Peninsula Securities Ltd v Dunnes Stores (Bangor) Ltd

This is another appeal which concerns the doctrine against restraint of trade. If a covenant falls within what I will simply call the doctrine, it is unenforceable against the covenantor unless it is reasonable. Last year, in Egon Zehnder Ltd v Tillman [2019] UKSC 32, [2020] AC 154, the court was required to address aspects of the doctrine. In para 29 it considered what it called the outer reaches of the doctrine, by reference in particular to the decision of the House of Lords in Esso Petroleum Co Ltd v Harper’s Garage (Stourport) Ltd [1968] AC 269. But, as it explained in para 30, there was no need for any closer study of those outer reaches in the light of the facts of that case. The present appeal permits no such escape.

A developer of a shopping centre leases part of it to a well-known retailer. He covenants with the retailer that he will not allow any substantial shop to be built on the rest of the centre in competition with the retailer. In due course he assigns his interest in the centre to a company. The company considers that the centre is ailing and that the covenant is stunting its ability to revive it. In these proceedings brought against the retailer, the company seeks a declaration that the covenant by which it is currently bound engages the doctrine; that it is unreasonable; and that it is therefore unenforceable. To date the courts have addressed only the first question raised by the company’s claim: does the covenant engage the doctrine? Yes, ruled the Court of Appeal in Northern Ireland (Stephens LJ, who delivered the judgment of the court, Sir Ronald Weatherup and Sir Richard McLaughlin) on 9 February 2018, [2018] NICA 7, when proceeding to remit the case to the High Court to consider whether the covenant was reasonable. So it is the retailer which now appeals to this court against that ruling.

Mr Shortall is a property developer. In 1979, in his own name, he bought land in Springtown, Londonderry, which, for planning purposes, had been zoned for retail use. At that time Londonderry was, in his own words, an economic and political wasteland. The site comprised about five and a half acres, defined, for land registry purposes, as Folio 25992 County Londonderry. In 1980 he obtained planning permission to develop the site so as to yield 32,000 square feet of gross retail space.

Mr Shortall sought an “anchor tenant”, a substantial and prestigious retail company which would lease a significant part of the site and whose presence would persuade other retailers to lease other parts of it, thus making the proposed centre as attractive as possible to shoppers. To this end, he approached Dunnes Stores (“Dunnes”), which comprised a group of companies based in Dublin and which operated a number of substantial retail outlets of high repute throughout Ireland. Early in 1980 Mr Ben Dunne met Mr Shortall at the site and expressed reservations about the economic viability of establishing a retail unit there. But at a further meeting in about May 1980 they orally agreed outline terms. These were that Mr Shortall would grant Dunnes a long lease of part of the site in consideration of its payment to him of a premium of £50,000 and a nominal ground rent. But Mr Dunne required Mr Shortall to promise not to cause or permit the establishment on any other part of the site of a unit measuring more than about 3,000 square feet for the sale of food or textiles. Mr Shortall agreed. In his evidence he said:

In November 1980 Mr Shortall and Dunnes signed Heads of Agreement. Dunnes decided that its Belfast company, Dunnes Stores (Bangor) Ltd, which is the appellant in this appeal, should sign the Heads and take the proposed lease. The main terms recorded in the Heads were that Dunnes should take a lease of the part of the site there delineated; that it should bear the cost of building its retail unit there; that Mr Shortall should construct at least six units on the rest of the site; and that Dunnes should contribute one third of the cost of constructing the roads, footpaths and car park on the site. Nothing turns on the omission from the Heads of Mr Shortall’s promise not to establish on the rest of the site any substantial unit in competition with Dunnes.

On 2 February 1981 the proposed lease was duly executed. Attached to it was a map of the land in Folio 25992, on which the area subject to the lease was edged in red. The area was said to comprise just in excess of an acre so the rest of the site will have comprised about four and a half acres. The lease was for 999 years in consideration of a premium of £50,000 (which Dunnes paid) and of an annual ground rent of £100. It was Mr Shortall, by his solicitors, who had proposed a term of that length. Upon the area subject to the lease Dunnes covenanted to erect a retail unit measuring at least 15,000 square feet at ground floor level within two years of the grant of detailed planning permission. As provided in the Heads of Agreement, it also covenanted to contribute one third of the cost of the construction of the common areas, in particular of the car park.

In the lease, as also foreshadowed in the Heads of Agreement, Mr Shortall covenanted to construct at least six shop units in an enclosed mall in a specified location adjoining the area subject to the lease. He also entered into the restrictive covenant which is the subject of these proceedings. The covenant, as I will call it, is in the following terms:

Dunnes duly constructed its store. Acting through one of his companies, Mr Shortall duly constructed the shop units in the mall; and he found tenants for them. He also constructed the car park, to the cost of which Dunnes duly contributed. In October 1982 the shopping centre opened. At first it was a great success.

Peninsula Securities Ltd (“Peninsula”), the respondent to this appeal, is another of Mr Shortall’s companies. Of the 100 issued shares in it, he holds 99 and his wife holds the other. He is also its managing director. It is a property holding company. By transfer registered on 27 April 1983, Mr Shortall assigned to Peninsula his freehold interest in all the land in Folio 25992, thus including not only his reversionary interest in the land leased to Dunnes but also his interest in all the other land in the folio which was subject to the covenant.

