Service Credits and Penalties: What Exactly is the Position?

January 1, 2003

It is common in outsourcing (or other services contracts) to see some clause such as the following in a Service Level Agreement (SLA) for the operation of a computerised system for running a factory:

“For each week in which the service is not offered according to the standards set out in the Service Level Agreement, the Service Provider will pay the User £1,000 by way of liquidated damages and not as penalty, up to a maximum of 8 weeks.”

Obviously, as anyone who works in the area knows, Service Level Agreements can be far more complex than this simple example, although such a simple example will serve to demonstrate some of the more open questions still unresolved about service credits and liquidated damages.

For someone who comes to such clauses more as a dispute resolver rather than draftsman, a number of questions immediately raise themselves. Let us say that the service provider in fact misses the target for 12 weeks running and as a result the factory has to close down causing loss of profits for the user in the order of £1 million. Even leaving to one side the old question of whether this clause is a penalty and unenforceable, this still leaves the following questions:

  • The clause is expressed to run for 8 weeks as a maximum, so is 8 times £1,000 the maximum sum recoverable by the user, or is it the maximum sum recoverable for that 8 week period, leaving the user able to claim general damages for losses occurring after those 8 weeks?
  • Given that the level of LDs is so small compared with the losses actually incurred, is there some principle akin to penalties, but operating in favour of the user, to enable it to overturn an liquidated damages provision which turns out to be too small in relation to the loss which has actually been incurred?
  • Does the Unfair Contract Terms Act 1977 (UCTA) come into play and potentially overturn a liquidated damages provision which is set too low?
  • Can the user simply waive the liquidated damages provision and claim general damages anyway?

As we shall see, the law on these topics is by no means so settled as the ubiquity of Service Level Agreements (SLAs) and service credits would suggest. In fact the classic question of penalties is raised by many SLAs and service credit provisions: for many “technical” breaches of the detailed provisions of an SLA (eg service dips below 98.75% availability in any period) there will be no measurable loss suffered by the user: if service was 98.74% available, the Service Credit is a true windfall in the user’s hands. An alternative view might be to say that any loss is simply impossible to assess and the parties have done their best by the agreement of the service credit. Whichever view is taken, it seems the service credit is there to “encourage” compliance by the service provider with the detailed provisions of the SLA. Such clauses thus – at first blush – fall within the prohibition against clauses stipulated in terrorem and where there is no genuine pre-estimate of loss. Does this mean that all Service Credits in SLAs are unenforceable?


Origins of penalties and basic law

The origin of the law on penalties is a curious one: in perhaps no other area does the common law apparently interfere with parties’ contracts. Some of the cases look at the origins of the rule, but it is not exactly clear where it derives from. Its origin seems to be in equity. Some support for this is to be found in Law v Local Board of Redditch [1892] 1 QB 127 where Kay LJ said:

“In early times it was decided by Courts of Equity that, where a sum of money was agreed to be paid as a penalty for non-performance of a collateral contract, equity would not allow the whole sum to be recovered; but, where the damages for non-performance of such contract could be estimated, would cut down the penalty to the amount of the actual damages sustained.”

Thus, in that case, where there was a contract for the construction of sewerage works by a certain date, the contractor should, in default of meeting that date, pay a sum of money for each period of 7 days delay. It was held that this was to be regarded as liquidated damages, and not a penalty.

While there are many cases on the subject, in recent times it appears that the doctrine has some life in it and could well be developed as a principle of more general application. So in recent years the courts seem to have been more willing to apply similar principles of law to forfeiture types of cases. For example, in Stockloser v Johnson [1954] 1 QB 476, there was a contract for the supply of plant and machinery with provision for instalment payments. It was further provided that, if the buyer defaulted, the seller might terminate the contract and forfeit the instalments already paid. A majority of the Court of Appeal found that the court had an equitable jurisdiction to relieve against forfeiture in cases where the clause was penal. Subsequent cases have shown some development with the House of Lords apparently keen to restrict the doctrine to cases where there is some element of the transfer or creation of proprietary or possessory interests (eg Scandinavian Trading Tanker Co AB v Flota Petrolera Ecuatoriana [1983] 2 AC 694). It seems that the category of clauses that might be “penalties” is not necessarily closed.

