“No Oral Modification” – does it mean what it says?

“Boilerplate clauses” are a standard part of most written contracts and are rarely given much thought.  They provide the basic provisions which are considered appropriate in nearly all contracts.  A common one provides that any variation to the agreement must be in writing and signed by or on behalf of the parties.  This is known as a “No Oral Modification” clause, or “NOM”.  Its purpose is to reduce the potential for future disputes where one party seeks to argue that the other had orally agreed to their departing in some way from the terms of the written contract.  This helps create certainty (which is the point of putting contracts in writing), but the problem is that in practice the parties don’t read their contracts (and especially not the boilerplate clauses, which are considered “legalese”) and so do sometimes actually agree such oral variations, which they then proceed to act upon.

Take this example:

“All variations to this Licence must be agreed, set out in writing and signed on behalf of both parties before they take effect.”

This wording was in a licence to occupy serviced offices in central London granted by MWB Business Exchange Centres Ltd to Rock Advertising Ltd.  Rock fell into arrears and claimed to have agreed a revised payment schedule over the phone with MWB’s credit controller.  The credit controller’s boss didn’t approve the proposed payment schedule, and MWB evicted Rock and claimed the arrears.  Rock counterclaimed for wrongful eviction.  The judge found there was indeed an oral agreement to vary the licence, which the credit controller had authority to conclude, but it was ineffective as it didn’t comply with the NOM.

The case went all the way up to the Supreme Court, as there was no clear authority under English law whether NOM clauses were effective.  The general view, supported by recent cases (and by the Court of Appeal in this case), was that they were not – because the parties had freedom to contract orally and so could agree a subsequent oral contract which would impliedly override the NOM.  But lawyers still included NOMs in contracts – just in case they did work.

The Supreme Court, in Rock Advertising Ltd v MWB Business Exchange Centres Ltd [2018] UKSC 24, held 4 to 1 that NOM clauses did indeed work (so we lawyers were right to include them all along).  Lord Sumption, delivering the lead judgment, explained his view that:

“What the parties to such a clause have agreed is not that oral variations are forbidden, but that they will be invalid. The mere fact of agreeing to an oral variation is not therefore a contravention of the clause. It is simply the situation to which the clause applies. It is not difficult to record a variation in writing, except perhaps in cases where the variation is so complex that no sensible businessman would do anything else. The natural inference from the parties’ failure to observe the formal requirements of a No Oral Modification clause is not that they intended to dispense with it but that they overlooked it. If, on the other hand, they had it in mind, then they were courting invalidity with their eyes open.”

Lord Briggs disagreed with this analysis, but agreed the appeal should be allowed.  He took the view it was theoretically possible to agree orally to dispense with a NOM clause, but the Courts would only imply that the parties had done so where “strictly necessary”, rather than as a matter of course just because they had not complied with the NOM.  He therefore agreed with the majority that the oral variation was ineffective in this case.

So we now have clear authority that NOMs work, and that you can’t agree to delete them except in writing.  This is good for legal certainty, but is likely to create problems in those cases where the parties have agreed an oral variation anyway and gone ahead and acted upon it.

In such cases, as Lord Sumption pointed out, “the safeguard against injustice lies in the various doctrines of estoppel”; i.e. if something is agreed orally and one party acts in reliance on it to their detriment, the party who allowed this to happen will be “estopped” from relying on the NOM.  You might think this is the same thing as allowing oral variation of NOMs, but the subtle legal difference is that estoppel is an equitable doctrine which allows the Courts to do justice in individual cases rather than a hard and fast rule that a party can always rely on.  So the contract remains as per the written terms, but that doesn’t mean you’ll be able to enforce it if you’ve allowed the other party to believe you agreed you wouldn’t.

Downloaded software is not "Goods"

The Commercial Agents (Council Directive) Regulations 1993 provide for the payment of compensation to a commercial agent whose agency agreement is terminated by the principal without cause, even when terminated under a notice clause in the agreement.  However an important limitation is that they only apply to to agents authorised to negotiate or conclude "the sale or purchase of goods" on behalf of their principal.  Agencies to negotiate the supply of services by the principal are not covered.

