Strike-out Appeals,Unjust Enrichment,and Discoverability: Insights from Kenya

The interlocutory/final ruling distinction on appeal

The Court of Appeal's holding that Alba ought to have immediately appealed Justice Mwera's refusal to strike out the claim as time-barred, and that the ruling could not be challenged following the trial, is problematic. It engages the age-old quandary of distinguishing interlocutory rulings from final rulings. This distinction is pivotal to the administration of justice. If every judicial ruling were immediately appealable the cogs of justice would stall from cost and delay. ‘Final’ rulings, which are appealable once handed down, must be distinguished from those ‘interlocutory’ rulings that are merely a step towards a final ruling that later will become appealable.

The High Court of Australia affirmed the interlocutory/final ruling distinction in Gerlach v Clifton Bricks Pty Ltd, relying upon a lauded decision of the Privy Council that was also cited in Alba, noting:

It is an established rule of practice, not recent but dating back at least to the early decisions of the Privy Council on appeal from colonial courts, that objections to interlocutory orders will not be lost if not immediately prosecuted by an interlocutory application for leave to appeal. Such complaints could be saved up to be raised in objection to the final judgment in the proceedings between the parties, provided the ground was still relevant to that judgment. ¹¹

Access to justice justifies saving objections to interlocutory orders until the appeal of a final judgment. A recent article analysing applications to appeal interlocutory orders in Canada concluded that ‘access to justice is more likely to be impeded when appeals are heard on interlocutory issues’. ¹² That is because, as the High Court of Australia canvassed:

Interlocutory appeals can often cause great injustice to parties who are less well resourced than those who pursue them. They can be misused by those with ‘a long purse’ to prevent others from securing justice in an early trial of the substantive issues. They can lead to a plethora of appeals and further interlocutory hearings that needlessly raise the costs and delay the conclusion of litigation. ¹³

The need for an interlocutory/final ruling distinction is not controversial. How the line is drawn is. Two landmark judgments of the Court of Appeal of England and Wales – both of which have precedential force to this day – illustrate the quandary. In Salaman v Warner, Lord Esher M.R. and Fry L.J. held that a ruling is final if it brings an end to the case by ‘finally dispos[ing] of the matter in dispute’ or ‘determin[ing] the action’. ¹⁴ A ruling is interlocutory if it allows the action to go on. Lord Justice Lopes in Salaman, though purportedly agreeing with this holding, framed an alternative test, opining that a ruling is final when ‘it would finally determine the rights of the parties’. ¹⁵ The Lopes L.J. test came to be adopted by the Court of Appeal in Bozson v Altrincham: Lord Alverstone C.J. held that a ruling is final if it ‘finally dispose[s] of the rights of the parties’, but it is interlocutory if it does not. ¹⁶ The Bozson inquiry into the effect on parties’ rights leads to a more expansive interpretation of ‘final’ rulings compared to the Salaman inquiry into whether a decision disposed of a case, since ‘almost all orders finally decide a right of the parties – even if it is a right concerning a procedural step’. ¹⁷ As a result, Bozson tends to generate more occasions on which to appeal, whereas Salaman tends towards the opposite.

The slippery interlocutory/final ruling distinction continues to challenge courts around the world. Various compromises between the positions of Salaman and Bozson have been sought through legislation and common law interpretation. The Kenyan Civil Procedure Act (CPA) distinguishes judicial ‘orders’ from judicial ‘decrees’. An ‘order’ is ‘the formal expression of any decision of a court which is not a decree’. ¹⁸ It is interlocutory. A ‘decree’ is ‘the formal expression of an adjudication’ (including a judgment), which ‘conclusively determines the rights of the parties with regard to all or any of the matters in controversy in the suit …’. ¹⁹ This definition reflects the language of Bozson. Pivotal to the distinction is the word ‘conclusively’. A decree can be either ‘preliminary’ or ‘final’. It is preliminary ‘when further proceedings have to be taken before the suit can be completely disposed of’. ²⁰ As the Supreme Court of India described in Shankar v Chandrakant, ‘[a] preliminary decree is one which declares the rights and liabilities of the parties leaving the actual result to be worked out in further proceedings’. ²¹ The correctness of a preliminary decree that is not appealed cannot later be disputed on appeal from a final decree. ²² So the time for appealing a preliminary decree is when it is handed down. Orders, meanwhile, presumptively cannot be appealed. ²³ The time for challenging orders is in the course of appealing a decree.

