Avoiding ‘Bizarre’ Consequences from Liability Caps

March 25, 2018

In Royal Devon and
Exeter NHS Foundation Trust v Atos IT Services UK Ltd
[2017] EWCA Civ 2196,
the Court of Appeal considered the interpretation of a liability cap in a
contract for the provision of IT systems and related services. It favoured the interpretation
with ‘the least bizarre consequences’.

As is not uncommon in IT services and outsourcing
contracts, the clause in question purported to apply different caps to the IT
provider’s liability depending on the stage of the contract at which defaults
occurred. The parties disagreed as to how the clause should be interpreted,
contributing to a lengthy dispute.

Contracting parties in the IT sphere and elsewhere should
be wary of finding themselves in similar situations when the drafting of
limitation clauses in their contracts is unclear.

Facts of the case

Royal Devon and Exeter Foundation Trust engaged Atos to
provide an IT system which would hold patient records online, so that
clinicians could view and amend patients’ records electronically. Atos agreed
to provide health record scanning services and electronic document management (EDM)
software and associated services.

The contract was based on the standard form NHS Conditions
of Contract for the Supply of IT Systems and the Provision of Associated Services,
with certain amendments. It had a five-year term and a contract price of

The project did not go well and there was a dispute as to
who was responsible. The Trust terminated the contract shortly before the end
of its term and issued proceedings in the Technology and Construction Court (TCC),
claiming damages for alleged breaches of contract by Atos. Atos counterclaimed,
asserting that the Trust was responsible for the problems.

The liability cap

The liability cap was one of the provisions negotiated by
the parties, rather than being part of the standard form contract. It provided
for the level of the cap to decline as the contract progressed, presumably to
reflect the parties’ perceptions that the project risks would reduce over time.
It was drafted in the following terms:

Clause 8.1.2(b):

‘(b) the aggregate liability of
either party under the Contract for all defaults, other than those [resulting
in direct loss of or damage to tangible property of the other party], shall not
exceed the amount stated in schedule G to be the limit of such liability’.

Paragraph 9.2 of Schedule G:

‘9.2       The aggregate liability of [Atos] in
accordance with sub-clause 8.1.2 paragraph (b) shall not exceed:

9.2.1     for any claim arising in the first 12
months of the term 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.’

Notably, paragraph 9.2 ambiguously combined the reference
to ‘aggregate liability’ with the phrases ‘for any claim’ and ‘for claims’.

The principles of

The legal principles of contractual interpretation are
well-established and will apply equally to a provision seeking to limit
liability. In summary:

  • The court will seek to ascertain the parties’
    intentions by reference to ‘what a reasonable person having all the background
    knowledge which would have been available to the parties would have understood
    them to mean’ (Lord Hoffmann in Chartbrook
    Ltd v Persimmon Homes Ltd
    [2009] UKHL 38, para 14).
  • The court will focus on the meaning of the words
    used in their documentary, factual and commercial context. In assessing the
    meaning of the language used, Lord Neuberger explained in Arnold v Britton [2015] UKSC 36 (at [15]) that the court will
    consider: (i) the natural and ordinary meaning of the clause, (ii) any other
    relevant provisions of the contract, (iii) the overall purpose of the clause
    and the contract, (iv) the facts and circumstances known or assumed by the
    parties at the time that the document was executed, (v) commercial common
    sense, but, (vi) it will disregard subjective evidence of any party’s
  • If there are two possible constructions, the
    court is entitled to prefer the construction which is consistent with business
    common sense (Lord Clarke in Rainy Sky SA v Kookmin Bank [2011] UKSC
    50, paras 21 to 30).
  • For limitation of liability provisions, there is
    no presumption against the parties having agreed to give up or limit their
    remedies for breach of contract. Provided that the words used are clear, the
    court will give effect to the commercial allocation of risk in the contract (Moore-Bick
    LJ in Tradigrain SA v Intertek Testing Services (ITS) Canada Ltd [2007] EWCA Civ
    154 at [46]; Briggs LJ in
    Nobahar-Cookson v The Hut Group Limited [2016]
    EWCA Civ 128 at [19]).
  • Although it has been held that the words used
    must be clear and unambiguous in order to limit the liability of a party for
    his own wrongdoing (see eg Dairy Containers ltd v Tasman Orient Line [2004]
    UKPC 22 at [12]), the courts are reluctant to find that a provision is void for
  • The court will strive, where possible, to give
    effect to all terms agreed by the parties (Moore-Bick LJ in Whitecap Leisure
    v John H Rundle Ltd
    [2008] EWCA Civ 429 at [21]-[22]).

The decision at
first instance

In the TCC decision (Royal
Devon and Exeter NHS Foundation Trust v Atos IT Services UK Ltd
[2017] EWCA
Civ 2196), Mrs Justice O’Farrell considered as a preliminary issue whether and
to what extent Atos’s liability was limited by paragraph 9.2 of Schedule G.

Applying the legal principles on interpretation referred
to above, she rejected the Trust’s submission that the provision was not
capable of being construed and should be declared unenforceable. Instead, she
held that it imposed a single cap which, depending on the timing of the first
default, would be either that set out in paragraph 9.2.1 or that in paragraph

She considered that the words ‘aggregate liability … shall
not exceed’ at the start of paragraph 9.2 indicated that the parties intended
to limit the total or composite liability of Atos. When read together with the
word ‘or’ between sub-paragraphs 9.2.1 and 9.2.2, she considered that this
indicated that the parties intended to agree that one cap would apply.

