Daily Journal - December 23, 1997

By Deborah A. Pitts

 Don't Expect Year 2000 Losses to Be Covered Under CGL Insurance Policies

 By now, just about everyone has heard about the year 2000 computer problem. Some speculate that it will be a disaster of unprecedented proportion, changing life as we know it. Others forecast that the turn of the century will be but a blip on the radar screen. Wherever the truth lies on that broad spectrum, one thing is certain: There will be litigation - and lots of it - as a result of year 2000 failures.

Where litigation happens, insurance claims - and insurance litigation - are sure to follow. And whether there is insurance for year 2000 losses under traditional liability coverage is bound to be a hotly contested area over the next several years.

The year 2000 problem is the outgrowth of a fateful decision by programmers at the dawn of the computer age several decades ago. To save money on memory, programmers used only a two-digit, as opposed to a four-digit, date field to specify the year. Consequently, at the turn of the century, uncorrected computers will base their date calculations on the assumption that the new year is 1900 instead of 2000. The problem is exacerbated by the widespread use of noncompliant embedded chips in a vast array of equipment, ranging from microwaves to medical devices.

While the estimates of how much this computer crisis will cost are daunting, the projected litigation costs alone are mind-boggling. In March, the Giga Information Group, a Connecticut consulting firm that specializes in studying year 2000 problems, estimated that the litigation arising out of the year 2000 problem could exceed $1 trillion. Needless to say, the price of no other litigation problem in the history of the world has approached this staggering figure. And the meter has started running. On Dec. 2, the first year 2000 class action in the country, Atlaz Int'l Ltd v. Software Business Tech. Inc., 172539, was filed by Milberg, Weiss, Bershad, Hynes & Lerach in Marin County. Others are sure to follow.

In response to the growing concern, many companies are beginning to analyze whether their insurance programs will cover prospective year 2000 losses. This analysis may soon be mandatory. Last month, Sen. Robert Bennett of Utah introduced the Year 2000 Computer Remediation and Shareholder (CRASH) Protection Act of 1997. This legislation, if enacted, will require all publicly traded corporations to make specific disclosures about the current and expected costs of their year 2000 problems, the anticipated litigation costs and whether their insurance policies cover year 2000 losses.

Many companies will turn first to their comprehensive (or commercial) general liability policies seeking defense and indemnification, but they may discover that there are serious questions as to whether typical year 2000 losses and liabilities will be covered by standard CGL policies. Typically, the insuring agreement in a CGL policy covers damages the insured is legally obligated to pay from an occurrence resulting in bodily injury or property damage during the policy period. Aside from possible exclusions, the insuring agreement contains potential obstacles to coverage in the year 2000 context.

First, only sums that qualify "as damages" are potentially covered by a CGL policy. See AIU Ins. Co. v. Superior Ct., 51 Cal.3d 807 (1990). Insurers will argue that the vast sums many companies are paying to investigate correct noncompliant code and test programs are not being incurred as damages as a result of property damage or bodily injury, and do not fall within the coverage grant of CGL policies.

The occurrence element of a CGL insuring agreement, which requires that the damage be neither expected nor intended from the standpoint of the insured, also poses problems for policyholders. To be an occurrence, both the injury-causing event and the resulting damage must be unexpected. See Collin v. American Empire Ins. Co., 21 Cal.App.4th 787 (1994). Insurers will take the position that in the year 2000 field, neither the injury-causing event nor the resulting harm is fortuitous. The original decision to code the year in two-digit fields was intentional and clearly done as a cost-saving device.

Moreover, the results of noncompliant code failure are both expected and certain. Computer programs will be functioning exactly as they were intended to. The event that will produce the failures - the simple passage of time - is far from unexpected. The year-by-year calendar progression has been happening as long as we can remember, and we can expect it to continue as far as we can foresee into the future.

Coverage issues are also raised by the requirement in CGL insuring agreements that the loss result in bodily injury or property damage. The large losses that are anticipated are unlikely to involve either. Most liability is expected to result from corrupted financial data and cause purely economic loss.

