Of Taxes and Stabilisation

How can an investor take the risk of concluding a long-term contract with a state counterparty that has the power at any time to change the legal regime applicable to its contract? And how can a state party reassure an investor counterparty sufficiently to take such a risk? The answer to both questions, in a word, is ‘stabilisation’. Clauses that ensure stability are, as a consequence, of great value to all participants (investors and state parties alike) in long-term upstream contracts. And nowhere do stabilisation clauses play a more important role than in the field of taxation: the area of government lawmaking power in which a state can most easily change the fiscal regime of a contract that it (or a state entity) has signed.

The ultimate balance struck between the investor and the host state (or state entity) on issues of stabilisation and taxation in a particular contract will, of course, depend on the identity of the participants, the nature of the project, and the particular legal, economic and political history of the relevant state. But certain trends have emerged, both in the terms of stabilisation clauses that investors and state parties are including in their contracts, and the way in which such clauses have been interpreted and applied by international tribunals over the years. This chapter identifies such trends, with a particular focus on stabilisation clauses specifically in the context of taxation.[2]

First, we define our terms: the phrase ‘stabilisation clause’ is used to cover a broad spectrum of contractual language of varying nature and effect, and the different subspecies deserve to be distinguished. We then briefly summarise the key historical jurisprudence on stabilisation clauses from the 1930s through to the 1990s, much of which involves the effect of a stabilisation clause in the event of a full expropriation. We then turn to more recent jurisprudence dealing with value-eroding tax measures, particularly Duke Energy v. Peru.[3] And we conclude by drawing together the strands and making some general observations about the law and practice of stabilisation clauses today.

Before we begin, a word on scope: this chapter addresses stabilisation clauses and taxation in the context of international commercial arbitration grounded in contracts. We do not consider the undoubted intersection of contract and investment treaty claims so far as they relate to changes in law, through umbrella clauses, fair and equitable treatment clauses or otherwise.[4]

Definitions: the spectrum of stabilisation clauses

As the name suggests, a stabilisation clause aims to ‘stabilise’, and thus to a greater or lesser extent fix, the fact or effect of the legal and regulatory regime applicable to the subject matter of the contract at the time of contracting.[5] From this general point of departure, stabilisation clauses can be organised into four different categories.

Freezing clauses, stabilisation clauses stricto sensu – classic stabilisation clauses

As defined in the seminal case of Amoco v. Iran, ‘freezing clauses’ freeze ‘the provisions of a national system of law chosen as the law of the contract as to the date of the contract in order to prevent the application to the contract of any future alterations of this system’.[6] In other words, such clauses incorporate into the agreement, as the applicable law, the law of the host state as it stands at a specific time, such as the law valid when the contract entered into force.[7] There are varying degrees of freezing clauses, with some, for example, limiting stability to a key fiscal element such as royalty or income tax.[8]

Freezing clauses were common in the early wave of contracts negotiated in the 1970s and 1980s, but are less common now given the extent to which they are seen as impeding a state’s sovereign ability to develop its own law. That is not to say that such clauses are extinct; partial forms of freezing clauses are still seen today in contracts awarded by host states, including Angola, Cambodia, Guyana, Iraq, Kazakhstan (with exceptions), Malta, Poland and Tunisia.[9]

A typical example of a freezing clause can be found in the 1989 Tunisian Model Production Sharing Contract:

The Contractor shall be subject to the provisions of this Contract as well as to all laws and regulations duly enacted by the Granting Authority and which are not incompatible or conflicting with the Convention and/or this Agreement. It is also agreed that no new regulations, modifications or interpretation which could be conflicting or incompatible with the provisions of this Agreement and/or the Convention shall be applicable.[10]

An example of a freezing clause limited to taxation appears in a Liberian minerals development agreement:

[T]he CONCESSIONAIRE and its Associates shall be subject to taxation under the provisions of the Minerals and Mining Law and the Code and all regulations, orders and decrees promulgated thereunder, all interpretations (written or oral) thereof and all methods of implementation and administration thereof by any agency or instrumentality of the GOVERNMENT (the Code and all such regulations, interpretations and methods of implementation collectively, the ‘Tax Corpus’), in each case as in effect as of the date of this Agreement . . .

For the avoidance of doubt, any amendments, additions, revisions, modifications or other changes to the Tax Corpus made after the Amendment Effective Date shall not be applicable to the CONCESSIONAIRE. Furthermore, any future amendment, additions, revisions, modifications or other changes to any Law (other than the Tax Corpus) applicable to the CONCESSIONAIRE or the Operations that would have the effect of imposing an additional or higher tax, duty, custom, royalty or similar charge on the CONCESSIONAIRE will not apply to the CONCESSIONAIRE to the extent it would require the CONCESSIONAIRE to pay such additional tax, duty, royalty or charge.[11]

Extending the stabilised regime expressly also to encompass ‘interpretations’ is particularly interesting, and is considered further in the discussion of the decision in Duke Energy v. Peru below.

Intangibility and inviolability clauses – prohibition on unilateral changes

Intangibility and inviolability clauses prohibit the unilateral modification of the parties’ contract, and only permit changes by agreement of all parties. In this way, changes in the law of the host state that might have the effect of changing the terms of the parties’ contract will not apply to that contract.[12] This is a subcategory of freezing clauses, but instead of ‘freezing’ the law, they focus their freeze on the contract.[13] As Professor Cameron notes, the advantage of this approach is ‘that it establishes a procedural mechanism for discussion (and probably negotiation) between the parties about the future of the agreement’.[14] An example of such a stabilisation clause is in the concession agreement at issue in the famous arbitration Texaco v. Libya,[15] discussed in detail in the next section.

The Government of Libya will take all steps necessary to ensure that the Company enjoys all the rights conferred by this Concession. The contractual rights expressly created by this concession shall not be altered except by mutual consent of the parties.
This Concession shall throughout the period of its validity be construed in accordance with the Petroleum Law and the Regulations in force on the date of execution. . . Any amendment to or repeal of such Regulations shall not affect the contractual rights of the Company without its consent.[16]

Economic equilibrium clauses

Economic equilibrium clauses, also known as balancing, rebalancing of benefits or adaptation clauses, do not freeze the legal regime applicable to a contract. Rather, they attempt to deal with the consequences of change by providing for the negotiation of amendments to the contract to reinstate the initial economic balance of the contract.[17] Such clauses vary widely, both in terms of defining the change that triggers the protection (a formal change in law, or a broader change in the application of existing law), and the nature and extent of protection offered (monetary compensation or compensatory revision of the underlying contract).

An example of such a stabilisation clause can be found in the 1997 Model Production Sharing Agreement for Petroleum Exploration and Production in Turkmenistan:

Where present or future laws or regulations of Turkmenistan or any requirements imposed on Contractor or its subcontractors by any Turkmen authorities contain any provisions not expressly provided for under this Agreement and the implementation of which adversely affects Contractor’s net economic benefits hereunder, the Parties shall introduce the necessary amendments to this Agreement to ensure that Contractor obtains the economic results anticipated under the terms and conditions of this Agreement.[18]

Similarly, the tax stabilisation clause at issue in Burlington v. Ecuador (a treaty claim, and thus publicly available), reads as follows:

Modification to the tax system and to the employment contribution: In the event of a modification to the tax system, the employment contribution or its interpretation, which have an impact on the economics of this Contract, a correction factor will be included in the production sharing percentages to absorb the increase or decrease in the tax. This adjustment will be approved by the Administrative Board on the basis of a study that the Contractor will present to that effect . . .

Contract amendments: There shall be negotiation and execution of contract amendments, with prior agreement of the Parties, particularly in the following cases: . . . (c) When the tax system . . . applicable to this type of Contract in the country is modified, in order to restore the economy of the Contract.[19]

In drafting such clauses, the parties should consider defining: the change of circumstances triggering the clause; the effect of the change on the contract; the objective of and procedure for the renegotiation; and the solution in cases of failure of the renegotiation process.[20]

Allocation of burden clauses

A variation on economic equilibrium clauses are ‘allocation of burden’ clauses. These provide that, rather than requiring the parties to negotiate revisions to the terms of the contract in order to restore the economic equilibrium, a state entity (or the state itself) will indemnify the foreign investor for any loss or damage resulting from a change in legislation.[21] An example of this is in the Kurdistan Regional Government Model PSA:

[T]he GOVERNMENT shall indemnify each CONTRACTOR Entity upon demand against any liability to pay any Taxes assessed or imposed upon such entity which relate to any of the exemptions granted by the GOVERNMENT under this Article 31.1, and under Articles 31.4 to 31.11 [exempting the investor from certain taxes].[22]


These four types of stabilisation clauses are neither exclusive nor mutually exclusive, and may be combined in different forms in one agreement. A 2008 study found that stabilisation clauses freezing all laws were only common in contracts relating to projects in, or investors from, East Asia and the Pacific, the Middle East and North Africa, and most of sub-Saharan Africa. Such freezing clauses were less common in projects in countries that are members of the Organisation for Economic Co-operation and Development (OECD),[23] where contracts tended to include different forms of economic equilibrium clauses.[24]

In recent years, some such clauses have also been subject to express carve-outs in relation to laws protecting health, the environment or human rights. One high-profile example is the ‘Human Rights Undertaking’ of the Baku-Tbilisi-Ceyhan pipeline (BTC Pipeline) consortium, regarding the contract’s stabilisation clause:

[The BTC Pipeline consortium] shall not seek compensation under the ‘economic equilibrium’ clause or other similar provisions . . . in such a manner as to preclude any action or inaction by the relevant Host Government that is reasonably required to fulfil the obligations of that Host Government under any international treaty on human rights (including the ECHR), labour or HSE (health, safety, environment) in force in the relevant Project State from time to time to which such Project State is then a party.[25]

Stabilisation clauses: the jurisprudence of international tribunals over time

Stabilisation clauses have a long heritage that stretches back to the early 20th century. At least one tribunal in the 1930s, and subsequently a number of tribunals in the 1970s and 1980s, have had occasion to interpret various stabilisation clauses in long-term energy contracts, although usually in the context of nationalisations as opposed to taxation changes. These key cases are briefly summarised in this section, as a historical backdrop to the modern application of stabilisation clauses.

Lena Goldfields, Ltd v. USSR (1930)[26]

In 1925, the government of the USSR granted Lena Goldfields (Lena) exclusive exploring and mining rights over ‘vast areas’ of the Soviet Union, for 30 to 50 years.[27] The concession agreement contained a simple stabilisation clause: Lena was to submit to all existing and future legislation only ‘in so far as special conditions are not provided in this agreement’ (Article 75). For its part, the government undertook ‘not to make any alteration in the Agreement by Order, Decree, or other unilateral act or at all except with Lena’s consent’ (Article 76). The result of Articles 75 and 76, as understood by the arbitral tribunal (consisting only of Dr Otto Stutzer and Sir Leslie Scott, after the government’s appointee,
Dr S B Chlenov, failed to participate), was ‘completely to protect Lena’s legal position – i.e., to prevent the mutual rights and obligations of the parties under the contract being altered by any act of the Government, legislative, executive, or fiscal, or by any action of local authorities or trade unions’.[28]

Nevertheless, in 1929, the Soviet government famously adopted its ‘Five-Year Plan’, the effective result of which was ‘to deprive the company of available cash resources, to destroy its credit, and generally to paralyse its activities’.[29] The tribunal found this new government policy, and associated legislation, ‘necessarily meant, when measured in terms of contractual obligation, the breach by the Government of many of the fundamental provisions, express and implied, of the Concession Agreement’, and awarded compensation.[30]

Texaco v. Libya (1977)[31]

Between 1955 and 1966, Texaco signed a number of petroleum concession agreements with the government of Libya. The contracts included ‘intangibility clauses’, which provided as follows:

The Government of Libya will take all steps necessary to ensure that the Company enjoys all the rights conferred by this Concession. The contractual rights expressly created by this concession shall not be altered except by mutual consent of the parties. . .
This Concession shall throughout the period of its validity be construed in accordance with the Petroleum Law and the Regulations in force on the date of execution . . . Any amendment to or repeal of such Regulations shall not affect the contractual rights of the Company without its consent.[32]

In the 1970s, Libya nationalised its oil industry, leading Texaco and others to launch arbitration proceedings in search of compensation.

In Texaco v. Libya, relying, inter alia, on the intangibility clause, the sole arbitrator, René-Jean Dupuy, concluded that Libya had breached the concession agreements:

The effect is . . . to ensure to the private contracting party a certain stability which is justified by the considerable investments which it makes in the country concerned. The investor must in particular be protected against legislative uncertainties, that is to say, the risks of the municipal law of the host country being modified, or against any government measures which would lead to an abrogation or rescission of the contract. Hence, the insertion, as in the present case, of so-called stabilization clauses: these clauses tend to remove all or part of the agreement from the internal law and to provide for its correlative submission to sui generis rules. . . or to a system which is properly an international law system . . .
[A] State cannot invoke its sovereignty to disregard commitments freely undertaken through the exercise of this same sovereignty and cannot, through measures belonging to its internal order, make null and void the rights of the contracting party which has performed its various obligations under the contract . . .
Thus, in respect of the international law of contracts, a nationalization cannot prevail over an internationalized contract, containing stabilization clauses, entered into between a State and a foreign private company.[33]

Revere Copper v. OPIC (1978)[34]

In 1967, Revere Copper entered into an agreement with the government of Jamaica in relation to the financing, construction and operation of a bauxite mining plant. The contract contained a stabilisation clause providing for security of investment, and also stated that ‘[n]o further taxes. . . burdens, levies. . . will be imposed .... For the purposes of taxation and royalties the provisions of this Agreement shall remain in force until the expiry of twenty five years’.[35] In 1974, the government introduced a bauxite levy and increased the royalties to be paid by Revere, while also forcing Revere to renegotiate the contract. In 1975, Revere was forced to shut down its plant ‘for economic reasons’.[36] In 1976, Revere unsuccessfully sued the government in the Supreme Court of Jamaica on the basis that the bauxite levy was a breach of the agreement.[37] The Jamaican Supreme Court declared the contractual prohibitions against increased taxes and royalties void ab initio.[38]

Revere then commenced AAA arbitration under a related political risk insurance contract with Overseas Private Investment Corporation (OPIC), alleging that the government had committed ‘expropriatory actions’ as defined in the OPIC contract.[39]

The majority of the AAA tribunal, consisting of G W Haight, Carroll R Wetzel and Francis Bergan (dissenting), noted the ‘contractual prohibitions against increasing taxes and royalties’ and the silence of the contract on the issue of applicable law.[40] In the absence of party agreement on the applicable law, the tribunal accepted the application of:

Jamaican law for all ordinary purposes of the Agreement, but we do not consider that its applicability for some purposes precludes the application of principles of public international law which govern the responsibility of States for injuries to aliens. We regard these principles as particularly applicable where the question is, as here, whether actions taken by a government contrary to and damaging to the economic interests of aliens are in conflict with undertakings and assurances given in good faith to such aliens as an inducement to their making the investments affected by the action.[41]

The majority further found that the international character of the contract arose from the fact that the contract was ‘part of a contemporary international process of economic development, particularly in the less developed countries’.[42] On this basis, the majority upheld the legality and binding nature of the clause, noting that under ‘international law the commitments made in favour of foreign nationals are binding notwithstanding the power of Parliament and other governmental organs under the domestic Constitution to override or nullify such commitments. . . To suggest that for the purposes of obtaining foreign private capital the Government could only issue contracts that were non-binding would be meaningless’.[43]

AGIP v. Congo (1979)[44]

In 1965, AGIP began oil distribution activities in the People’s Republic of the Congo. In 1974, the government nationalised the oil products distribution sector in the Congo, affecting all companies except for AGIP, which 10 days earlier had signed an agreement with the government. This agreement provided for AGIP to sell 50 per cent of its shares to the government, but AGIP would otherwise continue to operate as a limited liability company under private law.[45] Articles 4 and 11 of this contract included stabilisation clauses pursuant to which the government undertook not to apply certain laws and decrees as well as ‘any other subsequent law or decree that aims to alter the Company’s status as a limited liability corporation in private law’, and ‘that if changes are made in the law concerning companies, appropriate measures will be taken to ensure that such changes do not affect the structure and composition of the organs of the Company as provided in the Agreement and in the Company’s statutes, which fix its duration at 99 years’.[46] The contract also provided for ICSID arbitration to resolve disputes.[47] In 1975, Congo nationalised the company.