Although there is some life in the doctrine, by the same token, it has to be said that courts do seem reluctant to find that clauses in the nature of liquidated damages clauses are in fact unenforceable penalties. The starting point is still Dunlop Pneumatic Tyre Co Ltd v New Garage and Motor Co Ltd [1915] AC 79 and the well known speech of Lord Dunedin (pp 86-88). At the risk of repeating well known propositions of law, he said:

“(1) Though the parties to a contract who use the words ‘penalty’ or ‘liquidated damages’ may prima facie be supposed to mean what they say, yet the expression used is not conclusive. The court must find out whether the payment stipulated is in truth a penalty or liquidated damages. …
(2) The essence of a penalty is a payment of money stipulated as in terrorem of the offending party; the essence of liquidated damages is a genuine covenanted pre-estimate of damage.
(3) The question whether a sum stipulated is a penalty or liquidated damages is a question of construction to be decided upon the terms and inherent circumstances of each particular contract, judged of as at the time of the making of the contract, not as at the time of the breach.
(4) To assist this task of construction various tests have been suggested, which if applicable to the case under consideration may prove helpful, or even conclusive. Such are:
(a) It will be held to be a penalty if the sum stipulated for is extravagant and unconscionable in amount in comparison with the greatest loss which could conceivably be proved to have followed from the breach.
(b) It will be held to be a penalty if the breach consists only in not paying a sum of money, and the sum stipulated is a sum greater than the sum which ought to have been paid. …
(c) There is a presumption (but no more) that it is a penalty when ‘a single lump sum is made payable by way of compensation, on the occurrence of one or more or all of several events, some of which may occasion serious and others but trifling damage’.
On the other hand:
(d) It is no obstacle to the sum stipulated being a genuine pre-estimate of damage, that the consequences of the breach are such as to make precise pre-estimation almost an impossibility. On the contrary, that is just the situation when it is probable that pre-estimated damage was the true bargain between the parties.”


Are all service credits penalties?

The last point in the extract from the Dunlop case can help us dispose quickly of one of the questions posed above: if service credits are agreed in circumstances where there could well be no (or only unquantifiable) loss, it must be at least arguable that pre-estimation of loss was impossible and that the parties must be held to their bargain. The fact that the service credits may be seen as a windfall in the user’s hands is nothing to the point. In favour of upholding the validity of such clauses is the very complexity of such provisions and the fact that the parties do (in general) devote a lot of thought (and time) to the drafting of such clauses with the effect that the amount of the service credits does reflect the seriousness of the breach. This is reinforced by the relatively modest amounts that are payable by way of service credits.

In favour of this proposition is the fact that the courts are in general reluctant to hold a clause unenforceable as a penalty. The case frequently cited as support for this proposition is Robophone Facilities Ltd v Blank [1966] 1 WLR 1428, where Diplock LJ said:

“Nevertheless the courts would be doing an ill turn to those whom the rule about ‘penalty clauses’ is designed to protect if they were to apply it so as to make it impracticable for parties to agree at the time when they enter into a contract upon a fair and easily ascertainable sum to become payable by one party to another as compensation for the loss which the latter will sustain as a consequence of its breach. It is good business sense that parties to a contract should know what will be the financial consequences to them of a breach on their part, for circumstances may arise when further performance of the contract may involve them in loss. And the more difficult it is likely to be to prove and assess the loss which a party will suffer in the event of a breach, the greater the advantages to both parties of fixing by the terms of the contract itself an easily ascertainable sum to be paid in that event. Not only does it enable the parties to know in advance what their position will be if a breach occurs and so avoid litigation at all, but if litigation cannot be avoided, it eliminates what may be the very heavy legal costs of proving the loss actually sustained which would have to be paid by the unsuccessful party. The court should not be astute to descry a ‘penalty clause’ in every provision of a contract which stipulates a sum to be payable by one party to the other in the event of a breach by the former.”

