Auditing of Accounts between the Netherlands and France

CASE CONCERNING THE AUDITING OF ACCOUNTS
BETWEEN
THE KINGDOM OF THE NETHERLANDS AND
THE FRENCH REPUBLIC PURSUANT TO THE ADDITIONAL PROTOCOL OF 25 SEPTEMBER 1991 TO THE CONVENTION ON THE PROTECTION OF THE RHINE AGAINST POLLUTION BY CHLORIDES OF 3 DECEMBER 1976
ARBITRAL AWARD OF 12 MARCH 2004
Unofficial English Translation

Additional Protocol of 25 September 1991 to the Convention of 3 December 1976 on the Protection of the Rhine against Pollution by Chlorides – Annex B to the Convention, on Arbitration – Annex III to the Additional Protocol, on “Financial Arrangements”.
Final auditing of accounts for the storage and removal from storage of residual salts on French territory – Dispute between the Parties as to the interpretation of point 4.2.1 of Annex III to the Additional Protocol on the auditing of accounts – Argument of the Netherlands that running costs must be calculated on the fixed basis of 61.5 French francs per ton of residual salts stored – Argument of France that the said costs must be calculated on the basis of actual costs per unit.
Interpretation of treaties in accordance with international law – Articles 31 and 32 of the Vienna Convention on the Law of Treaties – Customary nature of these provisions.
Constituent elements of the rule in Article 31 of the Vienna Convention making an integral whole – Importance of one element of the rule in relation to the others depends on the particular case – Intention and common will of the parties to be sought objectively and rationally starting from the text – Notion of “text of the treaty” distinct from, and broader than, that of “terms” – Ordinary and particular meaning of terms.
Good faith – Fundamental nature of principle of good faith in treaty interpretation – No reason in this respect to doubt the interpretations put forward by France and the Netherlands.
Point 4.2.1 providing that the accounts shall be audited by “comparing the actual expenditure, calculated according to the conditions specified in points 1.2.3, 1.2.4, and 1.2.6” with the spending limits – Point 1.2.6 referring to a rate of “61.5 French francs … per ton stored” – Ordinary meaning of the terms favours the interpretation of the Netherlands – Phrase “payments are not payments in full discharge” does not support the French position – Meaning of the term “comparaison” and the term “difference” used in point 3.2.3 of Annex III – Meaning of the expressions “depenses engagees” and “depenses effectuees “. Arguments of France on these points not upheld.
Context of point 4.2.1 – Other provisions of the Protocol – Article 4 of the Protocol not conclusive as to whether the audit must be based on 61.5 French francs or on actual costs – Other elements of the context – Declaration of the Heads of Delegation made on 25 September 1991 of limited usefulness.
Object and purpose of the Protocol – Joint nature of the system chosen by the Parties for implementation of the Protocol – Joint financing of the steps to be taken by France and the Netherlands – Identification of the object and purpose leaves open the question of how this financing was to function.
Subsequent practice in implementing the Protocol – Annual reports and provisional balance sheet at the end of 1996, prepared by France, showing the existence of a common practice and agreement as to the interpretation of point 3.2.2 of Annex III alone – Cessation of payments by Switzerland and Germany not determinative as between France and the Netherlands.
Relevant rule of international law applicable to relations between France and the Netherlands – Principle that “polluter pays” not used in Additional Protocol and not part of general international law – Geographical situation of the Netherlands of no relevance.
Resort to supplementary methods referred to in Article 32 of the Vienna Convention in the interpretation process – Analysis of preparatory works showing how the amount of 61.5 French francs was calculated using both fixed and variable elements but adopted by the Parties as a lump sum.
Calculation of final auditing – Determination of surplus amount paid to France – Calculation of interest due up to 31 December 1998 in accordance with point 3.2.3 of Annex III – Calculation of interest falling due after 31 December 1998 – Distinction between storage expenditures and expenditures on removal from storage – Cost overruns for removal from storage resulting from increase in prices after 1998.
THE TRIBUNAL, COMPOSED OF:
HE. KRZYSZTOF SKUBISZEWSKI, President, H E. PIETER KOOIJMANS, Member, H.E. GILBERT GUILLAUME, Member, Ms. ANNE WALLEMACQ, Registrar,
Renders the following Arbitral Award:
Chapter I – Introduction
1.    Following negotiations held within the framework of the International Commission for the Protection of the Rhine against Pollution (hereinafter called the “ICPR”), the governments of the French Republic, the Federal Republic of Germany, the Grand Duchy of Luxembourg, the Kingdom of the Netherlands and the Swiss Confederation (hereinafter respectively called “France”, “Germany”, “Luxembourg”, “the Netherlands” and “Switzerland”) signed the Convention on the Protection of the Rhine against Pollution by Chlorides (hereinafter called the “Convention”) on 3 December 1976 in Bonn. The Convention was amended in 1983 by an exchange of letters initiated by France. Those amendments are not relevant to the present case. On 25 September 1991, in Brussels, the Contracting Parties signed the Additional Protocol to the Convention (hereinafter called the “Protocol”).
The objective of the Convention, as set out in Article 1, is to combat the pollution of the Rhine by chlorides. To do this, the Contracting Parties undertook to strengthen their cooperation. The objectives of the Protocol, as defined in its Preamble, are to improve the quality of the waters of the Rhine, facilitate the supply of drinking water from the Rhine and the IJsselmeer, and reach a definitive solution to the problem of reducing chloride concentrations in the Rhine.
2.    The Convention provided that France must reduce discharges of chlorides from the Alsace Potassium Mines in two phases. The first phase consisted of reducing discharges into the Rhine by 20 kgs of chloride ions per second. In the second phase, in addition to the reduction of 20 kgs already achieved, France was to reduce chloride discharges by 40 kgs per second. These provisions were amended by the Protocol, under which measures were to be taken:
–    in French territory, to reduce chloride discharges and temporarily store the resulting chlorides on land when chloride concentration in the Rhine at the German-Netherlands border exceeded 200 mg/l; and
–    in Netherlands territory, to limit chloride concentrations in the waters of the IJsselmeer by discharging the briny waters from the Wieringermeer polder previously discharged into the IJsselmeer into the Waddenzee.
3.    The Protocol provided that the financing of these measures was to be apportioned as follows: 30% was to be borne by France, 30% by Germany, 34% by the Netherlands and 6% by Switzerland. The parties agreed on maximum expenditures of 400 million French francs for the 1991-1998 period for the measures to be taken in France, including the costs of investment – which were themselves limited to 40 million French francs – and running costs for storage and removal from storage of the chlorides. That amount constituted a spending limit beyond which France was released from its storage obligations.
4. A dispute arose between the Netherlands and France concerning the interpretation and implementation of the Protocol. Pursuant to Article 13 of the Convention, reproduced in Article 7 of the Protocol, “[a]ny dispute between the Contracting Parties relating to the interpretation or implementation of this Convention which cannot be settled by negotiation shall, unless the Parties to the dispute decide otherwise, be submitted, at the request of one of the Parties, to arbitration in accordance with the provisions of Annex B. The latter … forms an integral part of this Convention”. Annex B, entitled “Arbitration”, lays down the composition and constitution of the Arbitral Tribunal and the process by which it must reach its decision, and provides that, with respect to other aspects of the proceedings, the Arbitral Tribunal shall establish its own rules of procedure. The dispute referred to above could not be settled by negotiation.
Chapter II – Procedural History
5.    On 21 October 1999, the Netherlands sent a note verbale from its Embassy in Paris to the French Ministry of Foreign Affairs requesting that, pursuant to Article 13 of the Convention and Article 7 of the Protocol, the dispute between them concerning:
the amount to be repaid by the French Republic to the Kingdom of the Netherlands
under the final auditing of accounts which was to take place not later than
31 December 1998 as provided in paragraph 4.2.1 of Annex III to the Additional
Protocol
be submitted to arbitration.
In its note verbale, the Netherlands stated its intention to appoint an arbitrator within two months of the date France received the note verbale containing the request, in accordance with the provisions of Annex B to the Convention, and invited the French authorities to appoint an arbitrator within the same two-month period in order that the two arbitrators thus appointed could by mutual agreement designate a third arbitrator, who would be President of the Arbitral Tribunal.
6.    By a note verbale from its Embassy in Paris dated 17 December 1999, the Netherlands informed the French Ministry of Foreign Affairs that it had appointed as arbitrator Mr. Pieter H. Kooijmans, Judge of the International Court of Justice.
7.    By a note verbale dated 17 December 1999 from its Ministry of Foreign Affairs, France informed the Embassy of the Netherlands in Paris that it had taken note of the Netherlands’ request for arbitration, and had appointed as arbitrator Mr. Gilbert Guillaume, Judge of the International Court of Justice.
8.    The two arbitrators thus appointed chose as the third arbitrator, by mutual agreement, Mr. Krzysztof Skubiszewski, President of the Iran-United States Claims Tribunal, who on 1 February 2000 agreed to serve as President of the Netherlands-France Arbitral Tribunal (hereinafter called the “Tribunal”).
9.    On 21 June 2000, the duly constituted Tribunal held, in consultation with the Parties, a meeting at the Peace Palace in The Hague. At that meeting, the Tribunal adopted its Rules of Procedure and proposed that the International Bureau of the Permanent Court of Arbitration (hereinafter called the “PCA”) should serve as the Registry of the Tribunal, and that Ms. Bette E. Shifman should serve as Registrar.
10.    The Netherlands having agreed on the same day to the appointment of the PCA as Registry and Ms. Bette E. Shifman as Registrar, and France having so agreed by letter of 1 July 2002, the Tribunal proceeded to make the said appointments. Pursuant to the Tribunal’s Rules of Procedure, all communications from or to the Tribunal from that date on, including all notifications and transmissions of documents, were dispatched through the intermediary of the Registry.
11.    In accordance with the Rules of Procedure, France informed the Tribunal by letter dated 12 July 2000 that it had appointed as its Agent Mr. Ronny Abraham, Director of Legal Affairs at the Ministry of Foreign Affairs. The Netherlands informed the Tribunal by letter dated 20 July 2000 that it had appointed as its Agent Mr. Johan G. Lammers, Legal Adviser to the Ministry of Foreign Affairs.
12.    By letters dated 4 August and 18 August 2000 respectively, the Netherlands and France informed the Tribunal that they had agreed to submit to it the following question:
Taking into account all relevant provisions of the Additional Protocol to the Convention of 3 December 1976 on the Protection of the Rhine against Pollution by Chlorides and its Annexes, interpreted in accordance with international law, what is the sum that must be transferred between the Parties to the dispute in order that the accounts be audited pursuant to point 4.2.1 of Annex III to the said Protocol?
13.    In accordance with the timetable set forth in the Tribunal’s Rules of Procedure, the Netherlands filed its Memorial on 28 February 2001. France filed its Counter-Memorial on 12 September 2001, after which the Netherlands filed its Reply on 11 January 2002. The Rejoinder of France was filed with the Registry on 8 April 2002, thus concluding the written phase of the proceedings.
14.    Following the resignation of the Registrar, the Tribunal appointed, at the suggestion of the Secretary-General of the International Bureau of the PCA and after consulting the Parties, Ms. Anne Wallemacq as its Registrar, with effect from 1 July 2002. On 24 July 2002, Ms. Catherine Cisse was appointed as Assistant Registrar.
15.    Following inquiries as to the Parties’ intentions with respect to a possible oral phase of proceedings, the Parties informed the Registrar, by letters dated 16 April and 20 May 2002, that after consultation they were unable to reach an agreement on whether the proceedings should consist only of a written phase, as allowed by the Tribunal’s Rules of Procedure.
16.    Consequently, and in accordance with the Rules of Procedure, a hearing was held in closed session on 3 October 2002, at the International Conference Centre in Paris. At that hearing, oral arguments and arguments in reply were submitted, with the benefit of simultaneous interpretation, by:
For the Netherlands: Mr. Johan G. Lammers, Legal Adviser, Ministry of Foreign Affairs;
Ms. Ineke van Bladel, Senior Lawyer, Ministry of Foreign Affairs; Mr. Robert H. Dekker, Director of International Water Policy, Ministry of Transport, Public Works and Water Management, and Head of the Netherlands delegation to the ICPR;
For France:    Mr. Ronny Abraham, Director of Legal Affairs, Ministry of Foreign
Affairs;
Professor Serge Sur, Adviser to the French Government.
Other members of the French delegation were:
Mr Besozzi, Rhine-Meuse Water Agency; Ms. Isidoro, Ministry of Foreign Affairs; Mr. Rulleau, Alsace Potassium Mines;
Ms. Taon, Legal Department, Ministry for Sustainable Development and Ecology;
Mr. Vincent, Ministry of the Economy, Finance and Industry.
Each of the Parties filed a note of its oral arguments at the hearings and presented documents that had not previously been produced. The Netherlands submitted a document entitled “Report to the ICPR on the technical aspects, costs and financial arrangements of the plan presented by the Netherlands with the agreement of France as an alternative to the measures envisaged for implementing the Chlorides Convention”, bearing the typewritten classification number C 2/90 rev. 21.03.90 and the handwritten classification number DELch 4/90 (hereinafter called document “DELch 4/90”), a document entitled “Record of Decisions of the 47th Meeting of the Heads of Delegation held at Haarlem on 4 May 1990”, bearing the handwritten classification number PLEN 1/90 (hereinafter called document “PLEN 1/90”), and a document entitled “Record of the Decisions of the 56th Plenary Meeting of the ICPR held on 2 July 1991 at Lenzbourg (Switzerland)”, bearing the classification number PLEN 30/91 (hereinafter referred to as document “PLEN 30/91”). France submitted a document entitled “Cost evaluation for the implementation of adjusted temporary storage in France with effect from 1 January 1991”, bearing the date 10 September 1990 and the handwritten classification number C 8/90 (hereinafter called document “C 8/90-Fr”), a document entitled “Note on the costs of implementation in France of the 1991 Additional Protocol to the Chlorides Convention”, bearing the date 24 March 1997 (hereinafter called the “Note on costs”) and a document entitled “Bonn Convention 2nd Phase – expenditures calculated in F 88” comprising three tables (hereinafter called “Bonn Convention – expenditures calculated in F 88”). After consulting the Parties, and since they had no objections, the Tribunal decided that the said documents would be included in the record. The Members of the Tribunal put questions to the Parties following a plan agreed by the Parties, and the Parties replied to these questions.
17.    On 23 October 2002, the Netherlands presented to the Registry a document entitled “Calculation Note on the implementation in France of the Additional Protocol to the Chlorides Convention” (hereinafter called the “Calculation Note”), to which the Netherlands delegation had referred at the hearing but not produced. On 6 November 2002, the Netherlands presented two other new documents to the Registry. The first of these was a letter dated 14 March 1990 from the Director of the ICPR, bearing the reference 7945/90, which was the covering letter for the preliminary draft of a report summarising the Franco-Dutch alternative for implementing the second phase of the Convention on the Protection of the Rhine from Pollution by Chlorides, bearing the classification number C 8/90, only the title page of which was attached (hereinafter called the “Covering letter of 14 March 1990”). The second document was entitled “Record of Decisions of the Working Group on Chlorides, held on 20 and 21.03.90 in Koblenz”, bearing the classification number C 11/90 (hereinafter called document “C 11/90”).
18.    Having invited France to provide its views on these submissions, the Tribunal issued Order No. 1 on 4 December 2002, based on its Rules of Procedure, authorising the production of the documents presented by the Netherlands and requesting the Parties to explain to it, by 7 January 2003, why the same classification number, C 8/90, appeared on two different documents, one submitted by France at the hearing and the other submitted by the Netherlands on 6 November 2002. The Tribunal further asked the Netherlands to submit to it, no later than 7 January 2003, the text and annexes of the above-mentioned preliminary draft report, and asked France to provide, within the same time period, all additional details concerning the circulation and discussion within the Working Group on Chlorides of the document C 8/90 it had submitted at the hearing.
19.    By a letter dated 30 December 2002, France requested an extension by ten days of the time limit for providing the explanations requested by the Tribunal in Order No. 1. By a letter dated 6 January 2003, the Netherlands filed its replies to the questions put by the Tribunal in Order No. 1, together with the relevant annexes, including the “Report of the Working Group on Chlorides on the alternative Franco-Dutch plan for the implementation of the second phase of the Chlorides Convention” dated March 1990 and bearing the classification number C 8/90 (hereinafter called document “C 8/90-Nl”), without its annexes, and stated that it had no objection to France being granted a ten day extension. By Order No. 2 of 8 January 2003, the Tribunal granted France a ten day extension, until 17 January 2003, to provide the explanations requested in Order No. 1.
20.    France duly delivered its reply within the extended time limit. Annexed to its reply, France submitted the following documents: a document entitled “Temporary storage of salt at the MPDA (Alsace Potassium Mines) during periods of heavy Cl concentration in the Rhine – Technical and Economic Study”, dated 13 November 1989 and marked C 10/89 (hereinafter called document “C 10/89”), a document entitled “Efficiency and cost of various scenarios for regulating the retention of residual salts by French potassium mines – documentation for the Ministerial Conference of States Bordering the Rhine of 30 November
1989    in Brussels”, dated November 1989 and marked C 11/89 (hereinafter called document “C 11/89”), a document entitled “Report to the ICPR on the technical aspects, costs and financial arrangements of the plan put forward by the Netherlands by agreement with France for the implementation of the second phase of the Chlorides Convention”, dated 1 March
1990    and bearing the handwritten classification number C 2/90 (hereinafter called document “C 2/90”), a document entitled “Draft: – Declaration of the Heads of Delegation of the Governments Parties to the Agreement on the International Commission for the Protection of the Rhine against Pollution”, bearing the handwritten classification number DELch 12/90 (hereinafter called document “DELch 12/90”), and its annex entitled “Explanatory Note by the French Delegation on Item 5 of the draft Declaration by the Heads of Delegation”, a document entitled “Dutch Draft Proposal for the subsequent implementation of the Convention on the Protection of the Rhine against Pollution by Chlorides”, dated 14 July 1989 and bearing the classification number DELch 17/89 (hereinafter called document “DELch 17/89”), and a document entitled “Studies for the Bonn Convention Second Phase – Temporary storage of residual salts near the Mines”, dated August 1987 (hereinafter called the “1987 Studies”).
21.    By the above-mentioned letter of 6 January 2003, the Netherlands asked the Tribunal for leave to comment on France’s reply to the questions in Order No. 1, in particular the submission of new documents by France.
22.    By Order No. 3 of 27 January 2003, having confirmed on the one hand that each Party had the right to present its case in full, and recalling on the other that the leave granted to them during and after the hearing to present new documents, and the questions put by the Tribunal in Order No. 1, were not to result in a de facto reopening of the record, the Tribunal authorised each Party, if it so wished, to submit its observations on the other Party’s reply to Order No. 1, including the documents annexed to those replies, by no later than 17 February 2003.
23.    In compliance with the terms of Order No. 3 of the Tribunal, France submitted its observations on the reply of the Netherlands by letter dated 31 January 2003, and the Netherlands submitted its observations on the French reply on 14 February 2003.
24.    By Order No. 4 of 26 February 2003, the Tribunal requested that France provide it, by 31 March 2003, with answers to two questions concerning the document entitled “Provisional Balance Sheet as at the end of 1996 on the implementation in France of the 1991 Additional Protocol to the Chlorides Convention” contained in Annex 8 to the Memorial of the Netherlands (hereinafter called the “provisional balance sheet as at the end of 1996”). The Tribunal further requested the Netherlands to comment, if it so wished, on France’s replies within ten days of receiving them.
25.    By letter dated 28 March 2003, France submitted its replies to the questions put by the Tribunal in Order No. 4. Stating that France’s replies were received on 31 March 2003, the Netherlands informed the Tribunal of its comments on France’s replies by letter of 9 April 2003.
26.    By Order No. 5 of 24 September 2003, the Tribunal requested the Parties to consult together and provide it, by 10 October 2003, with the missing page – only the second page having been supplied – of the annual reports for the years 1994, 1995, 1996 and 1997 annexed to the Memorial of the Netherlands (see annexes 5.3, 5.4, 5.5 and 5.6). Recalling that, under Annex III to the Protocol, work was to continue up to December 1998 and France was bound to submit information as to the quantities stored and the related costs each year, the Tribunal requested the Parties to provide it with the annual report for the year 1998, which the Parties had not produced. The Tribunal also requested France to provide it, by 10 October 2003, with any information useful in determining the applicable interest rate that was to replace the Credit national rate. Finally, having found that no indication had been given as to the coefficients of price increases to be applied for the period after 31 December 1998, the Tribunal requested France to provide it, within the same time period, with any information useful in determining the coefficients of price increases applicable after that date.
27.    After consulting with the Netherlands, France sent the Tribunal the documents requested by letter dated 9 October 2003, in which it suggested that the Tribunal should replace the Credit national rate not by the T 4 M rate, as it had suggested in its Counter- Memorial, since that was a short term rate, but rather by the borrowing rate on French 10-year State bonds. Annexed to its reply, France submitted the annual reports for the years 1994 to 1998 in their entirety, a table showing the progression in the borrowing rate on French 10- year State bonds between 1991 and 2002, and a table summarising the annual progression in the price index from 1988 to 1992.
28.    With reference to the above-mentioned reply of France, and acting at the Tribunal’s request, the Registrar, by letter of 12 November 2003, requested that France provide her with any information useful in determining the rate of interest for the year 2003.
29.    By letter of 18 November 2003, France sent the Tribunal a table summarising the progression of the borrowing rate on French 10-year State bonds for the months of January to October 2003, explaining that these were average monthly rates.