One court has interpreted California law to require an insurer to defend a policyholder in a suit based on faulty computer equipment. See Centennial Ins. Co. v. Applied Health Care Sys. Inc., 710 F.2d 1288 (7th Cir. 1983). However, few courts, and no California courts, have squarely addressed the question of whether malfunctioning computer code is property damage within the meaning of insurance policies. Insurers are likely to contend that purely economic losses caused by computers coded with two-digit data fields are not "physical injury to tangible property" or "loss of use" of tangible property. See, e.g., Waller v. Truck Ins. Exchange, Inc., 11 Cal.4th 1 (1995).

Even if courts should find that faulty computer data constitutes property damage, policyholders may still not be entitled to CGL coverage unless the noncompliant code causes damage to other property. Liability insurance covers only losses arising from negligence or product liability, and not actions based on breach of contract or warranty. California courts steadfastly hold that a plaintiff cannot recover under negligence or product-liability theories unless the defect causes damage to property other than the defective property. See, e.g., Zamora v. Shell Oil Co., 55 Cal.App.4th 204 (1997). Accordingly, CGL coverage may not be available unless the noncompliant code damages other property, which may be rare.

One of the thorniest areas for policyholders attempting to obtain coverage for year 2000 losses under CGL policies is the trigger issue, i.e., the operative event that activates the insurer's defense and indemnity obligations. A CGL policy only covers damage that takes place during its policy period. See, e.g., Montrose Chemical Corp. of Calif. v. Admiral Ins. Co., 10 Cal.4th 645 (1995). Insurers will point out that under the law in California and many other states, the policy that is "on the risk" when the damage or injury occurs -- not when the "dangerous condition" is created -- is responsible for the loss. See, e.g., Hallmark Ins. Co. v. Superior Court, Cal.App.3d 1014 (1988). If the noncompliant programming is considered to be a dangerous condition, then the policy on the risk when the failures actually take place would be the likely policy to respond, assuming all other obstacles to coverage were removed. This would place the coverage, if any, in 2000 and after.

Policyholders are likely to argue that this classic trigger principle should be rejected, and that the policies on the risk when the computers were originally coded with two-digit date fields should be triggered. They will make these arguments because the language in these old policies is fixed, while the CGL policies in effect when the failures actually occur have yet to be written. It has been predicted that many other policies commencing in 2000 may contain restrictive language or exclude year 2000 computer liabilities altogether.

Policyholders will base their trigger argument on Armstrong v. World Indus. V. Aetna Casualty & Sur. Co., 45 Cal.App.4th 1 (1996). Armstrong dealt with the unique situation of asbestos bodily injury and property damage. The court relied on extensive evidence that buildings were damaged when asbestos was installed, when the fibers were released and when settled releases were "re-entrained" into the air. Based on this evidence, the court held that, among other insurance policies, the policy on the risk when the defective product was installed was triggered.

This ruling is unlikely to be expanded to year 2000 computer failures for several reasons. First, the Armstrong court expressly limited its holding to asbestos products, and no California court has applied the "installation as trigger" principle to any product other than asbestos. Second, the decision was based on evidence and allegations in the underlying actions that injury actually occurred when the asbestos products were installed.

It is difficult to imagine how it can be argued that any property damage took place when computers were originally coded with two-digit date fields in the 1970s and 1980s. Even if one considers faulty code to be property damage, the code was operating correctly. No damage will take place until noncompliant code or embedded chips incorrectly calculate dates. Unless policyholders can convince courts that property damage actually took place during the periods of past policies, which appears doubtful, it is unlikely that CGL policies in existence when the code was originally programmed will be triggered by year 2000 losses.

Posted with permission from the Daily Journal.

Hancock Rothert & Bunshoft has formed a Year 2000 Team to assist companies with related legal problems. If you would like more information on Hancock's Year 2000 Team, or on the firm in general, please contact: Vito C. Peraino on Tel: 213-623-7777 or E-mail: Click Contact Link or visit the Hancock Rothert & Bunshoft website at Click Contact Link

Visit the Year 2000 website at Click Contact Link

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.