The Investment Treaty Arbitration Review: Country Risk

I Introduction

Among damages experts, there is wide agreement that the risks arising from an asset's geographical location – otherwise known as country risk – can significantly affect the asset's value. There is also a broad consensus that investors consider these types of risks in their investment decisions.

Despite this consensus, the analysis of country risk is often little more than an afterthought in the damages analysis. As we discuss in this chapter, a proper assessment of country risk should carefully consider the exposure of a specific investment to the risks of the host country. However, it is quite common for valuation experts to account for country risk by applying simple shortcuts, which in many cases could prove inaccurate and lack theoretical support.

Accordingly, in this chapter, we:

  1. define what we mean by country risk;
  2. explain what financial theory says about accounting for country risk;
  3. discuss commonly applied techniques to account for country risk in valuation and their potential issues; and
  4. describe an alternative method that considers country risk more rigorously.

We conclude with a discussion of an area of particular interest in the context of investment arbitration: whether – and how – to account for risks protected under investment treaties versus unprotected risks.

II Defining country risk

Country risk refers to the incremental adverse political and economic risks specific to the country that hosts the asset relative to the level of risk that would prevail if the investment were located in a safer jurisdiction, often taken to be the United States or another Member State of the Organisation for Economic Co-operation and Development (OECD). There is broad consensus that risks specific to the host country matter: it makes a difference whether an investment is in Switzerland or Swaziland.2 The presence of country risk – like any other risk – reduces the market value of an investment. Market value, as we use it here, refers to fair market value, which is the most commonly applied standard in international arbitration cases.3

When valuation practitioners use the term 'country risk', they refer to the risk that 'a foreign government action will negatively affect the cash flows of a company conducting an international investment'.4 Country risk also encompasses the risk that 'imperfections in the country's executive, legislative, or judicial institutions [may] adversely affect the value of an investment by a foreign firm in that country'.5 This risk can manifest in several ways: asset seizures, creeping expropriation through unexpected taxes or royalties on profits, the instability of relevant government policies, a weak legal system, and internal and external conflicts such as general strikes, terrorism and war.6

It is an accepted finance principle that an asset's ability to generate future cash flows determines its market value. In recent years, the discounted cash flow (DCF) method has gained acceptance by tribunals and has become the primary valuation approach.7 In the DCF method, the valuation practitioner forecasts an asset's cash flows into the future and derives the market value by summing all future expected (or probability-weighted) cash flows, adjusted for timing and risk through the application of an appropriate discount rate. In the remainder of this discussion, we assume that assets are valued using the DCF method.8

The question we address next is how we should account for country risk in a DCF valuation.

III The theory: financial principles and country risk

A DCF model is based on expected or risk-adjusted cash flows. Hence, financial theory tells us to adjust a project's cash flows to account for country risk. In other words, we should identify the country risks relevant to the project's cash flows, estimate their likelihood and effects and adjust the cash flows accordingly.

The question then is whether the discount rate should also be adjusted. This depends on whether country risk is a systematic risk (also referred to as non-diversifiable risk). Systematic risk refers to risk that is inherent to the economic system and that affects all assets. Examples of systematic risks include inflation, movements of interest rates and performance of the overall economy.

In contrast, non-systematic (or diversifiable) risks are not related to the economic system as a whole but rather affect specific assets. Many projects in the mining and resource sectors face multiple project-specific, non-systematic risks – the probability of finding gold in a gold project, for example, depends on the geological and technical properties of the project itself rather than the performance of the overall economy. Other examples of non-systematic risks include the probability that the US Food and Drug Administration approves a new medication and the risk that a company can be sued for a defective product.

Finance theory prescribes that we should account for diversifiable risks in a project's forecast of cash flows and not in the discount rate.9 In their leading corporate finance textbook, Principles of Corporate Finance, Professors Richard Brealey, Stewart Myers and Franklin Allen advise not to 'try to compensate for political risks by adding casual fudge factors to discount rates. Fudge factors spawn bias and confusion'.10 This implies that country risk is mostly diversifiable. In support of this idea, there are actions that investors can take to mitigate it. For example, investors can purchase insurance against adverse political events or involve local banks in financing a project to increase the government's incentives not to harm the project.11

If country risk is diversifiable, we should account for it by adjusting cash flows, not by bumping up the discount rate. We discuss the implications of this finding below.

