The International Arbitration Review: Financial Debt and Damages in Investor–State Arbitration
Most submissions in international arbitration cases cite Chorzów Factory and the standard of full reparation under customary international law. The standard of full reparation enjoys widespread support. Yet its implementation always provokes intense debate. The devil lies in the details.
One of those details is debt financing. Many of the investor–state disputes recorded by UNCTAD have involved capital-intensive industries such as energy, mining, water and financial services, where the use of extensive debt financing is typical.2 Accounting for outstanding debt is relevant to loss quantification in these cases, because international claimants tend to be shareholders and the damages claimed reflect shareholder loss. That is, a shareholder claims damages in its affected investment; the affected investment does not present its independent claim.3
This chapter considers several consequences of shareholders' pursuit of international reflective loss claims.4
One consequence is the emergence of allegations of financial imprudence, typically directed by respondent states towards claimant shareholders. A common allegation is that claimants themselves were irresponsible in burdening an investment with excessive debt, prompting inevitably poor financial performance and a slide into financial distress. Such claims are relevant to an analysis of causation and may become more frequent after covid-19, given a general increase in debt levels and bankruptcies. The appropriate economic framework to assess allegations of financial imprudence is discussed below.
A second consequence concerns the magnitude of any shareholder reflective loss. Debt enjoys a priority right of payment, so a reliable damages analysis must first consider the impact of the measures at issue on the debtholders in an investment before quantifying any impact on shareholders. I explain that the presence of extensive debt typically reduces shareholder losses.
A third consequence relates to the incentives of both investors and a host state in the events leading to an arbitration. Extensive debt can sometimes lead to the escalation of a dispute, rendering arbitration the inevitable outcome. An analysis of financial incentives can often help explain the actions of the parties.
A common allegation by host states in international arbitrations is that claimant shareholders are the authors of their own misfortune. A claimant shareholder was imprudent, investing too little equity, while burdening an investment with too much debt. Counsel for the state may accuse claimants of inappropriately treating international arbitration like investment insurance, when the problem was always a claimant's own financing choices and the inherent vulnerabilities.
Available evidence might confirm the presence of extensive debt financing. Accounting statements might reveal deteriorating financial performance, substantial debt and high financial leverage, which refers to the proportion of debt funding out of total investment funding. The extent of leverage might exceed that observed elsewhere and may even have led to a restructuring or bankruptcy, either before or after covid-19.
Such evidence is informative and likely to be undisputed. However, it is insufficient by itself to indicate imprudent financing choices or excessive debt. Such conclusions require an analysis of causation. Did the measures at issue cause the observed deterioration in financial performance and the slide towards financial distress? Or was it just bad luck, or an inevitable consequence of under-investment and excessive risk-taking by a claimant?
Assessing causation requires a detailed analysis of the financial impact of the measures at issue. The relevant analysis must reconstruct the financial performance of the investment in the absence of (but for) the measures at issue,5 and compare reconstructed performance to reality. Reconstructing financial performance can demand significant modelling effort, depending on the complexity of the investment in question and the terms of the relevant contracts or concessions. The modelling effort should aim to trace the evolution of key financial ratios such as financial leverage6 and debt service coverage ratios,7 and ultimately to identify if sufficient additional cash flows would have been available to satisfy outstanding debt obligations.
If so, then the company or project could have avoided bankruptcy in the absence of the measures at issue, and the measures at issue were the cause of the financial problems. If not, then financial distress was inescapable notwithstanding the measures at issue, and was either the result of bad luck unrelated to the legal claims or the inevitable consequence of poor financing decisions.
An analysis of claimant imprudence also needs to consider the original expectations of both the claimants and lenders when they undertook the loans.8 The available information at the time should inform the claimant's financing choices; it would not be reasonable to second-guess them in the light of hindsight in general and the changed world following covid-19 in particular.9
An investment's debt capacity depends on the magnitude and certainty of expected cash flows.10 More debt is typically appropriate for activities with larger and relatively predictable cash flows; less debt for activities with smaller and highly volatile cash flows. More debt can provide significant financial benefits, including the imposition of business discipline and the opportunity to reduce a project's overall tax bill, as debt interest is tax deductible in most jurisdictions. Business discipline includes a commitment to stay with a defined business, and to return the proceeds of the business to lenders instead of investing in new projects. However, the various advantages of debt can come at the expense of potential financial problems down the road.