An ICSID tribunal consisting of Jørgen Trolle, René-Jean Dupuy (the sole arbitrator in Texaco v. Libya) and Fuad Rouhani, considered the compatibility of the nationalisation decree with the contract’s stabilisation clause (providing for ‘stabilization of the Company’s legal status’[48]), and concluded as follows:

Congo had a contractual relationship with AGIP which under Congolese law obliged it not to alter the company’s status unilaterally . . . It cannot be denied that the measures taken. . . ignored the obligation of the contracting State to perform the contract . . .
The unilaterally-decided dissolution . . . represented a repudiation of these stability clauses, whose applicability results not from the automatic play of the sovereignty of the contracting State but from the common will of the parties expressed at the level of international juridical order.
These stabilization clauses, freely accepted by the Government, do not affect the principle of its sovereign legislative and regulatory powers, since it retains both in relation to those, whether nationals or foreigners, with whom it has not entered into such obligations, and that, in the present case, changes in the legislative and regulatory arrangements stipulated in the agreement simply cannot be invoked against the other contracting party . . .
It is sufficient to focus the examination of the compatibility of the nationalization with international law to the stabilization clauses. It is in fact in regard to such clauses that the principles of international law supplement the rules of Congolese law. The reference to international law is enough to demonstrate the irregular nature, under this law, of the act of nationalization which occurred in this case. Consequently, the Government must compensate AGIP for the damage it suffered from the nationalization . . . [49]

Kuwait v. Aminoil (1982)[50]

In 1948, Aminoil was granted a concession for the exploration and exploitation of petroleum and natural gas in Kuwait for a period of 60 years. Article 17 of the concession agreement contained a stabilisation clause in the following terms:

The Shaikh shall not by general or special legislation or by administrative measures or by any other act whatever annul this Agreement except as provided in article 11 [relating to early termination by the Shaikh due to Aminoil’s actions]. No alteration shall be made in the terms of this Agreement by either the Shaikh or the Company except in the event of the Shaikh and the Company jointly agreeing that it is desirable in the interest of both parties to make certain alterations, deletions or additions to this Agreement.[51]

The concession agreement was supplemented and amended on numerous occasions. In 1974, OPEC countries adopted the ‘Abu Dhabi formula’, which effectively raised taxes on the oil that Aminoil was producing. In 1977, the state nationalised the project and terminated Aminoil’s contract. In 1979, the parties agreed to submit the dispute to arbitration, to resolve various issues relating to alleged payments due from each party to the other.[52] Following long-running arbitral proceedings, a majority of the tribunal rejected Aminoil’s argument that a ‘straightforward and direct reading of [the stabilisation clause] can lead to the conclusion that they prohibit any nationalisation’,[53] holding instead that:

It seems fair to say that what the Parties had in mind in drafting the stabilisation clauses. . . was anything which, by reason of its confiscatory character, might cause serious financial prejudice to the interests of the Company ....
No doubt contractual limitations on the State’s right to nationalise are juridically possible, but what that would involve would be a particularly serious undertaking which would have to be expressly stipulated for, and be within the regulations governing the conclusion of State contracts; and it is to be expected that it should cover only a relatively limited period ....
A limitation on the sovereign rights of the State is all the less to be presumed where the concessionaire is in any event in possession of important guarantees regarding its essential interests in the shape of a legal right to eventual compensation.
Such is the case here, – for if the Tribunal thus holds that it cannot interpret [the stabilisation clause] as absolutely forbidding nationalisation, it is nevertheless the fact that these provisions are far from having lost all their value and efficacity on that account since, by impliedly requiring that nationalisation shall not have any confiscatory character, they re-inforce the necessity for a proper indemnification as a condition of it.[54]

In his separate opinion, Gerald Fitzmaurice QC disagreed with the majority as regards the effect of the stabilisation clause, finding, inter alia, that compensation alone was not what a company seeks but rather that the breach not occur in the first place:

It is an illusion to suppose that monetary compensation alone, even on a generous scale, necessarily removes the confiscatory element from a take-over, whether called nationalisation or something else. It is like paying compensation to a man who has lost his leg. Unfortunately it does not restore the leg. When a Company such as Aminoil procures the insertion in its Concession of a clause like Article 17 [the stabilisation clause], its aim is not to obtain money if the Article is breached, but to guarantee if possible that it is not breached. What the Company wants is to be able to go on operating its Concession for the agreed term, not to be compensated for having to cease doing so against its will . . .
In consequence,. . . [I] conclude that although the nationalisation of Aminoil’s undertaking may otherwise have been perfectly lawful, considered simply in its aspect of being an act of the State, it was nevertheless irreconcilable with the stabilization clauses of a Concession that was still in force at the moment of the take-over.[55]

Amoco v. Iran (1987)[56]

In 1966, Amoco and the Iranian National Petrochemical Company (NPC) entered into a contract forming a joint venture company, Khemco Chemical Company Limited (Khemco) for the building and operating of a plant for the production and marketing of sulphur, natural gas liquids and liquefied petroleum gas derived from natural gas. The contract provided that ‘[m]easures of any nature to annul, amend or modify the provisions of this Agreement shall only be made possible by the mutual consent of NPC and Amoco’.[57] In 1979, NPC effectively terminated Amoco’s involvement in Khemco.

The Iran-US Claims Tribunal, consisting of Michel Virally, Charles N Brower and Parviz Ansari Moin, followed Aminoil in concluding that a stabilisation clause in a contract should only be understood as a renunciation on the part of the host state of its right to expropriate a concession in limited circumstances.[58] However, the Amoco tribunal ultimately concluded that the clauses in this contract were not true stabilisation clauses.[59] In particular, the tribunal found that while the clause bound NPC, it did not bind the Iranian government, which was not a party to the agreement. As such, the clause did not restrict the Iranian government from enacting legislative or regulatory measures. Moreover, even if the clause did bind the Iranian government, the clause did not expressly prohibit nationalisation of the contract.[60]

LETCO v. Liberia (1989)[61]

In 1970, Liberian Eastern Timber Corporation (LETCO) and the government of Liberia entered into a 20-year concession contract for the exploitation of timber reserves in Liberia. The contract included the following stabilisation clause:

Except as otherwise provided in this Agreement, no amendment or repeal of any law or regulation governing this Agreement or any part thereof shall affect the rights and duties of the CONCESSIONAIRE without its consent.[62]

After signing the agreement, in 1970, 1971 and 1977, the government withdrew portions of LETCO’s concession. In 1979, the government requested a renegotiation of the concession and in 1980, unilaterally halved the concession area ‘with immediate effect’.[63] LETCO ultimately suspended its operations because ‘we have no forest to operate’, and commenced ICSID arbitration against the government.[64]