30.    At the Tribunal’s request, by letter of 3 December 2003, the Registrar requested France to provide her, as soon as possible, with the study carried out in April 1988 referred to in document C 10/89 annexed to France’s reply to the Tribunal’s Order No. 1, as well as the definitive rate of interest for the year 2003. In response to this request, France sent a letter to the Tribunal on 15 December 2003 enclosing the 1987 Studies, explaining that, while that document had been compiled in August 1987, the amounts shown in it had been updated in April 1988, though this did not appear on the face of the document. As to the definitive rate of interest for the year 2003, France stated that it would communicate this information at the beginning of January 2004, and it did so by letter on 9 January 2004.
Chapter III – The Arguments of the Parties
1. The Netherlands
31.    The Netherlands contends that the question put to the Tribunal must be examined on the basis of the principles of international law on the interpretation of treaties, which are expressed in Article 31 of the Vienna Convention on the Law of Treaties of 23 May 1969 (hereinafter called “the Vienna Convention”). This means that the interpretation must be made in good faith, following the ordinary meaning of the terms in their context and in the light of the object and purpose of the Protocol. The Netherlands points out that the examination should not proceed on the basis of Article 32 of the Vienna Convention unless, after an examination under the rule laid down in Article 31, the meaning remains ambiguous or obscure, or leads to a result which is manifestly absurd or unreasonable.
32.    The Netherlands recalls, first, that under point 4.2.1 of Annex III to the Protocol:
4.2.1. For expenditure in French territory, payments are not payments in full discharge, and the accounts shall be audited by 31 December 1998 at the latest by comparing the actual expenditure, calculated according to the conditions specified in points 1.2.3, 1.2.4 and 1.2.6 above, with the spending limit set out in point 2 above, increased by any amount carried over pursuant to point 3.2.3 above. In the event that actual expenditure by France is less than the amount referred to in point 1.2, France undertakes to return the surplus received plus 11/12 of the yearly interest on this amount at the annual long-term interest rate on national loans. In this connection, the rate of price rises shall also be taken into account.
It adds that under point 1.2.6 of Annex III:
1.2.6. Each year, France shall be released from its storage obligations once the expenditure in that year has reached the spending limit as determined by point 2 and point 3.2.3. For this purpose, the running costs for France shall be calculated at a rate of 61.5 French francs (1988 French francs adjusted for inflation) per ton stored. For the first year, investment costs shall be added on (40 million 1988 French francs adjusted for inflation).
33.    In the view of the Netherlands, the passages in point 4.2.1 that are particularly important in answering the question before the Tribunal are those mentioning “the actual expenditure, calculated according to the conditions specified in points 1.2.3, 1.2.4 and 1.2.6”, “11/12 of the yearly interest on this amount at the annual long-term interest rate on national loans”, and “the rate of price rises” [“rate of inflation”]. In calculating the expenditures, the interest and the rate of inflation, points 4.2.1, 1.2.6 and 1.2.1 are relevant for the expenditures, points 4.2.1 and 3.2.3 for the interest, and 4.2.1 for the rate of inflation.
34.    According to the Netherlands, the “ordinary meaning to be given to the terms of a treaty in their context” referred to in Article 31 of the Vienna Convention implies that the “expenditures”, “interest” and “rate of price rises” mentioned in point 4.2.1 must be calculated as set out below.
35.    Expenditures, in the meaning of point 4.2.1, comprise the investment costs set at 40 million French francs (1988 French francs adjusted for inflation), an amount which is not in dispute, and the running costs. The running costs, pursuant to point 1.2.1, comprise the costs of storage and removal from storage. Given the relationship between the first and second sentences of point 4.2.1 on the one hand, and point 1.2.6, to which point 4.2.1 expressly refers, on the other, the running costs must be calculated by multiplying the quantities of chlorides actually stored by 61.5 French francs per ton (1988 French francs adjusted for inflation). In other words, the method indicated in the second sentence of point 1.2.6 is meant not only to enable the amount of the expenditures referred to in the first sentence of that paragraph to be calculated, but also to enable the calculation of expenditures by France in the context of the final auditing of accounts. The Netherlands emphasises, moreover, that the French interpretation amounts to breaking down the amount of 61.5 French francs into fixed costs, which must be multiplied by 4.55 million tons of chlorides, and into variable costs, which must be multiplied by the quantity of chlorides actually stored. Thus, the amount of the fixed costs that France could have charged to the other Parties could have been determined as early as 1991. The fact that no such determination was made argues in favour of the Netherlands’ interpretation.
The Netherlands disputes France’s contention that the notions of “depenses engagees” as used in the first sentence of point 4.2.1 and “depenses effectuees” as used in the second sentence are capable of two different meanings.  It draws attention to the fact that the expression “depenses effectuees”, which according to France’s argument refers to actual expenditures, also appears in point 1.2.6, which relates to calculations and carry-overs that are not to be done on the basis of actual costs. Furthermore, the French interpretation would mean giving a specific meaning to the term “depenses” [expenditures], because, according to France, only the actual expenditures of France could be taken into consideration. Finally, in reply to the argument of France that, for the Wierengermeer project, the Netherlands conducted the final audit referred to in point 4.1.1 on the basis of actual expenditures, the Netherlands contends that the Parties had reached agreements on the measures to be taken and the auditing of the accounts by France as well as the Netherlands, and that there is no inconsistency between the provisions of points 4.1.1 and 4.1.2.
In response to the argument of France derived from the use of the term “a cette fin” [“for this purpose”] in point 1.2.6, the Netherlands points out that this expression does not appear in the Dutch or German texts. Referring in this regard to Article 33, paragraph 3, of the Vienna Convention, according to which the terms of a treaty are presumed to have the same meaning in each authentic text, it concludes that the French, Dutch and German texts confirm that its interpretation is correct, while France’s attempt to reconcile the texts when implementing the provision would be tantamount to making the French text prevail over the other two.
In response to the argument of France based on the use of the term “difference” [“difference”] in point 3.2.3 and of the term “comparaison” [“comparison”]  in point 4.2.1, the Netherlands contends that the term “difference””, as it appears in point 3.2.3 dealing with carry-overs, is used correctly, as that point envisages a situation where the actual storage costs incurred are below the limit set for the year concerned. For this reason, according to the Netherlands, no link can be established between point 3.2.3 and the first sentence of point 4.2.1 which states that the accounts shall be audited by a “comparison” between the expenditures “calculated according to the conditions specified in points 1.2.3, 1.2.4 and 1.2.6” and “the spending limit set out in point 2”. Indeed, where the auditing is concerned, it is not possible in practice to start from the hypothesis that the said expenditures are below the spending limit indicated in point 2: there must first be a “comparison”. If it appears that the actual expenditures of France are below the amount shown in point 1.2, France will then be bound to return the surplus received (the “difference”).
36.    The Netherlands goes on to argue that, pursuant to Article 31 of the Vienna Convention, a treaty must be interpreted “in the light of its object and purpose”, and that the Preamble to the Protocol is especially important in this regard. It contends that it is clear from the Preamble that the object and purpose of the Protocol are to improve the quality of the waters of the Rhine and facilitate the supply of drinking water. From this it deduces that its interpretation of the relevant terms of the Protocol in the light of their ordinary meaning is compatible with the object and purpose of the Protocol.
The Netherlands moreover denies having established a relationship between its geographical situation and the “polluter pays” principle, on the one hand, and the object and purpose of the Protocol on the other, or even the meaning that must be given to the terms of the Protocol. France is therefore incorrect, it claims, in alleging that the Netherlands was seeking to use those arguments to obtain preferential treatment.
37.    Referring to paragraph 2 a) of Article 31 of the Vienna Convention, the Netherlands contends that the “Declaration of the Heads of Delegation of the Governments Parties to the Agreement on the International Commission for the Protection of the Rhine against Pollution” (hereinafter called the “Declaration of Heads of Delegation”), signed at the same time as the Protocol, must be considered as an “agreement” within the meaning of that paragraph, and that the said Declaration would not lead to any changes in the interpretation favoured by the Netherlands.
38.    The Netherlands states that in the present case there is no “subsequent agreement between the parties” within the meaning of Article 31, paragraph 3 a) of the Vienna Convention. On the other hand, account has to be taken of the “subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation”, as referred to in paragraph 3 b) of Article 312 of the Vienna Convention. In this regard, the Netherlands emphasises that, in the annual reports submitted by France for the 1992-1997 period, the running costs are calculated by multiplying the quantities of chlorides stored by 61.5 French francs per ton. The interest rates applied in the annual reports are, with the exception of the 1997 rate, the annual long-term interest rates of the Credit national, also taking account of the rate of inflation. The Netherlands concludes from this that the undisputed approach of France in implementing its obligation to provide the information set out in Article 1, paragraph 2 of the Protocol and point 3.2.2 of Annex III to the Protocol must be considered as an additional argument in support of the correctness of its interpretation.
In addition, the Netherlands does not agree that the discussions that took place between the Parties to the Protocol regarding the final auditing can be described as “subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation”. According to the Netherlands, there was no such agreement in practice. Proof of this is to be found in these proceedings, the very object of which is to produce a final auditing. The Netherlands notes, in this regard, that Germany and Switzerland have taken the same view.
39.    The Netherlands also maintains that the interpretation given by it to the calculation of expenditures, interest and the rate of inflation must be considered as an interpretation in good faith of point 4.2.1, point 1.2.6 read with point 1.2.1, and point 3.2.3 of Annex III to the Protocol. It observes that good faith is not an element that can be used after the event to deduce a different, autonomous meaning from the terms of the Convention.
40.    The Netherlands goes on to analyse the preparatory works, making it clear at the outset that it does so for one reason only: that these works can, under Article 32 of the Vienna Convention, serve to confirm the meaning resulting from an application of Article 31.
The Netherlands disputes the relevance of document DELch C 2/90 annexed to the Rejoinder of France, which dates from the spring of 1990, in other words one and a half years before the Protocol was signed. The Netherlands further contends that the said document is incomplete, and that, contrary to what is claimed by France, it was a revised version of that document, headed DELch 4/90, that was discussed by the Heads of Delegation when they met on 4 May 1990. Moreover, as appears from the record of the decisions of that meeting, the revised document had been referred to the Working Groups on Chlorides and on Legal Affairs for their examination. Reference is made in this connection to document PLEN 1/90.
The Netherlands adds that one document that is, however, relevant to the case is document PLEN 30/91, which was drawn up at the end of the negotiations on the Protocol. Of particular importance is paragraph 8.2 of that document, entitled “Details concerning financial arrangements”, as well as Annex 4 to which paragraph 8.2 refers. According to the Netherlands, it appears from that document that France was to base the final auditing of the accounts on the calculation for the year 1998, in which, in line with the reports France had drawn up for the preceding years, the quantities of chlorides stored were multiplied by 61.5 French francs and the amounts carried over from previous years were taken into account. Furthermore, no distinction was made between fixed costs and variable costs. According to the Netherlands, this distinction was discussed for the first time in the ICPR coordination group, when France presented its provisional balance sheet as at the end of 1996, a document France does not dispute having put forward spontaneously, with no obligation. Furthermore, the Netherlands adds, it appears from this last document that certain items were in fact 45% lower than France had calculated. This finding raises the question of whether, in calculating the fixed costs, a 45% reduction for the dismantling of the installations should be factored in, insofar as this calculation is relevant.
The Netherlands also raises doubts as to whether the amount of 61.5 French francs was calculated by taking a hypothetical amount stored of 4.55 million tons, and referring back in this regard to the Calculation Note, from which it emerges that France requested facilities for only half that amount, namely 2.7 million tons.
As to document C 8/90 submitted by France at the hearing in support of its contention that the amount of 61.5 French francs must be broken down into fixed costs of storage and removal from storage, and variable costs of storage and removal from storage, the Netherlands maintains that the document does not form part of the “preparatory works” of the Protocol. Indeed, it was not produced by the ICPR but by a French body, and was discussed neither at the meeting of the Working Group on Chlorides on 13 September 1990 nor at the meeting of the Heads of Delegation on 31 October 1990. In support of this statement, the Netherlands presented another document also bearing the classification number C 8/90, which this time had been drafted within the framework of the ICPR. It explains, however, that this latter document was not discussed by the Working Group on Chlorides, nor was it part of the “preparatory works”.
The Netherlands points out that the documents C 10/89, C 11/89, C 2/90, DELch 12/90, DELch 17/89 and the 1987 Bonn Convention Studies, attached as annexes to France’s reply to the questions put by the Tribunal in Order No. 1, date from a very early stage of the negotiations and were not discussed in the Plenary Session as the possible basis of a decision. In other words, these were unilateral proposals from the Parties and discussion documents, the contents of which could not be taken to reflect the intentions of the Parties as to the interpretation of point 4.2.1 or other provisions of the Protocol or Annex III. As to the breakdown into fixed costs and variable costs, the Netherlands adds that while the Parties had necessarily talked about the cost structure required by the measures to be taken, this did not mean that they wanted that structure to be taken into consideration in the final auditing. France was therefore wrong to claim that the above-mentioned documents related to negotiations “leading to the adoption of the Protocol”.
The Netherlands also offers the following comments. The 1987 Studies formed part of an overall Plan for the execution of the second phase of the Convention, a plan which had been found unacceptable by the Parties, casting doubt on its value as part of the preparatory works. Moreover, it is of no avail to France to refer to paragraphs 51 to 54 of document C 10/89 to support its argument that a distinction must be drawn between fixed costs and variable costs. Those paragraphs do not in fact mention “fixed costs”, but rather proportionate expenditures, which are also called running costs to distinguish them from investment costs. Finally, the Netherlands emphasises that document C 2/90, on page 10 and page 3 of its Annex, refers to document C 12/89, which was not submitted by France, confirming the thesis of the Netherlands whereby the “actual expenditures” or “expenditures incurred” correspond to the costs mentioned by France in the annual reports that must be used in the final auditing. Indeed, this document states: “For expenditures in French territory, payments are not payments in full discharge, and the accounts shall be audited by 31 December 1998 at the latest by comparison with the actual expenditures, for which the French delegation shall, each year, under the ICPR, submit a detailed report on the implementation of the plan, showing in particular the quantities of salt stored and the expenditures incurred”.
41. On the basis of the foregoing, the Netherlands submits that expenditures must be calculated according to the following methods for the purposes of the final auditing.
The running costs, comprising the costs of storage and subsequent removal from storage, must be calculated by multiplying the quantities of chlorides stored, being 960,000 tons, by 61.5 French francs per ton. This would amount to running costs of 59.04 million French francs. The fixed investment costs of 40 million French francs must also be added to this amount. The expenditures incurred would therefore amount to 99.04 million French francs.
According to the apportionment formula set out in Article 4 of the Protocol, 34% of this amount, or 33.67 million French francs, is to be borne by the Netherlands. Since France has received a total of 135.7 million French francs in payments from the Netherlands, which corresponds to 34% of 400 million French francs, France must reimburse the Netherlands the sum of 102.33 million French francs.
Furthermore, France must pay interest on the amounts carried over each year, on the surplus received as from 1 January 1999, being 102.33 million French francs, and on the amounts reserved for removal from storage after 31 December 1998, the interest being 11/12 of the year at the annual long-term interest rate on national loans. Finally, the rate of inflation affecting the costs of storage and subsequent removal from storage must be taken into account.
42.    The Netherlands requests the Tribunal to take these items into account when calculating the amount that must be reimbursed to it by France in the final auditing.
2. France
43.    France takes the view that the codification of the rules of customary law governing the interpretation of treaties in Articles 31 to 33 of the Vienna Convention expresses the state of the customary law in force, so that the rules it sets out apply to all States, including France, which is not a party to the Convention. However, France explains, the fact that these rules are applicable as a matter of customary law does not mean that the Vienna Convention can be applied with the same kind of minute and analytical rigour as would be the case if it were itself binding as between the Parties. In the present case, only customary international law is binding on the Parties. France adds that the interpretation of these rules should not be rigidly confined within the compass of those articles, and that, furthermore, the reference to the text of the treaty that the Netherlands attempts to present as the essential basis for their interpretation does not have, under international law, the exclusive character that the Netherlands seeks to give it.
In addition, as appears from the wording of Article 31, paragraph 1 of the Vienna Convention, “the ordinary meaning to be given to the terms of the treaty” must be taken in a wider context to include good faith, the text of the treaty as a whole, the context of the treaty, and its object and purpose. Good faith, in particular, is the guiding principle that governs the interpretation and application of treaties. France emphasises that this principle requires that there be a reasonable interpretation of the Protocol and its Annexes.
Finally, France states, it is important, as recalled in Article 32 of the Convention, that the meaning arrived at does not lead to a result that is ambiguous or obscure, or even manifestly absurd or unreasonable.
44.    According to France, the object and purpose of the Protocol, as can be seen from the circumstances in which it was signed, was to promote solidarity among the States bordering the Rhine, each of which has an equal interest in the quality of its waters, and the activities of which contribute in their different ways to the pollution of the river. In concrete terms, this solidarity takes the form of actions taken by each of the States bordering the river on its own behalf, and also of collective financing of necessary measures. France concludes from this that to impose a greater burden on one of the parties than the one it must assume on the basis of the agreed provisions would be to undermine the solidarity the Protocol has put in place, and thus to disregard its object and purpose. In other words, the solution to the dispute cannot, France states, be based on preferential treatment of the parties, in this case the Netherlands, on the grounds that, in accepting the terms of the Protocol, that country had derogated from the “polluter pays” principle. In this regard, France explains that it made both a technical and material contribution – there having been practical reasons why the storage on land was done in France – and a financial contribution that was in no degree less than that of the other parties, and that it was in no sense favoured by the Protocol. The Protocol, incidentally, did not impose any abnormal burden on the Netherlands.
45.    France maintains that all the financial terms of the Protocol and its Annexes are “relevant terms” as referred to in the question put to the Tribunal, and that, in this respect, Article 4 of the Protocol is essential.
46.    According to France, the terms of Article 4 have a dual purpose. On the one hand, they set out the spending limits for the territories of France and the Netherlands, and on the other they show how these expenditures shall be apportioned between the parties to the Protocol, with France bearing 30% of the expenditures, Germany bearing 30%, the Netherlands bearing 34% and Switzerland bearing 6%.
Furthermore, in France’s view, Article 4 has two aspects. The first part of the reasoning was to fix spending limits that must not be exceeded, with an overall ceiling of 400 million French francs, and intermediate limits set for each year of storage according to the apportionment formula set forth in point 2.1.1 of Annex III. In determining under what conditions the limits would be reached, a lump sum of 61.5 French francs per ton stored – which also includes a provision for the removal from storage – permits the regularisation of the Protocol’s operation as a function of the spending limits it sets out. However, the way these expenses were calculated was no more than as a provisional lump sum. This is proved by point 4.2.1 of Annex III, which provides that these payments “are not payments in full discharge”, and that the accounts were to be “audited” when the contemplated operations were completed, and no later than 31 December 1998. The second aspect was to lay down the terms on which the auditing was to be carried out, this being a different notion involving a distinct process, one that was retrospective and final. Here, the calculation based on a lump sum of 61.5 French francs was no longer to be used. On the contrary, it was to be based on actual expenditures, without which the notion of auditing, as distinct from prior payments and annual carry-overs, would be meaningless.
47.    France goes on to state that a mere reading of the terms of Annex III, and point 4.2.1 in particular, in accordance with their ordinary meaning shows that, contrary to what the Netherlands contends, the amount of 61.5 French francs cannot, as such, be the amount used in the final calculation. In support of this position, France puts forward the following arguments.
France emphasises that the amount of 61.5 French francs appears only in point 1 of Annex III, which deals with “Spending limits”. The terms of point 1.2.6, and in particular the use of the words “a cette fin” [“for this purpose”], unambiguously show that the said amount has a meaning specific to the context of the workings of the Protocol from one year to the next, and that its sole purpose was to establish whether the spending limit set for each year had been reached, and whether, as a result, France was discharged from its storage obligations. In this respect, France refers to paragraph 4 of Article 33 of the Vienna Convention, which provides that, where a comparison of authentic texts discloses a difference of meaning, “the meaning which best reconciles the texts, having regard to the object and purpose of the treaty, shall be adopted”. France considers that it is not inconsistent with the other authentic texts to take into consideration the restriction that those words import into the French version of the Protocol. In fact, nothing changes in the way point 1.2.6 is applied, and its purpose is to limit the transposition of the calculation mechanism used in point 4.2.1, which is not incompatible with the German or Dutch texts, both of which draw a distinction between the year-to-year implementation and the final adjustment. On the other hand, to treat these words as having no meaning, as the Netherlands does, amounts to regarding them as never having been written, which is contrary to the correct interpretation of the treaty as it makes it impossible to reconcile the different texts.
France points out, furthermore, that the same language is not used in point 4.2.1 as in point 3.2.3, which deals with annual carry-overs, calculated on the basis of the lump sum of 61.5 French francs. Point 3.2.3 speaks of the “difference” between the storage costs (based on 61.5 French francs) and the amounts previously paid by the parties, a term referring back to a mathematical calculation or result the purpose of which was to ensure the provisional implementation of the Protocol. By contrast, the term “comparison” used in point 4.2.1 refers “to a method or procedure the outcome of which cannot be given in advance as an exact figure, because it will depend upon the evaluation of final data”: what is more, while points 3.2.3 and 4.2.1 both refer to the same rate of interest having to be applied to the amount corresponding to the “difference” or the “surplus received”, point 4.2.1 also mentions the rate of inflation, which underlines the fact that these are different processes to which different methods apply.
