Smart Flexibility Clauses in International Investment Agreements

A major challenge for investment treaty designers and adjudicators is to separate opportunistic behavior by host states that should be sanctioned under international law from bona fide public policy measures that should not. This article suggests that International Investment Agreements (IIAs) need to be both ‘smarter’ and more ‘flexible’ to better make that distinction. It draws on economic contract theory as a basic framework, and political economy theory for fine-tuning.

Economic Contract Theory

Complete contracts (or treaties) are impossible to draft since it is impossible to foresee and describe adequately the contractual outcome for all possible states of the future.[1] Contract theory distinguishes between three types of uncertainty:

  1. Uncertainty about the future (unforeseeability)
  2. Uncertainty about the actions of other players (asymmetrical information)
  3. Uncertainty about the meaning and scope of contractual provisions (i.e. textual ambiguity and legal indeterminateness)

Information asymmetries between the parties pose the biggest problem: each party has exclusive information about itself, giving rise to potential opportunism. Contract theory analyzes the resulting problems: adverse selection, moral hazard, and verification. These problems arise equally in investment law.  Contract theory also highlights the possibilities of solving them by separating behavior that should be allowed intra-contractually from behavior that should lead to punishment in order for the contract to be ex post efficient.[2] This separation serves as background for efficiently drafting flexibility clauses in IIAs.

Unanticipated contingencies, had they been known ex ante, would have changed the initial content of the contract. Regret occurs when unforeseen circumstances are not provided for in the (incomplete) contract, thus (erroneously) mandating performance, whereas the (ideal) ‘complete contingent contract’ would have excused such performance. Mandating performance or damages for a regulatory measure taken in good faith and without a discriminatory basis as, for example, in CMS v. Argentina,[3] would be ex post inefficient. It’s for that reason that contracts usually try to distinguish between opportunistic behavior taken in bad faith and unforeseen contingencies taken in good faith, and mandate different legal consequences accordingly.[4] Regret contingency is different from opportunism in that the former is ex post welfare-enhancing for both parties, whereas opportunism is only welfare-enhancing for one party. It follows that the former should be permitted in the contract whereas the latter should not. This, however, is not the case in investment law: good faith regulations or good faith reactions to crises have led tribunals to order damage payments.

Political Economy Theory

Political economy can give guidance on instances of opportunistic behavior. It has been successfully applied to trade-policy formation,[5] but not yet to investment law. However, there are similarities: as Gene M. Grossman and Elhanan Helpman note, “(t)he domestic industry has the same incentive to lobby for barriers to investment as it has to lobby for impediments to trade.”[6] Thus, being the result of maximizing political support, instead of maximizing welfare, some regulation will be opportunistic.[7] Their approach allows us to look for typical opportunistic behavior, such as protectionist and discriminatory measures that seek to maximize either the support of the relevant national industry[8] or one’s own finances (i.e., corruption) via expropriation.

Smart Flexibility Clauses

Smart flexibility clauses can be defined as clauses that take into account unforeseen contingencies, in order to distinguish between host state measures that are in the public interest, for which compensation is not required, and opportunistic measures, for which compensation is necessary. One dimension of smart flexibility concerns the textual scope of the clause (i.e., the substantive dimension). Flexibility clauses are to be found on a continuum and range from essential security clauses to exceptions for certain regulatory goals. A second dimension of flexibility is grounded in the strictness of scrutiny of those clauses; in other words, the scope provided to tribunals to review government actions. Again, these are to be found on a continuum and range from self-judging clauses to strict scrutiny by tribunals. Furthermore, there are institutional mechanisms to be considered for a separation of opportunistic and good faith behavior (i.e., the institutional dimension).

With a view to the substantive dimension, clauses can be drafted such that particular measures taken in good faith and non-discriminatory are explicitly excluded from a breach and thus allow for state measures taken with a view to sustainable development goals. They can also include guidance to the tribunal concerning the allowed scope of review, thus delegating more or less judgment to the tribunal.

