Insurance News

P-C Insurers Passed A Real ERM Test Last Year, Analysts Say

Posted on: June 17, 2009

BROOKLYN , N.Y. The only real way to test the effectiveness of an organization’s enterprise risk management system is with real-life disasters, and on that score, property-casualty insurers passed with flying colors in 2008, according to analysts.

Speaking at the S&P 2009 Insurance Conference last week, investment analysts who research p-c insurance companies were highly upbeat when S&P Managing Director Mark Puccia asked how they evaluate ERM procedures for the companies they follow.

Analysts for life insurance companies said that sector had done a poor job of risk management.

Jay Cohen, a managing director in the equity research division of Bank of America-Merrill Lynch, said companies’ ERM processes are “only tested if something happens,” noting that in 2008, p-c insurers were tested by the combination of a disastrous credit market, a terrible equity market, and Hurricanes Ike and Gustav.

These were “big, big events, and yet no one went out of business. Some companies even grew book value,” he said.

While he conceded there may have been some luck involved, Mr. Cohen said he believes insurers were thinking about some of these potential issues as they set up their balance sheets and put their underwriting standards in place.

“I would say the p-c industry as a whole,” with some exceptions, “did pretty damn good last year. It’s got to say something about their enterprise risk management,” Mr. Cohen said.

David Havens, managing director of Hexagon Securities, gave the example of Bermuda-based reinsurer, Montpelier Re, to underscore Mr. Cohen’s point. Using the same words, Mr. Havens said that company “did pretty damn good” in 2008, considering it specializes in catastrophe reinsurance.

Montpelier Re had terrible problems after the 2005 hurricanes, Mr. Havens said, noting that while the second worst cat-year on record meant an unprofitable third-quarter for Montpelier last year, it managed to record a profit for the entire year.

In 2005, Montpelier saw 66 percent of its June 30 capital wiped out by nearly $1 billion in third-quarter storms, according to historical earnings records maintained by National Underwriter.

The company’s problems were compounded in 2005 by the fact that it had returned capital to shareholders with a special dividend of nearly $400 million before the storms hit.

Montpelier had said the dividend was predicated on a reduced risk profile, prompting rating analysts to question the company’s risk management capabilities after the 2005 storms.

In contrast to 2005, when Montpelier’s combined ratio came in over 200, in 2008, the company, which has since diversified beyond a property-cat reinsurance focus with forays into Lloyd’s and into the U.S. excess and surplus lines market, reported a combined ratio of 91 for the year.

“The proof is in the pudding,” Mr. Haven’s said, expanding his commentary to refer to ERM practices of the p-c insurance market generally.

Jimmy Bhullar, a senior analyst for J.P. Morgan, who covers the life insurance sector, said some risk management processes of the life and mortgage insurance industries failed in 2008, highlighting the management of risk correlations as one area of failure.

“If you are a mortgage insurance company, you should not own a bunch of RMBS [residential mortgage-backed securities] in your [investment] portfolio,” he said. Similarly, he said, life insurers should not have had the financial sector comprising the largest portion of their corporate bond holdings.

“Everyone underestimated the correlation risk in the financial services industry overall,” he said, noting that insurers stock values declined at the same time as the bond values declined for other financial services companies.

He also said the life insurance industry “did a poor job of managing capital overall” by buying back a lot of stock, with the analyst community urging and supporting these returns of capital to investors.

“As unusual as the environment was” in the equity and credit markets, he said “if this happens every 50, 60 or 70 years,” then you’ll see a recurrence of the situation that occurred at “a company like Hartford [which was] at one point distressed to the point that where they were just folding their hands and praying” that the market would go up.

“You don’t want to be running your company at the mercy at the markets, and many of these companies have done that,” Mr. Bhullar said.

Mr. Cohen said insurers on the p-c side understand that 1-in-50-year events do occur. “It’s what they do for a living.”

They “are always asking the question ‘What if’—‘What if we get the big storm? What if the credit markets implode,’” he said. “I think it protected them pretty well,” he said, contrasting other financial services firms where he believes “there was way too much reliance on…models.”

“The insurance industry has learned lessons over the years,” he concluded, noting that they knew from the aftermath of Hurricane Katrina, for example, how capital requirements were increased, models had to be changed, and companies had to become disciplined in how they interpret the results of analytical models.

Mr. Havens also warned against an over-reliance on models in place of human intuition within an ERM framework, underscoring his admonition with a story he read in the business press about a hedge fund.

The fund, he related, had three consecutive days of volatility in its stock portfolio that each exceeded the level of volatility its models predicted would only occur in 1 day out of 10,000 years.

“I think it’s important that you build models that people understand,” he added, identifying some of the people who need to understand as the ones that sit on boards of directors and manage companies.

© Copyright 2009 National Underwriter Property & Casualty. A Summit Business Media publication. All Rights Reserved.

%d bloggers like this: