Les demanderesses réclamaient des dommages-intérêts pour le préjudice qu'elles affirmaient avoir subi du fait qu'elles n'avaient pu exploiter un gisement de gaz (Y) dans l'Etat défendeur. Après avoir rendu une sentence intérimaire sur la responsabilité, le tribunal arbitral s'est prononcé dans sa sentence finale sur la question du quantum. Pour ce faire, il s'est appuyé sur le prix auquel le gaz produit par le gisement aurait pu être vendu.

'Price

290. The Claimants suggest that an appropriate across-the-board netback value for exported gas from [the field] should be $40; the Defendant says $27. Whether one accepts one or the other of these figures, or something in between, or yet again a value above $40 or below $27, is not a matter of scientific precision, but one of the choices to be made in establishing a value by selecting among a myriad possible indirect methods, each of which is perforce based on assumptions.

291. Thus, for example, to establish the netback value of [the] gas from [capital city of customer State], [an economist and energy-industry consultant called by Claimants (Expert A)] selected as a basis of comparison the price in [that city] of heavy fuel oil used in conventional steam generation of electricity. Granted, one must find some basis for setting a price, since [the] gas has never been sold anywhere, and no [State X] gas has ever been sold in [customer State]. Moreover, the choice of heavy fuel is rational, because oil-burning power stations may be converted to burn gas, and so will switch to gas if the price is competitive. But there are a host of questions with respect to which reasonable experts could differ: is the source for price information with respect to heavy fuel oil reliable? does it cover an appropriate reference period? does it contain taxes which do not apply to gas, or vice versa? what is the prospect for heavy fuel oil on a 5-year horizon? or 10? or 20? what might be the effect of subsidies in [customer State]? or price supports? would the exchange rate . . . pose different dangers for gas than for heavy fuel oil? will dollar-denominated contracts be signed with [customer State]?

. . . . . . . . .

294. It must be understood that the Claimants' tariff projection, and [an economist and energy-industry consultant called by Respondent (Expert B)]'s critique thereof, are based on numerous assumptions involving international diplomatic relations and macroeconomics-in addition to the more ordinary, but no less daunting, task of predicting the market evolution in a number of consuming and producing countries.

295. The Tribunal is not in a position to reconstruct tariff forecasts of its own. To "adjust" the experts' contradictory assumptions one by one, accepting some while rejecting others, would do no more than to create the illusion of a method. In light of its more general conclusions with respect to the marketability of [State X] gas in general and [the gas field's] gas in particular, it is sufficient to conclude that the Claimants' projections cannot be taken as preferable to those of the Defendant. Accordingly, the Tribunal is not satisfied that [the gas field] could generate netback sales approaching those postulated by the Claimants.

296. This being said, the Tribunal has no objection to [Expert A]'s method as such, as long as it is not claimed that it is "objective" or "scientific" in the sense of yielding a demonstrably correct answer.

. . . . . . . . .

Discount rate

310. . . . The Claimants have projected lost profits over 35 years, on the assumption that a successful venture would likely have been extended beyond its initial term. Since Article 9 of the JV Agreement provides for a 25-year term with no assurances of any extension, the time frame for assessing lost profits must be limited to 25 years. (This is not a matter of great moment, as the present value of net cashflow from Year 25 to Year 35 is a small fraction of the aggregate claim.) If income is projected over that period with a significant discount factor to establish a present value, cash flows anticipated more than 10 years into the future become relatively insignificant. Thus, the present value of US$ 1,000,000 to be received 10 years from now is US$ 631,000 if one discounts at 4.5%; US$ 348,678 if one discounts at 10%; and US$ 107,374 if one discounts at 20%.

311. The Claimants are acutely aware of this phenomenon. While their experts . . . project annual flows from [gas field] in excess of 28 bcm/yr, there is none at all in the first two years, less than 14 bcm/yr in aggregate for the first five years, and indeed annual amounts inferior to 28 bcm until the 10th year.