The success of the shopping centre at Springtown has declined. The reasons for its decline are disputed and, at any rate at this stage, are irrelevant. Peninsula blames the covenant for causing the decline and for stunting its ability to reverse it. Dunnes disagrees.

In 2010 Peninsula made a reference to the Lands Tribunal pursuant to the Property (Northern Ireland) Order 1978 (SI 1978/459) (“the 1978 Order”). It asked the tribunal to declare under article 4 of the order that the covenant represented an impediment to enjoyment of its land and, in that (so Peninsula said) the impediment was unreasonable, to order under article 5 that it should be modified or extinguished. The suggested modification was to substitute for the reference in the covenant to 3,000 square feet a reference to 55,000 square feet. But Peninsula also included in its reference to the tribunal a claim for a declaration that the covenant was unenforceable at common law as being in unreasonable restraint of trade. In due course Peninsula accepted that the tribunal lacked jurisdiction to determine its common law claim. So instead it made the claim in proceedings which it brought in the Queen’s Bench Division of the High Court of Justice in Northern Ireland. By successive amendments to its statement of claim, it added two further claims to the court proceedings. The first was a claim that the covenant was void under section 2 of the Competition Act 1998. On receipt of Dunnes’ expert evidence in answer, Peninsula withdrew that claim and substituted its claim, already made to the Lands Tribunal, under the 1978 Order, which accordingly had to be recast so as to fall within article 6 of it. But Peninsula there made clear that it relied on its claim under the 1978 Order only in the event of the failure of its common law claim. By its amended Defence, Dunnes disputed both claims and counterclaimed that, in the event of any modification or extinguishment of the covenant under the 1978 Order, it should be awarded compensation. The proceedings in the Lands Tribunal have therefore come to an end.

It was therefore agreed that Peninsula’s claim at common law should be determined in advance of its claim under the 1978 Order. In relation to the common law claim, it seems that it was McBride J herself, at first instance ([2017] NIQB 59), who raised the question whether, in particular in relation to an assignee such as Peninsula, the covenant engaged the doctrine at all. Such was the only question which she proceeded to address and which therefore the Court of Appeal addressed. No court has yet addressed the question whether, if the covenant engages the doctrine, the restraint is unreasonable and is therefore unenforceable by Dunnes against Peninsula.

It follows that, were it to be determined on this appeal either that the covenant did not engage the doctrine even at the time when Mr Shortall entered into it or that, although it initially engaged the doctrine, it ceased to engage it when Peninsula became subject to it, the High Court would proceed to hear the claim and counterclaim under the 1978 Order.

In her judgment dated 25 May 2017, [2017] NIQB 59, McBride J sought faithfully to apply the decision in the Esso case, cited (as there will be no need to repeat) in para 1 above. She correctly took the view that, in the opinion of a majority in the appellate committee, a covenant restrictive of the use of land engaged the doctrine only if the covenantor had, by entry into it, surrendered a pre-existing freedom to use the land as he wished. She reasoned that, from the date in 1979 of his purchase of the land in the folio until the date in 1981 of the lease to Dunnes, Mr Shortall had, subject only to planning permission, been free to build retail units of any size on the land and therefore that he had, by entry into the covenant, surrendered a pre-existing freedom. She therefore concluded that, while the land had been held by Mr Shortall, the covenant had engaged the doctrine. She noted however that Peninsula had become bound by the covenant at the same time as it had begun to hold the land and she therefore concluded that it had not, by subjecting itself to the covenant, surrendered a pre-existing freedom. She therefore held that from then onwards the covenant had no longer engaged the doctrine.

On Peninsula’s appeal, the Court of Appeal, at para 56, agreed with McBride J that Mr Shortall had surrendered a pre-existing freedom and therefore, in accordance with the decision in the Esso case, that, while the land had been held by him, the covenant had engaged the doctrine. But at para 45 it disagreed with her conclusion that, upon the assignment of the land to Peninsula, the covenant had ceased to engage the doctrine. It conceded at para 43 that a literal application of the decision in the Esso case would yield her conclusion. But it questioned why in logic the doctrine should at the point of assignment no longer be engaged. It asked at para 44 whether, and if so why, it would no longer be engaged if, instead of assigning the covenant, the original covenantor died or became bankrupt. The court had reminded itself at para 37 that the doctrine was based on public policy; and, observing at para 45 that public policy was a surer foundation for inquiry into the continuing engagement of the doctrine following the assignment of the covenant to Peninsula, it held that there was no reason of public policy why it should not have continued to be engaged.

In this court our duty is to look more closely at the decision in the Esso case in the light of the questions of logic and public policy on which the Court of Appeal touched. That court applied its questions only to the later part of the history: why, in terms of logic and public policy, should Peninsula not have continued to enjoy the benefit of the doctrine just because it had surrendered no pre-existing freedom? But we are required also to ask: why, in terms of logic and public policy, should Mr Shortall have enjoyed the benefit of the doctrine in the first place just because he had surrendered a pre-existing freedom? So we confront an awkward question: is the surrender of a pre-existing freedom an acceptable criterion for engagement of the doctrine?