Thus, it seems that service credits of modest proportion are beyond the reach of the law on penalties. However, there are other questions which are harder to answer. Reverting to the simple example given at the start of this note, it is proposed to look at some harder questions.

It is also notable that Lord Dunedin gives some weight to the description used by the parties, suggesting that there is some value in using the well worn expression “as liquidated damages and not as penalty”. The modern view on the interpretation of contracts is of course Investors Compensation Scheme Ltd v West Bromwich Building Society [1998] 1 WLR 896 which gives a more fluid view of how to interpret the actual words used by the parties. This being so, it is worth spelling out exactly what the consequences are intended to be – the court will not assume that the parties meant by “penalties” what a lawyer understands by the term.

Service credits and limits of liability

One of the most difficult points arises when there is a converse of the services credits point – in that case, the user has, arguably, received a windfall – but in the converse case, its losses vastly exceed the amount provided for in the liquidated damages provision. Can the user simply ignore the liquidated damages provision and sue for the full loss?

The well known case of Suisse Atlantique v Rotterdamsche Kolen Centrale [1967] 1 AC 361 is of relevance here. This case is normally taken as authority for the writing out of the textbooks of any doctrine of “fundamental breach”, but it does deal directly with the present question. Suisse Atlantique was concerned with a charterparty where the charterers of a vessel failed to use it to carry out the number of voyages which might have been expected. Demurrage was payable at the rate of US$1,000 a day and the charterers admitted 150 days delay when they detained the vessel without useful work. Liability for US$150,000 demurrage was therefore admitted. The owners claimed that their actual loss was US$772,866.92 or alternatively US$476,490.92, based on the number of voyages which might have been expected if the charterers had proceeded with proper despatch.

The House of Lords dismissed the owner’s claim and limited it to the amount of demurrage. The point stressed in all the speeches is that it was all a question of construction of the clause in question. In this case, demurrage is the term used for agreed or liquidated damages for delay in operating a vessel; as this had been agreed, it was not open for the innocent party to disregard it and sue for its actual loss. This seems to provide, therefore, direct authority for the proposition that, where there is a valid liquidated damages clause, the user cannot of its own will disregard it and sue for its actual loss where this is greater. This should make the possible payee of any service credit provision consider carefully whether the amount is actually such as will provide proper compensation.

Can the user treat liquidated damages as a penalty?

However, can the user claim that the liquidated damages provision is in fact a penalty, so that the user can disregard the LDs provision and claim its full loss? This is a rather harder question which did not directly arise in Suisse Atlantique and needs to be considered.

In Cellulose Acetate Silk Co Ltd v Widnes Foundry (1925) Ltd [1931] 2 KB 393 and [1933] AC 20, there was a contract for the delivery and erection of acetone recovery plant. The contract contained a provision for the payment of £20 for each week of delay. The contractors sued for moneys due under the contract and were met with a counterclaim for 30 weeks’ delay amounting to £5,850. The contractors said that the liquidated damages provision meant that that maximum amount of the counterclaim was £600.

In fact the Court of Appeal held that this was a perfectly valid liquidated damages provision, not a penalty, and thus bound the parties. The actual decision was affirmed by the House of Lords. Scrutton LJ found difficulty in holding that an estimate of damages which was too low could be characterised as a penalty in terrorem. However, the overall effect of the judgment is that the party suing on the LDs provision is effectively limited to recovering the sums provided for in that provision, and cannot “waive” the wording in his favour and sue for his actual (and greater) loss. A closer look at the judgments is justified.

The House of Lords did little more than affirm the decision of the Court of Appeal. Lord Atkin, with whom the other lords agreed, said:

“… the Silk Company are only entitled to recover 20L. a week as agreed damages; and that the decision of the Court of Appeal was correct and should be affirmed. In these circumstances I find it unnecessary to consider what would be the position if this were a penalty. It was argued by the appellants that if this were a penalty they would have the option either to sue for the penalty or for damages for breach of the promise as to time of delivery. I desire to leave open the question whether, where a penalty is plainly less in amount than the prospective damages, there is any legal objection to suing on it, or in a suitable case ignoring it and suing for damages. …”