The LETCO tribunal, consisting of Bernardo Cremades, Jorge Gonçalves Pereira and Alan Redfern, observed that stabilisation clauses, such as the one before it, are ‘commonly found in long-term development contracts and. . . [are] meant to avoid the arbitrary actions of the contracting government. This clause must be respected, especially in this type of agreement. Otherwise, the contracting state may easily avoid its contractual obligations by legislation’.[65] The tribunal held that:

By its failure to follow the procedure laid down in the Concession Agreement [in relation to revoking the agreement for cause], as well as by its subsequent actions, the Government of Liberia has acted in plain breach of the terms of the Concession Agreement. Its breach of the Agreement entitles LETCO to the recovery of damages.[66]


Surveying this body of case law as a whole, Professor Cameron has offered the following insight:

The failure of early stabilization clauses to act as a deterrent in the oil nationalizations in the Middle East and North Africa led to the emergence of more pragmatically designed stability mechanisms . . . Yet a striking feature of the jurisprudence and scholarly writing on this subject is the extent to which it has continued to rely upon the generation of awards made following the events of the 1970s and 1980s . . . that is, a series of awards issued before such innovations [in the form of economic equilibrium clauses] were widely used. The reason for this historical focus is not hard to discern: until very recently, the only awards that tested contractual stabilization clauses were the ones that followed the wave of unilateral state actions of that period. The more flexible, ‘modern’ clauses have not yet been the subject of review by arbitral tribunals, even though they constitute one of the principal ways in which energy investors prepared for another wave of anti-investor sentiment among host states . . . [67]

Stabilisation clauses today

Against this historical backdrop, we now turn to an analysis of more recent arbitral interpretations of stabilisation clauses. These remain few and far between and, avoiding comment on cases that the authors have personally been involved in involving stabilisation provisions in Kazakh, Nigerian and East Timorese state contracts, we focus on the important 2008 award in Duke Energy v. Peru.

In 1996, the Republic of Peru entered into a number of legal stability agreements (LSAs) in relation to an electricity generation project, Egenor. The LSAs were concluded as part of a broader push by the government to promote and protect foreign investment in Peru.[68] Under the LSAs, Peru guaranteed legal stability as follows:

By virtue of this Agreement,. . . the STATE guarantees legal stability for DUKE ENERGY INTERNATIONAL, according to the following terms:
Stability of the tax regime with respect to the Income Tax. . . in effect at the time this Agreement was executed, according to which dividends and any other form of distribution of profits, are not taxed . . .
This Legal Stability Agreement shall have an effective term of ten (10) years as from the date of its execution. As a consequence, it may not be amended unilaterally by any of the parties during this period, even in the event that Peruvian law is amended, or if the amendments are more beneficial or detrimental to any of the parties than those set forth in this Agreement.[69]

In 1999, Duke Energy acquired a stake in Egenor, and executed additional foreign investor LSAs relating to the project.[70] At this time, the government’s taxation regime, as interpreted and applied by Peru’s tax authority, SUNAT, granted various benefits to Egenor and its owners.

In 2000, SUNAT initiated a tax audit of Egenor, and in 2001, SUNAT imposed a tax liability on Egenor on the basis of alleged tax underpayments, plus interest and penalties.[71] There were two bases for this tax assessment: (1) a 1996 merger was a sham transaction concluded solely to take improper advantage of tax benefits provided for under the Merger Revaluation Law (merger revaluation assessment); and (2) Egenor should have depreciated the assets transferred to it during privatisation using a special decelerated rate rather than the general statutory rate (depreciation assessment).[72] In 2001, Egenor filed administrative complaints with SUNAT, and these were rejected in 2002. Egenor then appealed to the Tax Court, but these appeals were at least partially rejected.[73]

In 2003, in addition to its local administrative challenges, Duke Energy commenced ICSID arbitration in accordance with the dispute resolution clause of the LSAs.[74] Duke Energy alleged that Peru had breached, inter alia, the guarantee of tax stabilisation in the LSAs, along with various other guarantees.[75]

The LSA did not specify the applicable substantive law, and thus the ICSID tribunal applied Peruvian law together with international law.[76]

The tribunal, comprising Yves Fortier, Guido Tawil and Pedro Nikken, found by majority that Peru was not liable for the depreciation assessment, but was liable for the merger revaluation assessment, because this was in breach of the guarantee of tax stabilisation under the LSA.[77]

The issue before the ICSID tribunal was ‘whether legal stability covers not only the formal text of the laws and regulations that were in place at the time the Egenor LSA was executed, but also their specific interpretation and application at that time’.[78]

The tribunal’s treatment of this interesting question is worth quoting in full:

The Tribunal begins its analysis of this difficult question with the principle that its jurisdiction does not include the power to review the correctness of SUNAT’s decisions and assessments or of the Tax Court’s decisions as a matter of Peruvian tax law . . . [The Tribunal] does not sit as the appellate division of the Tax Court.
The Tribunal’s jurisdiction, under this particular guarantee, is limited to determining whether the relevant decisions or interpretations of SUNAT and/or the Tax Court, be they right or wrong, are consistent with the tax regime stabilized for Claimant in the. . . LSA. The Tribunal’s standard is therefore comparative in nature, rather than absolute. In other words, the Tribunal does not opine on the correctness of the relevant decision or interpretation, but only determines whether such decision of SUNAT or of the Tax Court in the present case represents a change from their respective decisions prior to the entry into force of the. . . LSA.
This comparative exercise is reasonably straightforward for legislation and regulations, where a change is objectively demonstrable. Claimant establishes an actionable change by proving (i) the existence of a pre-existing law or regulation (or absence thereof) at the time the tax stability guarantee was granted, and (ii) a law or regulation passed or issued after the LSA that changed the pre-existing regime.
The exercise is considerably more difficult where the Tribunal must analyze changes in the interpretation or application of a law or regulatory instrument, which could give rise to a finding of breach of the stability guaranteed by the Respondent.[79]