The Substantive Dimension

Substantively, the concepts of ‘good faith’ and ‘non-discrimination’ are particularly important to smart flexibility clauses. Currently, it is contentious whether a good faith measure by a host state would breach an IIA, especially when fair and equitable treatment or indirect expropriation is discussed. Whereas bad faith would indicate a breach for all tribunals, it is unclear whether good faith spares a host state from violating a treaty. From a flexibility point of view, a measure taken in good faith should be more readily accepted as legitimate by tribunals. In these cases there is a need to isolate special circumstances in which damages would be warranted, such as a bona fide direct expropriation where compensation has not been paid, or an indirect expropriation where a special sacrifice was demanded by the claimant. A good faith measure tends to be the result of evolving factual and legal circumstances (for example, the ratification of an international human rights treaty or an enhanced environmental policy) rather than of opportunistic behavior by the government.  Both the United States and Canadian Model BITs explicitly state that certain measures with a good faith purpose, such as a public welfare objective, and taken in a non-discriminatory fashion, are not compensable.[9] Thus, smart flexibility clauses not only explicitly describe permissible measures, but also exclude non-discriminatory measures. Furthermore, they provide the tribunal with the discretion to award damages in the rare circumstances in which a good faith measure would amount to a treaty breach.

Arguments on non-discriminatory measures are more principled. From an economic point of view, non-discrimination is the gold standard since it guarantees a level playing field for all investors (national and international). Discriminatory measures are therefore supposed to breach an IIA in almost all circumstances,[10] since there is reasonable suspicion that the motivation behind the policy measure is protective. If investment is also, as trade, about competitive opportunities, discriminatory conduct of the state ought to be scrutinized, whereas non-discriminatory conduct should not lead to a finding of a violation of an IIA, except in rare circumstances. Discriminatory conduct in post-establishment settings is, in other words, a strong indicator of opportunistic behavior. However, intent plays an exculpatory role since even discriminatory conduct may be permissible if the intent is benevolent. That could, for example, be the case in times of financial crises with a limited budget to bail out banks: there, discriminatory bail-outs or state guarantees for national banks may be excused. Furthermore, good faith measures which are taken by the government in emergency situations in an attempt to solve the crisis should not be compensable if the same measures are applied to nationals. A strong hint would be a national emergency law that also imposes hardship on national industry and citizens.

The Review Dimension

With respect to scope of review, self-judging clauses provide states with the greatest flexibility, while preventing tribunals from differentiating between opportunistic and welfare enhancing good faith measures. They could therefore easily lead to abuse by self-interested politicians. They also create a legal void for investors and may impede legal security since any review of the host state’s conduct is discarded. However, full review seems inadequate given that oftentimes, particularly in emergency or national security cases, a margin of appreciation should be provided to the government. In addition, the tribunal may lack the expertise to properly assess the government’s response to the crisis. In order to avoid opportunistic behavior, a good faith review suffices. This is what national administrative courts do when they control discretion of the administration: they control for good faith limits. In other words, the legality of the behavior of the government is checked, but if the behavior stays within good faith limits, expediency considerations are not scrutinized. If, for example, the government takes a land zoning measure, courts would scrutinize the legality of the procedure and whether the government has taken into account certain interests prescribed by law, but would not scrutinize the economic efficiency of the decision.

Institutional Mechanisms

Finally, institutional mechanisms can help distinguish between opportunistic behavior and desired behavior by delegating certain questions to expert bodies. For example, the 2012 United States Model BIT contains an institutional innovation concerning procedures for prudential measures related to the financial market. Similar procedures may also be applicable to other constellations. Here, states take back the delegation from the tribunal to solve certain questions – such as emergency measures in financial crises—and delegates them to specialized experts, normally non-political agencies of the state parties. This has the potential to mitigate opportunistic behavior by governments. When both agencies agree that a measure taken by the host state was prudential, a tribunal will likely have difficulties in finding opportunistic behavior. When considering the introduction of such procedures in their IIAs, states should decide whether they: 1) allow them to apply only above a certain threshold of damage claims (cost-benefit analysis); 2) bind tribunals by the views issued by the state parties’ agencies; and 3) include expert consultation as a general principle.


Smart flexibility clauses such as those incorporated in the US and Canadian Model BITs might help solve a fundamental problem of drafting and applying investment law: the distinction between opportunistic (and thus compensable), or good faith (and thus non-compensable) behavior of host states. Drawing on economic theory helps to illuminate the underlying problem structure and may serve as a guide to treaty-makers and tribunals alike. The best option is to include smart flexibility clauses in IIAs but this will take a long time. In the meanwhile, tribunals could draw on economic contract theory as a background for help in distinguishing opportunistic behavior from permissible reactions to unforeseen contingencies.