At the hearing before us, no doubt fortified by early judicial encouragement, Mr Humphreys QC on behalf of Dunnes presented a preliminary argument. It was founded on the fact that Mr Shortall was and is a property developer and that Peninsula was and is a property holding company. Neither of them was or is a trader. How then, ran the argument, could any restraint on them amount to a restraint of trade? On analysis, however, the argument appears to be too narrow. In para 64 below Lord Carnwath argues persuasively that, notwithstanding its conventional description, the doctrine extends to restraints not only of trade but also, more generally, of business, thus including that of a developer. In any event, however, the covenant does restrain trade because it restrains Peninsula (and still also Mr Shortall under the law of contract) from causing or permitting a trade in specified goods in a retail unit of a specified size on the site. In, for example, The British Motor Trade Association v Gray 1951 SC 586 the Inner House of the Court of Session addressed a covenant which the petitioning association required its trade members to extract from all purchasers of new cars. Following the Second World War there was a shortage of new and nearly new cars; and speculators were operating a black market in them. In an attempt by the trade association to combat it, the requisite covenant obliged ordinary members of the public who purchased a new car not to sell it for the first two years. But ordinary members of the public were not traders in cars. So one question was whether their covenant was in restraint of trade. Lord Russell at p 602 expressed the opinion, with which Lord Keith at p 604 was inclined to agree, that the covenant did restrain the trade of dealers in second-hand cars; but, forming the majority, they proceeded to hold that it was in any event reasonable.

The prelude to the Esso case was the decision of the Court of Appeal in Petrofina (Gt Britain) Ltd v Martin [1966] Ch 146. The owners of a petrol filling station in Chesterfield had entered into a covenant to buy from Petrofina all the petrol to be sold at the station (“a solus agreement”) and to require any purchaser of the site to covenant likewise. They agreed to sell the site to Mr Martin, who proposed to form a company to own it and to operate the filling station. He entered into an identical covenant with Petrofina and began to operate the station. Within less than two months, however, he had begun to sell petrol which instead he had bought from Esso. Once his company was incorporated, he agreed to sell the site to it. Petrofina’s claim for an injunction against Mr Martin and his company failed on the ground that the covenant was in restraint of trade and that Petrofina had failed to establish that it was reasonable. It is not easy to identify within the three judgments a common basis for the court’s conclusion that Mr Martin’s covenant engaged the doctrine. But the law reporter was probably correct to suggest in the headnote that its basis was that the covenant restricted the ability to trade on land in which Petrofina “had no interest by way of mortgage, lease or sale”. That such was the basis of the decision seems to follow in particular from passages in the judgment of Harman LJ at p 177 (where he also explained that a reference to the covenantee’s interest in the case of a sale related only to a situation in which the vendor retained other land which could benefit from the restraint) and at p 178; and in the judgment of Diplock LJ at pp 179 and 187. We should note therefore that, in the case of a restriction on the use of land, the focus of the decision was on the covenantee, namely whether Petrofina retained an interest in the use of the land; and also that the doctrine was held to apply to the covenant even though neither Mr Martin nor his company had enjoyed any pre-existing freedom to trade at the site.

In the Esso case Mr Harper (or possibly Mr and Mrs Harper) owned the site of a filling station in Stourport. At first Mr Harper himself operated the trade in petrol there; and in due course he entered into a solus agreement with Esso. Later he allowed the respondent company, which he and his wife owned, to operate it; and so the company entered into the solus agreement. In 1962 the company entered into a revised agreement with Esso for the supply, at a price to be fixed by Esso, of all petrol to be sold there for 21 years. One term of it, similar to a term of the agreement in the Petrofina case, was that the company should keep the filling station open at all reasonable hours throughout the period of the agreement; the effect of it was that, even if it was trading at a loss, the company had to continue to trade there unless it found a purchaser willing to assume its obligations under the agreement. Mr Harper wanted to sell the site to the company. Esso lent to the company funds which enabled it to make the purchase; and the company granted to Esso a mortgage over the site, in which the terms of the solus agreement were repeated. In 1963 the company bought both the site and the business of a second filling station, which was near Kidderminster. The vendor had entered into a solus agreement with Esso on similar terms. At the time of the company’s purchase the agreement was to subsist for about five further years; and the company agreed to be bound by it for that remaining period.

The company soon repudiated the solus agreements in respect of both filling stations and Esso sought an injunction to require it to abide by them. The trial judge held that the doctrine was not engaged by covenants which restrained the use of land and he granted the injunction. The Court of Appeal, constituted by the three judges who had decided the Petrofina case, allowed the company’s appeal. It reasoned that, apart from the incorporation of the terms of the solus agreement in the mortgage on the Stourport property, which, so it held, should not affect its conclusion, such a result was mandated by its earlier decision. Upon Esso’s appeal the appellate committee of the House of Lords agreed with the Court of Appeal that the company’s covenants engaged the doctrine. But it held that, whereas the restraint for 21 years on the Stourport property had not been shown to be reasonable and was therefore unenforceable, the restraint for about five years on the Kidderminster property had been shown to be reasonable and that, to that extent, the trial judge’s injunction should therefore be restored.