The ICSID tribunal found that in this ‘more difficult’ scenario, the claimant must prove:

(i) a stable interpretation or application at the time the tax stability guarantee was granted, and (ii) a decision or assessment after the LSA that modified that stable interpretation or application.
Thus, if, at the time when the guarantee was granted, the application of the existing rules resulted in a consistent interpretation, such interpretation must be deemed to be incorporated into the guaranteed stability. In a broad sense, stability is the standard by which the legal order prevailing on the date on which the guarantee is granted is perpetuated, including the consistent and stable interpretation in force at the time the LSA is concluded. The Tribunal is convinced that the maintenance of such stable interpretations of the law, existing at the time the LSA was executed, is part of ‘the continuity of the existing rules’.[80]

The tribunal emphasised that the burden was on the claimant to prove both the laws and regulations in force on a given date, and ‘the prevalence of a particular, consistent and stable interpretation’ based on ‘compelling evidence’ such as ‘[c]lear case law and/or well-established practice’ or ‘generally accepted legal doctrine’.[81]

But the Duke Energy tribunal went further. It also held that where there had been ‘arguably insufficient time for the development of stable interpretations or applications of the relevant law and regulations’, a tribunal ‘may depart from a strictly comparative analysis’ and will evaluate the decision or assessment ‘against a reasonableness standard’, based on local law alone.[82] The tribunal emphasised that, in doing so, it would avoid acting as a Peruvian court of appeal, and thus:

[I]n order to preserve the proper balance of fairness between the parties in this arbitration, it must be demonstrated, absent a demonstrable change of law or a change to a stable prior interpretation or application, that the application of the law to DEI Egenor was patently unreasonable or arbitrary.[83]

The tribunal summarised its conclusions as follows:

[T]ax stabilization guarantees that: (a) laws or regulations that form part of the tax regime at the time the LSA is executed will not be amended or modified to the detriment of the investor, (b) a stable interpretation or application that is in place at the time the LSA is executed will not be changed to the detriment of the investor, and (c) even in the absence of (a) and (b), stabilized laws will not be interpreted or applied in a patently unreasonable or arbitrary manner.[84]

Thus, the tribunal found that not only the text of the law but also the interpretation of it was effectively frozen by the stabilisation clause. For his part, Dr Tawil issued a partial dissenting opinion, indicating that he would have gone further and adopted a broader view of the tribunal’s scope to construe Peruvian law and review the correctness (as opposed to the reasonableness) of the decisions and assessments of SUNAT and the Tax Court.[85]

In this way, the tribunal in Duke Energy v. Peru recognised that there are different methods by which a state party can erode the stability of a contract. While it is not an arbitral tribunal’s mission to sit as a tax appeals court, it can take notice of a change in interpretation of a tax law that can have the same effect on an investor as a formal change in legislation. Indeed, the tribunal went further still, and indicated that a clear misapplication of local tax law, in departing from the provisions of the stabilised regime, in and of itself can amount to an actionable change. Such a view has compelling logic: if a state has undertaken to stay faithful to a stabilised legal regime, why should it face the consequences of its actions when formally amending the law, but remain immune when it takes the arguably easier step of simply misapplying its existing law?[86]


Stabilisation clauses have a long and distinguished provenance. Rather than amounting to a forgoing of sovereignty, they have been recognised as an important example of the exercise of sovereignty by states that have a self-interest in offering undertakings that will encourage investors to make long-term commitments to challenging projects in their territory.

As to the form that such stabilisation undertakings take, the modern trend appears in some places to favour economic equilibrium clauses over classic freezing clauses. State parties are less willing today to bind themselves not to develop their own law, than to undertake to compensate investors one way or the other for the consequences of such development.

As to the approach taken by international tribunals to the interpretation and application of such clauses, the first observation to make is trite, but no less important for this: such clauses are applied on their precise terms, which vary in almost all contracts. Stabilisation clauses are not boilerplate clauses, and tribunals rightly approach their interpretation with very careful regard to the precise language agreed by the parties.

The second observation is that, in determining the legality and binding nature of such clauses, international tribunals will look beyond local legal standards. Intended to protect an investor in the event of a change in law, international tribunals rightly look with scepticism at any local law arguments to the effect that such clauses were – or have become – illegal as a matter of local law.

The third and final observation is that, in applying such clauses and determining whether a change has happened, tribunals are less and less likely to allow state participants in the international economy to benefit from undue formalism. Whether the change is effected by a new tax law, or a new interpretation of an existing law, or most simply by a clear misapplication of existing law, state parties who have given stability undertakings should expect international tribunals now to require them to shoulder the contractual consequences.


[1] Constantine Partasides QC is a partner at Three Crowns LLP in London. Lucy Martinez is an independent arbitrator and counsel, and former counsel at Three Crowns LLP in London. The authors wish to thank Sonja Sreckovic, Stephen Bartels and Matteo Angelini for their original research assistance in relation to this chapter.