Finally, according to France, the choice of terms in point 4.2.1 confirms the conclusion that this provision refers to actual expenditures and not lump sums. Reference is made in the first sentence to “depenses engagees” calculated at a lump sum rate on the bases defined for temporary year-to-year application, and in the second sentence, which sets out the conditions for reimbursement of the surplus received, reference is made to “depenses effectuees”, calculated according to their actual amount and corresponding to the “costs of the measures taken” referred to in Article 4 of the Protocol.  Moreover, the terms “depenses effectuees” do not have the same meaning in points 1.2.6 and 4.2.1, because they are used in different contexts. In point 1.2.6, these terms refer to the quantities actually stored, but not their actual cost, whereas in point 4.2.1 they refer not only to the quantities stored, but also to the costs incurred. Finally, the terms “depenses effectuees” also appear in point 4.1.1, which deals with the auditing of accounts for the Netherlands, which, because the amounts it actually spent had exceeded the spending limit, considers that it is under no obligation to repay any surplus.
The inescapable conclusion, according to France, is that Annex III is based on two strands of financial reasoning, which clearly appear if the text is taken in its ordinary meaning. The first strand, as used in points 1, 2 and 3, has to do with the year-to-year functioning of the Protocol: it is based on a lump sum, which is provisional. The second, which appears in point 4.2.1, is relevant only to the final auditing, which is based on actual quantities and costs corresponding to expenses incurred. And, as Article 4 of the Protocol makes clear, the costs of the measures taken must be apportioned according to the formula laid down therein.
48.    As to subsequent practice in the implementation of the treaty, France maintains that the suspension of payments by certain parties, the financial statements presented by France and the fact that no legal dispute had crystallised prior to this case must be taken into account.
The suspension of payment of contributions by certain parties, which had not elicited any protest on the part of France, shows, in France’s opinion, that the lump sum based system of operation for Annex III could not continue. This was due to the increased coefficient of flow of the Rhine and the fact that the true costs could not be ignored for much longer. The contributions paid had in fact been sufficient to cover the expenses for which France was responsible, because of the annual carry-overs.
The mere fact that the calculations appearing in the provisional reports submitted by France were based on an amount of 61.5 French francs per ton does not mean that France agreed to the interpretation upheld by the Netherlands, nor is there is any question here of estoppel. In reality, these are provisional calculations, which cannot bind France as far as the final auditing is concerned. Moreover, these are documents of a technical nature, which cannot be imputed to any authority capable of binding the French Republic internationally.
France claims, finally, that no conclusions can be drawn from the positions the Parties respectively took before the dispute was submitted to arbitration. Those positions had no legal character; still less were they legally binding on the Parties. On the contrary, these discussions show that the Parties had disagreed all along about how the auditing should be carried out.
49.    France considers that the interpretation by the Netherlands mixes two distinct processes, wrongly assimilates terms that are in formal respects different, ignores certain significant elements of the subsequent practice and distorts others, disregards the rule that limits the amount to be borne by France to 30% of the cost of the measures taken, and leads to a result that is unreasonable. France adds that the interpretation put forward by the Netherlands would result in the imposition of a financial burden on France corresponding to 75% of the costs, and making it alone bear the consequences of a level of coefficient of flow of the Rhine greater than that which the parties originally predicted. Such a result would be in contradiction of the terms of the Protocol, its object and purpose, and the subsequent practice of the parties in its application. In this regard, France explains that the amount of 61.5 French francs was based on hypothetical levels of storage of 4.5 million tons, which was considerably greater than the amounts actually stored, as a result of a coefficient of flow of the Rhine that was greater than had been predicted. Consequently, the amount of fixed costs was in fact increased by the decrease in actual storage. In this regard, France refers to the table annexed to the Note on Costs presented at the hearing, noting that this document was itself annexed to the provisional balance sheet as at the end of 1996, submitted by France in 1997.
50.    France also refers to the preparatory works, simply in order to confirm the meaning resulting from the application of Article 31 of the Vienna Convention. First, France refers to document DELch C 2/90 annexed to its Rejoinder, which describes an alternative plan allegedly put forward by the Netherlands, with France’s agreement, at the meeting of the
Heads of Delegation of 4 May 1990. According to France, that document confirms its position on three different levels: by distinguishing between a fixed limit and variable actual expenditures; by distinguishing between the operation of the Protocol based on lump sums, and a final auditing based on actual expenditures; and by using a forecast quantity of chlorides to be stored for the calculation of the lump sum.
In support of its argument whereby the amount of 61.5 French francs must be broken down into fixed costs and variable costs, France puts forward document C 8/90-Fr, dated 10 September 1990, which shows a distinction between the costs of fitting out the storage areas, which according to France are fixed costs, and the costs of storage, which are variable. France explains that this document did not come from the ICPR archives, but rather from the archives of the Rhine-Meuse Water Agency, and that it had been circulated as necessary to enable it to be discussed. France adds that document C 8/90-Nl of March 1990, presented by the Netherlands, is, compared to the French document, in the form of a summary, because in addition to the contents of the French document C 8/90, it contains the measures recommended by the Netherlands for its territory. It is France’s view that the September 1990 document reflects the outcome of prior discussions. In order to shed more light on its contents, France puts forward documents C 10/89, C 11/89 and the 1987 Studies.
According to France, Document C 10/89, which was examined by the “Chlorides” Group on 15 December 1989, and in particular paragraphs 51 to 54, shows that it had already been planned at the negotiation phase of the Protocol that the amount of 61.5 French francs would be broken down into fixed and proportionate costs. France adds that the reason the costs of removal from storage are not specifically broken down in that document is that they used the figures appearing in the 1987 Studies, which were taken as known. Furthermore, France says, document C 10/89 was restated in document C 2/90, this time with the inclusion of the measures envisaged for the territory of the Netherlands, and these documents were examined at the meetings of the “Chlorides” Group on 20 and 21 March, 23 May, 13 and 14 September and 22 November 1990, and discussed at the meetings of the Heads of Delegation on 3 and 4 May in Haarlem, 31 October in Koblenz and 12 December in Luxembourg, in different versions.
Document C 11/89, drawn up by the Netherlands, shows that different scenarios were studied based on variations in the coefficient of flow of the Rhine. Document C 10/89 also pointed out that “the forecast amount will depend on the hypotheses used for the coefficient of flow of the Rhine during the years of storage”.
The 1987 Studies form part of the overall Plan presented by France to achieve a reduction of 60 kg of chloride ions per second, which was rejected. Nevertheless, document C 10/89 makes extensive reference to the 1987 Studies for estimating the costs of storage and removal from storage, notably in its preamble and especially in Annex 9.
France concludes from these three documents that the breaking down of the amount of 61.5 French francs and the variability of the proportional costs as a function of the coefficient of flow of the Rhine were a constant factor in the negotiations of the Protocol. On this basis, the French document C 8/90 of 10 September 1990 cannot be excluded from the relevant “preparatory works” because the price calculation elements it contains were, as the above- mentioned documents show, discussed on numerous occasions.
In conclusion, France draws attention to document DELch 12/90, in particular to its Annex and the heading “Investments” in the section devoted to cost estimates. According to France, it appears from that heading that the fitting out of the storage areas was valued on the basis of an estimate of a given quantity of salts being stored, with the result that the proportionate costs included the forecast costs of the storage areas.
51.    Based on the foregoing considerations, France explains that the method to be used for the final auditing must take into account France’s “depenses effectuees” (“actual expenditures”), which are composite in terms of their nature, how they were spread out over time, whether they were fixed or variable, and how they were adjusted as a function of inflation and the interest rate applicable to the surplus received. Furthermore, in the terms of Article 4 of the Protocol, “the cost of the measures taken in French territory” was to be apportioned between the Parties, with 30% to be borne by France. Basing itself on this method, France proposes to calculate the amount of the surplus received as follows.
In order to determine the running costs, France starts from the calculations used to justify the choice of 61.5 French francs as the lump sum. While this figure, when used in point 1.2.6, is a lump sum based on forecasts, in point 4.2.1 it applies to actual expenditures, the amount of which must be determined as accurately as possible. For this reason, a distinction must be made between the four separate elements in this amount: fixed storage costs: 12 francs per ton; variable storage costs: 12 francs per ton stored; fixed costs of removal from storage: 17.8 francs per ton; variable costs of removal from storage: 19.7 francs per ton removed from storage. Furthermore, the lump sum value placed on each of these elements was arrived at on the hypothesis that 4.55 million tons would be stored. Thus, France makes a distinction between the fixed costs of storage and removal from storage, and the variable costs of storage and removal from storage. The first two, which do not depend on the quantities actually stored, were calculated on the basis of the initial hypothesis of a quantity of 4.55 million tons, resulting respectively in the following amounts: 12 francs x 4.55 million = 54.6 million francs for the fixed costs of storage and 17.8 francs x 4.55 million = 81 million francs for the fixed costs of removal from storage. The second two, derived from the tonnage actually stored under the Protocol, namely 960,199 tons, are calculated by multiplying the corresponding portion of the lump sum by the number of tons stored, resulting in the following amounts: 12 francs x 960,199 = 11.5 million francs for the variable costs of storage, and 19.7 x 960,199 = 18.9 million francs for the variable costs of removal from storage. This results in running costs amounting to 54.6 + 81 + 11.5 + 18.9 = 166 million French francs. The fixed investment costs of 40 million French francs must also be added to this amount. As a result, the amount of actual expenditures by France is 206 million French francs at 1988 rates.
An inflation rate, arrived at by taking the official national rate of inflation in France for the period in question as a reference, must be added to the amount of actual expenditures. According to France, this results in an updated amount of 234,919,608 French francs. France must bear 30% of this amount, and the Netherlands 34%, which according to France corresponds to 79,921,104 French francs. Since France has received payments from the Netherlands totalling 135,714,883 French francs, the amount that France is obliged to reimburse to the Netherlands is 55,793,779 French francs. Interest must be added to this amount at the rate to be fixed by the Tribunal.
52.    In reply to the questions raised by the Tribunal in Order No. 4 concerning expenditures incurred for the purposes of fitting out the storage areas and the costs of removal from storage mentioned in Chapters II B 2 and II B 4 of the provisional balance sheet as at the end of 1996, France confirmed that the amounts given in that statement were correct. It confirmed in particular that the fitting out of the storage areas had only been partially completed, and that the cost of dismantling the installations and restoring the sites to their former state would be less than initially forecast. France explained that this had been factored into the calculations presented in its submissions and reproduced above, and it reaffirmed its submissions.
53. For all these reasons, France requests that the Tribunal answer the question before it in accordance with the above submissions. In addition, France refers to Article 14 of the Tribunal’s Rules of Procedure, which provides that the Tribunal may “appoint experts, question witnesses and ask the agents, counsel and experts of the Parties for explanations”, and places itself in this regard in the hands of the Tribunal.
Chapter IV – The Interpretation of Treaties in Accordance with International Law
54.    The question the Netherlands and France agreed to submit to the Tribunal (see paragraph 12 above) refers to “the relevant terms of the Additional Protocol … interpreted in accordance with international law”. Before interpreting the terms of the Additional Protocol and its annexes, the law applicable to treaty interpretation should first be recalled.
55.    The Parties invoke Articles 31 and 32 of the Vienna Convention, which provide:
Article 31: General rule of interpretation
1.    A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.
2.    The context for the purpose of the interpretation of a treaty shall comprise, in addition to the text, including its preamble and annexes:
(a)    any agreement relating to the treaty which was made between all the parties in connection with the conclusion of the treaty;
(b)    any instrument which was made by one or more parties in connection with the conclusion of the treaty and accepted by the other parties as an instrument related to the treaty.
3.    There shall be taken into account, together with the context:
(a)    any subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions;
(b)    any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation;
(c)    any relevant rules of international law applicable in the relations between the parties.
of the Vienna Convention, the ordinary meaning of its terms and its object and purpose, as well as the practice of the contracting parties.
57.    While recognising that Articles 31 and 32 constitute the customary law in force, France raises the possibility that the customary rules might differ somewhat from the provisions of the Vienna Convention in their application or content. As France is not a party to the Convention, it submits that “in the present case, it is only international custom that is the law as between the two parties” (see Counter-Memorial, paragraph 6). France also states that interpretation cannot be narrowly confined within the limits of Articles 31 and 32 of the Vienna Convention, and that other provisions of that Convention also have a bearing on the interpretation.
58.    The Tribunal finds that the Parties are, in general terms, in agreement as to the rules applicable to the interpretation of the Protocol. Both Parties consider that Articles 31 and 32 of the Vienna Convention reflect the customary rules on treaty interpretation, even though France advances that this view should be somewhat nuanced (see paragraph 43 above).
59.    The Tribunal notes that no further discussion is necessary on the question of whether the rules of interpretation set forth in Articles 31 and 32 of the Vienna Convention apply in the present case. The Tribunal further notes that the International Court of Justice and other tribunals have affirmed on a number of occasions that these provisions are a codification of customary law.
60.    In the case of Oil Platforms (Islamic Republic of Iran v. United States of America), Preliminary Objections (I.C.J. Reports 1996, (II), p. 812, para. 23), the International Court of Justice recalled that:
according to customary international law as expressed in Article 31 of the Vienna Convention on the Law of Treaties of 23 May 1969, a treaty must be interpreted in good faith in accordance with the ordinary meaning to be given to its terms in their context and in the light of its object and purpose. Under Article 32, recourse may be had to supplementary means of interpretation such as the preparatory work and the circumstances in which the treaty was concluded.
This passage reproduces similar holdings made on previous occasions, and it has subsequently been reiterated in a number of cases, for example: Territorial Dispute (Libyan Arab Jamahiriya/Chad), I.C.J. Reports 1994, p. 21, para. 41; Maritime Delimitation and Territorial Questions between Qatar and Bahrain (Qatar v. Bahrain), Jurisdiction and Admissibility, I.C.J. Reports 1995, p. 18, para. 33; I.C.J. Reports 1996 (II), para. 23; Kasikili/Sedudu Island (Botswana/Namibia), I.C.J. Reports 1999, p. 1059, para. 18; LaGrand (Germany v. United States of America), I.C.J. Reports 2001, p. 501, para. 99; Sovereignty over Pulau Litigan and Pulau Sipadan (Indonesia/Malaysia), Judgment of 17 December 2002, para. 37.
61.    The Tribunal would also add that arbitral awards or decisions along these lines were handed down after the adoption of the Vienna Convention. In its decision of 16 May 1980 in the case of The Agreement on German External Debts, the Arbitral Tribunal set up under that Agreement based its reasoning on the Vienna Convention, despite the fact that it was not applicable by virtue of Article 4 of the Agreement, stating that:
following the opinions expressed in case-law and doctrine, which are in agreement, the Vienna Convention reflects the situation at that time as it does the current situation in public international law with regard to treaties, as, at least where treaty interpretation is concerned, it does no more than codify the customary law in force. Not only have all the parties to the present case expressly adopted that view, but the Tribunal itself has already expressed it in its previous decisions.
(Revue generate de droit international public (hereinafter “RGDIP”), vol. 84, 1980, p. 1184, para. 18 of the reasons).
The Tribunal notes that the Iran-United States Claims Tribunal has systematically referred to Articles 31 and 32 as applicable law since it began its work in 1981, in spite of the fact that neither Iran nor the United States are parties to the Convention (Iran v. United States of America, Case A/1, Iran-United States Claims Tribunal Reports, vol. 1, p. 190; cf. ibid, p. 200 and Case A/2, p. 109). The number of such references in the jurisprudence of that Tribunal is considerable. As an example, in The Islamic Republic of Iran v. The United States of America (Case A/21) (decision of 4 May 1987), the Claims Tribunal stated (ibid., vol. 14, p. 328, para. 8):
[t]he task of the Tribunal is to ascertain the nature and content of the obligations undertaken by the respective States Parties to the Algiers Declarations. The means to be employed in the process of interpretation of an international agreement of this nature are set out in the Vienna Convention on the Law of Treaties.
Finally, the present Tribunal recalls that in its Award of 14 February 1985, the Arbitral Tribunal constituted to decide the Delimitation of the Guinea/Guinea Bissau Maritime Boundary case held that “it is not disputed between the two States Parties that, while neither of them is a party to the Vienna Convention … , Articles 31 and 32 of that Convention constitute the relevant rules of international law … ; the Tribunal can do no other than base its decision on those Articles .. ,.”  In fact, those provisions are applicable “as an international custom recognised between States” (RGDIP, vol. 89, p. 503, para. 41).
62.    It can thus be seen that international jurisprudence has adhered to the general rule of interpretation codified in Article 31 of the Vienna Convention. The Tribunal considers that this rule should be viewed as forming an integral whole, the constituent elements of which cannot be separated. Moreover, this is the approach that is now taken by the International Court of Justice and by certain international arbitral bodies. All the elements of the general rule of interpretation provide the basis for establishing the common will and intention of the parties by objective and rational means.
63.    In the Territorial Dispute (Libyan Arab Jamahiriya/Chad) case, the Court stated that “interpretation must be based above all upon the text of the treaty” (para. 41, cited above). In this regard, the Tribunal emphasises that the “text of the treaty” is a notion distinct from, and broader than, the notion of “terms”. Relying on the text does not mean relying solely, or mainly, on the ordinary meaning of the terms. Such a solution would effectively ignore the references to good faith, the context, and the object and purpose of the treaty. The ordinary meaning of the terms is even itself determined as a function of the context, object and purpose of the treaty. Lastly, as paragraph 2 of Article 31 of the Vienna Convention provides, the text of the treaty (including the preamble and annexes) is itself part of the context for the purposes of interpretation.
64.    It is clear from what has just been said that the Parties’ divergent positions as to the applicable rules of interpretation mostly concern the importance to be attached to each of the various elements of the general rule of interpretation. For the Netherlands, the determining element is the ordinary meaning of the terms used, while for France, these terms are only the point of departure of the process of interpretation; other factors must also be taken into account such as good faith, the context of the provision interpreted, and the object and purpose of the treaty. The Tribunal recalls that all these elements are mentioned in Article 31 and that they are interrelated. According to the International Law Commission, the heading of Article 31 uses the singular form (“rule”) because the provisions of this article “form a single, closely integrated rule” (ILC Yearbook, 1966, vol. II, p. 239). It follows that factors other than the ordinary meaning of the terms cannot be cast aside; in the same way, neither can that meaning be ignored. While the text of the treaty remains the basis of interpretation, this does not mean that preference should be given to a purely literal interpretation. This applies equally to the logic and the economy of drafting of Article 31 of the Vienna Convention, accepted by both Parties for the purposes of this case. The ordinary meaning of the terms must be determined in good faith, in the light of the context, as well as of the object and purpose of the treaty. The importance of one element in relation to the others will of course vary depending on the case. As the Arbitral Tribunal stated in the Lac Lanoux (Spain/France) case (Award of 16 November 1957, United Nations Recueil des Sentences Arbitrales (hereinafter “UNRSA”), vol. XII, p. 281), international law “does not sanction any absolute and rigid method of interpretation”.[7] 65.    In this context, the Tribunal emphasises that it fully recognises the fundamental role of good faith and how it dominates the interpretation and application of the entire body of international law, not only the interpretation of treaties. The fundamental rule of pacta sunt servanda rests on the principle of good faith. This does not alter the fact that when interpreting a treaty, the text must always be taken as the starting point. It is the duty of the Tribunal to do so here.
66.    Consequently, the Tribunal will apply the provisions of the first paragraph of Article 31 concerning “the ordinary meaning to be given to the terms of the treaty”. Nonetheless, the ordinary meaning of the terms is not always the only means of interpretation that is relevant.
67.    The Tribunal notes that the ordinary meaning of the terms does not always have to be adopted. In fact, according to Article 31, paragraph 4, “[a] special meaning shall be given to a term if it is established that the parties so intended”. It is however for the party invoking such a meaning to make a convincing case for it, as is clear from the judgment of the Permanent Court of International Justice in the case concerning the Legal Status of Eastern Greenland (P.C.I.J. Series A/B No. 53, p. 49), the Advisory Opinion of the International Court of Justice on the Western Sahara (I.C.J. Reports 1975, p. 53), and the judgment by a
Chamber of the Court in the Land, Island and Maritime Frontier Dispute (El
Salvador/Honduras: Nicaragua intervening) (ibid, 1992, p. 585, para. 377).
*
68.    The Tribunal must also examine the relationship between Articles 31 and 32 of the Vienna Convention. The Netherlands argues that recourse may only be had to the supplementary means of Article 32 where the application of the means in Article 31 leaves the meaning ambiguous or obscure, or leads to a result which is manifestly absurd or unreasonable (see Memorial, paragraph 3.9). That would not be the case here. France, on its part, submits that the provisions of Article 32 “cannot be dissociated from the general rules of interpretation”, and that they “in a sense constitute the criteria for testing whether an interpretation is correct” (see Rejoinder, paragraph 7).
69.    The Tribunal has already referred to the jurisprudence of the International Court of Justice, in which the Court acknowledged the customary nature of the provisions of Article 32 (paragraph 60, above). The Tribunal has also referred to arbitral decisions (paragraph 61 above). Furthermore, in the case of the Arbitral Award of 31 July 1989 (Guinea-Bissau v. Senegal) (I.C.J. Reports 1991, p. 69, para. 48), the International Court of Justice paid particular attention to the role played by the supplementary means in the interpretation process. Citing its own decisions, the Court stated:
the first duty of a tribunal which is called upon to interpret and apply the provisions of a treaty, is to endeavour to give effect to them in their natural and ordinary meaning in the context in which they occur. If the relevant words in their natural and ordinary meaning make sense in their context, that is an end of the matter. If, on the other hand, the words in their natural and ordinary meaning are ambiguous or lead to an unreasonable result, then and then only, must the Court, by resort to other methods of interpretation, seek to ascertain what the parties really did mean when they used these words. (Competence of the General Assembly for the Admission of a State to the United Nations, Advisory Opinion, I.C.J. Reports 1950, p. 8).