Author: Anne van Aaken is Professor of Law and Economics, Legal Theory, Public International Law and European Law at the University of St. Gallen. This article is based on a longer paper entitled Smart Flexibility Clauses in International Investment Treaties and Sustainable Development: A Functional View (March 21, 2013). U. of St. Gallen Law & Economics Working Paper No. 2013-18. Available at SSRN: or

[1] Robert E. Scott and Paul B. Stephan, The Limits of Leviathan. Contract Theory and the Enforcement of International Law. Cambridge: Cambridge University Press (2006), p. 76. Jean Tirole, Incomplete Contracts: Where do We Stand?, 67 Econometrica 741 (1994), defines on page 743 an incomplete contract as one that “does not exhaust the contracting possibilities envisioned in the complete contract”.

[2] Contract theorists distinguish between observable and verifiable information. The former can be observed by the two parties but it may still be that the information is not verifiable in the sense that the observing party is unable to establish the fact sufficiently to convince a neutral third party at a reasonable cost. For details, see Scott and Stephan, supra note 2, p. 71-72.

[3] CMS Gas Transmission Company v. Argentina, May 12th, 2005 – ICSID Case No. Arb/01/8.

[4] If damages awarded are higher than expectation damages, “efficient breach” is discouraged and as they may lead to inefficient behavior by the host state. For a discussion of contracts covered by BITs, see Fabrizio Marrella and Irmgard Marboe, “Efficient Breach” and Economic Analysis of International Investment Law, 4 Transnational Dispute Settlement, Online Journal 6 (2007).

[5] Gene M. Grossman and Elhanan Helpman, Protection for Sale, 84 American Economic Review 833 (1994); Gene M. Grossman and Elhanan Helpman, Trade Wars and Trade Talks, 103 Journal of Political Economy 675 (1995).

[6] Gene M. Grossman and Elhanan Helpman, Foreign Investment with Endogenous Protection, in Robert C. Feenstra, Gene M. Grossman and Douglas A. Irwin (eds), The Political Economy of Trade Policy: Papers in Honor of Jagdish Bhagwati. Cambridge, Mass.: MIT Press (1996), p. 199, p. 216. See also Anne van Aaken and Tobias Lehmann, Sustainable Development and International Investment Law: An Harmonious View from Economics, in Roberto Echandi and Pierre Sauvé (eds), Prospects in International Investment Law and Policy, Cambridge: Cambridge University Press 2013, pp. 317.

[7] The assumption in the Grossman/Helpman model is that governments always act so as to maximize political support/probability of reelection: it assumes that signing an international economic treaty was equivalent to maximizing political support. Hence, it is crucial whether signing the treaty would still have maximized political support ex ante if the exclusion of some measure (e.g. protectionism of some specific industry) had been on the table during the treaty negotiations. If not, then it amounts to opportunistic behavior ex post.

[8] UNCTAD World Development Report, 2012UNCTAD, World Investment Report, 2012, p. 101 describes the problem as follows: “Despite the fact that international policy forums at the highest level (e.g. the G-202) frequently make reference to ‘investment protectionism’, there is no universally agreed definition of the term. Different schools of thought take different approaches. Broadly, protectionist measures related to investment would include: (1) measures directed at foreign investors that explicitly or ‘de facto’ discriminate against them (i.e. treating them differently from domestic investors) and that are designed to prevent or discourage them from investing in, or staying in, the country. And (2) measures directed at domestic companies that require them to repatriate assets or operations to the home country or that discourage new investments abroad. In this context, ‘measures’ refer to national regulatory measures, but also include the application of administrative procedures or, even less tangible, political pressure.”

[9] See e.g. Art. 20 (2) (a) US Model BIT on Financial Services: “Nothing in this Treaty applies to non-discriminatory measures of general application taken by any public entity in pursuit of monetary and related credit policies or exchange rate policies.”  More known is Annex 2, 4 (b): “Except in rare circumstances, non-discriminatory regulatory actions by a Party that are designed and applied to protect legitimate public welfare objectives, such as public health, safety, and the environment, do not constitute indirect expropriations.” Canada includes those issues in Annex B 13 (1) of the Canadian Model BIT from 2004: “Except in rare circumstances, such as when a measure or series of measures are so severe in the light of their purpose that they cannot be reasonably viewed as having been adopted and applied in good faith, non-discriminatory measures of a Party that are designed and applied to protect  legitimate public welfare objectives, such as health, safety and the environment, do not constitute indirect expropriation.”

[10] There might be acceptable reasons for discriminating between third country investors (MFN): for example in sensitive sectors (like the financial market sector), where a specific treatment is only granted to investors of a country with the same level of supervision. The same rationale as in the GATS Annex on Financial Services applies.