So what is the position when an agency for the supply of software is terminated?  Is software "goods" for this purpose?  The Regulations, and the EU Directive which they implemented, do not define "goods".

This question came up in the case of Computer Associates UK Ltd v The Software Incubator Ltd [2018] EWCA Civ 518 decided by the Court of Appeal on 19 March 2018.  In that case Computer Associates had terminated an agency agreement to resell their release automation software, which was supplied by electronic download only, and not on disks, by way of perpetual licence.  The judge held that the Regulations should be interpreted so that "goods" included downloadable software, and that Computer Associates had wrongly terminated the agency, as the agent was not in breach of contract.  He therefore awarded the agent £475,000 in compensation for the loss of its future income stream.

Lady Justice Gloster, delivering the judgment of the Court of Appeal agreed that Computer Associates had not been entitled to terminate the agency, but disagreed that downloadable software was "goods" under the Regulations.  She referred to the earlier St. Albans and Your Response cases, which had made a distinction between software provided on physical disks and software provided by electronic download, and held that only the former constituted "goods".  She noted that the Consumer Rights Act 2015 (which implements the EU Consumer Rights Directive and now governs the sale of goods to consumers - though not to businesses) accepted this distinction as being the existing law and provided for a new category of "digital content" to give consumers equivalent rights for downloaded content to those they had for physical goods.  As the software here was not "goods", the Regulations therefore did not apply, the agent's £475,000 compensation was disallowed, and it was left with the £15,000 the judge had awarded as damages for breach of contract.

This case confirms the orthodox understanding that packaged software sold on physical disks is "goods" but software downloaded from the internet is not.  The reality nowadays is that almost all software is sold by download.  Consumers have the protection of the digital content provisions of the Consumer Rights Act 2015, but those do not apply to businesses, who cannot therefore claim that downloaded software is not of satisfactory quality under the Sale of Goods Act 1979.

In any case, much software is now supplied as a Cloud-based service, especially in a B2B context.  This will definitely not be "goods" when the sale is negotiated by an agent, but is more likely to be considered as a "service" given the way Cloud subscription agreements are typically structured.  The Commercial Agents Regulations will not apply in such cases, but the implied warranty that the supplier has used reasonable skill and care in the provision of the services under the Supply of Goods and Services Act 1982 would apply if not contractually excluded.

Restrictive Covenants - how long and how wide?

This is one of the hardest questions to advise upon when drafting contracts of employment for employer clients (or advising employee clients whether their covenants are enforceable).  It depends what is "reasonable" and cases are of limited use, as each turns on its own facts.  In other words, you have to second guess what a judge might think.

Here's an example:

"13.2. You shall not without the prior written consent of the Company directly or indirectly, either alone or jointly with or on behalf of any third party and whether as principal, manager, employee, contractor, consultant, agent or otherwise howsoever at any time within the period of six months from the Termination Date:
[...]
13.2.3 directly or indirectly engage or be concerned or interested in any business carried on in competition with any of the businesses of the Company or any Group Company which were carried on at the Termination Date or during the period of twelve months prior to that date and with which you were materially concerned during such period;"

So, a non-compete clause for 6 months and apparently with no territorial limitation.  This is taken from the recent case of Egon Zehnder Ltd v Tillman [2017] EWHC 1278 (Ch) and was for a fairly high-powered executive headhunter in the financial services sector (a former European managing director and COO of JP Morgan before taking a career break).  The reasonableness of restrictive covenants has to be assessed at the date the contract was entered into (rather than when the employee leaves), and here she had started at the most junior "consultant" level (albeit on a guaranteed minimum of £220k p.a.) but eventually left at the most senior "partner" level.  The Group operated worldwide.