There is nothing to indicate what Lord Atkin might think was a “suitable case”. Certainly, the judges in the Court of Appeal had little difficulty in holding the defendants to their bargain, and awarding only the amount of the LDs. Scrutton LJ, as might be expected, referred to some of the older cases on charterparties, where historically there had been a provision providing “penalty for non-performance of this agreement estimated amount of freight”. As Scrutton LJ observed, this was quickly held to be a penalty and unenforceable, since it provided a single remedy for any type of breach of the agreement. Two cases held in the context of similar provisions in charterparties that parties might sue for actual loss, though this might be more than the estimated freight (see Wall v Rederaktiebolaget Luggude [1915] 3 KB 66 and Watts, Watts & Co Ltd v Mitsui & Co Ltd [1917] AC 227). Scrutton LJ explained the difference between those cases and a case such as Widnes by saying that the charterparty cases were concerned with one remedy for every type of breach, whereas the Widnes case was concerned with one remedy for one type of breach. Greer LJ in the Widnes case was, however, perhaps more ambivalent, and said:

“there is something to be said for the view … that the whole origin of the right to object to recovery of a penalty is the equitable doctrine that a plaintiff suing for a penalty would be restrained by the Court of Chancery from seeking to enforce it in respect of any figure beyond the amount of the actual damages, and that it has no application to the case where the alleged penalty is something less than the actual damages. But if I had to decide the case on that ground I should require a little more time to consider my judgment and to investigate with considerable care the somewhat doubtful origin and history of the law with regard to the enforcement of penalties.”

Slesser LJ in the same case was more forthright, and said:

“But when the sum is less, I think it is reasonable to assume that the parties intended to limit the damages to a certain amount, and that is an agreement perfectly within their power to make. It neither invades any equitable doctrine on the ground of unconscionableness, nor … does it lead to an absurdity, but on the contrary is a very reasonable and natural thing.”

Whether a limitation of limitation can be seen as a “very reasonable and natural thing” we will review in a moment when we come to look at UCTA.

Differing views

Chitty is somewhat equivocal as to whether this principle is established as a general proposition of law applicable to all contracts (28th edn, paras 27-117 and 27-118). As Treitel observes (10th edn, p 933), it would be strange if a clause could be both penal and unenforceable as well as an effective limitation of liability; Treitel therefore concludes that Wall v Rederaktiebolaget Luggude may well be right. The editors of Halsbury’s Laws (vol 12(1) para 1073) offer more solid advice:

“Where a clause is penal, the innocent party is at liberty either to sue to enforce the clause itself, which remains in the contract and capable of being sued upon, or to disregard the clause and sue for damages at large. The value of the former option is limited for two converse reasons. First, the stated penal sum becomes the maximum sum recoverable, regardless of the true measure of the innocent party’s loss. Secondly, a penalty clause will not be enforced beyond the sum which represents actual loss, that is (1) the actual loss of the party seeking enforcement if the breach consists of the non-payment of money; or (2) the recoverable loss of the party seeking enforcement if the breach consists of failure to perform some other obligation. The wiser course is to disregard the penalty clause and sue for damages at large, which can exceed the sum stipulated in the clause. It follows that little is lost by incorporating a penalty clause in the contract.”

Wall v Redeiaktiebolaget Luggude is given as authority for some of these propositions (although, as we have seen, Scrutton LJ sought to distinguish this and other similar cases in the Widnes Foundry case and the point seems to have been left open, a point conceded in footnote 11 to para 1073 of Halsbury). The other case frequently cited in the footnotes to this paragraph as supporting these propositions is Jobson v Johnson [1989] 1 WLR 1026, and it is therefore necessary to look at what can be derived from this case.

Jobson v Johnson concerned an agreement for the sale of a football club. The Rubins brothers inherited 62,666 ordinary shares in Southend United FC from their father (this was some 44.9% of the issued share capital). By two contracts of August 1983, they sold these shares to the defendant. The defendant agreed to pay the Rubins brothers further sums by instalments but, in the event of default in these payments, there was an obligation to transfer to the Rubins brothers 44.9% of the issued share capital or to make a cash payment instead (the defendant had the option of transferring the shares or making the cash payment). This was the basic contractual scheme, although there were some contractual variations (and the Rubins brothers assigned their rights to the plaintiff in the action).

