- In a recent English case, the court accepted that various types of loss did not constitute ‘consequential loss’ for the purposes of an exclusion clause in an IT contract for an automated billing system.
- It is unclear whether the same conclusion would be reached in Australia, where courts have adopted a different approach in relation to the interpretation of consequential loss.
- In light of such uncertainty, it is still important to clearly identify the types of losses intended to be recoverable and the types of losses intended to be covered by exclusion clauses for consequential loss.
IT system and service failures are becoming a serious business risk for companies. The most recently publicised example of such a failure occurred last month when hardware failures caused the Accenture-owned Navitaire’s New Skies reservation system used by Virgin Blue to crash. The resulting chaos meant that numerous Virgin Blue customers were either delayed or missed their flights.
Virgin Blue’s attempts to mitigate the brand damage included giving free pizza and accommodation to affected customers. No doubt both Virgin Blue and Navitaire have also embarked on a very careful review of their contract and in particular any consequential loss exclusion clauses, to determine how much of Virgin Blue’s losses are recoverable from Navitaire.
Given the potential damage to both brand and revenue, it is little wonder that both IT suppliers and their customers are starting to give more attention to consequential loss exclusion clauses. Ultimately, the proper allocation of risk arising from IT system and service failures is a commercial decision. A key question is the extent to which the drafting that parties agree upon in their commercial negotiations will in fact lead to liability outcomes that both sides would expect.
A good example of a situation where a consequential loss clause did not meet the expectations of at least the supplier is the recent English case of GB Gas Holdings Ltd v Accenture (UK) Ltd  EWCA Civ 912.1 Whether the result would have been the same under Australian law is an interesting question. On balance, we think the answer is probably not. What all of this highlights is that parties need to think very carefully about the losses that might flow from IT system or service failures. They also need to be very wary of treating exclusion clauses as mere ‘boilerplates’. Failure to properly identify and document any exclusions from liability may result in a party being disappointed with the outcome if a failure occurs and a claim is made.
The law on consequential loss: England v Australia
The English approach to consequential loss has traditionally drawn upon the two-limb test for the remoteness of damage set out in Hadley v Baxendale (1854) 156 ER 145. According to this test, losses are recoverable where they:
- arise naturally, in the usual course of things, from the breach (the first limb); or
- were contemplated by both parties at the time of making the contract (the second limb).
Exclusion clauses for consequential loss have been interpreted by English courts to only exclude liability for losses falling within the second limb.
This was the approach also adopted by courts in Australia until the decision in Environmental Systems Pty Ltd v Peerless Holdings Pty Ltd (2008) 19 VR 358;  VSCA 26 (Peerless).2 In Peerless, the Victorian Court of Appeal held that the relevant distinction was instead between:
- the ‘normal loss’ that every plaintiff in a like situation will suffer, and
- anything beyond the normal measure of loss, which it regarded as including losses such as lost profits or expenses incurred through breach.
According to this approach, consequential loss refers to the second type of loss, ie anything beyond the loss that every plaintiff in a like situation will suffer. This interpretation of consequential loss in the context of exclusion clauses necessarily excludes liability for some losses falling within the first limb of Hadley v Baxendale. In other words, the position in Australia is likely to result in more types of losses being considered consequential and therefore not recoverable where the parties have agreed to exclude liability for consequential loss.
GB Holdings v Accenture: an English example
This case related to an agreement between GB Gas Holdings Ltd (Centrica), a supplier of gas and electricity to residential customers in England and Wales, and Accenture plc (Accenture) for the design, supply, installation and maintenance of a new IT system by Accenture for Centrica, including an automated billing system based on pre-packaged SAP IS-U software.
Centrica began migrating its customer accounts onto the new billing system from December 2005, but problems with the system began to emerge around June 2006. This resulted in a large number of ‘exceptions’ being generated in the system during 2006 and 2007. These ‘exceptions’ required manual intervention before a bill could be issued, and the massive backlog of unresolved exceptions led to numerous customer accounts going unbilled.
In February 2007, Centrica notified Accenture of what it considered were a series of ‘fundamental defects’ under the agreement, but Accenture disputed that the defects were fundamental and accordingly refused to follow the agreed regime for dealing with fundamental defects. As a result, Centrica commenced proceedings against Accenture claiming damages for various losses flowing from Accenture’s breaches of the agreement, including warranty breaches relating to functionality, design and compliance with specifications.
Accenture contended that its liability for a number of losses claimed by Centrica was excluded by clause 16.2 of the agreement, which provided that neither party was liable for:
- loss of profits or of contracts arising directly or indirectly
- loss of business or of revenues arising directly or indirectly, or
- losses or damages to the extent that they are indirect or consequential or punitive.
Accenture failed both at first instance and on appeal.
Losses in dispute
Centrica’s claims appear to have been analysed on the basis that the agreement was akin to a building contract. The normal losses recoverable for defective work under a building contract would be the cost of rectification. These costs were not however discussed in this case, presumably because such losses were not in dispute.
The losses which were in dispute comprised five other categories of loss and the issue was whether clause 16.2 had the effect of excluding liability for these losses. The trial judge held that these losses were not consequential or indirect within clause 16.2, and therefore would be payable by Accenture, assuming they could be proved. The English Court of Appeal affirmed the trial judge’s decision, although it did not address each category of loss individually.
A summary of the courts’ findings is set out in the table below.
Click here to view.
It seems doubtful that the losses alleged to be suffered by Centrica (as described above) would be losses that every plaintiff in a like situation would suffer.
It would be open to an Australian court to hold that none of these losses fell within the concept of ‘normal’ loss. Some, if not all, of the losses would appear to meet the description of ‘expenses incurred through breach’, which was cited in Peerless as a specific example of a loss ‘going beyond the normal measure of loss’. On this basis, Centrica would not have recovered any of the five losses discussed above.
This case also illustrates the importance of the difference between a contract to supply IT services and a contract to design, supply and install a customised IT system. In relation to the former, it is likely that an Australian court would take the approach of applying the normal measure of loss for breach of a warranty of quality (as was at issue in Peerless). The normal measure of loss in such circumstances is the difference in value between a functioning and a defective system. But where the contract is to design, supply and install a customised system, breach of IT warranties relating to functionality, design and compliance with specifications is more akin to breach of warranties under a building contract. The normal measure of loss for such breaches would be the cost of remedying defects in the works.
Given the approach in Peerless, customers of IT goods and services in Australia would be well advised when negotiating exclusion clauses to clearly identify the types of losses intended to be recoverable and the types of losses intended to be excluded through the use of the term ‘consequential loss’. Failure to identify loss types which are excluded may otherwise result in a broader range of losses being found to be consequential than anticipated.