premium and loss statistics and business demographics for more than 900 industries at the county, state, and countrywide levels.

6. ISO's Property Claim Services® (PCSTM) unit defines catastrophes as events resulting in $25 million or more in insured losses and affecting significant numbers of insureds and insurers. The $5.5 billion in catastrophe losses included in insurers' financial results for 2002 is the net result of $5.9 billion in losses from catastrophes that occurred in 2002 and downward revisions to losses from catastrophes in prior periods. The $17.1 billion in catastrophe losses included in insurers' financial results for 2001 includes $9.3 billion in net losses after reinsurance from the terrorist attack on September 11, 2001, based on ISO's analysis of the information included in U.S. insurers' statutory financial statements for the year. Other catastrophes in 2001 caused $7.8 billion in losses. As of June 30, 2003, PCS estimated that property and business interruption losses from the terrorist attack on September 11, 2001, will ultimately amount to $20.3 billion for all insurers and reinsurers worldwide. For all lines combined, ISO estimates that losses from the attack may ultimately range from $30 billion to $50 billion. Using the midpoint of that range, $40 billion, and excluding losses covered by foreign insurers and reinsurers, U.S. insurers and reinsurers may ultimately bear $20 billion in losses from the attack.

7. For purposes of this analysis, the long-tailed lines of insurance comprise commercial auto liability, general liability (including products liability), medical malpractice, personal auto liability, workers' compensation, and reinsurance. Because claims for those lines of business take a relatively long period of time to settle, insurers hold proportionately more reserves for those lines.

8. Operating cash flow is the sum of net premiums collected, net investment income received, other underwriting income received, and other income received, minus net loss and loss adjustment expenses paid, other underwriting expenses paid, policyholder dividends paid, and taxes paid.

9. ISO estimated the industry's operating cash flow for 2002 based on the historical relationships between items on insurers' income and cash flow statements.

10. This study classifies an insurer as "large" if the insurer accounts for more than 0.5% of the industry's net written premium in a given year. In 2002, thirty-eight large insurers each wrote more than $1.8 billion in premiums.

11. As of year-end 2002, the S&P Property & Casualty Insurance Index was based on data for eleven insurers that collectively accounted for 19.2% of total industry net written premium in 2002 and 28.4% of total stock company net written premium in 2001 (the latest year for which separate data on stock company premiums was available).

12. Investors in the S&P life/health index earned a compound average total rate of return of 11.5% per year from year-end 1982 to year-end 2002.

13. The number and value of the property/casualty mergers and acquisitions exclude the 1998 merger of Travelers Group, Inc. with Citicorp Inc., a noninsurer, and Citigroup Inc.'s acquisition of Associates First Capital Corporation in 2000.