Counsel for Esso submitted to the appellate committee that the trial judge had been correct to rule that restraints on the use of land did not engage the doctrine. It was a powerful argument because counsel were able to point to the long history whereby, in the interests of other land which they retained, vendors of land had required their purchasers to covenant not to use it for a specified purpose, including not to trade there whether in specified respects or indeed at all; and whereby lessors had required lessees to enter into analogous covenants. It was common ground that such covenants did not engage the doctrine. But their argument went further. For they were able to cite a common situation somewhat akin to the solus agreements in issue: it was that of the tied public house, in which a brewery company leased, or occasionally sold, premises to the publican on terms which disabled him from selling any beer there other than beer which it had sold to him itself. It was, again, common ground that the tie of a public house did not engage the doctrine.

It is clear that, because of what it perceived to be the oppressive nature of the restraints in many solus agreements relating to the sale of petrol, including in the agreement relating to the Stourport property, the appellate committee was minded to hold that they engaged the doctrine and therefore that, unless they were reasonable, they were unenforceable. But how was the committee to rationalise their engagement of the doctrine in circumstances in which restrictive covenants on the part of purchasers and lessees of land, and in particular the ties to which publicans bound themselves when leasing or buying pubs, did not engage it? In the Petrofina case the Court of Appeal had focussed upon the position of the covenantee: it was when he retained an interest in the land that the doctrine was not engaged.

In the Esso case, by contrast, the majority of the committee focussed upon the position of the covenantor. Lord Reid said at p 298:

So the criterion favoured by Lord Reid for distinguishing between restraints on land which engage the doctrine and those which do not do so has come to be called the “pre-existing freedom” test. It is clear that this is the test which enjoyed majority support within the committee and so constitutes the basis of its decision. Lord Morris of Borth-y-Gest said at p 309:

In the Esso case what criterion did Lord Pearce favour? In Cleveland Petroleum Co Ltd v Dartstone Ltd [1969] 1 WLR 116, decided less than two years after the decision of the appellate committee, the Court of Appeal of England and Wales, at pp 118-119, regarded Lord Pearce as having subscribed to the pre-existing freedom test, which that court then proceeded loyally to apply. It may be, however, that his subscription to it was less than comprehensive. It is true that he said at p 325:

In the Esso case Lord Wilberforce gave the fifth and final speech. It is clear that he did not subscribe to the pre-existing freedom test. He said at p 331 that the common law had often thrived on ambiguity; that, even if it were possible, it would be mistaken to try to crystallise the rules of the doctrine into neat propositions; and that the doctrine had to be applied to factual situations with a broad and flexible rule of reason. He observed at pp 332-333 that provisions of contracts which reflected the accepted and normal currency of commercial relations had come to fall outside the scope of the doctrine because, moulded under the pressures of negotiation, competition and public opinion, they had assumed a form which satisfied the test of public policy as understood at that time. Then, adverting to restrictive covenants imposed on the sale or lease of land, and in particular to the ties imposed on publicans, he said at p 335:

Since the covenant in issue in this court today was made by a lessor, it is worthwhile to note that, in his analysis of covenants in relation to land which had generally been found acceptable and necessary, Lord Wilberforce referred at pp 334-335 to covenants by lessors, and by vendors in relation to land retained by them, as well as by lessees and purchasers in relation to the land leased or conveyed to them. Indeed, as an example of a lessor’s covenant, he cited Hinde v Gray (1840) 1 Man & G 195. There the defendant leased a brewery in Sheffield to the claimants. The defendant, who also owned and operated a public house in Sheffield called The Punch Bowl, covenanted that he would not sell beer in that pub other than as supplied to him by the claimants; and, when later he granted a lease of the pub, he caused the lessee to covenant likewise. One of the claims made in an action brought by the claimants in the Court of Common Pleas was of a breach of that covenant. That claim was rejected because the claimants had failed to establish that the beer sold in The Punch Bowl had not been supplied by them at least indirectly even if not directly. There was no suggestion that the covenant engaged the doctrine and so was required to be reasonable. There is nothing in the jurisprudence, ancient or modern, to indicate that covenants by lessors, and by vendors in relation to retained land, engage the doctrine by reference to any criterion different from that which applies to covenants by lessees and purchasers.

An intriguing question, irrelevant to the search for legal principle, is why, by reference to the criteria which they favoured, their lordships in the Esso case unanimously held that both of the solus agreements entered into by the company engaged the doctrine. In relation to the Kidderminster property, what was the basis on which the majority considered that the company had enjoyed a pre-existing freedom to trade there? In relation to the Stourport property, what was the basis on which they considered that the company had enjoyed a pre-existing freedom to trade there or, if such was their thinking, that it sufficed that Mr Harper had enjoyed that freedom? Indeed, when Lord Reid observed at p 304 that, while he did not subscribe to all of the Court of Appeal’s reasoning, its decision in the Petrofina case had been correct, what was the basis on which he considered that either Mr Martin or his company had enjoyed a pre-existing freedom to trade at the Chesterfield property? How did Lord Pearce persuade himself that the effect of the solus agreements in favour of Esso had been not to absorb the company’s services but, rather, to sterilise them? And, in the light of the evidence, noted by Lord Hodson at p 315, that, out of 36,000 filling stations in the UK, 35,000 had become subject to solus agreements with oil companies, how did Lord Wilberforce feel able to conclude that, “on balance” (as he said at p 337), the agreements in issue had not become acceptable and necessary as part of the structure of a trading society?