In Alba, the Court of Appeal held that Justice Mwera's strike-out judgment was a preliminary decree and that Alba was precluded from reopening the issue of limitation on appeal from the final trial judgment. ²⁴ The Court acknowledged comparative authorities holding that a refusal to strike out a claim as time-barred was interlocutory and not immediately appealable. ²⁵ But it found that the point was ‘debatable’ and that the balance of considerations favoured treating Justice Mwera's ruling as conclusive. ²⁶ Regrettably, it seems that the leading Kenyan authority on this issue was overlooked. In Mandavia v Ratta Singh, Law J.A. set out a preferable interpretation of the interlocutory/final ruling distinction:

[W]hen a suit is disposed of on a preliminary point, an appeal will lie from the decree dismissing the suit, and where an issue such as liability is tried as a preliminary issue and finally disposed of at first instance, a preliminary decree arises from which an appeal lies; but where a preliminary issue alleging misjoinder, limitation, lack of jurisdiction, or res judicata fails, no preliminary decree arises from which the unsuccessful party has a right of appeal. ²⁷

This passage gives context to the word ‘conclusively’ in the CPA. A strike-out application only conclusively determines the rights of the parties when it succeeds, because it brings the plaintiff's action to an end. In such circumstances, an immediate right of appeal is warranted. Alba's strike-out application, however, failed. Because it failed Justice Mwera's ruling did not conclusively determine the rights of the parties. Rather, the action continued to trial, at which point the High Court heard and resolved factual disputes concerning whether Alba had committed fraud or illegality, and when Caltex's mistake could have been discovered. Applying Mandavia, the strike-out judgment would have been a preliminary decree had it succeeded. Since it failed, it should have been characterised as an interlocutory order that remained challengeable following trial.

Precedent from New Zealand is instructive. Like the Kenyan CPA, in New Zealand, a judicial decision is understood to be ‘final and not interlocutory if it constitutes a final disposition of the rights of the parties in the proceeding, whether or not there has been consideration of the substantive merits’. ²⁸ A decision granting summary judgment ‘is a final decision because it determines the claim and brings the proceeding to an end’. ²⁹ It generates a right of appeal. Conversely, a decision refusing summary judgment is ‘not a final disposition’ since it does not bring the proceeding to an end. ³⁰ It is interlocutory and not immediately appealable. The parties’ procedural rights thus depend upon the outcome of judicial rulings.

The wisdom of this approach can be seen in considering hypothetically what might have happened had Alba immediately appealed Justice Mwera's ruling in 2002. Alba would have had to argue that Justice Mwera's ruling was a preliminary decree that ought to be overturned. Caltex would have reason to argue that Justice Mwera's ruling, which was favourable to Calex, was merely an interlocutory order and that the Court of Appeal lacked jurisdiction to hear the appeal at that stage. Caltex could cite Mandavia in support. ³¹ Both parties would be compelled to invest considerable time and expense arguing an appeal that might ultimately be ruled premature and remitted to the High Court for trial. A case that already took over 12 years from commencement to trial judgment might become more prolonged, further exacerbating Kenya's substantial court case backlog. ³² It is for this reason that commentators and courts have tended to favour an expansive understanding of interlocutory orders that minimises unnecessary appeals. ³³

A final relevant authority of note is the United Kingdom Supreme Court's second judgment in Test Claimants in the FII Group Litigation v Revenue and Customs Comrs. ³⁴ Litigation of the Test Claimants’ long-running claims for restitution of overpaid taxes had been divided into two primary stages. All throughout the first stage trial and two appeals, as well as the second stage trial, the parties had proceeded on the basis that the question of limitation was governed by Deutsche Morgan, a decision of the House of Lords that was also applied in Alba. Only at the second-stage Court of Appeal hearing did the Revenue indicate its intention to seek leave to challenge the authority of Deutsche Morgan. The Test Claimants complained that to reopen the issue of limitation at such a late stage in the proceedings would cause them ‘enormous prejudice’. ³⁵ The Supreme Court nevertheless found that the equities weighed in favour of hearing the issue. Courts should not lightly dismiss important questions of law where they have not already been aired on appeal.

