Tackling terrorism coverage: a government backstop will have limited value since most claims likely will fall within net retentions in the Terrorism Risk Insurance Act. (Underwriting Insight: Property/Casualty).

Now that the insurance industry has been required to deal with the issue of providing terrorism insurance, the question is how can insurers tackle a peril that many underwriters consider entirely uninsurable by private industry.

Premium rates are essentially a function of loss frequency and severity. Of course, historic records that can document the frequency of terrorist attacks in the past do exist. However, no one believes that in this case past performance will be a guarantee for future results.

While large attacks on the scale of the World Trade Center catastrophe are certainly possible, they are not probable. Government vigilance will make attacks that require extensive logistics, involving a large number of terrorists, increasingly difficult to perform. Instead, terrorist groups are likely to resort to smaller, targeted assaults that are easier to carry out. If we want to determine the most likely modes of attack, we have only to look at the militant tactics employed in Israel, the West Bank and most recently, in Indonesia and Kenya. Car bombs or suicide bombings would have a devastating effect on a population's psyche and would undermine an already shaky economy.

The insured loss caused by these targeted attacks will likely be on the scale of Irish Republican Army bombings of the early 1990s. Those insured losses varied between $259 million that resulted from the IRA bomb attack in South Key Docklands, London, in 1996, to $907 million following the IRA bombing near NatWest Tower, London, in 1993. This, of course, means that the government backstop will have very limited value to insurers, as most claims will fall well within the net retentions defined under the Terrorism Risk Insurance Act. Proponents of the law cite that insurers would have to pay "only" between $10 billion to $15 billion a year before the law kicks in and provides reinsurance for 90% of the losses in excess of that.

It does not take a rocket scientist to understand that multiple $15 billion hits in consecutive years would have a devastating impact on the property/casualty industry. Single years of large catastrophic losses, such as Hurricane Andrew in 1992, at $18.6 billion, or the Northridge Earthquake in 1994, at $13.7 billion, led to reduced coverage availability and higher premiums as investors withdrew from the market. Losses on a larger magnitude would certainly cause insolvencies, especially with carriers that are already experiencing a weakened capital base.

So how do insurers price a product where frequency is essentially an unknown and even severity moves in a fairly large band? Here are a few ideas on tackling the problem:

* To make terrorism insurance affordable to most commercial buyers make the peril an integral mandatory part of every policy. A high take-up rate of the coverage is essential to allow the law of large numbers to function. After all, a company s retention under the act does not change if only few insureds buy the coverage extension.

* Determine your company's net retention according to the new law and establish premium rates that allow your company to fund the retention over a very short time--two to three years at most. In filing rates, allow for a large experience modifier paving the way for immediate prospective premium adjustments in case of large losses.

* Use excess and surplus lines paper whenever possible to limit the ability of state regulators to influence pricing for political reasons.

* Decline the underwriting of target risks--essentially any place where large masses of people congregate. Leaving it up to customers to choose which locations they want to purchase coverage for will clearly lead to adverse selection and an insufficient premium pool to fund company retentions.

* Exclude nuclear, biological and chemical terrorist attacks. The new law requires terrorism coverage that does "not differ materially from the terms, amounts and other coverage limitations" applicable to losses arising from events other than acts of terrorism. Most property policies carry nuclear and pollution/contamination exclusions that would have preempted recovery for terrorist attacks of this type in the past. There is no reason to give up this protection, especially considering that these attacks defy any reasonable accumulation control.

* Invest in multiline aggregation software. Some of it will be commercially available in the spring of 2003. Since the terrorism risk law will automatically expire in 2005, there is no guarantee that catastrophic excess reinsurance will be available in key metropolitan centers.

Klaus Gebhardt, a Best's Review columnist, is senior vice president of Ace Bermuda Insurance Ltd. He can be reached at [e-mail].