“Homemade” limitation clause re-construed

February 19, 2018
Papers and Signature

Limitation of liability is invariably a key point of negotiation in any contract.  Each party – alive to the fact that things can and often do go wrong in commercial relationships – seeks to limit its potential exposure as far as possible and/or (where the more likely to suffer damage) to maximise the losses it can recover.

A recent Court of Appeal decision[1] illustrates that hours spent deliberating over how liability caps are to be calculated and when they will apply could prove worthless if what has been agreed is not plainly conveyed in the drafting that ultimately finds its way into the contract.

In this case, what Lord Justice Jackson referred to as a “homemade” limitation clause was interpreted on the basis of its natural meaning, which was also the meaning that yielded the “least bizarre consequences” – even though the “homemade” clause was a linguistic hodgepodge and there were alternative interpretations available that also appeared to make commercial sense.


ATOS was appointed by the Royal Devon & Exeter NHS Trust in November 2011 under a five-year contract to provide an IT system for patient records to be held online.  The plan was for the system to operate first in two small departments and then to be rolled out to the Trust’s other 45 departments.

Things did not go to plan, and the system roll-out never occurred.  The Trust gave notice to terminate the contract in April 2016.  The Trust then issued proceedings, claiming damages for numerous alleged breaches of contract.


The key issue was the extent to which ATOS’s liability for the alleged breaches was limited.  The relevant provision read as follows:

“The aggregate liability of the Contractor … shall not exceed:

9.2.1     for any claim arising in the first 12 months of the terms of the Contract, the Total Contract Price as set out in section 1.1; or

9.2.2     for claims arising after the first 12 months of the Contract, the total Contract Charges paid in the 12 months prior to the date of that claim.”

The Trust argued that this clause was incapable of interpretation and so unenforceable.  ATOS accepted that it was poorly drafted, but maintained that there were two possible interpretations: (a) it imposed a single cap which, depending on the circumstances, would either be the one set out in clause 9.2.1 or the one set out in clause 9.2.2; and (b) it imposed two separate caps, the first for defaults occurring in the first 12 months of the contract and the second for subsequent defaults.

High Court ruling

The judge at first instance held that the provision was not unenforceable (as the Trust attempted to argue), nor should it be read as providing for multiple caps (it being unlikely that the parties would have agreed to a cap that could be multiple times the total contract price of nearly £5 million).  She held that the first interpretation put forward by ATOS applied: there was a single cap, and its level would be determined by the timing of the first default.

The Trust appealed this decision, abandoning its original stance and instead arguing there were two caps (i.e. ATOS’s second interpretation).

Court of Appeal decision

The Court of Appeal held that, while the High Court judge’s reasoning in relation to multiple caps was sound, the language of the provision pointed “emphatically” at there being two caps, and that there was nothing surprising about this arrangement.  ATOS was doing high-value work in the first 12 months of the contract, during which time defaults could have expensive consequences.  If, however, a major default had occurred in the first year and used up the whole cap, there was no reason why ATOS should have a free ride in subsequent contract years – instead, the cap at clause 9.2.2 should apply.  Jackson LJ also pointed out that, if there was held to be only one cap (as per the judgment at first instance), it would have to be determined which of the two levels of cap should apply.  Although the judge at first instance had held that this would be determined by the date of the first default, there was nothing in clause 9.2 making the date of the first default a critical factor in choosing between the two caps.  Accordingly, the Trust’s appeal was allowed.


There is a clear drafting lesson here: avoid messy language, take a step back, and consider how plainly the wording used actually conveys the meaning intended.  In his judgment, Jackson LJ noted that it was not necessary to embark on “an Odyssey” through the “galaxy of high appellate guidance” on contractual interpretation if the natural meaning of the words used (in this case pointing emphatically to two separate caps) accorded with business common sense.  A badly drafted provision is not an invitation for the court to re-draft, particularly where its natural meaning (despite poor drafting) is otherwise clear.  If the natural meaning of a contractual provision is commercially sensible, the court is likely to apply it.

[1] The Royal Devon & Exeter NHS Foundation Trust v ATOS IT Services UK Ltd [2017] EWCA Civ 2196 (20 December 2017).

Helena FranklinHelena Franklin
Helena Franklin
Helena Franklin

News & Insights