Before the Court of Appeal, it was not challenged that the effect of the clause providing for transfer of shares of payment was indeed a penalty clause. However, it is a complex case, not least because the clause in question involved not just a penalty payment but also the transfer of shares (ie a transfer of property as opposed to a payment of money). Matters were also complicated by the obvious procedural complexities in that case, as well as by the fact that there was a suggestion that, by the time of the appeal, the value of the shares had increased considerably.

With this background, perhaps the clearest exposition of the law is to be found in the judgment of Nicholls LJ. The historical review of the law relating to penalties is extensive, but the core of it is as follows,

“An obligation to make a money payment stipulated in terrorem will not be enforced beyond the sum which represents the actual loss of the party seeking payment, namely, principal, interest and, if appropriate, costs, in those cases where … the primary obligation is to pay money, or where the primary obligation is to perform some other obligation, beyond the sum recoverable as damages for breach of that obligation. … Hence normally there is no advantage in suing on the penalty clause.

Strictly the legal position is that by the clause remains in the contract and can be sued upon, but will not be enforced in the court beyond the sum which represents, in the events which have happened, the actual loss of the party seeking payment.”

The majority view was therefore that the law provided for a “scaling down” (as Nicholls LJ put it) – bringing the award of damages down from the penal level to the level of the actual loss. The fact that there was property involved meant that the process was therefore somewhat more involved than if it was a straight money payment, but that was the essence of what the court should do. It is not clear that the Court of Appeal was seeking to change the law regarding liquidated damages, but perhaps the best explanation of the decision is that it illustrates the court’s approach where the penalty is something non-pecuniary, in that the law more closely approaches the equitable rules on forfeiture.

The case, taken as a whole, does not seem to be authority for the proposition that any party can, at its option, disregard a penalty provision, and sue for its whole loss where this is larger than the penalty provided for in the contract. The matter therefore seems to be open, at least in this jurisdiction.

What do the liquidated damages cover?

One issue that needs to be considered in every liquidated damages provision is exactly what, as a straight question of interpretation, falls within the ambit of the clause. This is useful, from the user’s point of view, in being able to sue for actual loss as opposed to being limited to the level of liquidated damages set out in the contract. It also highlights the importance in all cases of being specific in the drafting of liquidated damages provisions.

In Aktieselskabet Reidar v Arcos Ltd [1927] 1 KB 352, there was a charterparty for the performance of a number of voyages, and which required the charterers to load a “full and complete cargo” at a contractual rate of loading. If the ship was delayed beyond certain contractually stipulated times, demurrage was to be paid at a daily rate. In fact, the ship did not load at the contractual rate in Archangel, and in consequence the date in October passed after which a ship bound for Britain could not load beyond a certain amount of cargo (these were safety regulations to ensure ships did not travel heavily laden in winter conditions). The shipowner therefore claimed the cargo lost from the limit on the amount of cargo, and the charterers said that they could only be liable for the contractual rate of demurrage. It will be seen that the facts are very similar to those of Suisse Atlantique. However, the Court of Appeal found that the demurrage clause was there only for the losses that would normally arise from delay, and not from the loss of freight which resulted from the separate breach of loading less than the full cargo. This case was applied in Total Transport Corporation v Amoco Trading Co. [1958] 1 Lloyd’s Rep 423 and see also Dias Compania Naviera SA v Louis Dreyfus Corporation [1976] 2 Lloyd’s Rep 395.

In Surrey Heath Borough Council v Lovell Construction Ltd (1988) 42 BLR 25, there was a contract for the construction of a new office building, with liquidated damages provisions and provisions for insurance for damage by fire (these were standard to the RIBA/JCT design and build standard form). A sub-contractor’s negligence caused a fire before completion, and the contractor re-built the office and achieved completion some 67 weeks late, as was allowed under the extension of time clause in the contract. However, the council claimed for a wide range of loss and damage related to the fire and its consequences.

