Earthquake insurance: problems of successive events and allocation of loss (published on 27 October 2015)


In 2010 and 2011 the Canterbury region of New Zealand suffered several major earthquakes, resulting in severe destruction. The most devastating quake in February 2011 caused the tragic loss of 185 lives when Christchurch, the country's second-largest city, was torn apart. Unsurprisingly, the insurance industry has spent the following years dealing with unprecedented property claims. The claims were on a scale not before seen in New Zealand and, in many cases, raised novel or complex issues requiring litigation to resolve.

Over recent months, the Supreme Court and the Court of Appeal decided a number of important cases regarding earthquake insurance. These rulings will shape New Zealand insurance law for decades to come and more cases remain in the pipeline for resolution.

In September 2010 the first earthquake, of magnitude 7.1, struck about 40 kilometres west of Christchurch. It triggered a series of aftershocks over many months, including four that were major earthquakes in their own right. The February 2011 earthquake was of magnitude 6.3 at shallow depth directly below the city and occurred in the middle of the working day. One-third of the central business district's buildings were destroyed and many more later suffered additional damage or required demolition due to aftershocks, which occurred in relatively quick succession and continued (with diminishing force) into 2012. Often, building damage from an earlier earthquake was unrepaired and still being assessed by the time the next earthquake hit.

This update examines three insurance cases that grappled with the difficulties created by the successive separate physical events.

Insurer liability issues in multiple events

Insurance policies covering material damage of premises usually allow for automatic reinstatement of full cover after one event, unless contrary notice is given. However, if a later event causes fresh damage while the building is still awaiting repairs or destroys a damaged building (ie, a total loss), does automatic reinstatement mean that successive payments to the insured are due?

In attempts to limit their liability under such policy clauses, some insurers argued that if partial loss is suffered in an earlier earthquake, followed by total loss in a later earthquake, insureds should be entitled only to the total loss payment in respect of the cumulative damage suffered across all events. These arguments typically took two forms:

  • The doctrine of merger, common in marine insurance cases, should apply to general insurance cases, meaning that an insured's partial loss claim would merge into any later total loss claim; or
  • The indemnity principle should apply to prevent an insured from also recovering for earlier partial losses where recovery is made for a later total loss.

Doctrine of merger

In Ridgecrest NZ Limited v IAG New Zealand Limited the Supreme Court considered whether the doctrine of merger can apply to insurance policies outside the marine insurance context. Earlier High Court decisions had diverged on this issue. The Supreme Court noted several important differences between marine policies and general insurance policies – notably that under marine insurance policies, a cause of action for unrepaired damage generally arises only when the risk expires (ie, at end of the policy year).

In the Ridgecrest policy, maximum liability in respect of any single "happening" affecting the commercial building was NZ$1.984 billion. Ridgecrest's building was damaged by the September and December 2010 earthquakes. Repairs were underway, but incomplete when it was then damaged beyond repair in later 2011 earthquakes. The true replacement cost exceeded the amount of the liability cap, leading the insured to argue for losses to be assessed event by event, with a cause of action accruing immediately after each event. IAG's policy wording provided that the liability cap was reset after each event, which was another factor that the court considered important in its assessment of whether the doctrine of merger applied.

The Supreme Court concluded that the doctrine of merger could not apply and that the insured should be paid for damage for each earthquake event. The court was careful to frame its conclusion on this point with particular reference to the policy that it was interpreting. However, its broad comments about the differences between marine and general (ie, non-marine) insurance policies may make it difficult for an insurer to limit the scope of the Supreme Court's Ridgecrest reasoning and argue for the doctrine of merger in liability policies that rely on 'event' or 'happening' wordings.

Later, in QBE Insurance (International) Limited v Wild South Holdings Limited, the Court of Appeal noted that although the Supreme Court did not go so far as to confine the doctrine of merger strictly to marine insurance, it did hold that the doctrine is a correlative of contractual practice in marine insurance, where a cause of action arises only when the risk expires. Accordingly, if any scope remains to argue that the doctrine of merger can still apply outside the marine insurance context, it is likely to be only where the policy provides that the insured's right of action arises only when the risk expires – which may be an unlikely scenario in many general insurance contracts.

Indemnity principle

Despite the difficulties in arguing that an insured's successive property damage claims should be merged, Ridgecrest allows more scope for an insurer to argue that the principle of indemnity is a method of limiting its liability under insurance policies. The indemnity principle provides that where an insurer has an obligation to indemnify for loss, the insured will be fully indemnified and made whole – but never more than indemnified or in receipt of a windfall from the contract.

In Ridgecrest, the Supreme Court noted the availability of the indemnity principle in general, notwithstanding its rejection of the doctrine of merger, and that the principle precludes recovery of more than the replacement value of the insured property. However, in this particular case IAG's policy wording prevented the indemnity principle from applying as a limit on the insurer's liability. The policy contemplated full replacement cover and the liability cap under the policy was less than the replacement value of the building. Further, the policy provided that the liability cap was reset after each event. To apply the indemnity principle to prevent the insured from claiming for more than one event would "involve a substantial rewriting of the liability cap provision" in favour of the insurer.