On first sight one might think 6 months is fine but no territorial limitation must be unreasonable.  But that's not quite how Mr Justice Mann approached it in the High Court, which shows how difficult these cases can be to advise upon.

The defendant did indeed try to argue the covenant was unenforceable due to its global reach, but didn't get very far.  The judge construed the restriction as only applying to businesses which competed with the businesses of Group Companies with which the employee had been materially concerned (and she had been concerned with some in other countries).  Thus there was "an in-built restriction on the global reach to the clause, deriving from the need for Mrs Tillman to have been involved locally."  This was therefore limited to what was reasonable to protect the Group's business.  Worth bearing in mind when drafting non-compete clauses for clients who say their business is global so they don't want a territorial restriction.  In reality they are unlikely to be operating in every part of the world, and involving the employee in those operations, so this formulation is a neat way of dealing with this.

The defendant also sought to argue that the clause prevented her from holding any shares in a competitor as an investment, and so went beyond what was reasonable to protect the employer's interests.  However there was another clause that expressly allowed shareholdings of up to 5% as an investment whilst still employed, so the Court held that the non-compete clause couldn't have been intended to have that effect after employment.  This shows the current approach of construing the wording to find the presumed intention of the parties, rather than construing any ambiguity against the party seeking to rely on the restrictive covenant.  It is also a reminder it's best to include an express exception for investments of this sort.

On the reasonableness of the 6 months, the discussion centered on how senior this employee really was.  The conclusion was that although she started at the junior consultant grade, she was clearly a high flyer whom the parties expected to move up the ranks (as proved to be the case).  So although her contract had never been updated, she was sufficiently senior that the covenant was enforceable against her. "Six months seems to me to be a reasonable period," said the Judge, "principally to allow the substitution of new relationships with the client and for the fading of confidentiality."  The implication here is that it could be tricky to argue for more than 6 months, even in fairly senior roles - though as ever it all depends on the facts of the particular case.

Textualism and Contextualism

Working out what an ambiguously drafted clause in a contract means (or what a court is most likely to decide it means) is one of the trickier tasks for us lawyers.  The case law provides guidelines on the principles of contractual interpretation, but they do sometimes seem to conflict.

In Rainy Sky SA v Kookmin Bank in 2011 the Supreme Court took the contextualist approach, looking at the factual context and preferring the interpretation most consistent with business common sense. In Arnold v Britton in 2015 (which I blogged about here) the Supreme Court took the textualist approach, preferring the literal interpretation of the words used even if they gave an unjust result to one party.  So have the courts now "rowed back" from contextualism towards textualism?

Not according to the Supreme Court in the latest case on the principles of contractual interpretation, Wood v Capita Insurance Services Ltd.  According to Lord Hodge, "Textualism and contextualism are not conflicting paradigms in a battle for exclusive occupation of the field of contractual interpretation. Rather, the lawyer and the judge, when interpreting any contract, can use them as tools to ascertain the objective meaning of the language which the parties have chosen to express their agreement."

Which tool you use depends on the contract.  Sophisticated contracts which have been professionally drafted are more likely to be interpreted textually.  But the court recognised that "negotiators of complex formal contracts may often not achieve a logical and coherent text because of, for example, the conflicting aims of the parties, failures of communication, differing drafting practices, or deadlines which require the parties to compromise in order to reach agreement".  (This all sounds familiar from my experience of negotiating deals.)  More informal contracts which have been drafted without professional assistance are more likely to be interpreted contextually (perhaps because they may not make sense if taken out of the context and read literally).

Having made all that clear, the Supreme Court in Wood proceeded to uphold the Court of Appeal's literal interpretation of the contract (overruling the first instance judge's more contextual interpretation), but explained that this also made sense in the context.  The clause in question was an indemnity in a share purchase agreement by which Capita acquired a motor insurance broker specialising in classic cars.  It read (my highlighting):

“The Sellers undertake to pay to the Buyer an amount equal to the amount which would be required to indemnify the Buyer and each member of the Buyer’s Group against all actions, proceedings, losses, claims, damages, costs, charges, expenses and liabilities suffered or incurred, and all fines, compensation or remedial action or payments imposed on or required to be made by the Company following and arising out of claims or complaints registered with the FSA, the Financial Services Ombudsman or any other Authority against the Company, the Sellers or any Relevant Person and which relate to the period prior to the Completion Date pertaining to any mis-selling or suspected mis-selling of any insurance or insurance related product or service.”