[2] Generally in relation to stabilisation clauses, see, e.g.: Redfern, Hunter, Blackaby and Partasides, Redfern and Hunter on International Arbitration, 6th edn, 2015, pp. 193-195; R. Doak Bishop et al (eds), Foreign Investment Disputes: Cases, Materials and Commentary, Kluwer, 2014, pp. 213–280; Katja Gehne and Romulo Brillo, Stabilization Clauses in International Investment Law: Beyond Balancing and Fair and Equitable Treatment, Working Paper No. 2013/46, Swiss National Centre of Competence in Research, January 2014 (Gehne & Brillo); Deloitte, Stabilisation Clauses in International Petroleum Contracts: Illusion or Safeguard?, 2014; Rudolf Dolzer and Christoph Schreuer, Principles of International Investment Law, 2nd edn, 2012 (Dolzer & Schreuer), pp. 82–85; Peter D. Cameron, International Energy Investment Law: The Pursuit of Stability, 2010 (Cameron); Andrea Shemberg, Stabilization Clauses and Human Rights, IFC/SRSG Research Paper, 27 May 2009 (Shemberg); Lorenzo Cotula, Regulatory Takings, Stabilization Clauses and Sustainable Development, OECD Global Forum on International Investment 2008 (Cotula); Piero Bernardini, Stabilization and adaptation in oil and gas investments, (2008) Vol. 1, No. 1, Journal of World Energy Law & Business, pp. 98-112 (Bernardini); J. Nna Emeka, Anchoring Stabilization Clauses in International Petroleum Contracts, (2008) Vol. 42, No. 4, The International Lawyer, pp. 1317-1338; Abdullah Al Faruque, Validity and Efficacy of Stabilisation Clauses: Legal Protection vs. Functional Value, (2006) Vol. 23 Issue 4, Journal of International Arbitration, pp. 317-336; Zeyad A. Al Qurashi, Renegotiation of International Petroleum Agreements, (2005) Vol. 22 Issue 4, Journal of International Arbitration, pp. 261-300; R. Doak Bishop, International Arbitration of Petroleum Disputes: The Development of a Lex Petrolea, XXIII Yearbook Comm. Arb’n 1131 (1998), pp. 1158-1160; Thomas W. Walde and George Ndi, Stabilizing International Investment Commitments: International Law Versus Contract Interpretation, 31 Texas Int’l L. J. 215 (1996).

[3] Duke Energy International Peru Investments No. 1, Ltd. v. Republic of Peru, ICSID Case No. ARB/03/28, Award, 18 August 2008 (Duke Energy v. Peru).

[4] On this issue, see, e.g., Occidental Petroleum Corp. & Another v. Republic of Ecuador, ICSID Case No. ARB/06/11, Award, 5 October 2012; Burlington Resources Inc. v. Ecuador, ICSID Case No. ARB/08/5 (Burlington v. Ecuador), Decision on Jurisdiction, 2 June 2010, Decision on Liability, 14 December 2012, and Decision on Reconsideration and Award, 7 February 2017; AES Summit Generation Limited & Another. v. Hungary, ICSID Case No. ARB/07/22, Award, 23 September 2010; Parkerings-Compagniet AS v. Lithuania, ICSID Case No. ARB/05/8, Award, 11 September 2007; LG&E Energy Corp. & Others. v. Argentina, Decision on Liability, 3 October 2006, ICSID Case No. ARB/02/1; CMS Gas Transmission v. Argentina, ICSID Case No. ARB/01/8, Award, 12 May 2005; Link-Trading Joint Stock Company v. Department for Customs Control of the Republic of Moldova, UNCITRAL, Award, 18 April 2002.

[5] See, e.g., Cotula, supra note 2, pp. 5-6; Dolzer & Schreuer, supra note 2, p. 82.

[6] Amoco International Finance Corporation v. The Government of the Islamic Republic of Iran & Ors, 15 Iran-US CTR 189, Case No. 56, Partial Award No. 310-56-3, 14 July 1987 (Amoco v. Iran), p. 239. This case is considered further below.

[7] Dolzer & Schreuer, supra note 2, p. 83; see also Cameron, supra note 2, p. 70 (‘such a clause prohibits the host state from changing its laws, and in a sense ‘handcuffs’ the host state so that it cannot exercise its sovereign rights to change its laws. In this way the investor creates an enclave arrangement for itself.’); Gehne & Brillo, supra note 2, p. 2.

[8] Cameron, supra note 2, p. 70.

[9] Id., pp. 71–72.

[10] Tunisian Model Production Sharing Contract, Article 24.1, quoted in Cameron, supra note 2, p. 71.

[11] Mineral Development Agreement between the Government of the Republic of Liberia and Mittal Steel Holdings NV dated 17 August 2005 and the Amendment thereto dated 28 December 2006, quoted in Cameron, supra note 2, pp. 70-71.

[12] Dolzer & Schreuer, supra note 2, p. 83.

[13] Cameron, supra note 2, p. 74.

[14] Id.

[15] Texaco Overseas Petroleum Company, et al. v. The Government of the Libyan Arab Republic, Award on the Merits, 19 January 1977, 17 Int’l Legal Materials 1 (Texaco v. Libya).

[16] Id., ¶¶ 3, 70. For other examples of intangibility clauses in contracts involving India, Yemen and Mozambique, see Cameron, supra note 2, p. 74.

[17] Cameron, supra note 2, pp. 59, 75; Gehne & Brillo, supra note 2, p. 2.

[18] Model Production Sharing Agreement for Petroleum Exploration and Production in Turkmenistan 1997, Article 16.1, available at faolex.fao.org/docs/texts/tuk81989e.doc (accessed 22 July 2015). For other examples involving Kurdistan, Nigeria, Egypt, Vietnam, Russia, Mozambique and Kazakhstan, see Cameron supra note 2, pp. 75–80.

[19] Burlington v. Ecuador, supra note 4, Decision on Liability, ¶¶ 328-329. This is a treaty claim, as opposed to a contract claim, and thus is not discussed further herein.

[20] See Bernardini, supra note 2, p. 103.

[21] See id., p. 102; Cameron, supra note 2, pp. 80-81.

[22] Available through www.krg.org. For other examples involving Egypt, Algeria and Azerbaijan, see Cameron, supra note 2, pp. 80-81.

[23] Shemberg, supra note 2, p. 19.

[24] Id. See also Dolzer and Schreuer, supra note 2, p. 85 (noting that the recent trend is towards economic equilibrium clauses as opposed to freezing clauses).

[25] Quoted in Gehne & Brillo, supra note 2, pp. 5–6. The Undertaking also excludes more generally claims against host State measures that are based on human rights, health, safety and environmental aspects, provided that domestic regulation is ‘reasonably required by international labor or human rights treaties to which the Host Government is a party’ and that ‘domestic law is no more stringent than the highest of European Union standards . . . ’. Id.