Our task in this court is, however, to analyse the validity in principle of the pre-existing freedom test favoured by the majority of our distinguished predecessors. They had alighted upon a distinction which served their purpose: for its effect was, for example, that a tie accepted by a publican upon entry into a lease remained excluded from the doctrine but that a solus agreement with which the operator of a filling station burdened his premises was brought within it. But was the distinction consonant with the doctrine? Or did it mask an attempt to square a circle?

The trouble is that the majority did not explain why a covenant restrictive of the use of land is more likely to offend public policy when the covenantor enjoyed a pre-existing freedom in relation to its use than when he enjoyed no such freedom. It is an explanation for which we must therefore search. Is there a ground for concluding, for example, that a covenantor’s pre-existing freedom places him in a weaker bargaining position than otherwise or in some other way renders his covenant more deserving of legal intervention?

Less than two years after the decision of the appellate committee in the Esso case, Mr J D Heydon, then a lecturer at Oxford University, wrote a coruscating criticism of it in an article entitled “The Frontiers of The Restraint of Trade Doctrine” (1969) 85 LQR 229. Later the author, who ultimately became a justice of the High Court of Australia, wrote a book entitled “The Restraint of Trade Doctrine”, now in its 4th ed published in 2018, in which he has consistently repeated much of what he said in the article. At p 281 of the article he suggested that the pre-existing freedom test reflected “a distinction based purely on form and not on substance at all”. He developed his suggestion as follows:

Until today neither the appellate committee nor, more recently, this court has had an opportunity to reconsider the committee’s decision in the Esso case. Indeed, apart from in the Cleveland Petroleum case, cited in para 25 above, it has received little attention even in the intermediate appeal courts of the UK. As so often, however, contributions of real value to us in this court are to be found in the judgments of other senior courts in the common law world.

The early decision of the Court of Appeal of New Zealand in Robinson v Golden Chips (Wholesale) Ltd [1971] NZLR 257 and the decision of the Supreme Court of Ireland in Sibra Building Co Ltd v Ladgrove Stores Ltd [1998] 2 IR 589 suggest that the pre-existing freedom test has been adopted in the law of both jurisdictions.

There are two Canadian decisions of great relevance. They even replicate the circumstances of the case before us, namely a covenant by the owner of a shopping centre in favour of a lessee of part of it.

In Russo v Field [1973] SCR 466 the third defendant company was the owner of a shopping centre in Toronto. It leased part of it to the claimants. In consideration of their covenant to conduct business as a hairdresser and beauty salon at those premises, the company covenanted not to permit any other store in the centre to conduct that business. The company then leased adjoining premises to the second defendant, Mrs Field, who had notice of the covenants and who commenced a business (which she ultimately discontinued) of selling wigs. In a judgment of the court delivered by Spence J, the Supreme Court of Canada held that the trial judge had been entitled to find that the sale of wigs had become an integral part of the business of a hairdresser and beauty salon; that he had rightly awarded damages to the claimants against both the company and Mrs Field; and, in that she had discontinued the business only later, he had also rightly enjoined her from continuing it. Spence J addressed the doctrine against restraint of trade at pp 486-487 as follows:

In F W Woolworth Co Ltd v Hudsons Bay Co, Zeller’s Inc and Burnac Leaseholds Ltd (1985) 61 NBR (2d) 403 the developer of a shopping mall in New Brunswick had, in the course of granting a lease of premises in it to Woolworths, covenanted that no other premises in the mall would be used as a junior department store, which meant a low-price value store such as Woolworths itself. Zellers, a junior department store in competition with Woolworths, took an assignment of premises in the mall. The New Brunswick Court of Appeal held that the developer’s covenant should be enforced by injunction against both it and Zellers. Hoyt JA, giving judgment on behalf of the court, quoted at para 31 from the speech of Lord Wilberforce in the Esso case and at para 35 the passage set out above in the judgment of Spence J in the Russo case. Hoyt JA continued as follows:

Five Australian authorities will help us; and we should address them in chronological order.

The first is the decision of the High Court in Amoco Australia Pty Ltd v Rocca Bros Motor Engineering Co Pty Ltd (1973) 133 CLR 288. By a majority the court held that a solus agreement entered into by the owner of a proposed filling station in favour of Amoco engaged the doctrine and was unreasonable and so unenforceable. When entering into the agreement, the owner had leased the property to Amoco and had taken back an underlease of it. The majority rejected Amoco’s contention that the owner had therefore enjoyed no pre-existing freedom to trade. In concluding that the covenant engaged the doctrine the majority therefore applied Lord Reid’s test in the Esso case. Nevertheless there were murmurs of doubt about it. Walsh J at p 304 expressed reluctance to accept that it provided a valid criterion for excluding covenants from engagement with the doctrine; and Gibbs J at p 313 expressly left that question open.