Capita claimed £2.4m under the indemnity re the cost of a remediation scheme required by the FSA (now the FCA) to compensate customers for mis-selling, but the problem was that it had not followed and arose out of complaints by customers but from self-reporting by the company in accordance with regulatory requirements.  On textual analysis of this "opaque" provision the Supreme Court held that it did not cover losses which followed or arose otherwise than out of complaints.  Also looking at the context, it noted that there were wider warranties which did cover the loss (but were subject to a 2 year time limit which Capita had, for some reason, missed) and "It is not contrary to business common sense for the parties to agree wide-ranging warranties, which are subject to a time limit, and in addition to agree a further indemnity, which is not subject to any such limit but is triggered only in limited circumstances."

This last comment is very true, but in my experience indemnities usually focus on particular potential liabilities against which the buyer requires specific protection, and are not (or should not be) dependent on whether the liability arises in a particular way.  This looks more like a case of poor drafting to me.  But, as the court pointed out, it is not their function to improve a bad bargain.

All in all this is a helpful case in explaining how to go about interpreting contracts and what are the tools for the job, and it is good to see the Supreme Court showing such understanding of the realities of negotiating and drafting share purchase agreements.

I'm thinking of having a bumper sticker made for my BMW: "No tools of textual or contextual exegesis are left in this vehicle overnight".

The Dangers of Going Ahead Before Agreeing the Contract

Samuel Goldwyn is famously misquoted as having said that a verbal contract isn't worth the paper it's written on, and there’s much truth in the saying.

Surprisingly often, solicitors are engaged to advise on contract terms but, due to commercial pressures or over-eagerness, the parties go ahead with the work before they’ve agreed all the terms, or they agree the terms but then don’t sign the contract.  The courts will usually find that there is a contract where there has been actual performance, but it won’t be a written one, and so the question arises what are its terms?

In the case of Arcadis Consulting (UK) Ltd v AMEC (BSC) Ltd [2016] EWHC 2509 Hyder carried out design works for Buchan, the sub-contractor on two large building projects.  Over 15 years later one of the designs proved to be defective and Buchan claimed £40M of damages from Hyder, who argued that their liability was subject to an agreed cap of £610,515.  The problem was that the parties had never reached agreement on the terms of their contract, but had gone ahead with the work anyway.  There were three competing versions of the terms and conditions, all of which included a cap on Hyder’s liability, but none of which had been agreed.

The judge held that that:

• there was a simple contract between the parties that Hyder would carry out design work and would be paid for that work by Buchan;
• the contract did not include any of the three different sets of proposed terms and conditions; and
• that there was no limitation on Hyder’s liability - despite the fact that every set of proposed terms and conditions included some sort of provision to that effect.

He was critical of Hyder’s unco-operative approach to negotiations and concluded his judgment by saying:

“This case starkly demonstrates the commercial truism that it is usually better for a party to reach a full agreement (which in this case would almost certainly have included some sort of cap on their liability) through a process of negotiation and give-and-take, rather than to delay and then fail to reach any detailed agreement at all.”

Where terms have been agreed but the contract just hasn’t been signed, the court is likely to find that the full unsigned terms apply, so long as the parties have acted consistently with them (at least up to the point the dispute arose).  But where one or both parties have made it clear that they do not agree to a term, the court will not find that they are bound by it just because they have gone ahead with the transaction.  A sneaky negotiator might therefore think there’s scope to dispute the terms you don’t like and to go ahead without signing the contract, on the basis that the remaining terms will apply.  The Arcadis case shows the dangers of this: you could end up with none of the terms applying - including ones in your favour that the other party was prepared to agree.

