Insurers warned to not rely on, nor abandon risk models

In light of the financial meltdown in 2008 of companies that relied heavily on risk modelling, companies should be cautious about using risk models, but they should not abandon their use altogether, Guy Carpenter & Company LLC’s global chief economist and risk strategist says.
Joan Lamm-Tennant of Guy Carpenter was speaking at Chubb Canada’s first annual Executive Women’s Forum, held in Toronto on June 3.
She was asked to comment on the utility of risk models, the use of which failed to prevent a number of financial failures by seemingly solid U.S. financial institutions.
Lamm-Tennant immediately referred to a public presentation by economist George Box. “All models are wrong,” she recalled Box as saying. He then added, “but some are useful” as he showed a slide of supermodel and actress Naomi Campbell.
Joking aside, Lamm-Tennant said, risk models “are wrong, but what they do is they paint the corners of the table. Without them, you’re going into a board room, a management decision, a strategy session with nothing but a gut feeling.”
Illustrating the danger of relying too heavily on models, she referred to a Washington Post story, ‘The Beautiful Machine,’ which reported that American International Group (AIG) relied very heavily on risk modeling prior to its near financial collapse (and subsequent federal bailout of more than US$160 billion).
The Post story notes that a consultant’s model on AIG’s financial services side predicted a 99.85% percent chance of never having to pay out for the credit default risks AIG was assuming. In addition, Lamm-Tennant said, citing the article, AIG’s property and casualty models never picked up on the risks associated with the credit default swaps because the P&C models were focused narrowly on strictly P&C risks.
Notwithstanding these dilemmas, the models do serve to determine the broad parameters of a company’s potential risks, Lamm-Tennant said.
“I initially was fearful that people would throw out all models as a result of A.I.G.,” she said. “But I’m seeing something quite different.
“I’m seeing some heavy investment in modelling, but I’m seeing a much more cautious, suspicious use of it.”