The rule of interpretation according to the natural and ordinary meaning of the words employed
is not an absolute one. Where such a method of interpretation results in a meaning incompatible with the spirit, purpose and context of the clause or instrument in which the words are contained, no reliance can be validly placed on it (South West Africa, Preliminary Objections, Judgment, I.C.J. Reports 1962, p. 336).
These principles are expressed in Articles 31 and 32 of the Vienna Convention on the Law of Treaties, which, in many respects, can be considered as a codification of the existing customary law on this point.
As to arbitrations dating from after the Vienna Convention, one decision that can be cited is that of the Iran-United States Claims Tribunal of 19 December 2000 (not yet published) in The United States of America and The Federal Bank of New York v. The Islamic Republic of Iran and Bank Markazi Iran (Case A/28), in which that Tribunal stated that it “had interpreted paragraph 7 [of the Algiers Declaration of 19 January 1981] in accordance with the rules laid down in Articles 31 and 32 of the Vienna Convention”. Given the jurisprudence of the Claims Tribunal with regard to the Vienna Convention (paragraph 61 above), there can be no doubt that the Claims Tribunal was, in that case, referring to the rules in Article 32 as customary norms.
70.    The Tribunal notes that Article 32 does not restrict the use of supplementary means of interpretation to cases in which the result of the application of the provisions of Article 31 would be ambiguous, obscure or manifestly absurd or unreasonable. Recourse may in fact be had to these means in order to “confirm the meaning resulting from the application of Article 31”. This is in line with several judgments of the International Court of Justice.
71.    In the Territorial Dispute (Libyan Arab Jamahiriya/Chad) case (cited above, para. 55) and the Maritime Delimitation and Territorial Questions between Qatar and Bahrain (Qatar v. Bahrain), Jurisdiction and Admissibility case (cited above, para. 40), the Court declared that while it was not strictly necessary to have recourse to them to determine the meaning of the treaties in dispute, the consideration of supplementary means, including the preparatory works, confirmed the conclusions reached by the Court. The Court also resorted to the supplementary means of Article 32 to confirm its interpretation in the Oil Platforms (Islamic Republic of Iran v. United States of America), Preliminary Objection case (cited above, para. 29), Kasikili/Sedudu Island (Botswana/Namibia) (cited above, para. 46) and the Case concerning Sovereignty over Pulau Litigan and Pulau Sipadan (Indonesia/Malaysia) (cited above, para. 53). In its judgment in the LaGrand (Germany v. United States of America) case (cited above, para. 104), the Court explained that, while it was not necessary to refer to the preparatory works to determine the meaning of the article in question, those preparatory works did not contradict the conclusion the Court had reached.
72.    As to arbitral practice, in the case concerning the Air Services Agreement of 27 March 1946 between the United States of America and France (Award of 9 December 1978), the Franco-American Arbitral Tribunal “tested” one of its conclusions by also taking into account “the overall context of international civil aviation in which the Agreement was negotiated” (UNRIAA, vol. 14, p. 434, para. 44). The Tribunal carried out that verification “even though the Parties had not referred to the negotiating history of the Agreement”, but, according to the Tribunal, “the broader context in which the Agreement was negotiated is relevant” (p. 440, para. 66). The Franco-American Tribunal found that the historical context confirmed the interpretation it had reached on the basis of the text of the Agreement (ibid. and p. 441, para. 69).
In another case, that of the Agreement on German External Debts, the Arbitral Tribunal conducted a detailed analysis of the preparatory works, and came to the conclusion that they “confirmed the result it had already arrived at by interpreting the disputed clause in accordance with Article 31 paragraph 1 of the Vienna Convention”[8] (cited above, para. 37 of the reasons).
Similarly, the Iran-United States Claims Tribunal made the following observation in Case No. A/28 (cited above, para.70):
In their pleadings, the Parties have dealt at length with the negotiating history of
paragraph 7 [of the 1981 Algiers Declarations]. Because the meaning of paragraph 7
is clear, there is no need for the Tribunal to resort to that history in the present
Decision. Nevertheless, the Tribunal finds that nothing in the negotiating history of paragraph 7 contradicts or weakens the interpretation adopted by the Tribunal.
73.    According to Article 32, recourse can be had to supplementary means of interpretation, including the preparatory work and the circumstances in which the treaty was concluded, when the interpretation pursuant to Article 31 leaves the meaning ambiguous or obscure or leads to a result which is manifestly absurd or unreasonable. This provision in the Vienna Convention has its origins in a wealth of long-standing and consistent arbitral jurisprudence that rejects any interpretation that leads to unreasonable results. The Tribunal will give some examples of this jurisprudence.
74.    In one such case, Georges Pinson (France) v. United Mexican States, the Franco-Mexican Claims Commission formulated “the following propositions”, among others, “as to the general principles of interpretation applicable” (Decision No. 1 of 19 October 1928, RSA, vol. V, p. 422, para. 50):
1.    Insofar as the text of the Convention is clear in itself, there is no need to look to the alleged contrary intentions of its authors, save in the exceptional case where the two parties in dispute recognise that the text does not correspond to what they jointly intended.
2.    Insofar as the text is insufficiently clear, recourse to the intentions of the contracting parties is permissible. If, in such a case, their intentions are clear and unanimous, they must prevail over any other possible interpretation. If, on the contrary, they differ or are not clear, the interpretation must be sought which, in the framework of the text, corresponds most closely either to a reasonable solution to the dispute, or to the impression that the offer by the party who took the initiative must reasonably and in good faith have made on the other party.
3.    To determine the meaning of the text of a treaty or the intentions of the contracting parties, the diplomatic negotiations that led to the conclusion of the agreement can be taken into account, unless the text finally adopted by the contracting parties was incompatible with the tenor of the negotiations, or they knowingly waived the right to invoke the diplomatic negotiations as an aid to interpretation.
And, as the Franco-Italian Conciliation Commission stated,
Effort must be made when interpreting a treaty to give a reasonable meaning to a
condition it lays down. […] A rational interpretation should either confirm or invalidate the interpretation arrived at in the particular case by a grammatical analysis of the text taken in isolation.
(Case of Societa Mineraria e Metallurgica di Petrusola, Decision No. 95 of 8 March 1951, ibid, vol. XIII, pp. 186 and 187.)
75.    There are a number of cases in which arbitrators or commissions have rejected interpretations that they believed produced an effect that was “monstrous” (Case of German
Reparations under Article 260 of the Treaty of Versailles, Award of 3 September 1924, ibid., vol. I, p. 439), or lead to results that were “absurd” (the Pinson case, cited above, p. 425), “unreasonable” (John W. Browne (UnitedStates) v. Panama, Decision of 26 June 1933, ibid., vol. VI, p. 334) or “not reasonable” (Naomi Russell, in Her own Right and as Administratrix and Guardian (USA) v. United Mexican States, Decision No. 5 of 24 April 1931, ibid, vol. IV, p. 820). In the Societa Mineraria case (cited above, p. 185), the Commission rejected an interpretation of a provision in an agreement which, in the words of the Commission, constituted “a nonsense or at least a tautology”. In the Baccharach case, the Italo-American Conciliation Commission discarded an interpretation the effect of which, it said, “would be to extend the ordinary meaning . beyond reasonable limits” (Decision No. 22 of 19 February 1954, ibid., vol. XIV, p. 189).
76.    The jurisprudence of the Permanent Court of International Justice and the International Court of Justice on this question has followed the arbitral awards. In the case concerning Polish Postal Services in Danzig (Advisory Opinion of 16 May 1925, P.C.I.J., Series B, No. 11, p. 39), the Permanent Court declared:
It is a cardinal principle of interpretation that words must be interpreted in the sense which they would normally have in their context, unless such interpretation would lead to something unreasonable or absurd.
The International Court of Justice recalled that finding in its Advisory Opinion of 3 March 1950 on the Competence of the General Assembly for the Admission of a State to the United Nations (I.C.J. Reports 1950, p. 7), arriving at an analogous decision (see also Competence of the General Assembly, I.C.J. Reports 1950, p. 8 and South-West Africa, cited above in paragraph 69). Since the conclusion of the Vienna Convention, the Court has on several occasions confirmed that recourse to supplementary means of interpretation under the conditions set forth in Article 32 was possible (Land, Island and Maritime Frontier Dispute (El Salvador/Honduras: Nicaragua intervening), I.C.J. Reports 1992, p. 584, para. 376; Territorial Dispute (Libyan Arab Jamahiriya/Chad), I.C.J. Reports 1994, p. 22, para. 44; Oil Platforms Case (Islamic Republic of Iran v. United States of America), Preliminary Objections, cited above, para. 29).
77.    As the Tribunal has already recalled, France agrees with the view that Articles 31 and 32 of the Vienna Convention reflect customary law in force, which is binding on all States, even those which, like France, are not parties to the Vienna Convention. France however emphasises that “customary law that [alone] binds the two parties does not allow the Convention to be applied with the same kind of minute and analytical rigour as would be the case if it were itself binding on the two Parties” (Counter-Memorial, paragraph 6). In this regard, France refers to the jurisprudence of the International Court of Justice (Counter- Memorial, paragraph 6):
In its Judgment of 27 June 1986 (merits) in the Military and Paramilitary Activities in and against Nicaragua case, the International Court clearly acknowledged that customary and conventional rules could be superimposed on one another. It follows from this that each of them retains its “separate applicability” and that “the substantive rules in which they are framed” do not have to be “identical in content” (I.C.J. Reports 1986, p. 94, para. 175). In the present case, the only law as between the Parties is international custom.
The Tribunal notes, however, that France has not drawn attention to any specific factors in support of its contention that customary rules might differ from Articles 31 and 32 of the Vienna Convention. Moreover, these articles are not as rigid in nature as France appears to claim; on the contrary, they leave the adjudicator a sufficient degree of latitude in the interpretation process. The Tribunal concludes from this that these articles must be taken as a faithful reflection of the current state of customary law. They are therefore applicable in the present case.
*
78.    The Tribunal must now consider the status of Article 33 of the Vienna Convention in customary law. This provision deals with the interpretation of treaties authenticated in several languages (of which the Protocol is one). Article 33 is worded as follows:
Art. 33: Interpretation of treaties authenticated in two or more languages
1.    When a treaty has been authenticated in two or more languages, the text is equally authoritative in each language, unless the treaty provides or the parties agree that, in case of divergence, a particular text shall prevail.
2.    A version of the treaty in a language other than one of those in which the text was authenticated shall be considered an authentic text only if the treaty so provides or the parties so agree.
3.    The terms of the treaty are presumed to have the same meaning in each authentic text.
4.    Except where a particular text prevails in accordance with paragraph 1, when a comparison of the authentic texts discloses a difference of meaning which the application of articles 31 and 32 does not remove, the meaning which best reconciles the texts, having regard to the object and purpose of the treaty, shall be adopted.
79.    Both Parties invoke Article 33 for the purposes of their argument. For reasons that will become apparent further on (see paragraph 89), the Tribunal does not consider it necessary to discuss the status of that article under customary law here. It should be noted, however, that the International Court of Justice recently observed that Article 33, paragraph 4 of the Vienna Convention reflected customary international law (LaGrand case, cited above, paragraph 101).
Chapter V – The Solution to the Dispute
80.    The dispute between the Netherlands and France concerns the financing of the operations contemplated by the Protocol. More precisely, the disagreement between the Netherlands and France relates to the calculation of the amount that France must repay to the Netherlands for the final auditing of the accounts (point 4.2.1 of Annex III). In order to understand fully this dispute, it is necessary to study the mechanism put in place under Article 4 and Annex III of the Protocol. This mechanism proved necessary because it was impossible to predict the quantities of chlorides that would have to be stored each year. As stated in point 1.2.5 of Annex III to the Protocol, “[i]n practice, running costs will vary according to the coefficient of flow of the Rhine”.
81.    Article 4 of the Protocol provides that the parties shall bear the costs of measures taken in French territory, up to a maximum of 400 million French francs, in the following proportions (see also Annex III, points 1.2.1 and 3.2.1):
Germany    30%
France    30%
The Netherlands 34% Switzerland    6%
Point 2.1.1 of Annex III provides for the payment of expenditures in French territory for each year during the 1991-1998 period. The maximum expenditures for 1991 were 90 million French francs. For the other years, maximum expenditures (in French francs) were 38 million for 1992, 27 million for 1993, 73 million for 1994, 36 million for 1995, 36 million for 1996, 50 million for 1997, and 50 million for 1998.
Germany, Switzerland and the Netherlands were to make annual advance payments to France in order to enable it to cover the costs of storage of chlorides and their removal from storage (point 1.2.2). Pursuant to point 2.1.3, these annual payments were to be apportioned according to the formula given in Article 4 of the Protocol. Pursuant to this article, the Netherlands was to pay an amount equivalent to 34% of 90 million French francs for the first year; for the second year, its contribution was to be 34% of 38 million French francs; for the third year, it was 34% of 27 million French francs, and so on.
82.    The amounts set out in point 2 of Annex III also served to limit France’s storage obligations. Indeed, under point 1.2.6:
1.2.6 Each year, France shall be released from its storage obligations once the expenditure in that year has reached the spending limit as determined by point 2 and point 3.2.3. For this purpose, the running costs for France shall be calculated at a rate of 61.5 French francs (1988 French francs adjusted for inflation) per ton stored. For the first year, investment costs shall be added on (40 million French francs adjusted for inflation).
The amounts set out in point 2 of Annex III therefore constituted spending limits. Where the expenditures incurred remained below those limits, the Protocol provided:
3.2.2    At the end of each year, France shall present a report on the quantities stored and the costs related thereto, calculated according to the conditions specified in point 1.2.6.
3.2.3    In the event that the actual storage costs calculated in this way are below the limit initially set for the year concerned (point 2.1.1), the difference (plus 11/12 of the yearly interest on this amount at the annual long-term interest rate on national loans) shall be carried over to the following year, thereby raising the spending limit for the following year accordingly.
83.    Only at the end of the period covered by the Protocol were the accounts to be finally audited:
4.2.1. For expenditure in French territory, payments are not payments in full discharge, and the accounts shall be audited by 31 December 1998 at the latest by comparing the actual expenditure, calculated according to the conditions specified in points 1.2.3, 1.2.4 and 1.2.6 above, with the spending limit set out in point 2 above, increased by any amount carried over pursuant to point 3.2.3 above. In the event that actual expenditure by France is less than the amount referred to in point 1.2, France undertakes to return the surplus received plus 11/12 of the yearly interest on this amount at the annual long-term interest rate on national loans. In this connection, the rate of price rises shall also be taken into account.
84.    Throughout the relevant period, the quantities of chlorides actually stored by France remained well below the limits set out by the parties. As a result, at the end of the period, France was obliged to repay to the other parties the surplus amounts that had been paid (the auditing of accounts). The Parties disagree, however, about the manner in which the surplus to be reimbursed should be calculated. For the Netherlands, the “depenses engagees” [“expenditures incurred”] by France are the costs of investment, plus the amount of 61.5 French francs stipulated in Annex III multiplied by the number of tons of chlorides stored and removed from storage. For France, the only purpose of the 61.5 French francs amount was to calculate the spending limits during the 1991-1998 period; by contrast, at the time of the final auditing, the actual costs of storage and removal from storage were to be used. France submits that the actual costs of storage and removal from storage were much higher than 61.5 French francs per ton, because of the small amounts of chlorides that were in practice stored and removed from storage. The Netherlands, applying its favoured method of calculation, claims reimbursement of the sum of 102.33 million French francs, not including interest, while France, based on the method it advances, considers that it must reimburse the sum of
55,793,779 French francs, not including interest (for detailed calculations, see paragraphs 133 ff).
85.    The Tribunal recalls that the question put to it by the Parties is the following:
Taking into account all relevant provisions of the Additional Protocol to the Convention of 3 December 1976 on the Protection of the Rhine against Pollution by Chlorides and its Annexes, interpreted in accordance with international law, what is the sum that must be transferred between the Parties to the dispute in order that the accounts be audited pursuant to point 4.2.1 of Annex III to the said Protocol?
In paragraphs 54 to 79 above, the Tribunal recalled the rules of international law governing the interpretation of treaties. It will now apply them in deciding the dispute between the Parties.
86.    The general rule in Article 31 of the Vienna Convention is that “[a] treaty shall be interpreted in good faith” (paragraph 1). France insisted on the importance of the principle of good faith in the interpretation of treaties. Such insistence is fully justified; the Tribunal itself acknowledged the fundamental nature of this principle (paragraph 65 above). But the Tribunal also notes that there is no reason to doubt the good faith of the interpretation of the Protocol advanced by the Netherlands. Likewise, there is no reason to doubt the good faith of the interpretation put forward by France. The Tribunal will therefore proceed to examine the terms of point 4.2.1, the meaning of which is in issue between the Parties, and which must be interpreted in good faith.
87.    Point 4.2.1 provides that the auditing of the accounts shall be done:
by comparing the actual expenditure [“par comparaison des depenses engagees”], calculated according to the conditions specified in points 1.2.3, 1.2.4 and 1.2.6 above, with the spending limit set out in point 2 above, increased by any amount carried over pursuant to point 3.2.3 above.
The disagreement between the Parties as to how the phrase “actual expenditure, calculated according to the conditions specified in points 1.2.3, 1.2.4 and 1.2.6” [“depenses engagees calculees selon les modalites prevues aux points 1.2.3, 1.2.4 et 1.2.6”] should be interpreted appears to be the main issue in dispute. Given the importance of point 1.2.6 in this regard, the Tribunal deems it necessary to recall that it reads:
Each year, France shall be released from its storage obligations once the expenditure [“depenses effectuees”] in that year has reached the spending limit as determined by point 2 and point 3.2.3. For this purpose, the running costs [“depenses engagees”] for France shall be calculated at a rate of 61.5 French francs (1988 French francs adjusted for inflation) per ton stored. For the first year, investment costs shall be added on (40 million 1988 French francs adjusted for inflation).
For the Netherlands, the meaning of the phrase in point 4.2.1 quoted above is clear: expenditures [“depenses engagees”] must be calculated by using the amount of 61.5 French francs per ton. France disputes that interpretation, emphasising that the amount of 61.5 French francs was to be used only to determine the annual spending limit and the carry-overs, but not in the final auditing calculation.
88.    At first sight, the French argument appears to be contrary to the actual text of point 4.2.1. Indeed, that point expressly refers back, for the purposes of the calculation of expenditures required for the final auditing, to point 1.2.6, which refers to the product of the quantities stored and the amount of 61.5 French francs per ton.
89.    France disputes this, however, relying on various textual arguments. It explains, first, that the second sentence of point 1.2.6 provides the methods for calculating the running costs only for the purposes stated in the first sentence, in other words for calculating the spending limits.
The Tribunal observes, first, that while the French version of point 1.2.6 uses the words “a cette fin” [“for this purpose”], upon which France relies, the same is not true of the German and Dutch versions. It does not consider it necessary, however, to resort to methods of interpretation such as those in paragraph 4 of Article 33 of the Vienna Convention in order to decide which text should be used. Indeed, point 4.2.1 refers back, for the purposes of auditing, to the calculation methods in point 1.2.6. In so doing, point 4.2.1 gives these methods a new purpose in addition to the one laid down in point 1.2.6. Put another way, this is not a case where the precise terms of one provision are incorporated into another.
90.    Point 4.2.1 also stipulates that “payments are not payments in full discharge”. According to France, this phrase attests to the fact that the Protocol contemplates two distinct mechanisms: one temporary, serving to calculate the spending limits and carry-overs and based on the lump sum amount of 61.5 French francs per ton; the other applying to the final auditing of accounts and based on actual expenditures.
In the present case, the use of the phrase cited above is somewhat incongruous. Indeed, as soon as the parties have made all their advance payments, they are discharged from all financial obligations because the advance payments serve as the maximum limit on the payments each of them is required to make each year. If payments are made each year pursuant to the Protocol, they will necessarily be equal to or greater than each party’s overall obligation, and there could never be a debt or obligation that survived those payments.
In fact, in these circumstances, if “payments are not payments in full discharge” for one of the parties, this can only be so for France, which, as is clearly stated in the second sentence of point 4.2.1, must return the surplus received in the event that its actual expenditures are less than the spending limits set out. The phrase in question does not, under these conditions, support the French position.
91.    The Parties also disagree as to the consequences flowing from the use of the term “comparaison”.  France contends that the use of this term in point 4.2.1 is in contrast with the use of the term “difference” in point 3.2.3, and that this difference in terminology provides a further indication that two distinct processes are contemplated by the Protocol. The Netherlands concedes that there is a difference between the two terms, but maintains that the use of the term “comparaison” in point 4.2.1 is easily explained if one considers that the amount of the expenditures [“depenses engagees”] is not necessarily lower than the spending limit set out. It could, in fact, be the same.
The Tribunal notes that the mere fact that a treaty uses two different terms (but which are very close in meaning) does not mean that it must immediately conclude, without further analysis, that the parties intended to create a significant distinction. Naturally, each treaty is presumed to be consistent in the way it uses its terms, but this presumption cannot be regarded as an absolute rule.
Furthermore, the term “comparaison” is a very broad term, which in some of its meanings comes very close to the notion of “mathematical difference” or “measure of difference”. The Petit Robert dictionary gives the following meanings for “comparaison” [in informal translation]:
1° The act of considering together (two or more objects of thought) together to find
their differences or similarities. See: to compare: analysis, judgment, closeness. To
draw a comparison between…; to make a comparison. To place one thing in
comparison with another. (See: balance, parallel, view). No comparison is possible.
To maintain the comparison. Term of comparison. See: Measure (common
measure). Elements of comparison (analogy, difference, connection, relationship, resemblance). Comparison of texts, writing. See: collation, confrontation, checking. Adverbs of comparison: indicating a greater, lesser or equal relationship (Thus, as many, as much, like same, likewise, less, more. – Degrees of comparison. See: Positive; comparative; superlative.
2° Physical: In comparison to. See: near, by, in regard, in proportion, price, in relation, relatively, vis-a-vis.(…)
3° Relationship established between a subject and another term, in language (…) [Emphasis added] The Tribunal is of the opinion that, in the present case, the immediate context supports the interpretation of the term “comparaison” as synonymous with “mathematical difference”: the two elements being compared in the auditing of the accounts are in practice amounts which can be put into figures. Without subscribing to the view of the Netherlands, the Tribunal concludes from this that France has not shown that the parties intended to invest the term “comparaison” with a particular meaning that referred back to a method that spoke of actual costs.
92. The Tribunal will now turn to study the expressions “depenses engagees” [“expenditures incurred”] and “depenses effectuees” [“actual expenditures”].  France suggests that there is a distinction to be drawn between the “depenses engagees”, “calculated according to the conditions specified in points 1.2.3, 1.2.4 and 1.2.6” (first sentence of point 4.2.1), and the “depenses effectuees” (in the second sentence of point 4.2.1). For France, the “depenses effectuees” are expenditures actually made, calculated according to their actual amount.
In the view of the Tribunal, the terms “engagees” and “effectuees” are very close in meaning, if not identical, and in the context of point 4.2.1, the ordinary meaning of expenditures “engagees” or “effectuees” is that of expenditures made or carried out. The two expressions thus seem to have been used synonymously. At the very least, if the parties had wished, by the use of these terms, to indicate different expenditures, they should have done so in a more explicit fashion.
What is more, the Tribunal observes at the outset that the terms “engagees” and “effectuees” seem to have been used interchangeably in the German, French and Dutch texts. Thus, while the French text of point 1.2.6 refers in the first sentence to “depenses effectuees”, and the German text to “getatigten Ausgaben”, in other words “actual expenditures”, the Dutch text uses only “uitgaven”, or “expenditures”, without qualifying the term. In the second sentence, the French text speaks of “depenses de fonctionnement engagees”, which corresponds to the German text, “eingegangene Betriebskosten”. On the other hand, the expression used in the Dutch text is “betaalde exploitatiekosten”, which corresponds to “depenses effectuees” in French.  The same lack of consistency in the use of the terms “engagees” and “effectuees” can be seen in point 4.2.1. Indeed, while the Dutch and German texts refer, in the first sentence, to a “comparison” [“comparaison”] of “expenditures” [“depenses”], the French text uses the terms “comparaison des depenses engagees”. Lastly, in the second sentence, it is the French and Dutch texts which, referring respectively to “depenses effectuees” and “gedane uitgaveri”, use similar expressions, while the German text uses the term “eingegangene Ausgaben”, which corresponds to the expression “depenses engagees”.
In addition, if the “depenses effectuees” in point 4.2.1 really meant actual expenditures, the phrase “par comparaison des depenses engagees calculees selon les modalites prevues aux points 1.2.3, 1.2.4 et 1.2.6 ci-dessus et les plafonds de depenses prevus au point 2 eventuellement augmentes des reports prevus au point 3.2.3 ci-dessus” (“by comparing the actual expenditure, calculated according to the conditions specified in points 1.2.3, 1.2.4 and 1.2.6 above, with the spending limit set out in point 2 above, increased by any amount carried over pursuant to point 3.2.3 above”) would be rendered meaningless, because the auditing is to be done by simple comparison (or difference) between, on the one hand, the actual expenditures and, on the other, the amount set out in point 1.2. The presumption of effectiveness thus militates against an interpretation that would draw a distinction between “depenses engagees” and “depenses effectuees'”.
Furthermore, point 1.2.6 of Annex III provides that “la France est liberee de ses obligations de stockage des lors que les depenses effectuees au cours de l’annee consideree atteignent le plafond de depenses” (“France shall be released from its storage obligations once the expenditure in that year has reached the spending limit”) [Emphasis added]. Point 1.2.6 provides that the “depenses effectuees” for each year shall be calculated by multiplying the number of tons stocked by 61.5 French francs with the addition, for the first year, of the sum of 40 million French francs for investment costs. In other words, the “depenses effectuees” are calculated, in this provision, not on their actual amount, but according to the stipulated amount of 61.5 French francs per ton. Thus, the expression “depenses effectuees” has been used in the Protocol to mean expenditures calculated on the basis of 61.5 French francs per ton stored.
The response of France to this argument is that the “depenses effectuees” do not have the same meaning in points 1.2.6 and 4.2.1 as the contexts of these two provisions are different: the processes are distinct, they are based on different logic and different methods of calculation. However, France’s response is premised on the recognition that the reasoning underlying the calculation of the carry-overs is different from that underlying the auditing. That is precisely what has to be shown here.
It follows that the argument derived from the use of the terms “depenses engagees” and “depenses effectuees” does not support the proposition that the actual expenditures rather than the amount of 61.5 French francs per ton should be used in the auditing.
93.    The Tribunal has already referred to points 1.2.3, 1.2.4 and 1.2.6, which are mentioned in point 4.2.1. In accordance with the requirements of paragraphs 2 and 3 of Article 31 of the Vienna Convention, the context of point 4.2.1, in other words the other provisions and elements in the context of the Protocol which are also relevant to the resolution of the present dispute, remains to be examined by the Tribunal.
94.    As to the other provisions of the Protocol, France attaches special importance to the apportionment formula that features in Article 4. According to France, the proportion of expenditures that each party must bear pursuant to this article forms the essential basis for all the financial dispositions in the Protocol and cannot be contradicted by any other provision of the Protocol or its Annexes. The Netherlands agrees that the apportionment provided in
Article 4 should be applied, but it adds that all the other terms of the Protocol must be respected when doing so.
The underlying issue here is whether the apportionment formula in Article 4 relates to the actual costs of the measures or the costs as stipulated by the parties in point 1.2.6. Article 4 does not expressly deal with this question. However, it refers back specifically to the payment terms in Annex III. In other words, Article 4 of the Protocol does not decide the issue of whether the calculation for the auditing must proceed on the basis of the lump sum amount of 61.5 (1988) French francs per ton or on the basis of actual costs: it is Annex III that is determinative in this regard. The result is that, while Annex III provides that the calculation of the final auditing must be based on the amount of 61.5 (1988) French francs per ton, Article 4 provides that the expenditures thus calculated must be allocated according to the proportions set forth in that article. Even if Article 4 took precedence over Annex III, the result would be the same, since Article 4 refers back to Annex III for the calculation of the amounts due.
Though the Parties have not raised the question, the Tribunal wishes to add that point 1.2.6 also provides for an amount of 40 million (1988) French francs for investment costs. This was an amount fixed in advance by the Parties, as was the amount of 61.5 (1988) French francs per ton. The Parties do not dispute that the auditing of investment costs must proceed on the basis of the said amount of 40 million (1988) French francs. In the light of that agreement, there appears to be no justification for auditing the running costs on the basis of a different criterion, specifically that of the actual costs.
95.    Turning now to the elements of context (other than the provisions of the Protocol and its Annexes), the Tribunal is not in a position to find that instruments exist that could be helpful in the present interpretation. This is the case, in particular, with the Declaration of the Heads of Delegation adopted on 25 September 1991. Under Article 31 2 b) of the Vienna Convention, “any instrument which was made by one or more parties in connection with the conclusion of the treaty” is part of the context for interpreting the treaty. The Declaration of the Heads of Delegation is incontestably an agreement in connection with the Protocol within the meaning of the provision of the Vienna Convention just cited, and must be taken into consideration in interpreting the Protocol. However, its usefulness in interpreting the Protocol is rather limited, as the two documents closely resemble each other. In the view of the Tribunal, taking the 1991 Declaration of the Heads of Delegation into account does not at this stage offer any assistance to the Tribunal in determining the method to be used to carry out the auditing.
96.    The Tribunal will now turn to the object and purpose of the Protocol. The Tribunal will in particular focus its attention on what significance this might have for the interpretation of point 4.2.1. Both Parties have in fact analysed these elements in their pleadings. The Netherlands, deriving its argument from the Preamble to the Protocol, maintained that the object and purpose of the Protocol was to improve the quality of the waters of the Rhine and facilitate the supply of drinking water from it. France contended that the object and purpose of the Protocol was to establish solidarity between the States bordering the Rhine, as the sources of pollution of that river were many, and in no sense limited to French territory. According to France, forcing it to bear a heavier burden than it had accepted would undermine the solidarity established under the treaty and disregard its object and purpose.
The Netherlands, too, recognises that this solidarity is relevant. Evidence of this can be found in document DELch 17/89 entitled “Dutch Draft Proposal for the subsequent implementation of the Convention on the Protection of the Rhine against Pollution by Chlorides”, dated 14 July 1989, which was in fact a draft proposal formulated by the German and Netherlands delegations. That document sets out measures to be taken “respecting the principle of solidarity”. In addition, document C 2/90 and its revised version, document DELch 4/90 (see below, paragraphs 114 and 122, for the relationship between those documents) add a further element of explanation. Paragraph 1.2 of these documents postulates:
After studying all the possible ways of reducing chloride discharges over the entire
length of the Rhine, the principle decided on was that of reducing levels at the
French Potassium Mines alone, on behalf of all the polluters, because of its better
cost-effectiveness ratio.
It thus appears from this document that the parties decided to opt for a system that established a solidarity between them.
97.    When the States bordering an international waterway decide to create a joint regime for the use of its waters, they are acknowledging a “community of interests” which leads to a “community of law” (to quote the notions used by the Permanent Court of International Justice in 1929 in the Case concerning Territorial Jurisdiction of the International Commission of the Oder (P.C.I.J. Series A, No. 23, p. 27). Solidarity between the bordering States is undoubtedly a factor in their community of interests.
98.    However, identifying the object and purpose of the Protocol is not decisive with regard to the question of what basis of calculation is to be used in the auditing. Both of the interpretations put forward by the Parties are compatible with the object and purpose as identified. Indeed, the Protocol, in the name of the solidarity that exists between the States bordering the Rhine, does organise the fight against the pollution of the river by chlorides by guaranteeing that the measures taken by France and the Netherlands will be jointly financed. The question is, what are the agreed terms of this financing?
99.    Next, in accordance with the general rule of interpretation in Article 31 of the Vienna Convention, the Tribunal must ask whether, in this case, there was any “subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation” (paragraph 3 b), in respect of the interpretation of point 4.2.1. The fact that both Parties have relied on the existence of such a practice makes it all the more important for the Tribunal to address this question.
100.    According to the Netherlands, the annual reports and the provisional balance sheet as at the end of 1996 confirm its interpretation of the Protocol. France opposed this line of argument, on the grounds that the annual reports were prepared in conformity with point 3.2.2 and did not in any way indicate agreement on its part to the interpretation put forward by the Netherlands relating to point 4.2.1 and the final auditing. France referred to the judgment of the Chamber of the International Court of Justice in the Delimitation of the Maritime Boundary in the Gulf of Maine Area case (I.C.J. Reports 1984, p. 246), and added that the annual reports invoked by the Netherlands could in no sense be imputed to any authority that had the capacity to bind the French Republic at international level. Referring again to the decision of the International Court of Justice in the Gulf of Maine case, France also suggested that the practice identified by the Netherlands could not reflect an agreement regarding the interpretation of point 4.2.1, because at the time of that practice, the legal dispute had not crystallised. Finally, France maintained that the suspension of annual payments by some of the parties constituted a practice reflecting the Parties’ acceptance of a flexible application of the Protocol.
101.    The question must be asked whether the method used in the annual reports and the provisional balance sheet as at the end of 1996, prepared by France, reflects agreement by the Parties as to the interpretation to be given to point 4.2.1. In the view of the Tribunal, this is not the case. The reports prepared by France were drawn up in accordance with point 3.2.2 of the Protocol, which provides that each year, France must present “a report on the quantities stored and the costs related thereto, calculated according to the conditions specified in point 1.2.6”. It cannot but be noted that there is no mention of the final auditing anywhere in these reports. As for the provisional balance sheet as at the end of 1996, France prepared this on its own initiative, and it has not been shown that it was submitted for discussion or decision. Whereas it does indeed raise the question of the final auditing, it cannot be taken as expressing a common practice of the Parties with regard to it. As a result, the annual reports and the provisional balance sheet as at the end of 1996 are evidence only of a common practice and an agreement as to how point 3.2.2 should be interpreted, but not as to the final auditing.
Therefore, it is not possible to conclude from the texts relied upon by the Netherlands that a practice existed which established the agreement of the Parties regarding the interpretation of point 4.2.1. There is thus no need to enter into a discussion as to the relevance of the Gulf of Maine case, nor to decide whether it was necessary for the legal dispute to have crystallised at the time of the practice relied upon.
The Tribunal also considers that the fact that Germany and Switzerland, which were parties to the Protocol, stopped making their annual payments around 1995-1996 does not show the existence of any practice as between France and the Netherlands, reflecting an agreement between the Parties to the present case that favoured the interpretation France gives to the Protocol.
102.    In carrying out its task, the Tribunal must also take into account “any relevant rules of international law applicable in the relations between the parties” (Article 31, paragraph 3 c) of the Vienna Convention).
The Tribunal notes that the Netherlands has referred to the “polluter pays” principle in support of its claim.
103.    The Tribunal observes that this principle features in several international instruments, bilateral as well as multilateral, and that it operates at various levels of effectiveness. Without denying its importance in treaty law, the Tribunal does not view this principle as being a part of general international law. The “polluter pays” principle does not appear anywhere in the Convention or the Protocol. The Protocol, furthermore, adopts a different solution. Besides, the Netherlands acknowledges that the Protocol derogates from the “polluter pays” principle. It follows that this principle is of no relevance for the interpretation of point 4.2.1.
104.    The Netherlands also claimed that it was in a particular geographical situation in relation to the other parties to the Protocol. The Netherlands put forward the view whereby
“by virtue of its geographical position on the lower reaches of the Rhine, it suffers damage caused by discharges of chlorides, and is therefore in a different position from that of Germany, Switzerland and France, as far as instituting the present proceedings is concerned” (Reply, paragraph 1.2).
It is not clear whether, by underlining its geographical position, the Netherlands is seeking to rely on any legal regime and its effect. The Netherlands has not specified the legal implications of the particular geographical situation it claims to have. In any event, the Protocol does not accord any such particular status to the Netherlands where the final auditing is concerned.
105.    Having regard to the foregoing analysis (see paragraphs 80 to 104), it is the opinion of the Tribunal that the calculation for the final auditing must be based on the amount of 61.5 (1988) French francs per ton.
*
106.    Nonetheless, the Tribunal considers that it must, in this case, do what is required under Article 32 of the Vienna Convention. That provision states that the Tribunal may have recourse “to supplementary means of interpretation, including the preparatory work of the treaty and the circumstances of its conclusion, in order to confirm the meaning resulting from the application of article 31, or to determine the meaning when the interpretation according to article 31: a) leaves the meaning ambiguous or obscure; or b) leads to a result which is manifestly absurd or unreasonable”.
In the course of the written proceedings, France and the Netherlands referred to certain documents that could be regarded as supplementary means of interpretation within the meaning of Article 32. During the hearing, both Parties submitted documents that had not previously been produced. After the hearing, the Netherlands put forward two other documents. The Tribunal authorised the production of these documents and, furthermore, asked the Parties for explanations concerning them, which led to additional documents being submitted by the Parties. Finally, the Tribunal itself ordered the Parties to submit certain documents to it (see paragraphs 17 to 30).
The Tribunal proposes to examine the various documents submitted to it by the Parties as “preparatory documents” for the Protocol of 25 September 1991. The Tribunal recalls that each Party has maintained that the said documents uphold its own interpretation. In particular, France contends that the interpretation of the Netherlands leads to an unreasonable result, and that the documents submitted to the Tribunal enable it to reject that interpretation.
107.    For the sake of clarity, these documents are identified below (in chronological order, according to the date they bear). Some chronological points of reference are also given:
– August 1987:    Document entitled “Studies for the Bonn Convention
– 14 July 1989:
Second Phase – Temporary storage of residual salts near the Mines”, prepared by the Mines de Potasse d’Alsace S.A. (the “1987 Studies”);
“Zendbrief’ and document entitled “Dutch Draft Proposal for the subsequent implementation of the
Convention on the Protection of the Rhine against Pollution by Chlorides”, drafted by the German and Netherlands parties at Kempen on 14 July 1989 and bearing the classification number DELch 17/89 (“DELch 17/89”);
Document entitled “Temporary storage of salts at the MPDA [Alsace Potassium Mines] during periods of heavy Cl concentration in the Rhine – Technical and Economic Study”, bearing the handwritten classification number C 10/89, prepared by the Environmental Department of the Mines de Potasse d’Alsace S.A. (“C 10/89”);
Document entitled “Efficiency and cost of various scenarios for regulating the retention of residual salts by French potassium mines – documentation for the Ministerial Conference of States Bordering the Rhine of 30 November 1989 in Brussels”, prepared by the Netherlands and bearing the handwritten classification number C 11/89 (“C 11/89”);
13 November 1989:
November 1989:
30 November 1989:
1 March 1990:
Conference of Ministers in Brussels, submission of the alternative plan by the Netherlands Minister, by agreement with the French Minister;
March 1990:
Covering memo and document entitled “Report to the ICPR on the technical aspects, costs and financing methods of the plan put forward by the Netherlands by agreement with France for the implementation of the second phase of the Chlorides Convention”, prepared by the rapporteur Patrick Weingertner of the Rhine-Meuse Water Agency, and bearing the handwritten classification number C 2/90 (“C 2/90”);
Document entitled “Report of the Working Group on Chlorides on the alternative Franco-Dutch plan for the implementation of the second phase of the Chlorides Convention” drawn up and sent to the parties on 14 March 1990 by the Secretariat of the ICPR and bearing the classification number C 8/90 (“C 8/90-Nl”);

Document entitled “Record of Decisions of the Working Group on Chlorides, held on 20 and 21.03.90 in Koblenz”, prepared by the ACPR and bearing the classification C 11/90 (“C 11/90”);
after 21 March 1990:
28 March 1990:
Document entitled “Report to the ICPR on the technical aspects, costs and financing methods of the plan presented by the Netherlands with the agreement of
Document entitled “Record of Decisions of the 47th Meeting of the Heads of Delegation held at Haarlem on 4 May 1990”, prepared by the ICPR and bearing the handwritten classification number PLEN 1/90 (“PLEN 1/90″);
after 4 May 1990:
10 July 1990: 10 September 1990:
5 October 1990:
France as an alternative to the measures envisaged for implementing the Chlorides Convention”, prepared by the rapporteur Patrick Weingertner of the Rhine-Meuse Water Agency, and bearing the typewritten classification number C 2/90 rev. 21.03.90 and the handwritten classification number DELch 4/90 (“DELch 4/90”);
– 4 May 1990:    47th meeting of the Heads of Delegation in Haarlem.
Plenary meeting of the ICPR held in Essen.
2 July 1991:
Document entitled “Cost evaluation for the implementation of adjusted temporary storage in France with effect from 1 January 1991”, prepared by the Rhine- Meuse Water Agency and bearing the handwritten classification number C 8/90 (“C 8/90-Fr”);
Document entitled “Draft – Declaration of the Heads of Delegation of the Governments Parties to the Agreement on the International Commission for the Protection of the Rhine against Pollution”, with an annex entitled “Explanatory Note by the French Delegation on Item 5 of the draft Declaration by the Heads of Delegation”, (“DELch 12/90”);
Plenary meeting of the ICPR held in Lenzbourg.

– after 2 July 1991:    Document entitled “Record of the Decisions of the 56th
Plenary Meeting of the ICPR held on 2 July 1991 at Lenzbourg (Switzerland)”, prepared by the ICPR and bearing the classification number PLEN 30/91 (“PLEN 30/91”).
108. Before proceeding to analyse the documents cited above, certain preliminary observations need to be made. First, different classification systems for the documents appear to coexist (C, DELch, PLEN). It seems likely that these classifications correspond to the body that prepared the document in question, or the one to which it was addressed (C = Working Group on Chlorides, DELch = Heads of Delegation, PLEN = plenary meeting). The Tribunal does not it think it necessary in the present case to consider whether documents bearing the classification number DELch or PLEN are of greater “value”, as such, for the purposes of interpreting the Protocol, than those marked with a C.
Next, it should be noted that there are “gaps” in the documentation: certain periods or meetings are not covered by the documents submitted by the Parties. Because of this, the
Tribunal has no information before it concerning the discussions that took place or the documents that were prepared between 5 October 1990 and the meeting of 2 July 1991.
109.    Turning to its analysis of the documents produced by the Parties, the Tribunal will start with the “1987 Studies” document. According to France, the amounts appearing in it were updated in April 1988. As a result, the “April 1988 Study”, to which certain later documents refer, and the 1987 Studies are one and the same document. This document must be treated with caution, because the studies it shows are based on the mechanism originally provided for in the 1976 Convention for the second phase of reduction of chlorides in the Rhine. That mechanism, which provided for continuous storage of the chlorides produced at the Alsace Potassium Mines, was not adopted in the Additional Protocol. The 1987 Studies envisaged the storage of 20.6 million tons of residual salts (or 16.3 million tons of NaCl), a figure that was subsequently reduced to a maximum of 5.6 million tons of NaCl. The Studies furthermore make a clear distinction between, on the one hand, the investment costs connected to storage and removal from storage, the former valued at 186 to 219 million French francs and the latter at 62 to 65 million (rounded up), and, on the other hand, the running costs for the operations of storage and removal from storage, respectively valued at 347 and 535 million French francs (rounded up). These figures, which feature in the 1987 Studies, are of limited interest in themselves, as the investments actually made and the quantities of salts to be stored were in the end very different from those anticipated in 1987. The Tribunal nonetheless notes that in these studies the running costs were estimated at 13 francs per ton of residual salts for their storage (Annex 9, page 3). As to the running costs for the removal from storage operations, details of which appear in the same Annex (page 5), an estimate of 19 francs per ton of residual salts is given, to which a cost of 6 francs must be added for the “Marie-Louise” mine and 12 francs for the “Amelie” mine, after that mine ceased to operate, resulting roughly in an amount of 26 francs per ton of residual salts (or 33 francs per ton of NaCl). The total cost of removal from storage (including the investments) is valued at approximately 37 francs per ton of NaCl (Annex 9, page 6).
110.    The first time an alternative plan was proposed was in document DELch 17/89, put forward by the Netherlands Government by agreement with the French Government, which served as the basis for the negotiations that were to come. The document does not provide details concerning the financing of the operations; it simply states that measures would be taken that respected the principle of solidarity, without setting out any detailed modalities.
111.    Document C 10/89 gives the first detailed technical and economic study of the envisaged alternative plan. Its preamble specifies that the general characteristics of the storage to be carried out at the Alsace Potassium Mines “are taken for the most part from the study done in April 1988 for the implementation of the second phase of the Bonn Convention”.
The preamble of document C 10/89 adds that “the total quantities of salts to be stored, and the corresponding expenditures, are to be determined depending on the number of days of storage, which is tied to the hypotheses used for the coefficient of flow of the Rhine during the period 1991/1998”. This sentence presents a problem of interpretation: does it mean that only the total amount will vary as a function of the quantities stored, or that the amount per ton itself will also vary as a function of the quantities actually stored?
Document C 10/89 provides that the investment expenditures would be 66.5 million (1988) French francs for the first year, of which 26.6 million (1988) were for setting up the necessary storage areas for the first year of operation (item 51). However, it was proposed to deduct the said amount of 26.6 million (1988) from the investment expenditures and include it in the proportionate storage expenditures (item 52), and to limit the total amount of investments, other than setting up the storage areas, to 40 million (1988) French francs. In addition, document C 10/89 reiterates that the proportionate storage expenditures had been fixed in April 1988 at 13 francs per ton of product stored. It corrects this figure, in order to include in these expenditures those that were necessary to set up the storage areas, and also to take account of the reduction in the quantities to be stored. It thus arrives at the figure of 19.75 French francs, representing the cost per ton of product stored (or 24 francs per ton of NaCl). This shows, France maintains, that the said amount was calculated using both fixed elements and variable elements. Lastly, item 51 of document C 10/89 states that these figures were calculated for the first year of operation on the basis of storage in a dry year, “or the equivalent of 228 days of storage maxi (ref. 1976)”. The Tribunal has not been able to determine the yearly quantity to which this corresponds. Also, no indication is given of the quantities expected to be stored after the first year.
As for removal from storage, document C 10/89 gives the figure of 31 French francs per ton of product stored (or 37.5 francs per ton of NaCl), based on the April 1988 study. This, as has been seen above, included expenditures for removal from storage of about 37 francs per ton of NaCl, broken down into approximately 33 francs for running costs and 4 to 5 francs for investment expenditures. Here again it thus appears that, as France maintains, the figure of 37.5 French francs per ton was calculated on the basis of both fixed and variable elements. And here, again, the yearly quantity of residual salts to be stored (which it appears had to be less than the 20.6 million tons initially forecast) is not provided, and the Tribunal has been unable to determine it.
In sum, it is clear from document C 10/89 prepared by the Environmental Department of the Alsace Potassium Mines that final investment expenditures of 40 million (1988) French francs were proposed for the first year. The proportionate expenditures for storage and removal from storage were calculated by including both fixed and variable costs. However, it is not possible, from these documents, to determine the share of each of these in the figures proposed of 24 francs per ton of NaCl for storage and 37.5 francs for removal from storage, giving a total amount of 61.5 francs per ton.
112. The purpose of document C 11/89 was to examine a certain number of scenarios for salt storage as a function of the coefficient of flow of the Rhine. The results are presented on the one hand in the form of averages over the entire period (1970-1988), which was considered a “representative” period, and, on the other, in the form of averages over the period 1971-1980 (a dry period used to give a “worst case” scenario). The quantity of salts to be stored varied, depending on the scenario, from 2.9 to 6.9 million tons. The expenditures fluctuated between 223 and 467 million francs. The results obtained are summarised in document C 11/89 in a table corresponding to the first three columns of the table that follows. The Tribunal used the table to calculate the cost per ton of salts of the operating expenditures.
Representative (1970-1988)    Storage of
salts
(millions of tons of NaCl)    Expenditures (millions of 1988
french francs)    Cost per ton (excluding 40 million French francs in investments)
Scenario 1    3.1    235    (235-40) – 3.1 = 62.9 Fr francs/ ton
Scenario 2    2.9    223    (223-40) – 2.9 = 63.1 Fr francs/ ton
Scenario 3    4.1    293    (293-40) – 4.1 = 61.7 Fr francs/ ton
Scenario 4    4.7    325    (325-40) – 4.7 = 60.6 Fr francs/ ton
Scenario 5    5.2    357    (357-40) – 5.2 = 61 Fr francs/ ton
“Worst case” (1971-1980)
Scenario 1    4.7    331    (331-40) – 4.7 = 61.9 Fr francs/ ton
Scenario 2    4.1    294    (294-40) – 4.1 = 62.0 Fr francs/ ton
Scenario 3    5.6    387    (387-40) – 5.6 = 62.0 Fr francs/ ton
Scenario 4    6.3    430    (430-40) – 6.3 = 61.9 Fr francs/ ton
Scenario 5    6.9    467    (467-40) – 6.9 = 61.9 Fr francs/ ton

The table shows that the cost per ton (excluding the investment costs of 40 million French francs) varies between 60.6 French francs and 63.1 French francs per ton. Under scenarios No. 3, which document C 11/89 proposes should be adopted, this cost amounts to 61.7 francs for storage of 4.1 million tons of residual salts and 62 francs for storage of 5.6 million tons. The Parties could therefore note that the variations in operating costs that resulted from the variations envisaged in the rate of flow of the Rhine were minimal.
113.    The alternative plan was put forward at the Ministerial Conference of 30 November 1989 in Brussels. The preamble of document C 2/90 states that, at that conference, the Environment Ministers asked the ICPR to provide more details on certain aspects of the alternative plan.
114.    Document C 2/90 was subsequently drafted. In that document, a choice was made between the various storage scenarios that had been put forward in document C 11/89. On the basis of that choice, it is explained on page 5 that:
The proportionate expenditures are linked in essence to the quantities that will actually be stored, at the rate of 61.5 French francs/ton of NaCl (of which 24 French francs/ton are for storage and 31 francs/ton for later removal from storage ).
The quantities stored will in practice vary according to the coefficient of flow of the Rhine.
Calculated on the basis of 8.5 years of provisional operations, the quantities stored are estimated at:
–    4.1 million tons of NaCl for an average coefficient of flow of the Rhine corresponding to the period 1970-1988.
–    5.6 million tons of NaCl for a low flow coefficient of the Rhine, corresponding to the period 1971-1980.
[Note by the Tribunal: A material error exists here: the amount for subsequent removal from storage is 37.5 French francs/ton and not 31 French francs/ton, the latter amount being the amount by ton of product stored, and not by ton of NaCl.] 2.2 – Costs
By contrast with the overall French plan of 1988 for which additional storage was carried out on a continuous basis over the entire forecast period, thus giving a total amount of expenditures that could be calculated precisely in advance, the costs under the alternative plan could not be established with any precision.
Indeed, for the occasional storage to be carried out at the Alsace Potassium Mines, the expenditures will vary according to the coefficient of flow of the Rhine in the years to come. This leads to the simultaneous proposals:
•    to make a forecast of expenditures on the basis of 5.1 million tons of NaCl stored in 8.5 years, which corresponds to costs of 355 MF (of which 40 MF are investment costs and 37 MF are proportionate annual expenditures).
•    to audit the closing accounts, on 31 December 1998, by comparison with actual expenditures;
•    to limit the expenditures in respect of storage at the Alsace Potassium Mines to a maximum of 400 MF.
The Tribunal notes that the estimate of the quantities to be stored given in this document corresponds to the two scenarios No. 3 in document C 11/89 reproduced above, that is 4.1 and 5.5 million tons. The corresponding costs came to 61.3 French francs per ton in the first case and 62 French francs in the second. However, rather than allowing this cost to vary according to the coefficient of the flow of the Rhine and the quantities to be stored, document C 2/90 instead chose to use a hypothetical average of 5.1 million tons of residual salts to be stored over 8.5 years to calculate the costs, corresponding to a cost of 355 million French francs – 40 million in investment, and 315 million in operating expenditures. On the basis of this hypothesis, the cost per ton would have been 61.8 French francs. The cost that was finally opted for (61.5 francs) was very close to this, and corresponds to the figure that had initially been proposed by the Alsace Potassium Mines in document C 10/89.
It appears that document C 2/90 (technical and financial arrangements prepared by France) was discussed at the meeting of the Working Group on Chlorides held on 20 and 21 March 1990 in Koblenz, because item 3.1 of document C 11/90 states that the proposal drafted by the French delegation, document C 2/90, was to be reworked by Mr. P. Weingertner on the basis of the discussions and the requests for amendments.
115.    The Tribunal will now examine document C 8/90. In Order No. 1, the Tribunal asked the Parties, among other things, to explain why the same classification number, C 8/90, appeared on two different documents, one of them submitted by France at the hearing and the other by the Netherlands after the hearing.
116.    In reply to the question put by the Tribunal, France explained that the C 8/90 document presented by the Netherlands (C 8/90-Nl) included the contents of document C 8/90 submitted by France (C 8/90-Fr) at the hearing as well as the measures recommended by the Government of the Netherlands for its territory. France added (page 1, fourth paragraph of its reply to Order No. 1):
These two documents have the same content in part, but they should in fact have borne different classification numbers. The document presented by the French party does not emanate from the ICPR (International Commission for the Protection of the Rhine) archives, unlike the one presented by the Netherlands party, but rather from the archives of the Rhine-Meuse Water Agency. The French government is not in a position to explain why two documents, the contents of which are partly the same and partly different, were given the same classification number.
France reverts to this question in its reply to Order No. 3, in the first paragraph of which it writes:
As to the reply of the Netherlands party of 6 January 2003 to Order No. 1 of the Tribunal, transmitted to the French party by letter from the Registry on 20 January 2003, the French party has no particular comments to make, but simply wishes to restate the position set forth in its letter of 17 January 2003. It is stated in that letter that the document C 8/90 mentioned by the Netherlands party has, as far as the present discussion is concerned, the same content as the document C 8/90 presented by the French party at the hearing on 3 October 2002. In any event, the formal aspects of the document have no bearing on the issue at hand. In this regard, the French party has already pointed out that the price calculation elements contained in that document had been the subject of lengthy discussions between the parties on several occasions. As is known, the notion of preparatory works is more a matter of substance than form, and the French party cannot understand how such a document could be excluded from them, given that it quite clearly dealt with an important point in the Protocol, and had been discussed between the parties, as the French party has shown. [references omitted] As to the circulation given to the French document C 8/90, France states (at the second page, first paragraph of its reply to Order No. 1):
it is clear that, like all documents archived by the ICPR, it was given the circulation necessary to enable it to be discussed. All the documents mentioned in and relevant to the present reply come from the ICPR archives, which means they have been brought to the attention of the parties to the Protocol, including the Netherlands party, in a timely manner.
117.    For its part, the Netherlands writes, in paragraphs 5 and 6 of its response to Order No. 1:
Inquiries conducted with the help of the ICPR Secretariat in Koblenz mean that the following arguments can now be put forward in support of the contention that document C 8/90 dated 10 September 1990 is not an ICPR document:
a)    A search of the file made in The Hague in October and November 2002, an account of which is given in the letter of 6 November 2002, did not succeed in locating document C 8/90 of 10 September 1990. Nor was the ICPR Secretariat able to trace this document in the ICPR archives in Koblenz;
b)    The same inquiry in October and November 2002 brought to light another document C 8/90, dated March 1990, this time drafted within the ICPR, as evidenced by the fact that:
i)    it was sent to the Parties with a letter from the ICPR Secretariat dated 14 March 1990, and
ii)    it was referred to at the meeting of the Working Group on Chlorides held on 20 and 21 March 1990, even though it was not discussed in detail.
It was the title page of this latter document that was forwarded to the Tribunal annexed to the letter of 6 November 2002, at the same time as the abovementioned letter to the ICPR Secretariat;
c)    No document with the classification number C 8/90 dated 10 September 1990 was discussed at the meeting of the Working Group on Chlorides on 13 September 1990, nor at the meeting of the Heads of Delegation on 31 October 1990, contrary to what the date of that document might suggest;
d)    Document C 8/90 of 10 September 1990 was drafted by a French body (the Rhine-Meuse Water Agency, International Mission). We would mention that the classification number C 8/90 was added by hand.
In the light of the foregoing, the Netherlands considers that document C 8/90 of 10 September 1990 does not form part of the travaux preparatoires of the Protocol. [italics in the original] In its response to Order No. 3, dated 14 February 2003, the Netherlands observes:
As for the document C 8/90 submitted by France, the Netherlands considers that it has been convincingly shown in its response to Order No. 1 of 6 January 2003 that the said document did not play any part in the negotiations on the Protocol. It considers, moreover, that the argument made by France regarding this document is ambivalent. France says, on the one hand, that this document does not come from the ICPR archives, but from those of the Rhine-Meuse Water Agency (point 1 of their response), and on the other, that that it was circulated as a document archived by the ICPR (point 2(a) of the response) and, in addition, that it reflects the outcome of prior discussions and was taken up in the subsequent ICPR negotiations (point 2(b) of the response). [underlining and italics in the original] 118. In the view of the Tribunal, a comparison of both C 8/90 documents shows that the Netherlands document (dated March 1990) and the French document (dated 10 September 1990) differ in several respects:
– the March 1990 document C 8/90 for the most part repeats what is in document C 2/90. Having mentioned several hypotheses for storage (4.1, 5.1 and 5.6 million tons), it proposes an expenditures forecast based on 5.1 million tons to be stored over 8.5 years. The 10 September 1990 document contemplates storage of 4.5 million tons, over a period of 7.5 years;
– document C 8/90 from the Netherlands repeats the calculations in document C 2/90 without going into further detail, while the French document deals with the proportionate yearly costs of setting up the storage areas.
The Tribunal also points out that, while document C 8/90 presented by France clearly states that the storage costs can be broken down into fixed costs and variable costs of storage (it is stated on page 2 of that document that each of these headings accounts for about 12 francs per ton stored), that same document does not break down the costs of removal from storage into fixed costs and variable costs. Indeed, that document only mentions the amount of 37.5 French francs per ton allocated for removal from storage.
119. The Tribunal will begin its analysis with document C 8/90 of March 1990 (the document submitted by the Netherlands). There is no doubt that this document emanated from the ICPR, more specifically from the Working Group on Chlorides. This document was later described as a “brief summary report on the facts … prepared by the Secretariat” (see document C 11/90, page 5).
The costs of storage operations at the Alsace Potassium Mines are discussed on page 4 of the document. The amount of 61.5 French francs per ton is broken down into two amounts, one for storage (24 francs per ton) and the other for removal from storage (31 francs per ton). The Tribunal notes, as it has done with regard to document C 2/90 (see paragraph 114 above), that the latter amount should have read 37.5 francs per ton and not 31 francs per ton. These amounts are not broken down into fixed costs and variable costs.
Document C 8/90-Nl hardly differs at all from document C 2/90 on the issues relevant to this dispute. It states, at page 4:
For the occasional storage to be carried out at the Alsace Potassium Mines, the expenditures will vary according to the coefficient of flow of the Rhine in the years to come: for 4.1 million tons of NaCl – which corresponds to an average (coefficient of flow) – costs will be 292 MF (40 MF in investment and 252 MF in proportionate expenditures) while the costs for storage of 5.6 million tons of NaCl – which corresponds to a low coefficient of flow – will be 384 MF (40+344 MF). This leads to the simultaneous proposals:
*    to make a forecast of expenditures based on 5.1 million tons of NaCl stored in 8.5 years, which corresponds to costs of 355 MF (of which 40 MF are investment costs and 37 MF are proportionate annual expenditures).
*    to audit the closing accounts on 31.12.1998, by comparison with actual expenditures.
*    to limit the expenditures in respect of storage at the Alsace Potassium Mines to a maximum of 400 MF.
When the costs per ton for these different coefficients of flow are calculated, the result is always close to 61.5 French francs per ton, the variations being insignificant:
252 MF – 4.1 M tons = 61.46 francs/ton 344 MF – 5.6 M tons = 61.43 francs/ton
315 MF -r 5.1 M tons = 61.76 francs/ton
In these circumstances, it can be understood, given the hypotheses put forward as to the coefficient of flow of the Rhine, how the figure of 61.5 French francs per ton was arrived at irrespective of the quantities to be stored.
Finally, the March 1990 document C 8/90 also provides the apportionment formula for expenditures (see page 5) and states that “for expenditures in French territory, payments are not payments in full discharge, and the accounts shall be audited by 31 December 1998 at the latest”.
120.    As for the 10 September 1990 document C 8/90 (in other words the one submitted by France), the Tribunal takes the view, having regard to the answers of the Parties to the question put by the Tribunal in Order No. 1 concerning the two documents bearing the classification number C 8/90 (see paragraphs 116-117), that France has not established with certainty that the 10 September 1990 document C 8/90 was an ICPR document or that it was circulated to all the parties.
Besides, France does not assert that this particular document was actually discussed. Rather, it suggests that it “reflects the outcome of previous discussions and was taken up again later in the negotiation” (paragraph 2, page 2 of the reply to Order No. 1). It then makes reference to a number of other documents. In these circumstances, the Tribunal considers that France has not established that the Parties discussed the document it produced. Therefore, aside from the incidental remarks that follow, the Tribunal takes the view that any further analysis of that document is superfluous.
That document appears to support the French position up to a point. It contains a statement that the running costs of 24 French francs per ton for storage included (item 2):
the proportionate annual costs of setting up the watertight storage areas (for a quantity of 4.5 million tons) and the direct expenses of storage, each of these two items representing about one half.
On the other hand, no indication is given as to how the figure of 37.5 French francs per ton for removal from storage is broken down.
Lastly, the Tribunal observes that the quantity to be stored envisaged in this document differs from the hypothetical amounts used previously, while the cost per ton remains the same. It will come back to this point at a later stage.
121.    The Tribunal will now examine document C 11/90. This document is the record of the decisions of the meeting of the Working Group on Chlorides held on 20 and 21 March 1990 in Koblenz. It states at page 2:
After the precise financial arrangements for the measures to be taken on French and Netherlands territory have been worked out, bilateral consultations will have to take place between the financial experts responsible.
There is no way of knowing whether these consultations between the financial experts ever took place.
Document C 11/90 also mentioned the reports that had been discussed (p. 5):
The Working Group rejected the proposal by the German delegation to submit a short summary report setting out the facts as in doc. C 8/90 drawn up by the Secretariat, supplemented by the issues that are still undecided. It agreed only to the presentation of the individual reports that had been discussed (C 2/90, C 3/90 and C 10/90, each of which bore the date of revision of 21.3.90). The delegations themselves undertook to inform their Heads of Delegation before the next Heads of Delegation meeting of the outcome of the discussions and the issues that remained undecided.
[Note by the Tribunal: document C 3/90 had been prepared by the Netherlands and dealt with ecological aspects; document C 10/90 had been prepared by Germany and dealt with the supply of drinking water.] 122. In paragraphs 37 and following of its Rejoinder, France refers to a document which, with its agreement, the Netherlands had supplemented at the hearing. This was document DELch 4/90. This document was a revised version of document C 2/90, and, like it, dealt with the alternative Franco-Dutch plan. It was presented to the Heads of Delegation of the ICPR on 4 May 1990.
As with document C 2/90, document DELch 4/90 contains a cost evaluation based on an estimate of 5.1 million tons of NaCl to be stored over a period of 8.5 years. The relevant passages, cited by France, are reproduced here:
the proportionate expenditures are essentially tied to the quantities to be stored, at 61.50 FF/T of NaCl (of which 24 F/T is for storage and 31 F/T for subsequent removal from storage).
By contrast with the overall French plan of 1988 for which additional storage was carried out on a continuous basis over the entire forecast period, thus giving a total amount of expenditures that could be calculated precisely in advance, the costs under the alternative plan could not be established with any precision.