As a lawyer I always try to achieve certainty for my clients. Unsigned contracts mean uncertainty, with increased risk of a dispute ending up in court, which is always costly even if you win the case at the end of the day.

Why bother with the formalities when you can "Duomatic" it?

I was recently asked to advise on some new Articles of Association a client company proposed to adopt.  When I asked about the Special Resolution of the shareholders to adopt the new Articles, I was told the Directors intended to agree them at their next Board meeting and they thought it was just "a paperwork exercise".  Well it usually is, but I do try to get the paperwork right.  I advised that the necessary resolution should be passed by a 75% majority at a General Meeting convened for the purpose or by circulating a Written Resolution for signature.  However, it turned out that there were only a small number of shareholders, all of whom were on the Board of Directors.  As a recent case illustrates, the "Duomatic principle" would in fact have validated the client’s informal procedure.  So did I really need to bother with the correct, formal advice?

It is a well-established principle of company law that where all shareholders who have a right to attend and vote at a general meeting of the company assent to some matter which a general meeting of the company could carry into effect, that assent is as binding as a resolution in general meeting would be (Re Duomatic Ltd [1969] 2 Ch 365).  There are a number of cases where amendments to the Articles of companies have been held valid in this way, despite a lack of the formalities prescribed by the Companies Act.

The recent case of Randhawa & Ors v Turpin & Anor [2016] EWHC 2156 (Ch) is a good illustration of just how far the Duomatic principle can go.  75% of the shares in the company were held by the sole Director as nominee for his father, who was disqualified from acting as a director, with the remaining 25% being registered in the name of an Isle of Man company which had been dissolved in 1996.  The sole Director had purported to hold a Board meeting at which he had appointed Administrators.  But the Articles provided that a sole director only had power to convene a general meeting or appoint an additional director.  A creditor (which had taken an assignment of the debt owed to the company’s solicitors) challenged the validity of the appointment of the Administrators.  However this was only after they had lost at a previous hearing seeking to challenge the amount of their fees, when they had not taken this point, so the judge was unsympathetic, saying that at best this smacked of abuse of process.

But he decided the case on the Duomatic principle.  From 2009 to the appointment of the Administrators in 2013 both the disqualified father (who was the beneficial owner of 75% of the shares and in reality in control of the company) and the son (who was sole director and registered holder of the 75% shareholding at the time) had acquiesced in the son exercising the full powers of the Board of Directors.  The father had also acquiesced in the appointment of the Administrators, though not actually present at the meeting appointing them.  This was held to be an effective variation of the Articles under the Duomatic principle to allow the sole Director to appoint the Administrators.  The 25% shareholder did not count because it had been dissolved, and so was unable to exercise its voting rights.

The Duomatic principle can be very useful to cure formal defects in procedure where the reality is that all the shareholders whose formal consent was needed had agreed to the matter.  But to go back to my original question, why bother with the formalities then?

The answer is that you do not want to rely on a Court decision to validate things.  It is always best to get it right first time, so there can be no argument about it.  You would need to prove the shareholders all agreed, and the best way to do this is to get them to pass a resolution in the first place.  If they don’t all agree then you will need to pass the resolution by the requisite majority, because Duomatic requires 100% consent or acquiescence (from those who exist and are entitled to vote at least).

Update: on 1 August 2017 the Court of Appeal allowed an appeal against the decision of the High Court in Randhawa v Turpin, disagreeing that the dissolved 25% shareholder could be ignored.  It remained a member of the company (despite not existing) and could not have given its informal consent when it did not exist.  Which supports my original point, that it is best to get the formalities right in the first place.

Warranties and Representations

It is common in share purchase agreements for the warranty clause (the first draft of which is prepared by the Buyer's solicitors) to use wording such as "The Sellers warrant, represent and undertake that…"  Whilst this may just be a case of lawyers preferring to use three words when one will do, there is often more to the use of such language than meets the eye.