312. This is an obvious feature of a remote project which remains to be developed, which is not connected to a transport system, and where only one out of 45 reef formations has even been perforated by drilling.

313. An eloquent and uncontradicted demonstration of the importance of timing was made before the Tribunal by reference to alterations of the net present value of the project if one postulated a five-year delay in development . . . If the discount rate were 4.5%, the present value of the revenues assumed by the Claimants over 20 years would in the event of such a five-year delay decrease only from $1,089 million to $913 million-a 16% decrease. If the discount rate were 10%, of course the starting point would be far lower in the first place-in fact $487 million-and the decrease on account of delay would also be steeper: to a present value of $333 million, i.e. 31.6% less. If the discount rate were 20%, the starting point would be a value of $55 million, decreasing to $27 million in the event of a five-year delay, or 51%.

314. With these consequences in mind, one can readily see why the Claimants searched for a theory which would allow them to apply a low discount rate. They thought they had found such a method, allowing them to put forward a discount rate of 4.5%, effectively treating the prospective earnings from [gas field] as if it were as safe as gilts. This result was advocated on the footing that their so-called "deterministic" approach1 neutralised the risk factor by using conservative input figures which the Claimants' experts referred to as "high probability cashflow" and which therefore, or so the Claimants argue, do not warrant further reduction on account of risk.

315. In the course of his oral testimony before the Tribunal, [Expert A] suggested that it was wrong for [Expert B] to propose a 27% discount rate simply because it so sharply reduces the present value of net cashflow postulated in the mid to distant future. Thus, after indicating that his own 4.5% proposed rate would yield an 88-cent present value of one dollar to be received five years hence, he noted that [Expert B]'s much higher rate would yield a present value of only 34 cents. He then said: "Thus [Expert B] is proposing for year 5 that the discount rate, or the amount of reduction of the cashflow, be roughly what you would expect to find using a time value of money rate." . . .

The Tribunal cannot understand why one would be looking, in this context, for the "time value of money rate". The object of the exercise is not to assess the present value of certain future revenues, but of hypothetical future revenues. To apply a high discount rate is no stranger than saying that one will only take a big risk if there is a big reward - in fact it is the same thing. Given the uncertain prospects of achieving profits from developing the [gas field], [Expert B] advances the coherent suggestion that one should not place a value on the income projected for year 5 in excess of 34% thereof.

316. Both [experts] made impressive and methodologically coherent expositions; there is nothing wrong with [Expert B]'s method just because his assessment of factual prospects of a profitable investment programme to develop the [gas field] is bleak.

317. Given the kinds of conjecture inherent in the exercise of evaluating remotely located gas reserves-involving political and economic developments on a truly international scale, and over more than a biblical generation-there must be a "soft number" somewhere. The Claimants are in no position to castigate [Expert B]'s 27% discount figure as "soft" because of its multiplicity of imprecise variables which must call for subject inputs, since their own apparently "harder" number of 4.5% is exposed as an illusion when one considers the multiple subjective inputs that went into the "high probability" net revenue streams to which the low discount rate is then applied.

318. The Claimants' costs forecasts include assumptions about pipeline tariffs which in turn are based on assumptions about debt/equity ratios which in turn require assessment of required rates of return for investment as opposed to loans. . . . This gave rise to a multitude of debates . . .

319. There is nothing inappropriate or surprising about such debates.

320. Nor is there anything inappropriate about the Claimants taking a view of the myriad parameters which affect the returns on important investments such as bringing the [gas field] into production. Nor indeed is it wrong for the Claimants to make optimistic assumptions; it is for the Tribunal to determine their plausibility in light of the evidence. . . .

321. What is inappropriate, however, is any suggestion that the Claimants' methodology, in "establishing" so-called high-probability financial outcomes and then applying an exceedingly low-risk discount rate to them, is somehow more scientific and less intuitive than the approach advocated by [Expert B].