IV In practice: increasing discount rate with sovereign spread

Some valuation practitioners argue that even if country risk were diversifiable, adjusting the cash flows to account for the risks tends to be quite a tall order. It can be difficult – though not impossible – to estimate the probability and consequences of specific events. As a result, valuation practitioners sometimes propose a shortcut and apply a country risk premium (CRP) to the discount rate in DCF models.

In practice, valuation practitioners typically equate the CRP with the sovereign yield spread. The sovereign spread is the difference between the yield to maturity on dollar-denominated government bonds issued by the host country of the investment and the yield to maturity on US government bonds, which are considered risk-free. Therefore, the sovereign spread reflects sovereign risk – that is, the risk of default of the foreign currency government bonds and the expected recovery in the event of default.

In most cases, applying the sovereign spread to the discount rate is unlikely to provide an accurate estimate of the effect of country risk on an asset's value. There are several reasons for this.

First, although country risk and sovereign risk are related, they are not the same. Using the sovereign spread to measure country risk implies that investments in private assets face the same risk as creditors to a government. However, intuitively, the two types of risks can be very different. For example, a government may default on its debt, leaving an oil-exporting business largely unaffected. Conversely, there could be incidents of civil unrest that adversely affect a project's operations but that do not affect a government's ability to pay its debts.

Using the sovereign spread as a proxy for country risk fails to recognise that different investments in the same country can have different degrees of exposure to country risk. That is, country risk is not only country-specific but also asset-specific.12 Depending on the asset, using the sovereign spread as a proxy for country risk could either overestimate or underestimate the effect of country risk on the asset's value.

For instance, consider a project exporting oil to the international market. One could argue that the project might be at least partially insulated from exposure to the local economy and internal conflicts. The ability to generate cash flows is likely to be largely unrelated to whether the host country can pay its debts. Valuing the oil project using a discount rate that includes the sovereign spread as a proxy for country risk could understate the valuation of the project.

In contrast, suppose that a project has a contract with a government, from which it derives most of its revenue. In this case, the revenue stream is highly exposed to the risk of a government default. Applying the sovereign spread to the discount rate could actually underestimate exposure to country risk and thus overstate its value.13

Further, although the sovereign spread approach is straightforward, it lacks any theoretical support. As noted above, country risk should generally be accounted for in the cash flows, not in the discount rate. The approach of adding the sovereign spread to the discount rate has drawn many other criticisms owing to its lack of theoretical and empirical support.14

An empirical study found that less than one-third of the variation in sovereign spread accounts for country risk.15 The rest includes elements that have already been included in the discount rate or are unrelated to the country risk faced by assets, such as liquidity premiums, in the sovereign bond markets. To the extent that the sovereign spread includes risks that are already included in the discount rate, adding it to the discount rate could underestimate the value of the asset in question.

Local macroeconomic risk factors, such as inflation, can affect sovereign spreads but, depending on the type of project, do not necessarily entail increased country-specific risk for foreign investments. Similarly, it is difficult to argue that there is a link between liquidity in the trading of government bonds – which is a factor that drives spreads – and country-specific risk for foreign investments. Including the sovereign spread in full in the discount rate can result in understated asset valuations.

Including the sovereign spread in the discount rate poses yet another problem: the sovereign spread is the difference between two promised returns (measured as yield to maturity), whereas the discount rate in a DCF model is based on expected returns. The yield to maturity assumes that the bond survives to its maturity and makes all coupon and principal payments on time – it is a promised return. In contrast, the expected return can be much lower, especially if the bond has a high risk of default. In this case, including the promised return in the discount rate will understate the valuation if cash flows have already been adjusted to account for country risk.

Despite the shortcomings, there are obvious reasons for the sovereign spread approach to have widespread use. The sovereign spread is simple to estimate16 and has at least some bearing on country risk. Accounting for country risk using the sovereign spread approach is as simple as selecting the CRP for the relevant country and adding it to the discount rate.

There may be cases where default risk – as reflected in the sovereign spread – is directly relevant to asset valuation, for example, in a company with a government contract. However, even in those cases, adding the sovereign spread to the discount rate is unlikely to account for country risk correctly. Accounting for country risk by applying the sovereign spread approach tends to bear out the words of journalist H L Mencken, who noted that 'there is always a well-known solution to every human problem – neat, plausible, and wrong'.17

V Can we fix the sovereign spread approach?

To sum up so far, there is some evidence that country risk is largely diversifiable and asset-specific, and so should be accounted for by adjusting forecast cash flows.18 However, it can often be difficult to estimate how country risk will affect cash flows directly.