Indeed, for the occasional storage to be carried out at the Alsace Potassium Mines, the expenditures will vary according to the coefficient of flow of the Rhine in the years to come. This leads to the simultaneous proposals:
*    to make a forecast of expenditures on the basis of 5.1 million tons of NaCl stored in 8.5 years, which corresponds to costs of 355 MF (of which 40 MF are investment costs and 37 MF are proportionate annual expenditures).
*    to audit the closing accounts, on 31 December 1998, by comparison with actual expenditures [“depenses effectives”].
*    to limit the expenditures in respect of storage and removal from storage at the Alsace Potassium Mines to an overall amount of 400 MF.
[Emphasis added] [Note by the Tribunal: the same material error as in documents C 2/90 and C 8/90 NL can be found in this document: the amount for subsequent removal from storage is 37.5 ff/t NaCl and not 31 ff/t.] According to France, this document confirms the French position in three ways (see Rejoinder, paragraph 39):
first, because it distinguishes between a lump sum spending limit and actual variable expenditures; then because it distinguishes between the working of the Protocol based on lump sums, on the basis of forecast expenditures, and a final auditing to be based on actual expenditures; and finally because, for the calculation of the lump sum, a forecast was made of a given quantity of chlorides to be stored (which in this document is 5.1 million tons of NaCl).
France adds that the fact that the final figure for the amount of chlorides to be stored was revised and lowered (from 5.1 million tons to 4.55 million tons) “has no significance here, especially since the figures are still of the same order of magnitude”.
123.    The Tribunal finds that, while this document gives a lump sum figure of 61.5 French francs per ton and breaks this amount down into costs of storage and costs of removal from storage, it does not, any more than documents C 2/90 and C 8/90-Nl, break down these figures themselves into fixed costs and variable costs of storage and removal from storage. The document only indicates that the amount of 61.5 French francs per ton includes, amongst other things, 24 francs per ton for storage and 31 francs [sic] per ton for subsequent removal from storage. Thus, the amounts mentioned in paragraph 67 of the Counter-Memorial (12 francs/ton for fixed storage costs; 12 francs/ton for variable storage costs; 17.8 francs/ton for fixed costs of removal from storage; 19.7 francs/ton for variable costs of removal from storage) do not appear in this document.
Furthermore, the use of the terms “depenses effectives” in the document does not constitute an indication that it was referring to actual costs. Indeed, one is left with the same problem of not knowing whether the “depenses effectives” refer to actual costs or to the number of tons stored multiplied by the amount of 61.5 French francs per ton.
Finally, while the forecast expenditures for the calculations in document DELch 4/90 were based on a hypothetical amount stored of 5.1 million tons, it should be pointed out here that this figure was not the one that was ultimately used, and it was later revised and lowered to 4.5 million tons without any change being made to the amount of 61.5 French francs per ton (see paragraph 125 below).
124.    Document PLEN 1/90 is the record of decisions of the 47th meeting of the Heads of Delegation held in Haarlem on 4 May 1990. Under point 3, it is mentioned that Dr. Dubois presented the results of the work of the Working Group on Chlorides, which the Ministers had charged with examining the feasibility and the impact of the alternative Netherlands plan which France supported. The summary of this presentation indicates that (p. 2):
because of fundamental divergences in the views expressed, it was not possible to reach agreement on a draft Additional Protocol to the present “Chlorides” Convention.
125.    Document DELch 12/90 is a draft Declaration for the adoption of the alternative plan. It sets out in paragraph 5 the total cost of the measures contemplated and for the details refers to an attached explanatory note from the French party that repeats some of the data provided in the 10 September 1990 document C 8/90. The first page of that explanatory note states that:
Given the time that has elapsed since the presentation of the Franco-Dutch alternative plan (approximately one year) and the shortening of the period during which storage will be implemented (from 8.5 years to 7.5 years), it seemed appropriate to revise the costs for the storage measures to be taken in Alsace.
In December 1989, the total cost had been estimated at 355 MF on the basis of a quantity of 5.1 million tons of NaCl stored over 8.5 years, and the spending limit had been fixed at 400 MF (margin of 12.5%).
To date, the total costs can be estimated at 317 MF based on a quantity of 4.5 million tons of NaCl stored over 7.5 years and the spending limit can be fixed at 355 MF (same margin of 12.5%).
The Parties thus had the opportunity to revise the costs relating to the alternative plan. However, in spite of the reduction contemplated in the quantity to be stored, which fell from 5.1 to 4.5 million tons, the amount of 61.5 French francs per ton remained unchanged (see the final paragraph of the explanatory note). The only thing adjusted was the total cost of the measures.
This solution was easily explained, because, here again, the variations in the quantity to be stored did not have much effect on the cost per unit, since:
(355 – 40): 5.1 = 61.76 francs/ton (317 – 40): 4.5 = 61.55 francs/ton.
126.    The Tribunal repeats what has already been pointed out in paragraph 108 above, namely that the Parties have given the Tribunal no information concerning developments between 5 October 1990 and 2 July 1991.
127.    Document PLEN 30/91 is the record of decisions of the 56th Plenary Meeting of the ICPR held on 2 July 1991. Item 8, on page 11, deals with the “Chlorides” Convention, and item 8.2 refers to Annex 4 of the record of decisions. The details concerning the financial arrangements for the alternative plan can be found in this annex. Chapter III of Annex 4 is especially important for the purposes of the present dispute, and deserves to be reproduced in full:
III. Methods for settling the accounts in 1998
At the end of the storage period (31.12.1998), France shall submit, under the ICPR, a report to the other Contracting Parties showing the sums to be reimbursed, if any.
Reimbursement will occur in the event that the sum of the expenditures made in France, the financial obligations undertaken for the removal from storage, and the Swiss contribution [reference omitted] to be taken into account, comes to less than 400 MF. It must be calculated as follows:
1. For the amounts corresponding to storage expenditures, any reimbursement shall correspond to the financing surplus, including amounts carried over from previous years, of the last year of storage (1998), including 11/12 of the annual interest generated by the surplus.
2. For the calculation of any amounts to be reimbursed that correspond to expenditures for removal from storage by or on behalf of France, the following method shall be used:
At the end of each year, the share of the payments by the parties that corresponds to the financing of the removal from storage of that year and the previous years shall be carried over to the following year, including interest on that amount for the year in question. For the calculation, the rate of interest to be applied is the actual rate of interest corresponding to the difference between the long-term national loan rate and the TP01 index. A final amount is thus arrived at for 31 December 1998, from which the cost of French obligations for removal from storage, calculated by multiplying the quantities stored since 1991 by 37 French francs/ton (1988 francs adjusted according to the TP01 index), will be subtracted. Any difference, if it is a positive number, will be the amount to be reimbursed by France.
Having regard to the risk to the French government of a rise in the costs of removal from storage, linked notably to the costs of the probable transfer of the operations of removal from storage to an operating company following the closure of the Alsace Potassium Mines, to the need for ensuring that the storage areas remain watertight, and to the implementation of various new environmental restrictions, there is no need to deduct non-accrued interest from the costs of removal from storage for the current period starting in 1999, as it will be compensated for by the cost overruns mentioned above. [Emphasis added] Chapter III of Annex 4 of document PLEN 30/91 seems to establish a relationship between the possible reimbursement and the calculation of the carry-overs (which is to be based on the amount of 61.5 francs per ton). First of all, for the amounts corresponding to storage expenses, the reimbursement is equal to “the financing surplus, including amounts carried over from previous years, of the last year of storage (1998)”. This surplus is to be calculated, as provided in Chapters I and II of that same Annex, on the basis of the stipulated amount of 61.5 French francs per ton (1988 francs). Thus, Chapter III of Annex 4 of document PLEN 30/91 provides that the calculation of any reimbursement, for the amounts corresponding to storage expenses, rests on an operation involving the amount of 61.5 French francs per ton (1988 francs). This runs counter to the argument whereby the amount of 61.5 francs per ton is not relevant to the final auditing.
More clearly still, the same Chapter III provides that the calculation of the surplus, for the share of the parties’ payments corresponding to financing the removal from storage, must be done by subtracting from the final surplus as at 31 December 1998 “the cost of French obligations for removal from storage, calculated by multiplying the quantities stored since 1991 by 37 French francs/ton (1988 francs adjusted according to the TP01 index).” Thus, what was to be taken into account when calculating the reimbursement of sums corresponding to the costs of removal from storage was incontestably the amount stipulated of 37 French francs per ton (1988 francs), and not the actual costs of restoring the areas to their former condition. Furthermore, the figure to be used in calculating the reimbursement for the amounts corresponding to storage expenses would, in the circumstances, inevitably be 24 French francs per ton.
The Tribunal raises one final point. In Annex 4 mentioned above, the Parties dealt in detail with the question of the risk of a rise in the costs of removal from storage (see Chapter III, paragraph 2, third subparagraph). They anticipated that this risk of an increase in costs (arising out of the costs of a probable transfer to an operating company, of ensuring the storage areas remained watertight and of implementing various new environmental restrictions) would be compensated by the interest that had not been running since 1999. They did not, on the other hand, anticipate any risk that costs would increase because of a reduction in the quantities actually stored.
For all these reasons, the Tribunal takes the view that Chapter III of Annex 4 of document PLEN 30/91 gives a clear indication that the Parties intended that the final auditing of the accounts be done on the basis of the stated amount of 61.5 French francs per ton (1988 francs). This conclusion is the more inevitable since document PLEN 30/91 is the preparatory document closest in time to the date of signature of the Protocol (which took place two months later, on 25 September 1991) and is therefore likely to be more reliable as a source of information regarding the commitments undertaken by the Parties in the Protocol.
128. To summarise, this discussion of the preparatory works has led to the following findings:
1)    In 1987 and 1988, the Alsace Potassium Mines had conducted studies in order to be able to store 16.3 million tons of NaCl over ten years (see paragraph 109 above). Those studies are of limited interest in the present case as far as storage costs are concerned. They set out the total cost of expenditures on removal from storage at 37.5 francs per ton of NaCl (33 francs of which was for running costs and 4 francs for investment costs).
2)    In 1989, the Alsace Potassium Mines prepared a study based on the alternative plan presented by the Netherlands in agreement with France (see paragraphs 110 and 111 above). It has not been possible for the Tribunal to determine precisely what storage quantities were used as the basis of that study. The study puts forward a figure of 24 French francs per ton of NaCl for storage (consisting partly of fixed and partly of variable elements). As to removal from storage, and in spite of the substantial reduction in the quantities involved, it referred to earlier studies and kept to the figure of 37.5 francs per ton. The total was 61.5 French francs per ton.
3)    In 1990, a study was carried out within the ICPR on the coefficient of flow of the Rhine. The resulting forecasts were such that it appeared that there would be limited variation in the costs of storage per ton (see paragraph 112 above). Rather than have this cost vary according to the quantities to be stored, it was suggested that an average hypothesis should be used in calculating the costs, which was that of 5.1 million tons of residual salts to be stored over 8.5 years. The cost per ton remained the one put forward by the Alsace Potassium Mines, namely 61.5 francs (see paragraph 114 above).
4)    These proposals were reflected in the various documents that followed, but, given the delay in getting the project off the ground, the decision was finally taken to calculate the expenditures on the basis of a quantity of 4.5 million tons of NaCl to be stored over 7.5 years. The cost per unit of 61.5 French francs per ton was thus not changed (see paragraph 125 above).
129.    The Tribunal finds, in these circumstances, that this amount, as France maintains, was indeed calculated on the basis of both fixed and variable elements. The Tribunal has not, however, been able to determine what proportion was finally allocated to each of these elements.
At a more fundamental level, the Tribunal finds that the figure of 37.5 French francs per ton for the removal from storage has been maintained substantially unchanged since 1988, while the quantities to be removed from storage fell from 16.3 million tons of NaCl to 5.1 million, and then to 4.5 million. The overall figure of 61.5 French francs per ton, adopted in 1990, also remained unchanged when the quantities to be stored were last reduced.
The Parties thus adopted the figure of 61.5 French francs per ton in the knowledge that it comprised both fixed costs and variable costs. They took the view, nonetheless, that, given the forecasts they had made of the coefficient of flow of the Rhine, there was no need when deciding on the cost per ton to take account of the variations that might occur in the quantities to be stored. Using those forecasts, they adopted the figure of 61.5 francs per ton as a lump sum figure, irrespective of those quantities, as the record of decisions of the ICPR of 2 July 1991 confirms (see paragraph 127 above).
In sum, the overriding concern of the Parties throughout this negotiation seems to have been to ensure that their forecasts of the quantities to be stored or the expenditures to be made were not exceeded as a result of the reduction in the coefficient of flow of the Rhine. They did not for one moment imagine that the coefficient of flow would increase and the resulting storage be so reduced as to throw their cost forecasts into disarray.
130.    Lastly, the Tribunal recalls that the phrase “payments are not payments in full discharge”, appearing in point 4.2.1 of Annex III to the Protocol, seems to be devoid of meaning if taken literally (see paragraph 90 above). The preparatory works contain some clues that might explain the presence of this phrase in point 4.2.1. An examination of these documents reveals that the parties had, at one time, contemplated making a forecast for expenditures (and, consequently, for payments) in the order of 355 million (1988) French francs, with a spending limit of 400 million (1988) French francs (see foot of page 5 of document C 2/90; p. 4 of document C 8/90-Nl and p. 6 of document DELch 4/90). Operating on such a hypothesis, the phrase “payments are not payments in full discharge” takes on some meaning, as it would have been possible for the total payments to be less than the total costs. However, the final text of the Protocol provides that the amount of payments shall be 400 million French francs; the spending limit is also fixed at 400 million French francs. Consequently, the Parties can never have obligations that go beyond what they have paid, and the phrase “payments are not payments in full discharge” was probably left in point 4.2.1 of Annex III to the Protocol by mistake.
131.    For all these reasons, the Tribunal concludes that the preparatory documents in its possession do not lead to a different conclusion from the one it has already reached by applying the general rule of interpretation in accordance with Article 31 of the Vienna Convention (paragraph 106 above).
132.    Following this analysis, the Tribunal will now turn, in the next paragraphs, to the calculation of the amounts owed by France to the Netherlands.
Chapter VI – Calculation of the Amounts to be Reimbursed
133.    France and the Netherlands have each proposed a method for calculation of the final auditing. It follows from the reasoning set out in paragraphs 80 to 132 that the Tribunal is unable to accept the method proposed by France, which is based on the actual costs of the operations of storage and removal from storage. Moreover, while the Tribunal considers, like the Netherlands, that the auditing calculation must be based on the lump sum amount of 61.5 French francs per ton, it cannot wholly subscribe to the method recommended by the Netherlands for calculating the sums to be reimbursed.
134.    According to the Netherlands, the operating expenses must be calculated by multiplying the quantities of chlorides stored, namely 960,000 tons (after rounding up), by the amount of 61.5 French francs. This gives an amount of 59.04 million French francs, to which must be added the 40 million French francs representing the investment expenses. This would give a total of expenditures of 99.04 million French francs, of which 34%, or 33.67 million, are to be borne by the Netherlands. Since the Netherlands has paid the sum of 136 million French francs to France, this would mean that France is obliged to reimburse the sum of 102.33 million French francs. In addition, France is obliged to pay interest on the annual carry-overs, on the surplus received from 1 January 1999 onwards, and on the amounts set aside for removal from storage. Lastly, according to the Netherlands, account must be taken of inflation.
135.    In the view of the Tribunal, this method is incomplete, if not erroneous. It takes no account of the fact that the expenditures – both for operation and investment – were to be calculated in 1988 French francs (point 1.2.6 of Annex III), while the payments were made in francs at current value. Furthermore, while it claims to be entitled to interest on the amounts carried over each year and the credit balance as at 1 January 1999, the Netherlands offers no indications or explanations as to how that interest should be calculated. It would seem that this calculation depends on various factors, one of which, with regard to the interest on the annual carry-overs, is the amount of the annual payments made by the Netherlands and of the carry-overs themselves. A further question arises in respect of the calculation of interest after 31 December 1998, namely whether interest must be calculated on the interest capitalised at that date and whether the amounts needed for removal from storage bear interest. Finally, it is clear that interest must be calculated as a function of the interest rate applicable to each year.
136.    Taking the above considerations into account, the Tribunal proposes to perform the auditing calculation in two stages. In the first stage, the Tribunal will calculate the surplus, in other words the amount of the excess paid by the Netherlands to France. In the second stage, the Tribunal will calculate the interest, drawing a distinction between interest due up to 31 December 1998 on the one hand, and interest after that date on the other. The Tribunal will proceed year by year, both for the calculation of the surplus and of the interest. Indeed, it is necessary to determine the amount of the surplus for each year in order to be able to then calculate the interest accrued on that amount as prescribed in point 3.2.3, and likewise, it is necessary to take into account the fact that the interest from previous years is added, each time, to the amount carried over on which interest must be calculated for the following year. Finally, the long-term annual rate of interest of the Credit National, which the Tribunal must apply by virtue of points 3.2.3 and 4.2.1, is different every year.
137.    In order to calculate the annual surplus, the first task is to determine the amounts actually expended by France, then calculate 34% of this amount, representing the
contribution of the Netherlands under Article 4 of the Protocol, and finally compare the resulting amounts with the payments made each year by the Netherlands. In the paragraphs that follow, the Tribunal will determine and value the data that is necessary for this calculation.
The Tribunal calculates the operating expenditures by multiplying the quantities stored annually by France by the amount of of 61.5 French francs, in accordance with the Tribunal’s interpretation of point 4.2.1. In order to determine the quantities stored, which, as can be seen from the introductory paragraph of Annex I to the Protocol, are quantities of residual salts, the Tribunal relies on the figures appearing in the annual reports submitted by France and the table relating to the “Updated actual expenditures for the final auditing” which appears in the document entitled “Bonn Convention – actual expenditures calculated in F 88”, that France submitted during the hearing. According to that table, the total quantity of salts stored was 960,199 tons, a figure with which both Parties agree. While the quantities of salts recorded in the annual reports and the above-mentioned table for the years 1992, 1993, 1997 and 1998 are identical, namely 177,917 tons for the year 1991-1992, 102,455 tons for the year 1993, 310,590 tons for the year 1997 and 98,799 tons for the year 1998, with respect to the year 1995, the Tribunal will use the quantities of residual salts as shown in the table, i.e. 270,438 tons, since the annual report for that year accounts only for the quantities of NaCl. For the years 1994 and 1996, the table does not mention any stored salts, while the annual reports show a storage of 8,300 and 8,500 tons of NaCl. The provisional balance sheet as at the end of 1996 indicates however that this storage was not required under the Bonn Convention, but by orders of the local prefecture, which explains why France made no mention of it in the table it submitted on 3 October 2003. The Tribunal notes that according to the annual reports for those two years, this storage was nonetheless useful for the consolidation of the platforms, and therefore considers it justified to take it into account. Noting that, as a general rule, one ton of NaCl corresponds to 1.31 tons of residual salts, the Tribunal will use the figures of 10,873 and 11,135 tons of residual salts as the amounts for 1994 and 1996.
Lastly, inflation must be taken into consideration when calculating the operating expenditures. Indeed, while point 3.2.3 is silent in this regard, the annual reports prepared by France take inflation into account. The Netherlands did not object to this being taken into consideration, and has expressly agreed to take it into account (see Reply, paragraph 4.3). The Tribunal has thus referred to the revaluation coefficients that appear in the annual reports and in the table submitted by France, and which are not disputed by the Netherlands. For the years 1991/1992 to 1998, these coefficients are respectively 1.1319%, 1.1556%, 1.1847%, 1.2096%, 1.2290%, 1.2500% and 1.2536%.
Taking into account these observations, the figures for the annual operating expenditures are as follows:
1991/1992:
12,385,132 francs 7,281,415 francs 792,196 francs 20,117,991 francs 841,622 francs 23,876,606 francs 7,617,047 francs
1993
1994
1995
1996
1997
1998

They amount to a total of 72,912,009 French francs.
In accordance with point 1.2.6, for the first year, the investment expenditures of 40 million 1988 French francs, revalued to reflect inflation, must be added to the operating expenditures. However, the Tribunal finds that according to the annual reports, 32,458,273 French francs were dedicated to investments in 1992 and 12,732,044 French francs were invested in 1993, a total of 45,190,317 French francs. In other words, 72% of the expenditures on investment were made in 1992, while 28% were made in 1993 (after rounding up the percentages). Thus, the next step is to revalue the lump sum amount of 40 million 1988 French francs by applying the 1992 price rise coefficient (1.1319%) to 72% of the said amount, or 28,800,000 French francs. Likewise, the price rise coefficient for 1993 (1.1556%) must be applied to 28% of the same lump sum amount, or 11,200,000 French francs. This means that the investment expenditures to be taken into account amount to 32,598,720 French francs for the year 1991/1992 and 12,942,720 French francs for 1993.
Since these amounts must be allocated to the years 1991/1992 and 1993, they should be added to the operating expenditures relating to those two years, which are 12,385,132 and 7,281,415 French francs, respectively. The total expenditures for the year 1991/1992 thus amount to 44,983,852 French francs, while the total expenditures for the year 1993 amount to 20,224,135 French francs.
As to the yearly payments made by the Netherlands, the Tribunal has used the amounts that appear in Annex 4 to the Memorial of the Netherlands, which France does not dispute. According to that annex, the Netherlands made the following payments:
24 December 1991:
26    February 1992: 5 March 1993:
3 February 1994: 2 February 1995: 21 August 1996: 19 January 1998:
27    November 1998:
30,600,000 francs 12,920,000 francs 9,180,000 francs 24,820,000 francs 12,240,000 francs 12,240,000 francs 16,850,000 francs 16,850,000 francs

The Tribunal notes that the Netherlands did not make a payment in 1997, while two payments were made in 1998. Both these payments were attributed to 1998. Furthermore, the payments made in 1991 and 1992 were added for the purposes of the calculation. The total amount paid by the Netherlands to France was 135,700,000 French francs.