Breach of contract or unjust enrichment?

Alba was argued and adjudicated as a breach-of-contract case. At the strike-out hearing, it was said there was ‘no dispute’ that the suit was ‘based on a contract of supply and purchase of petroleum products in 1991’. ³⁶ The Court of Appeal described the suit as a dispute over ‘petroleum products sold and supplied to the appellant, Alba Petroleum Limited, on 4th December 1991’. ³⁷ This does not seem right.

The facts as found by Justice Kimaru at trial do not suggest that Alba had contracted to pay for the diesel Caltex supplied to Shell or that Alba had breached an express contract with Caltex in respect of that transaction. Alba had pre-purchased 10,000 MT of diesel stored with Caltex. The parties’ business relationship ‘did not involve the sale of any petroleum products in credit’. ³⁸ The only contract that arose on 4 December 1991 was Caltex's agreement to transfer one-quarter of Alba's stored diesel entitlement to Shell. That contract was satisfied when Caltex executed it. There was no requirement or expectation of payment. Caltex's bookkeeping error did not give rise to or affect any new contract with Alba. What it did was cause Caltex to labour under a mistake. As Justice Kimaru stated, the failure correctly to record the transfer gave Caltex the ‘mistaken belief that it still owed the defendant … a further quantity equivalent to 2,500 MT of [diesel]’. ³⁹ As a result, rather than capping Alba's remaining diesel entitlement at 7500 MT, Caltex ‘allowed the defendant to withdraw petroleum products equivalent to [2,500 MT of diesel] in the mistaken belief that the defendant was at the time still entitled to the said petroleum products’. ⁴⁰ The essence of the dispute was not a prima facie breach of contract over the initial transfer of 2500 MT of diesel on 4 December 1991. It was a prima facie unjust enrichment over the final 2500 MT of diesel withdrawn by Alba after Alba had exhausted its own stored diesel supply. The case concerned not an express contract but an alleged contract implied by law: a quasi-contract imposed by the law of unjust enrichment. This insight has ramifications both for the theory of Caltex's case and the timeliness of its cause of action.

It seems surprising that unjust enrichment was not addressed given Justice Kimaru's findings of fact. The Kenyan Court of Appeal first recognised ‘the doctrine of unjust enrichment’ in 1978 in a judgment that drew upon landmark English, Canadian, and American precedents. ⁴¹ The Court endorsed the canonical passage from Goff and Jones's treatise that:

Most mature systems of law have found it necessary to provide, outside the fields of contract and civil wrongs, for the restoration of benefits on grounds of unjust enrichment. There are many circumstances in which a defendant may find himself in possession of a benefit which, in justice, he should restore to the plaintiff. Obvious examples are where the plaintiff has himself conferred the benefit on the defendant through mistake or compulsion. To allow the defendant to retain such a benefit would result in his being unjustly enriched at the plaintiff's expense, and this, subject to certain defined limits, the law will not allow. ⁴²

The Kenyan courts have since reaffirmed and applied unjust enrichment as part of Kenyan common law on a number of occasions. ⁴³

The facts found in Alba prima facie support a restitutionary claim by Caltex. The object of the claim would not be the initial 2500 MT of diesel transferred to Shell on 4 December 1991, but rather the final 2500 MT of diesel ultimately withdrawn by Alba. The reason is that, following the initial transfer, Alba still had 7500 MT of diesel stored with Caltex. Since it was entitled to withdraw up to that amount such withdrawal could not be unjust. But Alba had no entitlement beyond that amount. The further transfer ⁴⁴ of 2500 MT of diesel enriched Alba at Caltex's expense. It provided Alba diesel that it had not paid for. Allowing Alba to retain the benefit of that enrichment would prima facie be unjust – the transfer being caused not by any lawful entitlement of Alba but by Caltex's mistaken bookkeeping. Aside from limitations, there were no obvious defences. ⁴⁵ This was a straightforward case of unjust enrichment. ⁴⁶