It was held in the TCC (Judge Fox-Andrews) that the right approach was to distinguish the types of loss, so that such items as furniture storage charges, loss of income from car parking charges and repairs to telephones in the old premises were “caused” by late completion, and thus fell within the liquidated damages provision. As there was an extension of time allowed by the contract, not even liquidated damages could be claimed. However, with regard to some other items, such as damage to property already installed before the fire and not covered by the insurance, these had nothing to do with late completion and so could be recovered.

One point seems to remain open. Many liquidated damages clauses (and service credits provisions) allow for a maximum period of payment – in the example given at the beginning of this article, the liquidated damages were expressed to be payable for a maximum of 8 weeks. What happens to accumulating loss and damage after this period of 8 weeks? The cases do not seem to say. It would come down, in line with general principles as exemplified by Suisse Atlantique, to a question of contractual interpretation. There does not appear to be a general rule to the effect that there is, in this case, a cut-off of 8 weeks’ worth of damages that can be claimed with any balance beyond the 8 weeks not allowed. For this reason, such expressions are risky without specific words setting out the intention as to recoverability of loss beyond the maximum period specified in the contract for the payment of liquidated damages: such a clause should either say expressly that no further damages are payable at all or that after 8 weeks the liquidated damages provision is of no further effect and that general damages are payable.

Is UCTA relevant to liquidated damages?

UCTA (unlike the consumer regulations) does not deal expressly with liquidated damages provisions. Indeed, it is unclear what the impact of UCTA exactly is in this context. In many cases of liquidated damages (and this is very common in the construction industry), liquidated damages are used not to provide a “genuine pre-estimate of loss” but in effect by way of limitation of liability for the consequences of delay. LDs provisions are agreed where the sums payable fall far short of the reasonably anticipated loss. As we have seen, the law is not entirely clear as to what the user’s options are in these cases, and it is by no means clear that it can simply disregard the liquidated damages, even if they are a penalty (in the true sense), and sue for damages at large.

Treitel is of the view (10th edn, p 933) that UCTA has no application to agreed damages provisions, the reason being that, since loss at a level lower than the liquidated damages is conceivable, such clauses are not exemption clauses within the ambit of UCTA. This would be an interesting conclusion indeed. With the IT industry so frequently on the receiving end of judicial attempts to undo “written standard terms of business” containing exclusions and limitations, the use of liquidated damages might be a way around the whole problem. However, it has to be admitted that the law in this area is not settled by authority, and a decision to rely on liquidated damages as a limit of liability should be carefully considered – witness the difficulties that liquidated damages have thrown up in the cases mentioned above.

Final observations

While the standard works on contract tend to have little specifically devoted to liquidated damages, this is not the situation in the case of books devoted to construction law. Liquidated damages have been widely used for many years in construction and engineering contracts, and are currently a feature of many of that industry’s current standard forms of contract. There would be a good deal to say to give a complete overview of liquidated damages, but perhaps I should end with a quick summary of the essential points as regards service credits and similar provisions in outsourcing contracts.

  • The Service Level Agreement’s standard approach of providing performance measurements and (relatively) small price adjustments for failure to meet these service levels is almost certainly not a penalty: calculation of loss is almost impossible and so the courts will likely respect the parties’ attempt to legislate in advance for the damages payable.
  • As against this, there is some authority at least for the proposition that if damages are agreed in advance (ie liquidated in the true sense of the term) then the victim of a breach which causes far greater loss than the amount of liquidated damages agreed is probably unable to sue for the whole loss but is restricted to the level of liquidated damages agreed.
  • It appears to be the case that, where the clause is a penalty in the true sense, a victim of a breach can sue for its actual loss which is less than the penalty stipulated in the contract.
  • Liquidated damages are subject to rules of contractual interpretation like any other clause – so the parties have to be precise as to what loss falls within and without the ambit of liquidated damages clauses.
  • It is by no means clear that UCTA will apply to regulate liquidated damages, so use may be made (subject to the caveat above) of liquidated damages clauses to function as effective limitations of liability.

Richard Stephens, Partner, Field Fisher Waterhouse