Based on these calculations, the difference between the annual payment made by the Netherlands and the amount that results when the French annual expenditures are multiplied by the 34% contribution of the Netherlands, or in other words, the excess payment made each year by the Netherlands, is as follows:
1991/1992:    28,225,490 francs
2,303,794 francs 24,550,653 francs 5,399,883 francs 11,953,849 francs – 8,118,046 francs
1993
1994
1995
1996
1997

31,110,204 francs 95,425,827 francs
1998: Total:

138.    The Tribunal will now proceed to calculate the interest. The Tribunal will start with the calculation of interest due up to 31 December 1998, and will then determine the interest for the period after that date. As noted above, the interest accrued over these two periods is calculated in different ways.
139.    The interest due up to 31 December 1998 must be calculated pursuant to, and in accordance with, point 3.2.3 of Annex III of the Protocol. This point sets out that: “In the event that the actual storage costs calculated in this way are below the limit initially set for the year concerned (point 2.1.1), the difference (plus 11/12 of the yearly interest on this amount at the annual long-term interest rate on national loans) shall be carried over to the following year, thereby raising the spending limit for the following year accordingly”.
By referring to point 2.1.1 of Annex III, this provision invokes the annual spending limits, which, under the Protocol, were to correspond exactly with the payments to be made. However, these payments were not made in accordance with point 2. Since it was clearly not the intention of the parties to the Protocol that amounts that were not paid should bear interest, it follows, in the Tribunal’s opinion, that the text of point 3.2.3 must be applied by calculating the interest on the actual annual balances resulting from the above calculation of the surpluses.
In addition, since point 3.2.3 provides that interest thus accrued prior to 31 December 1998 shall be capitalised, it follows that, aside from the balance for the year 1991/1992, which is equal to the amount of the surplus for that year, i.e. 28,225,490 French francs, the sums carried over are the sum not only of the balance of the previous year and the current year, but also of the interest on the previous year’s balance.
According to point 3.2.3, interest shall be paid on 11/12 of the year. This provision is explained, in the Tribunal’s view, by the fact that, under point 2.1.2, the Parties were required to pay their contribution “by a lump-sum annual advance payment”. Since these payments had to be made not later than 31 January of each year (see point 3.2.1), it was normal to stipulate that the amounts in question should bear interest for 11 months out of 12. Similarly, to the extent that the said payments were not “advance” payments, a different coefficient was to be applied to them. Based on the information provided by the Netherlands in Annex 4 to its Memorial, the Netherlands made the payments as envisaged from 1991 to 1996. However, it must be noted that, for the years 1992, 1993 and 1996, there were delays in payment. The payment for the year 1992 was made on 26 February, followed by the 1993 payment on 5 March, and finally, the 1996 payment was made a semester late, on 21 August. Consequently, the Tribunal considers it justified to apply a coefficient of 10/12 for 1992, a coefficient of 9/12 for 1993, and a coefficient of 4/12 for 1996.
Again based on Annex 4 of its Memorial, the Netherlands stopped making payments in 1997 and made two payments in 1998. At the beginning of 1997, France had produced its annual report for 1996, and an overall provisional balance sheet, the provisional balance sheet as at the end of 1996. It emerged from this balance sheet that “the hydrological conditions and a general reduction in chloride concentrations in the Rhine meant that the conditions for storage to be carried out were very rarely fulfilled”. The balance sheet noted, more specifically, that “over a period of five full years, the actual total storage (436,439 tons) is less
than one year according to the forecasts (600,000 tons)”. It therefore became apparent that any further payments to France for the implementation of the Protocol would be futile. Switzerland and Germany likewise stopped all payments with effect from 1995 and 1996 respectively, and France, it appears, quite naturally accepted this. As noted above, the Netherlands halted all payments in 1997, but had a change of attitude in 1998 and on 16 January it paid its contribution of 16.85 million francs for 1997, which was followed by its 1998 contribution, of the same amount, on 27 November 1998. These payments, made late and during the last year the Protocol was in effect, are the source of some confusion. They could be regarded as intending to regularise the situation created by the delay in payment in 1996 and the fact that no payment at all was made in 1997. However, by January 1998 and, a fortiori, by November 1998, the Netherlands must have known that the amounts in question were unlikely to be used in the fight against chloride pollution of the Rhine pursuant to the Protocol. These amounts would necessarily have to be reimbursed to the Netherlands, after bearing interest at the long-term interest rate specified in Annex III. This was in effect an investment. France could have, and perhaps should have, refused this investment. However, at the time, it was contemplating an extension of the period of implementation of the Protocol, a proposal that its partners rejected. From the French perspective, the payments by the Netherlands could be justified. The fact remains, however, that these payments could not have been intended to provide France with the advances it required for the depollution of the Rhine. The payments were unreasonable. They must certainly be reimbursed to the Netherlands, but given the circumstances of the case, the Tribunal does not consider that amounts paid in this way by the Netherlands should bear interest.
Finally, the rate of interest to be applied, according to point 3.2.3, is the long-term annual rate of the Credit National. Since the Credit National long-term interest rate was no longer in existence for the years 1997 and 1998, the Tribunal, after consulting with the Parties, applied the interest rates on French 10-year State bonds, which were provided by France and not disputed by the Netherlands, namely 5.56% for 1997 and 4.65% for 1998.
In the light of the foregoing considerations, the interest and the amounts carried over are as follows:
28,225,490 francs 32,521,533 francs 59,499,950 francs 69,399,517 francs 86,601,704 francs 82,509,755 francs 117,136,937 francs
interest for previous year    balance carried over
1991/1992:
1993:    1,992,249 francs
1994:    2,427,764 francs
1995:    4,499,684 francs
1996:    5,248,338 francs
1997:    4,026,097 francs
1998:    3,516,978 francs
As at 31 December 1998, the total balance carried forward was thus 117,136,937 French francs, and the capitalised interest was 21,711,110 French francs.
140. On the basis of the amounts thus determined, the Tribunal will now proceed to calculate the interest due after 31 December 1998, as required by point 4.2.1 of Annex III to the Protocol.
According to the Tribunal, a distinction should be made here between expenditures for storage and those for removal from storage. Document PLEN 30/91, which the Tribunal examined in paragraph 47, and more specifically Chapter III of Annex 4 to that document, in fact states, with regard to removal from storage, that “on 31 December 1998 the cost of French obligations for removal from storage [shall be] calculated by multiplying the quantities stored by 37 French francs/ton (1988 francs adjusted as a function of the TPOI index)”. It adds: “Having regard to the risk to the French government of a rise in the costs of removal from storage, linked notably to the costs of the probable transfer of the operations of removal from storage to an operating company following the closure of the Alsace Potassium Mines, to the need of ensuring that the storage areas remain watertight, and to the implementation of various new environmental restrictions, there is no need to deduct non- accrued interest from the costs of removal from storage for the current period starting in 1999, as it will be compensated for by the cost overruns mentioned above.” Pursuant to that provision, since the quantities stored amounted to 982,207 tons (see paragraph 137 above) and the price rise coefficient for the year 1998 was 1.2536%, the costs of removal from storage on 31 December 1998 came to 45,557,904 million French francs. The share of the Netherlands amounts to 15,489,687 French francs, in other words 34% of these costs. Pursuant to the provision, and contrary to what is claimed by the Netherlands, that amount should not bear interest for the current period starting in 1999.
Furthermore, and as the Tribunal pointed out above, the amount of 31,110,204 francs, being the amount resulting from the payments made by the Netherlands in 1998, should not bear interest from 31 December 1998 either.
Therefore, in order to calculate future interest, the above-mentioned amounts of 15,489,687 French francs and 31,110,204 French francs must be deducted from the balance at 31 December 1998, namely 117,136,937 French francs. This gives a balance at 31 December 1998 of 70,537,046 French francs.
The Tribunal notes, moreover, that in its Reply (paragraph 4.3), the Netherlands claims the payment of interest only on the capital that was due at that date. On the other hand, it does not claim interest after 31 December 1998, on the interest accrued at that date. Since the Tribunal’s decision may not be ultrapetita, in calculating the interest after 31 December 1998 it must also subtract from the balance at 31 December 1998 the interest due by that date. That interest, as has been seen, amounted to 21,711,110 French francs. Consequently, the balance at 31 December 1998, for the purposes of calculating interest after that date, was 48,825,936 French francs.
In accordance with point 4.2.1 of Annex III to the Protocol, and in the same way that applies to the calculation of interest due before 31 December 1998, this amount shall bear interest at the rate of interest on French 10-year State bonds, since the long-term Credit National interest rate no longer exists. The French State bond rates, provided by France and not disputed by the Netherlands, were 4.61% for 1999, 5.42% for 2000, 4.96% for 2001, 4.88% for 2002 and lastly 4.13% for 2003. The Tribunal would again point out that, bearing in mind the requirements of point 4.2.1 and the fact that the Netherlands has not submitted that compound interest should be paid (see Reply, paragraph 4.3), there is no need to capitalise the interest.
The interest due up to the end of the year 2003 therefore amounts to 2,250,876 French francs for 1999, 2,646,366 French francs for 2000, 2,421,766 French francs for 2001, 2,382,706 French francs for 2002, and 2,016,511 French francs for 2003. In total, this amounts to 11,718,225 French francs.
141.    These calculations dictate that the surplus that France is obliged to reimburse to the Netherlands is 128,855,162 French francs, being the total amount carried forward as at 31 December 1998, the sum of the amount of 117,136,937 French francs and the interest of 11,718,225 French francs.
142.    However, pursuant to the provisions of Chapter III of Annex 4 of document PLEN 30/91, the cost overruns on removal from storage resulting from the rise in prices after 1998 must be deducted from that amount. In this connection, while it has been possible for the Tribunal to establish the cost of removal from storage as at 31 December 1998, namely 45,557,904 French francs, it finds that it has no means of determining the cost overruns resulting from the rise in prices after 1998, since these depend on the rate at which product is removed from storage and the rate of inflation. It will be up to the Parties to determine the amount of these cost overruns. Nonetheless, the Tribunal considers that it is appropriate at the present time to set aside an amount to finance these cost overruns. This amount, which the Tribunal sets at 10 million French francs, will be deducted from the amount that France is bound to transfer to the Netherlands.
143.    To conclude, France’s obligation of payment towards the Netherlands is at present in the amount of 128,855,162 French francs, or 19,643,843 Euros (1 Euro = 6.55957 French francs). At this time, France is obliged to pay the Netherlands the amount of 118,855,162 French francs, or 18,119,353 Euros. It must also set aside the sum of 10 million French francs, or 1,524,490 Euros, to cover the rise in prices of removal from storage after 1998. Finally, when the Parties have determined what these cost overruns were, France must deduct them from the amount set aside for that purpose, and transfer the remaining balance to the Netherlands, with no interest payable.
Chapter VII – Dispositif
For these reasons,
the TRIBUNAL unanimously decides that:
1.    The amount of funds that must be transferred for the auditing of the accounts between France and the Netherlands as provided in point 4.2.1 of Annex III to the Protocol is 128,855,162 French francs, or 19,643,843 Euros, subject to the deduction from that sum of the amount of the cost overruns resulting from the rise in the price of removal from storage after 1998;
2.    To enable it to carry these cost overruns, France shall keep in reserve the sum of 10 million French francs, or 1,524,490 Euros, which sum shall not bear interest;
3.    France is bound to reimburse immediately to the Netherlands the sum of 118,855,162 French francs, or 18,119,353 Euros;
4.    As soon as the removals from storage are completed, France and the Netherlands shall determine the cost overruns resulting from the rise in the price of removal from storage after 1998;
5.    After deducting the amounts of the cost overruns referred to in point 4 of this dispositif, France shall be obliged to repay to the Netherlands the balance of the sum reserved pursuant to point 2 of this dispositif.
DONE in French at the Peace Palace, The Hague, on 12 March two thousand and four, in three originals, one of which shall be deposited in the archives of the International Bureau of the PCA and the others transmitted to the Government of the Kingdom of the Netherlands and the Government of the French Republic respectively.
(signed)
Krzysztof Skubiszewski President
(signed)    (signed)
Pieter H. Kooijmans    Gilbert Guillaume
(signed)
Anne Wallemacq Registrar
Mr. Gilbert Guillaume appends a Declaration to this Award.
Declaration of Mr. Gilbert Guillaume
The Parties are in disagreement as to the methods to be used in the final auditing of accounts for the storage and removal from storage of salts on French territory pursuant to the Additional Protocol to the Convention on the Protection of the Rhine against Pollution by Chlorides. The Netherlands takes the view that, independently of the initial investments, operating expenditures must be calculated in accordance with the combined provisions of points 4.2.1 and 1.2.6 of Annex III to the Protocol, namely on the basis of 61.5 French francs per ton of residual salts stored. France contends that this calculation should be done on the basis of the actual cost per unit.
The essence of the French argument relies on Article 4 of the Protocol, which provides, inter alia, that:
The cost of the measures taken on French territory under articles 1 and 2, up to a
maximum of 400 million French francs … shall be apportioned as follows:
Federal Republic of Germany    30%
French Republic    30%
Kingdom of the Netherlands    34%
Swiss Confederation    6%
According to France, the application of the provisions invoked by the Netherlands would result in France having to bear 75% of the total expenditures, which would be contrary to Article 4 of the Protocol.
The Tribunal has rightly observed that Article 4 does no more than lay down a formula for allocating the expenditures, and refers back to Annex III for the methods by which these same expenditures are to be paid. In fact, with regard to the final auditing of the accounts, point 4.2.1 of Annex III itself refers back to point 1.2.6, which prescribes the figure of 61.5 francs per ton. This makes it difficult to reconcile the French argument with the text of the Protocol itself.
The calculation performed in accordance with the Protocol nevertheless produces a result that is, if not absurd, at the very least unreasonable. It leads to the total expenditures being calculated at 99.04 million 1988 francs (of which 33.67 million must be borne by the Netherlands). However, according to France, these expenditures amount to around 206 million francs. This figure is clearly excessive, as France did not carry out all the work originally contemplated. The fact remains, however, that even after making the corrections required by the 1996 balance sheet presented by France and the information it provided at the hearings and in its letter of 28 March 2003, the total expenditures would still be in the region of 160 million 1988 francs. If the Netherlands were to pay only 33.67 million, its contribution to these expenditures would be not 34% but 21%, and France would bear 57% of these expenditures, instead of the 30% set out in the Protocol. Such a result is unreasonable.
The abundant arbitral jurisprudence referred to by the Tribunal shows that, pursuant to customary law as codified in Article 32 of the Vienna Convention on the Law of Treaties, it is appropriate in such a situation to ask whether the result of a literal application of the text is not contrary to the will of the parties as revealed by the preparatory works.
The Arbitral Tribunal analysed these documents, and concluded that the sum of 61.5 francs per ton had indeed been calculated, as France maintains, on the basis of both fixed and variable elements. It took the view that it was not in a position to determine what share was finally attributed to each of these elements. Where the storage expenditures are concerned, I do not, for my part, see any reason to question the half-half allocation of these costs, as is done in document C 8/90 of 10 September 1990. This more or less corresponds, in fact, to the calculations appearing in the earlier documents. For removal from storage, on the other hand, I recognise that the calculation is a more difficult one. Moreover, those figures are liable to be revised because France has not carried out all the site works envisaged.
But that is not the real issue. The real question is whether it can be concluded from the preparatory works that the figure of 61.5 francs per ton was chosen by the drafters of the Protocol as a lump sum figure irrespective of the quantities stored.
The Tribunal considered that this was indeed the case, in the framework of the forecasts the Parties made regarding the coefficient of flow of the Rhine. I share its view on this point. In practice, the Parties’ forecasts would have led to the storage of a minimum of 3.1 million tons of salts and a maximum of 6.9 million. However, they ruled out these extreme forecasts and took the view that variations in the coefficient of flow might be such as to cause storage to vary between 4.1 and 5.6 million tons. Furthermore, they calculated the cost per unit on the basis of 5.1 million tons to be stored. It is clear from the preparatory works that this figure was a lump sum figure, irrespective of the quantities stored, as long as those quantities remained within the forecast limits.
In reality, the overriding concern of the Parties – and France in particular – during the negotiations was to avoid any additional expenditures in the event the coefficient of flow of the Rhine fell below the forecast levels. These concerns are expressed in the provisions of points 1.2.1 and 1.2.6 of Annex III which set an absolute spending limit of 400 million 1988 French francs and determined the spending limits for each year, above which France was released from all storage obligations. At no time did the negotiators envisage that the tonnage to be stored would be so low as to call their cost calculations into question.
In these circumstances, I do not think that any common intention of the Parties can be read into the preparatory works that goes against the text of the Protocol itself and allows the application of the text to be avoided, however unreasonable it may be.
The fundamental difficulty of this case arises from the fact that the forecasts made in 1989 about the coefficient of flow of the Rhine turned out to be completely wrong. The question then becomes whether this is a fundamental change of circumstances such as to permit the Tribunal of its own motion to invoke the “rebus sic stantibus” clause.
In this regard, it should be recalled that, according to Article 62(1) of the Vienna Convention on the Law of Treaties:
b) the effect of the change is radically to transform the extent of obligations still to be performed under the treaty.
The International Court of Justice has recognised that these provisions have a customary character in its Judgments of 2 February 1973 in the Fisheries Jurisdiction (Federal Republic of Germany v. Iceland) case, I.C.J. Reports 1973, p. 63, para. 36, and of 25 September 1997 in the Gabcikovo – Nagymaros (Hungary/Slovakia) case, I.C.J. Reports 1997, p. 64, para. 104. The Iran-United States Claims Tribunal even viewed this as a general principle of law (Questech Inc. v. The Ministry of National Defence of the Islamic Republic of Iran, Case No. 59, 20 September 1985, Award No. 191-59-1, Iran-United States Claims Tribunal Reports, Vol. 9 -1987, p. 122).
Those provisions are therefore applicable in the present case, notwithstanding the fact that France is not a party to that Convention. It is my view that in this case the conditions laid down in Article 62(1) are fulfilled. It seems to me that the coefficient of flow of the Rhine was fundamentally different during the years 1991 to 1998 from what it had been in previous years. This resulted in the quantity of salts stored being a fifth of the quantity that had served as the basis for the cost calculations. France would certainly never have agreed to costs being fixed at a lump sum of 61.5 French francs per ton if it had known that this would be the case. This was an essential basis of its consent, as indeed it was for the other Parties. The change that came about had the effect of radically transforming the obligations that remained to be performed. This being the case, the conditions in Article 62(1) are satisfied (the exceptions in Article 62(2) not being applicable). In the case of the Gabcikovo-Nagymaros dam, the International Court of Justice made it clear that “the estimated profitability of the Project might have appeared less in 1992 than in 1997”, but it added that the profitability was not “bound to diminish to such an extent that the treaty obligations of the parties would have been radically transformed as a result” (I.C.J. Reports 1997, p. 64, para. 104). The situation in the present case is different.
What are the consequences to be drawn from such a situation? In domestic law, when the theory of fundamental change of circumstances is accepted, it can lead either to the termination of the contract,  or to revision by the adjudicator (see for the latter case the application of the theory of “imprevision” in French administrative law;  in German law, see Article 242 of the Civil Code and the theory of the “Wegfall der Geschaftsgrundlage'”;’ in Swiss law, see the jurisprudence of the Federal Court).  This latter solution has sometimes
been considered, but never adopted, in the law of international contracts falling under the “lex mercatoria” (see the International Chamber of Commerce arbitral award in Case No. 2508, Reports of Arbitral Awards of the International Chamber of Commerce, 1976, p. 939).   On the other hand, it has never been applied in public international law. In relations between States, the “rebus sic stantibus” clause implies that where there is a fundamental change of circumstances, the parties have a duty to negotiate, and then, if negotiations do not bear fruit, they may be allowed to withdraw from the convention or suspend its implementation (Vienna Convention on the Law of Treaties, Article 62(3)).  Revision by the judge or arbitrator is permitted only where the Parties have granted such a power. That is not the case here.
This situation seems to me to be thoroughly regrettable, insofar as it has led the Tribunal to adopt an unreasonable solution that should have been reviewed in the light of the new circumstances by good faith negotiations between the Parties. One might have hoped that, in the relations between two friendly neighbouring countries, such negotiations would have been possible. It is a matter of regret that no such negotiations took place, and one may only hope that this process will be used to find a reasonable solution to the problems that remain outstanding between the Parties, as well as any that might arise between France on the one hand and Germany or Switzerland on the other.
(signed) Gilbert GUILLAUME
4 Translator’s note: for cases cited in this Chapter, the authentic English text is used where one exists. In other cases, an unofficial translation is given. Where the translation is unofficial, the reference provided is to the original French source.
1. A fundamental change of circumstances which has occurred with regard to those existing at the time of the conclusion of a treaty, and which was not foreseen by the parties, may not be invoked as a ground for terminating or withdrawing from the treaty unless:
a) the existence of those circumstances constituted an essential basis of the consent of the parties to be bound by the treaty; and
3    Pursuant to this theory based on good faith, German jurisprudence considers that the creditor may no longer require the obligation to be fulfilled when it has become, in equity, intolerable for the debtor. This criterion leaves a wide margin of discretion to the judge.