The sovereign spread is simple to estimate, contains useful information on default risk and has at least some bearing on country risk. However, in many cases, adding the sovereign spread to the discount rate will not provide an accurate estimate of the effect of country risk on asset value.

Given these factors, is there anything that can salvage the sovereign spread approach?

Some authors have attempted to refine the sovereign spread approach in an attempt to better capture the specific exposure of a particular asset to the factors that drive the sovereign spread. For example, some models adjust the spread for differences in the volatility of equity and bonds in the host country or differences in the volatility of equities in the host country and in developed markets.19 Other models replace the sovereign spread with a credit default spread.20 There have been efforts to incorporate asset-specific exposure to country risks, for example, by assuming that the country risk an asset faces is proportional to the sovereign spread multiplied by the proportion of revenues or profits derived in the host country.21

However, these refinements typically suffer from similar problems as the simple sovereign spread approach. They are subjective and arbitrary and have neither theoretical nor empirical support.22

VI An alternative approach

Given the weakness of the sovereign spread approach and its variants, is there another way to account for country risk, one that can somehow extract the useful data from the sovereign spread and use it to adjust the cash flows? It turns out that there is.

Geert Bekaert, Campbell Harvey, Christian Lundblad and Stephan Siegel (Bekaert et al.) have developed a framework to estimate the fraction of the sovereign spread that is attributable to what the authors call political risk. The approach produces the 'political risk spread'.23 Bekaert et al. measure political risk through a political risk rating produced by the PRS Group, a firm that provides political and country risk forecasts and ratings.24 The political risk rating is a score that combines several indicators of political risk, including the investment profile, government stability, corruption, and law and order.25 The investment profile reflects indicators that are of particular interest in international arbitration, including contract viability, expropriation and profits repatriation.26

Bekaert et al. show that, by using a regression analysis, it is possible to break down the sovereign spread into various elements, including the liquidity of the bonds, global and local macroeconomic risks, volatility of bonds returns and, of course, political risk. For example, for a given sovereign spread of 9 per cent, it could be that only 3 per cent relates directly to political risk – the political risk spread. The other 6 per cent relates to the other listed factors, such as liquidity or macroeconomic risk, that are already included in the discount rate.

We can then convert the political risk spread into a political risk probability, which can be interpreted as the likelihood of an adverse political event. For example, we may calculate that a 3 per cent political risk premium implies a 2 per cent probability of an adverse political risk event.

Next, we need to translate the adverse political event into an effect on the asset's cash flows. It is up to the valuation practitioner to judge this effect, and it is likely to vary by asset. The effect of political risk could be the reduction of revenues, an increase in costs or, in the case of expropriation, a complete loss of all cash flows. These are just a few examples as many other consequences could be possible.

To illustrate the approach, suppose that the annual risk of an adverse political event is 2 per cent. Also, suppose that we conclude that if a political risk event occurred, there would be a complete loss of the business. From the perspective of a party claiming for the loss of the business, this would probably be the most conservative assumption. We would then reduce the cash flows in the first year by 2 per cent to account for the risk of an adverse political event.

Note that the risk of an adverse political event is cumulative over time. Therefore, if the risk is 2 per cent each year, then the chance that an adverse political event would not have happened in the first year is 98 per cent. The chance that no adverse political event would not take place in the second year would be 98 per cent of 98 per cent, or about 96 per cent, and so on.

Finally, we would then estimate the value of the business using the same discount rate used to value a business in the US or another OECD country, but we would apply a 'haircut' to the projected cash flows, as described above, to reflect political risks.

This approach has a number of key practical advantages relative to the direct application of the sovereign spread or its variants. First, the method correctly adjusts cash flows rather than modifying the discount rate in line with evidence that suggests political risk is largely diversifiable. Second, the method recognises that not all the sovereign spread is relevant to political risk. By extracting only the relevant part of the sovereign spread, the method avoids the major drawback of sovereign spread-based models – namely, double-counting risks or including risks that are not relevant to the valuation of the asset in question.

Of course, the technique is best suited to assets with broad exposure to political risk. The technique could still overstate political risk for an asset focused exclusively on exports in a stable currency, such as the US dollar. In this case, we are likely to arrive at a more accurate valuation if we consider the relatively narrow range of political risks that affect the project and adjust the project's cash flows accordingly.