A warranty is a contractual promise that something is so; e.g. that the Company does not have any liability in respect of a particular matter.  If that turns out not to have been so, the Buyer has a claim for breach of contract.  Such claims are subject to carefully-negotiated limitations of the Buyer’s liability under the share purchase agreement; typically a cap of the amount of purchase price received, a minimum threshold for claims, and a time limit for bringing claims of 1 to 3 years (or 6 or 7 years where tax is involved).

A representation is a statement of fact that induces a party to enter into a contract.  If it turns out to have been untrue, the other party may claim damages for misrepresentation.  This is a claim in tort (a legal wrong), not a claim for breach of contract, and the damages are calculated differently.  The limitations of liability in the share purchase agreement are not usually drafted to cover liability for misrepresentation.  This is why buyers’ solicitors try to include the language of representation, and sellers’ solicitors seek to delete it.  Such deletions are usually accepted without serious argument - though private equity investors' solicitors may take a tougher line, and be in a stronger negotiating position.

An undertaking is a contractual promise to do something (or not to do it).  This is completely inappropriate language for the warranties in a share purchase agreement, and is either bad drafting or a cunning attempt to hide the word "represent” amongst some apparent bad drafting.

Acting for the Seller, one therefore always seeks to avoid the language of representation, and to include the usual boilerplate "entire agreement" clause to the effect that this is the entire agreement between the parties, it supersedes all prior negotiations, and the Buyer acknowledges it is not relying on any previous representations.  This language is intended to exclude liability for misrepresentation, and with an express exception for liability for fraudulent misrepresentation is generally considered to be reasonable if negotiated at arms’ length between commercial parties with the benefit of legal advice (This is important because under the Misrepresentation Act 1967 liability for misrepresentation can only legally be excluded to the extent the exclusion is reasonable.)

In the case of Idemitsu Kosan Co Ltd v Sumitomo Corporation [2016] EWHC 1909 (Comm) (27 July 2016), Idemitsu were out of time for bringing a contractual claim for breach of warranty under a share purchase agreement (the agreed 18 month time limit for non-tax warranty claims having expired without a claim having been made), so they sought to get round this by bringing a claim for misrepresentation under s.2(1) of the Misrepresentation Act 1967.  Sumitomo responded with an application under CPR Part 24 for summary judgment dismissing that claim, on the basis that it had no real prospect of success and there was no other compelling reason why it should be disposed of at a trial.

Such cases always depend on the wording of the agreement, and Idemitsu were in some difficulty there, as the relevant clauses only used the language of warranties.  The word "representation" only appeared in the entire agreement clause, apparently intended to exclude them.  However there were conflicting previous cases on the point: in one Arnold J. had decided that warranties could of themselves amount to representations, and in another Mann J. had decided that they could not.  Both are eminent judges.  Idemitsu’s Counsel also ran a clever argument that by putting forward the agreement with the warranties for execution, Sumitomo had made representations inducing Idemitsu to enter into the agreement.  The target Company had interests in North Sea oil and gas fields, had been sold for US$575M, and Idemitsu was seeking to recover a claimed loss of US$105.9M (as against a contractual cap on warranty claims of US$1,5M) relating to liability for sharing the operating expenses of a floating production storage and offshore loading vessel.  So it must have seemed worth a try.

Andrew Baker QC sitting as a judge of the High Court was unconvinced by Idemitsu’s Counsel’s arguments, and preferred the reasoning of Mann J. from the previous cases.  He held that a warranty is (without further language) a contractual promise: nothing less, but nothing more.  He upheld the entire agreement clause as effective to exclude misrepresentations, of which there were none.  He therefore gave judgment for the defendant.

The case is welcome confirmation that my deletions of representation wording when acting for sellers were not in vain, and should be of comfort to sellers that their limitations on liability do mean what they thought they did.  However, given the amount at stake and the conflicting first-instance decisions, it may still go to the Court of Appeal.