The presence of financial risk per se is not evidence of undue risk taking. A major theory of finance defines the optimal debt level with reference to a trade-off between the benefits of reduced tax payments on one side and the costs of potential financial distress on the other.11 From the perspective of this theory, eliminating financial risk altogether would be undesirable, because it would needlessly sacrifice project and shareholder value.
If debt is sizeable, third-party lenders will have had a natural financial incentive to perform due diligence on the borrower in question and to design the financing package to ensure the best possible chance of repayment. Prudent lenders will typically consider the legal rights and obligations of the borrower; analyse major business, market and technical risks; and develop a detailed financial model to forecast project cash flows that helps assess a project's ability to meet its scheduled debt service.
Public bond offerings can also attract scrutiny from independent ratings agencies and investors. These sorts of considerations ultimately determine loan pricing; elevated risks naturally prompt higher interest rates.12
Analysing debt issuance and contemporaneous lender expectations is therefore likely to cast light on allegations of financial imprudence, in addition to but-for analysis. Lender expectations represent an important and independent reference point to assess the reasonableness of financing choices and, more broadly, a claimant's overall expectations.13
ii Debtholder losses
The most that a shareholder can lose is the value of its equity in an investment.14 For example, if a state were to expropriate a house, the owner would lose only the value of its equity in the house, and not the entire value of the house itself.15
Allegations of overleverage primarily concern liability: the claimant shareholder caused its own downfall, not the host state. However, allegations of overleverage also have consequences for damages. Debt has a priority right to payment, so more debt implies that a larger share of project value must flow to debtholders before any residual value can flow to shareholders, including the claimant. Reliable assessments of shareholder damages must consider the priority payment of debt.16
Suppose a project were worth US$100 million, but that the measures at issue destroyed US$70 million of economic value, reducing the project's value to US$30 million. Suppose also that the project was prudently financed with US$50 million in debt and US$50 million in equity. The project would face bankruptcy because of state measures, because the project's value (US$30 million) would fall below the face value of the outstanding debt (US$50 million). As a result, I would expect debtholders to capture all of the US$30 million in remaining value from the project after the measures. Debtholders incur a US$20 million loss, while the shareholder loses the entirety of its US$50 million investment.
Suppose that a shareholder then responds by initiating an international arbitration, but that the debtholders do not do likewise. This assumption reflects our experience that shareholder claims predominate in international arbitration, while debtholder claims are less common, in part because project lenders often are domestic banks that lack standing to claim protection from an international investment treaty. The shareholder would likely advance claims under the relevant treaty and pursue damages equal to the entire US$50 million value of its lost equity.
The US$50 million claim for shareholder reflective loss would be necessarily lower than the US$70 million of enterprise value destroyed by the measures at issue. Any damages claim for shareholder reflective loss must first account for the debtholders' priority right to payment and deduct the US$20 million in value lost to the debtholders. With only a shareholder claim and no corresponding debtholder claim in our example, a state could take a total of US$70 million in economic value, for which it would owe only US$50 million in shareholder damages.
The consequence of overleverage is to reduce the compensation owed by a state even further. Suppose that a shareholder had financed our US$100 million project with US$90 million of debt and US$10 million of equity. The ensuing shareholder arbitration would likely involve allegations of overleverage and imprudence, which could affect liability. However, the resulting shareholder damages would relate only to the shareholder's US$10 million investment, after proper accounting of the US$90 million in outstanding debt. US$10 million is less than the damages available to a comparable claimant utilising much less debt financing (US$50 million, for example), and far less than the total economic harm caused by the measures at issue (US$70 million).
The possibility of debtholder losses can arise even before there has been an outright event of default or insolvency, which arises when the value of the assets falls below outstanding liabilities. I distinguish between the book or face value of debt, and its market value. Book or face values indicate the amount of debt outstanding at any point in time. Market values depend on the current value of prospective scheduled interest and principal payments, which depends on the risk of default, considering the returns available elsewhere given prevailing market conditions. Debt market values can fall prior to events of default or insolvency, and these falls can represent losses suffered by debtholders because of the measures at issue.17 A reliable damages analysis needs to consider the possibility that debtholders shared in the overall economic losses because of the measures at issue.18
The presence of extensive debt financing affects not just the analysis of liability and damages in an international arbitration, but also the incentives of both investors and a host state in the lead up to the arbitration. Extensive debt can render early settlement less attractive to both investor and state and leave arbitration as the inevitable outcome. A careful analysis of debt and financial incentives can help illuminate the actions of the parties and the events leading to the dispute, with potential consequences for both liability and damages.