In sum, it remains crucial to carefully assess the specific political risk exposure of the asset we are valuing and to apply the appropriate technique. This could include:

  1. recognising that the revenues of a project are exposed to government default risk, in the case that a project derives most of its revenues from the host government;
  2. assessing the political risk spread and annual adverse political event probability, in the case that the asset has relatively broad political risk exposure; and
  3. judging that, because the project exports its product and sells it in a stable currency, it may have little exposure to political risk. In this case, all that may be required are some minor adjustments to forecast cash flows to account for any remaining political risks.

VII Identifying country risk protected by investment treaties

In investment arbitrations, foreign investors usually claim that the applicable treaty protects them27 from certain government actions, such as direct expropriation and other creeping forms of expropriation, including unexpected taxes, royalties, some forms of currency restrictions and some types of regulations. If a claimant prevails on this point, then the asset's value should not be reduced to account for these protected risks. In other words, the risk of an unlawful28 uncompensated expropriation should not reduce the value of the claimant's asset.

Of course, this is a legal issue about which we have no opinion. However, economists must consider that there will be questions about how the valuation should incorporate the parties' positions and, ultimately, the tribunal's decision. This debate is far from a purely theoretical exercise, as tribunals have adopted different views on the issue.29

On one side of the argument, the claimant should enjoy a blanket protection that assumes that the government will meet its legal obligations.30 This view considers adverse government actions to be unlawful. Under this position, the fair market value of the asset should exclude all government conduct from which the investment is protected. This is the position embraced by the tribunal in Gold Reserve v. Venezuela, which decided that 'it is not appropriate to increase the country risk premium to reflect the market's perception that a State might have a propensity to expropriate investment in breach of BIT obligations'.31

Proponents of the opposite side of the argument claim that the market value of an asset reflects all political risks as market participants incorporate the potential for negative government measures in their investment decisions.32 In that case, there are no protected risks and adverse government action is considered lawful. This legal position requires damage to be assessed accounting for all political risks. This is the view that the tribunal in Venezuela Holdings v. Venezuela adopted. The award notes that:

The Tribunal finds that it is precisely at the time before an expropriation (or the public knowledge of an impending expropriation) that the risk of a potential expropriation would exist and this hypothetical buyer would take into account when determining the amount he would be willing to pay in that moment. The Tribunal considers that the confiscation risk remains part of the country risk and must be taken into account in the determination of the discount rate.33

Some tribunals have opted for an intermediate view in which damages should assume that only the specific violation being arbitrated did not occur, and that future increases in risk of negative government conduct should reduce the claimant's compensation.34 For example, the tribunal in Flughafen v. Venezuela decided that protected risks should be set at the level prevailing at the time of the investment. The tribunal's position was that any increase in political risk occurring afterwards is not protected. The tribunal opined that:

[The claimant's expert] argues that legal, regulatory and political risks should not be incorporated into the model, because a government cannot create risks under its control before an expropriation, thus significantly reducing compensation.
. . .
The Tribunal agrees with this assessment by [claimant's expert]. A government that through the adoption of new political attitudes, adopted after the investment was made increases country risk, cannot benefit from a wrongful act, to reduce compensation.35

These divergent views on risks of government actions from which investors are or are not protected raise the question of whether there is a principled but also pragmatic way for valuation practitioners to incorporate them into their valuations. Fortunately, the answer is a tempered but optimistic 'yes'.

It should not be surprising that the weaknesses of the sovereign spread model and its variants discussed earlier carry over here, and perhaps become even more evident when trying to separate political risks that may be protected by investment treaties from risks that may not be. Similarly, the political risk spread approach proposed by Bekaert et al. bundles protected and unprotected risks. However, it turns out that the Bekaert et al. framework offers a promising road map that can be used to extend it to separate protected and unprotected risks.

The approach involves classifying the components of the political risk rating as protected and unprotected risks according to a specific legal instruction. For example, the tribunal could ask the valuation expert to assess damages under the assumption that risks of expropriation and profit repatriation are protected under the treaty. The valuation expert would perform an analysis similar to the one proposed by Bekaert et al., but extract the political risk spread that arises only from unprotected political risks. This unprotected political risk spread can be converted into a political risk probability and used to assess the effects of the unprotected adverse political events on cash flows in the same way as described earlier.

To be sure, the approach does not remove the unavoidable need for the judgement of the valuation practitioner. However, in our experience, it forces quantum experts to consider the relevant issues carefully and methodically. Hence, the method provides a sensible, pragmatic and principled framework that valuation experts can apply to assist tribunals in quantifying the effect of political risk on compensation according to the parties' or the tribunals' views regarding the level of protection an investment treaty offers.

