EJIL: Talk! | 26 January 2021
Against DCF valuation in ISDS: on the inflation of awards and the need to rethink the calculation of compensation for the loss of future profits
by Toni Marzal
One of the most noticeable facts in recent ISDS is the spectacular inflation of compensation awarded to investors. The overall increase in the amounts is well documented. Of the (now more than) 50 known awards in excess of USD 100 million, none was rendered before 2000 and only 11 before 2010. What is most attention-grabbing, however, are the dizzying multi-billion numbers reached in some extractive industry cases (e.g. $8.7 billion in ConocoPhillips v Venezuela, $6.6 billion in P&ID v Nigeria, $5.9 billion in Tethyan Copper v Pakistan). These awards have made headlines. More importantly, they impose an impossibly heavy burden on the finances of the States concerned.
What explains this trend? The single most important factor is probably the now widespread use by tribunals of the DCF (discounted cash flows) valuation method to determine the value of the investment affected by the State’s breach. It consists of estimating the worth of a business as equal to the present value of the income it is expected to generate in the future (see here). It is widely used in financial practice, particularly where the income stream can be reliably predicted. It thus stands in stark contrast to cost-based methods, where value derives from the amounts historically invested. The stakes involved in choosing between income and cost-based methods can be fantastically high. In the Tethyan Copper award, the investor’s sunk costs added up to no more than USD 300 million, whereas the DCF calculation produced over USD 4 billion. Indeed, most of the highest awards tend to rely on DCF.
Why is the use of DCF so widespread in investment arbitration? The first thing to remember is that it is a very recent development. Arbitrators used to be highly suspicious of this method. Because it is necessarily based on predictions about the future, it was perceived as fundamentally speculative (no less than ‘the prophecies of Nostradamus’, as one scholar forcefully put it), and thus contrary to the principle that speculative losses are not compensable. Its use remained rare all the way until the 2000s (and amounts awarded more moderate).
Why then has arbitral practice so radically shifted, to enthusiastically embrace DCF? The key argument for DCF is not a legal one, but an analogy with financial practice, where this method is prevalent (see e.g. here). Simply put, because real-life investors estimate the value of an investment via the DCF method, this is also how damages should be quantified in ISDS. The beginning of this shift can be traced back to Judge Brower’s separate opinion in the 1987 Amoco decision of the Iran-US Claims Tribunal. The tribunal majority made it clear that, even if prospective investors commonly rely on DCF to make investment decisions, this does not mean that tribunals should do the same. It even considered it doubtful whether DCF, given its inherently speculative nature, could ever be validly used in international adjudication. It therefore chose to employ a cost-based valuation method, with an additional sum to account for lost profits during a limited period of time. Judge Brower, however, was highly critical of this approach. In his separate opinion, he accepted that the non-compensability of speculative losses is indeed a valid principle of international law, but emphasised that it concerns only the existence of compensable harm (which if established will trigger liability), not the determination of its magnitude (which should follow business practice, even if the process is inherently speculative). Moreover, the majority’s rejection of the DCF, he argued powerfully, came from a ‘misunderstanding of economics’. A division of labour is thus established in Judge Brower’s opinion: determining whether the State has acted wrongfully is essentially a legal question, but once that is established the calculation of compensation will primarily be a matter of economic and financial expertise.
Even if Judge Brower ‘lost the battle’ of Amoco, he ‘won the war’ of valuation methods. His separate opinion has proved far more influential, and it is now commonplace to view the establishment of liability as entirely separate from the calculation of compensation, with the former as a legal question and the latter economic (e.g. here and here). The Amoco majority’s resistance to DCF is seen as an example of lawyers failing to grasp basic economics (e.g. here and here). The multi-billion Tethyan Copper award takes this logic to the extreme: against Pakistan’s protestations that the use of a certain approach to DCF advocated by the investor had never before been validated by an investment tribunal, the tribunal stated that legal authority is entirely irrelevant to the choice of valuation methods.
Is the dominance of DCF in investment arbitration truly unassailable? In reality, the analogy with financial practice, on which the current consensus is so solidly built, is a false one. Leaving aside the issue of the (very) different incentives that operate in the two contexts (see here), there is a fundamental difference between the reasoning of a prospective investor and that of an arbitral tribunal. The former must only carry out a purely factual assessment – estimate the extent and probability of a future income stream. The job of the tribunal goes much further than that – it must also establish whether the expectation of such an income, where it does exist as a matter of fact, is protected under the investment treaty. Thus, how it chooses to approach the calculation of damages will be determinative of the true extent of the rights of investors (and obligations of States). This is also true with respect to the choice of valuation method: a cost-based valuation means that the investor’s entitlement is limited to the measure of its reliance on treaty-conforming behaviour by the State, whereas an income-based one implies that the entitlement extends to expectations of future profitability. Valuation is not therefore a mere technical appendix to the actual legal process, but absolutely central to it.