She went on to say that, if the provision were interpreted
as providing multiple caps where there were multiple claims, this could lead to
a liability cap that was multiple times the contract price if there were
several defaults during the first year. Similarly, she said that if there were
one breach in the first 12 months and another, one day after the first 12
months, the level of the cap would approach double the contract price. She
considered it ‘very unlikely’ that the parties would have agreed to such a term
because ‘the potential level of the cap would render it devoid of any real

The decision of the
Court of Appeal

Both parties appealed. The Trust appealed against the judge’s
decision that paragraph 9.2 imposed one cap rather than two. Atos’s
cross-appeal was asserted on the basis that, if the Trust’s appeal was
successful, the last four words of sub-paragraph 9.2.2 should be construed as
meaning ‘the date when that claim was notified’.

Jackson LJ and Lewison LJ in the Court of Appeal
considered that the central issue for them to decide was the natural meaning of
the words used, applying the test of a reasonable person who had all the
background knowledge of the parties. Jackson LJ observed that both parties were
well-resourced, commercial organisations with ready access to legal advice,
such that there was no reason to depart from the natural meaning of the words
used, once that meaning had been ascertained.

Jackson LJ disagreed with the judge’s view that the
provision imposed one rather than two caps. In his view, the phrase ‘aggregate
liability’ did not suggest one cap as it could also refer to the aggregate of
the sums referred to in the sub-paragraphs below. He also said that the word ‘or’
could have a ‘conjunctive’ as well as a ‘disjunctive’ meaning. A good reason
for using it (in the conjunctive sense) between sub-paragraphs 9.2.1 and 9.2.2
was that the sub-paragraphs were mutually exclusive in that each referred to a
discrete period and the two periods did not overlap.

He considered that the language of paragraph 9.2 pointed
emphatically to there being two separate caps. For any default or defaults
occurring during the first 12 months, Jackson LJ held that Atos’s liability was
capped at the contract price. For any default or defaults occurring during the
remainder of the contract term, Atos’s liability was capped at the amount of
the contract charges paid in the preceding 12 months. If defaults occurred
during both periods, both caps would apply: the cap of the contract price to
defaults in the first 12 months, the cap of the last 12 months’ charges to
defaults in the subsequent period.

Jackson LJ considered that there was nothing surprising
about this arrangement. Atos was expected to do the high-value work in the
first year, when defaults could have very expensive consequences. Lower-value
work would be undertaken in the subsequent years, with less expensive
consequences. He accepted that this interpretation could still lead to some odd
results, depending on when defaults occurred; however, he considered that the
natural meaning, which yielded the least bizarre consequences, was that the
provision imposed two separate caps.

Atos’s cross-appeal was refused. The parties had already
agreed that the words ‘claim’ and ‘claims’ at the beginning of each of the
sub-paragraphs of paragraph 9.2 meant ‘default’ or ‘defaults’, respectively. If
the phrase ‘claims arising’ at the beginning of sub-paragraph 9.2.2 referred to
‘defaults arising’, Jackson LJ considered that the phrase ‘that claim’ at the
end of that sub-paragraph must be a reference to a default occurring, rather
than a claim being communicated.

liability caps with clear and precise wording

The case serves as a useful reminder of the importance of
negotiating clauses with clear and precise wording. Limitation of liability
provisions are often one of the most heavily negotiated parts of IT and
outsourcing contracts and need careful consideration to ensure that they can be
relied upon if problems occur.

When negotiating liability caps, parties should consider:

If liability is to be linked to the value of
charges paid under their contract, whether the cap will apply multiple times on
a ‘per claim’ basis or only once as an ‘umbrella’ for all claims arising within
a particular period.

How any connected claims will be treated; and
whether the cap will be calculated by reference to charges paid or charges payable
under the contract.

Whether any time boundary for defining the value
of a cap will be triggered by (i) the date on which the claim first arose, (ii)
the date when the claim was first discovered by a party, or (iii) the date on
which the claim was formally notified by a party.

Where there are several contracts governing a
relationship, whether there are any conflicting provisions purporting to cap or
otherwise limit liability.

Whether a cap should apply to certain liabilities
only. For example, will it include payments made under an indemnity, liquidated
damages, service credits or interest on any amount awarded by a court or
arbitration tribunal?

The level of a cap – if the Unfair Contract
Terms Act 1977 applies, eg in a business to business contract where one party
is dealing on the other’s standard terms, a term purporting to limit liability
for breach of contract will apply in so far as it is ‘reasonable’. 

In interpreting a liability cap, although the court will
consider commercial common sense and the surrounding circumstances, its focus
will be on the language used. This is something that contracting parties have
control over, and they should exercise it carefully.

Where the words of a provision are unclear, the more ready
the court will be to depart from their natural meaning. As the divergent views
of the judges in the TCC and the Court of Appeal in this case demonstrate, a
confusing clause can be interpreted in different ways. This can lead to
uncertainty and the potential for costly and time-consuming disputes. 

James Farrell is a Partner at Hebert Smith Freehills LLP
based in London.

Sophie Jones is a Senior Associate there.