VIII Conclusion

There is broad agreement that country risk can significantly affect asset values and quantum of damages in investment arbitrations. Accordingly, valuation experts, counsel and tribunals should carefully consider the nature and extent of political risk facing a given asset. Even assets within the same country can face quite different levels of political risk, depending on the source of their revenues and profits. An analysis of country risk should adequately account for these differences.

One of the most common – if not the most common – approaches to account for country risk in a DCF valuation is increasing the sovereign spread's discount rate. Despite its simplicity, which is almost certainly what gives it its appeal, this approach is neither backed by financial principles nor does it have empirical support. In most cases, adding the sovereign spread to the discount rate is unlikely to provide an accurate estimate of the effect of political risk on the value of an asset.

In this chapter, we have explained the problems with the method and described an alternative and robust method for quantifying the effect of a broad exposure to political risk.

Although the extent to which claimants' asset values should include protected risks is a debatable legal question, the method we have set out provides valuation experts with a mechanism to quantify the effect of alternative legal views on the issue.


Footnotes

1 Dan Harris is a principal, Fabricio Nuñez is an economist and Ilinca Popescu is a senior associate at The Brattle Group.

2 See Himpurna California Energy Ltd. v PT. (Persero) Perusahaan Listruik Negara, Award dated 4 May 1999, paragraph 358, in which the tribunal noted that 'the fact remains that it is riskier to enter into a 30-year venture in Indonesia than in more mature economies. And it is no answer to say that the contract has allocated 99% of the risk to the Indonesian side. After all, there are documents which by their terms allot 100% of the risk to the debtor: bonds. Although they may be denominated in US dollars, and although they may stipulate absolute obligations to pay, it still makes a difference whether the issuer is Switzerland or Swaziland'.

3 Fair market value is defined as 'the estimated amount for which an asset or liability should exchange on the valuation date between a willing buyer and a willing seller in an arm's length transaction, after proper marketing and where the parties each acted knowledgeably, prudently and without compulsion'. International Valuation Standards 2020, paragraph 30.1.

4 Geert Bekaert, Campbell R Harvey, Christian T Lundblad and Stephan Siegel, 'Political risk and international valuation', Journal of Corporate Finance, 37 (2016), p. 1.

5 ibid., p. 4.

6 id.

7 Of course, if the asset of the company being valued is itself publicly traded or has been traded close to the valuation date and under similar conditions to the valuation situation, then the observed market value typically takes priority. However, in the context of international arbitration, even when the market value is directly observed, the valuation should consider that the market value may reflect a discount for government actions that may be protected under a treaty (see Florin Dorobantu, Natasha Dupont and M Alexis Maniatis, 'Country Risk and Damages in Investment Arbitration', ICSID Review, Vol. 31 (2016), pp. 219, 220–21). We discuss the issue of protected political risks later in this chapter.

8 Issues with incorporating political risk into valuation are found in other valuation methods, such as market comparables. See Dorobantu, Dupont and Maniatis, op. cit. note 7, above, pp. 229–30. See also M Alexis Maniatis, Fabricio Nuñez, Ilinca Popescu and Jack Stirzaker, 'Accounting-Based Valuation Approach', Global Arbitration Review: Damages in International Arbitration (4th ed.), p. 268, for a discussion on the need to account for country risk in the context of an accounting based valuation.

9 Richard A Brealey, Stewart C Myers and Franklin Allen, Principles of Corporate Finance (10th ed., McGraw-Hill, 2011), Chapter 9, p. 213: 'A project's cost of capital [so the discount rate] depends only on market risk. Diversifiable risk can affect project cash flows but does not increase the cost of capital. Also don't be tempted to add arbitrary fudge factors to discount rates.'

10 ibid., Chapter 27, Section 27-5.

11 id. Alternatively, investors can set up operational or corporate structures between the parent and foreign subsidiaries to minimise exposure to country risk.

12 Dorobantu, Dupont and Maniatis, op. cit. note 7, above, p. 221 and footnote 8.

13 This is because the discount rate applies to the cash flows, which are the difference between revenues and costs. However, the presence of costs magnifies any risk to revenues for the cash flows. For example, suppose government bond yields were 10 per cent. If we think of a revenue stream for a project with a government contract as analogous to a stream of payments from the government, the discount rate for the revenues would also be about 10 per cent. But if the revenues have a discount rate of 10 per cent, then, because cash flows are the difference between costs and revenues and costs do not necessarily face the same risks, the risk to the overall cash flows could be higher.