The second is the decision of the High Court in Quadramain Pty Ltd v Sevastapol Investments Pty Ltd [1975-1976] 133 CLR 390. X owned adjacent parcels of land in New South Wales. On the first parcel it operated a hotel. It sold the second parcel to Y for use as part of a shopping centre. Y covenanted on behalf of itself and its successors not to apply for a liquor licence there. X assigned the first parcel to Quadramain, which continued to operate the hotel. Sevastapol became the lessee of the second parcel and it applied for a liquor licence there. By a majority the court held that Y’s covenant did not engage the doctrine and should be enforced against Sevastapol. Y had surrendered no pre-existing freedom to use the second parcel; and the majority was willing to reach its conclusion by reference to Lord Reid’s test. But there were further murmurs of discontent about it, louder than in the Amoco case. Gibbs J, with whom Stephen and Mason JJ agreed, observed at p 401 that Lord Wilberforce’s test was more flexible than the pre-existing freedom test and might in time come to be preferred; and, in a dissenting judgment with which Murphy J agreed, Jacobs J suggested at p 414 that the distinction which formed the basis of the pre-existing freedom test presented difficulties unmatched in Lord Wilberforce’s test.

The third is the decision of the Full Federal Court of the Australian Capital Territory in Australian Capital Territory v Munday [2000] FCA 653. Mr Munday traded in articles of waste. The public authority which operated a waste disposal tip changed the contractual terms of his admission to the tip so as to rescind his licence to solicit members of the public to give articles to him before they abandoned them there. The court rejected his claim that the rescission was in restraint of trade and unenforceable. Application of the pre-existing freedom test might well have yielded a conclusion that the doctrine was engaged. But, in a careful judgment with which the other members of the court agreed, Heerey J, after addressing the Amoco and Quadramain cases and also the Woolworth case in Canada, concluded at para 105 that the trading society test should be adopted; and that it yielded a conclusion that the term which prohibited Mr Munday from soliciting for articles did not engage the doctrine.

The fourth is the decision of the High Court in Peters (WA) Ltd v Petersville Ltd (2001) 205 CLR 126. The respondents, which manufactured ice-cream across Australia, sold their business in Western Australia to the appellant. They covenanted not to sell any ice-cream in Western Australia which they had manufactured. The court upheld a ruling that the covenant was in restraint of trade and unenforceable. Although the covenant did not relate to the use of land, the decision is interesting. For the appellant argued that the covenant absorbed, rather than sterilised, the respondents’ capacity to service the market for ice-cream and that, by application of Lord Pearce’s test in the Esso case, it therefore failed to engage the doctrine. In a joint judgment Gleeson CJ and Gummow, Kirby and Hayne JJ held at para 35 that Lord Pearce’s test involved “the application of criteria of particular indeterminacy” and at para 39 that it “should not be accepted in Australian common law”. In passing the judges had also, at para 22, noted criticisms of the pre-existing freedom test, including in Treitel on “The Law of Contract”, 10th ed (1999), p 434. Indeed four months later, in Maggbury Pty Ltd v Hafele Australia Pty Ltd (2001) 210 CLR 181, three of those four judges, in the course of holding that a confidentiality agreement had been in restraint of trade, suggested at para 55 that the court in the Peters case had gone so far as to reject the pre-existing freedom test.

The fifth is the decision of the Supreme Court of Victoria, Court of Appeal, in Specialist Diagnostic Services Pty Ltd v Healthscope Ltd (2012) 41 VR 1. The case is of particular interest because its subject was a covenant by a lessor of part of its premises in relation, among other things, to the use of other parts of them. The appellant conducted a pathology business. It took a lease of premises within two hospitals owned by the predecessors of the first respondent and within a third hospital owned by the second respondent. In granting the leases the owners of the hospitals covenanted (a) not to be concerned in a business similar to that which the appellant was to conduct in the hospitals and (b) not to grant any right to occupy any other part of the hospitals to any third party for the conduct of such business. The respondents breached the covenants. In upholding the appeal the court found that the covenants were reasonable and should be enforced against the respondents. But, perhaps unnecessarily in the light of that finding, the court proceeded to consider whether the covenants engaged the doctrine. By reference to the Australian jurisprudence it resolved to assume that the issue fell to be decided by application of the trading society test. It held as follows:

The passage of half a century since the appellate committee in the Esso case established the pre-existing freedom test has not generated a reasoned defence of it. Mr Heydon’s early criticisms of it remain unrebutted. The commentary on it in the 10th ed (1999) of Treitel on “The Law of Contract”, quoted in the judgment of the majority in the Peters case cited at para 41 above, is replicated, almost word for word, in the 15th ed (2020) of the book at para 11-151, which reads as follows:

In terms of public policy, which is the foundation of the doctrine, there is no explanation why a restraint should engage the doctrine if the covenantor enjoyed a pre-existing freedom but why an identical restraint should not engage it if he did not do so. Surely our conclusion, respectful to our predecessors yet also firm, has to be that the test does not deserve its place in the doctrine.