If I am right that the nature of Caltex's claim was unjust enrichment and not, as was contended, breach of contract, is the claim a misfire? Should Alba have prevailed? I think not. It is not entirely surprising that unjust enrichment was overlooked when Caltex initiated proceedings in 1999, given the doctrine was very much in a state of development at that time. Whether Caltex had a valid legal claim did not depend upon Caltex identifying the ‘correct’ legal cause of action. The forms of action have long been abandoned. What matters is whether Caltex sufficiently pleaded a factual situation that would entitle it to obtain from the court a remedy against Alba. ⁴⁷ Plaintiffs plead facts, not law. While the judgments in Alba do not set out Caltex's complaint in full, ⁴⁸ the Court of Appeal's judgment does identify the pleading that ‘[a]s a consequence of the aforesaid … mistake … by means of which the said debt was concealed the plaintiff suffered loss in the sum of Ksh. 14,778,257/70 and therefore the plaintiff seeks relief therefrom by way of this suit….’. ⁴⁹ Arguably, Caltex did sufficiently plead facts that satisfy a claim of restitution for unjust enrichment on the basis of mistake. ⁵⁰ If the pleaded facts support such a claim it makes no difference that Caltex failed to recognise that and misidentified its action as breach of contract. The courts can recognise the correct legal basis for providing relief.

This case can be distinguished from the Privy Council decision in Samsoondar v Capital Insurance Company Ltd. ⁵¹ That case, too, was pleaded and tried as an alleged breach of contract. On appeal, the question arose as to whether the case ought to have proceeded as an alleged unjust enrichment. The Court of Appeal of Trinidad and Tobago was satisfied that – despite the plaintiff not having framed its case as such – there were sufficiently pleaded facts and proven evidence to support a successful unjust enrichment claim. ⁵² The Privy Council on appeal overturned. Emphasising fairness to defendants in knowing the case against them, Lord Burrows on behalf of the Committee opined that ‘[t]he ideal pleading of a statement of case by the claimant should indicate that the claim is for restitution of unjust enrichment and should identify facts that satisfy each of those three elements [of enrichment, at the claimant's expense, which was unjust]’. In Samsoondar, the plaintiff's failure sufficiently to identify a plausible unjust factor in its pleadings doomed its case.

Lord Burrows’ speech in Samsoondar brings to mind the aphorism of the ideal being the enemy of the good. Courts rarely receive ideal pleadings. As Hedley observes, given the practical and budgetary constraints under which many litigants and courts operate justice often merits ‘a quick and authoritative decision’. ⁵³ We should heed Lord Diplock's caution in Letang v Cooper against allowing ‘the old forms of action to rule us from their graves’. ⁵⁴ Judges should not abandon pragmatism in pursuit of idealism.

In the case of Alba, Samsoondar seems distinguishable. Caltex's pleading identified relevant loss and the unjust factor of mistake, as well as a gain to Alba of 2500 MT of diesel. It would have been open to the courts to entertain a cause of action in unjust enrichment, notwithstanding the case having been argued as a breach of contract. The facts indicate that such a pleading would have succeeded. Accordingly, it would seem the courts were right to find Caltex had a valid cause of action (even if not the action they identified). As for the remedy, ceteris paribus, the value of restitution for the mistakenly withdrawn final amount of diesel would seem to be comparable to the value of damages awarded over the initial amount of diesel (given the quantities were the same), although valuation should be informed by the diesel market price at the dates(s) when the further 2500 MT was drawn down and on-sold. The interest calculation would not change since interest was awarded with effect only from the date of filing.

Extending limitation for mistake

Alba contended that Caltex's claim was time-barred by the ordinary six-year limitation period and that the extension in s 26(c) of the KLAA did not apply. Had the claim progressed as one of unjust enrichment there would have been no question as to the applicability of s 26(c): a claim for restitution of a benefit mistakenly conferred is an action ‘for relief from the consequences of a mistake’. It makes no difference whether the mistake was due to a plaintiff's carelessness. ⁵⁵ Furthermore, on an unjust enrichment theory of the case time would commence from when the final 2500 MT of diesel was transferred (rather than from the date of the initial transfer). If that transfer occurred after 29 April 1993 then the claim would be in time under the ordinary limitation period. In this respect, the KLAA is more prudently drafted than comparative in pari materia limitation statutes. The EWLA, for instance, does not mention quasi-contract or unjust enrichment, resulting in courts straining the language of the statute to capture such claims. ⁵⁶ Meanwhile, Singapore's Limitation Act – while in many respects identical to the KLAA (both being progeny of the EWLA template) – similarly omits language explicitly capturing unjust enrichment actions. As a result, Singaporean courts have held no limitation period applies to such actions at all. ⁵⁷ Section 4(1) of the KLAA, by contrast, includes an additional paragraph (e) applying the ordinary six-year limitation period to ‘actions … for which no other period of limitation is provided by this Act or by any other written law’. This simple provision closes a loophole that has generated so much litigation abroad.

