Poor relations (a rant about aggregation)

Go Back 13.2.19 INSIGHTS

The Supreme Court of New South Wales has become the latest judicial body forced to try not to appear baffled when confronted with the word “related” in an aggregation clause (Bank of Queensland v AIG Australia Ltd [2018] NSWSC 1689).

In 2017 the UK Supreme Court just about covered its tracks when placing the issue on the “too difficult pile” by asserting that “Determining whether transactions are related is therefore an acutely fact sensitive exercise”, thereby returning the conundrum to a first instance judge (AIG Europe v Woodman [2017] UKSC 18).

In previous posts I have expressed reservations about insurance practitioners fetishizing individual words. But this word needs singling out because the way it is used in aggregation clauses is uniquely bad.

The problem is this: using “related” in an aggregation clause only begs the question the clause is supposed to answer. Two cases of the highest authority show this to be true.

In Lloyds TSB General Insurance Holdings Ltd v Lloyds Bank Insurance Co ([2003] 4 All ER 43), Lord Hoffman cited with approval the following description of the purpose of an aggregation clause:

“…to enable two or more separate losses covered by the policy to be treated as a single loss for deductible or other purposes when they are linked by a unifying factor of some kind.” (emphasis added).

Whereas in Woodman we were told how one aggregates by reference to “related”:

“Use of the word ‘related’ implies that there must be some interconnection between the matters or transactions, or in other words they must in some way fit together” (emphasis added).

Spot the difference.

All that the word “related” tells you is that a link exists between two or more things. However, an aggregation clause needs to tell you what the required link is, so you can seek to apply the specified link to the facts in order to determine if aggregation is achieved. Confronted with the search for “a related series of acts or omissions” or “a series of related Wrongful Acts” you are left not with an answer but a question: “related by what?”.

“Related” as a word is entirely open ended. It applies to a tenuous connection just as much as to a close tie. One way or another just about anything can be “related” to just about anything else: you merely have to identify the type of relationship that supplies the link. The parties to the policies in the reported cases have not identified for the court the linking factor it should apply. They have left it entirely to the court. But it is not the court’s job to decide for the parties the nature and extent of the linking relationship they should have agreed.

Confronted with what is, in effect, an absolute discretion as to how, and so whether, to aggregate it is not surprising that the courts have decided to “go narrow” and aggregate to a limited degree, if at all, when confronted with “related”. It is easier to rationalise a limited approach than go out on a limb, by using an extended notion of “related”, and thereby risk being accused of re-writing the parties’ bargain.

If the parties to a policy want to use “related” then that is fine provided they specify the nature and extent of the relationship that will effect aggregation. If the drafting fails to answer that question then the result may be expensive litigation, with no real reference point for the question the court has to answer and, by reference to experience to date, a result which limits or defeats aggregation.

Stuart Hill

Stuart is a founding partner of Wynterhill LLP.

This post is intended to provide guidance of a practical nature but does not contain legal advice or advice as to what action you should or should not take specific to your insurance needs or those of your business, or with regard to any particular situation.