But is the trading society test any more defensible? At first sight it appears unattractive. It seems to concede that the law follows where many might expect it to lead. Is the law (one might ask) to be determined as if by a weathercock which answers only to the direction of the wind? But such criticisms fail to recognise the nature of the common law. It is a law built by the judges on behalf of the people over seven centuries. It has been generated from below, not imposed from above. Over time bits have been added here, discarded there; enlarged here, confined there; strengthened here, diluted there. Bits have been re-interpreted; bits have withered away as a result of disuse; and bits have been abrogated by statute. In Kleinwort Benson Ltd v Lincoln City Council [1999] 2 AC 349 Lord Goff of Chieveley said at p 377:

Under the trading society test a covenant which restrains the use of land does not engage the doctrine if, in the words of Lord Wilberforce in the Esso case at p 333, it is of a type which has “passed into the accepted and normal currency of commercial or contractual or conveyancing relations” and which may therefore be taken to have “assumed a form which satisfies the test of public policy”. But the proper rooting of Lord Wilberforce’s test in public policy itself generates a need to qualify it. In giving the judgment of the Judicial Committee of the Privy Council in Vancouver Malt and Sake Brewing Co v Vancouver Breweries Ltd [1934] AC 181 Lord Macmillan observed at p 189:

I conclude that, unlike the pre-existing freedom test, the trading society test is consonant with the doctrine.

This conclusion places this court in an acutely uncomfortable position.

In 1966 the appellate committee recognised a facility for it to depart from one of its previous decisions: Practice Statement (Judicial Precedent) [1966] 1 WLR 1234. This court has inherited the facility to do so. Nevertheless in the Practice Statement Lord Gardiner, the Lord Chancellor, stressed the importance of certainty in the law. A sudden change in the law is likely to destabilise it. Negotiations for contractual restraints on the use of land may well have been conducted with the pre-existing freedom test in mind. Past litigants, actual or potential, whose contentions failed or would have failed by virtue of that test would rightly resent a departure from it which would have given them saliency. Future such litigants, whose contentions would fail by virtue of departure from it, would resent it in equal measure. Subsequent opinions of the appellate committee stressed the high degree of caution with which it should address a request for departure. In Horton v Sadler [2007] 1 AC 307 Lord Bingham of Cornhill said at para 29:

It is therefore with appropriate hesitation that I propose that this court should depart from the test favoured by the majority in the Esso case. To adapt Lord Bingham’s words, the objection to it is not just that the issue in the Esso case should have been resolved differently or the principle formulated differently there. Apart from the fact that even at the time Lord Wilberforce chose not to associate himself with it, the objections to the test are that it has no principled place within the doctrine; that it has been consistently criticised for over 50 years and, although in some quarters loyally applied, the reasoning behind it has, to the best of my knowledge, scarcely been defended; and that the common law has been limping between the continuing authority of the test in our jurisdiction and its rejection in Australia and in parts of Canada.

The application of the trading society test to the facts of the present case is straightforward; there is no need to remit the case for inquiry into it. At para 39 of her judgment McBride J addressed the evidence of Mr Crothers, the chartered surveyor who gave evidence on behalf of Dunnes. She said:

It will be recalled that an interesting question caused division between McBride J and the Court of Appeal. It was whether, if a covenant were to engage the doctrine because the covenantor (Mr Shortall) enjoyed a pre-existing freedom, it would continue to engage it following an assignment of the burden of the covenant to an assignee (Peninsula) which enjoyed no such freedom. It follows that, were my colleagues to agree with this judgment, the question would no longer arise.

I propose that Dunnes’ appeal should be allowed and that Peninsula’s common law claim should be dismissed.

The possible availability to Peninsula of an alternative remedy in relation to the covenant has played no part in the conclusion that it has at no time engaged the doctrine. But the conclusion is fortified by the possibility of relief pursuant to Peninsula’s alternative claim under the 1978 Order, which should now proceed to be heard. Nothing in this judgment should be taken to influence the determination of any issue which will then arise.

The 1978 Order is loosely based on section 84 of the Law of Property Act 1925 which, by section 209(3), extends only to England and Wales. Section 84 is entitled “Power to discharge or modify restrictive covenants affecting land”. In its report entitled “Making Land Work: Easements, Covenants and Profits À Prendre” (2011) (Law Com No 327) the Law Commission of England and Wales explained the background to section 84 as follows:

The 1978 Order confers a wider, more flexible, jurisdiction than that conferred by section 84 even as amended. If Peninsula were to establish that the covenant represents an impediment to the enjoyment of land under article 3 and that the impediment was unreasonable under article 5(1), the court under article 6(2)(a), like the Lands Tribunal under article 5(1), would have a wide discretion whether to make an order modifying or extinguishing the impediment and, if so, whether under article 5(6) to substitute a different impediment and/or to award compensation to Dunnes. Article 5(5) requires the discretion whether to make the order to be exercised by reference to seven specified factors and to any other material circumstances.