By framing its claim in breach of contract Caltex put in issue the scope of s 26(c) of the KLAA. The leading authority on point, Phillips-Higgins v Harper, construed the equivalent English provision narrowly, holding that it ‘applies only where the mistake is an essential ingredient of the cause of action’. ⁵⁸ This interpretation could capture a claim in respect of the final 2500 MT of diesel withdrawn by Alba, on the theory that the withdrawal amounted to ‘a contract entered into in consequence of a mistake’. ⁵⁹ Yet, the object of Caltex's claim seemed to be the initial transfer of 2500 MT of diesel. Alba would seem to have the stronger argument that mistake was not an ‘essential ingredient’ of that transaction. That is because the alleged contract arising therefrom was not entered into by mistake: Caltex knowingly received and executed Alba's telex. At most, it was Alba's failure to pay – and Caltex's failure to demand the same – that was mistaken. But the alleged breach accrued immediately. It precipitated Caltex's mistake, but it was not a consequence of its mistake, and so on the basis of Phillips-Higgins s 26(c) of the KLAA ought not to have applied.

Although brief in their disposal of the issue, it can be inferred that the Kenyan Court of Appeal rejected the reasoning in Phillips-Higgins. ⁶⁰ The Court cited Deutsche Morgan in favour of applying the limitation extension. Though Deutsche Morgan has since been superseded by the United Kingdom Supreme Court's judgments in the FII cases, ⁶¹ there are good arguments for giving s 26(c) of the KLAA a generous reading. Two Law Lords in Deutsche Morgan had challenged the narrow interpretation restricting the provision to cases where mistake was an element of the cause of action. ⁶² Edelman and Zuckerman have argued that Phillips-Higgins was wrongly decided. ⁶³ Alba illustrates why. A party who mistakenly fails to realise they are owed a debt does not sleep on their rights. Time ought not to run until they can reasonably discover their mistake. Joseph Unger once proclaimed that (the equivalent) s 26 of the EWLA ‘ensures the victory of Equity over common law’. ⁶⁴ This sentiment seems to be reflected in the Kenyan Court of Appeal's acquiescence to the notion that ‘[t]he law did not intend to allow what is properly owed under contract to be lost unless there was everything against suing for it…’. ⁶⁵

Reasonable discoverability of mistakes

Though the Kenyan Court of Appeal's finding that s 26(c) applied is justifiable, the Court erred in its interpretation of the provision. Relying upon Deutsche Morgan, the Court held that ‘a limitation period in a claim for mistake starts to run when the mistake is discovered’. ⁶⁶ This is an error because it treats s 26 as a subjective enquiry. It is unfortunate that Alba was decided before Deutsche Morgan's demise. In its landmark departure from Deutsche Morgan, the United Kingdom Supreme Court emphasised that discoverability entails an objective test. ⁶⁷ The pertinent question is when could Caltex ‘with reasonable diligence have discovered’ its mistake?

The objective enquiry ensures that limitation is not postponed ‘for an indefinite period of time’. ⁶⁸ Any interpretation that would have that effect is suspect. ⁶⁹ In the context of Alba, s 26(c) would not forgive Caltex for resting on its laurels. That is because it would not be reasonable for a company such as Caltex never to conduct an internal audit. Rather, we should look to when a company carrying on a similar business to Caltex would reasonably have conducted an audit. ⁷⁰ That is a question of evidence that cannot be answered in this note. Nonetheless, if (it may be speculated) it were found to be reasonable to have an 18-month gap between internal audits then on the facts Caltex's claim would still have been timely.