For example, the measures at issue may have left shareholders with little or no remaining value, while covid-19 may have administered a further blow. And with little left to lose, shareholders may prefer to escalate a dispute and run the risks of an investment arbitration rather than to pursue negotiations through the underlying project company. Negotiations between the underlying project company and the host state are likely to require the involvement and consent of lenders, and any resulting settlement value could largely flow to them in any event. Escalating a dispute in the hope of triggering a response from the host state and proceeding to arbitration provides a better chance of obtaining at least some equity return. Of course, shareholder damages in an arbitration would need to consider the priority payment of debt, as explained above, but at least the arbitration process might proceed directly between the shareholder and the state, without the complications of lender involvement.19
At the same time, extensive debt financing could create a disincentive for host states to seek alternative solutions. Suppose that a host state displayed some willingness to negotiate with the project company and even to provide compensation (albeit partial). The state would logically consider whether compensation would benefit foreign shareholders sufficiently to avoid arbitration.
The state might foresee that lenders could capture a large part of any compensation, leaving shareholders with little. The state might therefore fear that an arbitration with a shareholder would emerge in spite of any realistic payment to the project company. A settlement with the project company would not therefore solve anything, and would only serve to compensate debtholders, the one class of investor unlikely to arbitrate anyway.
These considerations do not relate to the possibility of multiple legal proceedings and the potential for double recovery. Table 1 illustrates a scenario in which payment of partial compensation to a project company would not impact the shareholder damages claim in a subsequent arbitration. The state makes a partial payment, but shareholder damages remain unchanged, not because of double recovery by shareholders, but because the state's partial payment represents an effective payoff to debtholders via the project company.
Table 1 Shareholder damages unchanged by partial compensation to project company
|But for [A]||Actual [B]||Actual plus partial compensation [C]|
|Total value|| +||100||60||70|
|Debt|| see note||80||60||70|
|Equity damages|| [A]–||20||20|
|Notes:  [A] and [B]: assumed|
 [C]: [B]+
The ability of shareholders to pursue reflective international claims can often give rise to claims of excessive debt, and a need to analyse the causes of financial distress. The relevant analysis involves detailed but-for reconstruction and a review of the claimant and lender due diligence undertaken at the time of any major financing decisions. Avoiding hindsight will be necessary given that higher debt and more bankruptcies will be a feature after covid-19. At the same time, quantification of shareholder reflective loss must consider whether debtholders have suffered a portion of any economic harm alongside equity holders. Extensive debt financing actually reduces the magnitude of shareholder damages, all else being equal. Extensive debt financing can also affect shareholder and state incentives, making the escalation of disputes more likely. Careful financial analysis can help to explain the incentives and actions leading to a dispute, with potential consequences for both liability and damages.
1 Richard Caldwell is a principal at the Brattle Group.
2 Out of a total of 942 registered cases, 214 related to electricity, gas, steam and air conditioning supply, and water supply. A further 92 cases related to financial and insurance services, 47 to real estate, and 152 to mining and quarrying (including crude oil production). Debt financing is common in these sectors, and such cases collectively account for close to 60 per cent of the registered cases.
3 I understand that foreign-controlled companies can pursue international treaty claims in some circumstances.
4 The ability to advance international claims for shareholder reflective loss forms a fundamental part of treaty protection, but it raises a theoretical risk of double recovery. In relation to the same conduct, a shareholder could pursue a claim for shareholder reflective loss before an international tribunal at the same time as an affected company or project pursued a domestic court action. Success in both an arbitration and domestic action could result in a state paying damages twice over in relation to the same breach. The theoretical possibility of double recovery motivates restrictions on shareholder claims in many domestic legal systems. This chapter does not discuss the relative merits of shareholder reflective loss, but examines several consequences of it.
5 The reconstruction should eliminate the impact of the measures at issue, but reflect the impact of independent factors, such as changes in market prices and circumstances unrelated to the claims at issue.
6 Debt to equity or debt to enterprise value.
7 Cash available for debt service in a given year divided by the debt service in that year.
8 The host state may also have set out its financing expectations.
9 Of course, that does not imply that hindsight is irrelevant to an analysis of causation. We could conclude that a claimant was prudent and that it sized debt appropriately based on an analysis of the original expectations. However, using hindsight, we might also find that market circumstances then moved in such a way as to cause financial difficulties and even an insolvency, before we consider the additional impact of the conduct at issue. In such an event, we would conclude that a claimant was prudent and at the same time that the conduct of the respondent state did not cause financial difficulties.