322. For example, [Expert C, called by Claimants] applied a range of landed prices in [customer State] from $115.85/mcm minimum to $155.76/mcm maximum for the purposes of their computer simulation and its thousands of iterations. This range of prices was then assumed reliable to generate revenues on gas exports of 12 bcm/yr in 2002 and 2003, and 16 bcm/yr thereafter. These figures were adjusted by assigning a 50% probability that the . . . Pipeline would be completed in 2002, a 75% probability that it would be completed in 2003, and so forth until a 100% probability of completion by 2006. Incorporating the "uncertainty element" of [consultant]'s forecast, [Expert C] then applied an 85% probability to the [customer State] market as a whole.

323. The method as such gives a semblance of being scientific, but is nevertheless based on assumptions that are anything but scientific (in the sense of being demonstrable, as opposed to being assumed by intuition, or the accumulation of anecdotal indications).

324. For what it boils down to, concretely, is that the Tribunal is asked to believe, if one takes the year 2004, that there is a practical and quantifiable certainty that [State X] exports to [customer State] will reach 16 bcm adjusted by a factor of 0.875 (no pipeline) and a further factor of 0.85 (overall market risk). As to the resulting "high probability" of 11.9 bcm, however, there is a 0% probability that the price would be less than $115.85, because even if the computer were asked to do a million iterations the inputs would not allow a lower outcome.

325. The debatable assumptions inherent in this approach are apparent and require little comment.

326. [It was] also established in cross-examination at the hearing that [Expert A] had not taken into account the effect of [customer State] taxes, which would have reduced the landed price to $78.25 even if one began with the average price of $135.95 he had forecasted . . . The difference is fundamental (unless one assumes profit margins in the order of 50%).

327. Similarly, [Expert C] assumed that the . . . pipeline would be operational at the earliest in 2005 and by the latest in 2009. [Expert A] forecasted average netback values of $59.55 for 10 bcm of [State X] exports to [State Y] in 1906 [sic], $71.30 for 15 bcm in 2007, and increasing to $76.99 for 20 bcm every year forward until 2020. This was subject to a 50% overall market risk.

328. The point is obvious: what about the risk that the . . . Pipeline would not be built at all through the rugged and war-torn country of [State Z]? Apparently that risk is purportedly incorporated into the 50% risk for the [State Y] market as a whole, in which case the Tribunal believes that figure to be wholly inadequate. The alternative is that it was omitted; if so, the Tribunal, on a no more nor less intuitive basis than that of the multiple forecasts on either side, would not give the pipeline any significant chance of materialising within the relevant time frame.

329. In truth, both sides have brought to bear highly professional disciplines on what perforce are intuitive assumptions. In having access to the valuable advice of such analysts as [Experts A and B], business executives all over the world make the age-old transition from guess to educated guess. But guesswork it remains; that is the essence of the concept of risk and venture. If this were not so-if it were somehow possible to purge the exercise of its fundamentally intuitive elements-there would be no way of explaining why otherwise highly successful enterprises led by highly competent managers time and again embark on projects which they are able to defend with ardour, eloquence, and scientific method, but which still end in disaster.

330. In spite of the preferences expressed by [Expert A and the author of an opinion on principles of discounting and risk analysis], who suggested that a unitary discount rate was not an appropriate way for corporate management to put a value on a sizeable venture involving multiple diversified risks, the Tribunal does not accept that such a rate is inherently inferior when evaluating a loss-of-bargain claim. The Defendant cited as an illustration a Morgan Stanley Dean Witter evaluation of Gazprom's "forecast cash flow", which was based on a 16.2% discount rate . . . Discounted cashflow was described by Morgan Stanley Dean Witter as "the only fair valuation method for Gazprom". Whether or not a more complex approach might have been used to evaluate a corporate investment opportunity, this kind of valuation bears considerable resemblance to the task of a court or a tribunal faced with a loss-of-bargain claim, as long as care is taken to focus on the benefit of which the wronged party has been deprived, rather than on what a third party would have paid to take over its position. The Tribunal believes that it has taken that care.'

1
To be precise, it was a hybrid method . . . using "probabilistic methods to determine the gas-in-place, and then deterministic estimates of other parameters to produce a final result" . . .