It would be absurd to consider that the doctrine against restraint of trade would have represented a vehicle for the resolution of the issues between Peninsula and Dunnes as satisfactory as that represented by the 1978 Order. It is this Order which properly reflects modern public policy in relation to the covenants to which it applies.

I begin by paying tribute to Lord Wilson’s characteristically compelling judgment, tinged with regret that this is likely to be his last substantive contribution to the jurisprudence of this court. It has been a privilege to work with him. I adopt with gratitude his clear and concise exposition of the legal and factual background, and of the relevant authorities. This enables me to express my own views relatively briefly. I do so in recognition of the importance of the case, and out of respect for the Court of Appeal whose decision we will be reversing.

I agree entirely with his analysis of the Esso Petroleum Company Ltd v Harper’s Garage (Stourport) Ltd [1968] AC 269 case, and that, for the reasons he gives, we should finally discard the much-criticised “pre-existing freedom test”. I also agree with his reasons for considering that this departure is within the scope and spirit of the 1966 Practice Statement. I note that many of the later criticisms had been foreshadowed at the time in the powerful reply of Robert Megarry QC before the House (pp 288-289). He had observed that the test now put forward by the respondents was “wholly novel”:

I also agree with Lord Wilson in preferring Lord Wilberforce’s so-called “trading society” approach: whether the restrictive covenants in question -

Lord Wilson (para 51) regards the application of Lord Wilberforce’s trading society test in the present case as “straightforward”, having regard to the unchallenged evidence of their not infrequent use in leases in favour of anchor tenants. On the other hand the parallels discussed by Lord Wilberforce might appear to suggest that he had in mind the need for a longer historical pedigree than that implied by the limited evidence in this case. However, as the passage cited above makes clear, the practice had to be looked at in the context of what was “essentially a trade agreement between traders”, rather than a transaction in property. Less important than history is whether, in the light of established practice, there is in the relevant context any public policy reason for interfering in the free process of negotiation between the parties, or seeking to redress the balance of interests between them. The doctrine is an exception to the ordinary principles of freedom of contract, and should not be extended without good justification beyond those categories already established by the case law, or indistinguishable in principle from them.

That approach is also consistent with the underlying approach of the majority in the Esso case. The case itself establishes no more than that a restraint may attract the doctrine even if it relates to the use of land. According to the majority in Esso, if the trader is giving up an existing freedom to trade, it matters not whether the covenant is a purely personal restraint or a restraint on the use of a particular piece of land. As Lord Reid explained (at pp 297G-298A), dismissing the argument that the respondents were left free to trade anywhere else:

The present case is quite different from Esso, or any of the other trading cases. The agreement is not in essence an agreement between traders, but a transaction in land. The only trade which might be inhibited is that of a potential future occupier, seeking to trade in textiles, provisions or groceries in some other part of the development. None of the authorities suggest that there is any public policy reason or legal basis for protecting that mere possibility.

I accept, as Lord Wilson says (paras 17-18), that the mere fact that the Peninsula is a developer, rather than a trader in the conventional sense, is not necessarily determinative. Common formulations of the doctrine refer to any restraints on “the free exercise of (a person’s) trade or business” (see eg Petrofina (Great Britain) Ltd v Martin [1966] Ch 146, 169C per Lord Denning MR). So it is necessary to focus on the nature of any restriction on Peninsula’s own business as a developer.

We have been referred to no case in which the doctrine has been held to apply to a restriction accepted by a developer as part of a development scheme such as the present. This is not surprising. The business of developing a shopping centre as in this case inevitably involves doing deals to regulate the use of the relevant land, and balance the competing interests, to advance the success of the centre as a whole. That was rightly recognised in the Canadian cases to which Lord Wilson has referred (paras 35-36). As is shown by those cases, along with the evidence in this case, there is nothing unusual in special terms being required to secure an appropriate anchor tenant. Indeed, if Dunnes had had reason to think that the covenant would prove unenforceable in law, the likely result would have been its withdrawal from the development, and the failure of the scheme. Thus, the ability to offer such terms does not restrict, but rather facilitates, the developer’s business. It can be seen, in Lord Wilberforce’s words, as “encouraging or strengthening …, rather than as limiting” that business. There is no public policy reason for interfering with such an arrangement.

There is a parallel between Mr Shortall’s position and that of the prospective lessee faced with choice of taking premises subject to a covenant, as Lord Morris explained it in the Esso case. He said at p 309B-C:

Like that hypothetical lessee, Mr Shortall was faced with a free but limited choice: to take Dunnes on the terms offered, or not to have an anchor tenant at all. As in the case of the lessee, in no way could it be said that the exercise of that choice interfered with the free exercise of his business as a developer or with his individual liberty of action. It was rather an intrinsic part of that business. In Lord Morris’ words:

For these reasons, I agree with Lord Wilson that the appeal should be allowed, and Peninsula’s common law claim dismissed. I do so the more readily having regard to the existence in the 1978 Order, of an alternative, and in many ways more satisfactory, vehicle for the resolution of the issues between the parties.