10 Richard A Brealey, Stewart C Myers and Frank Allen, Principles of Corporate Finance, tenth edition, McGraw-Hill Irwin (2011), pp. 440–70.
11 See for example Jonathan Berk and Peter DeMarzo, Corporate Finance, third edition, Pearson (2013), pp. 550–2.
12 Roughly 80 per cent of public bond issues in the US market are at an investment grade rating. Investment grade ratings are given to the largest, most creditworthy companies and projects. The remaining 20 per cent or so of public bond issues fall into the high yield category. High yield issuers are often moderately sized companies without the size advantage or long history of large corporations. A high yield issuance involves additional risk compared to investment grade, but high yield is not a signal of an unreasonable or imprudent financial choice.
13 However, the expectations of shareholders and lenders may legitimately differ, in part reflecting the distinct interests of shareholders and lenders in an investment project. Lenders are concerned with the ability of borrowers to service and repay a debt. Lenders are therefore likely to adopt conservative assumptions, and consider downside risks that could prompt loan losses, but largely ignore potential upsides from which they would not benefit. In contrast, shareholders will logically consider both the potential downsides (where they stand to lose money) and upsides (where they stand to gain).
14 A shareholder could experience other contingent losses, such as a loss of reputation or difficulties refinancing outside of the host state. I have ignored such contingent losses for ease of presentation. An equity holder could lose more than its equity in an investment if it also guaranteed the debt financing.
15 I assume that the house was financed with a mortgage, and that the mortgage was non-recourse, meaning that the lending bank could only pursue the borrower until the foreclosure of the house in the event of a default, and could not continue to pursue the owner for any shortfall in payment thereafter.
16 The most common valuation approach is the weighted average cost of capital (WACC) valuation method, which is an indirect approach in that it estimates equity in two steps: discounting project free cashflows at the WACC to estimate the overall enterprise value and deducting the value of outstanding debt from the enterprise value to determine the equity value. A direct alternative is a dividend discount model, which is a type of flow-to-equity method that estimates the value of equity directly by discounting projected cash flows to equity (dividends) at an appropriate cost of equity. See for example Brealey (see footnote 10), p. 479, and Berk (see footnote 11), chapter 18.
17 Consider a hypothetical project with US$80 in outstanding debt. Assuming that the project had an equal likelihood of generating cash flows with a present value of US$110 in a favourable scenario, and US$90 in a less favourable scenario, the average expected outcome across both scenarios would therefore be US$100. The project could pay back all the debt even in the less favourable scenario (as the US$90 of project value still exceeds the US$80 of debt). The value of the equity would be US$20, representing the US$100 of project value less the US$80 of outstanding debt.
Now suppose that the measures at issue reduced the present value of expected cash flows by US$20, permitting the project to earn only US$90 in the favourable scenario, and US$70 in the other. The average project value would now fall to US$80. The debt would likely suffer an impairment in the less favourable scenario, as the debtholders could at most capture the project's value of US$70 in that scenario. The debt would therefore be worth only US$75 in total, calculated as the average of receiving US$80 in the favourable scenario and only US$70 in the other. The corresponding equity value would be US$5, calculated as US$80 minus US$75, equal to the average of retaining US$10 in the favourable scenario (after repaying US$80 in debt), and retaining zero in the other scenario.
It would be a mistake to calculate an asset value of US$80 after the measures at issue, and then to subtract the full US$80 of the face value of the debt. An illusion would arise that the equity had no value (US$80 – US$80 = zero) and that the debtholders suffered no loss. The correct analysis reveals a decline in asset value of US$20 (US$100 – US$80), which breaks down into a decline of US$15 in the value of equity (US$20 – US$5) and a decline of US$5 in the market value of debt (US$80 – US$75).
18 An analysis may conclude that the measures at issue did not have a significant impact on debt market values, and thus that the shareholders bore the entirety of the harm. However, the alternative is also possible, and will depend on the extent of both debt financing and the economic impact of the measures at issue.
19 The economics literature highlights the presence of incentive problems for highly indebted firms. For example, equity holders in highly indebted firms have an incentive to pursue high-risk strategies. See Berk (see footnote 11), pp. 553–7.