14 See, e.g., Lutz Kruschwitz, Andreas Löffler and Gerwald Mandl, 'Damodaran's Country Risk Premium: A Serious Critique', Business Valuation Review, Vol. 31, Issue 2-3 (2012). See also Aswath Damodaran, 'Response to Damodaran's Country Risk Premium: A Serious Critique', Business Valuation Review (same).

15 Bekaert, Harvey, Lundblad and Siegel, op. cit. note 4, p. 2.

16 There are several sources that calculate country risk premiums (CRPs) as part of their data offering. For instance, Kroll publishes CRPs at https://www.kroll.com/en/cost-of-capital. Also see CRPs published by Professor Damodaran, available at https://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/ctryprem.html (web pages last accessed 6 Apr. 2022).

17 H L Mencken, 'The Divine Afflatus' in Prejudices: Second Series (Jonathan Cape, 1920).

18 We note that some political risk is systematic so it should be captured in the discount rate.

19 See, e.g., Bekaert, Harvey, Lundblad and Siegel, op. cit. note 4, above, pp. 3–5. See also Aswath Damodaran, 'Country Risk: Determinants, Measures and Implications – The 2021 Edition', NYU Stern School of Business Forthcoming (5 Jul. 2021).

20 Bekaert, Harvey, Lundblad and Siegel, op. cit. note 4, above, pp. 3–5. Damodaran, op. cit. note 19, above.

21 Damodaran, op. cit. note 19, above.

22 Bekaert, Harvey, Lundblad and Siegel, op. cit. note 4, above, p. 5.

23 ibid., p.2.

24 https://www.prsgroup.com/. The political risk rating and its components are discussed in the International Country Risk Guide – see https://www.prsgroup.com/explore-our-products/international-country-risk-guide/. There are other country and political risk ratings that may be used in a similar analysis, including: Economist Intelligence Unit: Country Risk Model, https://www.euromoneycountryrisk.com/methodology">https://www.euromoneycountryrisk.com/methodology; Organisation for Economic Co-operation and Development, Country Risk Classification, https://www.oecd.org/trade/topics/export-credits/arrangement-and-sector-understandings/financing-terms-and-conditions/country-risk-classification/; Bertelsmann Stiftung's Political Transformation Index, https://bti-project.org/en/index/political-transformation (web pages last accessed 25 Mar. 2022).

25 The political risk rating and its components are discussed in the International Country Risk Guide, op. cit. note 24, above.

26 id.

27 In this context, we mean the investor is protected in that, if the government does expropriate the asset, it should provide prompt compensation based on the market value of the asset at the time of expropriation. The precise form of protection will vary by treaty.

28 The tribunal must have found that the claimant has proven liability for a discussion on damages to be relevant. Hence, we assume we are valuing the asset on the basis that the tribunal has found for the claimant on liability.

29 See Dorobantu, Dupont and Maniatis, op. cit. note 7, above, which contains a very exhaustive treatment of these issues. The first portion of this section relies heavily on that article.

30 Dorobantu, Dupont and Maniatis, op. cit. note 7, above, p. 219.

31 Gold Reserve Inc v. Bolivarian Republic of Venezuela, ICSID Case No. ARB(AF)/09/1, Award dated 22 September 2014, Paragraph 841.

32 Dorobantu, Dupont and Maniatis, op. cit. note 7, above, p. 220.

33 Venezuela Holdings BV and others (formerly Mobil Corporation and others) v. Bolivarian Republic of Venezuela, Case No. ARB/07/27, Award dated 9 October 2014, Paragraph 365. A similar position was taken by the tribunal in St. Gobain v Bolivarian Republic of Venezuela, Case No. ARB/12/13, Award dated 30 December 2016, Paragraphs 718–19, 723. The tribunal in Tidewater Investment SRL and Tidewater Caribe C.A. v Bolivarian Republic of Venezuela, ICSID Case No. ARB/10/5, Award dated 13 March 2015 determined the expropriation to be lawful (i.e., included expropriation risk in the valuation).

34 Dorobantu, Dupont and Maniatis, op. cit. note 7, above, p. 220.

35 Flughafen Zurich AG and Gestión e Ingeniería IDC SA v. Bolivarian Republic of Venezuela, ICSID Case No. ARB/10/19, Award dated 18 November 2014, Paragraphs 904–05 (authors' translation) (emphasis added).

The Law Reviews content