The effect of today’s tendency to analogise valuation in arbitration to real-market valuation is to treat it as a given that investors are indeed entitled, as a legal matter, to compensation for the loss of the expectation of future profits. This leads to a massive inflation of awards. Tribunals have ceased enquiring if such an entitlement exists, focusing instead on the purely evidentiary question (see here) of whether, as a matter of fact, there is sufficient evidence that the investment would have been profitable and, if so, come to the most reasonable estimation of that income stream (see e.g. Bankswitch v Ghana, Devas v India). By resorting so widely to DCF, tribunals are acting as if the right to future profits necessarily exists, without however justifying that it does.
In reality, it is perfectly appropriate that tribunals avoid DCF to estimate the value of an investment in the quantum assessment, even though a real-life investor would have resorted to this method. If arbitrators are to take seriously the idea that quantum issues are rights issues, their reasoning should be broken down into two steps. In the first, they should determine whether such an entitlement to expected profitability exists, which is far from obvious. To establish this, it is not enough to argue that it is probable that profits would have been generated but for the State’ breach. That loss must also be shown to be legally compensable, and for that it is indispensable that the applicable law grants investors some entitlement against this specific deprivation. This is typically the case where the investor has been deprived of a right to property, since property of an asset usually carries, from a legal point of view, the right to reap its fruits. Conversely, such an entitlement is lacking where the existence of profits depends on some future discretionary act by the State, i.e. when the right to those profits has not yet been acquired. For instance, where the State’s breach has resulted in the loss of the opportunity for the investor to enter into a profitable contract or licence, but an actual right to that contract or licence was still lacking, it is wrong to award compensation for the loss of those profits (Lemire v Ukraine and again the very recent Devas v India are examples of this mistake).
Where the expectation of future profits is indeed legally protected, tribunals should then assess, as the second step of the reasoning, the exact measure of profitability to which the investor is entitled. Again, this is not a purely factual assessment. Here it is key to recognise that profitability is largely in the hands of States, who are free to drastically reduce the profit margins of businesses (and thus their economic value) through legitimate regulatory action (e.g. by increasing taxes). Thus, the fact that the investor has a legitimate expectation of reaping some future profits does not mean that it is entitled to the entire value of the profitability that happens to be projected at the moment of the State’s breach (or other valuation date), as tribunals commonly understand (see e.g. Quiborax v Bolivia). Instead, the tribunal’s task is to reconcile the investor’s legitimate claim to future profits with the State’s legitimate prerogative to decide on their appropriate level in the pursuit of the public interest. To treat the calculation of compensation for the loss of future profits as a purely factual assessment leads to absurdities such as considering that oil companies should be compensated against environmental legislation that renders it too costly for them to extract oil, with the benefits they would have obtained from that extraction (see here and here).
So how should tribunals quantify the loss of future profits? Such an assessment necessarily involves some form of equitable judgment (see here): investors must be guaranteed no more than reasonable profits. This notion can serve as a compromise between the competing interests at stake (where the investor’s is indeed a protected one). It has also been discussed at length in the recent saga of cases against Spain, arising from the abrogation of a regulatory framework that guaranteed investors in renewables an above-market premium. Some tribunals (see here) have limited compensation for the loss of future profits to only where these fall below a reasonable rate of return, given that this was found to be the rate that Spain had guaranteed. Such decisions are already heading in a promising direction, by showing firstly that the assessment of quantum is inseparable from that of that of the rights of investors, and secondly that an alternative to DCF valuation based on a reasonable rate of return is perfectly workable (as had already been shown by some economists).
It is true those cases fall short of what we advocate, inasmuch as the tribunals’ reliance on a reasonable rate of return was said to derive from a specific commitment by Spain, rather than out of the need to reconcile the interests of investors with those of the public. It is nevertheless submitted that, for the reasons explained, and regardless of possible treaty reforms to curtail the inflation of awards (see here), a reasonable return should already serve by default as the appropriate measure of compensation for the loss of future profits. It is to be hoped that tribunals will move in this direction, approaching valuation seriously as central to the definition of investors’ rights, and putting an end to the legally unjustifiable abuse of DCF.
For hyperlinks and notes, view on original page