The indemnity principle was further considered in QBE v Wild South by the Court of Appeal, which determined that if an insured has suffered damage from an earlier event that remains unrepaired at the time of a later event, the insurer is liable to indemnify only for the cumulative damage of both events. The cost of reinstating the property after the later earthquake represents the insured's actual loss, although it may also claim for any expenses actually incurred in relation to the earlier damage.

The Court of Appeal rejected arguments that such an application of the principle denies the insured its right to indemnification for the unrepaired damage from earlier events. In most cases the insured will never actually incur the expense of remedying damage from the first event. To allow separate recovery for unrepaired earlier damage would allow insureds to profit from the policy.

Wild South involved three appeals heard together, raising similar issues of interpretation. One of the insurers sought leave to appeal further, but was refused by the Supreme Court. The Supreme Court said that there was no obvious error by the Court of Appeal, which had applied orthodox interpretation to the specific wording of the automatic reinstatement and average clauses, and that the decision did not conflict with its earlier Ridgecrest principles.

Computer modelling for allocating losses

Subsequently, in Vero Insurance v Morrison the Court of Appeal evaluated ways in which insurers and insureds might try to quantify and apportion damage between multiple earthquakes – in particular, the evidential value of expert computer-modelling techniques.

The insureds owned an industrial-commercial building in Christchurch which was severely damaged by several earthquakes. Their experts said that it had suffered separate cumulative damage in each of the five major earthquakes. The insurance policy provided indemnity cover for the building on an event-b y-event basis, with automatic reinstatement for each "event or series of events arising from the same cause" in any 72-hour period. Vero said that it was liable only for damage arising from two of the earthquakes.

Actual witness evidence or building test data was sparse in this case. Vero's experts had not visited the site. In order to advance its claim, the insured's structural engineering expert employed modelling that measured the ground-shaking intensity of each earthquake to assist in identifying the events of loss and quantifying the extent of the damage caused by each event. The expert used it to identify the likely percentage of damage resulting from each event, along with a photographic essay of the building at different times.

The High Court had found the modelling reliable and substantially helpful as an input for the assessment of reasonable allocation of damage and repair costs per event. Vero challenged this on appeal. The Court of Appeal rejected Vero's argument that the evidence was inadmissible, finding it plainly relevant, meeting the reliability threshold and "of some probative value". It stated that, due to the deficiencies in documentation or testing of damage, this was an appropriate case in which to use modelling to supplement the evidence. The court referred with some approval to an approach taken in UK cases involving London Market Excess of Loss spiral reinsurance, which used actuarial models as a pragmatic solution to proving losses with less than "scientific exactitude" – although it did not formally decide whether such approach may extend to other types of insurance.

However, the Court of Appeal was more sceptical of the value of the modelling evidence than the High Court. It agreed with Vero that the model had flaws and limitations, such that less weight should have been accorded to the evidence. Direct evidence of the extent of the damage would usually be preferable. Together with the photographs, the insured could prove there was some additional physical damage attributable to later earthquakes; however, the limitations of the model meant that the extent of the damage (and the cost of repairing it) was relatively minor – perhaps even nominal.

Damaged or destroyed?

In both Wild South and Morrison, a separate issue arose as to when the building should be treated as destroyed for the purposes of the insurance policy – particularly where it may have still had some physical utility or functionality, but has become uneconomical to repair.

While this is closely a question of fact in every situation, the principles applicable to the meaning of 'destroyed' usually include an objective assessment by the court, informed by considerations which may include:

  • any special features of the building;
  • the insured's intentions for it (so long as they are not eccentric or unreasonable); and
  • the respective costs of reinstatement and replacement.

The test is not simply what the insured would do if it were spending its own money, and is not defined only by the economics of repair.

In Wild South it was important that the policy distinguished between 'destroyed' and 'damaged' buildings. In particular, different measures of indemnity were applied to each: the cost of rebuilding was the measure applied where the property was destroyed. That situation differed in two respects from the Morrison policy, which specifically defined 'destroyed' and contained a constructive total loss provision. In both cases the building was still functional to some extent and the insured wanted to preserve and use what it could. In Morrison conflicting evidence existed regarding the practicalities and economic viability of repair. More weight was seemingly given to the owner's subjective intentions, and against the insurer's choice of wording – which could have included 'economic destruction' language, but did not.


These cases demonstrate the significant influence that appeal courts are having in shaping the insurance sector's response to the Canterbury earthquakes. The decisions clarify the principles of liability of insurers and entitlements of insured property owners in both the residential and commercial sectors, and should assist towards settlement of other similar unresolved cases.

However, the decisions also show that outcomes depend on the specific policy wording used, and that allocation of specific damage across several events of loss is an extremely complicated evidential problem. While general principles have been established relating to the application of the indemnity principle and destroyed buildings, cases often turn on the particular policy wording. Even so, these rulings may have far-reaching consequences. In particular, the Supreme Court's rejection of the doctrine of merger makes it difficult to argue that the doctrine can apply to general insurance cases in New Zealand.

A version of this article appeared in the International Law Office newsletter on 27 October 2015