[26] Lena Goldfields, Ltd v. USSR, Award, 3 September 1930, 36 Cornell Law Quarterly 31 (1951) (Lena Goldfields v. USSR). In an interesting historical footnote, Professor Albert Einstein was considered as the presiding arbitrator (‘super-arbitrator’, probably mistranslated from German to Russian to English) in this case. See V.V. Veeder, The Lena Goldfields Arbitration: The Historical Roots of Three Ideas, (1998) 47 International & Comparative Law Quarterly 747, pp. 759 (fn. 35), 774. While some arbitrators, in the authors’ experience, can behave as if they are ‘Einsteins’ in their own right, it must have been interesting for the parties and counsel alike to behold Albert Einstein’s actual name on a list of possible arbitrators.

[27] Lena Goldfields v. USSR, supra note 26, p. 44.

[28] Id., p. 46.

[29] Id., p. 50.

[30] Id., pp. 46-47.

[31] Texaco v. Libya, supra note 15.

[32] Id., ¶¶ 3, 70.

[33] Id., ¶¶ 45, 68, 73.

[34] Revere Copper and Brass, Inc. v. Overseas Private Investment Corporation, AAA, Award, 24 August 1978, 17 ILM 1321 (1978).

[35] Id., pp. 5, 23, 33.

[36] Id., p. 11.

[37] Id., p. 16.

[38] Id., p. 18.

[39] There were four weeks of hearings on liability issues, during which 10 witnesses were heard. Id., p. 16.

[40] Id., pp. 8, 20.

[41] Id., p. 20.

[42] Id., p. 21.

[43] Id., pp. 43-44. The dissenting arbitrator concluded that international law ‘is not the standard, and indeed is quite irrelevant, to the rules of law which should guide this arbitration [against OPIC] to which neither [the local Revere entity] nor Jamaica is a party.’ Id., p. 102.

[44] AGIP Company v. People’s Republic of the Congo, Award, 30 November 1979, 21 ILM 726 (1982).

[45] Id., ¶¶ 17-18.

[46] Id., ¶¶ 18, 69-70.

[47] Id., ¶ 1.

[48] Id., ¶ 48.

[49] Id., ¶¶ 76-77, 85-88.

[50] The Government of the State of Kuwait v. The American Independent Oil Company (AMINOIL), Final Award, 24 March 1982, 21 ILM 976 (1982).

[51] Id., pp. 990-991 and 1020, quoting Article 17 of the contract.

[52] Id., p. 979.

[53] Id., p. 1020. The majority consisted of Professors Paul Reuter and Hamed Sultan.

[54] Id., pp. 1022–1023.

[55] Id., pp. 1052–1053. See also id., p. 1052, fn. 7, discussing the history of stabilisation clauses and their introduction ‘into concessionary contracts, particularly by American Companies in view of their Latin-American experiences, and for the express purpose of ensuring that Concessions would run their full term, except where the case was one for which the Concession itself gave a right of earlier termination’.

[56] Amoco v. Iran, supra note 6.

[57] Id., ¶ 168.

[58] Id., ¶ 179.

[59] Id., ¶¶ 172–173.

[60] Id., ¶ 180.

[61] Liberian Eastern Timber Corporation (LETCO) v. Government of the Republic of Liberia, Award, 31 March 1989, ICSID Reports, 1989 Volume 2, pp. 343-396.

[62] Id., p. 368.

[63] Id., p. 344.

[64] Id.

[65] Id., p. 368.

[66] Id., p. 369. The tribunal noted that the laws of Liberia had not been changed so as to affect the Concession Agreement. Id., p. 368.

[67] Cameron, supra note 2, pp. 59–60 (emphasis in original).

[68] Duke Energy v. Peru, supra note 3, ¶¶ 37–44.

[69] Id., ¶¶ 186–187, quoting Clauses 3 and 5 of the LSA.

[70] Id., ¶¶ 53–54, 65, 72.

[71] Id., ¶ 128.

[72] Id., ¶¶ 129-130.

[73] Id., ¶¶ 132, 137.

[74] Id., ¶ 63.

[75] Id., ¶¶ 138–139.

[76] Id., ¶ 144.

[77] Id., ¶¶ 142, 300.

[78] Id., ¶ 210.

[79] Id., ¶¶ 215–218.

[80] Id., ¶¶ 218–219 (emphasis in original).

[81] Id., ¶ 220.

[82] Id., ¶¶ 222–224.

[83] Id., ¶ 226.

[84] Id., ¶ 227.

[85] Id., Partial Dissenting Opinion of Arbitrator Dr. Guido Santiago Tawil, ¶¶ 1, 8-9, 16, 26. Dr. Nikken also issued a separate partial dissenting opinion on the issue of estoppel.

[86] In December 2008, Peru applied to annul the award, on the bases that, inter alia: (1) the tribunal’s analysis of the tax stabilisation issue was ‘wholly unaccompanied by reasons or explanation’; and (2) the ‘absence of any reference’ to applicable Peruvian law meant that the tribunal’s analysis on this issue had been developed ‘ “from whole cloth”, and not by the application of law’. Duke Energy v. Peru, Decision of the Ad Hoc Committee, 1 March 2011, ¶¶ 61-62. An ad hoc committee (Professor Campbell McLachlan QC, as president, and Judge Dominique Hascher and Judge Peter Tomka) dismissed Peru’s annulment application in its entirety, finding that the part of the award dealing with tax stability had been ‘adequately reasoned’ and was ‘arrived at after consideration of the evidence of Peruvian law’. Id., ¶ 221. Although the tribunal had not cited ‘legal authorities or passages from the expert evidence in support of its reasons’, the committee stressed that the tribunal ‘was, in large measure, engaged in a process of logical reasoning which did not require citation of authority’. Id., ¶¶ 217-218.

Get unlimited access to all Global Arbitration Review content